Fipy-3 1 3

𝜙
𝜋
FiPy Manual
Release 3.1.3
Jonathan E. Guyer
Daniel Wheeler
James A. Warren
Materials Science and Engineering Division

and the Center for Theoretical and Computational Materials Science
Material Measurement Laboratory
January 17, 2017

This software was developed at the National Institute of Standards and Technology by employees of the Federal
Government in the course of their official duties. Pursuant to title 17 section 105 of the United States Code this
software is not subject to copyright protection and is in the public domain. FiPy is an experimental system. NIST
assumes no responsibility whatsoever for its use by other parties, and makes no guarantees, expressed or implied,
about its quality, reliability, or any other characteristic. We would appreciate acknowledgement if the software is used.
This software can be redistributed and/or modified freely provided that any derivative works bear some notice that
they are derived from it, and any modified versions bear some notice that they have been modified.
Certain commercial firms and trade names are identified in this document in order to specify the installation and usage
procedures adequately. Such identification is not intended to imply recommendation or endorsement by the National
Institute of Standards and Technology, nor is it intended to imply that related products are necessarily the best available
for the purpose.
Contents
I Introduction 1
1 Overview 3
1.1 Even if you don’t read manuals... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 What’s new in version 3.1.3? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Download and Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Conventions and Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Installation 7
2.1 Recommended Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Installing Python Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Obtaining FiPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Installing FiPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5 Required Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.6 Optional Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.7 Level Set Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.8 Platform-Specific Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.9 Development Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Solvers 13
3.1 PySparse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 SciPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 PyAMG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.4 Trilinos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4 Viewers 17
4.1 Matplotlib . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Mayavi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5 Using FiPy 19
5.1 Testing FiPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.2 Command-line Flags and Environment Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.3 Solving in Parallel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4 Meshing with Gmsh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.5 Coupled and Vector Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.6 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.7 Running under Python 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.8 Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6 Theoretical and Numerical Background 29
i
6.1 General Conservation Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.2 Finite Volume Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.3 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.4 Linear Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.5 Numerical Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
7 Design and Implementation 37

7.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Programming Language39subsection.7.1.6
7.2 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
8 Frequently Asked Questions 41

8.1 How do I represent an equation in FiPy? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
8.2 How can I see what I’m doing? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
8.3 Iterations, timesteps, and sweeps? Oh, my! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
8.4 Why the distinction between CellVariable and FaceVariable coefficients? . . . . . . . . . . 46
8.5 How do I represent boundary conditions? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.6 What does this error message mean? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.7 How do I change FiPy’s default behavior? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.8 How can I tell if I’m running in parallel? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.9 Why don’t my scripts work anymore? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.10 What if my question isn’t answered here? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
9 Glossary 49
II Examples 51
10 Diffusion Examples 55
10.1 examples.diffusion.mesh1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
10.2 examples.diffusion.coupled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
10.3 examples.diffusion.mesh20x20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
10.4 examples.diffusion.circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
10.5 examples.diffusion.electrostatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
10.6 examples.diffusion.nthOrder.input4thOrder1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
10.7 examples.diffusion.anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
11 Convection Examples 89
11.1 examples.convection.exponential1D.mesh1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
11.2 examples.convection.exponential1DSource.mesh1D . . . . . . . . . . . . . . . . . . . . . . . . . . 90
11.3 examples.convection.robin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
11.4 examples.convection.source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
12 Phase Field Examples 95

12.1 examples.phase.simple . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
12.2 examples.phase.binaryCoupled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
12.3 examples.phase.quaternary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
12.4 examples.phase.anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
12.5 examples.phase.impingement.mesh40x1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
12.6 examples.phase.impingement.mesh20x20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
12.7 examples.phase.polyxtal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
12.8 examples.phase.polyxtalCoupled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
13 Level Set Examples 139

13.1 examples.levelSet.distanceFunction.mesh1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
ii
13.2 examples.levelSet.distanceFunction.circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
13.3 examples.levelSet.advection.mesh1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
13.4 examples.levelSet.advection.circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
13.5 Superconformal Electrodeposition Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
13.6 examples.levelSet.electroChem.simpleTrenchSystem . . . . . . . . . . . . . . . . . . . . . . . . . . 145
13.7 examples.levelSet.electroChem.gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
13.8 examples.levelSet.electroChem.leveler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
13.9 examples.levelSet.electroChem.howToWriteAScript . . . . . . . . . . . . . . . . . . . . . . . . . . 153
14 Cahn Hilliard Examples 161

14.1 examples.cahnHilliard.mesh2DCoupled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
14.2 examples.cahnHilliard.sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
15 Fluid Flow Examples 167

15.1 examples.flow.stokesCavity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
16 Reactive Wetting Examples 173

16.1 examples.reactiveWetting.liquidVapor1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
17 Updating FiPy 179

17.1 examples.updating.update2_0to3_0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
III fipy Package Documentation 189

18 How to Read the Modules Documentation 191
18.1 package.subpackage package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
19 fipy.boundaryConditions package 195

19.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
19.2 fipy.boundaryConditions.boundaryCondition module . . . . . . . . . . . . . . . . . . . . . . . . . . 195
19.3 fipy.boundaryConditions.constraint module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
19.4 fipy.boundaryConditions.fixedFlux module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
19.5 fipy.boundaryConditions.fixedValue module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
19.6 fipy.boundaryConditions.nthOrderBoundaryCondition module . . . . . . . . . . . . . . . . . . . . . 196
19.7 fipy.boundaryConditions.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
19.8 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
20 fipy.matrices package 199

20.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.2 fipy.matrices.offsetSparseMatrix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.3 fipy.matrices.pysparseMatrix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.4 fipy.matrices.scipyMatrix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.5 fipy.matrices.sparseMatrix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.6 fipy.matrices.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.7 fipy.matrices.trilinosMatrix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.8 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
21 fipy.meshes package 201

21.1 Subpackages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
21.2 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
21.3 fipy.meshes.abstractMesh module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
21.4 fipy.meshes.cylindricalGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
iii
21.5 fipy.meshes.cylindricalGrid2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
21.6 fipy.meshes.cylindricalNonUniformGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . 207
21.7 fipy.meshes.cylindricalNonUniformGrid2D module . . . . . . . . . . . . . . . . . . . . . . . . . . 208
21.8 fipy.meshes.cylindricalUniformGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
21.9 fipy.meshes.cylindricalUniformGrid2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
21.10 fipy.meshes.factoryMeshes module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
21.11 fipy.meshes.gmshMesh module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
21.12 fipy.meshes.grid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
21.15 fipy.meshes.mesh module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
21.16 fipy.meshes.mesh1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
21.17 fipy.meshes.mesh2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
21.18 fipy.meshes.nonUniformGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
21.21 fipy.meshes.periodicGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
21.22 fipy.meshes.periodicGrid2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
21.23 fipy.meshes.skewedGrid2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
21.24 fipy.meshes.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
21.25 fipy.meshes.tri2D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
21.26 fipy.meshes.uniformGrid module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
21.27 fipy.meshes.uniformGrid1D module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
21.30 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
22 fipy.solvers package 233

22.1 Subpackages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
22.2 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
22.3 fipy.solvers.pysparseMatrixSolver module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
22.4 fipy.solvers.solver module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
22.5 fipy.solvers.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
22.6 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
23 fipy.steppers package 247

23.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
23.2 fipy.steppers.pidStepper module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
23.3 fipy.steppers.pseudoRKQSStepper module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
23.4 fipy.steppers.stepper module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
23.5 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
24 fipy.terms package 251

24.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.2 fipy.terms.abstractBinaryTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.3 fipy.terms.abstractConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.4 fipy.terms.abstractDiffusionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.5 fipy.terms.abstractUpwindConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.6 fipy.terms.advectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
24.7 fipy.terms.asymmetricConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
24.8 fipy.terms.binaryTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
24.9 fipy.terms.cellTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
24.10 fipy.terms.centralDiffConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
24.11 fipy.terms.coupledBinaryTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
iv
24.12 fipy.terms.diffusionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
24.13 fipy.terms.diffusionTermCorrection module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
24.14 fipy.terms.diffusionTermNoCorrection module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
24.15 fipy.terms.explicitDiffusionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
24.16 fipy.terms.explicitSourceTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
24.17 fipy.terms.explicitUpwindConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
24.18 fipy.terms.exponentialConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
24.19 fipy.terms.faceTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
24.20 fipy.terms.firstOrderAdvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
24.21 fipy.terms.hybridConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
24.22 fipy.terms.implicitDiffusionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
24.23 fipy.terms.implicitSourceTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
24.24 fipy.terms.nonDiffusionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
24.25 fipy.terms.powerLawConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
24.26 fipy.terms.residualTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
24.27 fipy.terms.sourceTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
24.28 fipy.terms.term module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
24.29 fipy.terms.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
24.30 fipy.terms.transientTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
24.31 fipy.terms.unaryTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
24.32 fipy.terms.upwindConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
24.33 fipy.terms.vanLeerConvectionTerm module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
24.34 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
25 fipy.tests package 287

25.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
25.2 fipy.tests.doctestPlus module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
25.3 fipy.tests.lateImportTest module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
25.4 fipy.tests.testBase module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
25.5 fipy.tests.testClass module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
25.6 fipy.tests.testProgram module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
25.7 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
26 fipy.tools package 289

26.1 Subpackages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
26.2 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
26.3 fipy.tools.copy_script module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
26.4 fipy.tools.debug module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
26.5 fipy.tools.decorators module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
26.6 fipy.tools.dump module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
26.7 fipy.tools.inline module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
26.8 fipy.tools.numerix module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
26.9 fipy.tools.parser module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
26.10 fipy.tools.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
26.11 fipy.tools.vector module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
26.12 fipy.tools.vitals module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
26.13 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
27 fipy.variables package 327

27.1 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
27.2 fipy.variables.addOverFacesVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
27.3 fipy.variables.arithmeticCellToFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . 327
27.4 fipy.variables.betaNoiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
27.5 fipy.variables.binaryOperatorVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
v
27.6 fipy.variables.cellToFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
27.7 fipy.variables.cellVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
27.8 fipy.variables.cellVolumeAverageVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
27.9 fipy.variables.constant module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
27.10 fipy.variables.constraintMask module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
27.11 fipy.variables.coupledCellVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
27.12 fipy.variables.distanceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
27.13 fipy.variables.exponentialNoiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
27.14 fipy.variables.faceGradContributionsVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . 340
27.15 fipy.variables.faceGradVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
27.16 fipy.variables.faceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
27.17 fipy.variables.gammaNoiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
27.18 fipy.variables.gaussCellGradVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
27.19 fipy.variables.gaussianNoiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
27.20 fipy.variables.harmonicCellToFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
27.21 fipy.variables.histogramVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
27.22 fipy.variables.interfaceAreaVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.23 fipy.variables.interfaceFlagVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.24 fipy.variables.leastSquaresCellGradVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.25 fipy.variables.levelSetDiffusionVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.26 fipy.variables.meshVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.27 fipy.variables.minmodCellToFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.28 fipy.variables.modCellGradVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.29 fipy.variables.modCellToFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.30 fipy.variables.modFaceGradVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.31 fipy.variables.modPhysicalField module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.32 fipy.variables.modularVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
27.33 fipy.variables.noiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
27.34 fipy.variables.operatorVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
27.35 fipy.variables.scharfetterGummelFaceVariable module . . . . . . . . . . . . . . . . . . . . . . . . . 347
27.36 fipy.variables.surfactantConvectionVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
27.37 fipy.variables.surfactantVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
27.38 fipy.variables.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
27.39 fipy.variables.unaryOperatorVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
27.40 fipy.variables.uniformNoiseVariable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
27.41 fipy.variables.variable module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
27.42 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
28 fipy.viewers package 385

28.1 Subpackages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
28.2 Submodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
28.3 fipy.viewers.multiViewer module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
28.4 fipy.viewers.test module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
28.5 fipy.viewers.testinteractive module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
28.6 fipy.viewers.tsvViewer module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
28.7 fipy.viewers.viewer module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
28.8 Module contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
Bibliography 423
Python Module Index 425
Index 429
vi
Part I
Introduction
1
Chapter 1
Overview
FiPy is an object oriented, partial differential equation (PDE) solver, written in Python, based on a standard finite vol-
ume (FV) approach. The framework has been developed in the Materials Science and Engineering Division (MSED)
and Center for Theoretical and Computational Materials Science (CTCMS), in the Material Measurement Laboratory
(MML) at the National Institute of Standards and Technology (NIST).
The solution of coupled sets of PDEs is ubiquitous to the numerical simulation of science problems. Numerous PDE
solvers exist, using a variety of languages and numerical approaches. Many are proprietary, expensive and difficult
to customize. As a result, scientists spend considerable resources repeatedly developing limited tools for specific
problems. Our approach, combining the FV method and Python, provides a tool that is extensible, powerful and freely
available. A significant advantage to Python is the existing suite of tools for array calculations, sparse matrices and
data rendering.
The FiPy framework includes terms for transient diffusion, convection and standard sources, enabling the solution of
arbitrary combinations of coupled elliptic, hyperbolic and parabolic PDEs. Currently implemented models include
phase field [2] [3] [4] treatments of polycrystalline, dendritic, and electrochemical phase transformations, as well as
drug eluting stents [5], reactive wetting [6], photovoltaics [7] and a level set treatment of the electrodeposition process
[8].
The latest information about FiPy can be found at http://www.ctcms.nist.gov/fipy/.
1.1 Even if you don’t read manuals...
...please read Installation, Using FiPy and Frequently Asked Questions, as well as
examples.diffusion.mesh1D.
1.2 What’s new in version 3.1.3?
The significant changes since version 3.0 are:

• Level sets are now handled by LSMLIB or Scikit-fmm solver libraries. These libraries are orders of magnitude
faster than the original, Python-only prototype.
• The Matplotlib streamplot() function can be used to display vector fields.
• Version control was switched to the Git distributed version control system. This system should make it much
easier for FiPy users to participate in development.
Tickets fixed in this release:
62 "Move 'ImplicitDiffusionTerm().solve(var) == 0' ""failure"" from examples.phase.simple to example
118 subscriber()._markStale() AttributeError
138 `numerix.dot` doesn't support tensors
3
FiPy Manual, Release 3.1.3
143 "Trying to ""solve"" an integer `CellVariable` should raise an error"

195 broken arithmetic face to cell distance calculations
197 ~binOp doesn't work on branches/version-2_0
305 add rhie chow correction term in stokes cavity example
321 Windows interactive plotting mostly broken
324 --pysparse configuration should never attempt MPI imports
341 Fix fipy.terms._BinaryTerm test failure in parallel
365 Rename GridXD
368 Error adding meshes
370 Epetra Norm2 failure in parallel
383 move FiPy to distributed version control
385 `diffusionTerm._test()` requires PySparse
391 efficiency_test chokes on liquidVapor2D.py
432 LSMLIB refactor
441 Explicit convetion terms should fail when the equation has no TransientTerm (dt=None)
445 getFaceCenters() should return a FaceVariable
448 Gmsh2D does not respect background mesh
452 Gmsh background mesh doesn't work in parallel
453 faceValue as FaceCenters gives inline failures
454 Assorted errors
456 Web page links seem to be broken
457 Make the citation links go to the DOI links
460 Clean up interaction between dependencies and installation process
461 SvnToGit clean up
462 Fix for test failures on loki
465 sign issues for equation with transient, convection and implicit terms
466 "multiplying equation by ""x"" changes the solution"
469 text in source:trunk/examples/convection/source.py is out of date
470 Include mailing list activity frame on front page
473 Gmsh importer can't read mesh elements with no tags
475 getVersion() fails on Py3k
477 Update Ohloh to point at git repo
480 link to mailing list is wrong
481 constrain should return a handle to the constraint for later deletion
484 NIST CSS changed
486 Using `Popen('gmsh ...', shell=True)` rather than `shell=False` security danger
490 Parallel bug in non-uniform grids and conflicting mesh class and factory function names
491 Rename communicator instances
492 unOps can't be pickled
493 Change documentation to promote use of stackoverflow
494 Viewers don't inline well in IPython notebook
496 FIPY_DISPLAY_MATRIX is broken
497 examples/phase/binary.py has problems
513 convection problem with cylindrical grid
539 Bug with numpy 1.7.0
557 NumPy 1.7.0 doesn't have _formatInteger
564 VanLeerConvectionTerm MinMod slope limiter is broken
638 numpy 1.7.1 test failures with physicalField.py
639 Neumann boundary conditions not clearly documented
641 Add support for Matplotlib streamplot
648 Peclet inequalities have the wrong sign
650 CylindricalNonUniformGrid2D doesn't make a FaceVariable for exteriorFaces
652 Documentation change for Ubuntu install
653 enable google analytics
654 Switch to sphinxcontrib-bibtex
655 Home page needs out-of-NIST redirects
4 Chapter 1. Overview
Warning: FiPy 3 brought unavoidable syntax changes from FiPy 2. Please see
examples.updating.update2_0to3_0 for guidance on the changes that you will need to make to
your FiPy 2.x scripts.
1.3 Download and Installation
Please refer to Installation for details on download and installation. FiPy can be redistributed and/or modified freely,
provided that any derivative works bear some notice that they are derived from it, and any modified versions bear some
notice that they have been modified.
1.4 Support
You can communicate with the FiPy developers and with other users via our mailing list and we wel-
come you to use the issue tracker for bugs, support requests, feature requests and patch submissions
<https://github.com/usnistgov/fipy/issues>. We also monitor StackOverflow for questions tagged with “fipy”. We
welcome collaborative efforts on this project.
FiPy is a member of MatForge, a project of the Materials Digital Library Pathway. This National Science Foundation
funded service provides a “wiki” space for public contributions of code snippets, discussions, and tutorials and hosts
our testing framework.
1.4.1 Mailing List
In order to discuss FiPy with other users and with the developers, we encourage you to sign up for the mailing list by
sending a subscription email:
To: [email protected]
Subject: (optional)
Body: subscribe
Once you are subscribed, you can post messages to the list simply by addressing email to mailto:[email protected].
If you are new to mailing lists, you may want to read the following resource about asking effective questions:
http://www.catb.org/~esr/faqs/smart-questions.html
To get off the list follow the instructions above, but place unsubscribe in the text body.
Send help in the text body to learn other mailing list configurations you can change.
List Archive
https://www.mail-archive.com/[email protected]/
Copies of messages sent to [email protected] are stored at The Mail Archive.
(note: we have also historically sent copies to http://dir.gmane.org/gmane.comp.python.fipy, but the GMANE site now
appears to be defunct.)
1.3. Download and Installation 5

1.5 Conventions and Notation
FiPy is driven by Python script files than you can view or modify in any text editor. FiPy sessions are invoked from a
command-line shell, such as tcsh or bash.
Throughout, text to be typed at the keyboard will appear like this. Commands to be issued from an interactive
shell will appear:
$ like this
where you would enter the text (“like this”) following the shell prompt, denoted by “$”.
Text blocks of the form:
>>> a = 3 * 4
>>> a
12
>>> if a == 12:
... print "a is twelve"
...
a is twelve
are intended to indicate an interactive session in the Python interpreter. We will refer to these as “interactive sessions”
or as “doctest blocks”. The text “>>>” at the beginning of a line denotes the primary prompt, calling for input of a
Python command. The text “...” denotes the secondary prompt, which calls for input that continues from the line
above, when required by Python syntax. All remaining lines, which begin at the left margin, denote output from the
Python interpreter. In all cases, the prompt is supplied by the Python interpreter and should not be typed by you.
Warning: Python is sensitive to indentation and care should be taken to enter text exactly as it appears in the
examples.
When references are made to file system paths, it is assumed that the current working directory is the FiPy distribution
directory, refered to as the “base directory”, such that:
examples/diffusion/steadyState/mesh1D.py
will correspond to, e.g.:

/some/where/FiPy-X.Y/examples/diffusion/steadyState/mesh1D.py
Paths will always be rendered using POSIX conventions (path elements separated by “/”). Any references of the form:
examples.diffusion.steadyState.mesh1D
are in the Python module notation and correspond to the equivalent POSIX path given above.
We may at times use a
Note: to indicate something that may be of interest
or a
Warning: to indicate something that could cause serious problems.
6 Chapter 1. Overview
Chapter 2
Installation
The FiPy finite volume PDE solver relies on several third-party packages. It is best to obtain and install those first
before attempting to install FiPy. This document explains how to install FiPy, not how to use it. See Using FiPy for
details on how to use FiPy.
Note: It may be useful to set up a Development Environment before beginning the installation process.
2.1 Recommended Method
Attention:
There are many ways (described further down) to obtain the software packages necessary to run FiPy,
but the most expedient way is with the conda package manager.
• install Miniconda on your computer
• run:
$ conda create --name MYFIPYENV --channel guyer --channel conda-forge fipy
Note: This command creates a self-contained conda environment and then downloads and populates the
environment with the prerequisites for FiPy from the channels guyer and conda-forge at https://anaconda.org.
• enable this new environment with:
$ source activate MYFIPYENV
• move on to Using FiPy.
Note: On Linux and Mac OS X, you should have a pretty complete system to run and visualize FiPy
simulations. On Windows, there are fewer packages available via conda, particularly amongst the sparse
matrix Solvers, but the system still should be functional.
2.2 Installing Python Packages
In general, it is best to use the following order of precedence when installing packages:
• Use the operating system package manager, if possible.
• Use the conda package management system, which handles both Python and non-Python packages and pro-
vides facilities for self-contained environments with different combinations of Python packages, libraries, and
7
applications.
• Use the pip installs python (pip) tool to obtain software from the Python Package Index (PyPI) repository:
$ pip install package
Warning: pip takes care of dependencies that are themselves Python packages. It does not deal with
non-Python dependencies.
• Download the packages manually, unpack and run:

$ python setup.py install
Further information about each setup.py script is available by typing:

$ python setup.py --help
Many of the packages listed below have prebuilt installers for different platforms (particularly for Windows). These
installers can save considerable time and effort compared to configuring and building from source, although they
frequently comprise somewhat older versions of the respective code. Whether building from source or using a
prebuilt installer, please read and follow explicitly any instructions given in the respective packages’ README and
INSTALLATION files.
2.3 Obtaining FiPy
FiPy is freely available for download via Git or as a compressed archive from
<http://www.ctcms.nist.gov/fipy/download>. Please see documentation:GIT for instructions on obtaining FiPy
with Git.
Warning: Keep in mind that if you choose to download the compressed archive you will then need to preserve
your changes when upgrades to FiPy become available (upgrades via Git will handle this issue automatically).
2.4 Installing FiPy
Details of the Required Packages and links are given below and in platform-specific instructions, but for the courageous
and the impatient, FiPy can be up and running quickly by simply installing the following prerequisite packages on your
system:
• Python
• NumPy
• At least one of the Solvers
• At least one of the Viewers (FiPy‘s tests will run without a viewer, but you’ll want one for any practical work)
Other Optional Packages add greatly to FiPy‘s capabilities, but are not necessary for an initial installation or to simply
run the test suite.
It is not necessary to formally install FiPy, but if you wish to do so and you are confident that all of the requisite
packages have been installed properly, you can install it by typing:
$ pip install fipy
or by unpacking the archive and typing:
8 Chapter 2. Installation
$ python setup.py install
at the command line in the base FiPy directory. You can also install FiPy in “development mode” by typing:
$ python setup.py develop
which allows the source code to be altered in place and executed without issuing further installation commands.
Alternatively, you may choose not to formally install FiPy and to simply work within the base directory instead. In this
case or if you are making a non-standard install (without admin privileges), read about setting up your Development
Environment before beginning the installation process.
2.5 Required Packages
2.5.1 Python
http://www.python.org/
FiPy is written in the Python language and requires a Python installation to run. Python comes pre-installed on many
operating systems, which you can check by opening a terminal and typing python, e.g.:
$ python
Python 2.3 (#1, Sep 13 2003, 00:49:11)
...
Type "help", "copyright", "credits" or "license" for more information.
>>>
If necessary, you can download and install it for your platform <http://www.python.org/download>.
Note: FiPy requires at least version 2.4.x of Python. See the specialized instructions if you wish to RunUnderPython3.
Python along with many of FiPy‘s required and optional packages is available with one of the following distributions.
conda
http://conda.pydata.org
This package manager provides a wide array of both Python-based and general scientific packages. In addition to the
default packages, many other developers (including us) provide “channels” to distribute their own builds of a variety
of software.
In a given conda environment, you can install FiPy with:
$ conda install --channel guyer --channel conda-forge fipy
Enthought Python Distribution
http://www.enthought.com/epd
This installer provides a very large number of useful scientific packages for Python, including NumPy, SciPy, Mat-
plotlib, Mayavi, and IPython, as well as a Python interpreter. Installers are available for Windows, Mac OS X and
RedHat Linux, Solaris, Ubuntu Linux, and OpenSuSE Linux.
2.5. Required Packages 9

Attention: PySparse and FiPy are not presently included in EPD, so you will need to separately install them
manually.
Python(x,y)
http://python-xy.github.io
Another comprehensive Python package installer for scientific applications, presently only available for Windows.
Attention: PySparse and FiPy are not presently included in python(x,y), so you will need to separately install
them manually.
2.5.2 NumPy
http://numpy.scipy.org
Obtain and install the NumPy package. FiPy requires at least version 1.0 of NumPy.
2.6 Optional Packages
2.6.1 Gmsh
http://www.geuz.org/gmsh/
Gmsh is an application that allows the creation of irregular meshes.
2.6.2 SciPy
http://www.scipy.org/
SciPy provides a large collection of functions and tools that can be useful for running and analyzing FiPy simula-
tions. Significantly improved performance has been achieved with the judicious use of C language inlining (see the
Command-line Flags and Environment Variables section for more details), via the scipy.weave module.
2.7 Level Set Packages
To use the level set components of FiPy one of the following is required.
2.7.1 Scikit-fmm
http://packages.python.org/scikit-fmm/
Scikit-fmm is a python extension module which implements the fast marching method.
2.7.2 LSMLIB
http://ktchu.serendipityresearch.org/software/lsmlib/index.html
The Level Set Method Library (LSMLIB) provides support for the serial and parallel simulation of implicit surface
and curve dynamics in two- and three-dimensions.
Install LSMLIB as per the instructions on the website. Additionally PyLSMLIB is required. To install, follow the
instructions on the website, https://github.com/ktchu/LSMLIB/tree/master/pylsmlib#pylsmlib.
2.8 Platform-Specific Instructions
FiPy is tested regularly on Mac OS X, Debian Linux, Ubuntu Linux, and Windows XP. We welcome reports of
compatibility with other systems, particularly if any additional steps are necessary to install (see Miscellaneous Build
Recipes for user contributed installation tips).
The only elements of FiPy that are likely to be platform-dependent are the Viewers but at least one viewer should work
on each platform. All other aspects should function on any platform that has a recent Python installation.
2.8.1 Mac OS X Installation
There is no official package manager for Mac OS X, but there are several third-party package managers that provide
many, but not all of FiPy‘s Required Packages and Optional Packages. Options include
Fink is based on the Debian package management system. It installs all of its dependencies into /sw.
MacPorts is a package manager originally part of OpenDarwin. It installs all of its dependencies into
/opt.
Homebrew is a recent, lightweight package manager based on Ruby scripts. It installs all of its depen-
dencies into /usr/local (although it can be directed not to).
In addition, there is an Enthought Python Distribution installer for Mac OS X.
Attention: PySparse and FiPy are not presently included in any of these package managers or installers, so you
will need to separately install them manually.
We presently find that the combination of Homebrew and pip is a pretty straightforward way to get most of FiPy‘s
prerequesites. See the Miscellaneous Build Recipes for up-to-date directions.
2.8.2 Windows Installation
There is no official package manager for Windows, but the Enthought Python Distribution and Python(x,y) installers
provide most of FiPy‘s prerequisites.
Attention: PySparse and FiPy are not presently included in EPD or python(x,y), so you will need to separately
install them manually.
2.8.3 Ubuntu/Debian Installation
FiPy now has a .deb for Ubuntu/Debian systems that can be downloaded from
<http://www.ctcms.nist.gov/fipy/download>. Simply run:
2.8. Platform-Specific Instructions 11

$ VERSION=x.y-z # choose the version you want

$ apt-get install gmsh libsuperlu3 python-central python-sparse
$ curl -O http://www.ctcms.nist.gov/fipy/download/python-fipy_${VERSION}_all.deb
$ dpkg -i python-fipy_${VERSION}_all.deb
to install. The .deb includes dependencies for all of the Required Packages and Optional Packages.
2.8.4 Miscellaneous Build Recipes
We often post miscellaneous installation instructions on the FiPy blog and wiki pages. The most useful of these
include:
• Installing FiPy on Mac OS X using Homebrew
• Building a 64-bit scientific python environment for FiPy from source
• Installing FiPy with pip
Note: We encourange you to contribute your own build recipes on the wiki if they are significantly different.
2.9 Development Environment
It is often preferable to not formally install packages in the system directories. The reasons for this include:
• developing or altering the package source code,
• trying out a new package along with its dependencies without violating a working system,
• dealing with conflicting packages and dependencies,
• or not having admin privileges.
The simplest way to use a Python package without installing it is to work in the base directory of the unpacked package
and set the PYTHONPATH environment variable to “.”. In order to work in an directory other than the package’s base
directory, the PYTHONPATH environment variable must be set to “~/path/to/package”. This method of working
is adequate for one package, but quickly becomes unmanageable with multiple Python packages. In order to manage
multiple packages, it is better to choose a standard location other than the default installation path.
If you do not have administrative privileges on your computer, or if for any reason you don’t want to tamper with your
existing Python installation, most packages (including FiPy) will allow you to install to an alternative location. Instead
of installing these packages with python setup.py install, you would use python setup.py install
--home=dir, where dir is the desired installation directory (usually “~” to indicate your home directory). You will
then need to append dir/lib/python to your PYTHONPATH environment variable. See the Alternate Installation
section of the Python document “Installing Python Modules” [1] for more information, such as circumstances in which
you should use --prefix instead of --home.
An alternative to setting the PYTHONPATH is to employ one of the utilities that manage packages and their depen-
dencies independently of the system package manager and the system directories. These utilities include Stow, Vir-
tualenv_ and zc.buildout, amongst others. Here we’ll describe the use of Virtualenv_, which we highly recommend.
Chapter 3
Solvers
FiPy requires either PySparse, SciPy or Trilinos to be installed in order to solve linear systems. From our experiences,
FiPy runs most efficiently in serial when PySparse is the linear solver. Trilinos is the most complete of the three solvers
due to its numerous preconditioning and solver capabilities and it also allows FiPy to run in parallel. Although less
efficient than PySparse and less capable than Trilinos, SciPy is a very popular package, widely available and easy to
install. For this reason, SciPy may be the best linear solver choice when first installing and testing FiPy (and it is the
only viable solver under Python 3.x).
FiPy chooses the solver suite based on system availability or based on the user supplied Command-line Flags and
Environment Variables. For example, passing --no-pysparse:
$ python -c "from fipy import *; print DefaultSolver" --no-pysparse
<class 'fipy.solvers.trilinos.linearGMRESSolver.LinearGMRESSolver'>
uses a Trilinos solver. Setting FIPY_SOLVERS to scipy:

$ FIPY_SOLVERS=scipy
$ python -c "from fipy import *; print DefaultSolver"
<class 'fipy.solvers.scipy.linearLUSolver.LinearLUSolver'>
uses a SciPy solver. Suite-specific solver classes can also be imported and instantiated overriding any other directives.
For example:
$ python -c "from fipy.solvers.scipy import DefaultSolver; \
> print DefaultSolver" --no-pysparse
<class 'fipy.solvers.scipy.linearLUSolver.LinearLUSolver'>
uses a SciPy solver regardless of the command line argument. In the absence of Command-line Flags and Environment
Variables, FiPy‘s order of precedence when choosing the solver suite for generic solvers is PySparse followed by
Trilinos, PyAMG and SciPy.
3.1 PySparse
http://pysparse.sourceforge.net
PySparse is a fast serial sparse matrix library for Python. It provides several sparse matrix storage formats and
conversion methods. It also implements a number of iterative solvers, preconditioners, and interfaces to efficient
factorization packages. The only requirement to install and use Pysparse is NumPy.
Warning: FiPy requires version 1.0 or higher of PySparse.
13
3.2 SciPy
http://www.scipy.org/
The scipy.sparse module provides a basic set of serial Krylov solvers, but no preconditoners.
3.3 PyAMG
http://code.google.com/p/pyamg/
The PyAMG package provides adaptive multigrid preconditioners that can be used in conjunction with the SciPy
solvers.
3.4 Trilinos
http://trilinos.sandia.gov
Trilinos provides a more complete set of solvers and preconditioners than either PySparse or SciPy. Trilinos precon-
ditioning allows for iterative solutions to some difficult problems that PySparse and SciPy cannot solve, and it enables
parallel execution of FiPy (see Solving in Parallel for more details).
Attention: Be sure to build or install the PyTrilinos interface to Trilinos.
Attention: FiPy runs more efficiently when PySparse is installed alongside Trilinos.
Attention: Trilinos is a large software suite with its own set of prerequisites, and can be difficult to set up. It is
not necessary for most problems, and is not recommended for a basic install of FiPy.
Trilinos requires cmake, NumPy, and swig. The following are the minimal steps to build and install Trilinos (with
PyTrilinos) for FiPy:
$ cd trilinos-X.Y/
$ SOURCE_DIR=`pwd`
$ mkdir BUILD_DIR
$ cd BUILD_DIR
$ cmake \
> -D CMAKE_BUILD_TYPE:STRING=RELEASE \
> -D Trilinos_ENABLE_PyTrilinos:BOOL=ON \
> -D BUILD_SHARED_LIBS:BOOL=ON \
> -D Trilinos_ENABLE_ALL_OPTIONAL_PACKAGES:BOOL=ON \
> -D TPL_ENABLE_MPI:BOOL=ON \
> -D Trilinos_ENABLE_TESTS:BOOL=ON \
> -D DART_TESTING_TIMEOUT:STRING=600 \
> ${SOURCE_DIR}
$ make
$ make install
Depending on your platform, other options may be helpful or necessary; see the Trilinos user guide available from
http://trilinos.sandia.gov/documentation.html, or http://trilinos.sandia.gov/packages/pytrilinos/faq.html for more in-
depth documentation.
14 Chapter 3. Solvers
Note: Trilinos can be installed in a non-standard location by adding the -D

CMAKE_INSTALL_PREFIX:PATH=${INSTALL_DIR} and -D PyTrilinos_INSTALL_PREFIX:PATH=${INSTALL_DIR}
flags to the configure step. If Trilinos is installed in a non-standard location, the path to the PyTrilinos site-
packages directory should be added to the PYTHONPATH environment variable; this should be of the form
${INSTALL_DIR}/lib/${PYTHON_VERSION }/site-packages/. Also, the path to the Trilinos lib
directory should be added to the LD_LIBRARY_PATH (on Linux) or DYLD_LIBRARY_PATH (on Mac OS X)
environment variable; this should be of the form ${INSTALL_DIR}/lib‘.
3.4.1 mpi4py
http://mpi4py.scipy.org/
For Solving in Parallel, FiPy requires mpi4py, in addition to Trilinos.
3.4. Trilinos 15
16 Chapter 3. Solvers
Chapter 4
Viewers
A viewer is required to see the results of FiPy calculations. Matplotlib is by far the most widely used Python based
viewer and the best choice to get FiPy up and running quickly. Matplotlib is also capable of publication quality plots.
Matplotlib has only rudimentary 3D capability, which FiPy does not attempt to use. Mayavi is required for 3D viewing.
4.1 Matplotlib
http://matplotlib.sourceforge.net
Matplotlib is a Python package that displays publication quality results. It displays both 1D X-Y type plots and
2D contour plots for both structured and unstructured data, but does not display 3D data. It works on all common
platforms.
4.2 Mayavi
http://code.enthought.com/projects/mayavi/
The Mayavi Data Visualizer is a free, easy to use scientific data visualizer. It displays 1D, 2D and 3D data. It is the
only FiPy viewer available for 3D data. Matplotlib is probably a better choice for 1D or 2D viewing.
Mayavi requires VTK, which can be difficult to build from source. Mayavi and VTK can be most easily obtained
through
• the Ubuntu or Debian package managers
• the Enthought Python Edition
• python(x,y)
• the Homebrew package manager for Mac OS X (VTK only, not Mayavi)
Note: MayaVi 1 is no longer supported.
17
18 Chapter 4. Viewers
Chapter 5
Using FiPy
This document explains how to use FiPy in a practical sense. To see the problems that FiPy is capable of solving, you
can run any of the scripts in the examples.
Note: We strongly recommend you proceed through the examples, but at the very least work through
examples.diffusion.mesh1D to understand the notation and basic concepts of FiPy.
We exclusively use either the unix command line or IPython to interact with FiPy. The commands in the examples
are written with the assumption that they will be executed from the command line. For instance, from within the main
FiPy directory, you can type:
$ python examples/diffusion/mesh1D.py
A viewer should appear and you should be prompted through a series of examples.
Note: From within IPython, you would type:

>>> run examples/diffusion/mesh1D.py
In order to customize the examples, or to develop your own scripts, some knowledge of Python syntax is required. We
recommend you familiarize yourself with the excellent Python tutorial [11] or with Dive Into Python [12].
As you gain experience, you may want to browse through the Command-line Flags and Environment Variables that
affect FiPy.
5.1 Testing FiPy
For a general installation, FiPy can be tested by running:

$ python -c "import fipy; fipy.test()"
This command runs all the test cases in FiPy’s modules, but doesn’t include any of the tests in FiPy’s examples. To
run the test cases in both modules and examples, use:
$ python setup.py test
in an unpacked FiPy archive. The test suite can be run with a number of different configurations depending on which
solver suite is available and other factors. See Command-line Flags and Environment Variables for more details.
FiPy will skip tests that depend on Optional Packages that have not been installed. For example, if Mayavi and Gmsh
are not installed, FiPy will warn:
19
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Skipped 131 doctest examples because `gmsh` cannot be found on the $PATH
Skipped 42 doctest examples because the `tvtk` package cannot be imported
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
We have a few known, intermittent failures:

#425 The test suite can freeze, usually in examples.chemotaxis, when running on multiple proces-
sors. This has never affected us in an actual parallel simulation, only in the test suite.
#430 When running in parallel, the tests for _BinaryTerm sometimes return one erroneous result. This
is not reliably reproducible and doesn’t seem to have an effect on actual simulations.
Although the test suite may show warnings, there should be no other errors. Any errors should be investigated or
reported on the issue tracker. Users can see if there are any known problems for the latest FiPy distribution by
checking FiPy’s automated test display.
Below are a number of common Command-line Flags for testing various FiPy configurations.
5.1.1 Parallel Tests
If FiPy is configured for Solving in Parallel, you can run the tests on multiple processor cores with:
$ mpirun -np {# of processors} python setup.py test --trilinos
or:
$ mpirun -np {# of processors} python -c "import fipy; fipy.test('--trilinos')"
5.2 Command-line Flags and Environment Variables
FiPy chooses a default run time configuration based on the available packages on the system. The Command-line
Flags and Environment Variables sections below describe how to override FiPy‘s default behavior.
5.2.1 Command-line Flags
You can add any of the following case-insensitive flags after the name of a script you call from the command line, e.g:
$ python myFiPyScript --someflag
--inline
Causes many mathematical operations to be performed in C, rather than Python, for improved performance.
Requires the scipy.weave package.
The following flags take precedence over the FIPY_SOLVERS environment variable:
--pysparse
Forces the use of the PySparse solvers.
--trilinos
Forces the use of the Trilinos solvers, but uses PySparse to construct the matrices.
--no-pysparse
Forces the use of the Trilinos solvers without any use of PySparse.
--scipy
Forces the use of the SciPy solvers.
20 Chapter 5. Using FiPy

--pyamg
Forces the use of the PyAMG preconditioners in conjunction with the SciPy solvers.
--lsmlib
Forces the use of the LSMLIB level set solver.
--skfmm
Forces the use of the Scikit-fmm level set solver.
5.2.2 Environment Variables
You can set any of the following environment variables in the manner appropriate for your shell. If you are not
running in a shell (e.g., you are invoking FiPy scripts from within IPython or IDLE), you can set these variables via
the os.environ dictionary, but you must do so before importing anything from the fipy package.
FIPY_DISPLAY_MATRIX
If present, causes the graphical display of the solution matrix of each equation at each call of solve() or
sweep(). Setting the value to “terms,” causes the display of the matrix for each Term that composes the
equation. Requires the Matplotlib package.
FIPY_INLINE
If present, causes many mathematical operations to be performed in C, rather than Python. Requires the
scipy.weave package.
FIPY_INLINE_COMMENT
If present, causes the addition of a comment showing the Python context that produced a particular piece of
scipy.weave C code. Useful for debugging.
FIPY_SOLVERS
Forces the use of the specified suite of linear solvers. Valid (case-insensitive) choices are “pysparse”,
“trilinos”, “no-pysparse”, “scipy” and “pyamg”.
FIPY_VERBOSE_SOLVER
If present, causes the linear solvers to print a variety of diagnostic information.
FIPY_VIEWER
Forces the use of the specified viewer. Valid values are any <viewer> from the
fipy.viewers.<viewer>Viewer modules. The special value of dummy will allow the script to
run without displaying anything.
FIPY_INCLUDE_NUMERIX_ALL
If present, causes the inclusion of all funcions and variables of the numerix module in the fipy namespace.
5.3 Solving in Parallel
FiPy can use Trilinos to solve equations in parallel. Most mesh classes in fipy.meshes can solve in parallel. This
includes all “...Grid...” and “...Gmsh...” class meshes. Currently, the only remaining serial-only meshes
are Tri2D and SkewedGrid2D.
Attention: Trilinos must be compiled with MPI support.
Attention: FiPy requires mpi4py to work in parallel. See the mpi4py installation guide.
Note: Parallel efficiency is greatly improved by installing PySparse in addition to Trilinos. If PySparse is not installed
be sure to use the --no-pysparse flag when running in parallel.
5.3. Solving in Parallel 21

It should not generally be necessary to change anything in your script. Simply invoke:
$ mpirun -np {# of processors} python myScript.py --trilinos
instead of:
$ python myScript.py
To confirm that FiPy and Trilinos are properly configured to solve in parallel, the easiest way to tell is to run one of
the examples, e.g.,:
$ mpirun -np 2 examples/diffusion/mesh1D.py
You should see two viewers open with half the simulation running in one of them and half in the other. If this does not
look right (e.g., you get two viewers, both showing the entire simultion), or if you just want to be sure, you can run a
diagnostic script:
$ mpirun -np 3 python examples/parallel.py
which should print out:

mpi4py: processor 0 of 3 :: PyTrilinos: processor 0 of 3 :: FiPy: 5 cells on processor 0 of 3
If there is a problem with your parallel environment, it should be clear that there is either a problem importing one of
the required packages or that there is some problem with the MPI environment. For example:
[my.machine.com:69815] WARNING: There were 4 Windows created but not freed.
indicates mpi4py is properly communicating with MPI and is running in parallel, but that Trilinos is not, and is running
three separate serial environments. As a result, FiPy is limited to three separate serial operations, too. In this instance,
the problem is that although Trilinos was compiled with MPI enabled, it was compiled against a different MPI library
than is currently available (and which mpi4py was compiled against). The solution is to rebuild Trilinos against the
active MPI libraries.
When solving in parallel, FiPy essentially breaks the problem up into separate sub-domains and solves them (some-
what) independently. FiPy generally “does the right thing”, but if you find that you need to do something with the
entire solution, you can use var.globalValue.
Note: Trilinos solvers frequently give intermediate output that FiPy cannot suppress. The most commonly encoun-
tered messages are
Gen_Prolongator warning : Max eigen <= 0.0 which is not significant to FiPy.
Aztec status AZ_loss: loss of precision which indicates that there was some diffi-
culty in solving the problem to the requested tolerance due to precision limitations, but usually does
not prevent the solver from finding an adequate solution.
Aztec status AZ_ill_cond: GMRES hessenberg ill-conditioned which indi-
cates that GMRES is having trouble with the problem, and may indicate that trying a different
solver or preconditioner may give more accurate results if GMRES fails.

Aztec status AZ_breakdown: numerical breakdown which usually indicates serious

problems solving the equation which forced the solver to stop before reaching an adequate solu-
tion. Different solvers, different preconditioners, or a less restrictive tolerance may help.
5.4 Meshing with Gmsh
FiPy works with arbitrary polygonal meshes generated by Gmsh. FiPy provides two wrappers classes (Gmsh2D and
Gmsh3D) enabling Gmsh to be used directly from python. The classes can be instantiated with a set of Gmsh style
commands (see examples.diffusion.circle). The classes can also be instantiated with the path to either a
Gmsh geometry file (.geo) or a Gmsh mesh file (.msh) (see examples.diffusion.anisotropy).
As well as meshing arbitrary geometries, Gmsh partitions meshes for parallel simulations. Mesh partitioning automat-
ically occurs whenever a parallel communicator is passed to the mesh on instantiation. This is the default setting for
all meshes that work in parallel including Gmsh2D and Gmsh3D.
Note: FiPy solution accuracy can be compromised with highly non-orthogonal or non-conjunctional meshes.
5.5 Coupled and Vector Equations
Equations can now be coupled together so that the contributions from all the equations appear in a single system
matrix. This results in tighter coupling for equations with spatial and temporal derivatives in more than one variable.
In FiPy equations are coupled together using the & operator:
>>> eqn0 = ...
>>> eqn1 = ...
>>> coupledEqn = eqn0 & eqn1
The coupledEqn will use a combined system matrix that includes four quadrants for each of the different variable
and equation combinations. In previous versions of FiPy there has been no need to specify which variable a given
term acts on when generating equations. The variable is simply specified when calling solve or sweep and this
functionality has been maintained in the case of single equations. However, for coupled equations the variable that a
given term operates on now needs to be specified when the equation is generated. The syntax for generating coupled
equations has the form:
>>> eqn0 = Term00(coeff=..., var=var0) + Term01(coeff..., var=var1) == source0
>>> eqn1 = Term10(coeff=..., var=var0) + Term11(coeff..., var=var1) == source1
>>> coupledEqn = eqn0 & eqn1
and there is now no need to pass any variables when solving:

>>> coupledEqn.solve(dt=..., solver=...)
In this case the matrix system will have the form

(︂ )︂ (︂ )︂ (︂ )︂
Term00 Term01 var0 source0
=
Term10 Term11 var1 source1
FiPy tries to make sensible decisions regarding each term’s location in the matrix and the ordering of the variable
column array. For example, if Term01 is a transient term then Term01 would appear in the upper left diagonal and
the ordering of the variable column array would be reversed.
The use of coupled equation is described in detail in examples.diffusion.coupled. Other ex-
amples that demonstrate the use of coupled equations are examples.phase.binaryCoupled,
5.4. Meshing with Gmsh 23

examples.phase.polyxtalCoupled and examples.cahnHilliard.mesh2DCoupled. As well

as coupling equations, true vector equations can now be written in FiPy (see examples.diffusion.coupled
for more details).
5.6 Boundary Conditions
5.6.1 Applying fixed value (Dirichlet) boundary conditions
To apply a fixed value boundary condition use the constrain() method. For example, to fix var to have a value of
2 along the upper surface of a domain, use
>>> var.constrain(2., where=mesh.facesTop)
Note: The old equivalent FixedValue boundary condition is now deprecated.
5.6.2 Applying fixed gradient boundary conditions (Neumann)
To apply a fixed Gradient boundary condition use the faceGrad.constrain() method. For example, to fix var to
have a gradient of (0,2) along the upper surface of a 2D domain, use
>>> var.faceGrad.constrain(((0,),(2,)), where=mesh.facesTop)
If the gradient normal to the boundary (e.g., 𝑛

ˆ · ∇𝜑) is to be set to a scalar value of 2, use
>>> var.faceGrad.constrain(2 * mesh.faceNormals, where=mesh.exteriorFaces)
5.6.3 Applying fixed flux boundary conditions
Generally these can be implemented with a judicious use of faceGrad.constrain(). Failing that, an exterior
flux term can be added to the equation. Firstly, set the terms’ coefficients to be zero on the exterior faces,
>>> diffCoeff.constrain(0., mesh.exteriorFaces)
>>> convCoeff.constrain(0., mesh.exteriorFaces)
then create an equation with an extra term to account for the exterior flux,
>>> eqn = (TransientTerm() + ConvectionTerm(convCoeff)
... == DiffusionCoeff(diffCoeff)
... + (mesh.exteriorFaces * exteriorFlux).divergence)
where exteriorFlux is a rank 1 FaceVariable.
Note: The old equivalent FixedFlux boundary condition is now deprecated.
5.6.4 Applying outlet or inlet boundary conditions
Convection terms default to a no flux boundary condition unless the exterior faces are associated with a constraint, in
which case either an inlet or an outlet boundary condition is applied depending on the flow direction.

5.6.5 Applying spatially varying boundary conditions
The use of spatial varying boundary conditions is best demonstrated with an example. Given a 2D equation in the
domain 0 < 𝑥 < 1 and 0 < 𝑦 < 1 with boundary conditions,
𝑥𝑦 on 𝑥 > 1/2 and 𝑦 > 1/2

{︃
𝜑=
⃗
⃗𝑛 · 𝐹 = 0 elsewhere
where 𝐹⃗ represents the flux. The boundary conditions in FiPy can be written with the following code,
>>> X, Y = mesh.faceCenters
>>> mask = ((X < 0.5) | (Y < 0.5))
>>> var.faceGrad.constrain(0, where=mesh.exteriorFaces & mask)
>>> var.constrain(X * Y, where=mesh.exteriorFaces & ~mask)
then
>>> eqn.solve(...)
Further demonstrations of spatially varying boundary condition can be found in

examples.diffusion.mesh20x20 and examples.diffusion.circle
5.6.6 Applying internal boundary conditions
Applying internal boundary conditions can be achieved through the use of implicit and explicit sources. An equation
of the form
>>> eqn = TransientTerm() == DiffusionTerm()
can be constrained to have a fixed internal value at a position given by mask with the following alterations
>>> eqn = TransientTerm() == DiffusionTerm() - ImplicitSourceTerm(mask * largeValue) + mask * largeVa
The parameter largeValue must be chosen to be large enough to completely dominate the matrix diagonal and the
RHS vector in cells that are masked. The mask variable would typically be a CellVariable boolean constructed
using the cell center values.
One must be careful to distinguish between constraining internal cell values during the solve step and simply applying
arbitrary constraints to a CellVariable. Applying a constraint,
>>> var.constrain(value, where=mask)
simply fixes the returned value of var at mask to be value. It does not have any effect on the implicit value of var
at the mask location during the linear solve so it is not a substitute for the source term machinations described above.
Future releases of FiPy may implicitly deal with this discrepancy, but the current release does not. A simple example
can be used to demonstrate this:
>>> m = Grid1D(nx=2, dx=1.)
>>> var = CellVariable(mesh=m)
Apply a constraint to the faces for a right side boundary condition (which works).
>>> var.constrain(1., where=m.facesRight)
Create the equation with the source term constraint described above
5.6. Boundary Conditions 25

>>> mask = m.x < 1.

>>> largeValue = 1e+10
>>> value = 0.25
>>> eqn = DiffusionTerm() - ImplicitSourceTerm(largeValue * mask) + largeValue * mask * value
and the expected value is obtained.

>>> eqn.solve(var)
>>> print var
[ 0.25 0.75]
However, if a constraint is used without the source term constraint an unexpected value is obtained
>>> var.constrain(0.25, where=mask)
>>> eqn = DiffusionTerm()
>>> eqn.solve(var)
>>> print var
[ 0.25 1. ]
although the left cell has the expected value as it is constrained.
5.7 Running under Python 3
It is possible to run FiPy scripts under Python 3, but there is admittedly little advantage in doing so at this time. We
still develop and use FiPy under Python 2.x. To use, you must first convert FiPy‘s code to Python 3 syntax. From
within the main FiPy directory:
$ 2to3 --write .
$ 2to3 --write --doctests_only .
You can expect some harmless warnings from this conversion.

The minimal prerequisites are:
• NumPy version 1.5 or greater.
• SciPy version 0.9 or greater.
• Matplotlib version 1.2 or greater (this hasn’t been released yet, and we haven’t been able to successfully test the
matplotlibViewer classes with their development code).
5.8 Manual
You can view the manual online at <http://www.ctcms.nist.gov/fipy> or you can download the latest manual from
<http://www.ctcms.nist.gov/fipy/download/>. Alternatively, it may be possible to build a fresh copy by issuing the
following command in the base directory:
$ python setup.py build_docs --pdf --html
Note: This mechanism is intended primarily for the developers. At a minimum, you will need at least version 1.1.2
of Sphinx, plus all of its prerequisites, although we build the documentation witih the latest development code (you
will need hg installed):
$ pip install --upgrade -e hg+https://bitbucket.org/birkenfeld/sphinx#egg=sphinx
We use several contributed Sphinx plugins:

$ hg clone https://bitbucket.org/birkenfeld/sphinx-contrib/
$ cd sphinx-contrib/traclinks $ python setup.py install

Bibliographic citations require the sphinxcontrib-bibtex package. For the moment, the development versions of several
packages are required to properly render our bibliography (you will need both bzr and git installed):
$ pip install -e bzr+lp:~pybtex-devs/pybtex/trunk
$ pip install -e [email protected]:mcmtroffaes/pybtex-docutils.git#egg=pybtex-docutils
$ pip install -e [email protected]:mcmtroffaes/sphinxcontrib-bibtex.git#egg=sphinxcontrib-bibtex
5.8. Manual 27

Chapter 6
Theoretical and Numerical Background
This chapter describes the numerical methods used to solve equations in the FiPy programming environment. FiPy
uses the finite volume method (FVM) to solve coupled sets of partial differential equations (PDEs). For a good
introduction to the FVM see Nick Croft’s PhD thesis [13], Patankar [18] or Versteeg and Malalasekera [19].
Essentially, the FVM consists of dividing the solution domain into discrete finite volumes over which the state variables
are approximated with linear or higher order interpolations. The derivatives in each term of the equation are satisfied
with simple approximate interpolations in a process known as discretization. The (FVM) is a popular discretization
technique employed to solve coupled PDEs used in many application areas (e.g., Fluid Dynamics).
The FVM can be thought of as a subset of the Finite Element Method (FEM), just as the Finite Difference Method
(FDM) is a subset of the FVM. A system of equations fully equivalent to the FVM can be obtained with the FEM
using as weighting functions the characteristic functions of FV cells, i.e., functions equal to unity [20]. Analogously,
the discretization of equations with the FVM reduces to the FDM on Cartesian grids.
6.1 General Conservation Equation
The equations that model the evolution of physical, chemical and biological systems often have a remarkably universal
form. Indeed, PDEs have proven necessary to model complex physical systems and processes that involve variations
in both space and time. In general, given a variable of interest 𝜑 such as species concentration, pH, or temperature,
there exists an evolution equation of the form
𝜕𝜑
= 𝐻(𝜑, 𝜆𝑖 ) (6.1)
𝜕𝑡
where 𝐻 is a function of 𝜑, other state variables 𝜆𝑖 , and higher order derivatives of all of these variables. Examples of
such systems are wide ranging, but include problems that exhibit a combination of diffusing and reacting species, as
well as such diverse problems as determination of the electric potential in heart tissue, of fluid flow, stress evolution,
and even the Schroedinger equation.
A general conservation equation, solved using FiPy, can include any combination of the following terms,
𝜕(𝜌𝜑) 𝑛
+ ∇ · (⃗𝑢𝜑) = [∇ · (Γ𝑖 ∇)] 𝜑 + 𝑆𝜑
𝜕𝑡
⏟ ⏞ ⏟ ⏞ ⏟ ⏞ ⏟ ⏞ (6.2)
transient convection diffusion source
where 𝜌, ⃗𝑢 and Γ𝑖 represent coefficients in the transient, convection and diffusion terms, respectively. These coeffi-
cients can be arbitrary functions of any parameters or variables in the system. The variable 𝜑 represents the unknown
quantity in the equation. The diffusion term can represent any higher order diffusion-like term, where the order is given
by the exponent 𝑛. For example, the diffusion term can represent conventional Fickean diffusion [i.e., ∇ · (Γ∇𝜑)]
when the exponent 𝑛 = 1 or a Cahn-Hilliard term [i.e., ∇ · (Γ1 ∇[∇ · Γ2 ∇𝜑)]) [14] [15] [16]] when 𝑛 = 2, or a phase
field crystal term [i.e., ∇ · (Γ1 ∇[∇ · Γ2 ∇{∇ · Γ3 ∇𝜑)})]) [17]] when 𝑛 = 3, although spectral methods are probably
a better approach. Higher order terms (𝑛 > 3) are also possible, but the matrix condition number becomes quite poor.
29
 . . . . . .   ..  =  .. 
,

 ., . . 
 .  .
ρφV − (ρφV )old
+
[($n · $u)Aφ]face = . [ΓA$ . n · ∇φ]face φn+ V Sφ bn
∆t . a nn
face face
' ()
6.2 Finite*Volume ' Method () * ' () * '()*
transient convection ∂(ρφ) diffusion source
To use the FVM, the solution domain must first be divided into+ non-overlapping
∇ · ($uφ)polyhedral = ∇elements · (Γ∇φ) or cells. +
A Sφ
* In the'FVM() * ' () * '()*
solution domain divided in such a way is generally∂t known as a mesh (as we will see, a Mesh is also a FiPy object).
A mesh consists of vertices, faces and cells (see'Figure
()Mesh). the variables of interest are averaged over
convection
control volumes (CVs). The CVs are either defined by the cells or are centered on the vertices. diffusion source
transient
+ + + +
∂(ρφ)
fa∂t dV + ($n · $u)φ dS = Γ($n · ∇φ) dS +
ce vertex
'V () * ' S () * 'S () * '
cell
transient convection diffusion
Fig. 6.1: , ,
ρφV − (ρφV )old Mesh
+ [($n · $u)Aφ] = [ΓA$n · ∇φ]f

A mesh consists of cells, faces and vertices. For the purposes of FiPy, the divider between two cells is known as a face for all
dimensions. face
∆t
face face
' () * ' () * ' ()
6.2.1 Cell Centered FVM (CC-FVM)
transient convection diffusion
In the CC-FVM the CVs are formed by the mesh cells with the cell center “storing” the average variable value in the
CV, (see Figure CV structure for an unstructured mesh). The face fluxes are approximated using the variable values
in the two adjacent cells surrounding the face. This low order approximation has the advantage of being efficient and
requiring matrices of low band width (the band width is equal to the number of cell neighbors plus one) and thus low
storage requirement. However, the mesh topology is restricted due to orthogonality and conjunctionality requirements.
The value at a face is assumed to be the average value over the face. On an unstructured mesh the face center may not
lie on the line joining the CV centers, which will lead to an error in the face interpolation. FiPy currently only uses the
CC-FVM.
6.2.2 Vertex Centered FVM (VC-FVM)
In the VC-FVM, the CV is centered around the vertices and the cells are divided into sub-control volumes that make
up the main CVs (see Figure CV structure for an unstructured mesh). The vertices “store” the average variable values
over the CVs. The CV faces are constructed within the cells rather than using the cell faces as in the CC-FVM. The
face fluxes use all the vertex values from the cell where the face is located to calculate interpolations. For this reason,
the VC-FVM is less efficient and requires more storage (a larger matrix band width) than the CC-FVM. However, the
mesh topology does not have the same restrictions as the CC-FVM. FiPy does not have a VC-FVM capability.
6.3 Discretization
The first step in the discretization of Equation (??) using the CC-FVM is to integrate over a CV and then make
appropriate approximations for fluxes across the boundary of each CV. In this section, each term in Equation (??) will
be examined separately.
30 Chapter 6. Theoretical and Numerical Background
1
Ω2
2
Ω3
Ω1
3
a
1
Ω4
(b)
2
Ωa
a
3
1
(a)
Fig. 6.2: CV structure for an unstructured mesh

(a) Ω𝑎 represents a vertex-based CV and (b) Ω1 , Ω2 , Ω3 and Ω4 represent cell centered CVs.
6.3. Discretization 31
6.3.1 Transient Term 𝜕(𝜌𝜑)/𝜕𝑡
For the transient term, the discretization of the integral over the volume of a CV is given by
∫︀
𝑉
𝜕(𝜌𝜑) (𝜌𝑃 𝜑𝑃 − 𝜌old old

𝑃 𝜑𝑃 )𝑉𝑃
∫︁
𝑑𝑉 ≃ (6.3)
𝑉 𝜕𝑡 ∆𝑡
where 𝜑𝑃 represents the average value of 𝜑 in a CV centered on a point 𝑃 and the superscript “old” represents the
previous time-step value. The value 𝑉𝑃 is the volume of the CV and ∆𝑡 is the time step size.
This term is represented in FiPy as
>>> TransientTerm(coeff=rho)
6.3.2 Convection Term ∇ · (⃗𝑢𝜑)
The discretization for the convection term is given by

∫︁ ∫︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 = (⃗𝑛 · ⃗𝑢)𝜑 𝑑𝑆
𝑉 𝑆
∑︁ (6.4)
≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑓
where we have∫︀ used the divergence theorem to transform the integral over
∑︀ the CV volume 𝑉 into an integral over the
∫︀
CV surface 𝑆 . The summation over the faces of a CV is denoted by 𝑓 and 𝐴𝑓 is the area of each face. The vector
⃗𝑛 is the normal to the face pointing out of the CV into an adjacent CV centered on point 𝐴. When using a first order
approximation, the value of 𝜑𝑓 must depend on the average value in adjacent cell 𝜑𝐴 and the average value in the cell
of interest 𝜑𝑃 , such that
𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 .
The weighting factor 𝛼𝑓 is determined by the convection scheme, described in Numerical Schemes.
>>> <SpecificConvectionTerm>(coeff=u)
where <SpecificConvectionTerm> can be any of CentralDifferenceConvectionTerm,

ExponentialConvectionTerm, HybridConvectionTerm, PowerLawConvectionTerm,
UpwindConvectionTerm, ExplicitUpwindConvectionTerm, or VanLeerConvectionTerm.
The differences between these convection schemes are described in Section Numerical Schemes. The velocity
coefficient u must be a rank-1 FaceVariable, or a constant vector in the form of a Python list or tuple, e.g.
((1,), (2,)) for a vector in 2D.
6.3.3 Diffusion Term ∇ · (Γ1 ∇𝜑)
The discretization for the diffusion term is given by

∫︁ ∫︁
∇ · (Γ∇{. . .})𝑑𝑉 = Γ(⃗𝑛 · ∇{. . .})𝑑𝑆
𝑉 𝑆
∑︁ (6.5)
≃ Γ𝑓 (⃗𝑛 · ∇{. . .})𝑓 𝐴𝑓
𝑓

{. . .} indicates recursive application of the specified operation on 𝜑, depending on the order of the diffusion term. The
estimation for the flux, (⃗𝑛 · ∇{. . .})𝑓 , is obtained via
{. . .}𝐴 − {. . .}𝑃
(⃗𝑛 · ∇{. . .})𝑓 ≃
𝑑𝐴𝑃
where the value of 𝑑𝐴𝑃 is the distance between neighboring cell centers. This estimate relies on the orthogonality of
the mesh, and becomes increasingly inaccurate as the non-orthogonality increases. Correction terms have been derived
to improve this error but are not currently included in FiPy [13].
>>> DiffusionTerm(coeff=Gamma1)
or
>>> ExplicitDiffusionTerm(coeff=Gamma1)
ExplicitDiffusionTerm is provided primarily for illustrative purposes, although

examples.diffusion.mesh1D demonstrates its use in Crank-Nicolson time stepping.
ImplicitDiffusionTerm is almost always preferred (DiffusionTerm is a synonym for
ImplicitDiffusionTerm to reinforce this preference). One can also create an explicit diffusion term
with
>>> (Gamma1 * phi.faceGrad).divergence
Higher order diffusion
Higher order diffusion expressions, such as ∇4 𝜑 or ∇ · (Γ1 ∇ (∇ · (Γ2 ∇𝜑))) for Cahn-Hilliard are represented as
>>> DiffusionTerm(coeff=(Gamma1, Gamma2))
The number of elements supplied for coeff determines the order of the term.
6.3.4 Source Term
Any term that cannot be written in one of the previous forms is considered a source 𝑆𝜑 . The discretization for the
source term is given by,
∫︁
𝑆𝜑 𝑑𝑉 ≃ 𝑆𝜑 𝑉𝑃 . (6.6)
𝑉
Including any negative dependence of 𝑆𝜑 on 𝜑 increases solution stability. The dependence can only be included in a
linear manner so Equation (6.6) becomes
𝑉𝑃 (𝑆0 + 𝑆1 𝜑𝑃 ),
where 𝑆0 is the source which is independent of 𝜑 and 𝑆1 is the coeficient of the source which is linearly dependent on
𝜑.
A source term is represented in FiPy essentially as it appears in mathematical form, e.g., 3𝜅2 +𝑏 sin 𝜃 would be written
>>> 3 * kappa**2 + b * numrix.sin(theta)
Note: Functions like sin() can be obtained from the fipy.tools.numerix module.
Warning: Generally, things will not work as expected if the equivalent function is used from the NumPy or SciPy
library.
6.3. Discretization 33
If, however, the source depends on the variable that is being solved for, it can be advantageous to linearize the source
and cast part of it as an implicit source term, e.g., 3𝜅2 + 𝜑 sin 𝜃 might be written as
>>> 3 * kappa**2 + ImplicitSourceTerm(coeff=sin(theta))
6.4 Linear Equations
The aim of the discretization is to reduce the continuous general equation to a set of discrete linear equations that can
then be solved to obtain the value of the dependent variable at each CV center. This results in a sparse linear system
that requires an efficient iterative scheme to solve. The iterative schemes available to FiPy are currently encapsulated
in the PySparse and PyTrilinos suites of solvers and include most common solvers such as the conjugate gradient
method and LU decomposition.
Combining Equations (6.3), (6.4), (6.5) and (6.6), the complete discretization for equation (??) can now be written for
each CV as
𝜌𝑃 (𝜑𝑃 − 𝜑old
𝑃 )𝑉𝑃
∑︁
+ (⃗𝑛 · ⃗𝑢)𝑓 𝐴𝑓 [𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 ) 𝜑𝐴 ]
∆𝑡
𝑓
∑︁ (𝜑𝐴 − 𝜑𝑃 )
= Γ 𝑓 𝐴𝑓 + 𝑉𝑃 (𝑆0 + 𝑆1 𝜑𝑃 ).
𝑑𝐴𝑃
𝑓
Equation (6.4) is now in the form of a set of linear combinations between each CV value and its neighboring values
and can be written in the form
∑︁
𝑎𝑃 𝜑𝑃 = 𝑎𝐴 𝜑𝐴 + 𝑏𝑃 , (6.7)
𝑓
where
𝜌𝑃 𝑉𝑃 ∑︁
𝑎𝑃 = + (𝑎𝐴 + 𝐹𝑓 ) − 𝑉𝑃 𝑆1 ,
∆𝑡
𝑓
𝑎𝐴 = 𝐷𝑓 − (1 − 𝛼𝑓 )𝐹𝑓 ,
𝜌𝑃 𝑉𝑃 𝜑old
𝑃
𝑏𝑃 = 𝑉𝑃 𝑆0 + .
∆𝑡
The face coefficients, 𝐹𝑓 and 𝐷𝑓 , represent the convective strength and diffusive conductance respectively, and are
given by
𝐹𝑓 = 𝐴𝑓 (⃗𝑢 · ⃗𝑛)𝑓 ,
𝐴𝑓 Γ 𝑓
𝐷𝑓 = .
𝑑𝐴𝑃
6.5 Numerical Schemes
The coefficients of equation (??) must remain positive, since an increase in a neighboring value must result in an
increase in 𝜑𝑃 to obtain physically realistic solutions. Thus, the inequalities 𝑎𝐴 > 0 and 𝑎𝐴 + 𝐹𝑓 > 0 must be
satisfied. The Peclet number 𝑃𝑓 ≡ 𝐹𝑓 /𝐷𝑓 is the ratio between convective strength and diffusive conductance. To
achieve physically realistic solutions, the inequality
1 1
> 𝑃𝑓 > − (6.8)
1 − 𝛼𝑓 𝛼𝑓

must be satisfied. The parameter 𝛼𝑓 is defined by the chosen scheme, depending on Equation (6.8). The various
differencing schemes are:
the central differencing scheme, where
1
𝛼𝑓 = (6.9)
2
so that |𝑃𝑓 | < 2 satisfies Equation (6.8). Thus, the central differencing scheme is only numerically stable for a
low values of 𝑃𝑓 .
the upwind scheme, where
{︃
1 if 𝑃𝑓 > 0,
𝛼𝑓 = (6.10)
0 if 𝑃𝑓 < 0.
Equation (6.10) satisfies the inequality in Equation (6.8) for all values of 𝑃𝑓 . However the solution over predicts
the diffusive term leading to excessive numerical smearing (“false diffusion”).
the exponential scheme, where
(𝑃𝑓 − 1) exp (𝑃𝑓 ) + 1

𝛼𝑓 = . (6.11)
𝑃𝑓 (exp (𝑃𝑓 ) − 1)
This formulation can be derived from the exact solution, and thus, guarantees positive coefficients while not
over-predicting the diffusive terms. However, the computation of exponentials is slow and therefore a faster
scheme is generally used, especially in higher dimensions.
the hybrid scheme, where
⎧ 𝑃𝑓 −1
⎨ 𝑃𝑓
⎪ if 𝑃𝑓 > 2,
𝛼𝑓 = 1
if |𝑃𝑓 | < 2, (6.12)
⎪2
− 𝑃1𝑓 if 𝑃𝑓 < −2.
⎩
The hybrid scheme is formulated by allowing 𝑃𝑓 → ∞, 𝑃𝑓 → 0 and 𝑃𝑓 → −∞ in the exponential scheme.

The hybrid scheme is an improvement on the upwind scheme, however, it deviates from the exponential scheme
at |𝑃𝑓 | = 2.
the power law scheme, where
⎧ 𝑃𝑓 −1
⎪
⎪ 𝑃𝑓 if 𝑃𝑓 > 10,
⎨ (𝑃𝑓 −1)+(1−𝑃𝑓 /10)5
⎪
if 0 < 𝑃𝑓 < 10,
⎪
𝑃𝑓
𝛼𝑓 = 5 (6.13)
⎪ (1−𝑃𝑓 /10) −1 if −10 < 𝑃𝑓 < 0,
⎪
⎪
⎪ 𝑃 𝑓
if 𝑃𝑓 < −10.
⎩ 1
− 𝑃𝑓
The power law scheme overcomes the inaccuracies of the hybrid scheme, while improving on the computational
time for the exponential scheme.
Warning: VanLeerConvectionTerm not mentioned and no discussion of explicit forms.
All of the numerical schemes presented here are available in FiPy and can be selected by the user.
6.5. Numerical Schemes 35


Chapter 7
Design and Implementation
The goal of FiPy is to provide a highly customizable, open source code for modeling problems involving coupled sets
of PDEs. FiPy allows users to select and customize modules from within the framework. FiPy has been developed to
address model problems in materials science such as poly-crystals, dendritic growth and electrochemical deposition.
These applications all contain various combinations of PDEs with differing forms in conjunction with other unusual
physics (over varying length scales) and unique solution procedures. The philosophy of FiPy is to enable customization
while providing a library of efficient modules for common objects and data types.
7.1 Design
7.1.1 Numerical Approach
The solution algorithms given in the FiPy examples involve combining sets of PDEs while tracking an interface
where the parameters of the problem change rapidly. The phase field method and the level set method are specialized
techniques to handle the solution of PDEs in conjunction with a deforming interface. FiPy contains several examples
of both methods.
FiPy uses the well-known Finite Volume Method (FVM) to reduce the model equations to a form tractable to linear
solvers.
7.1.2 Object Oriented Structure
FiPy is programmed in an object-oriented manner. The benefit of object oriented programming mainly lies in encap-
sulation and inheritance. Encapsulation refers to the tight integration between certain pieces of data and methods that
act on that data. Encapsulation allows parts of the code to be separated into clearly defined independent modules that
can be re-applied or extended in new ways. Inheritance allows code to be reused, overridden, and new capabilities
to be added without altering the original code. An object is treated by its users as an abstraction; the details of its
implementation and behavior are internal.
7.1.3 Test Based Development
FiPy has been developed with a large number of test cases. These test cases are in two categories. The lower level tests
operate on the core modules at the individual method level. The aim is that every method within the core installation
has a test case. The high level test cases operate in conjunction with example solutions and serve to test global solution
algorithms and the interaction of various modules.
With this two-tiered battery of tests, at any stage in code development, the test cases can be executed and errors can
be identified. A comprehensive test base provides reassurance that any code breakages will be clearly demonstrated
with a broken test case. A test base also aids dissemination of the code by providing simple examples and knowledge
of whether the code is working on a particular computer environment.
37
7.1.4 Open Source
In recent years, there has been a movement to release software under open source and associated unrestrictied licenses,
especially within the scientific community. These licensing terms allow users to develop their own applications with
complete access to the source code and then either contribute back to the main source repository or freely distribute
their new adapted version.
As a product of the National Institute of Standards and Technology, the FiPy framework is placed in the public domain
as a matter of U. S. Federal law. Furthermore, FiPy is built upon existing open source tools. Others are free to use
FiPy as they see fit and we welcome contributions to make FiPy better.
7.1.5 High-Level Scripting Language
Programming languages can be broadly lumped into two categories: compiled languages and interpreted (or scripting)
languages. Compiled languages are converted from a human-readable text source file to a machine-readable binary
application file by a sequence of operations generally referred to as “compiling” and “linking.” The binary application
can then be run as many times as desired, but changes will provoke a new cycle of compiling and linking. Interpreted
languages are converted from human-readable to machine-readable on the fly, each time the script is executed. Because
the conversion happens every time 1 , interpreted code is usually slower when running than compiled code. On the other
hand, code development and debugging tends to be much easier and fluid when it’s not necessary to wait for compile
and link cycles after every change. Furthermore, because the conversion happens in real time, it is possible to have
interactive sessions in a scripting language that are not generally possible in compiled languages.
Another distinction, somewhat orthogonal, but closely related, to that between compiled and interpreted languages,
is between low-level languages and high-level languages. Low-level languages describe actions in simple terms that
are closer to the way the computer actually functions. High-level languages describe actions in more complex and
abstract terms that are closer to the way the programmer thinks about the problem at hand. This increased complexity
in the meaning of an expression renders simpler code, because the details of the implementation are hidden away in
the language internals or in an external library. For example, a low-level matrix multiplication written in C might be
rendered as
if (Acols != Brows)
error "these matrix shapes cannot be multiplied";
C = (float *) malloc(sizeof(float) * Bcols * Arows);
for (i = 0; i < Bcols; i++) {

for (j = 0; j < Arows; j++) {
C[i][j] = 0;
for (k = 0; k < Acols; k++) {
C[i][j] += A[i][k] * B[k][j];
}
}
}
Note that the dimensions of the arrays must be supplied externally, as C provides no intrinsic mechanism for deter-
mining the shape of an array. An equivalent high-level construction might be as simple as
C = A * B
All of the error checking, dimension measuring, and space allocation is handled automatically by low-level code that
is intrinsic to the high-level matrix multiplication operator. The high-level code “knows” that matrices are involved,
how to get their shapes, and to interpret ‘*‘ as a matrix multiplier instead of an arithmetic one. All of this allows the
programmer to think about the operation of interest and not worry about introducing bugs in low-level code that is not
unique to their application.
1 ... neglecting such common optimizations as byte-code interpreters.
38 Chapter 7. Design and Implementation

Although it needn’t be true, for a variety of reasons, compiled languages tend to be low-level and interpreted languages
tend to be high-level. Because low-level languages operate closer to the intrinsic “machine language” of the computer,
they tend to be faster at running a given task than high-level languages, but programs written in them take longer to
write and debug. Because running performance is a paramount concern, most scientific codes are written in low-level
compiled languages like FORTRAN or C.
A rather common scenario in the development of scientific codes is that the first draft hard-codes all of the problem
parameters. After a few (hundred) iterations of recompiling and relinking the application to explore changes to the
parameters, code is added to read an input file containing a list of numbers. Eventually, the point is reached where it is
impossible to remember which parameter comes in which order or what physical units are required, so code is added
to, for example, interpret a line beginning with ‘#‘ as a comment. At this point, the scientist has begun developing a
scripting language without even knowing it. Unfortunately for them, very few scientists have actually studied computer
science or actually know anything about the design and implementation of script interpreters. Even if they have the
expertise, the time spent developing such a language interpreter is time not spent actually doing research.
In contrast, a number of very powerful scripting languages, such as Tcl, Java, Python, Ruby, and even the vener-
able BASIC, have open source interpreters that can be embedded directly in an application, giving scientific codes
immediate access to a high-level scripting language designed by someone who actually knew what they were doing.
We have chosen to go a step further and not just embed a full-fledged scripting language in the FiPy framework,
but instead to design the framework from the ground up in a scripting language. While runtime performance is
unquestionably important, many scientific codes are run relatively little, in proportion to the time spent developing
them. If a code can be developed in a day instead of a month, it may not matter if it takes another day to run instead of
an hour. Furthermore, there are a variety of mechanisms for diagnosing and optimizing those portions of a code that
are actually time-critical, rather than attempting to optimize all of it by using a language that is more palatable to the
computer than to the programmer. Thus FiPy, rather than taking the approach of writing the fast numerical code first
and then dealing with the issue of user interaction, initially implements most modules in high-level scripting language
and only translates to low-level compiled code those portions that prove inefficient.
7.1.6 Python Programming Language
Acknowledging that several scripting languages offer a number, if not all, of the features described above, we have
selected Python for the implementation of FiPy. Python is
• an interpreted language that combines remarkable power with very clear syntax,
• freely usable and distributable, even for commercial use,
• fully object oriented,
• distributed with powerful automated testing tools (doctest, unittest),
• actively used and extended by other scientists and mathemeticians (SciPy, NumPy, ScientificPython, PySparse).
• easily integrated with low-level languages such as C (weave, blitz, PyRex).
7.2 Implementation
The Python classes that make up FiPy are described in detail in fipy Package Documentation, but we give a brief
overview here. FiPy is based around three fundamental Python classes: Mesh, Variable, and Term. Using the
terminology of Theoretical and Numerical Background:
A Mesh object represents the domain of interest. FiPy contains many different specific mesh classes to describe
different geometries.
7.2. Implementation 39
A Variable object represents a quantity or field that can change during the problem evolution. A particular type of
Variable, called a CellVariable, represents 𝜑 at the centers of the cells of the Mesh. A CellVariable
describes the values of the field 𝜑, but it is not concerned with their geometry; that role is taken by the Mesh.
An important property of Variable objects is that they can describe dependency relationships, such that:
>>> a = Variable(value = 3)
>>> b = a * 4
does not assign the value 12 to b, but rather it assigns a multiplication operator object to b, which depends on
the Variable object a:
>>> b
(Variable(value = 3) * 4)
>>> a.setValue(5)
>>> b
(Variable(value = 5) * 4)
The numerical value of the Variable is not calculated until it is needed (a process known as “lazy evaluation”):
>>> print b
20
A Term object represents any of the terms in Equation (??) or any linear combination of such terms. Early in the
development of FiPy, a distinction was made between Equation objects, which represented all of Equation
(??), and Term objects, which represented the individual terms in that equation. The Equation object has
since been eliminated as redundant. Term objects can be single entities such as a DiffusionTerm or a linear
combination of other Term objects that build up to form an expression such as Equation (??).
Beyond these three fundamental classes of Mesh, Variable, and Term, FiPy is composed of a number of related
classes.
Cell
Mesh Variable Viewer
Face
Vertex Solver
Term
BoundaryCondition
SparseMatrix
Fig. 7.1: Primary object relationships in FiPy.
A Mesh object is composed of cells. Each cell is defined by its bounding faces and each face is defined by its
bounding vertices. A Term object encapsulates the contributions to the _SparseMatrix that defines the solution
of an equation. BoundaryCondition objects are used to describe the conditions on the boundaries of the Mesh,
and each Term interprets the BoundaryCondition objects as necessary to modify the _SparseMatrix. An
equation constructed from Term objects can apply a unique Solver to invert its _SparseMatrix in the most
expedient and stable fashion. At any point during the solution, a Viewer can be invoked to display the values of the
solved Variable objects.
At this point, it will be useful to examine some of the example problems in Examples. More classes are introduced in
the examples, along with illustrations of their instantiation and use.
40 Chapter 7. Design and Implementation

Chapter 8
Frequently Asked Questions
8.1 How do I represent an equation in FiPy?
As explained in Theoretical and Numerical Background, the canonical governing equation that can be solved by FiPy
for the dependent CellVariable 𝜑 is
𝜕(𝜌𝜑) 𝑛
+ ∇ · (⃗𝑢𝜑) = [∇ · (Γ𝑖 ∇)] 𝜑 + 𝑆𝜑
𝜕𝑡
⏟ ⏞ ⏟ ⏞ ⏟ ⏞ ⏟ ⏞
transient convection diffusion source
and the individual terms are discussed in Discretization.

A physical problem can involve many different coupled governing equations, one for each variable. Numerous specific
examples are presented in Part Examples.
8.1.1 Is there a way to model an anisotropic diffusion process or more generally to

represent the diffusion coefficient as a tensor so that the diffusion term takes
the form 𝜕𝑖 Γ𝑖𝑗 𝜕𝑗 𝜑?
Terms of the form 𝜕𝑖 Γ𝑖𝑗 𝜕𝑗 𝜑 can be posed in FiPy by using a list, tuple rank 1 or rank 2 FaceVariable to represent
a vector or tensor diffusion coefficient. For example, if we wished to represent a diffusion term with an anisotropy
ratio of 5 aligned along the x-coordinate axis, we could write the term as,
>>> DiffusionTerm([[[5, 0], [0, 1]]])
which represents 5𝜕𝑥2 + 𝜕𝑦2 . Notice that the tensor, written in the form of a list, is contained within a list. This is
because the first index of the list refers to the order of the term not the first index of the tensor (see Higher order
diffusion). This notation, although succinct can sometimes be confusing so a number of cases are interpreted below.
>>> DiffusionTerm([[5, 1]])
This represents the same term as the case examined above. The vector notation is just a short-hand representation for
the diagonal of the tensor. Off-diagonals are assumed to be zero.
>>> DiffusionTerm([5, 1])
)︀2
This simply represents a fourth order isotropic diffusion term of the form 5 𝜕𝑥2 + 𝜕𝑦2 .
(︀
>>> DiffusionTerm([[1, 0], [0, 1]])
Nominally, this should represent a fourth order diffusion term of the form 𝜕𝑥2 𝜕𝑦2 , but FiPy does not currently support
anisotropy for higher order diffusion terms so this may well throw an error or give anomalous results.
41
>>> x, y = mesh.cellCenters
>>> DiffusionTerm(CellVariable(mesh=mesh,
... value=[[x**2, x * y], [-x * y, -y**2]])
This represents an anisotropic diffusion coefficient that varies spatially so that the term has the form 𝜕𝑥 (𝑥2 𝜕𝑥 +𝑥𝑦𝜕𝑦 )+
𝜕𝑦 (−𝑥𝑦𝜕𝑥 − 𝑦 2 𝜕𝑦 ) ≡ 𝑥𝜕𝑥 − 𝑦𝜕𝑦 + 𝑥2 𝜕𝑥2 − 𝑦 2 𝜕𝑦2 .
Generally, anisotropy is not conveniently aligned along the coordinate axes; in these cases, it is necessary to apply
a rotation matrix in order to calculate the correct tensor values, see examples.diffusion.anisotropy for
details.
8.1.2 How do I represent a ... term that doesn’t involve the dependent variable?
It is important to realize that, even though an expression may superficially resemble one of those shown in Discretiza-
tion, if the dependent variable for that PDE does not appear in the appropriate place, then that term should be treated
as a source.
How do I represent a diffusive source?
If the governing equation for 𝜑 is
𝜕𝜑
= ∇ · (𝐷1 ∇𝜑) + ∇ · (𝐷2 ∇𝜉)
𝜕𝑡
then the first term is a TransientTerm and the second term is a DiffusionTerm, but the third term is simply an
explicit source, which is written in Python as
>>> (D2 * xi.faceGrad).divergence
Higher order diffusive sources can be obtained by simply nesting the references to faceGrad and divergence.
Note: We use faceGrad, rather than grad, in order to obtain a second-order spatial discretization of the diffusion
term in 𝜉, consistent with the matrix that is formed by DiffusionTerm for 𝜑.
How do I represent a convective source?
The convection of an independent field 𝜉 as in
𝜕𝜑
= ∇ · (⃗𝑢𝜉)
𝜕𝑡
can be rendered as
>>> (u * xi.arithmeticFaceValue).divergence
when ⃗𝑢 is a rank-1 FaceVariable (preferred) or as

>>> (u * xi).divergence
if ⃗𝑢 is a rank-1 CellVariable.
42 Chapter 8. Frequently Asked Questions

How do I represent a transient source?
The time-rate-of change of an independent variable 𝜉, such as in
𝜕(𝜌1 𝜑) 𝜕(𝜌2 𝜉)
=
𝜕𝑡 𝜕𝑡
does not have an abstract form in FiPy and should be discretized directly, in the manner of Equation (??), as
>>> TransientTerm(coeff=rho1) == rho2 * (xi - xi.old) / timeStep
This technique is used in examples.phase.anisotropy.
8.1.3 What if my term involves the dependent variable, but not where FiPy puts it?
Frequently, viewing the term from a different perspective will allow it to be cast in one of the canonical forms. For
example, the third term in
𝜕𝜑
= ∇ · (𝐷1 ∇𝜑) + ∇ · (𝐷2 𝜑∇𝜉)
𝜕𝑡
might be considered as the diffusion of the independent variable 𝜉 with a mobility 𝐷2 𝜑 that is a function of the
dependent variable 𝜑. For FiPy‘s purposes, however, this term represents the convection of 𝜑, with a velocity 𝐷2 ∇𝜉,
due to the counter-diffusion of 𝜉, so
>>> eq = TransientTerm() == (DiffusionTerm(coeff=D1)
... + <Specific>ConvectionTerm(coeff=D2 * xi.faceGrad))
Note: With the advent of Coupled and Vector Equations in FiPy 3.x, it is now possible to represent both terms with
DiffusionTerm.
8.1.4 What if the coefficient of a term depends on the variable that I’m solving for?
A non-linear coefficient, such as the diffusion coefficient in ∇ · [Γ1 (𝜑)∇𝜑] = ∇ · [Γ0 𝜑(1 − 𝜑)∇𝜑] is not a problem
for FiPy. Simply write it as it appears:
>>> diffTerm = DiffusionTerm(coeff=Gamma0 * phi * (1 - phi))
Note: Due to the nonlinearity of the coefficient, it will probably be necessary to “sweep” the solution to convergence
as discussed in Iterations, timesteps, and sweeps? Oh, my!.
8.2 How can I see what I’m doing?
8.2.1 How do I export data?
The way to save your calculations depends on how you plan to make use of the data. If you want to save it
for “restart” (so that you can continue or redirect a calculation from some intermediate stage), then you’ll want
to “pickle” the Python data with the dump module. This is illustrated in examples.phase.anisotropy,
examples.phase.impingement.mesh40x1, examples.phase.impingement.mesh20x20, and
examples.levelSet.electroChem.howToWriteAScript.
8.2. How can I see what I’m doing? 43

On the other hand, pickled FiPy data is of little use to anything besides Python and FiPy. If you want to import your
calculations into another piece of software, whether to make publication-quality graphs or movies, or to perform some
analysis, or as input to another stage of a multiscale model, then you can save your data as an ASCII text file of
tab-separated-values with a TSVViewer. This is illustrated in examples.diffusion.circle.
8.2.2 How do I save a plot image?
Some of the viewers have a button or other mechanism in the user interface for saving an image file. Also, you can
supply an optional keyword filename when you tell the viewer to plot(), e.g.
>>> viewer.plot(filename="myimage.ext")
which will save a file named myimage.ext in your current working directory. The type of image is determined by
the file extension “.ext”. Different viewers have different capabilities:
Matplotlib accepts “.eps,” “.jpg” (Joint Photographic Experts Group), and “.png” (Portable Network Graphics).
Attention: Actually, Matplotlib supports different extensions, depending on the chosen backend, but our
MatplotlibViewer classes don’t properly support this yet.
What if I only want the saved file, with no display on screen?
To our knowledge, this is only supported by Matplotlib, as is explained in the Matplotlib FAQ on image backends.
Basically, you need to tell Matplotlib to use an “image backend,” such as “Agg” or “Cairo.” Backends are discussed
at http://matplotlib.sourceforge.net/backends.html.
8.2.3 How do I make a movie?
FiPy has no facilities for making movies. You will need to save individual frames (see the previous question) and then
stitch them together into a movie, using one of a variety of different free, shareware, or commercial software packages.
The guidance in the Matplotlib FAQ on movies should be adaptable to other Viewers.
8.2.4 Why doesn’t the Viewer look the way I want?
FiPy‘s viewers are utilitarian. They’re designed to let you see something with a minimum of effort. Because different
plotting packages have different capabilities and some are easier to install on some platforms than on others, we have
tried to support a range of Python plotters with a minimal common set of features. Many of these packages are capable
of much more, however. Often, you can invoke the Viewer you want, and then issue supplemental commands for
the underlying plotting package. The better option is to make a “subclass” of the FiPy Viewer that comes closest to
producing the image you want. You can then override just the behavior you wan to change, while letting FiPy do most
of the heavy lifting. See examples.phase.anisotropy and examples.phase.polyxtal for examples of
creating a custom Matplotlib Viewer class; see examples.cahnHilliard.sphere for an example of creating
a custom Mayavi Viewer class.
8.3 Iterations, timesteps, and sweeps? Oh, my!
Any non-linear solution of partial differential equations is an approximation. These approximations benefit from
repetetive solution to achieve the best possible answer. In FiPy (and in many similar PDE solvers), there are three
layers of repetition.

iterations This is the lowest layer of repetition, which you’ll generally need to spend the least time thinking about.
FiPy solves PDEs by discretizing them into a set of linear equations in matrix form, as explained in Discretiza-
tion and Linear Equations. It is not always practical, or even possible, to exactly solve these matrix equations
on a computer. FiPy thus employs “iterative solvers”, which make successive approximations until the linear
equations have been satisfactorily solved. FiPy chooses a default number of iterations and solution tolerance,
which you will not generally need to change. If you do wish to change these defaults, you’ll need to create a
new Solver object with the desired number of iterations and solution tolerance, e.g.
>>> mySolver = LinearPCGSolver(iterations=1234, tolerance=5e-6)
:
:
>>> eq.solve(..., solver=mySolver, ...)
Note: The older Solver steps= keyword is now deprecated in favor of iterations= to make the role
clearer.
Solver iterations are changed from their defaults in examples.flow.stokesCavity and

examples.updating.update0_1to1_0.
sweeps This middle layer of repetition is important when a PDE is non-linear (e.g., a diffusivity that depends on
concentration) or when multiple PDEs are coupled (e.g., if solute diffusivity depends on temperature and thermal
conductivity depends on concentration). Even if the Solver solves the linear approximation of the PDE to
absolute perfection by performing an infinite number of iterations, the solution may still not be a very good
representation of the actual non-linear PDE. If we resolve the same equation at the same point in elapsed time,
but use the result of the previous solution instead of the previous timestep, then we can get a refined solution to
the non-linear PDE in a process known as “sweeping.”
Note: Despite references to the “previous timestep,” sweeping is not limited to time-evolving problems. Non-
linear sets of quasi-static or steady-state PDEs can require sweeping, too.
We need to distinguish between the value of the variable at the last timestep and the value of the variable at the
last sweep (the last cycle where we tried to solve the current timestep). This is done by first modifying the way
the variable is created:
>>> myVar = CellVariable(..., hasOld=True)
and then by explicitly moving the current value of the variable into the “old” value only when we want to:
>>> myVar.updateOld()
Finally, we will need to repeatedly solve the equation until it gives a stable result. To clearly distinguish that a
single cycle will not truly “solve” the equation, we invoke a different method “sweep():
>>> for sweep in range(sweeps):
... eq.sweep(var=myVar, ...)
Even better than sweeping a fixed number of cycles is to do it until the non-linear PDE has been solved satisfac-
torily:
>>> while residual > desiredResidual:
... residual = eq.sweep(var=myVar, ...)
Sweeps are used to achieve better solutions in examples.diffusion.mesh1D,

examples.phase.simple, examples.phase.binaryCoupled, and
examples.flow.stokesCavity.
timesteps This outermost layer of repetition is of most practical interest to the user. Understanding the time evolu-
tion of a problem is frequently the goal of studying a particular set of PDEs. Moreover, even when only an
8.3. Iterations, timesteps, and sweeps? Oh, my! 45

equilibrium or steady-state solution is desired, it may not be possible to simply solve that directly, due to non-
linear coupling between equations or to boundary conditions or initial conditions. Some types of PDEs have
fundamental limits to how large a timestep they can take before they become either unstable or inaccurate.
Note: Stability and accuracy are distinctly different. An unstable solution is often said to “blow up”, with
radically different values from point to point, often diverging to infinity. An inaccurate solution may look
perfectly reasonable, but will disagree significantly from an analytical solution or from a numerical solution
obtained by taking either smaller or larger timesteps.
For all of these reasons, you will frequently need to advance a problem in time and to choose an appropriate
interval between solutions. This can be simple:
>>> timeStep = 1.234e-5
>>> for step in range(steps):
... eq.solve(var=myVar, dt=timeStep, ...)
or more elaborate:
>>> timeStep = 1.234e-5
>>> elapsedTime = 0
>>> while elapsedTime < totalElapsedTime:
... eq.solve(var=myVar, dt=timeStep, ...)
... elapsedTime += timeStep
... timeStep = SomeFunctionOfVariablesAndTime(myVar1, myVar2, elapedTime)
A majority of the examples in this manual illustrate time evolving behavior. Notably, boundary conditions
are made a function of elapsed time in examples.diffusion.mesh1D. The timestep is chosen based on
the expected interfacial velocity in examples.phase.simple. The timestep is gradually increased as the
kinetics slow down in examples.cahnHilliard.mesh2DCoupled.
Finally, we can (and often do) combine all three layers of repetition:
>>> myVar = CellVariable(..., hasOld=1)
:
:
>>> mySolver = LinearPCGSolver(iterations=1234, tolerance=5e-6)
:
:
>>> while elapsedTime < totalElapsedTime:
... myVar.updateOld()
... while residual > desiredResidual:
... residual = eq.sweep(var=myVar, dt=timeStep, ...)
... elapsedTime += timeStep
8.4 Why the distinction between CellVariable and FaceVariable

coefficients?
FiPy solves field variables on the cell centers. Transient and source terms describe the change in the value of a field at
the cell center, and so they take a CellVariable coefficient. Diffusion and convection terms involve fluxes between
cell centers, and are calculated on the face between two cells, and so they take a FaceVariable coefficient.
Note: If you supply a CellVariable var when a FaceVariable is expected, FiPy will automatically substitute
var.arithmeticFaceValue. This can have undesirable consequences, however. For one thing, the arithmetic
face average of a non-linear function is not the same as the same non-linear function of the average argument, e.g., for

𝑓 (𝑥) = 𝑥2 ,
1+2 9 𝑓 (1) + 𝑓 (2) 5

𝑓( ) = ̸= =
2 4 2 2
This distinction is not generally important for smoothly varying functions, but can dramatically affect the solution
when sharp changes are present. Also, for many problems, such as a conserved concentration field that cannot be
allowed to drop below zero, a harmonic average is more appropriate than an arithmetic average.
If you experience problems (unstable or wrong results, or excessively small timesteps), you may need to explicitly
supply the desired FaceVariable rather than letting FiPy assume one.
8.5 How do I represent boundary conditions?
See the Boundary Conditions section for more details.
8.6 What does this error message mean?
ValueError: frames are not aligned This error most likely means that you have provided a
CellVariable when FiPy was expecting a FaceVariable (or vice versa).
MA.MA.MAError: Cannot automatically convert masked array to Numeric because data is masked
This not-so-helpful error message could mean a number of things, but the most likely explanation is that the
solution has become unstable and is diverging to ±∞. This can be caused by taking too large a timestep or by
using explicit terms instead of implicit ones.
repairing catalog by removing key This message (not really an error, but may cause test failures) can
result when using the scipy.weave package via the --inline flag. It is due to a bug in SciPy that
has been patched in their source repository: http://www.scipy.org/mailinglists/mailman?fn=scipy-dev/2005-
June/003010.html.
numerix Numeric 23.6 This is neither an error nor a warning. It’s just a sloppy message left in SciPy:
http://thread.gmane.org/gmane.comp.python.scientific.user/4349.
8.7 How do I change FiPy’s default behavior?
FiPy tries to make reasonable choices, based on what packages it finds installed, but there may be times that you wish
to override these behaviors. See the Command-line Flags and Environment Variables section for more details.
8.8 How can I tell if I’m running in parallel?
See Solving in Parallel.
8.9 Why don’t my scripts work anymore?
FiPy has experienced three major API changes. The steps necessary to upgrade older scripts are discussed in Updating
FiPy.
8.5. How do I represent boundary conditions? 47

8.10 What if my question isn’t answered here?
Please post your question to the mailing list <http://www.ctcms.nist.gov/fipy/mail.html> or file an issue at
<https://github.com/usnistgov/fipy/issues/new>.

Chapter 9
Glossary
Buildbot The BuildBot is a system to automate the compile/test cycle required by most software projects to validate
code changes. See http://trac.buildbot.net/.
FiPy The eponymous software package. See http://www.ctcms.nist.gov/fipy.
Gmsh A free and Open Source 3D (and 2D!) finite element grid generator. It also has a CAD engine and post-
processor that FiPy does not make use of. See http://www.geuz.org/gmsh.
IPython An improved Python shell that integrates nicely with Matplotlib. See http://ipython.scipy.org/.
Matplotlib matplotlib Python package displays publication quality results. It displays both 1D X-Y type plots
and 2D contour plots for structured data. It does not display unstructured 2D data or 3D data. It works on
all common platforms and produces publication quality hard copies. See http://matplotlib.sourceforge.net and
Matplotlib.
Mayavi The mayavi Data Visualizer is a free, easy to use scientific data visualizer. It displays 1D, 2D and 3D data.
It is the only FiPy viewer available for 3D data. Other viewers are probably better for 1D or 2D viewing. See
http://code.enthought.com/projects/mayavi and Mayavi.
MayaVi The predecessor to Mayavi. Yes, it’s confusing.
numarray An archaic predecessor to NumPy.
Numeric An archaic predecessor to NumPy.
NumPy The numpy Python package provides array arithmetic facilities. See http://www.scipy.org/NumPy.
pip “pip installs python” is a tool for installing and managing Python packages, such as those found in PyPI. See
http://www.pip-installer.org.
PyAMG A suite of python-based preconditoners. See http://code.google.com/p/pyamg/ and PyAMG.
PyPI The Python Package Index is a repository of software for the Python programming language. See
http://pypi.python.org/pypi.
Pyrex A mechanism for mixing C and Python code. See http://www.cosc.canterbury.ac.nz/greg.ewing/python/Pyrex/.
PySparse The pysparse Python package provides sparse matrix storage, solvers, and linear algebra routines. See
http://pysparse.sourceforge.net and PySparse.
Python The programming language that FiPy (and your scripts) are written in. See http://www.python.org/.
Python 3 The (likely) future of the Python programming language. Third-party packages are slowly being adapted,
but many that FiPy uses are not yet available. See http://docs.python.org/py3k/ and PEP 3000.
PyTrilinos Python wrapper for Trilinos. See http://trilinos.sandia.gov/packages/pytrilinos/.
PyxViewer A now defunct python viewer.
ScientificPython A collection of useful utilities for scientists. See http://dirac.cnrs-
orleans.fr/plone/software/scientificpython.
49
SciPy The scipy package provides a wide range of scientific and mathematical operations. FiPy can use
scipy.weave for enhanced performance with C language inlining and Scipy‘s solver suite for linear solu-
tions. See http://www.scipy.org/. and SciPy.
Sphinx The tools used to generate the FiPy documentation. See http://sphinx.pocoo.org/.
Trilinos This package provides sparse matrix storage, solvers, and preconditioners, and can be used instead of PyS-
parse. Trilinos preconditioning allows for iterative solutions to some difficult problems that PySparse cannot
solve. See http://trilinos.sandia.gov and Trilinos.
50 Chapter 9. Glossary
Part II
Examples
51
Note: Any given module “example.something.input” can be found in the file

“examples/something/input.py”.
These examples can be used in at least four ways:

• Each example can be invoked individually to demonstrate an application of FiPy:
$ python examples/something/input.py
• Each example can be invoked such that when it has finished running, you will be left in an interactive Python
shell:
$ python -i examples/something/input.py
At this point, you can enter Python commands to manipulate the model or to make queries about the example’s
variable values. For instance, the interactive Python sessions in the example documentation can be typed in
directly to see that the expected results are obtained.
• Alternatively, these interactive Python sessions, known as doctest blocks, can be invoked as automatic tests:
$ python setup.py test --examples
In this way, the documentation and the code are always certain to be consistent.
• Finally, and most importantly, the examples can be used as a templates to design your own problem scripts.
Note: The examples shown in this manual have been written with particular emphasis on serving as
both documentation and as comprehensive tests of the FiPy framework. As explained at the end of
examples/diffusion/steadyState/mesh1D.py, your own scripts can be much more succint, if you
wish, and include only the text that follows the “>>>” and “...” prompts shown in these examples.
To obtain a copy of an example, containing just the script instructions, type:
$ python setup.py copy_script --From x.py --To y.py
In addition to those presented in this manual, there are dozens of other files in the examples/ directory, that demon-
strate other uses of FiPy. If these examples do not help you construct your own problem scripts, please contact us.
53
54
Chapter 10
Diffusion Examples
examples.diffusion.mesh1D Solve a one-dimensional diffusion equation under different conditi

examples.diffusion.coupled Solve the biharmonic equation as a coupled pair of diffusion equat
examples.diffusion.mesh20x20 Solve a two-dimensional diffusion problem in a square domain.
examples.diffusion.circle Solve the diffusion equation in a circular domain meshed with trian
examples.diffusion.electrostatics Solve the Poisson equation in one dimension.
examples.diffusion.nthOrder.input4thOrder1D Solve a fourth-order diffusion problem.
examples.diffusion.anisotropy Solve the diffusion equation with an anisotropic diffusion coefficie
10.1 examples.diffusion.mesh1D
Solve a one-dimensional diffusion equation under different conditions.

To run this example from the base FiPy directory, type:
$ python examples/diffusion/mesh1D.py
at the command line. Different stages of the example should be displayed, along with prompting messages in the
terminal.
This example takes the user through assembling a simple problem with FiPy. It describes different approaches to a 1D
diffusion problem with constant diffusivity and fixed value boundary conditions such that,
𝜕𝜑
= 𝐷∇2 𝜑. (10.1)
𝜕𝑡
The first step is to define a one dimensional domain with 50 solution points. The Grid1D object represents a linear
structured grid. The parameter dx refers to the grid spacing (set to unity here).
>>> from fipy import *
>>> nx = 50
>>> dx = 1.
>>> mesh = Grid1D(nx=nx, dx=dx)
FiPy solves all equations at the centers of the cells of the mesh. We thus need a CellVariable object to hold the
values of the solution, with the initial condition 𝜑 = 0 at 𝑡 = 0,
>>> phi = CellVariable(name="solution variable",
... mesh=mesh,
... value=0.)
We’ll let
55
>>> D = 1.
for now.
The set of boundary conditions are given to the equation as a Python tuple or list (the distinction is not generally
important to FiPy). The boundary conditions
{︃
0 at 𝑥 = 1,
𝜑=
1 at 𝑥 = 0.
are formed with a value

>>> valueLeft = 1
>>> valueRight = 0
and a set of faces over which they apply.
Note: Only faces around the exterior of the mesh can be used for boundary conditions.
For example, here the exterior faces on the left of the domain are extracted by mesh.facesLeft. The boundary
conditions is applied using phi. constrain() with tthese faces and a value (valueLeft).
>>> phi.constrain(valueRight, mesh.facesRight)
>>> phi.constrain(valueLeft, mesh.facesLeft)
Note: If no boundary conditions are specified on exterior faces, the default boundary condition is equivalent to a zero
gradient, equivalent to ⃗𝑛 · ∇𝜑|someFaces = 0.
If you have ever tried to numerically solve Eq. (10.1), you most likely attempted “explicit finite differencing” with
code something like:
for step in range(steps):
for j in range(cells):
phi_new[j] = phi_old[j] \
+ (D * dt / dx**2) * (phi_old[j+1] - 2 * phi_old[j] + phi_old[j-1])
time += dt
plus additional code for the boundary conditions. In FiPy, you would write
>>> eqX = TransientTerm() == ExplicitDiffusionTerm(coeff=D)
The largest stable timestep that can be taken for this explicit 1D diffusion problem is ∆𝑡 ≤ ∆𝑥2 /(2𝐷).
We limit our steps to 90% of that value for good measure
>>> timeStepDuration = 0.9 * dx**2 / (2 * D)
>>> steps = 100
If we’re running interactively, we’ll want to view the result, but not if this example is being run automatically as a
test. We accomplish this by having Python check if this script is the “__main__” script, which will only be true if
we explicitly launched it and not if it has been imported by another script such as the automatic tester. The factory
function Viewer() returns a suitable viewer depending on available viewers and the dimension of the mesh.
>>> phiAnalytical = CellVariable(name="analytical value",
... mesh=mesh)
>>> if __name__ == '__main__':

... viewer = Viewer(vars=(phi, phiAnalytical),
56 Chapter 10. Diffusion Examples

... datamin=0., datamax=1.)

... viewer.plot()
√
In a semi-infinite domain, the analytical solution for this transient diffusion problem is given by 𝜑 = 1−erf(𝑥/2 𝐷𝑡).
If the SciPy library is available, the result is tested against the expected profile:
>>> x = mesh.cellCenters[0]
>>> t = timeStepDuration * steps
>>> try:
... from scipy.special import erf
... phiAnalytical.setValue(1 - erf(x / (2 * numerix.sqrt(D * t))))
... except ImportError:
... print "The SciPy library is not available to test the solution to \
... the transient diffusion equation"
We then solve the equation by repeatedly looping in time:

... eqX.solve(var=phi,
... dt=timeStepDuration)
... if __name__ == '__main__':
... viewer.plot()
>>> print phi.allclose(phiAnalytical, atol = 7e-4)

1
>>> if __name__ == '__main__':

... raw_input("Explicit transient diffusion. Press <return> to proceed...")
1
solution variable
analytical value
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50
10.1. examples.diffusion.mesh1D 57
Although explicit finite differences are easy to program, we have just seen that this 1D transient diffusion problem is
limited to taking rather small time steps. If, instead, we represent Eq. (10.1) as:
phi_new[j] = phi_old[j] \
+ (D * dt / dx**2) * (phi_new[j+1] - 2 * phi_new[j] + phi_new[j-1])
it is possible to take much larger time steps. Because phi_new appears on both the left and right sides of the
equation, this form is called “implicit”. In general, the “implicit” representation is much more difficult to program
than the “explicit” form that we just used, but in FiPy, all that is needed is to write
>>> eqI = TransientTerm() == DiffusionTerm(coeff=D)
reset the problem

>>> phi.setValue(valueRight)
and rerun with much larger time steps

>>> timeStepDuration *= 10
>>> steps //= 10
... eqI.solve(var=phi,
... if __name__ == '__main__':
... viewer.plot()

1
>>> if __name__ == '__main__':

... raw_input("Implicit transient diffusion. Press <return> to proceed...")

1
solution variable
analytical value
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50
Note that although much larger stable timesteps can be taken with this implicit version (there is, in fact, no limit to
how large an implicit timestep you can take for this particular problem), the solution is less accurate. One way to
achieve a compromise between stability and accuracy is with the Crank-Nicholson scheme, represented by:
phi_new[j] = phi_old[j] + (D * dt / (2 * dx**2)) * \
((phi_new[j+1] - 2 * phi_new[j] + phi_new[j-1])
+ (phi_old[j+1] - 2 * phi_old[j] + phi_old[j-1]))
which is essentially an average of the explicit and implicit schemes from above. This can be rendered in FiPy as easily
as
>>> eqCN = eqX + eqI
We again reset the problem

>>> phi.setValue(valueRight)
and apply the Crank-Nicholson scheme until the end, when we apply one step of the fully implicit scheme to drive
down the error (see, e.g., section 19.2 of [22]).
>>> for step in range(steps - 1):
... eqCN.solve(var=phi,
... if __name__ == '__main__':
... viewer.plot()
>>> eqI.solve(var=phi,
>>> if __name__ == '__main__':
... viewer.plot()

1
>>> if __name__ == '__main__':

... raw_input("Crank-Nicholson transient diffusion. Press <return> to proceed...")
As mentioned above, there is no stable limit to how large a time step can be taken for the implicit diffusion problem.
In fact, if the time evolution of the problem is not interesting, it is possible to eliminate the time step altogether by
omitting the TransientTerm. The steady-state diffusion equation
𝐷∇2 𝜑 = 0
is represented in FiPy by
>>> DiffusionTerm(coeff=D).solve(var=phi)
>>> if __name__ == '__main__':

... viewer.plot()
The analytical solution to the steady-state problem is no longer an error function, but simply a straight line, which we
can confirm to a tolerance of 10−10 .
>>> L = nx * dx
>>> print phi.allclose(valueLeft + (valueRight - valueLeft) * x / L,
... rtol = 1e-10, atol = 1e-10)
1
>>> if __name__ == '__main__':

... raw_input("Implicit steady-state diffusion. Press <return> to proceed...")
1
solution variable
analytical value
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50

Often, boundary conditions may be functions of another variable in the system or of time.
For example, to have
{︃
(1 + sin 𝑡)/2 on 𝑥 = 0
𝜑=
0 on 𝑥 = 𝐿
we will need to declare time 𝑡 as a Variable

>>> time = Variable()
and then declare our boundary condition as a function of this Variable

>>> del phi.faceConstraints
>>> valueLeft = 0.5 * (1 + numerix.sin(time))
>>> phi.constrain(0., mesh.facesRight)
>>> eqI = TransientTerm() == DiffusionTerm(coeff=D)
When we update time at each timestep, the left-hand boundary condition will automatically update,
>>> dt = .1
>>> while time() < 15:
... time.setValue(time() + dt)
... eqI.solve(var=phi, dt=dt)
... if __name__ == '__main__':
... viewer.plot()
>>> if __name__ == '__main__':

... raw_input("Time-dependent boundary condition. Press <return> to proceed...")
1
solution variable
analytical value
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50
Many interesting problems do not have simple, uniform diffusivities. We consider a steady-state diffusion problem
∇ · (𝐷∇𝜑) = 0,
with a spatially varying diffusion coefficient

⎧
⎨1
⎪ for 0 < 𝑥 < 𝐿/4,
𝐷 = 0.1 for 𝐿/4 ≤ 𝑥 < 3𝐿/4,
for 3𝐿/4 ≤ 𝑥 < 𝐿,
⎪
1
⎩
and with boundary conditions 𝜑 = 0 at 𝑥 = 0 and 𝐷 𝜕𝜑 𝜕𝑥 = 1 at 𝑥 = 𝐿, where 𝐿 is the length of the solution domain.
Exact numerical answers to this problem are found when the mesh has cell centers that lie at 𝐿/4 and 3𝐿/4, or when
the number of cells in the mesh 𝑁𝑖 satisfies 𝑁𝑖 = 4𝑖 + 2, where 𝑖 is an integer. The mesh we’ve been using thus far is
satisfactory, with 𝑁𝑖 = 50 and 𝑖 = 12.
Because FiPy considers diffusion to be a flux from one cell to the next, through the intervening face, we must define
the non-uniform diffusion coefficient on the mesh faces
>>> D = FaceVariable(mesh=mesh, value=1.0)
>>> X = mesh.faceCenters[0]
>>> D.setValue(0.1, where=(L / 4. <= X) & (X < 3. * L / 4.))
The boundary conditions are a fixed value of

>>> valueLeft = 0.
to the left and a fixed flux of

>>> fluxRight = 1.
to the right:
>>> phi = CellVariable(mesh=mesh)
>>> phi.faceGrad.constrain([fluxRight], mesh.facesRight)
We re-initialize the solution variable

>>> phi.setValue(0)
and obtain the steady-state solution with one implicit solution step
>>> DiffusionTerm(coeff = D).solve(var=phi)
The analytical solution is simply

⎧
⎨𝑥
⎪ for 0 < 𝑥 < 𝐿/4,
𝜑 = 10𝑥 − 9𝐿/4 for 𝐿/4 ≤ 𝑥 < 3𝐿/4,
for 3𝐿/4 ≤ 𝑥 < 𝐿,
⎪
𝑥 + 18𝐿/4
⎩
or
>>> phiAnalytical.setValue(x)
>>> phiAnalytical.setValue(10 * x - 9. * L / 4. ,
... where=(L / 4. <= x) & (x < 3. * L / 4.))
>>> phiAnalytical.setValue(x + 18. * L / 4. ,
... where=3. * L / 4. <= x)
>>> print phi.allclose(phiAnalytical, atol = 1e-8, rtol = 1e-8)
1
And finally, we can plot the result

>>> if __name__ == '__main__':
... Viewer(vars=(phi, phiAnalytical)).plot()
... raw_input("Non-uniform steady-state diffusion. Press <return> to proceed...")
solution variable
300
solution variable
250
200
150
100
50
0
0 10 20 30 40 50
Often, the diffusivity is not only non-uniform, but also depends on the value of the variable, such that
𝜕𝜑
= ∇ · [𝐷(𝜑)∇𝜑]. (10.2)
𝜕𝑡
With such a non-linearity, it is generally necessary to “sweep” the solution to convergence. This means that each
time step should be calculated over and over, using the result of the previous sweep to update the coefficients of
the equation, without advancing in time. In FiPy, this is accomplished by creating a solution variable that explicitly
retains its “old” value by specifying hasOld when you create it. The variable does not move forward in time until it
is explicity told to updateOld(). In order to compare the effects of different numbers of sweeps, let us create a list
of variables: phi[0] will be the variable that is actually being solved and phi[1] through phi[4] will display the
result of taking the corresponding number of sweeps (phi[1] being equivalent to not sweeping at all).
>>> valueLeft = 1.
>>> valueRight = 0.
>>> phi = [
... CellVariable(name="solution variable",
... mesh=mesh,
... value=valueRight,
... hasOld=1),
... CellVariable(name="1 sweep",
... mesh=mesh),
... CellVariable(name="2 sweeps",
... mesh=mesh),
... mesh=mesh),

... mesh=mesh)
... ]
If, for example,
𝐷 = 𝐷0 (1 − 𝜑)
we would simply write Eq. (10.2) as

>>> D0 = 1.
>>> eq = TransientTerm() == DiffusionTerm(coeff=D0 * (1 - phi[0]))
Note: Because of the non-linearity, the Crank-Nicholson scheme does not work for this problem.
We apply the same boundary conditions that we used for the uniform diffusivity cases
>>> phi[0].constrain(valueRight, mesh.facesRight)
>>> phi[0].constrain(valueLeft, mesh.facesLeft)
Although this problem does not have an exact transient solution, it can be solved in steady-state, with
√︂
𝑥
𝜑(𝑥) = 1 −
𝐿
>>> phiAnalytical.setValue(1. - numerix.sqrt(x/L))
We create a viewer to compare the different numbers of sweeps with the analytical solution from before.
>>> if __name__ == '__main__':
... viewer = Viewer(vars=phi + [phiAnalytical],
... viewer.plot()
As described above, an inner “sweep” loop is generally required for the solution of non-linear or multiple equation
sets. Often a conditional is required to exit this “sweep” loop given some convergence criteria. Instead of using
the solve() method equation, when sweeping, it is often useful to call sweep() instead. The sweep() method
behaves the same way as solve(), but returns the residual that can then be used as part of the exit condition.
We now repeatedly run the problem with increasing numbers of sweeps.
>>> for sweeps in range(1,5):
... phi[0].setValue(valueRight)
... for step in range(steps):
... # only move forward in time once per time step
... phi[0].updateOld()
...
... # but "sweep" many times per time step
... for sweep in range(sweeps):
... res = eq.sweep(var=phi[0],
... if __name__ == '__main__':
... viewer.plot()
...
... # copy the final result into the appropriate display variable
... phi[sweeps].setValue(phi[0])
... if __name__ == '__main__':
... viewer.plot()
... raw_input("Implicit variable diffusity. %d sweep(s). \

... Residual = %f. Press <return> to proceed..." % (sweeps, (abs(res))))
As can be seen, sweeping does not dramatically change the result, but the “residual” of the equation (a measure of
how accurately it has been solved) drops about an order of magnitude with each additional sweep.
Attention: Choosing an optimal balance between the number of time steps, the number of sweeps, the number of
solver iterations, and the solver tolerance is more art than science and will require some experimentation on your
part for each new problem.
Finally, we can increase the number of steps to approach equilibrium, or we can just solve for it directly
>>> eq = DiffusionTerm(coeff=D0 * (1 - phi[0]))
>>> phi[0].setValue(valueRight)
>>> res = 1e+10
>>> while res > 1e-6:
... res = eq.sweep(var=phi[0],
>>> print phi[0].allclose(phiAnalytical, atol = 1e-1)

1
>>> if __name__ == '__main__':

... viewer.plot()
... raw_input("Implicit variable diffusity - steady-state. \
... Press <return> to proceed...")
1
solution variable
1 sweep
2 sweeps
3 sweeps
0.8 4 sweeps
analytical value
0.6
0.4
0.2
0
0 10 20 30 40 50

Fully implicit solutions are not without their pitfalls, particularly in steady state. Consider a localized block of material
diffusing in a closed box.
>>> phi = CellVariable(mesh=mesh, name=r"$\phi$")
>>> phi.value = 0.
>>> phi.setValue(1., where=(x > L/2. - L/10.) & (x < L/2. + L/10.))
>>> if __name__ == '__main__':
... viewer = Viewer(vars=phi, datamin=-0.1, datamax=1.1)
φ
φ
1.0
0.8
0.6
0.4
0.2
0.0
0 10 20 30 40 50
We assign no explicit boundary conditions, leaving the default no-flux boundary conditions, and solve
𝜕𝜑/𝜕𝑡 = ∇ · (𝐷∇𝜑)
>>> D = 1.
>>> eq = TransientTerm() == DiffusionTerm(D)
>>> dt = 10. * dx**2 / (2 * D)

>>> steps = 200

... eq.solve(var=phi, dt=dt)
... if __name__ == '__main__':
... viewer.plot()
>>> if __name__ == '__main__':
... raw_input("No-flux - transient. \
φ
φ
1.0
0.8
0.6
0.4
0.2
0.0
0 10 20 30 40 50
and see that 𝜑 dissipates to the expected average value of 0.2 with reasonable accuracy.
>>> print numerix.allclose(phi, 0.2, atol=1e-5)
True
If we reset the initial condition

>>> phi.value = 0.

>>> if __name__ == '__main__':

... viewer.plot()
and solve the steady-state problem

>>> if __name__ == '__main__':
... viewer.plot()
>>> if __name__ == '__main__':
... raw_input("No-flux - stead-state failure. \
>>> print numerix.allclose(phi, 0.0)

True
φ
φ
1.0
0.8
0.6
0.4
0.2
0.0
0 10 20 30 40 50
we find that the value is uniformly zero! What happened to our no-flux boundary conditions?
The problem is that in the implicit discretization of ∇ · (𝐷∇𝜑) = 0,

⃒ 𝐷
⃒ 2 − 𝐷2
⃒
⃒ ⃒ ⃒
⃒ Δ𝑥 Δ𝑥 ⃒ ⃒ new ⃒ ⃒0⃒
⃒ ⃒ ⃒ 𝜑0 ⃒ ⃒ ⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ .. .. ..
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ . . . ⃒ ⃒ . ⃒ ⃒ .. ⃒
⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒ . ⃒ ⃒.⃒
⃒⃒ . ⃒ ⃒ ⃒
⃒
⃒
𝐷 2𝐷 𝐷
− Δ𝑥 − Δ𝑥
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ 2 Δ𝑥2 2 ⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒ ⃒ 𝜑new ⃒ ⃒ ⃒
⃒ ⃒ ⃒ 𝑗−1 ⃒ ⃒0⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
𝐷 2𝐷 𝐷
⃒
⃒ − Δ𝑥 2 Δ𝑥2
− Δ𝑥 2
⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒ new ⃒ ⃒ ⃒
⃒ ⃒ ⃒ 𝜑𝑗 ⃒ ⃒ ⃒
⃒ ⃒⃒ ⃒ = ⃒0⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
𝐷 2𝐷 𝐷
⃒ ⃒⃒ ⃒ ⃒ ⃒
⃒ − Δ𝑥 2 Δ𝑥2
− Δ𝑥 2
⃒⃒ ⃒ ⃒ ⃒
⃒ ⃒ new ⃒ ⃒ ⃒
⃒ ⃒ 𝜑𝑗+1 ⃒ ⃒ ⃒
⃒
⃒ ⃒0⃒
⃒
⃒ ⃒⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
.. .. .. ⃒ ⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
. . . ⃒⃒ . ⃒ ⃒ ⃒
⃒ ⃒ ⃒ ⃒
⃒ ⃒ . ⃒ ⃒.⃒
⃒
⃒ ⃒ . ⃒ ⃒.⃒
⃒
⃒ ⃒.⃒
⃒
⃒ ⃒⃒
⃒ ⃒⃒ ⃒ ⃒ ⃒
𝐷 𝐷 ⃒⃒
⃒ − Δ𝑥 2 Δ𝑥2
⃒ ⃒ ⃒
⃒ ⃒𝜑new ⃒ ⃒ ⃒
⃒ ⃒ ⃒ ⃒ ⃒ ⃒
⃒
⃒ ⃒ 𝑁 −1 ⃒0⃒
⃒ ⃒
the initial condition 𝜑old no longer appears and 𝜑 = 0 is a perfectly legitimate solution to this matrix equation.
The solution is to run the transient problem and to take one enormous time step
>>> phi.value = 0.
>>> if __name__ == '__main__':
... viewer.plot()
>>> (TransientTerm() == DiffusionTerm(D)).solve(var=phi, dt=1e6*dt)

>>> if __name__ == '__main__':
... viewer.plot()
>>> if __name__ == '__main__':
... raw_input("No-flux - steady-state. \
>>> print numerix.allclose(phi, 0.2, atol=1e-5)

True

φ
φ
1.0
0.8
0.6
0.4
0.2
0.0
0 10 20 30 40 50
If this example had been written primarily as a script, instead of as documentation, we would delete every line that
does not begin with either “>>>” or “...”, and then delete those prefixes from the remaining lines, leaving:
#!/usr/bin/env python
## This script was derived from

## 'examples/diffusion/mesh1D.py'
nx = 50
dx = 1.
mesh = Grid1D(nx = nx, dx = dx)
phi = CellVariable(name="solution variable",
mesh=mesh,
value=0)
eq = DiffusionTerm(coeff=D0 * (1 - phi[0]))
phi[0].setValue(valueRight)
res = 1e+10
while res > 1e-6:

res = eq.sweep(var=phi[0],
dt=timeStepDuration)
print phi[0].allclose(phiAnalytical, atol = 1e-1)

# Expect:
# 1
#
if __name__ == '__main__':
viewer.plot()
raw_input("Implicit variable diffusity - steady-state. \
Press <return> to proceed...")
Your own scripts will tend to look like this, although you can always write them as doctest scripts if you choose. You
can obtain a plain script like this from some .../example.py by typing:
$ python setup.py copy_script --From .../example.py --To myExample.py
at the command line.

Most of the FiPy examples will be a mixture of plain scripts and doctest documentation/tests.
10.2 examples.diffusion.coupled
Solve the biharmonic equation as a coupled pair of diffusion equations.

FiPy has only first order time derivatives so equations such as the biharmonic wave equation written as
𝜕4𝑣 𝜕2𝑣
+ =0
𝜕𝑥4 𝜕𝑡2
cannot be represented as a single equation. We need to decompose the biharmonic equation into two equations that
are first order in time in the following way,
𝜕 2 𝑣0 𝜕𝑣1
2
+ =0
𝜕𝑥 𝜕𝑡
𝜕 2 𝑣1 𝜕𝑣0
− =0
𝜕𝑥2 𝜕𝑡
Historically, FiPy required systems of coupled equations to be solved successively, “sweeping” the equations to con-
vergence. As a practical example, we use the following system
𝜕𝑣0
= 0.01∇2 𝑣0 − ∇2 𝑣1
𝜕𝑡
𝜕𝑣1
= ∇2 𝑣0 + 0.01∇2 𝑣1
𝜕𝑡
subject to the boundary conditions
𝑣0 |𝑥=0 = 0 𝑣0 |𝑥=1 = 1
𝑣1 |𝑥=0 = 1 𝑣1 |𝑥=1 = 0
This system closely resembles the pure biharmonic equation, but has an additional diffusion contribution to improve
numerical stability. The example system is solved with the following block of code using explicit coupling for the
cross-coupled terms.

>>> from fipy import Grid1D, CellVariable, TransientTerm, DiffusionTerm, Viewer
>>> m = Grid1D(nx=100, Lx=1.)
>>> v0 = CellVariable(mesh=m, hasOld=True, value=0.5)

>>> v1 = CellVariable(mesh=m, hasOld=True, value=0.5)
>>> v0.constrain(0, m.facesLeft)

>>> v0.constrain(1, m.facesRight)
>>> v1.constrain(1, m.facesLeft)

>>> v1.constrain(0, m.facesRight)
>>> eq0 = TransientTerm() == DiffusionTerm(coeff=0.01) - v1.faceGrad.divergence

>>> eq1 = TransientTerm() == v0.faceGrad.divergence + DiffusionTerm(coeff=0.01)
>>> vi = Viewer((v0, v1))
>>> for t in range(100):

... v0.updateOld()
... v1.updateOld()
... res0 = res1 = 1e100
... while max(res0, res1) > 0.1:
... res0 = eq0.sweep(var=v0, dt=1e-5)
... res1 = eq1.sweep(var=v1, dt=1e-5)
... if t % 10 == 0:
... vi.plot()
The uncoupled method still works, but it can be advantageous to solve the two equations simultaneously. In this case,
by coupling the equations, we can eliminate the explicit sources and dramatically increase the time steps:
>>> v0.value = 0.5
>>> v1.value = 0.5
>>> eqn0 = TransientTerm(var=v0) == DiffusionTerm(0.01, var=v0) - DiffusionTerm(1, var=v1)

>>> eqn1 = TransientTerm(var=v1) == DiffusionTerm(1, var=v0) + DiffusionTerm(0.01, var=v1)
>>> eqn = eqn0 & eqn1

... v0.updateOld()
... v1.updateOld()
... eqn.solve(dt=1.e-3)
... vi.plot()
It is also possible to pose the same equations in vector form:

>>> v = CellVariable(mesh=m, hasOld=True, value=[[0.5], [0.5]], elementshape=(2,))
>>> v.constrain([[0], [1]], m.facesLeft)

>>> v.constrain([[1], [0]], m.facesRight)
>>> eqn = TransientTerm([[1, 0],

... [0, 1]]) == DiffusionTerm([[[0.01, -1],
... [1, 0.01]]])
>>> vi = Viewer((v[0], v[1]))
10.2. examples.diffusion.coupled 73

... v.updateOld()
... eqn.solve(var=v, dt=1.e-3)
... vi.plot()
Whether you pose your problem in coupled or vector form should be dictated by the underlying physics. If 𝑣0 and
𝑣1 represent the concentrations of two conserved species, then it is natural to write two seperate governing equations
and to couple them. If they represent two components of a vector field, then the vector formulation is obviously more
natural. FiPy will solve the same matrix system either way.
10.3 examples.diffusion.mesh20x20
Solve a two-dimensional diffusion problem in a square domain.

This example solves a diffusion problem and demonstrates the use of applying boundary condition patches.
>>> nx = 20
>>> ny = nx
>>> dx = 1.
>>> dy = dx
>>> L = dx * nx
>>> mesh = Grid2D(dx=dx, dy=dy, nx=nx, ny=ny)
We create a CellVariable and initialize it to zero:

>>> phi = CellVariable(name = "solution variable",
... mesh = mesh,
... value = 0.)
and then create a diffusion equation. This is solved by default with an iterative conjugate gradient solver.
>>> D = 1.
>>> eq = TransientTerm() == DiffusionTerm(coeff=D)
We apply Dirichlet boundary conditions

>>> valueTopLeft = 0
>>> valueBottomRight = 1
to the top-left and bottom-right corners. Neumann boundary conditions are automatically applied to the top-right and
bottom-left corners.
>>> facesTopLeft = ((mesh.facesLeft & (Y > L / 2))
... | (mesh.facesTop & (X < L / 2)))
>>> facesBottomRight = ((mesh.facesRight & (Y < L / 2))
... | (mesh.facesBottom & (X > L / 2)))
>>> phi.constrain(valueTopLeft, facesTopLeft)

>>> phi.constrain(valueBottomRight, facesBottomRight)
We create a viewer to see the results

>>> if __name__ == '__main__':
... viewer = Viewer(vars=phi, datamin=0., datamax=1.)
... viewer.plot()

and solve the equation by repeatedly looping in time:

>>> timeStepDuration = 10 * 0.9 * dx**2 / (2 * D)
>>> steps = 10
... eq.solve(var=phi,
... if __name__ == '__main__':
... viewer.plot()
solution variable
20 1.0
0.9
0.8
15
0.7
0.6
solution variable
10 0.5
0.4
0.3
5
0.2
0.1
0 0.0
0 5 10 15 20
We can test the value of the bottom-right corner cell.

>>> print numerix.allclose(phi(((L,), (0,))), valueBottomRight, atol = 1e-2)
1
>>> if __name__ == '__main__':

... raw_input("Implicit transient diffusion. Press <return> to proceed...")
We can also solve the steady-state problem directly

>>> DiffusionTerm().solve(var=phi)
>>> if __name__ == '__main__':
... viewer.plot()
10.3. examples.diffusion.mesh20x20 75
solution variable
20 1.0
0.9
0.8
15
0.7
0.6
solution variable
10 0.5
0.4
0.3
5
0.2
0.1
0 0.0
0 5 10 15 20
and test the value of the bottom-right corner cell.

>>> print numerix.allclose(phi(((L,), (0,))), valueBottomRight, atol = 1e-2)
1
>>> if __name__ == '__main__':

... raw_input("Implicit steady-state diffusion. Press <return> to proceed...")
10.4 examples.diffusion.circle
Solve the diffusion equation in a circular domain meshed with triangles.

This example demonstrates how to solve a simple diffusion problem on a non-standard mesh with varying boundary
conditions. The Gmsh package is used to create the mesh. Firstly, define some parameters for the creation of the mesh,
>>> cellSize = 0.05
>>> radius = 1.
The cellSize is the preferred edge length of each mesh element and the radius is the radius of the circular mesh domain.
In the following code section a file is created with the geometry that describes the mesh. For details of how to write
such geometry files for Gmsh, see the gmsh manual.
The mesh created by Gmsh is then imported into FiPy using the Gmsh2D object.
>>> mesh = Gmsh2D('''
... cellSize = %(cellSize)g;

... radius = %(radius)g;

... Point(1) = {0, 0, 0, cellSize};
... Point(2) = {-radius, 0, 0, cellSize};
... Point(3) = {0, radius, 0, cellSize};
... Point(4) = {radius, 0, 0, cellSize};
... Point(5) = {0, -radius, 0, cellSize};
... Circle(6) = {2, 1, 3};
... Circle(7) = {3, 1, 4};
... Circle(8) = {4, 1, 5};
... Circle(9) = {5, 1, 2};
... Line Loop(10) = {6, 7, 8, 9};
... Plane Surface(11) = {10};
... ''' % locals())
Using this mesh, we can construct a solution variable

>>> phi = CellVariable(name = "solution variable",
... mesh = mesh,
... value = 0.)
We can now create a Viewer to see the mesh

>>> viewer = None
>>> if __name__ == '__main__':
... try:
... viewer = Viewer(vars=phi, datamin=-1, datamax=1.)
... viewer.plotMesh()
... raw_input("Irregular circular mesh. Press <return> to proceed...")
... except:
... print "Unable to create a viewer for an irregular mesh (try Gist2DViewer, Matplotlib2DVie
10.4. examples.diffusion.circle 77
1.0
0.5
0.0
−0.5
−1.0
−1.0 −0.5 0.0 0.5 1.0
We set up a transient diffusion equation
>>> D = 1.
>>> eq = TransientTerm() == DiffusionTerm(coeff=D)
The following line extracts the 𝑥 coordinate values on the exterior faces. These are used as the boundary condition
fixed values.
>>> phi.constrain(X, mesh.exteriorFaces)
We first step through the transient problem

>>> timeStepDuration = 10 * 0.9 * cellSize**2 / (2 * D)
>>> steps = 10
... eq.solve(var=phi,
... if viewer is not None:
... viewer.plot()

1
1.0
0.5 0.5
solution variable
0.0 0
−0.5 −0.5
−1.0
−1
−1.0 −0.5 0.0 0.5 1.0
If we wanted to plot or analyze the results of this calculation with another application, we could export tab-separated-
values with
TSVViewer(vars=(phi, phi.grad)).plot(filename="myTSV.tsv")
x y solution variable solution variable_grad_x solution variable_grad_y

0.975559734792414 0.0755414402612554 0.964844363287199 -0.229687917881182
0.0442864953037566 0.79191893162384 0.0375859836421991 -0.773936613923853
0.0246775505084069 0.771959648896982 0.020853932412869 -0.723540342405813
0.223345558247991 -0.807931073108895 0.203035857140125 -0.777466238738658
-0.00726763301939488 -0.775978916110686 -0.00412895434496877 -0.650055516507232
-0.0220279064527904 -0.187563765977912 -0.012771874945585 -0.35707168379437
0.111223320911545 -0.679586798311355 0.0911595298310758 -0.613455176718145
-0.78996770899909 -0.0173672729866294 -0.693887874335319 -1.00671109050419
-0.703545986179876 -0.435813500559859 -0.635004192597412 -0.896203033957194
0.888641841567831 -0.408558914368324 0.877939107374768 -0.32195762184087
0.38212257821916 -0.51732949653553 0.292889724306196 -0.854466141879776 0
-0.359068256998365 0.757882581524374 -0.323541041763627 -0.870534227755687
-0.459673905457569 -0.701526587772079 -0.417577664032421 -0.725460726303266
-0.338256179134518 -0.523565732643067 -0.254030052182524 -0.923505840608445
0.87498754712638 0.174119064688993 0.836057900916614 -1.11590500805745 -0
-0.484106960369249 0.0705987421869745 -0.319827850867342 -0.867894407968447
-0.0221203060940465 -0.216026820080053 -0.0152729438559779 -0.341246696530392
The values are listed at the cell centers. Particularly for irregular meshes, no specific ordering should be relied upon.
Vector quantities are listed in multiple columns, one for each mesh dimension.
10.4. examples.diffusion.circle 79
This problem again has an analytical solution that depends on the error function, but it’s a bit more complicated due
to the varying boundary conditions and the different horizontal diffusion length at different vertical positions
>>> t = timeStepDuration * steps
>>> phiAnalytical = CellVariable(name="analytical value",

... mesh=mesh)
>>> x0 = radius * numerix.cos(numerix.arcsin(y))

>>> try:
... from scipy.special import erf
... ## This function can sometimes throw nans on OS X
... ## see http://projects.scipy.org/scipy/scipy/ticket/325
... phiAnalytical.setValue(x0 * (erf((x0+x) / (2 * numerix.sqrt(D * t)))
... - erf((x0-x) / (2 * numerix.sqrt(D * t)))))
... print "The SciPy library is not available to test the solution to \
... the transient diffusion equation"

1
>>> if __name__ == '__main__':

... raw_input("Transient diffusion. Press <return> to proceed...")
As in the earlier examples, we can also directly solve the steady-state diffusion problem.
The values at the elements should be equal to their x coordinate

>>> print phi.allclose(x, atol = 0.03)
1
Display the results if run as a script.

>>> if viewer is not None:
... viewer.plot()
... raw_input("Steady-state diffusion. Press <return> to proceed...")

1
1.0
0.5 0.5
solution variable
0.0 0
−0.5 −0.5
−1.0
−1
−1.0 −0.5 0.0 0.5 1.0
10.5 examples.diffusion.electrostatics
Solve the Poisson equation in one dimension.

The Poisson equation is a particular example of the steady-state diffusion equation. We examine a few cases in one
dimension.
>>> nx = 200
>>> dx = 0.01
>>> L = nx * dx
>>> mesh = Grid1D(dx = dx, nx = nx)
Given the electrostatic potential 𝜑,

>>> potential = CellVariable(mesh=mesh, name='potential', value=0.)
the permittivity 𝜖,
>>> permittivity = 1
the concentration 𝐶𝑗 of the 𝑗 th component with valence 𝑧𝑗 (we consider only a single component 𝐶e− with valence
with 𝑧e− = −1)
10.5. examples.diffusion.electrostatics 81
>>> electrons = CellVariable(mesh=mesh, name='e-')

>>> electrons.valence = -1
and the charge density 𝜌,

>>> charge = electrons * electrons.valence
>>> charge.name = "charge"
The dimensionless Poisson equation is

𝑛
∑︁
∇ · (𝜖∇𝜑) = −𝜌 = − 𝑧𝑗 𝐶𝑗
𝑗=1
>>> potential.equation = (DiffusionTerm(coeff = permittivity)

... + charge == 0)
Because this equation admits an infinite number of potential profiles, we must constrain the solution by fixing the
potential at one point:
>>> potential.constrain(0., mesh.facesLeft)
First, we obtain a uniform charge distribution by setting a uniform concentration of electrons 𝐶e− = 1.
>>> electrons.setValue(1.)
and we solve for the electrostatic potential

>>> potential.equation.solve(var=potential)
This problem has the analytical solution
𝑥2
𝜓(𝑥) = − 2𝑥
2
>>> analytical = CellVariable(mesh=mesh, name="analytical solution",
... value=(x**2)/2 - 2*x)
which has been satisifactorily obtained

>>> print potential.allclose(analytical, rtol = 2e-5, atol = 2e-5)
1
If we are running the example interactively, we view the result

>>> if __name__ == '__main__':
... viewer = Viewer(vars=(charge, potential, analytical))
... viewer.plot()
... raw_input("Press any key to continue...")

0
charge
potential
analytical solution
-0.5
-1
-1.5
-2
0 0.5 1 1.5 2
Next, we segregate all of the electrons to right side of the domain

{︃
0 for 𝑥 ≤ 𝐿/2,
𝐶e− =
1 for 𝑥 > 𝐿/2.
>>> electrons.setValue(1., where=x > L / 2.)
and again solve for the electrostatic potential

which now has the analytical solution

{︃
−𝑥 for 𝑥 ≤ 𝐿/2,
𝜓(𝑥) = (𝑥−1)2
2 −𝑥 for 𝑥 > 𝐿/2.
>>> analytical.setValue(-x)
>>> analytical.setValue(((x-1)**2)/2 - x, where=x > L/2)

1
and again view the result
10.5. examples.diffusion.electrostatics 83
>>> if __name__ == '__main__':

... viewer.plot()
... raw_input("Press any key to continue...")
0
charge
potential
-0.2 analytical solution
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
0 0.5 1 1.5 2
Finally, we segregate all of the electrons to the left side of the domain
{︃
1 for 𝑥 ≤ 𝐿/2,
𝐶e− =
0 for 𝑥 > 𝐿/2.
>>> electrons.setValue(0., where=x > L / 2.)
and again solve for the electrostatic potential

which has the analytical solution

{︃
𝑥2
2 − 𝑥 for 𝑥 ≤ 𝐿/2,
𝜓(𝑥) =
− 21 for 𝑥 > 𝐿/2.
We again verify that the correct equilibrium is attained

>>> analytical.setValue((x**2)/2 - x)
>>> analytical.setValue(-0.5, where=x > L / 2)


1
and once again view the result

>>> if __name__ == '__main__':
... viewer.plot()
0
charge
potential
analytical solution
-0.2
-0.4
-0.6
-0.8
-1
0 0.5 1 1.5 2
10.6 examples.diffusion.nthOrder.input4thOrder1D
Solve a fourth-order diffusion problem.

This example uses the DiffusionTerm class to solve the equation
𝜕4𝜑
=0
𝜕𝑥4
on a 1D mesh of length
>>> L = 1000.
We create an appropriate mesh

>>> nx = 500
>>> dx = L / nx
>>> mesh = Grid1D(dx=dx, nx=nx)
10.6. examples.diffusion.nthOrder.input4thOrder1D 85
and initialize the solution variable to 0

>>> var = CellVariable(mesh=mesh, name='solution variable')
For this problem, we impose the boundary conditions:
𝜑 = 𝛼1 at 𝑥 = 0
𝜕𝜑
= 𝛼2 at 𝑥 = 𝐿
𝜕𝑥
𝜕2𝜑
= 𝛼3 at 𝑥 = 0
𝜕𝑥2
𝜕3𝜑
= 𝛼4 at 𝑥 = 𝐿.
𝜕𝑥3
or
>>> alpha1 = 2.
>>> alpha2 = 1.
>>> alpha3 = 4.
>>> alpha4 = -3.
>>> BCs = (NthOrderBoundaryCondition(faces=mesh.facesLeft, value=alpha3, order=2),

... NthOrderBoundaryCondition(faces=mesh.facesRight, value=alpha4, order=3))
>>> var.faceGrad.constrain([alpha2], mesh.facesRight)
>>> var.constrain(alpha1, mesh.facesLeft)
We initialize the steady-state equation

>>> eq = DiffusionTerm(coeff=(1, 1)) == 0
and use the LinearLUSolver for stability.

We perform one implicit timestep to achieve steady state
>>> eq.solve(var=var,
... boundaryConditions=BCs,
... solver=GeneralSolver())
The analytical solution is:

𝛼4 3 𝛼3 2 (︁ 𝛼4 2 )︁
𝜑= 𝑥 + 𝑥 + 𝛼2 − 𝐿 − 𝛼3 𝐿 𝑥 + 𝛼1
6 2 2
or
>>> analytical = CellVariable(mesh=mesh, name='analytical value')
>>> analytical.setValue(alpha4 / 6. * x**3 + alpha3 / 2. * x**2 + \
... (alpha2 - alpha4 / 2. * L**2 - alpha3 * L) * x + alpha1)
>>> print var.allclose(analytical, rtol=1e-4)

1
If the problem is run interactively, we can view the result:

>>> if __name__ == '__main__':
... viewer = Viewer(vars=(var, analytical))
... viewer.plot()

x1e9
1.0
solution variable
analytical value
0.8
0.6
0.4
0.2
0.0
0 200 400 600 800 1000
10.7 examples.diffusion.anisotropy
Solve the diffusion equation with an anisotropic diffusion coefficient.

We wish to solve the problem
𝜕𝜑
= 𝜕𝑗 Γ𝑖𝑗 𝜕𝑖 𝜑
𝜕𝑡
on a circular domain centred at (0, 0). We can choose an anisotropy ratio of 5 such that
(︂ )︂
0.2 0
Γ′ =
0 1
A new matrix is formed by rotating Γ′ such that
(︂ )︂
cos 𝜃 sin 𝜃
𝑅=
− sin 𝜃 cos 𝜃
and
Γ = 𝑅Γ′ 𝑅𝑇
In the case of a point source at (0, 0) a reference solution is given by,
(︁ (︁ 2 )︁)︁
1 𝑋 𝑌2
exp − 4𝑡 Γ′00 + Γ′11
𝜑 (𝑋, 𝑌, 𝑡) = 𝑄 √︀
4𝜋𝑡 Γ′00 Γ′11
𝑇 𝑇
where (𝑋, 𝑌 ) = 𝑅 (𝑥, 𝑦) and 𝑄 is the initial mass.
10.7. examples.diffusion.anisotropy 87
Import a mesh previously created using Gmsh.

>>> import os
>>> mesh = Gmsh2D(os.path.splitext(__file__)[0] + '.msh', communicator=serialComm)
Set the centermost cell to have a value.

>>> var = CellVariable(mesh=mesh, hasOld=1)
>>> var[numerix.argmin(x**2 + y**2)] = 1.
Choose an orientation for the anisotropy.

>>> theta = numerix.pi / 4.
>>> rotationMatrix = numerix.array(((numerix.cos(theta), numerix.sin(theta)), \
... (-numerix.sin(theta), numerix.cos(theta))))
>>> gamma_prime = numerix.array(((0.2, 0.), (0., 1.)))
>>> DOT = numerix.NUMERIX.dot
>>> gamma = DOT(DOT(rotationMatrix, gamma_prime), numerix.transpose(rotationMatrix))
Make the equation, viewer and solve.

>>> eqn = TransientTerm() == DiffusionTermCorrection((gamma,))
>>> if __name__ == '__main__':

... viewer = Viewer(var, datamin=0.0, datamax=0.001)
>>> mass = float(var.cellVolumeAverage * numerix.sum(mesh.cellVolumes))

>>> time = 0
>>> dt=0.00025
>>> for i in range(20):

... var.updateOld()
... res = 1.
...
... while res > 1e-2:
... res = eqn.sweep(var, dt=dt)
...
... if __name__ == '__main__':
... viewer.plot()
... time += dt
Compare with the analytical solution (within 5% accuracy).

>>> X, Y = numerix.dot(mesh.cellCenters, CellVariable(mesh=mesh, rank=2, value=rotationMatrix))
>>> solution = mass * numerix.exp(-(X**2 / gamma_prime[0][0] + Y**2 / gamma_prime[1][1]) / (4 * time)
>>> print max(abs((var - solution) / max(solution))) < 0.08
True

Chapter 11
Convection Examples
examples.convection.exponential1D.mesh1D Solve the steady-state convection-diffusion equation in one dim

examples.convection.exponential1DSource.mesh1D Solve the steady-state convection-diffusion equation with a co
examples.convection.robin Solve an advection-diffusion equation with a Robin boundary
examples.convection.source Solve a convection problem with a source.
11.1 examples.convection.exponential1D.mesh1D
Solve the steady-state convection-diffusion equation in one dimension.

This example solves the steady-state convection-diffusion equation given by
∇ · (𝐷∇𝜑 + ⃗𝑢𝜑) = 0
with coefficients 𝐷 = 1 and ⃗𝑢 = 10ß̂, or

>>> diffCoeff = 1.
>>> convCoeff = (10.,)
We define a 1D mesh
>>> L = 10.
>>> nx = 10
>>> mesh = Grid1D(dx=L / nx, nx=nx)
>>> valueLeft = 0.
>>> valueRight = 1.
The solution variable is initialized to valueLeft:

>>> var = CellVariable(mesh=mesh, name="variable")
and impose the boundary conditions

{︃
0 at 𝑥 = 0,
𝜑=
1 at 𝑥 = 𝐿,
with
>>> var.constrain(valueLeft, mesh.facesLeft)
>>> var.constrain(valueRight, mesh.facesRight)
89
The equation is created with the DiffusionTerm and ExponentialConvectionTerm. The scheme used by
the convection term needs to calculate a Peclet number and thus the diffusion term instance must be passed to the
convection term.
>>> eq = (DiffusionTerm(coeff=diffCoeff)
... + ExponentialConvectionTerm(coeff=convCoeff))
More details of the benefits and drawbacks of each type of convection term can be found in Numerical Schemes. Es-
sentially, the ExponentialConvectionTerm and PowerLawConvectionTerm will both handle most types
of convection-diffusion cases, with the PowerLawConvectionTerm being more efficient.
We solve the equation
>>> eq.solve(var=var)
and test the solution against the analytical result
1 − exp(−𝑢𝑥 𝑥/𝐷)
𝜑=
1 − exp(−𝑢𝑥 𝐿/𝐷)
or
>>> axis = 0
>>> x = mesh.cellCenters[axis]
>>> CC = 1. - numerix.exp(-convCoeff[axis] * x / diffCoeff)
>>> DD = 1. - numerix.exp(-convCoeff[axis] * L / diffCoeff)
>>> analyticalArray = CC / DD
>>> print var.allclose(analyticalArray)
1

>>> if __name__ == '__main__':
... viewer = Viewer(vars=var)
... viewer.plot()
11.2 examples.convection.exponential1DSource.mesh1D
Solve the steady-state convection-diffusion equation with a constant source.

Like examples.convection.exponential1D.mesh1D this example solves a steady-state convection-
diffusion equation, but adds a constant source, 𝑆0 = 1, such that
∇ · (𝐷∇𝜑 + ⃗𝑢𝜑) + 𝑆0 = 0.
>>> diffCoeff = 1.
>>> sourceCoeff = 1.
We define a 1D mesh
>>> nx = 1000
>>> L = 10.
>>> mesh = Grid1D(dx=L / 1000, nx=nx)
90 Chapter 11. Convection Examples

>>> valueLeft = 0.
>>> valueRight = 1.

>>> var = CellVariable(name="variable", mesh=mesh)

{︃
0 at 𝑥 = 0,
𝜑=
1 at 𝑥 = 𝐿,
with
>>> var.constrain(valueLeft, mesh.facesLeft)
>>> var.constrain(valueRight, mesh.facesRight)
We define the convection-diffusion equation with source

>>> eq = (DiffusionTerm(coeff=diffCoeff)
... + ExponentialConvectionTerm(coeff=convCoeff)
... + sourceCoeff)
>>> eq.solve(var=var,
... solver=DefaultAsymmetricSolver(tolerance=1.e-15, iterations=10000))
and test the solution against the analytical result:

(︂ )︂
𝑆0 𝑥 𝑆0 𝑥 1 − exp(−𝑢𝑥 𝑥/𝐷)
𝜑=− + 1+
𝑢𝑥 𝑢𝑥 1 − exp(−𝑢𝑥 𝐿/𝐷)
or
>>> axis = 0
>>> x = mesh.cellCenters[axis]
>>> AA = -sourceCoeff * x / convCoeff[axis]
>>> BB = 1. + sourceCoeff * L / convCoeff[axis]
>>> analyticalArray = AA + BB * CC / DD
>>> print var.allclose(analyticalArray, rtol=1e-4, atol=1e-4)
1

>>> if __name__ == '__main__':
... viewer = Viewer(vars=var)
... viewer.plot()
11.3 examples.convection.robin
Solve an advection-diffusion equation with a Robin boundary condition.

This example demonstrates how to apply a Robin boundary condition to an advection-diffusion equation. The equation
11.3. examples.convection.robin 91
we wish to solve is given by,
𝜕2𝐶 𝜕𝐶
0= 2
−𝑃 − 𝐷𝐶 0<𝑥<1
𝜕𝑥 𝜕𝑥
𝜕𝐶
𝑥=0:𝑃 =− + 𝑃𝐶
𝜕𝑥
𝜕𝐶
𝑥=1: =0
𝜕𝑥
The analytical solution for this equation is given by,
2𝑃 exp 𝑃2𝑥 (𝑃 + 𝐴) exp 𝐴2 (𝑥 − 1) − (𝑃 − 𝐴) exp − 𝐴2 (𝑥 − 1)

(︀ )︀ [︀ (︀ )︀ (︀ )︀]︀
𝐶 (𝑥) = 2 2
(𝑃 + 𝐴) exp 𝐴2 − (𝑃 − 𝐴) exp − 𝐴2
(︀ )︀ (︀ )︀
where
√︀
𝐴= 𝑃 + 4𝐷2

>>> nx = 100
>>> dx = 1.0 / nx

>>> C = CellVariable(mesh=mesh)
>>> D = 2.0
>>> P = 3.0
>>> C.faceGrad.constrain([-P + P * C.faceValue], mesh.facesLeft)

>>> C.faceGrad.constrain([0], mesh.facesRight)
>>> eq = PowerLawConvectionTerm((P,)) == \
... DiffusionTerm() - ImplicitSourceTerm(D)
>>> A = numerix.sqrt(P**2 + 4 * D)
>>> CAnalytical = CellVariable(mesh=mesh)
>>> CAnalytical.setValue(2 * P * numerix.exp(P * x / 2) * ((P + A) * numerix.exp(A / 2 * (1 - x))
... - (P - A) * numerix.exp(-A / 2 *(1 - x)))/
... ((P + A)**2*numerix.exp(A / 2)- (P - A)**2 * numerix.exp(-A / 2)))
>>> if __name__ == '__main__':

... C.name = 'C'
... viewer = Viewer(vars=(C, CAnalytical))
>>> if __name__ == '__main__':

... restol = 1e-5
... anstol = 1e-3
... else:
... restol = 0.5
... anstol = 0.15
>>> res = 1e+10

>>> while res > restol:

... res = eq.sweep(var=C)
... if __name__ == '__main__':
... viewer.plot()
>>> print C.allclose(CAnalytical, rtol=anstol, atol=anstol)

True
11.4 examples.convection.source
Solve a convection problem with a source.

This example solves the equation
𝜕𝜑
+ 𝛼𝜑 = 0
𝜕𝑥
with 𝜑 (0) = 1 at 𝑥 = 0. The boundary condition at 𝑥 = 𝐿 is an outflow boundary condition requiring the use of
an artificial constraint to be set on the right hand side faces. Exterior faces without constraints are considered to have
zero outflow. An ImplicitSourceTerm object will be used to represent this term. The derivative of 𝜑 can be
represented by a ConvectionTerm with a constant unitary velocity field from left to right. The following is an
example code that includes a test against the analytical result.
>>> L = 10.
>>> nx = 5000
>>> dx = L / nx
>>> phi0 = 1.0
>>> alpha = 1.0
>>> phi = CellVariable(name=r"$\phi$", mesh=mesh, value=phi0)
>>> solution = CellVariable(name=r"solution", mesh=mesh, value=phi0 * numerix.exp(-alpha * mesh.cellC
>>> if __name__ == "__main__":

... viewer = Viewer(vars=(phi, solution))
... viewer.plot()
... raw_input("press key to continue")
>>> phi.constrain(phi0, mesh.facesLeft)

>>> ## fake outflow condition
>>> phi.faceGrad.constrain([0], mesh.facesRight)
>>> eq = PowerLawConvectionTerm((1,)) + ImplicitSourceTerm(alpha)

>>> eq.solve(phi)
>>> print numerix.allclose(phi, phi0 * numerix.exp(-alpha * mesh.cellCenters[0]), atol=1e-3)
True
>>> if __name__ == "__main__":

... viewer = Viewer(vars=(phi, solution))
... viewer.plot()
... raw_input("finished")
11.4. examples.convection.source 93

Chapter 12
Phase Field Examples
examples.phase.simple Solve a phase-field (Allen-Cahn) problem in one-dimension.

examples.phase.binaryCoupled Simultaneously solve a phase-field evolution and solute diffusion problem in
examples.phase.quaternary Solve a phase-field evolution and diffusion of four species in one-dimension.
examples.phase.anisotropy Solve a dendritic solidification problem.
examples.phase.impingement.mesh40x1 Solve for the impingement of two grains in one dimension.
examples.phase.impingement.mesh20x20 Solve for the impingement of four grains in two dimensions.
examples.phase.polyxtal Solve the dendritic growth of nuclei and subsequent grain impingement.
examples.phase.polyxtalCoupled Simultaneously solve the dendritic growth of nuclei and subsequent grain imp
12.1 examples.phase.simple
Solve a phase-field (Allen-Cahn) problem in one-dimension.

To run this example from the base FiPy directory, type python examples/phase/simple/input.py at the
command line. A viewer object should appear and, after being prompted to step through the different examples, the
word finished in the terminal.
This example takes the user through assembling a simple problem with FiPy. It describes a steady 1D phase field
problem with no-flux boundary conditions such that,
1 𝜕𝜑 𝜕𝑓
= 𝜅𝜑 ∇2 𝜑 − (12.1)
𝑀𝜑 𝜕𝑡 𝜕𝜑
For solidification problems, the Helmholtz free energy is frequently given by

𝑊 𝑇 − 𝑇𝑀
𝑓 (𝜑, 𝑇 ) = 𝑔(𝜑) + 𝐿𝑣 𝑝(𝜑)
2 𝑇𝑀
where 𝑊 is the double-well barrier height between phases, 𝐿𝑣 is the latent heat, 𝑇 is the temperature, and 𝑇𝑀 is the
melting point.
One possible choice for the double-well function is
𝑔(𝜑) = 𝜑2 (1 − 𝜑)2
and for the interpolation function is
𝑝(𝜑) = 𝜑3 (6𝜑2 − 15𝜑 + 10).
We create a 1D solution mesh

95
>>> L = 1.
>>> nx = 400
>>> dx = L / nx
We create the phase field variable

>>> phase = CellVariable(name = "phase",
... mesh = mesh)
and set a step-function initial condition

{︃
1 for 𝑥 ≤ 𝐿/2
𝜑= at 𝑡 = 0
0 for 𝑥 > 𝐿/2
>>> phase.setValue(1.)
>>> phase.setValue(0., where=x > L/2)
If we are running interactively, we’ll want a viewer to see the results

>>> if __name__ == '__main__':
... viewer = Viewer(vars = (phase,))
... viewer.plot()
... raw_input("Initial condition. Press <return> to proceed...")
We choose the parameter values,

>>> kappa = 0.0025
>>> W = 1.
>>> Lv = 1.
>>> Tm = 1.
>>> T = Tm
>>> enthalpy = Lv * (T - Tm) / Tm
96 Chapter 12. Phase Field Examples

We build the equation by assembling the appropriate terms. Since, with 𝑇 = 𝑇𝑀 we are interested in a steady-state
solution, we omit the transient term (1/𝑀𝜑 ) 𝜕𝜑
𝜕𝑡 .
The analytical solution for this steady-state phase field problem, in an infinite domain, is
[︃ ]︃
1 𝑥 − 𝐿/2
𝜑= 1 − tanh √︀ (12.2)
2 2 𝜅/𝑊
or
>>> analyticalArray = 0.5*(1 - numerix.tanh((x - L/2)/(2*numerix.sqrt(kappa/W))))
We treat the diffusion term 𝜅𝜑 ∇2 𝜑 implicitly,
Note: “Diffusion” in FiPy is not limited to the movement of atoms, but rather refers to the spontaneous spreading of
any quantity (e.g., solute, temperature, or in this case “phase”) by flow “down” the gradient of that quantity.
The source term is

𝜕𝑓 𝑊 𝑇 − 𝑇𝑀 ′
𝑆=− = − 𝑔 ′ (𝜑) − 𝐿 𝑝 (𝜑)
𝜕𝜑 2 𝑇𝑀
[︂ ]︂
𝑇 − 𝑇𝑀
= − 𝑊 𝜑(1 − 𝜑)(1 − 2𝜑) + 𝐿 30𝜑2 (1 − 𝜑)2
𝑇𝑀
= 𝑚𝜑 𝜑(1 − 𝜑)
where 𝑚𝜑 ≡ −[𝑊 (1 − 2𝜑) + 30𝜑(1 − 𝜑)𝐿 𝑇 −𝑇

𝑇𝑀 ].
𝑀
The simplest approach is to add this source explicitly

>>> mPhi = -((1 - 2 * phase) * W + 30 * phase * (1 - phase) * enthalpy)
>>> S0 = mPhi * phase * (1 - phase)
>>> eq = S0 + DiffusionTerm(coeff=kappa)
After solving this equation

>>> eq.solve(var = phase, solver=DummySolver())
we obtain the surprising result that 𝜑 is zero everywhere.

>>> print phase.allclose(analyticalArray, rtol = 1e-4, atol = 1e-4)
0
>>> if __name__ == '__main__':
... viewer.plot()
... raw_input("Fully explicit source. Press <return> to proceed...")
12.1. examples.phase.simple 97
On inspection, we can see that this occurs because, for our step-function initial condition, 𝑚𝜑 = 0 everwhere, hence
we are actually only solving the simple implicit diffusion equation 𝜅𝜑 ∇2 𝜑 = 0, which has exactly the uninteresting
solution we obtained.
The resolution to this problem is to apply relaxation to obtain the desired answer, i.e., the solution is allowed to relax
in time from the initial condition to the desired equilibrium solution. To do so, we reintroduce the transient term from
Equation (12.1)
>>> eq = TransientTerm() == DiffusionTerm(coeff=kappa) + S0

... eq.solve(var = phase, dt=1.)
... if __name__ == '__main__':
... viewer.plot()
After 13 time steps, the solution has converged to the analytical solution
1
>>> if __name__ == '__main__':
... raw_input("Relaxation, explicit. Press <return> to proceed...")

Note: The solution is only found accurate to ≈ 4.3 × 10−5 because the infinite-domain analytical solution (12.2) is
not an exact representation for the solution in a finite domain of length 𝐿.
Setting fixed-value boundary conditions of 1 and 0 would still require the relaxation method with the fully explicit
source.
Solution performance can be improved if we exploit the dependence of the source on 𝜑. By doing so, we can make
the source semi-implicit, improving the rate of convergence over the fully explicit approach. The source can only be
semi-implicit because we employ sparse linear algebra routines to solve the PDEs, i.e., there is no fully implicit way
⃗ − ⃗𝑏 = 0.
to represent a term like 𝜑4 in the linear set of equations M𝜑
By linearizing a source as 𝑆 = 𝑆0 − 𝑆1 𝜑, we make it more implicit by adding the coefficient 𝑆1 to the matrix diagonal.
For numerical stability, this linear coefficient must never be negative.
There are an infinite number of choices for this linearization, but many do not converge very well. One choice is that
used by Ryo Kobayashi:
>>> S0 = mPhi * phase * (mPhi > 0)
>>> S1 = mPhi * ((mPhi < 0) - phase)
>>> eq = DiffusionTerm(coeff=kappa) + S0 \
... + ImplicitSourceTerm(coeff = S1)
Note: Because mPhi is a variable field, the quantities (mPhi > 0) and (mPhi < 0) evaluate to variable fields
of True and False, instead of single boolean values.
This expression converges to the same value given by the explicit relaxation approach, but in only 8 sweeps (note that
because there is no transient term, these sweeps are not time steps, but rather repeated iterations at the same time step
to reach a converged solution).
Note: We use solve() instead of sweep() because we don’t care about the residual. Either function would work,
but solve() is a bit faster.

... eq.solve(var = phase)
1
>>> if __name__ == '__main__':
... viewer.plot()
... raw_input("Kobayashi, semi-implicit. Press <return> to proceed...")
In general, the best convergence is obtained when the linearization gives a good representation of the relationship
between the source and the dependent variable. The best practical advice is to ⃒ perform a Taylor expansion of the source
⃒
about the previous value of the dependent variable such that 𝑆 = 𝑆old + 𝜕𝜑 ⃒ (𝜑 − 𝜑old ) = (𝑆 − 𝜕𝑆
𝜕𝑆 ⃒ 𝜕𝑆 ⃒
𝜕𝜑 𝜑)old + 𝜕𝜑 ⃒ 𝜑.
⃒ old old
Now, if our source term is represented by 𝑆 = 𝑆0 + 𝑆1 𝜑, then 𝑆1 = 𝜕𝜑 ⃒ and 𝑆0 = (𝑆 − 𝜕𝜑 𝜑)old = 𝑆old − 𝑆1 𝜑old .
𝜕𝑆 ⃒ 𝜕𝑆
old
In this way, the linearized source will be tangent to the curve of the actual source as a function of the dependendent
variable.
For our source, 𝑆 = 𝑚𝜑 𝜑(1 − 𝜑),
𝜕𝑆 𝜕𝑚𝜑
= 𝜑(1 − 𝜑) + 𝑚𝜑 (1 − 2𝜑)
𝜕𝜑 𝜕𝜑
and
𝜕𝑚𝜑 𝑇 − 𝑇𝑀
= 2𝑊 − 30(1 − 2𝜑)𝐿 ,
𝜕𝜑 𝑇𝑀
or
>>> dmPhidPhi = 2 * W - 30 * (1 - 2 * phase) * enthalpy
>>> S1 = dmPhidPhi * phase * (1 - phase) + mPhi * (1 - 2 * phase)
>>> S0 = mPhi * phase * (1 - phase) - S1 * phase
>>> eq = DiffusionTerm(coeff=kappa) + S0 \
... + ImplicitSourceTerm(coeff = S1)
Using this scheme, where the coefficient of the implicit source term is tangent to the source, we reach convergence in
only 5 sweeps

... eq.solve(var = phase)
1
>>> if __name__ == '__main__':
... viewer.plot()
... raw_input("Tangent, semi-implicit. Press <return> to proceed...")
Although, for this simple problem, there is no appreciable difference in run-time between the fully explicit source and
the optimized semi-implicit source, the benefit of 60% fewer sweeps should be obvious for larger systems and longer
iterations.
This example has focused on just the region of the phase field interface in equilibrium. Problems of interest, though,
usually involve the dynamics of one phase transforming to another. To that end, let us recast the problem using physical
parameters and dimensions. We’ll need a new mesh

>>> nx = 400
>>> dx = 5e-6 # cm
>>> L = nx * dx
and thus must redeclare 𝜑 on the new mesh

>>> phase = CellVariable(name="phase",
... mesh=mesh,
... hasOld=1)
We choose the parameter values appropriate for nickel, given in [27]

>>> Lv = 2350 # J / cm**3
>>> Tm = 1728. # K
>>> T = Variable(value=Tm)
>>> enthalpy = Lv * (T - Tm) / Tm # J / cm**3
The parameters of the phase field model can be related to the surface energy 𝜎 and the interfacial thickness 𝛿 by
𝜅 = 6𝜎𝛿
6𝜎
𝑊 =
𝛿
𝑇𝑚 𝛽
𝑀𝜑 = .
6𝐿𝛿
We take 𝛿 ≈ ∆𝑥.
>>> delta = 1.5 * dx
>>> sigma = 3.7e-5 # J / cm**2
>>> beta = 0.33 # cm / (K s)
>>> kappa = 6 * sigma * delta # J / cm
>>> W = 6 * sigma / delta # J / cm**3
>>> Mphi = Tm * beta / (6. * Lv * delta) # cm**3 / (J s)
>>> if __name__ == '__main__':

... displacement = L * 0.1
... else:
... displacement = L * 0.025
>>> analyticalArray = CellVariable(name="tanh", mesh=mesh,

... value=0.5 * (1 - numerix.tanh((x - (L / 2. + displacement))
... / (2 * delta))))
and make a new viewer

>>> if __name__ == '__main__':
... viewer2 = Viewer(vars = (phase, analyticalArray))
... viewer2.plot()
Now we can redefine the transient phase field equation, using the optimal form of the source term shown above


>>> eq = TransientTerm(coeff=1/Mphi) == DiffusionTerm(coeff=kappa) \
... + S0 + ImplicitSourceTerm(coeff = S1)
In order to separate the effect of forming the phase field interface from the kinetics of moving it, we first equilibrate at
the melting point. We now use the sweep() method instead of solve() because we require the residual.
>>> timeStep = 1e-6
... phase.updateOld()
... res = 1e+10
... res = eq.sweep(var=phase, dt=timeStep)
>>> if __name__ == '__main__':
... viewer2.plot()
and then quench by 1 K

>>> T.setValue(T() - 1)
In order to have a stable numerical solution, the interface must not move more than one grid point per time step, we
thus set the timestep according to the grid spacing ∆𝑥, the linear kinetic coefficient 𝛽, and the undercooling |𝑇𝑚 − 𝑇 |
Again we use the sweep() method as a replacement for solve().
>>> velocity = beta * abs(Tm - T()) # cm / s
>>> timeStep = .1 * dx / velocity # s
>>> elapsed = 0
>>> while elapsed < displacement / velocity:
... res = 1e+10
... res = eq.sweep(var=phase, dt=timeStep)
... elapsed += timeStep
... if __name__ == '__main__':
... viewer2.plot()
A hyperbolic tangent is not an exact steady-state solution given the quintic polynomial we chose for the 𝑝 function,
but it gives a reasonable approximation.
>>> print phase.allclose(analyticalArray, rtol = 5, atol = 2e-3)
1
If we had made another common choice of 𝑝(𝜑) = 𝜑2 (3 − 2𝜑), we would have found much better agreement, as
that case does give an exact tanh solution in steady state. If SciPy is available, another way to compare against the
expected result is to do a least-squared fit to determine the interface velocity and thickness
>>> try:
... def tanhResiduals(p, y, x, t):
... V, d = p
... return y - 0.5 * (1 - numerix.tanh((x - V * t - L / 2.) / (2*d)))
... from scipy.optimize import leastsq
... x = mesh.cellCenters[0]
... (V_fit, d_fit), msg = leastsq(tanhResiduals, [L/2., delta],
... args=(phase.globalValue, x.globalValue, elapsed))
... V_fit = d_fit = 0
... print "The SciPy library is unavailable to fit the interface \
... thickness and velocity"

>>> print abs(1 - V_fit / velocity) < 4.1e-2

True
>>> print abs(1 - d_fit / delta) < 2e-2
True
>>> if __name__ == '__main__':

... raw_input("Dimensional, semi-implicit. Press <return> to proceed...")
phase
1
phase
0.8
0.6
0.4
0.2
0
0 5e-4 1e-3 1.5e-3 2e-3
12.2 examples.phase.binaryCoupled
Simultaneously solve a phase-field evolution and solute diffusion problem in one-dimension.

It is straightforward to extend a phase field model to include binary alloys. As in examples.phase.simple, we
will examine a 1D problem
>>> nx = 400
>>> dx = 5e-6 # cm
>>> L = nx * dx
The Helmholtz free energy functional can be written as the integral [2] [4] [26]
𝜅𝜑 𝜅𝐶
∫︁ {︁ }︁
ℱ (𝜑, 𝐶, 𝑇 ) = 𝑓 (𝜑, 𝐶, 𝑇 ) + |∇𝜑|2 + |∇𝐶|2 𝑑𝑉
𝒱 2 2
12.2. examples.phase.binaryCoupled 103

over the volume 𝒱 as a function of phase 𝜑 1

>>> phase = CellVariable(name="phase", mesh=mesh, hasOld=1)
composition 𝐶
>>> C = CellVariable(name="composition", mesh=mesh, hasOld=1)
2
and temperature 𝑇
>>> T = Variable(name="temperature")
Frequently, the gradient energy term in concentration is ignored and we can derive governing equations
(︂ )︂
𝜕𝜑 𝜕𝑓
= 𝑀 𝜑 𝜅 𝜑 ∇2 𝜑 − (12.3)
𝜕𝑡 𝜕𝜑
for phase and

(︂ )︂
𝜕𝐶 𝜕𝑓
= ∇· 𝑀𝐶 ∇ (12.4)
𝜕𝑡 𝜕𝐶
for solute.
The free energy density 𝑓 (𝜑, 𝐶, 𝑇 ) can be constructed in many different ways. One approach is to construct free
energy densities for each of the pure components, as functions of phase, e.g.
𝑊𝐴
𝑓𝐴 (𝜑, 𝑇 ) = 𝑝(𝜑)𝑓𝐴𝑆 (𝑇 ) + (1 − 𝑝(𝜑)) 𝑓𝐴𝐿 (𝑇 ) + 𝑔(𝜑)
2
where 𝑓𝐴𝐿 (𝑇 ), 𝑓𝐵𝐿 (𝑇 ), 𝑓𝐴𝑆 (𝑇 ), and 𝑓𝐵𝑆 (𝑇 ) are the free energy densities of the pure components. There are a variety of
choices for the interpolation function 𝑝(𝜑) and the barrier function 𝑔(𝜑),
such as those shown in examples.phase.simple
>>> def p(phi):
... return phi**3 * (6 * phi**2 - 15 * phi + 10)
>>> def g(phi):

... return (phi * (1 - phi))**2
The desired thermodynamic model can then be applied to obtain 𝑓 (𝜑, 𝐶, 𝑇 ), such as for a regular solution,
𝑓 (𝜑, 𝐶, 𝑇 ) = (1 − 𝐶)𝑓𝐴 (𝜑, 𝑇 ) + 𝐶𝑓𝐵 (𝜑, 𝑇 )

+ 𝑅𝑇 [(1 − 𝐶) ln(1 − 𝐶) + 𝐶 ln 𝐶] + 𝐶(1 − 𝐶) [Ω𝑆 𝑝(𝜑) + Ω𝐿 (1 − 𝑝(𝜑))]
where
>>> R = 8.314 # J / (mol K)
is the gas constant and Ω𝑆 and Ω𝐿 are the regular solution interaction parameters for solid and liquid.
Another approach is useful when the free energy densities 𝑓 𝐿 (𝐶, 𝑇 ) and 𝑓 𝑆 (𝐶, 𝑇 ) of the alloy in the solid and liquid
phases are known. This might be the case when the two different phases have different thermodynamic models or
when one or both is obtained from a Calphad code. In this case, we can construct
[︂ ]︂
𝑊𝐴 𝑊𝐵
𝑓 (𝜑, 𝐶, 𝑇 ) = 𝑝(𝜑)𝑓 𝑆 (𝐶, 𝑇 ) + (1 − 𝑝(𝜑)) 𝑓 𝐿 (𝐶, 𝑇 ) + (1 − 𝐶) +𝐶 𝑔(𝜑).
2 2
1 We will find that we need to “sweep” this non-linear problem (see e.g. the composition-dependent diffusivity example in
examples.diffusion.mesh1D), so we declare 𝜑 and 𝐶 to retain an “old” value.
2 we are going to want to examine different temperatures in this example, so we declare 𝑇 as a Variable

When the thermodynamic models are the same in both phases, both approaches should yield the same result.
We choose the first approach and make the simplifying assumptions of an ideal solution and that
𝑓𝐴𝐿 (𝑇 ) = 0
(︀ 𝐴
)︀
𝐿𝐴 𝑇 − 𝑇𝑀
𝑓𝐴𝑆 (𝑇 ) − 𝑓𝐴𝐿 (𝑇 ) = 𝐴
𝑇𝑀
and likewise for component 𝐵.

>>> LA = 2350. # J / cm**3
>>> LB = 1728. # J / cm**3
>>> TmA = 1728. # K
>>> TmB = 1358. # K
>>> enthalpyA = LA * (T - TmA) / TmA

>>> enthalpyB = LB * (T - TmB) / TmB
This relates the difference between the free energy densities of the pure solid and pure liquid phases to the latent heat
𝐿𝐴 and the pure component melting point 𝑇𝑀 𝐴
, such that
(︀ 𝐴
)︀
𝐿𝐴 𝑇 − 𝑇𝑀 𝑊𝐴
𝑓𝐴 (𝜑, 𝑇 ) = 𝐴
𝑝(𝜑) + 𝑔(𝜑).
𝑇𝑀 2
With these assumptions

𝜕𝑓 𝜕𝑓𝐴 𝜕𝑓𝐵
= (1 − 𝐶) +𝐶
𝜕𝜑 𝜕𝜑 𝜕𝜑
{︃ (︀ )︀ (︀ )︀ }︃
𝐴 𝐵
𝐿𝐴 𝑇 − 𝑇𝑀 𝐿𝐵 𝑇 − 𝑇𝑀
{︂ }︂
′ 𝑊𝐴 𝑊𝐵
= (1 − 𝐶) 𝐴
+𝐶 𝐵
𝑝 (𝜑) + (1 − 𝐶) +𝐶 𝑔 ′ (𝜑)
𝑇𝑀 𝑇𝑀 2 2
and
[︂ ]︂ [︂ ]︂
𝜕𝑓 𝑅𝑇 𝑅𝑇
= 𝑓𝐵 (𝜑, 𝑇 ) + ln 𝐶 − 𝑓𝐴 (𝜑, 𝑇 ) + ln(1 − 𝐶)
𝜕𝐶 𝑉𝑚 𝑉𝑚
= [𝜇𝐵 (𝜑, 𝐶, 𝑇 ) − 𝜇𝐴 (𝜑, 𝐶, 𝑇 )] /𝑉𝑚
where 𝜇𝐴 and 𝜇𝐵 are the classical chemical potentials for the binary species. 𝑝′ (𝜑) and 𝑔 ′ (𝜑) are the partial derivatives
of of 𝑝 and 𝑔 with respect to 𝜑
>>> def pPrime(phi):
... return 30. * g(phi)
>>> def gPrime(phi):

... return 2. * phi * (1 - phi) * (1 - 2 * phi)
𝑉𝑚 is the molar volume, which we take to be independent of concentration and phase

>>> Vm = 7.42 # cm**3 / mol
On comparison with examples.phase.simple, we can see that the present form of the phase field equation is
identical to the one found earlier, with the source now composed of the concentration-weighted average of the source
for either pure component. We let the pure component barriers equal the previous value
>>> deltaA = deltaB = 1.5 * dx
>>> sigmaA = 3.7e-5 # J / cm**2
>>> sigmaB = 2.9e-5 # J / cm**2
>>> betaA = 0.33 # cm / (K s)

>>> betaB = 0.39 # cm / (K s)

>>> kappaA = 6 * sigmaA * deltaA # J / cm
>>> kappaB = 6 * sigmaB * deltaB # J / cm
>>> WA = 6 * sigmaA / deltaA # J / cm**3
>>> WB = 6 * sigmaB / deltaB # J / cm**3
and define the averages

>>> W = (1 - C) * WA / 2. + C * WB / 2.
>>> enthalpy = (1 - C) * enthalpyA + C * enthalpyB
We can now linearize the source exactly as before

Using the same gradient energy coefficient and phase field mobility
>>> kappa = (1 - C) * kappaA + C * kappaB
>>> Mphi = TmA * betaA / (6 * LA * deltaA)
we define the phase field equation

>>> phaseEq = (TransientTerm(1/Mphi, var=phase) == DiffusionTerm(coeff=kappa, var=phase)
... + S0 + ImplicitSourceTerm(coeff=S1, var=phase))
When coding explicitly, it is typical to simply write a function to evaluate the chemical potentials 𝜇𝐴 and 𝜇𝐵 and then
perform the finite differences necessary to calculate their gradient and divergence, e.g.,:
def deltaChemPot(phase, C, T):
return ((Vm * (enthalpyB * p(phase) + WA * g(phase)) + R * T * log(1 - C)) -
(Vm * (enthalpyA * p(phase) + WA * g(phase)) + R * T * log(C)))
for j in range(faces):
flux[j] = ((Mc[j+.5] + Mc[j-.5]) / 2) \
* (deltaChemPot(phase[j+.5], C[j+.5], T) \
- deltaChemPot(phase[j-.5], C[j-.5], T)) / dx
for j in range(cells):
diffusion = (flux[j+.5] - flux[j-.5]) / dx
where we neglect the details of the outer boundaries (j = 0 and j = N) or exactly how to translate j+.5 or j-.5
into an array index, much less the complexities of higher dimensions. FiPy can handle all of these issues automatically,
so we could just write:
chemPotA = Vm * (enthalpyA * p(phase) + WA * g(phase)) + R * T * log(C)
chemPotB = Vm * (enthalpyB * p(phase) + WB * g(phase)) + R * T * log(1-C)
flux = Mc * (chemPotB - chemPotA).faceGrad
eq = TransientTerm() == flux.divergence
Although the second syntax would essentially work as written, such an explicit implementation would be very slow.
In order to take advantage of FiPy‘s implicit solvers, it is necessary to reduce Eq. (12.4) to the canonical form of Eq.
(??), hence we must expand Eq. (12.2) as
[︃ (︀ )︀ (︀ )︀ ]︃
𝐵 𝐴
𝜕𝑓 𝐿𝐵 𝑇 − 𝑇𝑀 𝐿𝐴 𝑇 − 𝑇𝑀 𝑅𝑇 𝑊𝐵 − 𝑊𝐴
= 𝐵
− 𝐴
𝑝(𝜑) + [ln 𝐶 − ln(1 − 𝐶)] + 𝑔(𝜑)
𝜕𝐶 𝑇𝑀 𝑇𝑀 𝑉𝑚 2

In either bulk phase, ∇𝑝(𝜑) = ∇𝑔(𝜑) = 0, so we can then reduce Eq. (12.4) to
(︂ {︂ }︂)︂
𝜕𝐶 𝑅𝑇
= ∇ · 𝑀𝐶 ∇ [ln 𝐶 − ln(1 − 𝐶)]
𝜕𝑡 𝑉𝑚
[︂ ]︂
𝑀𝐶 𝑅𝑇
= ∇· ∇𝐶
𝐶(1 − 𝐶)𝑉𝑚
and, by comparison with Fick’s second law
𝜕𝐶
= ∇ · [𝐷∇𝐶] ,
𝜕𝑡
we can associate the mobility 𝑀𝐶 with the intrinsic diffusivity 𝐷𝐶 by 𝑀𝐶 ≡ 𝐷𝐶 𝐶(1 − 𝐶)𝑉𝑚 /𝑅𝑇 and write Eq.
(12.4) as
𝜕𝐶
= ∇ · (𝐷𝐶 ∇𝐶)
𝜕𝑡 (︃ {︃[︃ (︀ 𝐵
)︀ (︀ 𝐴
)︀ ]︃ }︃)︃
𝐷𝐶 𝐶(1 − 𝐶)𝑉𝑚 𝐿𝐵 𝑇 − 𝑇𝑀 𝐿𝐴 𝑇 − 𝑇𝑀 𝑊𝐵 − 𝑊𝐴
+ ∇· 𝐵
− 𝐴
∇𝑝(𝜑) + ∇𝑔(𝜑) .
𝑅𝑇 𝑇𝑀 𝑇𝑀 2
= ∇ · (𝐷𝐶 ∇𝐶)
(︃ {︃[︃ (︀ )︀ (︀ )︀ ]︃ }︃ )︃
𝐵 𝐴
𝐷𝐶 𝐶(1 − 𝐶)𝑉𝑚 𝐿𝐵 𝑇 − 𝑇𝑀 𝐿𝐴 𝑇 − 𝑇𝑀 ′ 𝑊𝐵 − 𝑊𝐴 ′
+ ∇· 𝐵
− 𝐴
𝑝 (𝜑) + 𝑔 (𝜑) ∇𝜑 .
The first term is clearly a DiffusionTerm in 𝐶. The second is a DiffusionTerm in 𝜑 with a diffusion coefficent
{︃[︃ (︀ )︀ (︀ )︀ ]︃ }︃
𝐵 𝐴
𝐷𝐶 𝐶(1 − 𝐶)𝑉𝑚 𝐿𝐵 𝑇 − 𝑇𝑀 𝐿𝐴 𝑇 − 𝑇𝑀 ′ 𝑊𝐵 − 𝑊𝐴 ′
𝐷𝜑 (𝐶, 𝜑) = 𝐵
− 𝐴
𝑝 (𝜑) + 𝑔 (𝜑) ,
such that
𝜕𝐶
= ∇ · (𝐷𝐶 ∇𝐶) + ∇ · (𝐷𝜑 ∇𝜑)
𝜕𝑡
or
>>> Dl = Variable(value=1e-5) # cm**2 / s
>>> Ds = Variable(value=1e-9) # cm**2 / s
>>> Dc = (Ds - Dl) * phase.arithmeticFaceValue + Dl
>>> Dphi = ((Dc * C.harmonicFaceValue * (1 - C.harmonicFaceValue) * Vm / (R * T))

... * ((enthalpyB - enthalpyA) * pPrime(phase.arithmeticFaceValue)
... + 0.5 * (WB - WA) * gPrime(phase.arithmeticFaceValue)))
>>> diffusionEq = (TransientTerm(var=C)

... == DiffusionTerm(coeff=Dc, var=C)
... + DiffusionTerm(coeff=Dphi, var=phase))
>>> eq = phaseEq & diffusionEq
We initialize the phase field to a step function in the middle of the domain
>>> phase.setValue(0., where=mesh.cellCenters[0] > L/2.)
and start with a uniform composition field 𝐶 = 1/2

>>> C.setValue(0.5)
In equilibrium, 𝜇𝐴 (0, 𝐶𝐿 , 𝑇 ) = 𝜇𝐴 (1, 𝐶𝑆 , 𝑇 ) and 𝜇𝐵 (0, 𝐶𝐿 , 𝑇 ) = 𝜇𝐵 (1, 𝐶𝑆 , 𝑇 ) and, for ideal solutions, we can
deduce the liquidus and solidus compositions as
(︂ )︂
𝐿𝐴 (𝑇 −𝑇𝑀𝐴
) 𝑉𝑚
1 − exp − 𝑇𝑀𝐴 𝑅𝑇
𝐶𝐿 = (︂ )︂ (︂ )︂
𝐿𝐵 (𝑇 −𝑇𝑀 𝐵
) 𝑉𝑚 𝐿𝐴 (𝑇 −𝑇𝑀
𝐴
) 𝑉𝑚
exp − 𝑇𝑀𝐵 𝑅𝑇 − exp − 𝑇𝑀𝐴 𝑅𝑇
(︃ (︀ )︀ )︃
𝐵
𝐿𝐵 𝑇 − 𝑇𝑀 𝑉𝑚
𝐶𝑆 = exp − 𝐵
𝐶𝐿
𝑇𝑀 𝑅𝑇
>>> Cl = (1. - numerix.exp(-enthalpyA * Vm / (R * T))) \

... / (numerix.exp(-enthalpyB * Vm / (R * T)) - numerix.exp(-enthalpyA * Vm / (R * T)))
>>> Cs = numerix.exp(-enthalpyB * Vm / (R * T)) * Cl
The phase fraction is predicted by the lever rule

>>> Cavg = C.cellVolumeAverage
>>> fraction = (Cl - Cavg) / (Cl - Cs)
For the special case of fraction = Cavg = 0.5, a little bit of algebra reveals that the temperature that leaves
the phase fraction unchanged is given by
>>> T.setValue((LA + LB) * TmA * TmB / (LA * TmB + LB * TmA))
In this simple, binary, ideal solution case, we can derive explicit expressions for the solidus and liquidus compositions.
In general, this may not be possible or practical. In that event, the root-finding facilities in SciPy can be used.
We’ll need a function to return the two conditions for equilibrium
(︀ 𝐴
)︀
𝐿𝐴 𝑇 − 𝑇𝑀
0 = 𝜇𝐴 (1, 𝐶𝑆 , 𝑇 ) − 𝜇𝐴 (0, 𝐶𝐿 , 𝑇 ) = 𝑉𝑚 + 𝑅𝑇 ln(1 − 𝐶𝑆 ) − 𝑅𝑇 ln(1 − 𝐶𝐿 )
𝑇𝐴
(︀ 𝑀 𝐵 )︀
𝐿𝐵 𝑇 − 𝑇𝑀
0 = 𝜇𝐵 (1, 𝐶𝑆 , 𝑇 ) − 𝜇𝐵 (0, 𝐶𝐿 , 𝑇 ) = 𝐵
𝑉𝑚 + 𝑅𝑇 ln 𝐶𝑆 − 𝑅𝑇 ln 𝐶𝐿
𝑇𝑀
>>> def equilibrium(C):

... return [numerix.array(enthalpyA * Vm
... + R * T * numerix.log(1 - C[0])
... - R * T * numerix.log(1 - C[1])),
... numerix.array(enthalpyB * Vm
... + R * T * numerix.log(C[0])
... - R * T * numerix.log(C[1]))]
and we’ll have much better luck if we also supply the Jacobian
[︃ 𝑆 𝐿 𝑆 𝐿 ]︃
𝜕(𝜇𝐴 −𝜇𝐴 ) 𝜕(𝜇𝐴 −𝜇𝐴 ) [︂ 1 1 ]︂
𝜕𝐶𝑆 𝜕𝐶𝐿 − 1−𝐶 1−𝐶𝐿
= 𝑅𝑇 1
𝑆
𝜕(𝜇𝑆 − 𝐶1𝐿
𝐿
𝐵 −𝜇𝐵 ) 𝜕(𝜇𝑆 𝐿
𝐵 −𝜇𝐵 )
𝜕𝐶𝑆 𝜕𝐶𝐿 𝐶𝑆
>>> def equilibriumJacobian(C):

... return R * T * numerix.array([[-1. / (1 - C[0]), 1. / (1 - C[1])],
... [ 1. / C[0], -1. / C[1]]])

>>> try:
... from scipy.optimize import fsolve
... CsRoot, ClRoot = fsolve(func=equilibrium, x0=[0.5, 0.5],
... fprime=equilibriumJacobian)
... ClRoot = CsRoot = 0
... print "The SciPy library is not available to calculate the solidus and \
... liquidus concentrations"
>>> print Cl.allclose(ClRoot)

1
>>> print Cs.allclose(CsRoot)
1
We plot the result against the sharp interface solution

>>> sharp = CellVariable(name="sharp", mesh=mesh)
>>> sharp.setValue(Cs, where=x < L * fraction)
>>> sharp.setValue(Cl, where=x >= L * fraction)
>>> if __name__ == '__main__':

... viewer = Viewer(vars=(phase, C, sharp),
... viewer.plot()
Because the phase field interface will not move, and because we’ve seen in earlier examples that the diffusion problem
is unconditionally stable, we need take only one very large timestep to reach equilibrium
>>> dt = 1.e5
Because the phase field equation is coupled to the composition through enthalpy and W and the diffusion equation
is coupled to the phase field through phaseTransformationVelocity, it is necessary sweep this non-linear
problem to convergence. We use the “residual” of the equations (a measure of how well they think they have solved
the given set of linear equations) as a test for how long to sweep. Because of the ConvectionTerm, the solution
matrix for diffusionEq is asymmetric and cannot be solved by the default LinearPCGSolver. Therefore, we
use a LinearLUSolver for this equation.
We now use the “sweep()” method instead of “solve()” because we require the residual.
>>> solver = LinearLUSolver(tolerance=1e-10)
>>> phase.updateOld()
>>> C.updateOld()
>>> res = 1.
>>> initialRes = None
>>> while res > 1e-4:

... res = eq.sweep(dt=dt, solver=solver)
... if initialRes is None:
... initialRes = res
... res = res / initialRes
>>> if __name__ == '__main__':

... viewer.plot()
... raw_input("stationary phase field")

1
phase
composition
sharp
0.8
0.6
0.4
0.2
0
0 5e-4 1e-3 1.5e-3 2e-3
We verify that the bulk phases have shifted to the predicted solidus and liquidus compositions
>>> print Cs.allclose(C.faceValue[X.value==0], atol=1e-2)
True
>>> print Cl.allclose(C.faceValue[X.value==L], atol=1e-2)
True
and that the phase fraction remains unchanged

>>> print fraction.allclose(phase.cellVolumeAverage, atol=2e-4)
1
while conserving mass overall

>>> print Cavg.allclose(0.5, atol=1e-8)
1
We now quench by ten degrees

>>> T.setValue(T() - 10.) # K
>>> sharp.setValue(Cs, where=x < L * fraction)

>>> sharp.setValue(Cl, where=x >= L * fraction)
Because this lower temperature will induce the phase interface to move (solidify), we will need to take much smaller
timesteps (the time scales of diffusion and of phase transformation compete with each other).
The CFL limit requires that no interface should advect more than one grid spacing in a timestep. We can get a rough
idea for the maximum timestep we can take by looking at the velocity of convection induced by phase transformation

in Eq. (12.2) (even though there is no explicit convection in the coupled form used for this example, the principle
remains the same). If we assume that the phase changes from 1 to 0 in a single grid spacing, that the diffusivity is Dl
at the interface, and that the term due to the difference in barrier heights is negligible:
𝐷𝜑
⃗𝑢𝜑 = ∇𝜑
𝐶 [︃ )︀ ]︃
𝐷𝑙 12 𝑉𝑚 𝐿𝐵 𝑇 − 𝑇𝑀
(︀ 𝐵
)︀ (︀ 𝐴
𝐿𝐴 𝑇 − 𝑇𝑀 1
≈ 𝐵
− 𝐴
𝑅𝑇 𝑇𝑀 𝑇𝑀 ∆𝑥
𝐷𝑙 12 𝑉𝑚 𝑇 𝐴 − 𝑇𝑀
𝐵
1
≈ (𝐿𝐵 + 𝐿𝐴 ) 𝑀𝐴 𝐵
𝑅𝑇 𝑇𝑀 + 𝑇𝑀 ∆𝑥
≈ 0.28 cm/s
To get a CFL = ⃗𝑢𝜑 ∆𝑡/∆𝑥 < 1, we need a time step of about 10−5 s.
>>> dt = 1.e-5
>>> if __name__ == '__main__':

... timesteps = 100
... else:
... timesteps = 10
>>> for i in range(timesteps):

... C.updateOld()
... res = 1e+10
... res = eq.sweep(dt=dt, solver=solver)
... if __name__ == '__main__':
... viewer.plot()
>>> if __name__ == '__main__':

... raw_input("moving phase field")

1
phase
composition
sharp
0.8
0.6
0.4
0.2
0
0 5e-4 1e-3 1.5e-3 2e-3
We see that the composition on either side of the interface approach the sharp-interface solidus and liquidus, but it will
take a great many more timesteps to reach equilibrium. If we waited sufficiently long, we could again verify the final
concentrations and phase fraction against the expected values.
12.3 examples.phase.quaternary
Solve a phase-field evolution and diffusion of four species in one-dimension.

The same procedure used to construct the two-component phase field diffusion problem in
examples.phase.binary can be used to build up a system of multiple components. Once again, we’ll
focus on 1D.
>>> nx = 400
>>> dx = 0.01
>>> L = nx * dx
We consider a free energy density 𝑓 (𝜑, 𝐶0 , . . . , 𝐶𝑁 , 𝑇 ) that is a function of phase 𝜑

>>> phase = CellVariable(mesh=mesh, name='phase', value=1., hasOld=1)
interstitial components 𝐶0 . . . 𝐶𝑀
>>> interstitials = [
... CellVariable(mesh=mesh, name='C0', hasOld=1)
... ]

substitutional components 𝐶𝑀 +1 . . . 𝐶𝑁 −1
>>> substitutionals = [
... CellVariable(mesh=mesh, name='C1', hasOld=1),
... CellVariable(mesh=mesh, name='C2', hasOld=1),
... ]
a∑︀“solvent” 𝐶𝑁 that is constrained by the concentrations of the other substitutional species, such that 𝐶𝑁 = 1 −
𝑁 −1
𝑗=𝑀 𝐶𝑗 ,
>>> solvent = 1
>>> for Cj in substitutionals:
... solvent -= Cj
>>> solvent.name = 'CN'
and temperature 𝑇
>>> T = 1000
The free energy density of such a system can be written as

𝑁 [︂ ]︂
∑︁ 𝐶𝑗
𝑓 (𝜑, 𝐶0 , . . . , 𝐶𝑁 , 𝑇 ) = 𝐶𝑗 𝜇∘𝑗 (𝜑, 𝑇 ) + 𝑅𝑇 ln
𝑗=0
𝜌
where
>>> R = 8.314 # J / (mol K)
is the gas constant. As in the binary case,

𝑊𝑗
𝜇∘𝑗 (𝜑, 𝑇 ) = 𝑝(𝜑)𝜇∘𝑆 ∘𝐿
𝑗 (𝑇 ) + (1 − 𝑝(𝜑)) 𝜇𝑗 (𝑇 ) + 𝑔(𝜑)
2
is constructed with the free energies of the pure components in each phase, given the “tilting” function
>>> def p(phi):
... return phi**3 * (6 * phi**2 - 15 * phi + 10)
and the “double well” function

>>> def g(phi):
... return (phi * (1 - phi))**2
We consider a very simplified model that has partial molar volumes 𝑉¯0 = · · · = 𝑉¯𝑀 = 0 for the “interstitials” and
𝑉¯𝑀 +1 = · · · = 𝑉¯𝑁 = 1 for the “substitutionals”. This approximation has been used in a number of models where
density effects are ignored, including the treatment of electrons in electrodeposition processes [28] [29]. Under these
constraints
𝑁
𝜕𝑓 ∑︁ 𝜕𝑓𝑗
= 𝐶𝑗
𝜕𝜑 𝑗=0 𝜕𝜑
𝑁
[︂ ]︂
∑︁ 𝑊𝑗 ′
= 𝐶𝑗 𝜇∘𝑆𝐿𝑗 (𝑇 )𝑝′
(𝜑) + 𝑔 (𝜑)
𝑗=0
2
[︂ ]︂
𝜕𝑓 𝐶𝑗
= 𝜇∘𝑗 (𝜑, 𝑇 ) + 𝑅𝑇 ln
𝜕𝐶𝑗 𝜌
= 𝜇𝑗 (𝜑, 𝐶𝑗 , 𝑇 ) for 𝑗 = 0 . . . 𝑀
12.3. examples.phase.quaternary 113

and
[︂ ]︂ [︂ ]︂
𝜕𝑓 𝐶𝑗 𝐶𝑁
= 𝜇∘𝑗 (𝜑, 𝑇 ) + 𝑅𝑇 ln − 𝜇∘𝑁 (𝜑, 𝑇 ) + 𝑅𝑇 ln
𝜕𝐶𝑗 𝜌 𝜌
= [𝜇𝑗 (𝜑, 𝐶𝑗 , 𝑇 ) − 𝜇𝑁 (𝜑, 𝐶𝑁 , 𝑇 )] for 𝑗 = 𝑀 + 1 . . . 𝑁 − 1
where 𝜇∘𝑆𝐿
𝑗 𝑗 (𝑇 ) − 𝜇𝑗 (𝑇 ) and where 𝜇𝑗 is the classical chemical potential of component 𝑗 for the binary
(𝑇 ) ≡ 𝜇∘𝑆 ∘𝐿
∑︀𝑀
species and 𝜌 = 1 + 𝑗=0 𝐶𝑗 is the total molar density.
>>> rho = 1.
>>> for Cj in interstitials:
... rho += Cj
𝑝′ (𝜑) and 𝑔 ′ (𝜑) are the partial derivatives of of 𝑝 and 𝑔 with respect to 𝜑
>>> def pPrime(phi):
... return 30. * g(phi)
>>> def gPrime(phi):

... return 2. * phi * (1 - phi) * (1 - 2 * phi)
We “cook” the standard potentials to give the desired solid and liquid concentrations, with a solid phase rich in
interstitials and the solvent and a liquid phase rich in the two substitutional species.
>>> interstitials[0].S = 0.3
>>> interstitials[0].L = 0.4
>>> substitutionals[0].S = 0.4
>>> substitutionals[0].L = 0.3
>>> substitutionals[1].S = 0.2
>>> substitutionals[1].L = 0.1
>>> solvent.S = 1.
>>> solvent.L = 1.
... solvent.S -= Cj.S
... solvent.L -= Cj.L
>>> rhoS = rhoL = 1.

... rhoS += Cj.S
... rhoL += Cj.L
>>> for Cj in interstitials + substitutionals + [solvent]:

... Cj.standardPotential = R * T * (numerix.log(Cj.L/rhoL)
... - numerix.log(Cj.S/rhoS))

... Cj.diffusivity = 1.
... Cj.barrier = 0.

... Cj.diffusivity = 1.
... Cj.barrier = R * T
>>> solvent.barrier = R * T

We create the phase equation

𝑁 [︂ ]︂
1 𝜕𝜑 2
∑︁
∘𝑆𝐿 ′ 𝑊𝑗 ′
= 𝜅𝜑 ∇ 𝜑 − 𝐶𝑗 𝜇𝑗 (𝑇 )𝑝 (𝜑) + 𝑔 (𝜑)
𝑀𝜑 𝜕𝑡 𝑗=0
2
with a semi-implicit source just as in examples.phase.simple and examples.phase.binary

>>> enthalpy = 0.
>>> barrier = 0.
>>> for Cj in interstitials + substitutionals + [solvent]:
... enthalpy += Cj * Cj.standardPotential
... barrier += Cj * Cj.barrier
>>> mPhi = -((1 - 2 * phase) * barrier + 30 * phase * (1 - phase) * enthalpy)

>>> dmPhidPhi = 2 * barrier - 30 * (1 - 2 * phase) * enthalpy
>>> phase.mobility = 1.
>>> phase.gradientEnergy = 25
>>> phase.equation = TransientTerm(coeff=1/phase.mobility) \
... == DiffusionTerm(coeff=phase.gradientEnergy) \
... + S0 + ImplicitSourceTerm(coeff = S1)
We could construct the diffusion equations one-by-one, in the manner of examples.phase.binary, but it is
better to take advantage of the full scripting power of the Python language, where we can easily loop over components
or even make “factory” functions if we desire. For the interstitial diffusion equations, we arrange in canonical form as
before:
𝜕𝐶𝑗
= 𝐷𝑗 ∇2 𝐶𝑗
𝜕𝑡
⏟ ⏞ ⏟ ⏞
transient diffusion
counter diffusion
⎧ ⎫
⎪ phase transformation ⏞ ⏟ ⎪
⎪
⎪ ⏞ [︂ ⏟ ⎪
⎪
⎪ ]︂ 𝑀 ⎪
𝐶𝑗 ⎨ 𝜌 ∘𝑆𝐿 𝑊 𝑗
∑︁ ⎬
+ 𝐷𝑗 ∇ · ∑︀𝑀 𝜇𝑗 ∇𝑝(𝜑) + ∇𝑔(𝜑) − ∇𝐶𝑖
1 + 𝑘=0 𝐶𝑘 ⎪
⎪ 𝑅𝑇 2 𝑖=0
⎪
⎪
𝑘̸=𝑗
⎪ ⎪
⎪
⎩ 𝑖̸=𝑗 ⎪
⎭
⏟ ⏞
convection

... phaseTransformation = (rho.harmonicFaceValue / (R * T)) \
... * (Cj.standardPotential * p(phase).faceGrad
... + 0.5 * Cj.barrier * g(phase).faceGrad)
...
... CkSum = CellVariable(mesh=mesh, value=0.)
... for Ck in [Ck for Ck in interstitials if Ck is not Cj]:
... CkSum += Ck
...
... counterDiffusion = CkSum.faceGrad
...
... convectionCoeff = counterDiffusion + phaseTransformation
... convectionCoeff *= (Cj.diffusivity
... / (1. + CkSum.harmonicFaceValue))
...

... Cj.equation = (TransientTerm()

... == DiffusionTerm(coeff=Cj.diffusivity)
... + PowerLawConvectionTerm(coeff=convectionCoeff))
The canonical form of the substitutional diffusion equations is
𝜕𝐶𝑗
= 𝐷𝑗 ∇2 𝐶𝑗
𝜕𝑡
⏟ ⏞ ⏟ ⏞
transient diffusion
phase transformation counter diffusion
⎧ ⎫
⎪
⎪ ⏞ ⏟ ⏞ ⏟ ⎪
⎪
⎪
⎪ [︂ ]︂ 𝑁 −1
⎪
⎪
𝐶𝑗 ⎨ 𝐶𝑁 (︀ ∘𝑆𝐿 ∘𝑆𝐿 𝑊 𝑗 − 𝑊 𝑁
∑︁ ⎬
+ 𝐷𝑗 ∇ ·
)︀
∑︀𝑁 −1 𝜇𝑗 − 𝜇𝑁 ∇𝑝(𝜑) + ∇𝑔(𝜑) + ∇𝐶𝑖
⎪ 𝑅𝑇
1 − 𝑘=𝑀 +1 𝐶𝑘 ⎪ 2 ⎪
⎪ 𝑖=𝑀 +1 ⎪
⎪
𝑘̸=𝑗 ⎪
⎩ 𝑖̸=𝑗 ⎪
⎭
⏟ ⏞
convection

... phaseTransformation = (solvent.harmonicFaceValue / (R * T)) \
... * ((Cj.standardPotential - solvent.standardPotential) * p(phase).faceGrad
... + 0.5 * (Cj.barrier - solvent.barrier) * g(phase).faceGrad)
...
... CkSum = CellVariable(mesh=mesh, value=0.)
... for Ck in [Ck for Ck in substitutionals if Ck is not Cj]:
... CkSum += Ck
...
... counterDiffusion = CkSum.faceGrad
...
... convectionCoeff = counterDiffusion + phaseTransformation
... convectionCoeff *= (Cj.diffusivity
... / (1. - CkSum.harmonicFaceValue))
...
... Cj.equation = (TransientTerm()
... == DiffusionTerm(coeff=Cj.diffusivity)
... + PowerLawConvectionTerm(coeff=convectionCoeff))
We start with a sharp phase boundary

{︃
1 for 𝑥 ≤ 𝐿/2,
𝜉=
0 for 𝑥 > 𝐿/2,
>>> phase.setValue(0., where=x > L / 2)
and with uniform concentration fields, initially equal to the average of the solidus and liquidus concentrations
>>> for Cj in interstitials + substitutionals:
... Cj.setValue((Cj.S + Cj.L) / 2.)
If we’re running interactively, we create a viewer

>>> if __name__ == '__main__':

... viewer = Viewer(vars=([phase]
... + interstitials + substitutionals
... + [solvent]),
... datamin=0, datamax=1)
... viewer.plot()
and again iterate to equilibrium

>>> solver = DefaultAsymmetricSolver(tolerance=1e-10)
>>> dt = 10000
... for field in [phase] + substitutionals + interstitials:
... field.updateOld()
... phase.equation.solve(var = phase, dt = dt)
... for field in substitutionals + interstitials:
... field.equation.solve(var = field,
... dt = dt,
... solver = solver)
... if __name__ == '__main__':
... viewer.plot()
1
phase
C0
C1
C2
CN
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
We can confirm that the far-field phases have remained separated

>>> print numerix.allclose(phase.faceValue[X.value==0], 1.0, rtol = 1e-5, atol = 1e-5)
True
>>> print numerix.allclose(phase.faceValue[X.value==L], 0.0, rtol = 1e-5, atol = 1e-5)

True
and that the concentration fields have appropriately segregated into their equilibrium values in each phase
>>> equilibrium = True
>>> for Cj in interstitials + substitutionals:
... equilibrium &= numerix.allclose(Cj.faceValue[X.value==0], Cj.S, rtol = 3e-3, atol = 3e-3).val
... equilibrium &= numerix.allclose(Cj.faceValue[X.value==L], Cj.L, rtol = 3e-3, atol = 3e-3).val
>>> print equilibrium
True
12.4 examples.phase.anisotropy
Solve a dendritic solidification problem.

To convert a liquid material to a solid, it must be cooled to a temperature below its melting point (known as “under-
cooling” or “supercooling”). The rate of solidification is often assumed (and experimentally found) to be proportional
to the undercooling. Under the right circumstances, the solidification front can become unstable, leading to dendritic
patterns. Warren, Kobayashi, Lobkovsky and Carter [9] have described a phase field model (“Allen-Cahn”, “non-
conserved Ginsberg-Landau”, or “model A” of Hohenberg & Halperin) of such a system, including the effects of
discrete crystalline orientations (anisotropy).
We start with a regular 2D Cartesian mesh
>>> dx = dy = 0.025
>>> if __name__ == '__main__':
... nx = ny = 500
... else:
... nx = ny = 20
and we’ll take fixed timesteps

>>> dt = 5e-4
We consider the simultaneous evolution of a “phase field” variable 𝜑 (taken to be 0 in the liquid phase and 1 in the
solid)
>>> phase = CellVariable(name=r'$\phi$', mesh=mesh, hasOld=True)
and a dimensionless undercooling ∆𝑇 (∆𝑇 = 0 at the melting point)

>>> dT = CellVariable(name=r'$\Delta T$', mesh=mesh, hasOld=True)
The hasOld flag causes the storage of the value of variable from the previous timestep. This is necessary for solving
equations with non-linear coefficients or for coupling between PDEs.
The governing equation for the temperature field is the heat flux equation, with a source due to the latent heat of
solidification
𝜕∆𝑇 𝜕𝜑
= 𝐷𝑇 ∇2 ∆𝑇 +
𝜕𝑡 𝜕𝑡
>>> DT = 2.25
>>> heatEq = (TransientTerm()
... == DiffusionTerm(DT)
... + (phase - phase.old) / dt)

The governing equation for the phase field is

𝜕𝜑
𝜏𝜑 = ∇ · D∇𝜑 + 𝜑(1 − 𝜑)𝑚(𝜑, ∆𝑇 )
𝜕𝑡
where
1 𝜅1
𝑚(𝜑, ∆𝑇 ) = 𝜑 − − arctan (𝜅2 ∆𝑇 )
2 𝜋
represents a source of anisotropy. The coefficient D is an anisotropic diffusion tensor in two dimensions
[︃ ]︃
𝜕𝛽
1 + 𝑐𝛽 −𝑐
D = 𝛼2 (1 + 𝑐𝛽) 𝜕𝛽
𝜕𝜓
𝑐 𝜕𝜓 1 + 𝑐𝛽
1−Φ2 𝜕𝜑/𝜕𝑦
where 𝛽 = 1+Φ2 , 2 𝜓 , 𝜓 = 𝜃 + arctan 𝜕𝜑/𝜕𝑥 , 𝜃 is the orientation, and 𝑁 is the symmetry.
(︀ 𝑁 )︀
Φ = tan
>>> alpha = 0.015
>>> c = 0.02
>>> N = 6.
>>> theta = numerix.pi / 8.
>>> psi = theta + numerix.arctan2(phase.faceGrad[1],
... phase.faceGrad[0])
>>> Phi = numerix.tan(N * psi / 2)
>>> PhiSq = Phi**2
>>> beta = (1. - PhiSq) / (1. + PhiSq)
>>> DbetaDpsi = -N * 2 * Phi / (1 + PhiSq)
>>> Ddia = (1.+ c * beta)
>>> Doff = c * DbetaDpsi
>>> I0 = Variable(value=((1,0), (0,1)))
>>> I1 = Variable(value=((0,-1), (1,0)))
>>> D = alpha**2 * (1.+ c * beta) * (Ddia * I0 + Doff * I1)
With these expressions defined, we can construct the phase field equation as
>>> tau = 3e-4
>>> kappa1 = 0.9
>>> kappa2 = 20.
>>> phaseEq = (TransientTerm(tau)
... == DiffusionTerm(D)
... + ImplicitSourceTerm((phase - 0.5 - kappa1 / numerix.pi * numerix.arctan(kappa2 * dT))
... * (1 - phase)))
We seed a circular solidified region in the center

>>> radius = dx * 5.
>>> C = (nx * dx / 2, ny * dy / 2)
>>> phase.setValue(1., where=((x - C[0])**2 + (y - C[1])**2) < radius**2)
and quench the entire simulation domain below the melting point
>>> dT.setValue(-0.5)
In a real solidification process, dendritic branching is induced by small thermal fluctuations along an otherwise smooth
surface, but the granularity of the Mesh is enough “noise” in this case, so we don’t need to explicitly introduce
randomness, the way we did in the Cahn-Hilliard problem.
FiPy’s viewers are utilitarian, striving to let the user see something, regardless of their operating system or installed
packages, so you won’t be able to simultaneously view two fields “out of the box”, but, because all of Python is
accessible and FiPy is object oriented, it is not hard to adapt one of the existing viewers to create a specialized display:
12.4. examples.phase.anisotropy 119

>>> if __name__ == "__main__":

... try:
... import pylab
... class DendriteViewer(Matplotlib2DGridViewer):
... def __init__(self, phase, dT, title=None, limits={}, **kwlimits):
... self.phase = phase
... self.contour = None
... Matplotlib2DGridViewer.__init__(self, vars=(dT,), title=title,
... cmap=pylab.cm.hot,
... limits=limits, **kwlimits)
...
... def _plot(self):
... Matplotlib2DGridViewer._plot(self)
...
... if self.contour is not None:
... for c in self.contour.collections:
... c.remove()
...
... mesh = self.phase.mesh
... shape = mesh.shape
... x, y = mesh.cellCenters
... z = self.phase.value
... x, y, z = [a.reshape(shape, order="FORTRAN") for a in (x, y, z)]
...
... self.contour = pylab.contour(x, y, z, (0.5,))
...
... viewer = DendriteViewer(phase=phase, dT=dT,
... title=r"%s & %s" % (phase.name, dT.name),
... datamin=-0.1, datamax=0.05)
... viewer = MultiViewer(viewers=(Viewer(vars=phase),
... Viewer(vars=dT,
... datamin=-0.5,
... datamax=0.5)))
and iterate the solution in time, plotting as we go,

>>> if __name__ == '__main__':
... steps = 10000
... else:
... steps = 10
>>> for i in range(steps):
... dT.updateOld()
... phaseEq.solve(phase, dt=dt)
... heatEq.solve(dT, dt=dt)
... if __name__ == "__main__" and (i % 10 == 0):
... viewer.plot()

The non-uniform temperature results from the release of latent heat at the solidifying interface. The dendrite arms
grow fastest where the temperature gradient is steepest.
We note that this FiPy simulation is written in about 50 lines of code (excluding the custom viewer), compared with
over 800 lines of (fairly lucid) FORTRAN code used for the figures in [9].
12.5 examples.phase.impingement.mesh40x1
Solve for the impingement of two grains in one dimension.

In this example we solve a coupled phase and orientation equation on a one dimensional grid. This is another aspect
of the model of Warren, Kobayashi, Lobkovsky and Carter [9]
>>> nx = 40
>>> Lx = 2.5 * nx / 100.
12.5. examples.phase.impingement.mesh40x1 121

>>> dx = Lx / nx
This problem simulates the wet boundary that forms between grains of different orientations. The phase equation is
given by
𝜕𝜑
𝜏𝜑 = 𝛼2 ∇2 𝜑 + 𝜑(1 − 𝜑)𝑚1 (𝜑, 𝑇 ) − 2𝑠𝜑|∇𝜃| − 𝜖2 𝜑|∇𝜃|2
𝜕𝑡
where
1
𝑚1 (𝜑, 𝑇 ) = 𝜑 − − 𝑇 𝜑(1 − 𝜑)
2
and the orientation equation is given by
[︂ (︂ )︂ ]︂
2 𝜕𝜃 𝑠
𝑃 (𝜖|∇𝜃|)𝜏𝜃 𝜑 = ∇· 𝜑 2 2
+ 𝜖 ∇𝜃
𝜕𝑡 |∇𝜃|
where
𝜇
𝑃 (𝑤) = 1 − exp (−𝛽𝑤) + exp (−𝛽𝑤)
𝜖
The initial conditions for this problem are set such that 𝜑 = 1 for 0 ≤ 𝑥 ≤ 𝐿𝑥 and
{︃
1 for 0 ≤ 𝑥 < 𝐿𝑥 /2,
𝜃=
0 for 𝐿𝑥 /2 ≤ 𝑥 ≤ 𝐿𝑥 .
Here the phase and orientation equations are solved with an explicit and implicit technique respectively.
The parameters for these equations are
>>> timeStepDuration = 0.02
>>> phaseTransientCoeff = 0.1
>>> thetaSmallValue = 1e-6
>>> beta = 1e5
>>> mu = 1e3
>>> thetaTransientCoeff = 0.01
>>> gamma= 1e3
>>> epsilon = 0.008
>>> s = 0.01
>>> alpha = 0.015
The system is held isothermal at

>>> temperature = 1.
and is initially solid everywhere

>>> phase = CellVariable(
... name='phase field',
... mesh=mesh,
... value=1.
... )
Because theta is an 𝑆 1 -valued variable (i.e. it maps to the circle) and thus intrinsically has 2𝜋-peridocity, we must
use ModularVariable instead of a CellVariable. A ModularVariable confines theta to −𝜋 < 𝜃 ≤ 𝜋
by adding or subtracting 2𝜋 where necessary and by defining a new subtraction operator between two angles.

>>> theta = ModularVariable(

... name='theta',
... mesh=mesh,
... value=1.,
... hasOld=1
... )
The left and right halves of the domain are given different orientations.
>>> theta.setValue(0., where=mesh.cellCenters[0] > Lx / 2.)
The phase equation is built in the following way.

>>> mPhiVar = phase - 0.5 + temperature * phase * (1 - phase)
The source term is linearized in the manner demonstrated in examples.phase.simple (Kobayashi, semi-
implicit).
>>> thetaMag = theta.old.grad.mag
>>> implicitSource = mPhiVar * (phase - (mPhiVar < 0))
>>> implicitSource += (2 * s + epsilon**2 * thetaMag) * thetaMag
The phase equation is constructed.

>>> phaseEq = TransientTerm(phaseTransientCoeff) \
... == ExplicitDiffusionTerm(alpha**2) \
... - ImplicitSourceTerm(implicitSource) \
... + (mPhiVar > 0) * mPhiVar * phase
The theta equation is built in the following way. The details for this equation are fairly involved, see J.A. Warren et
al.. The main detail is that a source must be added to correct for the discretization of theta on the circle.
>>> phaseMod = phase + ( phase < thetaSmallValue ) * thetaSmallValue
>>> phaseModSq = phaseMod * phaseMod
>>> expo = epsilon * beta * theta.grad.mag
>>> expo = (expo < 100.) * (expo - 100.) + 100.
>>> pFunc = 1. + numerix.exp(-expo) * (mu / epsilon - 1.)
>>> phaseFace = phase.arithmeticFaceValue

>>> phaseSq = phaseFace * phaseFace
>>> gradMag = theta.faceGrad.mag
>>> eps = 1. / gamma / 10.
>>> gradMag += (gradMag < eps) * eps
>>> IGamma = (gradMag > 1. / gamma) * (1 / gradMag - gamma) + gamma
>>> diffusionCoeff = phaseSq * (s * IGamma + epsilon**2)
The source term requires the evaluation of the face gradient without the modular operator.
thetagetFaceGradNoMod() evelautes the gradient without modular arithmetic.
>>> thetaGradDiff = theta.faceGrad - theta.faceGradNoMod
>>> sourceCoeff = (diffusionCoeff * thetaGradDiff).divergence
Finally the theta equation can be constructed.

>>> thetaEq = TransientTerm(thetaTransientCoeff * phaseModSq * pFunc) == \
... DiffusionTerm(diffusionCoeff) \
... + sourceCoeff
If the example is run interactively, we create viewers for the phase and orientation variables.

>>> if __name__ == '__main__':

... phaseViewer = Viewer(vars=phase, datamin=0., datamax=1.)
... thetaProductViewer = Viewer(vars=theta,
... datamin=-pi, datamax=pi)
... phaseViewer.plot()
... thetaProductViewer.plot()
we iterate the solution in time, plotting as we go if running interactively,

>>> steps = 10
... theta.updateOld()
... thetaEq.solve(theta, dt = timeStepDuration)
... phaseEq.solve(phase, dt = timeStepDuration)
... if __name__ == '__main__':
The solution is compared with test data. The test data was created with steps = 10 with a FORTRAN code written
by Ryo Kobayashi for phase field modeling. The following code opens the file mesh40x1.gz extracts the data and
compares it with the theta variable.
>>> import os
>>> testData = numerix.loadtxt(os.path.splitext(__file__)[0] + '.gz')
>>> testData = CellVariable(mesh=mesh, value=testData)
>>> print theta.allclose(testData)
1
12.6 examples.phase.impingement.mesh20x20
Solve for the impingement of four grains in two dimensions.

In the following examples, we solve the same set of equations as in
examples.phase.impingement.mesh40x1 with different initial conditions and a 2D mesh:
>>> from fipy.tools.parser import parse
>>> numberOfElements = parse('--numberOfElements', action = 'store',

... type = 'int', default = 400)
>>> numberOfSteps = parse('--numberOfSteps', action = 'store',
... type = 'int', default = 10)
>>> steps = numberOfSteps

>>> N = int(numerix.sqrt(numberOfElements))
>>> L = 2.5 * N / 100.
>>> dL = L / N
>>> mesh = Grid2D(dx=dL, dy=dL, nx=N, ny=N)
The initial conditions are given by 𝜑 = 1 and
for 𝑥2 − 𝑦 2 < 𝐿/2,

⎧ 2𝜋
⎪
⎪ 3
⎨ −2𝜋 for (𝑥 − 𝐿)2 − 𝑦 2 < 𝐿/2,
⎪
𝜃 = −2𝜋 3
⎪ + 0.3 for 𝑥2 − (𝑦 − 𝐿)2 < 𝐿/2,
⎩ 3
⎪
⎪
2𝜋
3 for (𝑥 − 𝐿)2 − (𝑦 − 𝐿)2 < 𝐿/2.

This defines four solid regions with different orientations. Solidification occurs and then boundary wetting occurs
where the orientation varies.
The parameters for this example are
>>> timeStepDuration = 0.02
>>> phaseTransientCoeff = 0.1
>>> beta = 1e5
>>> mu = 1e3
>>> thetaTransientCoeff = 0.01
>>> gamma= 1e3
>>> epsilon = 0.008
>>> s = 0.01
>>> alpha = 0.015
The system is held isothermal at

and is initialized to liquid everywhere

>>> phase = CellVariable(name='phase field', mesh=mesh)
The orientation is initialized to a uniform value to denote the randomly oriented liquid phase
>>> theta = ModularVariable(
... name='theta',
... mesh=mesh,
... value=-numerix.pi + 0.0001,
... hasOld=1
... )
Four different solid circular domains are created at each corner of the domain with appropriate orientations
>>> for a, b, thetaValue in ((0., 0., 2. * numerix.pi / 3.),
... (L, 0., -2. * numerix.pi / 3.),
... (0., L, -2. * numerix.pi / 3. + 0.3),
... (L, L, 2. * numerix.pi / 3.)):
... segment = (x - a)**2 + (y - b)**2 < (L / 2.)**2
... phase.setValue(1., where=segment)
... theta.setValue(thetaValue, where=segment)
The phase equation is built in the following way. The source term is linearized in the manner demonstrated in
examples.phase.simple (Kobayashi, semi-implicit). Here we use a function to build the equation, so that it
can be reused later.
>>> def buildPhaseEquation(phase, theta):
...
... mPhiVar = phase - 0.5 + temperature * phase * (1 - phase)
... thetaMag = theta.old.grad.mag
... implicitSource = mPhiVar * (phase - (mPhiVar < 0))
... implicitSource += (2 * s + epsilon**2 * thetaMag) * thetaMag
...
... return TransientTerm(phaseTransientCoeff) == \
... ExplicitDiffusionTerm(alpha**2) \
... - ImplicitSourceTerm(implicitSource) \
... + (mPhiVar > 0) * mPhiVar * phase

>>> phaseEq = buildPhaseEquation(phase, theta)
al.. The main detail is that a source must be added to correct for the discretization of theta on the circle. The source
term requires the evaluation of the face gradient without the modular operators.
>>> def buildThetaEquation(phase, theta):
...
... phaseMod = phase + ( phase < thetaSmallValue ) * thetaSmallValue
... phaseModSq = phaseMod * phaseMod
... expo = epsilon * beta * theta.grad.mag
... expo = (expo < 100.) * (expo - 100.) + 100.
... pFunc = 1. + numerix.exp(-expo) * (mu / epsilon - 1.)
...
... phaseFace = phase.arithmeticFaceValue
... phaseSq = phaseFace * phaseFace
... gradMag = theta.faceGrad.mag
... eps = 1. / gamma / 10.
... gradMag += (gradMag < eps) * eps
... IGamma = (gradMag > 1. / gamma) * (1 / gradMag - gamma) + gamma
... diffusionCoeff = phaseSq * (s * IGamma + epsilon**2)
...
... thetaGradDiff = theta.faceGrad - theta.faceGradNoMod
... sourceCoeff = (diffusionCoeff * thetaGradDiff).divergence
...
... return TransientTerm(thetaTransientCoeff * phaseModSq * pFunc) == \
... DiffusionTerm(diffusionCoeff) \
... + sourceCoeff
>>> thetaEq = buildThetaEquation(phase, theta)
If the example is run interactively, we create viewers for the phase and orientation variables. Rather than viewing the
raw orientation, which is not meaningful in the liquid phase, we weight the orientation by the phase
>>> if __name__ == '__main__':
... phaseViewer = Viewer(vars=phase, datamin=0., datamax=1.)
... thetaProd = -numerix.pi + phase * (theta + numerix.pi)
... thetaProductViewer = Viewer(vars=thetaProd,
... datamin=-numerix.pi, datamax=numerix.pi)
The solution will be tested against data that was created with steps = 10 with a FORTRAN code written by
Ryo Kobayashi for phase field modeling. The following code opens the file mesh20x20.gz extracts the data and
compares it with the theta variable.
>>> import os
>>> testData = numerix.loadtxt(os.path.splitext(__file__)[0] + '.gz').flat
We step the solution in time, plotting as we go if running interactively,

... thetaEq.solve(theta, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, tolerance=1e-
... phaseEq.solve(phase, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, tolerance=1e-
... if __name__ == '__main__':

The solution is compared against Ryo Kobayashi’s test data

>>> print theta.allclose(testData, rtol=1e-7, atol=1e-7)
1
The following code shows how to restart a simulation from some saved data. First, reset the variables to their original
values.
>>> phase.setValue(0)
>>> theta.setValue(-numerix.pi + 0.0001)
>>> for a, b, thetaValue in ((0., 0., 2. * numerix.pi / 3.),
... (L, 0., -2. * numerix.pi / 3.),
... (0., L, -2. * numerix.pi / 3. + 0.3),
... (L, L, 2. * numerix.pi / 3.)):
... segment = (x - a)**2 + (y - b)**2 < (L / 2.)**2
... phase.setValue(1., where=segment)
... theta.setValue(thetaValue, where=segment)
Step through half the time steps.

>>> for i in range(steps // 2):
... thetaEq.solve(theta, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, tolerance=1e-
... phaseEq.solve(phase, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, tolerance=1e-
We confirm that the solution has not yet converged to that given by Ryo Kobayashi’s FORTRAN code:
>>> print theta.allclose(testData)
0
We save the variables to disk.

>>> (f, filename) = dump.write({'phase' : phase, 'theta' : theta}, extension = '.gz')
and then recall them to test the data pickling mechanism

>>> data = dump.read(filename, f)
>>> newPhase = data['phase']
>>> newTheta = data['theta']
>>> newThetaEq = buildThetaEquation(newPhase, newTheta)
>>> newPhaseEq = buildPhaseEquation(newPhase, newTheta)
and finish the iterations,

>>> for i in range(steps // 2):
... newTheta.updateOld()
... newThetaEq.solve(newTheta, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, toleran
... newPhaseEq.solve(newPhase, dt=timeStepDuration, solver=GeneralSolver(iterations=2000, toleran
The solution is compared against Ryo Kobayashi’s test data

>>> print newTheta.allclose(testData, rtol=1e-7)
1
12.7 examples.phase.polyxtal
Solve the dendritic growth of nuclei and subsequent grain impingement.
12.7. examples.phase.polyxtal 127

>>> dx = dy = 0.025
>>> if __name__ == "__main__":
... nx = ny = 200
... else:
... nx = ny = 200

>>> dt = 5e-4
solid)
a dimensionless undercooling ∆𝑇 (∆𝑇 = 0 at the melting point)

and an orientation −𝜋 < 𝜃 ≤ 𝜋

>>> theta = ModularVariable(name=r'$\theta$', mesh=mesh, hasOld=True)
>>> theta.value = -numerix.pi + 0.0001
solidification
𝜕∆𝑇 𝜕𝜑
= 𝐷𝑇 ∇2 ∆𝑇 + + 𝑐 (𝑇0 − 𝑇 )
𝜕𝑡 𝜕𝑡
>>> DT = 2.25
>>> q = Variable(0.)
>>> T_0 = -0.1
>>> heatEq = (TransientTerm()
... == DiffusionTerm(DT)
... + (phase - phase.old) / dt
... + q * T_0 - ImplicitSourceTerm(q))

𝜕𝜑
𝜏𝜑 = ∇ · D∇𝜑 + 𝜑(1 − 𝜑)𝑚(𝜑, ∆𝑇 ) − 2𝑠𝜑|∇𝜃| − 𝜖2 𝜑|∇𝜃|2
𝜕𝑡
where
1 𝜅1
𝑚(𝜑, ∆𝑇 ) = 𝜑 − − arctan (𝜅2 ∆𝑇 )
2 𝜋

[︃ ]︃
𝜕𝛽
2 1 + 𝑐𝛽 −𝑐 𝜕𝜓
D = 𝛼 (1 + 𝑐𝛽) 𝜕𝛽
𝑐 𝜕𝜓 1 + 𝑐𝛽
1−Φ2 𝜕𝜑/𝜕𝑦
(︀ 𝑁 )︀
Φ = tan
>>> alpha = 0.015
>>> c = 0.02
>>> N = 4.
>>> psi = theta.arithmeticFaceValue + numerix.arctan2(phase.faceGrad[1],

>>> PhiSq = Phi**2
>>> beta = (1. - PhiSq) / (1. + PhiSq)
>>> Ddia = (1.+ c * beta)

>>> I0 = Variable(value=((1,0), (0,1)))
>>> I1 = Variable(value=((0,-1), (1,0)))
>>> D = alpha**2 * Ddia * (Ddia * I0 + Doff * I1)
>>> tau_phase = 3e-4
>>> kappa1 = 0.9
>>> kappa2 = 20.
>>> epsilon = 0.008
>>> s = 0.01
>>> thetaMag = theta.grad.mag
>>> phaseEq = (TransientTerm(tau_phase)
... == DiffusionTerm(D)
... + ImplicitSourceTerm((phase - 0.5 - kappa1 / numerix.pi * numerix.arctan(kappa2 * dT))
... * (1 - phase)
... - (2 * s + epsilon**2 * thetaMag) * thetaMag))
The governing equation for orientation is given by

[︂ (︂ )︂ ]︂
𝜕𝜃 𝑠
𝑃 (𝜖|∇𝜃|)𝜏𝜃 𝜑2 = ∇ · 𝜑2 + 𝜖2 ∇𝜃
𝜕𝑡 |∇𝜃|
where
𝜇
𝑃 (𝑤) = 1 − exp (−𝛽𝑤) + exp (−𝛽𝑤)
𝜖
>>> tau_theta = 3e-3
>>> mu = 1e3
>>> gamma = 1e3
>>> beta_theta = 1e5
>>> expo = epsilon * beta_theta * theta.grad.mag
>>> expo = (expo < 100.) * (expo - 100.) + 100.
>>> Pfunc = 1. + numerix.exp(-expo) * (mu / epsilon - 1.)

>>> gradMagTheta = theta.faceGrad.mag

>>> eps = 1. / gamma / 10.
>>> gradMagTheta += (gradMagTheta < eps) * eps
>>> IGamma = (gradMagTheta > 1. / gamma) * (1 / gradMagTheta - gamma) + gamma
>>> v_theta = phase.arithmeticFaceValue * (s * IGamma + epsilon**2)
>>> D_theta = phase.arithmeticFaceValue**2 * (s * IGamma + epsilon**2)
thetagetFaceGradNoMod() evaluates the gradient without modular arithmetic.
>>> thetaEq = (TransientTerm(tau_theta * phaseMod**2 * Pfunc)
... == DiffusionTerm(D_theta)
... + (D_theta * (theta.faceGrad - theta.faceGradNoMod)).divergence)

>>> numSeeds = 10
>>> numerix.random.seed(12345)
>>> for Cx, Cy, orientation in numerix.random.random([numSeeds, 3]):
... radius = dx * 5.
... seed = ((x - Cx * nx * dx)**2 + (y - Cy * ny * dy)**2) < radius**2
... phase[seed] = 1.
... theta[seed] = numerix.pi * (2 * orientation - 1)
packages, so you the default color scheme of grain orientation won’t be very informative “out of the box”. Because
all of Python is accessible and FiPy is object oriented, it is not hard to adapt one of the existing viewers to create a
specialized display:
>>> if __name__ == "__main__":
... try:
... class OrientationViewer(Matplotlib2DGridViewer):
... def __init__(self, phase, orientation, title=None, limits={}, **kwlimits):
... Matplotlib2DGridViewer.__init__(self, vars=(orientation,), title=title,
... limits=limits, colorbar=None, **kwlimits)
...
... # make room for non-existent colorbar
... # stolen from matplotlib.colorbar.make_axes
... # https://github.com/matplotlib/matplotlib/blob
... # /ec1cd2567521c105a451ce15e06de10715f8b54d/lib
... # /matplotlib/colorbar.py#L838
... fraction = 0.15
... pb = self.axes.get_position(original=True).frozen()
... pad = 0.05
... x1 = 1.0-fraction
... pb1, pbx, pbcb = pb.splitx(x1-pad, x1)
... panchor = (1.0, 0.5)
... self.axes.set_position(pb1)
... self.axes.set_anchor(panchor)

...
... # make the gnomon
... fig = self.axes.get_figure()
... self.gnomon = fig.add_axes([0.85, 0.425, 0.15, 0.15], polar=True)
... self.gnomon.set_thetagrids([180, 270, 0, 90],
... [r"$\pm\pi$", r"$-\frac{\pi}{2}$", "$0$", r"$+\frac{\p
... frac=1.3)
... self.gnomon.set_theta_zero_location("N")
... self.gnomon.set_theta_direction(-1)
... self.gnomon.set_rgrids([1.], [""])
... N = 100
... theta = numerix.arange(-numerix.pi, numerix.pi, 2 * numerix.pi / N)
... radii = numerix.ones((N,))
... bars = self.gnomon.bar(theta, radii, width=2 * numerix.pi / N, bottom=0.0)
... colors = self._orientation_and_phase_to_rgb(orientation=numerix.array([theta]), p
... for c, t, bar in zip(colors[0], theta, bars):
... bar.set_facecolor(c)
... bar.set_edgecolor(c)
...
... def _reshape(self, var):
... '''return values of var in an 2D array'''
... return numerix.reshape(numerix.array(var),
... var.mesh.shape[::-1])[::-1]
...
... @staticmethod
... def _orientation_and_phase_to_rgb(orientation, phase):
... from matplotlib import colors
...
... hsv = numerix.empty(orientation.shape + (3,))
... hsv[..., 0] = (orientation / numerix.pi + 1) / 2.
... hsv[..., 1] = 1.
... hsv[..., 2] = phase
...
... return colors.hsv_to_rgb(hsv)
...
... @property
... def _data(self):
... '''convert phase and orientation to rgb image array
...
... orientation (-pi, pi) -> hue (0, 1)
... phase (0, 1) -> value (0, 1)
... '''
... orientation = self._reshape(self.vars[0])
... phase = self._reshape(self.phase)
...
... return self._orientation_and_phase_to_rgb(orientation, phase)
...
... self.image.set_data(self._data)
...
... from matplotlib import pyplot
... pyplot.ion()
... w, h = pyplot.figaspect(1.)
... fig = pyplot.figure(figsize=(2*w, h))
... timer = fig.text(0.1, 0.9, "t = %.3f" % 0, fontsize=18)
...
... viewer = MultiViewer(viewers=(MatplotlibViewer(vars=dT,
... cmap=pyplot.cm.hot,
... datamin=-0.5,

... datamax=0.5,
... axes=fig.add_subplot(121)),
... OrientationViewer(phase=phase,
... orientation=theta,
... title=theta.name,
... axes=fig.add_subplot(122))))
... viewer = MultiViewer(viewers=(Viewer(vars=dT,
... datamin=-0.5,
... datamax=0.5),
... Viewer(vars=phase,
... datamin=0.,
... datamax=1.),
... Viewer(vars=theta,
... datamin=-numerix.pi,
... datamax=numerix.pi)))
... viewer.plot()

>>> if __name__ == "__main__":
... total_time = 2.
... else:
... total_time = dt * 10
>>> elapsed = 0.
>>> save_interval = 0.002
>>> save_at = save_interval
>>> while elapsed < total_time:

... if elapsed > 0.3:
... q.value = 100
... dT.updateOld()
... thetaEq.solve(theta, dt=dt)
... phaseEq.solve(phase, dt=dt)
... heatEq.solve(dT, dt=dt)
... elapsed += dt
... if __name__ == "__main__" and elapsed >= save_at:
... timer.set_text("t = %.3f" % elapsed)
... viewer.plot()
... save_at += save_interval

12.8 examples.phase.polyxtalCoupled
Simultaneously solve the dendritic growth of nuclei and subsequent grain impingement.
>>> dx = dy = 0.025
>>> if __name__ == "__main__":
... nx = ny = 200
... else:
... nx = ny = 200

>>> dt = 5e-4
solid)
a dimensionless undercooling ∆𝑇 (∆𝑇 = 0 at the melting point)

and an orientation −𝜋 < 𝜃 ≤ 𝜋
12.8. examples.phase.polyxtalCoupled 133

>>> theta = ModularVariable(name=r'$\theta$', mesh=mesh, hasOld=True)

>>> theta.value = -numerix.pi + 0.0001
solidification
𝜕∆𝑇 𝜕𝜑
= 𝐷𝑇 ∇2 ∆𝑇 + + 𝑐 (𝑇0 − 𝑇 )
𝜕𝑡 𝜕𝑡
>>> DT = 2.25
>>> q = Variable(0.)
>>> T_0 = -0.1
>>> heatEq = (TransientTerm(var=dT)
... == DiffusionTerm(coeff=DT, var=dT)
... + TransientTerm(var=phase)
... + q * T_0 - ImplicitSourceTerm(coeff=q, var=dT))
𝜕𝜑
𝜏𝜑 = ∇ · D∇𝜑 + 𝜑(1 − 𝜑)𝑚(𝜑, ∆𝑇 ) − 2𝑠𝜑|∇𝜃| − 𝜖2 𝜑|∇𝜃|2
𝜕𝑡
where
1 𝜅1
𝑚(𝜑, ∆𝑇 ) = 𝜑 − − arctan (𝜅2 ∆𝑇 )
2 𝜋
[︃ ]︃
𝜕𝛽
2 1 + 𝑐𝛽 −𝑐 𝜕𝜓
D = 𝛼 (1 + 𝑐𝛽) 𝜕𝛽
𝑐 𝜕𝜓 1 + 𝑐𝛽
1−Φ2 𝜕𝜑/𝜕𝑦
(︀ 𝑁 )︀
Φ = tan
>>> alpha = 0.015
>>> c = 0.02
>>> N = 4.
>>> psi = theta.arithmeticFaceValue + numerix.arctan2(phase.faceGrad[1],

>>> PhiSq = Phi**2
>>> beta = (1. - PhiSq) / (1. + PhiSq)
>>> Ddia = (1.+ c * beta)

>>> I0 = Variable(value=((1,0), (0,1)))
>>> I1 = Variable(value=((0,-1), (1,0)))
>>> D = alpha**2 * Ddia * (Ddia * I0 + Doff * I1)
>>> tau_phase = 3e-4
>>> kappa1 = 0.9

>>> kappa2 = 20.

>>> epsilon = 0.008
>>> s = 0.01
>>> thetaMag = theta.grad.mag
>>> phaseEq = (TransientTerm(coeff=tau_phase, var=phase)
... == DiffusionTerm(coeff=D, var=phase)
... + ImplicitSourceTerm(coeff=((phase - 0.5 - kappa1 / numerix.pi * numerix.arctan(kappa2
... * (1 - phase)
... - (2 * s + epsilon**2 * thetaMag) * thetaMag),
... var=phase))
The governing equation for orientation is given by

[︂ (︂ )︂ ]︂
2 𝜕𝜃 𝑠
𝑃 (𝜖|∇𝜃|)𝜏𝜃 𝜑 = ∇· 𝜑 2 2
+ 𝜖 ∇𝜃
𝜕𝑡 |∇𝜃|
where
𝜇
𝑃 (𝑤) = 1 − exp (−𝛽𝑤) + exp (−𝛽𝑤)
𝜖
>>> tau_theta = 3e-3
>>> mu = 1e3
>>> gamma = 1e3
>>> beta_theta = 1e5
>>> expo = epsilon * beta_theta * theta.grad.mag
>>> expo = (expo < 100.) * (expo - 100.) + 100.
>>> Pfunc = 1. + numerix.exp(-expo) * (mu / epsilon - 1.)
>>> gradMagTheta = theta.faceGrad.mag

>>> eps = 1. / gamma / 10.
>>> gradMagTheta += (gradMagTheta < eps) * eps
>>> IGamma = (gradMagTheta > 1. / gamma) * (1 / gradMagTheta - gamma) + gamma
>>> v_theta = phase.arithmeticFaceValue * (s * IGamma + epsilon**2)
>>> D_theta = phase.arithmeticFaceValue**2 * (s * IGamma + epsilon**2)
thetagetFaceGradNoMod() evaluates the gradient without modular arithmetic.
>>> thetaEq = (TransientTerm(coeff=tau_theta * phaseMod**2 * Pfunc, var=theta)
... == DiffusionTerm(coeff=D_theta, var=theta)
... + PowerLawConvectionTerm(coeff=v_theta * (theta.faceGrad - theta.faceGradNoMod), var=p

>>> numSeeds = 10
>>> for Cx, Cy, orientation in numerix.random.random([numSeeds, 3]):
... radius = dx * 5.
... seed = ((x - Cx * nx * dx)**2 + (y - Cy * ny * dy)**2) < radius**2
... phase[seed] = 1.
... theta[seed] = numerix.pi * (2 * orientation - 1)

packages, so you the default color scheme of grain orientation won’t be very informative “out of the box”. Because
all of Python is accessible and FiPy is object oriented, it is not hard to adapt one of the existing viewers to create a
specialized display:
>>> if __name__ == "__main__":
... try:
... class OrientationViewer(Matplotlib2DGridViewer):
... def __init__(self, phase, orientation, title=None, limits={}, **kwlimits):
... Matplotlib2DGridViewer.__init__(self, vars=(orientation,), title=title,
... limits=limits, colorbar=None, **kwlimits)
...
... # make room for non-existent colorbar
... # stolen from matplotlib.colorbar.make_axes
... # https://github.com/matplotlib/matplotlib/blob
... # /ec1cd2567521c105a451ce15e06de10715f8b54d/lib
... # /matplotlib/colorbar.py#L838
... fraction = 0.15
... pb = self.axes.get_position(original=True).frozen()
... pad = 0.05
... x1 = 1.0-fraction
... pb1, pbx, pbcb = pb.splitx(x1-pad, x1)
... panchor = (1.0, 0.5)
... self.axes.set_position(pb1)
... self.axes.set_anchor(panchor)
...
... # make the gnomon
... fig = self.axes.get_figure()
... self.gnomon = fig.add_axes([0.85, 0.425, 0.15, 0.15], polar=True)
... self.gnomon.set_thetagrids([180, 270, 0, 90],
... [r"$\pm\pi$", r"$-\frac{\pi}{2}$", "$0$", r"$+\frac{\p
... frac=1.3)
... self.gnomon.set_theta_zero_location("N")
... self.gnomon.set_theta_direction(-1)
... self.gnomon.set_rgrids([1.], [""])
... N = 100
... theta = numerix.arange(-numerix.pi, numerix.pi, 2 * numerix.pi / N)
... radii = numerix.ones((N,))
... bars = self.gnomon.bar(theta, radii, width=2 * numerix.pi / N, bottom=0.0)
... colors = self._orientation_and_phase_to_rgb(orientation=numerix.array([theta]), p
... for c, t, bar in zip(colors[0], theta, bars):
... bar.set_facecolor(c)
... bar.set_edgecolor(c)
...
... def _reshape(self, var):
... '''return values of var in an 2D array'''
... return numerix.reshape(numerix.array(var),
... var.mesh.shape[::-1])[::-1]
...
... @staticmethod
... def _orientation_and_phase_to_rgb(orientation, phase):

... from matplotlib import colors

...
... hsv = numerix.empty(orientation.shape + (3,))
... hsv[..., 0] = (orientation / numerix.pi + 1) / 2.
... hsv[..., 1] = 1.
... hsv[..., 2] = phase
...
... return colors.hsv_to_rgb(hsv)
...
... @property
... def _data(self):
... '''convert phase and orientation to rgb image array
...
... orientation (-pi, pi) -> hue (0, 1)
... phase (0, 1) -> value (0, 1)
... '''
... orientation = self._reshape(self.vars[0])
... phase = self._reshape(self.phase)
...
... return self._orientation_and_phase_to_rgb(orientation, phase)
...
... self.image.set_data(self._data)
...
... from matplotlib import pyplot
... pyplot.ion()
... w, h = pyplot.figaspect(1.)
... fig = pyplot.figure(figsize=(2*w, h))
... timer = fig.text(0.1, 0.9, "t = %.3f" % 0, fontsize=18)
...
... viewer = MultiViewer(viewers=(MatplotlibViewer(vars=dT,
... cmap=pyplot.cm.hot,
... datamin=-0.5,
... datamax=0.5,
... axes=fig.add_subplot(121)),
... OrientationViewer(phase=phase,
... orientation=theta,
... title=theta.name,
... axes=fig.add_subplot(122))))
... viewer = MultiViewer(viewers=(Viewer(vars=dT,
... datamin=-0.5,
... datamax=0.5),
... Viewer(vars=phase,
... datamin=0.,
... datamax=1.),
... Viewer(vars=theta,
... datamin=-numerix.pi,
... datamax=numerix.pi)))
... viewer.plot()

>>> eq = thetaEq & phaseEq & heatEq
>>> if __name__ == "__main__":

... total_time = 2.
... else:

... total_time = dt * 10
>>> elapsed = 0.
>>> save_interval = 0.002
>>> save_at = save_interval
>>> while elapsed < total_time:

... if elapsed > 0.3:
... q.value = 100
... dT.updateOld()
... eq.solve(dt=dt)
... elapsed += dt
... if __name__ == "__main__" and elapsed >= save_at:
... timer.set_text("t = %.3f" % elapsed)
... viewer.plot()
... save_at += save_interval

Chapter 13
Level Set Examples
examples.levelSet.distanceFunction.mesh1D Create a level set variable in one dimension.

examples.levelSet.distanceFunction.circle Solve the level set equation in two dimensions for a circle.
examples.levelSet.advection.mesh1D Solve the distance function equation in one dimension and then advec
examples.levelSet.advection.circle Solve a circular distance function equation and then advect it.
13.1 examples.levelSet.distanceFunction.mesh1D
Create a level set variable in one dimension.

The level set variable calculates its value over the domain to be the distance from the zero level set. This can be
represented succinctly in the following equation with a boundary condition at the zero level set such that,
𝜕𝜑
=1
𝜕𝑥
with the boundary condition, 𝜑 = 0 at 𝑥 = 𝐿/2.
The solution to this problem will be demonstrated in the following script. Firstly, setup the parameters.
>>> dx = 0.5
>>> nx = 10
Construct the mesh.

>>> from fipy.tools import serialComm
>>> mesh = Grid1D(dx=dx, nx=nx, communicator=serialComm)
Construct a distanceVariable object.

>>> var = DistanceVariable(name='level set variable',
... mesh=mesh,
... value=-1.,
... hasOld=1)
>>> var.setValue(1, where=x > dx * nx / 2)
Once the initial positive and negative regions have been initialized the calcDistanceFunction() method can be used to
recalculate var as a distance function from the zero level set.
>>> var.calcDistanceFunction()
The problem can then be solved by executing the solve() method of the equation.
139
>>> if __name__ == '__main__':

... viewer = Viewer(vars=var, datamin=-5., datamax=5.)
... viewer.plot()
The result can be tested with the following commands.

>>> print numerix.allclose(var, x - dx * nx / 2)
1
13.2 examples.levelSet.distanceFunction.circle
Solve the level set equation in two dimensions for a circle.

The 2D level set equation can be written,
|∇𝜑| = 1
and the boundary condition for a circle is given by, 𝜑 = 0 at (𝑥 − 𝐿/2)2 + (𝑦 − 𝐿/2)2 = (𝐿/4)2 .
>>> dx = 1.
>>> dy = 1.
>>> nx = 11
>>> ny = 11
>>> Lx = nx * dx
>>> Ly = ny * dy
Construct the mesh.

>>> mesh = Grid2D(dx=dx, dy=dy, nx=nx, ny=ny, communicator=serialComm)

... mesh=mesh,
... value=-1.,
... hasOld=1)
>>> var.setValue(1, where=(x - Lx / 2.)**2 + (y - Ly / 2.)**2 < (Lx / 4.)**2)
>>> var.calcDistanceFunction(order=1)
>>> if __name__ == '__main__':

... viewer.plot()

>>> dY = dy / 2.
>>> dX = dx / 2.
>>> mm = min (dX, dY)
>>> m1 = dY * dX / numerix.sqrt(dY**2 + dX**2)
>>> def evalCell(phix, phiy, dx, dy):
140 Chapter 13. Level Set Examples

... aa = dy**2 + dx**2

... bb = -2 * ( phix * dy**2 + phiy * dx**2)
... cc = dy**2 * phix**2 + dx**2 * phiy**2 - dx**2 * dy**2
... sqr = numerix.sqrt(bb**2 - 4. * aa * cc)
... return ((-bb - sqr) / 2. / aa, (-bb + sqr) / 2. / aa)
>>> v1 = evalCell(-dY, -m1, dx, dy)[0]
>>> v2 = evalCell(-m1, -dX, dx, dy)[0]
>>> v3 = evalCell(m1, m1, dx, dy)[1]
>>> v4 = evalCell(v3, dY, dx, dy)[1]
>>> v5 = evalCell(dX, v3, dx, dy)[1]
>>> MASK = -1000.
>>> trialValues = CellVariable(mesh=mesh, value= \
... numerix.array((
... MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK,
... MASK, MASK, MASK, MASK,-3*dY,-3*dY,-3*dY, MASK, MASK, MASK, MASK,
... MASK, MASK, MASK, v1, -dY, -dY, -dY, v1, MASK, MASK, MASK,
... MASK, MASK, v2, -m1, m1, dY, m1, -m1, v2, MASK, MASK,
... MASK, -dX*3, -dX, m1, v3, v4, v3, m1, -dX,-dX*3, MASK,
... MASK, -dX*3, -dX, dX, v5, MASK, v5, dX, -dX,-dX*3, MASK,
... MASK, -dX*3, -dX, m1, v3, v4, v3, m1, -dX,-dX*3, MASK,
... MASK, MASK, v2, -m1, m1, dY, m1, -m1, v2, MASK, MASK,
... MASK, MASK, MASK, v1, -dY, -dY, -dY, v1, MASK, MASK, MASK,
... MASK, MASK, MASK, MASK,-3*dY,-3*dY,-3*dY, MASK, MASK, MASK, MASK,
... MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK, MASK), 'd'))
>>> var[numerix.array(trialValues == MASK)] = MASK

>>> print numerix.allclose(var, trialValues)
True
13.3 examples.levelSet.advection.mesh1D
Solve the distance function equation in one dimension and then advect it.
This example first solves the distance function equation in one dimension:
|∇𝜑| = 1
with 𝜑 = 0 at 𝑥 = 𝐿/5.
The variable is then advected with,
𝜕𝜑
+ ⃗𝑢 · ∇𝜑 = 0
𝜕𝑡
The scheme used in the FirstOrderAdvectionTerm preserves the var as a distance function.
>>> velocity = 1.
>>> dx = 1.
>>> nx = 10
>>> timeStepDuration = 1.
>>> steps = 2
>>> L = nx * dx
>>> interfacePosition = L / 5.
13.3. examples.levelSet.advection.mesh1D 141

Construct the mesh.

>>> mesh = Grid1D(dx=dx, nx=nx, communicator=serialComm)

... mesh=mesh,
... value=-1.,
... hasOld=1)
>>> var.setValue(1., where=mesh.cellCenters[0] > interfacePosition)
The advectionEquation is constructed.

>>> advEqn = TransientTerm() + FirstOrderAdvectionTerm(velocity)
The problem can then be solved by executing a serious of time steps.

>>> if __name__ == '__main__':
... viewer.plot()
... var.updateOld()
... advEqn.solve(var, dt=timeStepDuration)
... viewer.plot()
The result can be tested with the following code:

... var.updateOld()
>>> distanceTravelled = timeStepDuration * steps * velocity
>>> answer = x - interfacePosition - timeStepDuration * steps * velocity
>>> answer = numerix.where(x < distanceTravelled,
... x[0] - interfacePosition, answer)
>>> print var.allclose(answer)
1
13.4 examples.levelSet.advection.circle
Solve a circular distance function equation and then advect it.

This example first imposes a circular distance function:
[︃(︂ )︂2 (︂ )︂2 ]︃1/2
𝐿 𝐿 𝐿
𝜑 (𝑥, 𝑦) = 𝑥− + 𝑦− −
2 2 4
The variable is advected with,

𝜕𝜑
+ ⃗𝑢 · ∇𝜑 = 0
𝜕𝑡
The scheme used in the FirstOrderAdvectionTerm preserves the var as a distance function. The solution to
this problem will be demonstrated in the following script. Firstly, setup the parameters.

>>> L = 1.
>>> N = 25
>>> velocity = 1.
>>> cfl = 0.1
>>> velocity = 1.
>>> distanceToTravel = L / 10.
>>> radius = L / 4.
>>> dL = L / N
>>> timeStepDuration = cfl * dL / velocity
>>> steps = int(distanceToTravel / dL / cfl)
Construct the mesh.

>>> mesh = Grid2D(dx=dL, dy=dL, nx=N, ny=N)

>>> var = DistanceVariable(
... name = 'level set variable',
... mesh = mesh,
... value = 1.,
... hasOld = 1)
Initialise the distanceVariable to be a circular distance function.

>>> initialArray = numerix.sqrt((x - L / 2.)**2 + (y - L / 2.)**2) - radius
>>> var.setValue(initialArray)
The advection equation is constructed.

>>> advEqn = TransientTerm() + FirstOrderAdvectionTerm(velocity)
The problem can then be solved by executing a serious of time steps.

>>> if __name__ == '__main__':
... viewer = Viewer(vars=var, datamin=-radius, datamax=radius)
... viewer.plot()
... var.updateOld()
... viewer.plot()

... var.updateOld()
>>> x = numerix.array(mesh.cellCenters[0])
>>> distanceTravelled = timeStepDuration * steps * velocity
>>> answer = initialArray - distanceTravelled
>>> answer = numerix.where(answer < 0., -1001., answer)
>>> solution = numerix.where(answer < 0., -1001., numerix.array(var))
>>> numerix.allclose(answer, solution, atol=4.7e-3)
1
If the advection equation is built with the AdvectionTerm() the result is more accurate,
13.4. examples.levelSet.advection.circle 143

>>> var.setValue(initialArray)
>>> advEqn = TransientTerm() + AdvectionTerm(velocity)
... var.updateOld()
>>> solution = numerix.where(answer < 0., -1001., numerix.array(var))
>>> numerix.allclose(answer, solution, atol=1.02e-3)
1
13.5 Superconformal Electrodeposition Examples
13.5.1 The Damascene Process
State of the art manufacturing of semiconductor devices involves the electrodeposition of copper for on-chip wiring
of integrated circuits. In the Damascene process interconnects are fabricated by first patterning trenches in a dielectric
medium and then filling by metal electrodeposition over the entire wafer surface. This metalization process, pioneered
by IBM, depends on the use of electrolyte additives that effect the local metal deposition rate.
13.5.2 Superfill
The additives in the electrolyte affect the local deposition rate in such a way that bottom-up filling occurs in trenches
or vias. This process, known as superconformal electrodeposition or superfill, is demonstrated in the following figure.
The figure shows sequential images of bottom-up superfilling of submicrometer trenches by copper deposition from
an electrolyte containing PEG-SPS-Cl. Preferential metal deposition at the bottom of the trenches followed by bump
formation above the filled trenches is evident.
•As a function of TIME
•As a function of ASPECT RATIO

13.5.3 The CEAC Mechanism
This process has been demonstrated to depend critically on the inclusion of additives in the electrolyte. Recent publi-
cations propose Curvature Enhanced Accelerator Coverage (CEAC) as the mechanism behind the superfilling process
[8]. In this mechanism, molecules that accelerate local metal deposition displace molecules that inhibit local metal
deposition on the metal/electrolyte interface. For electrolytes that yield superconformal filling of fine features, this
buildup happens relatively slowly because the concentration of accelerator species is muchFig.more
1
dilute compared to
the inhibitor species in the electrolyte. The mechanism that leads to the increased rate of metal deposition along the
bottom of the filling trench is the concurrent local increase of the accelerator coverage due to decreasing local surface
area, which scales with the local curvature (hence the name of the mechanism). A good overview of this mechanism
can be found in [32].
13.5.4 Using FiPy to model Superfill
Example ?? provides a simple way to use FiPy to model the superfill process. The example includes a detailed
description of the governing equations and feature geometry. It requires the user to import and execute a function at
the python prompt. The model parameters can be passed as arguments to this function. In future all superfill examples

will be provided with this type of interface. Example ?? has the same functionality as ?? but demonstrates how to
write a new script in the case where the existing suite of scripts do not meet the required needs.
In general it is a good idea to obtain the Mayavi plotting package for which a specialized superfill viewer class has been
created, see Installation. The other standard viewers mentioned in Installation are still adequate although they do not
give such clear images that are tailored for the superfill problem. The images below demonstrate the Mayavi viewing
capability. Each contour represents sequential positions of the interface and the color represents the concentration of
accelerator as a surfactant. The areas of high surfactant concentration have an increased deposition rate.
examples.levelSet.electroChem.simpleTrenchSystem Model electrochemical superfill using the CEAC mechanism

examples.levelSet.electroChem.gold Model electrochemical superfill of gold using the CEAC m
examples.levelSet.electroChem.leveler Model electrochemical superfill of copper with leveler and
examples.levelSet.electroChem.howToWriteAScript Tutorial for writing an electrochemical superfill script.
13.6 examples.levelSet.electroChem.simpleTrenchSystem
Model electrochemical superfill using the CEAC mechanism.

This input file is a demonstration of the use of FiPy for modeling electrodeposition using the CEAC mechanism.
The material properties and experimental parameters used are roughly those that have been previously published
[NIST:damascene:2003]].
To run this example from the base fipy directory type:
$ python examples/levelSet/electroChem/simpleTrenchSystem.py
at the command line. The results of the simulation will be displayed and the word finished in the terminal at the end
of the simulation. To run with a different number of time steps change the numberOfSteps argument as follows,
>>> runSimpleTrenchSystem(numberOfSteps=2, displayViewers=False)
1
Change the displayViewers argument to True if you wish to see the results displayed on the screen. Example
examples.levelSet.electroChem.simpleTrenchSystem gives explanation for writing new scripts or
modifying existing scripts that are encapsulated by functions.
Any argument parameter can be changed. For example if the initial catalyst coverage is not 0, then it can be reset,
>>> runSimpleTrenchSystem(numberOfSteps=2, catalystCoverage=0.1, displayViewers=False)
0
The following image shows a schematic of a trench geometry along with the governing equations for mod-
eling electrodeposition with the CEAC mechanism. All of the given equations are implemented in the
examples.levelSet.electroChem.simpleTrenchSystem.runSimpleTrenchSystem() function.
As stated above, all the parameters in the equations can be changed with function arguments.
13.6. examples.levelSet.electroChem.simpleTrenchSystem 145

S
cm = c∞
m cθ = c∞
θ
metal bulk equation

∂cm
= Dm ∇2 cm
∂t catalyst bulk equation
∂cθ
= Dθ ∇2 cθ
∂t
δ
deposition rate
! "
cim −αF
iΩ i = i0 exp η
v= c∞
m RT catalyst boundary condition
nF
i0 = b 0 + b 1 θ
W
v
−Dm!n · ∇cm = −Dθ !n · ∇cθ = Γkcθ (1 − θa )
Ω
dθ
= κvθ + kcθ (1 − θ)
D dt
metal boundary condition k = k0 + k3 η 3
catalyst surfactant equation
The following table shows the symbols used in the governing equations and their corresponding arguments to the
runSimpleTrenchSystem() function. The boundary layer depth is intentionally small in this example in order
not to complicate the mesh. Further examples will simulate more realistic boundary layer depths but will also have
more complex meshes requiring the gmsh software.

Symbol Description Keyword Argument Value Unit

Deposition Rate Parameters
𝑣 deposition rate m s−1
𝑖 current density A m−2
Ω molar volume molarVolume 7.1×10−6 m3 mol−1
𝑛 ion charge charge 2
𝐹 Faraday’s constant faradaysConstant 9.6×10−4 C mol−1
𝑖0 exchange current density A m−2
𝛼 transfer coefficient transferCoefficient 0.5
𝜂 overpotential overpotential -0.3 V
𝑅 gas constant gasConstant 8.314 J K−1 mol−1
𝑇 temperature temperature 298.0 K
𝑏0 current density for 𝜃0 currentDensity0 0.26 A m−2
𝑏1 current density for 𝜃 currentDensity1 45.0 A m−2
Metal Ion Parameters
𝑐𝑚 metal ion concentration metalConcentration 250.0 mol m−3
𝑐∞
𝑚 far field metal ion concentration metalConcentration 250.0 mol m−3
𝐷𝑚 metal ion diffusion coefficient metalDiffusion 5.6×10−10 m2 s−1
Catalyst Parameters
𝜃 catalyst surfactant concentration catalystCoverage 0.0
𝑐𝜃 bulk catalyst concentration catalystConcentration 5.0×10−3 mol m−3
𝑐∞
𝜃 far field catalyst concentration catalystConcentration 5.0×10−3 mol m−3
𝐷𝜃 catalyst diffusion coefficient catalystDiffusion 1.0×10−9 m2 s−1
Γ catalyst site density siteDensity 9.8×10−6 mol m−2
𝑘 rate constant m3 mol−1 s−1
𝑘0 rate constant for 𝜂 0 rateConstant0 1.76 m3 mol−1 s−1
𝑘3 rate constant for 𝜂 3 rateConstant3 -245.0×10−6 m3 mol−1 s−1 V−3
Geometry Parameters
𝐷 trench depth trenchDepth 0.5×10−6 m
𝐷/𝑊 trench aspect ratio aspectRatio 2.0
𝑆 trench spacing trenchSpacing 0.6×10−6 m
𝛿 boundary layer depth boundaryLayerDepth 0.3×10−6 m
Simulation Control Parameters
computational cell size cellSize 0.1×10−7 m
number of time steps numberOfSteps 5
whether to display the viewers displayViewers True
If the MayaVi plotting software is installed (see Installation) then a plot should appear that is updated every 20 time
steps and will eventually resemble the image below.
13.6. examples.levelSet.electroChem.simpleTrenchSystem 147

13.7 examples.levelSet.electroChem.gold
Model electrochemical superfill of gold using the CEAC mechanism.

This input file is a demonstration of the use of FiPy for modeling gold superfill. The material properties and experi-
mental parameters used are roughly those that have been previously published [24].
$ python examples/levelSet/electroChem/gold.py
at the command line. The results of the simulation will be displayed and the word finished in the terminal at the
end of the simulation. The simulation will only run for 10 time steps. To run with a different number of time steps
change the numberOfSteps argument as follows,
>>> runGold(numberOfSteps=10, displayViewers=False)
1
Change the displayViewers argument to True if you wish to see the results displayed on the screen. This
example has a more realistic default boundary layer depth and thus requires gmsh to construct a more complex mesh.
There are a few differences between the gold superfill model presented in this example and in
examples.levelSet.electroChem.simpleTrenchSystem. Most default values have changed to account

for a different metal ion (gold) and catalyst (lead). In this system the catalyst is not present in the electrolyte but in-
stead has a non-zero initial coverage. Thus quantities associated with bulk catalyst and catalyst accumulation are not
defined. The current density is given by,
𝑐𝑚
𝑖= (𝑏0 + 𝑏1 𝜃) .
𝑐∞
𝑚
The more common representation of the current density includes an exponential part. Here it is buried in 𝑏0 and 𝑏1 .
The governing equation for catalyst evolution includes a term for catalyst consumption on the interface and is given
by
𝜃˙ = 𝐽𝑣𝜃 − 𝑘𝑐 𝑣𝜃
where 𝑘𝑐 is the consumption coefficient (consumptionRateConstant). The trench geometry is also given a
slight taper, given by taperAngle.
If the MayaVi plotting software is installed (see Installation) then a plot should appear that is updated every 10 time
steps and will eventually resemble the image below.
13.8 examples.levelSet.electroChem.leveler
Model electrochemical superfill of copper with leveler and accelerator additives.

This input file is a demonstration of the use of FiPy for modeling copper superfill with leveler and accelerator additives.
The material properties and experimental parameters used are roughly those that have been previously published [25].
13.8. examples.levelSet.electroChem.leveler 149

$ python examples/levelSet/electroChem/leveler.py
end of the simulation. The simulation will only run for 200 time steps. To run with a different number of time steps
change the numberOfSteps argument as follows,
>>> runLeveler(numberOfSteps=10, displayViewers=False, cellSize=0.25e-7)
1
Change the displayViewers argument to True if you wish to see the results displayed on the screen. This
example requires gmsh to construct the mesh.
This example models the case when suppressor, accelerator and leveler additives are present in the electrolyte. The
suppressor is assumed to absorb quickly compared with the other additives. Any unoccupied surface sites are im-
mediately covered with suppressor. The accelerator additive has more surface affinity than suppressor and is thus
preferential adsorbed. The accelerator can also remove suppressor when the surface reaches full coverage. Similarly,
the leveler additive has more surface affinity than both the suppressor and accelerator. This forms a simple set of
assumptions for understanding the behavior of these additives.
The following is a complete description of the equations for the model described here. Any equations that have been
omitted are the same as those given in examples.levelSet.electroChem.simpleTrenchSystem. The
current density is governed by
[︂ (︂ )︂]︂
𝑐𝑚 ∑︁ −𝛼𝑗 𝐹 𝜂 (1 − 𝛼𝑗 ) 𝐹 𝜂
𝑖= ∞ 𝑖𝑗 𝜃𝑗 exp − exp
𝑐𝑚 𝑗 𝑅𝑇 𝑅𝑇
where 𝑗 represents 𝑆 for suppressor, 𝐴 for accelerator, 𝐿 for leveler and 𝑉 for vacant. This model assumes a linear
interpolation between the three cases of complete coverage for each additive or vacant substrate. The governing
equations for the surfactants are given by,
𝜃˙𝐿 = 𝜅𝑣𝜃𝐿 + 𝑘𝑙+ 𝑐𝐿 (1 − 𝜃𝐿 ) − 𝑘𝐿

−
𝑣𝜃𝐿 ,
𝜃˙𝑎 = 𝜅𝑣𝜃𝐴 + 𝑘𝐴+ − 𝑞−1
𝑐𝐴 (1 − 𝜃𝐴 − 𝜃𝐿 ) − 𝑘𝐿 𝑐𝐿 𝜃𝐴 − 𝑘𝐴 𝜃𝐴 ,
𝜃𝑆 = 1 − 𝜃𝐴 − 𝜃𝐿
𝜃𝑉 = 0.
It has been found experimentally that 𝑖𝐿 = 𝑖𝑆 .

If the surface reaches full coverage, the equations do not naturally prevent the coverage rising above full coverage due
to the curvature terms. Thus, when 𝜃𝐿 + 𝜃𝐴 = 1 then the equation for accelerator becomes 𝜃˙𝐴 = −𝜃˙𝐿 and when
𝜃𝐿 = 1, the equation for leveler becomes 𝜃˙𝐿 = −𝑘𝐿 −
𝑣𝜃𝐿 .
−
The parameters 𝑘𝐴
+
, 𝑘𝐴 and 𝑞 are both functions of 𝜂 given by,
+ + −𝛼𝑘 𝐹 𝜂
𝑘𝐴 = 𝑘𝐴0 exp ,
𝑅𝑇
− 𝐴
𝑘𝐴 = 𝐵𝑑 + + exp (𝐵𝑏 (𝜂 + 𝑉𝑑 ))
exp (𝐵𝑎 (𝜂 + 𝑉𝑑 ))
𝑞 = 𝑚𝜂 + 𝑏.
The following table shows the symbols used in the governing equations and their corresponding arguments for the

runLeveler() function.
Symbol Description Keyword Argument Value Unit

Deposition Rate Parameters
𝑣 deposition rate m s−1
𝑖𝐴 accelerator current density i0Accelerator A m−2
𝑖𝐿 leveler current density i0Leveler A m−2
Ω molar volume molarVolume 7.1×10−6 m3 mol−1
𝑛 ion charge charge 2
𝐹 Faraday’s constant faradaysConstant 9.6×10−4 C mol−1
𝑖0 exchange current density A m−2
𝛼𝐴 accelerator transfer coefficient alphaAccelerator 0.4
𝛼𝑆 leveler transfer coefficient alphaLeveler 0.5
𝜂 overpotential overpotential -0.3 V
𝑅 gas constant gasConstant 8.314 J K mol−1
𝑇 temperature temperature 298.0 K
Ion Parameters
𝑐𝐼 ion concentration ionConcentration 250.0 mol m−3
𝑐∞
𝐼 far field ion concentration ionConcentration 250.0 mol m−3
𝐷𝐼 ion diffusion coefficient ionDiffusion 5.6×10−10 m2 s−1
Accelerator Parameters
𝜃𝐴 accelerator coverage acceleratorCoverage 0.0
𝑐𝐴 accelerator concentartion acceleratorConcentration 5.0×10−3 mol m−3
𝑐∞
𝐴 far field accelerator concentration acceleratorConcentration 5.0×10−3 mol m−3
𝐷𝐴 catalyst diffusion coefficient catalystDiffusion 1.0×10−9 m2 s−1
Γ𝐴 accelerator site density siteDensity 9.8×10−6 mol m−2
𝑘𝐴+
accelerator adsorption m3 mol−1 s−1
+
𝑘𝐴0 accelerator adsorption coeff kAccelerator0 2.6×10−4 m3 mol−1 s−1
𝛼𝑘 accelerator adsorption coeff alphaAdsorption 0.62
−
𝑘𝐴 accelerator consumption coeff
𝐵𝑎 experimental parameter Bd -40.0
𝐵𝑏 experimental parameter Bd 60.0
𝑉𝑑 experimental parameter Bd 9.8×10−2
𝐵𝑑 experimental parameter Bd 8.0×10−4
Geometry Parameters
𝐷 trench depth trenchDepth 0.5×10−6 m
𝐷/𝑊 trench aspect ratio aspectRatio 2.0
𝑆 trench spacing trenchSpacing 0.6×10−6 m
𝛿 boundary layer depth boundaryLayerDepth 0.3×10−6 m
Simulation Control Parameters
computational cell size cellSize 0.1×10−7 m
number of time steps numberOfSteps 5
whether to display the viewers displayViewers True
The following images show accelerator and leveler contour plots that can be obtained by running this example.
13.8. examples.levelSet.electroChem.leveler 151


13.9 examples.levelSet.electroChem.howToWriteAScript
Tutorial for writing an electrochemical superfill script.

This input file demonstrates how to create a new superfill script if the existing suite of scripts do not meet the required
needs. It provides the functionality of examples.levelSet.electroChem.simpleTrenchSystem.
$ python examples/levelSet/electroChem/howToWriteAScript.py --numberOfElements=10000 --numberOfSteps=
end of the simulation. To obtain this example in a plain script file in order to edit and run type:
$ python setup.py copy_script --From examples/levelSet/electroChem/howToWriteAScript.py --To myScript
in the base FiPy directory. The file myScript.py will contain the script.
The following is an explicit explanation of the input commands required to set up and run the problem. At the top of
the file all the parameter values are set. Their use will be explained during the instantiation of various objects and are
the same as those explained in examples.levelSet.electroChem.simpleTrenchSystem.
The following parameters (all in S.I. units) represent,
• physical constants,
13.9. examples.levelSet.electroChem.howToWriteAScript 153

>>> faradaysConstant = 9.6e4

>>> gasConstant = 8.314
>>> transferCoefficient = 0.5
• properties associated with the catalyst species,

>>> rateConstant0 = 1.76
>>> rateConstant3 = -245e-6
>>> catalystDiffusion = 1e-9
>>> siteDensity = 9.8e-6
• properties of the cupric ions,

>>> molarVolume = 7.1e-6
>>> charge = 2
>>> metalDiffusionCoefficient = 5.6e-10
• parameters dependent on experimental constraints,

>>> overpotential = -0.3
>>> bulkMetalConcentration = 250.
>>> catalystConcentration = 5e-3
>>> catalystCoverage = 0.
• parameters obtained from experiments on flat copper electrodes,

>>> currentDensity0 = 0.26
>>> currentDensity1 = 45.
• general simulation control parameters,

>>> cflNumber = 0.2
>>> numberOfCellsInNarrowBand = 10
>>> cellsBelowTrench = 10
>>> cellSize = 0.1e-7
• parameters required for a trench geometry,

>>> trenchDepth = 0.5e-6
>>> aspectRatio = 2.
>>> trenchSpacing = 0.6e-6
>>> boundaryLayerDepth = 0.3e-6
The hydrodynamic boundary layer depth (boundaryLayerDepth) is intentionally small in this example to keep
the mesh at a reasonable size.
Build the mesh:
>>> from fipy.tools.parser import parse
>>> numberOfElements = parse('--numberOfElements', action='store',
... type='int', default=-1)
>>> numberOfSteps = parse('--numberOfSteps', action='store',
... type='int', default=2)
>>> if numberOfElements != -1:

... pos = trenchSpacing * cellsBelowTrench / 4 / numberOfElements
... sqr = trenchSpacing * (trenchDepth + boundaryLayerDepth) \
... / (2 * numberOfElements)

... cellSize = pos + numerix.sqrt(pos**2 + sqr)

... else:
... cellSize = 0.1e-7
>>> yCells = cellsBelowTrench \

... + int((trenchDepth + boundaryLayerDepth) / cellSize)
>>> xCells = int(trenchSpacing / 2 / cellSize)
>>> from metalIonDiffusionEquation import buildMetalIonDiffusionEquation

>>> from adsorbingSurfactantEquation import AdsorbingSurfactantEquation
>>> from fipy import serialComm

>>> mesh = Grid2D(dx=cellSize,
... dy=cellSize,
... nx=xCells,
... ny=yCells,
... communicator=serialComm)
A distanceVariable object, 𝜑, is required to store the position of the interface.

The distanceVariable calculates its value so that it is a distance function (i.e. holds the distance at any point in
the mesh from the electrolyte/metal interface at 𝜑 = 0) and |∇𝜑| = 1.
First, create the 𝜑 variable, which is initially set to -1 everywhere. Create an initial variable,
>>> narrowBandWidth = numberOfCellsInNarrowBand * cellSize
>>> distanceVar = DistanceVariable(
... name='distance variable',
... mesh= mesh,
... value=-1.,
... hasOld=1)
The electrolyte region will be the positive region of the domain while the metal region will be negative.
>>> bottomHeight = cellsBelowTrench * cellSize
>>> trenchHeight = bottomHeight + trenchDepth
>>> trenchWidth = trenchDepth / aspectRatio
>>> sideWidth = (trenchSpacing - trenchWidth) / 2
>>> distanceVar.setValue(1., where=(y > trenchHeight)
... | ((y > bottomHeight)
... & (x < xCells * cellSize - sideWidth)))
>>> distanceVar.calcDistanceFunction(order=2)
The distanceVariable has now been created to mark the interface. Some other variables need to be created that
govern the concentrations of various species.
Create the catalyst surfactant coverage, 𝜃, variable. This variable influences the deposition rate.
>>> catalystVar = SurfactantVariable(
... name="catalyst variable",
... value=catalystCoverage,
... distanceVar=distanceVar)
Create the bulk catalyst concentration, 𝑐𝜃 , in the electrolyte,

>>> bulkCatalystVar = CellVariable(
... name='bulk catalyst variable',

... mesh=mesh,
... value=catalystConcentration)
Create the bulk metal ion concentration, 𝑐𝑚 , in the electrolyte.

>>> metalVar = CellVariable(
... name='metal variable',
... mesh=mesh,
... value=bulkMetalConcentration)
The following commands build the depositionRateVariable, 𝑣. The depositionRateVariable is given

by the following equation.
𝑖Ω
𝑣=
𝑛𝐹
where Ω is the metal molar volume, 𝑛 is the metal ion charge and 𝐹 is Faraday’s constant. The current density is given
by
𝑐𝑖𝑚
(︂ )︂
−𝛼𝐹
𝑖 = 𝑖0 ∞ exp 𝜂
𝑐𝑚 𝑅𝑇
where 𝑐𝑖𝑚 is the metal ion concentration in the bulk at the interface, 𝑐∞ 𝑚 is the far-field bulk concentration of metal
ions, 𝛼 is the transfer coefficient, 𝑅 is the gas constant, 𝑇 is the temperature and 𝜂 is the overpotential. The exchange
current density is an empirical function of catalyst coverage,
𝑖0 (𝜃) = 𝑏0 + 𝑏1 𝜃
The commands needed to build this equation are,
>>> expoConstant = -transferCoefficient * faradaysConstant \
... / (gasConstant * temperature)
>>> tmp = currentDensity1 \
... * catalystVar.interfaceVar
>>> exchangeCurrentDensity = currentDensity0 + tmp
>>> expo = numerix.exp(expoConstant * overpotential)
>>> currentDensity = expo * exchangeCurrentDensity * metalVar \
... / bulkMetalConcentration
>>> depositionRateVariable = currentDensity * molarVolume \
... / (charge * faradaysConstant)
Build the extension velocity variable 𝑣ext . The extension velocity uses the extensionEquation to spread the
velocity at the interface to the rest of the domain.
>>> extensionVelocityVariable = CellVariable(
... name='extension velocity',
... mesh=mesh,
... value=depositionRateVariable)
Using the variables created above the governing equations will be built. The governing equation for surfactant conser-
vation is given by,
𝜃˙ = 𝐽𝑣𝜃 + 𝑘𝑐𝑖𝜃 (1 − 𝜃)
where 𝜃 is the coverage of catalyst at the interface, 𝐽 is the curvature of the interface, 𝑣 is the normal velocity of the
interface, 𝑐𝑖𝜃 is the concentration of catalyst in the bulk at the interface. The value 𝑘 is given by an empirical function
of overpotential,
𝑘 = 𝑘0 + 𝑘3 𝜂 3
The above equation is represented by the AdsorbingSurfactantEquation in FiPy:

>>> surfactantEquation = AdsorbingSurfactantEquation(

... surfactantVar=catalystVar,
... distanceVar=distanceVar,
... bulkVar=bulkCatalystVar,
... rateConstant=rateConstant0 \
... + rateConstant3 * overpotential**3)
The variable 𝜑 is advected by the advectionEquation given by,

𝜕𝜑
+ 𝑣ext |∇𝜑| = 0
𝜕𝑡
and is set up with the following commands:
>>> advectionEquation = TransientTerm() + AdvectionTerm(extensionVelocityVariable)
The diffusion of metal ions from the far field to the interface is governed by,
𝜕𝑐𝑚
= ∇ · 𝐷∇𝑐𝑚
𝜕𝑡
where,
{︃
𝐷𝑚 when 𝜑 > 0,
𝐷=
0 when 𝜑 ≤ 0
The following boundary condition applies at 𝜑 = 0,

𝑣
𝑛 · ∇𝑐 =
𝐷ˆ .
Ω
The metal ion diffusion equation is set up with the following commands.
>>> metalEquation = buildMetalIonDiffusionEquation(
... ionVar=metalVar,
... depositionRate=depositionRateVariable,
... diffusionCoeff=metalDiffusionCoefficient,
... metalIonMolarVolume=molarVolume,
... )
>>> metalVar.constrain(bulkMetalConcentration, mesh.facesTop)
The surfactant bulk diffusion equation solves the bulk diffusion of a species with a source term for the jump from the
bulk to an interface. The governing equation is given by,
𝜕𝑐
= ∇ · 𝐷∇𝑐
𝜕𝑡
where,
{︃
𝐷𝜃 when 𝜑 > 0
𝐷=
0 when 𝜑 ≤ 0
The jump condition at the interface is defined by Langmuir adsorption. Langmuir adsorption essentially states that the
ability for a species to jump from an electrolyte to an interface is proportional to the concentration in the electrolyte,
available site density and a jump coefficient. The boundary condition at 𝜑 = 0 is given by,
𝑛 · ∇𝑐 = −𝑘𝑐(1 − 𝜃).
𝐷ˆ
The surfactant bulk diffusion equation is set up with the following commands.

>>> from surfactantBulkDiffusionEquation import buildSurfactantBulkDiffusionEquation

>>> bulkCatalystEquation = buildSurfactantBulkDiffusionEquation(
... bulkVar=bulkCatalystVar,
... surfactantVar=catalystVar,
... diffusionCoeff=catalystDiffusion,
... rateConstant=rateConstant0 * siteDensity
... )
>>> bulkCatalystVar.constrain(catalystConcentration, mesh.facesTop)
If running interactively, create viewers.

>>> if __name__ == '__main__':
... try:
... from mayaviSurfactantViewer import MayaviSurfactantViewer
... viewer = MayaviSurfactantViewer(distanceVar,
... catalystVar.interfaceVar,
... zoomFactor=1e6,
... datamax=1.0,
... datamin=0.0,
... smooth=1)
... except:
... viewer = MultiViewer(viewers=(
... Viewer(distanceVar, datamin=-1e-9, datamax=1e-9),
... Viewer(catalystVar.interfaceVar)))
... from fipy.models.levelSet.surfactant.matplotlibSurfactantViewer import MatplotlibSurfacta
... viewer = MatplotlibSurfactantViewer(catalystVar.interfaceVar)
... else:
... viewer = None
The levelSetUpdateFrequency defines how often to call the distanceEquation to reinitialize the
distanceVariable to a distance function.
>>> levelSetUpdateFrequency = int(0.8 * narrowBandWidth \
... / (cellSize * cflNumber * 2))
The following loop runs for numberOfSteps time steps. The time step is calculated with the CFL number and the
maximum extension velocity. 𝑣 to 𝑣ext throughout the whole domain using ∇𝜑 · ∇𝑣ext = 0.
>>> for step in range(numberOfSteps):
...
... if viewer is not None:
... viewer.plot()
...
... if step % levelSetUpdateFrequency == 0:
... distanceVar.calcDistanceFunction(order=2)
...
... extensionVelocityVariable.setValue(depositionRateVariable())
...
... distanceVar.updateOld()
... distanceVar.extendVariable(extensionVelocityVariable, order=2)
... dt = cflNumber * cellSize / extensionVelocityVariable.max()
... advectionEquation.solve(distanceVar, dt=dt)
... surfactantEquation.solve(catalystVar, dt=dt)
... metalEquation.solve(var=metalVar, dt=dt)
... bulkCatalystEquation.solve(var=bulkCatalystVar, dt=dt, solver=GeneralSolver())
The following is a short test case. It uses saved data from a simulation with 5 time steps. It is not a test for accuracy

but a way to tell if something has changed or been broken.

>>> import os
>>> filepath = os.path.join(os.path.split(__file__)[0],

... "simpleTrenchSystem.gz")
>>> ##numerix.savetxt(filepath, numerix.array(catalystVar))
>>> print catalystVar.allclose(numerix.loadtxt(filepath), rtol=1e-4)
1
>>> if __name__ == '__main__':

... raw_input('finished')


Chapter 14
Cahn Hilliard Examples
examples.cahnHilliard.mesh2DCoupled Solve the Cahn-Hilliard problem in two dimensions.

examples.cahnHilliard.sphere Solves the Cahn-Hilliard problem on the surface of a sphere.
14.1 examples.cahnHilliard.mesh2DCoupled
Solve the Cahn-Hilliard problem in two dimensions.

The spinodal decomposition phenomenon is a spontaneous separation of an initially homogenous mixture into two
distinct regions of different properties (spin-up/spin-down, component A/component B). It is a “barrierless” phase
separation process, such that under the right thermodynamic conditions, any fluctuation, no matter how small, will
tend to grow. This is in contrast to nucleation, where a fluctuation must exceed some critical magnitude before it
will survive and grow. Spinodal decomposition can be described by the “Cahn-Hilliard” equation (also known as
“conserved Ginsberg-Landau” or “model B” of Hohenberg & Halperin)
(︂ )︂
𝜕𝜑 𝜕𝑓
= ∇ · 𝐷∇ 2 2
−𝜖 ∇ 𝜑 .
𝜕𝑡 𝜕𝜑
where 𝜑 is a conserved order parameter, possibly representing alloy composition or spin. The double-well free energy
function 𝑓 = (𝑎2 /2)𝜑2 (1 − 𝜑)2 penalizes states with intermediate values of 𝜑 between 0 and 1. The gradient energy
term 𝜖2 ∇2 𝜑, on the other hand, penalizes sharp changes of 𝜑. These two competing effects result in the segregation of
𝜑 into domains of 0 and 1, separated by abrupt, but smooth, transitions. The parameters 𝑎 and 𝜖 determine the relative
weighting of the two effects and 𝐷 is a rate constant.
We can simulate this process in FiPy with a simple script:
(Note that all of the functionality of NumPy is imported along with FiPy, although much is augmented for FiPy‘s
needs.)
>>> if __name__ == "__main__":
... nx = ny = 20
... else:
... nx = ny = 10
>>> mesh = Grid2D(nx=nx, ny=ny, dx=0.25, dy=0.25)
>>> phi = CellVariable(name=r"$\phi$", mesh=mesh)
>>> psi = CellVariable(name=r"$\psi$", mesh=mesh)
We start the problem with random fluctuations about 𝜑 = 1/2

>>> noise = GaussianNoiseVariable(mesh=mesh,
... mean=0.5,
... variance=0.01).value
161
>>> phi[:] = noise
FiPy doesn’t plot or output anything unless you tell it to:

>>> if __name__ == "__main__":
... viewer = Viewer(vars=(phi, psi)) # , datamin=0., datamax=1.)
We factor the Cahn-Hilliard equation into two 2nd-order PDEs and place them in canonical form for FiPy to solve
them as a coupled set of equations.
𝜕𝜑
= ∇ · 𝐷∇𝜓
𝜕𝑡
𝜕2𝑓 𝜕𝑓
𝜓= (𝜑 − 𝜑old ) + − 𝜖2 ∇2 𝜑
𝜕𝜑2 𝜕𝜑
We need to perform the partial derivatives
𝜕𝑓
= (𝑎2 /2)2𝜑(1 − 𝜑)(1 − 2𝜑)
𝜕𝜑
𝜕2𝑓
= (𝑎2 /2)2 [1 − 6𝜑(1 − 𝜑)]
𝜕𝜑2
manually.
>>> D = a = epsilon = 1.
>>> dfdphi = a**2 * 2 * phi * (1 - phi) * (1 - 2 * phi)
>>> dfdphi_ = a**2 * 2 * (1 - phi) * (1 - 2 * phi)
>>> d2fdphi2 = a**2 * 2 * (1 - 6 * phi * (1 - phi))
>>> eq1 = (TransientTerm(var=phi) == DiffusionTerm(coeff=D, var=psi))
>>> eq2 = (ImplicitSourceTerm(coeff=1., var=psi)
... == ImplicitSourceTerm(coeff=-d2fdphi2, var=phi) - d2fdphi2 * phi + dfdphi
... - DiffusionTerm(coeff=epsilon**2, var=phi))
>>> eq3 = (ImplicitSourceTerm(coeff=1., var=psi)
... == ImplicitSourceTerm(coeff=dfdphi_, var=phi)
... - DiffusionTerm(coeff=epsilon**2, var=phi))
>>> eq = eq1 & eq2
Because the evolution of a spinodal microstructure slows with time, we use exponentially increasing time steps to keep
the simulation “interesting”. The FiPy user always has direct control over the evolution of their problem.
>>> dexp = -5
>>> elapsed = 0.
>>> if __name__ == "__main__":
... duration = .5e-1
... else:
... duration = .5e-1
>>> while elapsed < duration:

... dt = min(100, numerix.exp(dexp))
... elapsed += dt
... dexp += 0.01
... eq.solve(dt=dt)
... if __name__ == "__main__":
... viewer.plot()
>>> if __name__ == '__main__':

... raw_input("Coupled equations. Press <return> to proceed...")
162 Chapter 14. Cahn Hilliard Examples

(a) t = 30 (b) t = 100 (c) t = 1000
These equations can also be solved in FiPy using a vector equation. The variables 𝜑 and 𝜓 are now stored in a single
variable
>>> var = CellVariable(mesh=mesh, elementshape=(2,))
>>> var[0] = noise
>>> if __name__ == "__main__":

... viewer = Viewer(vars=(var[0], var[1]))
>>> D = a = epsilon = 1.
>>> v0 = var[0]
>>> dfdphi = a**2 * 2 * v0 * (1 - v0) * (1 - 2 * v0)
>>> dfdphi_ = a**2 * 2 * (1 - v0) * (1 - 2 * v0)
>>> d2fdphi2 = a**2 * 2 * (1 - 6 * v0 * (1 - v0))
The source terms have to be shaped correctly for a vector. The implicit source coefficient has to have a shape of (2, 2)
while the explicit source has a shape (2,)
>>> source = (- d2fdphi2 * v0 + dfdphi) * (0, 1)
>>> impCoeff = -d2fdphi2 * ((0, 0),
... (1., 0)) + ((0, 0),
... (0, -1.))
This is the same equation as the previous definition of eq, but now in a vector format.
>>> eq = TransientTerm(((1., 0.), 1
... (0., 0.))) == DiffusionTerm([((0., D),
... (-epsilon**2, 0.))]) + ImplicitSourceTerm(impCo
>>> dexp = -5
>>> elapsed = 0.

... elapsed += dt
... dexp += 0.01
... eq.solve(var=var, dt=dt)
... if __name__ == "__main__":
... viewer.plot()
>>> print numerix.allclose(var, (phi, psi))

True
14.1. examples.cahnHilliard.mesh2DCoupled 163

14.2 examples.cahnHilliard.sphere
Solves the Cahn-Hilliard problem on the surface of a sphere.

This phenomenon canoccur on vesicles (http://www.youtube.com/watch?v=kDsFP67_ZSE).
The only difference from examples.cahnHilliard.mesh2D is the declaration of mesh.

>>> mesh = Gmsh2DIn3DSpace('''
... radius = 5.0;
... cellSize = 0.3;
...
... // create inner 1/8 shell
... Point(1) = {0, 0, 0, cellSize};
... Point(4) = {0, 0, radius, cellSize};
... Circle(1) = {2, 1, 3};
... Circle(2) = {4, 1, 2};
... Circle(3) = {4, 1, 3};
... Line Loop(1) = {1, -3, 2} ;
... Ruled Surface(1) = {1};
...
... // create remaining 7/8 inner shells
... t1[] = Rotate {{0,0,1},{0,0,0},Pi/2} {Duplicata{Surface{1};}};
... t2[] = Rotate {{0,0,1},{0,0,0},Pi} {Duplicata{Surface{1};}};
... t3[] = Rotate {{0,0,1},{0,0,0},Pi*3/2} {Duplicata{Surface{1};}};
... t4[] = Rotate {{0,1,0},{0,0,0},-Pi/2} {Duplicata{Surface{1};}};
... t5[] = Rotate {{0,0,1},{0,0,0},Pi/2} {Duplicata{Surface{t4[0]};}};
... t6[] = Rotate {{0,0,1},{0,0,0},Pi} {Duplicata{Surface{t4[0]};}};
... t7[] = Rotate {{0,0,1},{0,0,0},Pi*3/2} {Duplicata{Surface{t4[0]};}};
...
... // create entire inner and outer shell
... Surface Loop(100)={1,t1[0],t2[0],t3[0],t7[0],t4[0],t5[0],t6[0]};
... ''', order=2).extrude(extrudeFunc=lambda r: 1.1 * r)
>>> phi = CellVariable(name=r"$\phi$", mesh=mesh)
We start the problem with random fluctuations about 𝜑 = 1/2

>>> phi.setValue(GaussianNoiseVariable(mesh=mesh,
... mean=0.5,
... variance=0.01))
FiPy doesn’t plot or output anything unless you tell it to: If MayaviClient is available, we can customize the view
with a sublcass of MayaviDaemon.
>>> if __name__ == "__main__":
... try:
... viewer = MayaviClient(vars=phi,
... datamin=0., datamax=1.,
... daemon_file="examples/cahnHilliard/sphereDaemon.py")
... except:
... viewer = Viewer(vars=phi,
... datamin=0., datamax=1.,
... xmin=-2.5, zmax=2.5)
For FiPy, we need to perform the partial derivative 𝜕𝑓 /𝜕𝜑 manually and then put the equation in the canonical form

by decomposing the spatial derivatives so that each Term is of a single, even order:
𝜕𝜑
= ∇ · 𝐷𝑎2 [1 − 6𝜑 (1 − 𝜑)] ∇𝜑 − ∇ · 𝐷∇𝜖2 ∇2 𝜑.
𝜕𝑡
FiPy would automatically interpolate D * a**2 * (1 - 6 * phi * (1 - phi)) onto the faces, where the
diffusive flux is calculated, but we obtain somewhat more accurate results by performing a linear interpolation from
phi at cell centers to PHI at face centers. Some problems benefit from non-linear interpolations, such as harmonic or
geometric means, and FiPy makes it easy to obtain these, too.
>>> PHI = phi.arithmeticFaceValue
>>> D = a = epsilon = 1.
>>> eq = (TransientTerm()
... == DiffusionTerm(coeff=D * a**2 * (1 - 6 * PHI * (1 - PHI)))
... - DiffusionTerm(coeff=(D, epsilon**2)))
Because the evolution of a spinodal microstructure slows with time, we use exponentially increasing time steps to keep
the simulation “interesting”. The FiPy user always has direct control over the evolution of their problem.
>>> dexp = -5
>>> elapsed = 0.
>>> if __name__ == "__main__":
... duration = 1000.
... else:
... duration = 1e-2
... elapsed += dt
... dexp += 0.01
... eq.solve(phi, dt=dt, solver=DefaultSolver(precon=None))
... if __name__ == "__main__":
... viewer.plot()
14.2. examples.cahnHilliard.sphere 165


Chapter 15
Fluid Flow Examples
examples.flow.stokesCavity Solve the Navier-Stokes equation in the viscous limit.
15.1 examples.flow.stokesCavity
Solve the Navier-Stokes equation in the viscous limit.

Many thanks to Benny Malengier <[email protected]> for reworking this example and actually making it work
correctly...see changeset:3799
This example is an implementation of a rudimentary Stokes solver on a collocated grid. It solves the Navier-Stokes
equation in the viscous limit,
∇𝜇 · ∇⃗𝑢 = ∇𝑝
and the continuity equation,
∇ · ⃗𝑢 = 0
where ⃗𝑢 is the fluid velocity, 𝑝 is the pressure and 𝜇 is the viscosity. The domain in this example is a square cavity
of unit dimensions with a moving lid of unit speed. This example uses the SIMPLE algorithm with Rhie-Chow
interpolation for collocated grids to solve the pressure-momentum coupling. Some of the details of the algorithm will
be highlighted below but a good reference for this material is Ferziger and Peric [33] and Rossow [rossow:2003]. The
solution has a high degree of error close to the corners of the domain for the pressure but does a reasonable job of
predicting the velocities away from the boundaries. A number of aspects of FiPy need to be improved to have a first
class flow solver. These include, higher order spatial diffusion terms, proper wall boundary conditions, improved mass
flux evaluation and extrapolation of cell values to the boundaries using gradients.
In the table below a comparison is made with the Dolfyn open source code on a 100 by 100 grid. The table shows the
frequency of values that fall within the given error confidence bands. Dolfyn has the added features described above.
When these features are switched off the results of Dolfyn and FiPy are identical.
% frequency of cells x-velocity error (%) y-velocity error (%) pressure error (%)
90 < 0.1 < 0.1 <5
5 0.1 to 0.6 0.1 to 0.3 5 to 11
4 0.6 to 7 0.3 to 4 11 to 35
1 7 to 96 4 to 80 35 to 179
0 > 96 > 80 > 179
To start, some parameters are declared.
>>> #from fipy.meshes.grid2D import Grid2D
167
>>> L = 1.0
>>> N = 50
>>> dL = L / N
>>> viscosity = 1
>>> U = 1.
>>> #0.8 for pressure and 0.5 for velocity are typical relaxation values for SIMPLE
>>> pressureRelaxation = 0.8
>>> velocityRelaxation = 0.5
>>> if __name__ == '__main__':
... sweeps = 300
... else:
... sweeps = 5
Build the mesh.

>>> mesh = Grid2D(nx=N, ny=N, dx=dL, dy=dL)
Declare the variables.

>>> pressure = CellVariable(mesh=mesh, name='pressure')
>>> pressureCorrection = CellVariable(mesh=mesh)
>>> xVelocity = CellVariable(mesh=mesh, name='X velocity')
>>> yVelocity = CellVariable(mesh=mesh, name='Y velocity')
The velocity is required as a rank-1 FaceVariable for calculating the mass flux. This is required by the Rhie-Chow
correction to avoid pressure/velocity decoupling.
>>> velocity = FaceVariable(mesh=mesh, rank=1)
Build the Stokes equations in the cell centers.

>>> xVelocityEq = DiffusionTerm(coeff=viscosity) - pressure.grad.dot([1.,0.])
>>> yVelocityEq = DiffusionTerm(coeff=viscosity) - pressure.grad.dot([0.,1.])
In this example the SIMPLE algorithm is used to couple the pressure and momentum equations. Let us assume we
have solved the discretized momentum equations using a guessed pressure field 𝑝* to obtain a velocity field ⃗𝑢* . That
is ⃗𝑢* is found from
∑︁
𝑎𝑃 ⃗𝑢*𝑃 = 𝑎𝐴 ⃗𝑢*𝐴 − 𝑉𝑃 (∇𝑝* )𝑃
𝑓
We would like to somehow correct these initial fields to satisfy both the discretized momentum and continuity equa-
tions. We now try to correct these initial fields with a correction such that ⃗𝑢 = ⃗𝑢* + ⃗𝑢′ and 𝑝 = 𝑝* + 𝑝′ , where ⃗𝑢 and
𝑝 now satisfy the momentum and continuity equations. Substituting the exact solution into the equations we obtain,
∇𝜇 · ∇⃗𝑢′ = 𝑝⃗′
and
∇ · ⃗𝑢* + ∇ · ⃗𝑢′ = 0
We now use the discretized form of the equations to write the velocity correction in terms of the pressure correction.
The discretized form of the above equation results in an equation for 𝑝 = 𝑝′ ,
∑︁
𝑎𝑃 ⃗𝑢′𝑃 = 𝑎𝐴 ⃗𝑢′𝐴 − 𝑉𝑃 (∇𝑝′ )𝑃
𝑓
where notation from Linear Equations is used. The SIMPLE algorithm drops the second term in the above equation
to leave,
𝑉𝑃 (∇𝑝′ )𝑃
⃗𝑢′𝑃 = −
𝑎𝑃
168 Chapter 15. Fluid Flow Examples

By substituting the above expression into the continuity equations we obtain the pressure correction equation,
𝑉𝑃
∇ · ∇𝑝′ = ∇ · ⃗𝑢*
𝑎𝑃
In the discretized version of the above equation 𝑉𝑃 /𝑎𝑃 is approximated at the face by 𝐴𝑓 𝑑𝐴𝑃 /(𝑎𝑃 )𝑓 . In FiPy the
pressure correction equation can be written as,
>>> ap = CellVariable(mesh=mesh, value=1.)
>>> coeff = 1./ ap.arithmeticFaceValue*mesh._faceAreas * mesh._cellDistances
>>> pressureCorrectionEq = DiffusionTerm(coeff=coeff) - velocity.divergence
Above would work good on a staggered grid, however, on a colocated grid as FiPy uses, the term
velocity.divergence will cause oscillations in the pressure solution as velocity is a face variable. We can
apply the Rhie-Chow correction terms for this. In this an intermediate velocity term 𝑢◇ is considered which does not
contain the pressure corrections:
𝑉𝑃 ∑︁ 𝑎𝐴
⃗𝑢◇𝑃 = ⃗𝑢*𝑃 + (∇𝑝* )𝑃 = ⃗𝑢*
𝑎𝑃 𝑎𝑃 𝐴
𝑓
This velocity is interpolated at the edges, after which the pressure correction term is added again, but now considered
at the edge:
(︂ )︂
1 𝑉
⃗𝑢𝑓 = (⃗𝑢◇𝐿 + ⃗𝑢◇𝑅 )) − (∇𝑝*𝑓 )
2 𝑎𝑃 avg L,R
(︁ )︁
where 𝑎𝑉𝑃 is assumed a good approximation at the edge. Here L and R denote the two cells adjacent to the
avg L,R
face. Expanding the not calculated terms we arrive at
(︂ )︂ (︂ )︂
1 * * 1 𝑉 * * 𝑉
⃗𝑢𝑓 = (⃗𝑢𝐿 + ⃗𝑢𝑅 )) + (∇𝑝𝐿 + ∇𝑝𝑅 ) − (∇𝑝*𝑓 )
2 2 𝑎𝑃 avg L,R 𝑎𝑃 avg L,R
where we have replaced the coefficients of the cell pressure gradients by an averaged value over the edge. This formula
has the consequence that the velocity on a face depends not only on the pressure of the adjacent cells, but also on the
cells further away, which removes the unphysical pressure oscillations. We start by introducing needed terms
>>> from fipy.variables.faceGradVariable import _FaceGradVariable
>>> volume = CellVariable(mesh=mesh, value=mesh.cellVolumes, name='Volume')
>>> contrvolume=volume.arithmeticFaceValue
And set up the velocity with this formula in the SIMPLE loop. Now, set up the no-slip boundary conditions
>>> xVelocity.constrain(0., mesh.facesRight | mesh.facesLeft | mesh.facesBottom)
>>> xVelocity.constrain(U, mesh.facesTop)
>>> yVelocity.constrain(0., mesh.exteriorFaces)
>>> pressureCorrection.constrain(0., mesh.facesLeft & (Y < dL))
Set up the viewers,

>>> if __name__ == '__main__':
... viewer = Viewer(vars=(pressure, xVelocity, yVelocity, velocity),
... xmin=0., xmax=1., ymin=0., ymax=1., colorbar=True)
Below, we iterate for a set number of sweeps. We use the sweep() method instead of solve() because we
require the residual for output. We also use the cacheMatrix(), getMatrix(), cacheRHSvector() and
getRHSvector() because both the matrix and RHS vector are required by the SIMPLE algorithm. Additionally,
the sweep() method is passed an underRelaxation factor to relax the solution. This argument cannot be passed
to solve().
15.1. examples.flow.stokesCavity 169

>>> for sweep in range(sweeps):

...
... ## solve the Stokes equations to get starred values
... xVelocityEq.cacheMatrix()
... xres = xVelocityEq.sweep(var=xVelocity,
... underRelaxation=velocityRelaxation)
... xmat = xVelocityEq.matrix
...
... yres = yVelocityEq.sweep(var=yVelocity,
... underRelaxation=velocityRelaxation)
...
... ## update the ap coefficient from the matrix diagonal
... ap[:] = -xmat.takeDiagonal()
...
... ## update the face velocities based on starred values with the
... ## Rhie-Chow correction.
... ## cell pressure gradient
... presgrad = pressure.grad
... ## face pressure gradient
... facepresgrad = _FaceGradVariable(pressure)
...
... velocity[0] = xVelocity.arithmeticFaceValue \
... + contrvolume / ap.arithmeticFaceValue * \
... (presgrad[0].arithmeticFaceValue-facepresgrad[0])
... velocity[1] = yVelocity.arithmeticFaceValue \
... + contrvolume / ap.arithmeticFaceValue * \
... (presgrad[1].arithmeticFaceValue-facepresgrad[1])
... velocity[..., mesh.exteriorFaces.value] = 0.
... velocity[0, mesh.facesTop.value] = U
...
... ## solve the pressure correction equation
... pressureCorrectionEq.cacheRHSvector()
... ## left bottom point must remain at pressure 0, so no correction
... pres = pressureCorrectionEq.sweep(var=pressureCorrection)
... rhs = pressureCorrectionEq.RHSvector
...
... ## update the pressure using the corrected value
... pressure.setValue(pressure + pressureRelaxation * pressureCorrection )
... ## update the velocity using the corrected pressure
... xVelocity.setValue(xVelocity - pressureCorrection.grad[0] / \
... ap * mesh.cellVolumes)
... yVelocity.setValue(yVelocity - pressureCorrection.grad[1] / \
... ap * mesh.cellVolumes)
...
... if __name__ == '__main__':
... if sweep%10 == 0:
... print 'sweep:',sweep,', x residual:',xres, \
... ', y residual',yres, \
... ', p residual:',pres, \
... ', continuity:',max(abs(rhs))
...
... viewer.plot()

Test values in the last cell.

>>> print numerix.allclose(pressure.globalValue[...,-1], 162.790867927)
1
>>> print numerix.allclose(xVelocity.globalValue[...,-1], 0.265072740929)
1
>>> print numerix.allclose(yVelocity.globalValue[...,-1], -0.150290488304)
1
15.1. examples.flow.stokesCavity 171


Chapter 16
Reactive Wetting Examples
examples.reactiveWetting.liquidVapor1D Solve a single-component, liquid-vapor, van der Waals system.
16.1 examples.reactiveWetting.liquidVapor1D
Solve a single-component, liquid-vapor, van der Waals system.

This example solves a single-component, liquid-vapor, van der Waals system as described by Wheeler et al. [6]. The
free energy for this system takes the form,
𝑒𝜌2
(︂ )︂
𝑅𝑇 𝜌
𝑓 =− 2 + ln (16.1)
𝑚 𝑚 𝑚 − 𝑣¯𝜌
where 𝜌 is the density. This free energy supports a two phase equilibrium with densities given by 𝜌𝑙 and 𝜌𝑣 in the
liquid and vapor phases, respectively. The densities are determined by solving the following system of equations,
𝑃 𝜌𝑙 = 𝑃 (𝜌𝑣 ) (16.2)
(︀ )︀
and
𝜇 𝜌𝑙 = 𝜇 (𝜌𝑣 ) (16.3)
(︀ )︀
where 𝜇 is the chemical potential,

𝜕𝑓
𝜇= (16.4)
𝜕𝜌
and 𝑃 is the pressure,
𝑃 = 𝜌𝜇 − 𝑓 (16.5)
One choice of thermodynamic parameters that yields a relatively physical two phase system is
>>> molarWeight = 0.118
>>> ee = -0.455971
>>> gasConstant = 8.314
>>> vbar = 1.3e-05
with equilibrium density values of

>>> liquidDensity = 7354.3402662299995
>>> vaporDensity = 82.855803327810008
The equilibrium densities are verified by substitution into Eqs. (16.2) and (16.3). Firstly, Eqs. (16.1), (16.4) and (16.5)
are defined as python functions,
173
>>> def f(rho):

... return ee * rho**2 / molarWeight**2 + gasConstant * temperature * rho / molarWeight * \
... numerix.log(rho / (molarWeight - vbar * rho))
>>> def mu(rho):

... return 2 * ee * rho / molarWeight**2 + gasConstant * temperature / molarWeight * \
... (numerix.log(rho / (molarWeight - vbar * rho)) + molarWeight / (molarWeight - vbar * r
>>> def P(rho):

... return rho * mu(rho) - f(rho)
The equilibrium densities values are verified with

>>> print numerix.allclose(mu(liquidDensity), mu(vaporDensity))
True
and
>>> print numerix.allclose(P(liquidDensity), P(vaporDensity))
True
In order to derive governing equations, the free energy functional is defined.

∫︁ [︂ ]︂
𝜖𝑇 2
𝐹 = 𝑓+ (𝜕𝑗 𝜌) 𝑑𝑉
2
Using standard dissipation laws, we write the governing equations for mass and momentum conservation,
𝜕𝜌
+ 𝜕𝑗 (𝜌𝑢𝑗 ) = 0 (16.6)
𝜕𝑡
and
𝜕 (𝜌𝑢𝑖 )
+ 𝜕𝑗 (𝜌𝑢𝑖 𝑢𝑗 ) = 𝜕𝑗 (𝜈 [𝜕𝑗 𝑢𝑖 + 𝜕𝑖 𝑢𝑗 ]) − 𝜌𝜕𝑖 𝜇𝑁 𝐶 (16.7)
𝜕𝑡
where the non-classical potential, 𝜇𝑁 𝐶 , is given by,
𝛿𝐹
𝜇𝑁 𝐶 = = 𝜇 − 𝜖𝑇 𝜕𝑗2 𝜌 (16.8)
𝛿𝜌
As usual, to proceed, we define a mesh
>>> Lx = 1e-6
>>> nx = 100
>>> dx = Lx / nx
and the independent variables.

>>> density = CellVariable(mesh=mesh, hasOld=True, name=r'$\rho$')
>>> velocity = CellVariable(mesh=mesh, hasOld=True, name=r'$u$')
>>> densityPrevious = density.copy()
>>> velocityPrevious = velocity.copy()
The system of equations is solved in a fully coupled manner using a block matrix. Defining 𝜇𝑁 𝐶 as an independent
variable makes it easier to script the equations without using higher order terms.
174 Chapter 16. Reactive Wetting Examples

>>> potentialNC = CellVariable(mesh=mesh, name=r'$\mu^{NC}$')
>>> epsilon = 1e-16

>>> freeEnergy = (f(density) + epsilon * temperature / 2 * density.grad.mag**2).cellVolumeAverage
In order to solve the equations numerically, an interpolation method is used to prevent the velocity and density fields
decoupling. The following velocity correction equation (expressed in discretized form) prevents decoupling from
occuring,
𝐴𝑓 𝑑𝑓 (︁ )︁
𝑢𝑐𝑖,𝑓 = 𝜌𝜕𝑖 𝜇𝑁 𝐶 𝑓 − 𝜌𝑓 𝜕𝑖,𝑓 𝜇𝑁 𝐶 (16.9)
𝑑𝑓
where 𝐴𝑓 is the face area, 𝑑𝑓 is the distance between the adjacent cell centers and 𝑎𝑓 is the momentum conservation
equation’s matrix diagonal. The overbar refers to an averaged value between the two adjacent cells to the face. The
notation 𝜕𝑖,𝑓 refers to a derivative evaluated directly at the face (not averaged). The variable 𝑢𝑐𝑖 is used to modify the
velocity used in Eq. (16.6) such that,
𝜕𝜌
+ 𝜕𝑗 (𝜌 [𝑢𝑗 + 𝑢𝑐𝑖 ]) = 0 (16.10)
𝜕𝑡
Equation (16.10) becomes
>>> matrixDiagonal = CellVariable(mesh=mesh, name=r'$a_f$', value=1e+20, hasOld=True)
>>> correctionCoeff = mesh._faceAreas * mesh._cellDistances / matrixDiagonal.faceValue
>>> massEqn = TransientTerm(var=density) \
... + VanLeerConvectionTerm(coeff=velocity.faceValue + correctionCoeff \
... * (density * potentialNC.grad).faceValue, \
... var=density) \
... - DiffusionTerm(coeff=correctionCoeff * density.faceValue**2, var=potentialNC)
where the first term on the LHS of Eq. (16.9) is calculated in an explicit manner in the VanLeerConvectionTerm
and the second term is calculated implicitly as a DiffusionTerm with 𝜇𝑁 𝐶 as the independent variable.
In order to write Eq. (16.7) as a FiPy expression, the last term is rewritten such that,
𝜌𝜕𝑖 𝜇𝑁 𝐶 = 𝜕𝑖 𝜌𝜇𝑁 𝐶 − 𝜇𝑁 𝐶 𝜕𝑖 𝜌
(︀ )︀
which results in
>>> viscosity = 1e-3
>>> ConvectionTerm = CentralDifferenceConvectionTerm
>>> momentumEqn = TransientTerm(coeff=density, var=velocity) \
... + ConvectionTerm(coeff=[[1]] * density.faceValue * velocity.faceValue, var=velocity
... == DiffusionTerm(coeff=2 * viscosity, var=velocity) \
... - ConvectionTerm(coeff=density.faceValue * [[1]], var=potentialNC) \
... + ImplicitSourceTerm(coeff=density.grad[0], var=potentialNC)
The only required boundary condition eliminates flow in or out of the domain.
>>> velocity.constrain(0, mesh.exteriorFaces)
As previously stated, the 𝜇𝑁 𝐶 variable will be solved implicitly. To do this the Eq. (16.8) is linearized in 𝜌 such that
(︂ )︂*
𝜕𝜇
𝜇 𝑁𝐶
=𝜇 +*
(𝜌 − 𝜌* ) − 𝜖𝑇 𝜕𝑗2 𝜌 (16.11)
𝜕𝜌
𝜕𝜇
The * superscript denotes the current held value. In FiPy, 𝜕𝜌 is written as,
16.1. examples.reactiveWetting.liquidVapor1D 175

>>> potentialDerivative = 2 * ee / molarWeight**2 + gasConstant * temperature * molarWeight / density
and 𝜇* is simply,
>>> potential = mu(density)
Eq. (16.11) can be scripted as

>>> potentialNCEqn = ImplicitSourceTerm(coeff=1, var=potentialNC) \
... == potential \
... + ImplicitSourceTerm(coeff=potentialDerivative, var=density) \
... - potentialDerivative * density \
... - DiffusionTerm(coeff=epsilon * temperature, var=density)
Due to a quirk in FiPy, the gradient of 𝜇𝑁 𝐶 needs to be constrained on the boundary. This is because
ConvectionTerm‘s will automatically assume a zero flux, which is not what we need in this case.
>>> potentialNC.faceGrad.constrain(value=[0], where=mesh.exteriorFaces)
All three equations are defined and an are combined together with
>>> coupledEqn = massEqn & momentumEqn & potentialNCEqn
The system will be solved as a phase separation problem with an initial density close to the average density, but with
some small amplitude noise. Under these circumstances, the final condition should be two separate phases of roughly
equal volume. The initial condition for the density is defined by
>>> density[:] = (liquidDensity + vaporDensity) / 2 * \
... (1 + 0.01 * (2 * numerix.random.random(mesh.numberOfCells) - 1))
Viewers are also defined.

>>> if __name__ == '__main__':
... viewers = Viewer(density), Viewer(velocity), Viewer(potentialNC)
... for viewer in viewers:
... viewer.plot()
... raw_input('arrange viewers')
... viewer.plot()
The following section defines the required control parameters. The cfl parameter limits the size of the time step so
that dt = cfl * dx / max(velocity).
>>> cfl = 0.1
>>> tolerance = 1e-1
>>> dt = 1e-14
>>> timestep = 0
>>> relaxation = 0.5
>>> if __name__ == '__main__':
... totalSteps = 1e10
... else:
... totalSteps = 10
In the following time stepping scheme a time step is recalculated if the residual increases between sweeps or the
required tolerance is not attained within 20 sweeps. The major quirk in this scheme is the requirement of updat-
ing the matrixDiagonal using the entire coupled matrix. This could be achieved more elegantly by calling
cacheMatrix() only on the necessary part of the equation. This currently doesn’t work properly in FiPy.

>>> while timestep < totalSteps:

...
... sweep = 0
... dt *= 1.1
... residual = 1.
... initialResidual = None
...
... density.updateOld()
... velocity.updateOld()
... matrixDiagonal.updateOld()
...
... while residual > tolerance:
...
... densityPrevious[:] = density
... velocityPrevious[:] = velocity
... previousResidual = residual
...
... dt = min(dt, dx / max(abs(velocity)) * cfl)
...
... coupledEqn.cacheMatrix()
... residual = coupledEqn.sweep(dt=dt)
...
... if initialResidual is None:
... initialResidual = residual
...
... residual = residual / initialResidual
...
... if residual > previousResidual * 1.1 or sweep > 20:
... density[:] = density.old
... velocity[:] = velocity.old
... matrixDiagonal[:] = matrixDiagonal.old
... dt = dt / 10.
... if __name__ == '__main__':
... print 'Recalculate the time step'
... timestep -= 1
... break
... else:
... matrixDiagonal[:] = coupledEqn.matrix.takeDiagonal()[mesh.numberOfCells:2 * mesh.numb
... density[:] = relaxation * density + (1 - relaxation) * densityPrevious
... velocity[:] = relaxation * velocity + (1 - relaxation) * velocityPrevious
...
... sweep += 1
...
... if __name__ == '__main__' and timestep % 10 == 0:
... print 'timestep: %i, dt: %1.5e, free energy: %1.5e' % (timestep, dt, freeEnergy)
... viewer.plot()
...
... timestep += 1
>>> if __name__ == '__main__':

... raw_input('finished')
>>> print freeEnergy < 1.5e9

True
16.1. examples.reactiveWetting.liquidVapor1D 177


Chapter 17
Updating FiPy
examples.updating.update2_0to3_0 How to update scripts from version 2.0 to 3.0.

17.1 examples.updating.update2_0to3_0
How to update scripts from version 2.0 to 3.0.

FiPy 3.0 introduces several syntax changes from FiPy 2.0. We appreciate that this is very inconvenient for our users,
but we hope you’ll agree that the new syntax is easier to read and easier to use. We assure you that this is not something
we do casually; it has been over two and a half years since our last incompatible change (when FiPy 2.0 superceded
FiPy 1.0).
All examples included with version 3.0 have been updated to use the new syntax, but any scripts you have written for
FiPy 2.0 will need to be updated. A complete listing of the changes needed to take the FiPy examples scripts from
version 2.0 to version 3.0 can be found at
http://www.matforge.org/fipy/wiki/upgrade2_0examplesTo3_0
but we summarize the necessary changes here. If these tips are not sufficient to make your scripts compatible with
FiPy 3.0, please don’t hesitate to ask for help on the mailing list.
The following items must be changed in your scripts
• We have reconsidered the change in FiPy 2.0 that included all of the functions of the numerix module in the
fipy namespace. You now must be more explicit when referring to any of these functions:
>>> y = numerix.exp(x)
>>> from fipy.tools.numerix import exp

>>> y = exp(x)
We generally use the first, but you may see us import specific functions if we feel it improves readability. You
should feel free to use whichever form you find most comfortable.
Note: the old behavior can be obtained, at least for now, by setting the FIPY_INCLUDE_NUMERIX_ALL
environment variable.
• If your equation contains a TransientTerm, then you must specify the timestep by passing a dt= argument
when calling solve() or sweep().
The remaining changes are not required, but they make scripts easier to read and we recommend them. FiPy may issue
a DeprecationWarning for some cases, to indicate that we may not maintain the old syntax indefinitely.
179
• “getter” and “setter” methods have been replaced with properties, e.g., use
instead of
>>> x, y = mesh.getCellCenters()
• Boundary conditions are better applied with the constrain() method than with the old FixedValue and
FixedFlux classes. See Boundary Conditions.
• Individual Mesh classes should be imported directly from fipy.meshes and not
fipy.meshes.numMesh.
• The Gmsh meshes now have simplified names: Gmsh2D instead of GmshImporter2D, Gmsh3D instead of
GmshImporter3D, and Gmsh2DIn3DSpace instead of GmshImporter2DIn3DSpace.

FiPy 2.0 introduces several syntax changes from FiPy 1.0. We appreciate that this is very inconvenient for our users,
but we hope you’ll agree that the new syntax is easier to read and easier to use. We assure you that this is not something
we do casually; it has been over three years since our last incompatible change (when FiPy 1.0 superceded FiPy 0.1).
All examples included with version 2.0 have been updated to use the new syntax, but any scripts you have written for
FiPy 1.0 will need to be updated. A complete listing of the changes needed to take the FiPy examples scripts from
version 1.0 to version 2.0 can be found at
http://www.matforge.org/fipy/wiki/upgrade1_0examplesTo2_0
but we summarize the necessary changes here. If these tips are not sufficient to make your scripts compatible with
FiPy 2.0, please don’t hesitate to ask for help on the mailing list.
The following items must be changed in your scripts
• The dimension axis of a Variable is now first, not last
>>> x = mesh.getCellCenters()[0]
instead of
>>> x = mesh.getCellCenters()[...,0]
This seemingly arbitrary change simplifies a great many things in FiPy, but the one most noticeable to the user
is that you can now write
>>> x, y = mesh.getCellCenters()
instead of
>>> x = mesh.getCellCenters()[...,0]
>>> y = mesh.getCellCenters()[...,1]
Unfortunately, we cannot reliably automate this conversion, but we find that searching for “...,” and “:,”
finds almost everything. Please don’t blindly “search & replace all” as that is almost bound to create more
problems than it’s worth.
Note: Any vector constants must be reoriented. For instance, in order to offset a Mesh, you must write
180 Chapter 17. Updating FiPy

>>> mesh = Grid2D(...) + ((deltax,), (deltay,))
or
>>> mesh = Grid2D(...) + [[deltax], [deltay]]
instead of
>>> mesh = Grid2D(...) + (deltax, deltay)
• VectorCellVariable and VectorFaceVariable no longer exist. CellVariable and and

FaceVariable now both inherit from MeshVariable, which can have arbitrary rank. A field of scalars
(default) will have rank=0, a field of vectors will have rank=1, etc. You should write
>>> vectorField = CellVariable(mesh=mesh, rank=1)
instead of
>>> vectorField = VectorCellVariable(mesh=mesh)
Note: Because vector fields are properly supported, use vector operations to manipulate them, such as
>>> phase.getFaceGrad().dot((( 0, 1),
... (-1, 0)))
instead of the hackish

>>> phase.getFaceGrad()._take((1, 0), axis=1) * (-1, 1)
• For internal reasons, FiPy now supports CellVariable and FaceVariable objects that contain integers,
but it is not meaningful to solve a PDE for an integer field (FiPy should issue a warning if you try). As a result,
when given, initial values must be specified as floating-point values:
>>> var = CellVariable(mesh=mesh, value=1.)
where they used to be quietly accepted as integers

>>> var = CellVariable(mesh=mesh, value=1)
If the value argument is not supplied, the CellVariable will contain floats, as before.
• The faces argument to BoundaryCondition now takes a mask, instead of a list of Face IDs. Now you
write
>>> X, Y = mesh.getFaaceCenters()
>>> FixedValue(faces=mesh.getExteriorFaces() & (X**2 < 1e-6), value=...)
instead of
>>> exteriorFaces = mesh.getExteriorFaces()
>>> X = exteriorFaces.getCenters()[...,0]
>>> FixedValue(faces=exteriorFaces.where(X**2 < 1e-6), value=...)
With the old syntax, a different call to getCenters() had to be made for each set of Face objects. It was also
extremely difficult to specify boundary conditions that depended both on position in space and on the current
values of any other Variable.
>>> FixedValue(faces=(mesh.getExteriorFaces()
... & (((X**2 < 1e-6)
17.2. examples.updating.update1_0to2_0 181

... & (Y > 3.))

... | (phi.getArithmeticFaceValue()
... < sin(gamma.getArithmeticFaceValue())))), value=...)
although it probably could have been done with a rather convoluted (and slow!) filter function passed to
where. There no longer are any filter methods used in FiPy. You now would write
>>> initialArray[(x < dx) | (x > (Lx - dx)) | (y < dy) | (y > (Ly - dy))] = 1.
instead of the much slower

>>> def cellFilter(cell):
... return ((cell.center[0] < dx)
... or (cell.center[0] > (Lx - dx))
... or (cell.center[1] < dy)
... or (cell.center[1] > (Ly - dy)))
>>> positiveCells = mesh.getCells(filter=cellFilter)

>>> for cell in positiveCells:
... initialArray[cell.ID] = 1.
Although they still exist, we find very little cause to ever call getCells() or
fipy.meshes.mesh.Mesh.getFaces().
• Some modules, such as fipy.solvers, have been significantly rearranged. For example, you need to change
>>> from fipy.solvers.linearPCGSolver import LinearPCGSolver
to either
>>> from fipy import LinearPCGSolver
or
>>> from fipy.solvers.pysparse.linearPCGSolver import LinearPCGSolver
• The numerix.max() and numerix.min() functions no longer exist. Either call max() and min() or
the max() and min() methods of a Variable.
• The Numeric module has not been supported for a long time. Be sure to use
>>> from fipy import numerix
instead of
>>> import Numeric
The remaining changes are not required, but they make scripts easier to read and we recommend them. FiPy may issue
a DeprecationWarning for some cases, to indicate that we may not maintain the old syntax indefinitely.
• All of the most commonly used classes and functions in FiPy are directly accessible in the fipy namespace.
For brevity, our examples now start with
instead of the explicit

>>> from fipy.meshes.grid1D import Grid1D
>>> from fipy.terms.powerLawConvectionTerm import PowerLawConvectionTerm
>>> from fipy.variables.cellVariable import CellVariable

imports that we used to use. Most of the explicit imports should continue to work, so you do not need to change
them if you don’t wish to, but we find our own scripts much easier to read without them.
All of the numerix module is now imported into the fipy namespace, so you can call numerix functions a
number of different ways, including:
>>> y = exp(x)
or
>>> y = numerix.exp(x)
or
>>> y = exp(x)
We generally use the first, but you may see us use the others, and should feel free to use whichever form you
find most comfortable.
Note: Internally, FiPy uses explicit imports, as is considered best Python practice, but we feel that clarity
trumps orthodoxy when it comes to the examples.
• The function fipy.viewers.make() has been renamed to fipy.viewers.Viewer(). All of the

limits can now be supplied as direct arguments, as well (although this is not required). The result is a
more natural syntax:
>>> from fipy import Viewer
>>> viewer = Viewer(vars=(alpha, beta, gamma), datamin=0, datamax=1)
instead of
>>> from fipy import viewers
>>> viewer = viewers.make(vars=(alpha, beta, gamma),
... limits={'datamin': 0, 'datamax': 1})
With the old syntax, there was also a temptation to write

>>> from fipy.viewers import make
>>> viewer = make(vars=(alpha, beta, gamma))
which can be very hard to understand after the fact (make? make what?).
• A ConvectionTerm can now calculate its Peclet number automatically, so the diffusionTerm argument
is no longer required
... == DiffusionTerm(coeff=diffCoeff)
... + PowerLawConvectionTerm(coeff=convCoeff))
instead of
>>> diffTerm = DiffusionTerm(coeff=diffCoeff)
... == diffTerm
... + PowerLawConvectionTerm(coeff=convCoeff, diffusionTerm=diffTerm))
• An ImplicitSourceTerm now “knows” how to partion itself onto the solution matrix, so you can write

>>> S0 = mXi * phase * (1 - phase) - phase * S1

>>> source = S0 + ImplicitSourceTerm(coeff=S1)
instead of
>>> S0 = mXi * phase * (1 - phase) - phase * S1 * (S1 < 0)
>>> source = S0 + ImplicitSourceTerm(coeff=S1 * (S1 < 0))
It is definitely still advantageous to hand-linearize your source terms, but it is no longer necessary to worry about
putting the “wrong” sign on the diagonal of the matrix.
• To make clearer the distinction between iterations, timesteps, and sweeps (see FAQ Iterations, timesteps, and
sweeps? Oh, my!) the steps argument to a Solver object has been renamed iterations.
• ImplicitDiffusionTerm has been renamed to DiffusionTerm.

It seems unlikely that many users are still running FiPy 0.1, but for those that are, the syntax of FiPy scripts changed
considerably between version 0.1 and version 1.0. We incremented the full version-number to stress that previous
scripts are incompatible. We strongly believe that these changes are for the better, resulting in easier code to write and
read as well as slightly improved efficiency, but we realize that this represents an inconvenience to our users that have
already written scripts of their own. We will strive to avoid any such incompatible changes in the future.
Any scripts you have written for FiPy 0.1 should be updated in two steps, first to work with FiPy
1.0, and then with FiPy 2.0. As a tutorial for updating your scripts, we will walk through updating
examples/convection/exponential1D/input.py from FiPy 0.1. If you attempt to run that script with
FiPy 1.0, the script will fail and you will see the errors shown below:
This example solves the steady-state convection-diffusion equation given by:
∇ · (𝐷∇𝜑 + ⃗𝑢𝜑) = 0
with coefficients 𝐷 = 1 and ⃗𝑢 = (10, 0), or

>>> diffCoeff = 1.
>>> convCoeff = (10.,0.)
We define a 1D mesh
>>> L = 10.
>>> nx = 1000
>>> ny = 1
>>> mesh = Grid2D(L / nx, L / ny, nx, ny)

{︃
0 at 𝑥 = 0,
𝜑=
1 at 𝑥 = 𝐿,
or

>>> valueLeft = 0.
>>> valueRight = 1.
>>> from fipy.boundaryConditions.fixedValue import FixedValue
>>> from fipy.boundaryConditions.fixedFlux import FixedFlux
>>> boundaryConditions = (
... FixedValue(mesh.getFacesLeft(), valueLeft),
... FixedValue(mesh.getFacesRight(), valueRight),
... FixedFlux(mesh.getFacesTop(), 0.),
... FixedFlux(mesh.getFacesBottom(), 0.)
... )

>>> var = CellVariable(
... name = "concentration",
... mesh = mesh,
... value = valueLeft)
The SteadyConvectionDiffusionScEquation object is used to create the equation. It needs to be passed a

convection term instantiator as follows:
>>> from fipy.terms.exponentialConvectionTerm import ExponentialConvectionTerm
>>> from fipy.solvers import *
>>> from fipy.equations.stdyConvDiffScEquation import SteadyConvectionDiffusionScEquation
Traceback (most recent call last):
...
ImportError: No module named equations.stdyConvDiffScEquation
>>> eq = SteadyConvectionDiffusionScEquation(
... var = var,
... diffusionCoeff = diffCoeff,
... convectionCoeff = convCoeff,
... solver = LinearLUSolver(tolerance = 1.e-15, steps = 2000),
... convectionScheme = ExponentialConvectionTerm,
... boundaryConditions = boundaryConditions
... )
...
NameError: name 'SteadyConvectionDiffusionScEquation' is not defined
More details of the benefits and drawbacks of each type of convection term can be found in the numerical section
of the manual. Essentially the ExponentialConvectionTerm and PowerLawConvectionTerm will both
handle most types of convection diffusion cases with the PowerLawConvectionTerm being more efficient.
We iterate to equilibrium
>>> from fipy.iterators.iterator import Iterator
>>> it = Iterator((eq,))
...
NameError: name 'eq' is not defined
>>> it.timestep()
...
NameError: name 'it' is not defined
and test the solution against the analytical result

1 − exp(−𝑢𝑥 𝑥/𝐷)
𝜑=
1 − exp(−𝑢𝑥 𝐿/𝐷)

or
>>> axis = 0
>>> x = mesh.getCellCenters()[:,axis]
>>> from fipy.tools import numerix
>>> analyticalArray = CC / DD
>>> numerix.allclose(analyticalArray, var, rtol = 1e-10, atol = 1e-10)
0

>>> if __name__ == '__main__':
... from fipy.viewers.grid2DGistViewer import Grid2DGistViewer
...
ImportError: No module named grid2DGistViewer
... viewer = Grid2DGistViewer(var)

... viewer.plot()
We see that a number of errors are thrown:

• ImportError: No module named equations.stdyConvDiffScEquation
• NameError: name ’SteadyConvectionDiffusionScEquation’ is not defined
• NameError: name ’eq’ is not defined
• NameError: name ’it’ is not defined
• ImportError: No module named grid2DGistViewer
As is usually the case with computer programming, many of these errors are caused by earlier errors. Let us update
the script, section by section:
Although no error was generated by the use of Grid2D, FiPy 1.0 supports a true 1D mesh class, so we instantiate the
mesh as
>>> L = 10.
>>> nx = 1000
>>> mesh = Grid1D(dx = L / nx, nx = nx)
The Grid2D class with ny = 1 still works perfectly well for 1D problems, but the Grid1D class is slightly more
efficient, and it makes the code clearer when a 1D geometry is actually desired.
Because the mesh is now 1D, we must update the convection coefficient vector to be 1D as well
>>> diffCoeff = 1.
The FixedValue boundary conditions at the left and right are unchanged, but a Grid1D mesh does not even have
top and bottom faces:
>>> valueLeft = 0.
>>> valueRight = 1.
>>> from fipy.boundaryConditions.fixedValue import FixedValue
>>> boundaryConditions = (
... FixedValue(mesh.getFacesLeft(), valueLeft),
... FixedValue(mesh.getFacesRight(), valueRight))

The creation of the solution variable is unchanged:

>>> var = CellVariable(name = "concentration",
... mesh = mesh,
... value = valueLeft)
The biggest change between FiPy 0.1 and FiPy 1.0 is that Equation objects no longer exist at all. Instead, Term ob-
jects can be simply added, subtracted, and equated to assemble an equation. Where before the assembly of the equation
occurred in the black-box of SteadyConvectionDiffusionScEquation, we now assemble it directly:
>>> from fipy.terms.implicitDiffusionTerm import ImplicitDiffusionTerm
>>> diffTerm = ImplicitDiffusionTerm(coeff = diffCoeff)
>>> from fipy.terms.exponentialConvectionTerm import ExponentialConvectionTerm

>>> eq = diffTerm + ExponentialConvectionTerm(coeff = convCoeff,
... diffusionTerm = diffTerm)
One thing that SteadyConvectionDiffusionScEquation took care of automatically was that a

ConvectionTerm must know about any DiffusionTerm in the equation in order to calculate a Peclet number.
Now, the DiffusionTerm must be explicitly passed to the ConvectionTerm in the diffusionTerm parameter.
The Iterator class still exists, but it is no longer necessary. Instead, the solution to an implicit steady-state problem
like this can simply be obtained by telling the equation to solve itself (with an appropriate solver if desired, although
the default LinearPCGSolver is usually suitable):
>>> eq.solve(var = var,
... solver = LinearLUSolver(tolerance = 1.e-15, steps = 2000),
... boundaryConditions = boundaryConditions)
Note: In version 0.1, the Equation object had to be told about the Variable, Solver, and
BoundaryCondition objects when it was created (and it, in turn, passed much of this information to the Term
objects in order to create them). In version 1.0, the Term objects (and the equation assembled from them) are abstract.
The Variable, Solver, and BoundaryCondition objects are only needed by the solve() method (and, in
fact, the same equation could be used to solve different variables, with different solvers, subject to different boundary
conditions, if desired).
The analytical solution is unchanged, and we can test as before

>>> numerix.allclose(analyticalArray, var, rtol = 1e-10, atol = 1e-10)
1
or we can use the slightly simpler syntax

>>> print var.allclose(analyticalArray, rtol = 1e-10, atol = 1e-10)
1
The ImportError: No module named grid2DGistViewer results because the Viewer classes have
been moved and renamed. This error could be resolved by changing the import statement appropriately:
>>> if __name__ == '__main__':
... from fipy.viewers.gistViewer.gist1DViewer import Gist1DViewer
... viewer = Gist1DViewer(vars = var)
... viewer.plot()
Instead, rather than instantiating a particular Viewer (which you can still do, if you desire), a generic “factory”
method will return a Viewer appropriate for the supplied Variable object(s):

>>> if __name__ == '__main__':

... import fipy.viewers
... viewer = fipy.viewers.make(vars = var)
... viewer.plot()
Please do not hesitate to contact us if this example does not help you convert your existing scripts to FiPy 1.0.

Part III
fipy Package Documentation
189
Chapter 18
How to Read the Modules Documentation
Each chapter describes one of the main sub-packages of the fipy package. The sub-package fipy.package
can be found in the directory fipy/package/. In a few cases, there will be packages within packages, e.g.
fipy.package.subpackage located in fipy/package/subpackage/. These sub-sub-packages will not
be given their own chapters; rather, their contents will be described in the chapter for their containing package.
18.1 package.subpackage package
18.1.1 Submodules
18.1.2 package.subpackage.base module
This module can be found in the file package/subpackage/base.py. You make it available to your script by
either:
import package.subpackage.base
in which case you refer to it by its full name of package.subpackage.base, or:

from package.subpackage import base
in which case you can refer simply to base.

class package.subpackage.base.Base
With very few exceptions, the name of a class will be the capitalized form of the module it resides in. Depending
on how you imported the module above, you will refer to either package.subpackage.object.Object
or object.Object. Alternatively, you can use:
from package.subpackage.object import Object
and then refer simply to Object. For many classes, there is a shorthand notation:
from fipy import Object
Python is an object-oriented language and the FiPy framework is composed of objects or classes. Knowledge of
object-oriented programming (OOP) is not necessary to use either Python or FiPy, but a few concepts are useful.
OOP involves two main ideas:
encapsulation an object binds data with actions or “methods”. In most cases, you will not work with an
object’s data directly; instead, you will set, retrieve, or manipulate the data using the object’s methods.
Methods are functions that are attached to objects and that have direct access to the data of those objects.
Rather than passing the object data as an argument to a function:
191
fn(data, arg1, arg2, ...)
you instruct an object to invoke an appropriate method:

object.meth(arg1, arg2, ...)
If you are unfamiliar with object-oriented practices, there probably seems little advantage in this reorder-
ing. You will have to trust us that the latter is a much more powerful way to do things.
inheritance specialized objects are derived or inherited from more general objects. Common behaviors or data
are defined in base objects and specific behaviors or data are either added or modified in derived objects.
Objects that declare the existence of certain methods, without actually defining what those methods do,
are called “abstract”. These objects exist to define the behavior of a family of objects, but rely on their
descendants to actually provide that behavior.
Unlike many object-oriented languages, Python does not prevent the creation of abstract objects, but we
will include a notice like
Attention: This class is abstract. Always create one of its subclasses.
for abstract classes which should be used for documentation but never actually created in a FiPy script.
method1()
This is one thing that you can instruct any object that derives from Base to do, by calling
myObjectDerivedFromBase.method1()
Parameters
• self : this special argument refers to the object that is being created.
Attention: self is supplied automatically by the Python interpreter to all methods.

You don’t need to (and should not) specify it yourself.
method2()
This is another thing that you can instruct any object that derives from Base to do.
18.1.3 package.subpackage.object module
class package.subpackage.object.Object(arg1, arg2=None, arg3=’string’)

Bases: package.subpackage.base.Base
This method, like all those whose names begin and end with “__” are special. You won’t ever need to call these
methods directly, but Python will invoke them for you under certain circumstances, which are described in the
Python Reference Manual: Special Method Names.
As an example, the __init__() method is invoked when you create an object, as in:
obj = Object(arg1=something, arg3=somethingElse, ...)
Parameters
• arg1: this argument is required. Python supports named arguments, so you must either list
the value for arg1 first:
obj = Object(val1, val2)
or you can specify the arguments in any order, as long as they are named:
192 Chapter 18. How to Read the Modules Documentation

obj = Object(arg2=val2, arg1=val1)
• arg2: this argument may be omitted, in which case it will be assigned a default value of
None. If you do not use named arguments (and we recommend that you do), all required
arguments must be specified before any optional arguments.
• arg3: this argument may be omitted, in which case it will be assigned a default value of
’string’.
method2()
Object provides a new definition for the behavior of method2(), whereas the behavior of method1()
is defined by Base.
18.1.4 Module contents
Each chapter describes one of the main sub-packages of the fipy package. The sub-package fipy.package
can be found in the directory fipy/package/. In a few cases, there will be packages within packages, e.g.
fipy.package.subpackage located in fipy/package/subpackage/. These sub-sub-packages will not
be given their own chapters; rather, their contents will be described in the chapter for their containing package.
18.1. package.subpackage package 193

194 Chapter 18. How to Read the Modules Documentation

Chapter 19
fipy.boundaryConditions package
19.1 Submodules
19.2 fipy.boundaryConditions.boundaryCondition module
class fipy.boundaryConditions.boundaryCondition.BoundaryCondition(faces, value)

Bases: object
Generic boundary condition base class.
Parameters
• faces: A list or tuple of exterior Face objects to which this condition applies.
• value: The value to impose.
19.3 fipy.boundaryConditions.constraint module
class fipy.boundaryConditions.constraint.Constraint(value, where=None)

Bases: object
Object to hold a Variable to value at where
see constrain()
19.4 fipy.boundaryConditions.fixedFlux module
class fipy.boundaryConditions.fixedFlux.FixedFlux(faces, value)

Bases: fipy.boundaryConditions.boundaryCondition.BoundaryCondition
The FixedFlux boundary condition adds a contribution, equivalent to a fixed flux (Neumann condition), to the
equation’s RHS vector. The contribution, given by value, is only added to entries corresponding to the specified
faces, and is weighted by the face areas.
Creates a FixedFlux object.
Parameters
• faces: A list or tuple of Face objects to which this condition applies.
195
19.5 fipy.boundaryConditions.fixedValue module
class fipy.boundaryConditions.fixedValue.FixedValue(faces, value)

The FixedValue boundary condition adds a contribution, equivalent to a fixed value (Dirichlet condition), to the
equation’s RHS vector and coefficient matrix. The contributions are given by −value × 𝐺face for the RHS
vector and 𝐺face for the coefficient matrix. The parameter 𝐺face represents the term’s geometric coefficient,
which depends on the type of term and the mesh geometry.
Contributions are only added to entries corresponding to the specified faces.
Parameters
19.6 fipy.boundaryConditions.nthOrderBoundaryCondition module
class fipy.boundaryConditions.nthOrderBoundaryCondition.NthOrderBoundaryCondition(faces,
value,
or-
der)
This boundary condition is generally used in conjunction with a ImplicitDiffusionTerm that has multiple coeffi-
cients. It does not have any direct effect on the solution matrices, but its derivatives do.
Creates an NthOrderBoundaryCondition.
Parameters
• order: The order of the boundary condition. An order of 0 corresponds to a FixedValue and
an order of 1 corresponds to a FixedFlux. Even and odd orders behave like FixedValue and
FixedFlux objects, respectively, but apply to higher order terms.
19.7 fipy.boundaryConditions.test module
Test numeric implementation of the mesh
19.8 Module contents
class fipy.boundaryConditions.Constraint(value, where=None)

Bases: object
Object to hold a Variable to value at where
see constrain()
196 Chapter 19. fipy.boundaryConditions package

class fipy.boundaryConditions.FixedFlux(faces, value)

The FixedFlux boundary condition adds a contribution, equivalent to a fixed flux (Neumann condition), to the
equation’s RHS vector. The contribution, given by value, is only added to entries corresponding to the specified
faces, and is weighted by the face areas.
Creates a FixedFlux object.
Parameters
class fipy.boundaryConditions.FixedValue(faces, value)
The FixedValue boundary condition adds a contribution, equivalent to a fixed value (Dirichlet condition), to the
equation’s RHS vector and coefficient matrix. The contributions are given by −value × 𝐺face for the RHS
vector and 𝐺face for the coefficient matrix. The parameter 𝐺face represents the term’s geometric coefficient,
which depends on the type of term and the mesh geometry.
Contributions are only added to entries corresponding to the specified faces.
Parameters
class fipy.boundaryConditions.NthOrderBoundaryCondition(faces, value, order)
This boundary condition is generally used in conjunction with a ImplicitDiffusionTerm that has multiple coeffi-
cients. It does not have any direct effect on the solution matrices, but its derivatives do.
Creates an NthOrderBoundaryCondition.
Parameters
• order: The order of the boundary condition. An order of 0 corresponds to a FixedValue and
an order of 1 corresponds to a FixedFlux. Even and odd orders behave like FixedValue and
FixedFlux objects, respectively, but apply to higher order terms.
19.8. Module contents 197

198 Chapter 19. fipy.boundaryConditions package

Chapter 20
fipy.matrices package
20.1 Submodules
20.2 fipy.matrices.offsetSparseMatrix module
fipy.matrices.offsetSparseMatrix.OffsetSparseMatrix(SparseMatrix, numberOfVari-
ables, numberOfEquations)
Used in binary terms. equationIndex and varIndex need to be set statically before instantiation.
20.3 fipy.matrices.pysparseMatrix module
20.4 fipy.matrices.scipyMatrix module
20.5 fipy.matrices.sparseMatrix module
20.6 fipy.matrices.test module
20.7 fipy.matrices.trilinosMatrix module
199
200 Chapter 20. fipy.matrices package

Chapter 21
fipy.meshes package
21.1 Subpackages
21.1.1 fipy.meshes.builders package
Submodules
fipy.meshes.builders.abstractGridBuilder module
fipy.meshes.builders.grid1DBuilder module
fipy.meshes.builders.periodicGrid1DBuilder module
fipy.meshes.builders.utilityClasses module
Module contents
21.1.2 fipy.meshes.numMesh package
Submodules
fipy.meshes.numMesh.cylindricalGrid1D module
fipy.meshes.numMesh.cylindricalGrid2D module
fipy.meshes.numMesh.cylindricalUniformGrid1D module
fipy.meshes.numMesh.cylindricalUniformGrid2D module
fipy.meshes.numMesh.deprecatedWarning module
fipy.meshes.numMesh.deprecatedWarning.numMeshDeprecated()
201
fipy.meshes.numMesh.gmshImport module
fipy.meshes.numMesh.grid1D module
fipy.meshes.numMesh.periodicGrid1D module
fipy.meshes.numMesh.periodicGrid2D module
fipy.meshes.numMesh.skewedGrid2D module
fipy.meshes.numMesh.tri2D module
fipy.meshes.numMesh.uniformGrid1D module
Module contents
21.1.3 fipy.meshes.representations package
Submodules
fipy.meshes.representations.abstractRepresentation module
fipy.meshes.representations.gridRepresentation module
fipy.meshes.representations.meshRepresentation module
Module contents
21.1.4 fipy.meshes.topologies package
Submodules
fipy.meshes.topologies.abstractTopology module
fipy.meshes.topologies.gridTopology module
fipy.meshes.topologies.meshTopology module
202 Chapter 21. fipy.meshes package

Module contents
21.2 Submodules
21.3 fipy.meshes.abstractMesh module
class fipy.meshes.abstractMesh.AbstractMesh(communicator, _RepresentationClass=<class

‘fipy.meshes.representations.abstractRepresentation._AbstractRepresenta
_TopologyClass=<class
‘fipy.meshes.topologies.abstractTopology._AbstractTopology’>)
Bases: object
A class encapsulating all commonalities among meshes in FiPy.
VTKCellDataSet
Returns a TVTK DataSet representing the cells of this mesh
VTKFaceDataSet
Returns a TVTK DataSet representing the face centers of this mesh
aspect2D
The physical y:x aspect ratio of a 2D mesh
cellCenters
cellDistanceVectors
cellFaceIDs
Topology properties
cellToFaceDistanceVectors
cellVolumes
extents
exteriorFaces
faceCenters
facesBack
Return list of faces on back boundary of Grid3D with the z-axis running from front to back.
>>> from fipy import Grid3D, numerix
>>> mesh = Grid3D(nx = 3, ny = 2, nz = 1, dx = 0.5, dy = 2., dz = 4.)
>>> print numerix.allequal((6, 7, 8, 9, 10, 11),
... numerix.nonzero(mesh.facesBack)[0])
True
facesBottom
Return list of faces on bottom boundary of Grid3D with the y-axis running from bottom to top.
>>> from fipy import Grid2D, Grid3D, numerix
>>> print numerix.allequal((12, 13, 14),
... numerix.nonzero(mesh.facesBottom)[0])
1
>>> x, y, z = mesh.faceCenters
>>> print numerix.allequal((12, 13),
... numerix.nonzero(mesh.facesBottom & (x < 1))[0])
1
21.2. Submodules 203

facesDown
Return list of faces on bottom boundary of Grid3D with the y-axis running from bottom to top.
... numerix.nonzero(mesh.facesBottom)[0])
1
>>> x, y, z = mesh.faceCenters
... numerix.nonzero(mesh.facesBottom & (x < 1))[0])
1
facesFront
Return list of faces on front boundary of Grid3D with the z-axis running from front to back.
>>> from fipy import Grid3D, numerix
>>> print numerix.allequal((0, 1, 2, 3, 4, 5),
... numerix.nonzero(mesh.facesFront)[0])
True
facesLeft
Return face on left boundary of Grid1D as list with the x-axis running from left to right.
>>> from fipy import Grid2D, Grid3D
... numerix.nonzero(mesh.facesLeft)[0])
True
>>> mesh = Grid2D(nx = 3, ny = 2, dx = 0.5, dy = 2.)
... numerix.nonzero(mesh.facesLeft)[0])
True
facesRight
Return list of faces on right boundary of Grid3D with the x-axis running from left to right.
... numerix.nonzero(mesh.facesRight)[0])
True
... numerix.nonzero(mesh.facesRight)[0])
True
facesTop
Return list of faces on top boundary of Grid3D with the y-axis running from bottom to top.
... numerix.nonzero(mesh.facesTop)[0])
True
True

facesUp
Return list of faces on top boundary of Grid3D with the y-axis running from bottom to top.
True
True
getCellCenters(*args, **kwds)
Deprecated since version 3.0: use the cellCenters property instead
getCellVolumes(*args, **kwds)
Deprecated since version 3.0: use the cellVolumes property instead
getDim(*args, **kwds)
Deprecated since version 3.0: use the dim property instead
getExteriorFaces(*args, **kwds)
Deprecated since version 3.0: use the exteriorFaces property instead
Return only the faces that have one neighboring cell.
getFaceCellIDs(*args, **kwds)
Deprecated since version 3.0: use the faceCellIDs property instead
getFaceCenters(*args, **kwds)
Deprecated since version 3.0: use the faceCenters property instead
getFacesBack(*args, **kwds)
Deprecated since version 3.0: use the facesBack property instead
getFacesBottom(*args, **kwds)
Deprecated since version 3.0: use the facesBottom property instead
getFacesDown(*args, **kwds)
Deprecated since version 3.0: use the facesBottom property instead
getFacesFront(*args, **kwds)
Deprecated since version 3.0: use the facesFront property instead
getFacesLeft(*args, **kwds)
Deprecated since version 3.0: use the facesLeft property instead
getFacesRight(*args, **kwds)
Deprecated since version 3.0: use the facesRight property instead
getFacesTop(*args, **kwds)
Deprecated since version 3.0: use the facesTop property instead
getFacesUp(*args, **kwds)
Deprecated since version 3.0: use the facesTop property instead
getInteriorFaceCellIDs(*args, **kwds)
Deprecated since version 3.0: use the interiorFaceCellIDs property instead
getInteriorFaceIDs(*args, **kwds)
Deprecated since version 3.0: use the interiorFaceIDs property instead
21.3. fipy.meshes.abstractMesh module 205

getInteriorFaces(*args, **kwds)
Deprecated since version 3.0: use the interiorFaces property instead
Return only the faces that have two neighboring cells.
getNearestCell(point)
getNumberOfCells(*args, **kwds)
Deprecated since version 3.0: use the numberOfCells property instead
getPhysicalShape(*args, **kwds)
Deprecated since version 3.0: use the physicalShape property instead
getScale(*args, **kwds)
Deprecated since version 3.0: use the scale property instead
getShape(*args, **kwds)
Deprecated since version 3.0: use the shape property instead
getVertexCoords(*args, **kwds)
Deprecated since version 3.0: use the vertexCoords property instead
interiorFaceCellIDs
interiorFaceIDs
interiorFaces
scale
scaledCellDistances
scaledCellToCellDistances
scaledCellVolumes
scaledFaceAreas
scaledFaceToCellDistances
setScale(*args, **kwds)
Deprecated since version 3.0: use the scale property instead
x
Equivalent to using cellCenters[0].
>>> print Grid1D(nx=2).x
[ 0.5 1.5]
y
>>> print Grid2D(nx=2, ny=2).y
[ 0.5 0.5 1.5 1.5]
>>> print Grid1D(nx=2).y
...
AttributeError: 1D meshes do not have a "y" attribute.
z


>>> print Grid3D(nx=2, ny=2, nz=2).z
[ 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5]
>>> print Grid2D(nx=2, ny=2).z
...
AttributeError: 1D and 2D meshes do not have a "z" attribute.
21.4 fipy.meshes.cylindricalGrid1D module
21.5 fipy.meshes.cylindricalGrid2D module
21.6 fipy.meshes.cylindricalNonUniformGrid1D module
1D Mesh
class fipy.meshes.cylindricalNonUniformGrid1D.CylindricalNonUniformGrid1D(dx=1.0,
nx=None,
ori-
gin=(0,
),
over-
lap=2,
com-
mu-
ni-
ca-
tor=DummyComm(),
*args,
**kwargs)
Bases: fipy.meshes.nonUniformGrid1D.NonUniformGrid1D
Creates a 1D cylindrical grid mesh.
>>> mesh = CylindricalNonUniformGrid1D(nx = 3)
>>> print mesh.cellCenters
[[ 0.5 1.5 2.5]]
>>> mesh = CylindricalNonUniformGrid1D(dx = (1, 2, 3))

[[ 0.5 2. 4.5]]
>>> print numerix.allclose(mesh.cellVolumes, (0.5, 4., 13.5))

True
>>> mesh = CylindricalNonUniformGrid1D(nx = 2, dx = (1, 2, 3))

...
IndexError: nx != len(dx)
>>> mesh = CylindricalNonUniformGrid1D(nx=2, dx=(1., 2.)) + ((1.,),)

[[ 1.5 3. ]]
21.4. fipy.meshes.cylindricalGrid1D module 207

>>> print numerix.allclose(mesh.cellVolumes, (1.5, 6))

True
21.7 fipy.meshes.cylindricalNonUniformGrid2D module
2D rectangular Mesh
class fipy.meshes.cylindricalNonUniformGrid2D.CylindricalNonUniformGrid2D(dx=1.0,
dy=1.0,
nx=None,
ny=None,
ori-
gin=((0.0,
),
(0.0,
)),
over-
lap=2,
com-
mu-
ni-
ca-
tor=DummyComm(),
*args,
**kwargs)
Creates a 2D cylindrical grid mesh with horizontal faces numbered first and then vertical faces.
21.8 fipy.meshes.cylindricalUniformGrid1D module
1D Mesh
class fipy.meshes.cylindricalUniformGrid1D.CylindricalUniformGrid1D(dx=1.0,
nx=1, ori-
gin=(0, ),
overlap=2,
communica-
tor=DummyComm(),
*args,
**kwargs)
Bases: fipy.meshes.uniformGrid1D.UniformGrid1D
Creates a 1D cylindrical grid mesh.
>>> mesh = CylindricalUniformGrid1D(nx = 3)
[[ 0.5 1.5 2.5]]
cellVolumes

21.9 fipy.meshes.cylindricalUniformGrid2D module
2D cylindrical rectangular Mesh with constant spacing in x and constant spacing in y

class fipy.meshes.cylindricalUniformGrid2D.CylindricalUniformGrid2D(dx=1.0,
dy=1.0,
nx=1, ny=1,
origin=((0,
), (0, )),
overlap=2,
communica-
tor=DummyComm(),
*args,
**kwargs)
Bases: fipy.meshes.uniformGrid2D.UniformGrid2D
Creates a 2D cylindrical grid in the radial and axial directions, appropriate for axial symmetry.
cellVolumes
21.10 fipy.meshes.factoryMeshes module
fipy.meshes.factoryMeshes.Grid3D(dx=1.0, dy=1.0, dz=1.0, nx=None, ny=None, nz=None,

Lx=None, Ly=None, Lz=None, overlap=2, communica-
tor=DummyComm())
Factory function to select between UniformGrid3D and NonUniformGrid3D. If Lx is specified the length of the
domain is always Lx regardless of dx.
Parameters
• dx: grid spacing in the horizontal direction
• dy: grid spacing in the vertical direction
• dz: grid spacing in the z-direction
• nx: number of cells in the horizontal direction
• ny: number of cells in the vertical direction
• nz: number of cells in the z-direction
• Lx: the domain length in the horizontal direction
• Ly: the domain length in the vertical direction
• Lz: the domain length in the z-direction
• overlap: the number of overlapping cells for parallel simulations. Generally 2 is adequate.
Higher order equations or discretizations require more.
• communicator: either fipy.tools.parallelComm or fipy.tools.serialComm. Select
fipy.tools.serialComm to create a serial mesh when running in parallel. Mostly used for
test purposes.
fipy.meshes.factoryMeshes.Grid2D(dx=1.0, dy=1.0, nx=None, ny=None, Lx=None, Ly=None,
overlap=2, communicator=DummyComm())
Parameters
21.9. fipy.meshes.cylindricalUniformGrid2D module 209


test purposes.
>>> print Grid2D(Lx=3., nx=2).dx
1.5
fipy.meshes.factoryMeshes.Grid1D(dx=1.0, nx=None, Lx=None, overlap=2, communica-

tor=DummyComm())
Parameters
• dx: grid spacing in the horizonal direction
• nx: number of cells in the horizonal direction
• Lx: the domain length in the horizonal direction
test purposes.
fipy.meshes.factoryMeshes.CylindricalGrid2D(dr=None, dz=None, nr=None, nz=None,
Lr=None, Lz=None, dx=1.0, dy=1.0,
nx=None, ny=None, Lx=None, Ly=None,
origin=((0, ), (0, )), overlap=2, communica-
tor=DummyComm())
Factory function to select between CylindricalUniformGrid2D and CylindricalNonUniformGrid2D. If Lx is
specified the length of the domain is always Lx regardless of dx.
Parameters
• dr or dx: grid spacing in the radial direction
• dz or dy: grid spacing in the vertical direction
• nr or nx: number of cells in the radial direction
• nz or ny: number of cells in the vertical direction
• Lr or Lx: the domain length in the radial direction
• Lz or Ly: the domain length in the vertical direction
• origin : position of the mesh’s origin in the form ((x,),(y,))

test purposes.
fipy.meshes.factoryMeshes.CylindricalGrid1D(dr=None, nr=None, Lr=None, dx=1.0,
nx=None, Lx=None, origin=(0, ), over-
lap=2, communicator=DummyComm())
Parameters
• origin : position of the mesh’s origin in the form (x,)
test purposes.
21.11 fipy.meshes.gmshMesh module
fipy.meshes.gmshMesh.openMSHFile(name, dimensions=None, coordDimensions=None, com-

municator=DummyComm(), order=1, mode=’r’, back-
ground=None)
Open a Gmsh MSH file
Parameters
• filename: a string indicating gmsh output file
• dimensions: an integer indicating dimension of mesh
• coordDimensions: an integer indicating dimension of shapes
• order: ???
• mode: a string beginning with ‘r’ for reading and ‘w’ for writing. The file will be created if
it doesn’t exist when opened for writing; it will be truncated when opened for writing. Add
a ‘b’ to the mode for binary files.
• background: a CellVariable that specifies the desired characteristic lengths of the mesh cells
fipy.meshes.gmshMesh.openPOSFile(name, communicator=DummyComm(), mode=’w’)
Open a Gmsh POS post-processing file
class fipy.meshes.gmshMesh.Gmsh2D(arg, coordDimensions=2, communicator=DummyComm(), or-
der=1, background=None)
Bases: fipy.meshes.mesh2D.Mesh2D
Construct a 2D Mesh using Gmsh
21.11. fipy.meshes.gmshMesh module 211

>>> radius = 5.
>>> side = 4.
>>> squaredCircle = Gmsh2D('''
... // A mesh consisting of a square inside a circle inside a circle
...
... // define the basic dimensions of the mesh
...
... cellSize = 1;
... side = %(side)g;
...
... // define the compass points of the inner circle
...
... Point(1) = {0, 0, 0, cellSize};
...
... // define the compass points of the outer circle
...
... Point(6) = {-2*radius, 0, 0, cellSize};
... Point(7) = {0, 2*radius, 0, cellSize};
... Point(8) = {2*radius, 0, 0, cellSize};
... Point(9) = {0, -2*radius, 0, cellSize};
...
... // define the corners of the square
...
... Point(10) = {side/2, side/2, 0, cellSize/2};
... Point(11) = {-side/2, side/2, 0, cellSize/2};
... Point(12) = {-side/2, -side/2, 0, cellSize/2};
... Point(13) = {side/2, -side/2, 0, cellSize/2};
...
... // define the inner circle
...
... Circle(1) = {2, 1, 3};
... Circle(2) = {3, 1, 4};
... Circle(3) = {4, 1, 5};
... Circle(4) = {5, 1, 2};
...
... // define the outer circle
...
... Circle(5) = {6, 1, 7};
... Circle(6) = {7, 1, 8};
... Circle(7) = {8, 1, 9};
... Circle(8) = {9, 1, 6};
...
... // define the square
...
... Line(9) = {10, 13};
... Line(10) = {13, 12};
... Line(11) = {12, 11};
... Line(12) = {11, 10};
...
... // define the three boundaries
...
... Line Loop(1) = {1, 2, 3, 4};
... Line Loop(2) = {5, 6, 7, 8};

... Line Loop(3) = {9, 10, 11, 12};

...
... // define the three domains
...
... Plane Surface(1) = {2, 1};
...
... // label the three domains
...
... // attention: if you use any "Physical" labels, you *must* label
... // all elements that correspond to FiPy Cells (Physical Surace in 2D
... // and Physical Volume in 3D) or Gmsh will not include them and FiPy
... // will not be able to include them in the Mesh.
...
... // note: if you do not use any labels, all Cells will be included.
...
... Physical Surface("Outer") = {1};
... Physical Surface("Middle") = {2};
... Physical Surface("Inner") = {3};
...
... // label the "north-west" part of the exterior boundary
...
... // note: you only need to label the Face elements
... // (Physical Line in 2D and Physical Surface in 3D) that correspond
... // to boundaries you are interested in. FiPy does not need them to
... // construct the Mesh.
...
... Physical Line("NW") = {5};
... ''' % locals())
It can be easier to specify certain domains and boundaries within Gmsh than it is to define the same domains
and boundaries with FiPy expressions.
Here we compare obtaining the same Cells and Faces using FiPy’s parametric descriptions and Gmsh’s labels.
>>> x, y = squaredCircle.cellCenters
>>> middle = ((x**2 + y**2 <= radius**2)

... & ~((x > -side/2) & (x < side/2)
... & (y > -side/2) & (y < side/2)))
>>> print (middle == squaredCircle.physicalCells["Middle"]).all()

True
>>> X, Y = squaredCircle.faceCenters
>>> NW = ((X**2 + Y**2 > (1.99*radius)**2)

... & (X**2 + Y**2 < (2.01*radius)**2)
... & (X <= 0) & (Y >= 0))
>>> print (NW == squaredCircle.physicalFaces["NW"]).all()

True
It is possible to direct Gmsh to give the mesh different densities in different locations
>>> geo = '''
... // A mesh consisting of a square
...


...
... Point(1) = {1, 1, 0, 1};
... Point(2) = {0, 1, 0, 1};
... Point(3) = {0, 0, 0, 1};
... Point(4) = {1, 0, 0, 1};
...
...
... Line(1) = {1, 2};
... Line(2) = {2, 3};
... Line(3) = {3, 4};
... Line(4) = {4, 1};
...
... // define the boundary
...
... Line Loop(1) = {1, 2, 3, 4};
...
... // define the domain
...
... '''
>>> from fipy import CellVariable, numerix
>>> std = []
>>> bkg = None
>>> for refine in range(4):
... square = Gmsh2D(geo, background=bkg)
... x, y = square.cellCenters
... bkg = CellVariable(mesh=square, value=abs(x / 4) + 0.01)
... std.append(numerix.std(numerix.sqrt(2 * square.cellVolumes) / bkg))
Check that the mesh is monotonically approaching the desired density

>>> print numerix.greater(std[:-1], std[1:]).all()
True
and that the final density is close enough to the desired density
>>> print std[-1] < 0.2
True
The initial mesh doesn’t have to be from Gmsh

>>> from fipy import Tri2D
>>> trisquare = Tri2D(nx=1, ny=1)

>>> x, y = trisquare.cellCenters
>>> bkg = CellVariable(mesh=trisquare, value=abs(x / 4) + 0.01)
>>> std1 = numerix.std(numerix.sqrt(2 * trisquare.cellVolumes) / bkg)
>>> square = Gmsh2D(geo, background=bkg)

>>> x, y = square.cellCenters
>>> bkg = CellVariable(mesh=square, value=abs(x / 4) + 0.01)
>>> std2 = numerix.std(numerix.sqrt(2 * square.cellVolumes) / bkg)
>>> print std1 > std2

True

Parameters
• arg: a string giving (i) the path to an MSH file, (ii) a path to a Gmsh geometry
(”.geo”) file, or (iii) a Gmsh geometry script
• order: ???
class fipy.meshes.gmshMesh.Gmsh2DIn3DSpace(arg, communicator=DummyComm(), order=1,

background=None)
Bases: fipy.meshes.gmshMesh.Gmsh2D
Create a topologically 2D Mesh in 3D coordinates using Gmsh
Parameters
• order: ???
class fipy.meshes.gmshMesh.Gmsh3D(arg, communicator=DummyComm(), order=1, back-
ground=None)
Bases: fipy.meshes.mesh.Mesh
Create a 3D Mesh using Gmsh
Parameters
• order: ???
class fipy.meshes.gmshMesh.GmshGrid2D(dx=1.0, dy=1.0, nx=1, ny=None, coordDimensions=2,
communicator=DummyComm(), order=1)
Should serve as a drop-in replacement for Grid2D.
class fipy.meshes.gmshMesh.GmshGrid3D(dx=1.0, dy=1.0, dz=1.0, nx=1, ny=None, nz=None, com-
municator=DummyComm(), order=1)
class fipy.meshes.gmshMesh.GmshImporter2D(arg, coordDimensions=2)
class fipy.meshes.gmshMesh.GmshImporter2DIn3DSpace(arg)
Bases: fipy.meshes.gmshMesh.Gmsh2DIn3DSpace
class fipy.meshes.gmshMesh.GmshImporter3D(arg)

21.12 fipy.meshes.grid1D module
21.15 fipy.meshes.mesh module
exception fipy.meshes.mesh.MeshAdditionError
Bases: exceptions.Exception
class fipy.meshes.mesh.Mesh(vertexCoords, faceVertexIDs, cellFaceIDs, communi-
cator=DummyComm(), _RepresentationClass=<class
‘fipy.meshes.representations.meshRepresentation._MeshRepresentation’>,
_TopologyClass=<class ‘fipy.meshes.topologies.meshTopology._MeshTopology’>)
Bases: fipy.meshes.abstractMesh.AbstractMesh
Generic mesh class using numerix to do the calculations
Meshes contain cells, faces, and vertices.
This is built for a non-mixed element mesh.
21.16 fipy.meshes.mesh1D module

class fipy.meshes.mesh1D.Mesh1D(vertexCoords, faceVertexIDs, cellFaceIDs, communi-
_TopologyClass=<class ‘fipy.meshes.topologies.meshTopology._Mesh1DTopology’>)
21.17 fipy.meshes.mesh2D module

class fipy.meshes.mesh2D.Mesh2D(vertexCoords, faceVertexIDs, cellFaceIDs, communi-
extrude(extrudeFunc=<function <lambda>>, layers=1)
This function returns a new 3D mesh. The 2D mesh is extruded using the extrudeFunc and the number of
layers.

Parameters
• extrudeFunc: function that takes the vertex coordinates and returns the displaced values
• layers: the number of layers in the extruded mesh (number of times extrudeFunc will be
called)
>>> from fipy.meshes.nonUniformGrid2D import NonUniformGrid2D
>>> print NonUniformGrid2D(nx=2,ny=2).extrude(layers=2).cellCenters
[[ 0.5 1.5 0.5 1.5 0.5 1.5 0.5 1.5]
[ 0.5 0.5 1.5 1.5 0.5 0.5 1.5 1.5]
[ 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5]]
>>> from fipy.meshes.tri2D import Tri2D

>>> print Tri2D().extrude(layers=2).cellCenters
[[ 0.83333333 0.5 0.16666667 0.5 0.83333333 0.5
0.16666667 0.5 ]
[ 0.5 0.83333333 0.5 0.16666667 0.5 0.83333333
0.5 0.16666667]
[ 0.5 0.5 0.5 0.5 1.5 1.5 1.5
1.5 ]]
21.18 fipy.meshes.nonUniformGrid1D module
1D Mesh
class fipy.meshes.nonUniformGrid1D.NonUniformGrid1D(dx=1.0, nx=None, overlap=2,
communicator=DummyComm(),
_BuilderClass=<class
‘fipy.meshes.builders.grid1DBuilder._NonuniformGrid1DBui
_RepresentationClass=<class
‘fipy.meshes.representations.gridRepresentation._Grid1DRep
‘fipy.meshes.topologies.gridTopology._Grid1DTopology’>)
Creates a 1D grid mesh.
>>> mesh = NonUniformGrid1D(nx = 3)
[[ 0.5 1.5 2.5]]
>>> mesh = NonUniformGrid1D(dx = (1, 2, 3))

[[ 0.5 2. 4.5]]
>>> mesh = NonUniformGrid1D(nx = 2, dx = (1, 2, 3))

...
IndexError: nx != len(dx)
2D rectangular Mesh
21.18. fipy.meshes.nonUniformGrid1D module 217

class fipy.meshes.nonUniformGrid2D.NonUniformGrid2D(dx=1.0, dy=1.0, nx=None,

ny=None, overlap=2, com-
municator=DummyComm(),
Creates a 2D grid mesh with horizontal faces numbered first and then vertical faces.
class fipy.meshes.nonUniformGrid3D.NonUniformGrid3D(dx=1.0, dy=1.0, dz=1.0, nx=None,

ny=None, nz=None, overlap=2,
3D rectangular-prism Mesh
X axis runs from left to right. Y axis runs from bottom to top. Z axis runs from front to back.
Numbering System:
Vertices: Numbered in the usual way. X coordinate changes most quickly, then Y, then Z.
Cells: Same numbering system as vertices.
Faces: XY faces numbered first, then XZ faces, then YZ faces. Within each subcategory, it is numbered in the
usual way.
21.21 fipy.meshes.periodicGrid1D module
Peridoic 1D Mesh
class fipy.meshes.periodicGrid1D.PeriodicGrid1D(dx=1.0, nx=None, overlap=2, *args,
**kwargs)
Creates a Periodic grid mesh.
>>> mesh = PeriodicGrid1D(dx = (1, 2, 3))
>>> print numerix.allclose(numerix.nonzero(mesh.exteriorFaces)[0],

... [3])
True
>>> print numerix.allclose(mesh.faceCellIDs.filled(-999),

... [[2, 0, 1, 2],
... [0, 1, 2, -999]])
True

>>> print numerix.allclose(mesh._cellDistances,

... [ 2., 1.5, 2.5, 1.5])
True
>>> print numerix.allclose(mesh._cellToCellDistances,

... [[ 2., 1.5, 2.5],
... [ 1.5, 2.5, 2. ]])
True
>>> print numerix.allclose(mesh.faceNormals,

... [[ 1., 1., 1., 1.]])
True
>>> print numerix.allclose(mesh._cellVertexIDs,

... [[1, 2, 2],
... [0, 1, 0]])
True
cellCenters
Defined outside of a geometry class since we need the CellVariable version of cellCenters; that is, the
cellCenters defined in fipy.meshes.mesh and not in any geometry (since a CellVariable requires a reference
to a mesh).
21.22 fipy.meshes.periodicGrid2D module
2D periodic rectangular Mesh

class fipy.meshes.periodicGrid2D.PeriodicGrid2D(dx=1.0, dy=1.0, nx=None,
ny=None, overlap=2, communica-
tor=DummyComm(), *args, **kwargs)
Bases: fipy.meshes.periodicGrid2D._BasePeriodicGrid2D
Creates a periodic2D grid mesh with horizontal faces numbered first and then vertical faces. Vertices and cells
are numbered in the usual way.
>>> mesh = PeriodicGrid2D(dx = 1., dy = 0.5, nx = 2, ny = 2)

... [ 4, 5, 8, 11])
True

... [[2, 3, 0, 1, 2, 3, 1, 0, 1, 3, 2, 3],
... [0, 1, 2, 3, -1, -1, 0, 1, -1, 2, 3, -1]])
True

... [ 0.5, 0.5, 0.5, 0.5, 0.25, 0.25, 1., 1., 0.5, 1., 1., 0.5])
True
>>> print numerix.allclose(mesh.cellFaceIDs,

... [[0, 1, 2, 3],
... [7, 6, 10, 9],
... [2, 3, 0, 1],
21.22. fipy.meshes.periodicGrid2D module 219

... [6, 7, 9, 10]])

True

... [[ 0.5, 0.5, 0.5, 0.5],
... [ 1., 1., 1., 1. ],
... [ 0.5, 0.5, 0.5, 0.5],
... [ 1., 1., 1., 1. ]])
True
>>> normals = [[0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],

... [1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0]]
>>> print numerix.allclose(mesh.faceNormals, normals)

True

... [[4, 5, 7, 8],
... [3, 4, 6, 7],
... [1, 2, 4, 5],
... [0, 1, 3, 4]])
True
class fipy.meshes.periodicGrid2D.PeriodicGrid2DLeftRight(dx=1.0, dy=1.0, nx=None,

*args, **kwargs)
class fipy.meshes.periodicGrid2D.PeriodicGrid2DTopBottom(dx=1.0, dy=1.0, nx=None,
*args, **kwargs)
21.23 fipy.meshes.skewedGrid2D module
class fipy.meshes.skewedGrid2D.SkewedGrid2D(dx=1.0, dy=1.0, nx=None, ny=1, rand=0, *args,

**kwargs)
Creates a 2D grid mesh with horizontal faces numbered first and then vertical faces. The points are skewed by
a random amount (between rand and -rand) in the X and Y directions.
physicalShape
Return physical dimensions of Grid2D.
shape
21.24 fipy.meshes.test module
Test implementation of the mesh

21.25 fipy.meshes.tri2D module
class fipy.meshes.tri2D.Tri2D(dx=1.0, dy=1.0, nx=1, ny=1, _RepresentationClass=<class

‘fipy.meshes.representations.gridRepresentation._Grid2DRepresentation’>,
This class creates a mesh made out of triangles. It does this by starting with a standard Cartesian mesh (Grid2D)
and dividing each cell in that mesh (hereafter referred to as a ‘box’) into four equal parts with the dividing lines
being the diagonals.
Creates a 2D triangular mesh with horizontal faces numbered first then vertical faces, then diagonal faces.
Vertices are numbered starting with the vertices at the corners of boxes and then the vertices at the centers of
boxes. Cells on the right of boxes are numbered first, then cells on the top of boxes, then cells on the left of
boxes, then cells on the bottom of boxes. Within each of the ‘sub-categories’ in the above, the vertices, cells and
faces are numbered in the usual way.
Parameters
• dx, dy: The X and Y dimensions of each ‘box’. If dx <> dy, the line segments connecting
the cell centers will not be orthogonal to the faces.
• nx, ny: The number of boxes in the X direction and the Y direction. The total number of
boxes will be equal to nx * ny, and the total number of cells will be equal to 4 * nx * ny.
physicalShape
shape
21.26 fipy.meshes.uniformGrid module
class fipy.meshes.uniformGrid.UniformGrid(communicator, _RepresentationClass=<class

‘fipy.meshes.representations.abstractRepresentation._AbstractRepresentatio
‘fipy.meshes.topologies.abstractTopology._AbstractTopology’>)
Bases: fipy.meshes.abstractMesh.AbstractMesh
Wrapped scaled geometry properties
21.27 fipy.meshes.uniformGrid1D module
1D Mesh
class fipy.meshes.uniformGrid1D.UniformGrid1D(dx=1.0, nx=1, origin=(0, ), over-
lap=2, communicator=DummyComm(),
‘fipy.meshes.representations.gridRepresentation._Grid1DRepresentati
Bases: fipy.meshes.uniformGrid.UniformGrid
Creates a 1D grid mesh.
21.25. fipy.meshes.tri2D module 221

>>> mesh = UniformGrid1D(nx = 3)

[[ 0.5 1.5 2.5]]
exteriorFaces
Geometry set and calc
faceCellIDs
faceNormals
vertexCoords
2D rectangular Mesh with constant spacing in x and constant spacing in y

class fipy.meshes.uniformGrid2D.UniformGrid2D(dx=1.0, dy=1.0, nx=1, ny=1, ori-
gin=((0, ), (0, )), overlap=2,
Creates a 2D grid mesh with horizontal faces numbered first and then vertical faces.
faceCellIDs
faceNormals
faceVertexIDs
vertexCoords
class fipy.meshes.uniformGrid3D.UniformGrid3D(dx=1.0, dy=1.0, dz=1.0, nx=1, ny=1,

nz=1, origin=[[0], [0], [0]], over-
lap=2, communicator=DummyComm(),
3D rectangular-prism Mesh with uniform grid spacing in each dimension.
X axis runs from left to right. Y axis runs from bottom to top. Z axis runs from front to back.
Numbering System:
Vertices: Numbered in the usual way. X coordinate changes most quickly, then Y, then Z.
* arrays are arranged Z, Y, X because in numerix, the final index is the one that changes the most quickly
*
Cells: Same numbering system as vertices.

Faces: XY faces numbered first, then XZ faces, then YZ faces. Within each subcategory, it is numbered in the
usual way.
faceCellIDs
faceNormals
faceVertexIDs
vertexCoords
fipy.meshes.Grid3D(dx=1.0, dy=1.0, dz=1.0, nx=None, ny=None, nz=None, Lx=None, Ly=None,

Lz=None, overlap=2, communicator=DummyComm())
Parameters
• dz: grid spacing in the z-direction
• nz: number of cells in the z-direction
• Lz: the domain length in the z-direction
test purposes.
fipy.meshes.Grid2D(dx=1.0, dy=1.0, nx=None, ny=None, Lx=None, Ly=None, overlap=2, communi-
cator=DummyComm())
Parameters

test purposes.
>>> print Grid2D(Lx=3., nx=2).dx
1.5
fipy.meshes.Grid1D(dx=1.0, nx=None, Lx=None, overlap=2, communicator=DummyComm())

Parameters
• dx: grid spacing in the horizonal direction
• nx: number of cells in the horizonal direction
• Lx: the domain length in the horizonal direction
test purposes.
fipy.meshes.CylindricalGrid2D(dr=None, dz=None, nr=None, nz=None, Lr=None, Lz=None,
dx=1.0, dy=1.0, nx=None, ny=None, Lx=None, Ly=None, ori-
gin=((0, ), (0, )), overlap=2, communicator=DummyComm())
Parameters
• dz or dy: grid spacing in the vertical direction
• nz or ny: number of cells in the vertical direction
• Lz or Ly: the domain length in the vertical direction
• origin : position of the mesh’s origin in the form ((x,),(y,))
test purposes.
fipy.meshes.CylindricalGrid1D(dr=None, nr=None, Lr=None, dx=1.0, nx=None, Lx=None, ori-
gin=(0, ), overlap=2, communicator=DummyComm())
Parameters


• origin : position of the mesh’s origin in the form (x,)
test purposes.
class fipy.meshes.PeriodicGrid1D(dx=1.0, nx=None, overlap=2, *args, **kwargs)
Creates a Periodic grid mesh.
>>> mesh = PeriodicGrid1D(dx = (1, 2, 3))

... [3])
True

... [[2, 0, 1, 2],
... [0, 1, 2, -999]])
True

... [ 2., 1.5, 2.5, 1.5])
True

... [[ 2., 1.5, 2.5],
... [ 1.5, 2.5, 2. ]])
True
>>> print numerix.allclose(mesh.faceNormals,

... [[ 1., 1., 1., 1.]])
True

... [[1, 2, 2],
... [0, 1, 0]])
True
cellCenters
Defined outside of a geometry class since we need the CellVariable version of cellCenters; that is, the
cellCenters defined in fipy.meshes.mesh and not in any geometry (since a CellVariable requires a reference
to a mesh).
class fipy.meshes.PeriodicGrid2D(dx=1.0, dy=1.0, nx=None, ny=None, overlap=2, communica-
tor=DummyComm(), *args, **kwargs)
Creates a periodic2D grid mesh with horizontal faces numbered first and then vertical faces. Vertices and cells
are numbered in the usual way.

>>> mesh = PeriodicGrid2D(dx = 1., dy = 0.5, nx = 2, ny = 2)

... [ 4, 5, 8, 11])
True

... [[2, 3, 0, 1, 2, 3, 1, 0, 1, 3, 2, 3],
... [0, 1, 2, 3, -1, -1, 0, 1, -1, 2, 3, -1]])
True

... [ 0.5, 0.5, 0.5, 0.5, 0.25, 0.25, 1., 1., 0.5, 1., 1., 0.5])
True
>>> print numerix.allclose(mesh.cellFaceIDs,

... [[0, 1, 2, 3],
... [7, 6, 10, 9],
... [2, 3, 0, 1],
... [6, 7, 9, 10]])
True

... [[ 0.5, 0.5, 0.5, 0.5],
... [ 1., 1., 1., 1. ],
... [ 0.5, 0.5, 0.5, 0.5],
... [ 1., 1., 1., 1. ]])
True
>>> normals = [[0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],

... [1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0]]
>>> print numerix.allclose(mesh.faceNormals, normals)

True

... [[4, 5, 7, 8],
... [3, 4, 6, 7],
... [1, 2, 4, 5],
... [0, 1, 3, 4]])
True
class fipy.meshes.PeriodicGrid2DLeftRight(dx=1.0, dy=1.0, nx=None, ny=None, overlap=2,

communicator=DummyComm(), *args, **kwargs)
class fipy.meshes.PeriodicGrid2DTopBottom(dx=1.0, dy=1.0, nx=None, ny=None, overlap=2,
communicator=DummyComm(), *args, **kwargs)
class fipy.meshes.SkewedGrid2D(dx=1.0, dy=1.0, nx=None, ny=1, rand=0, *args, **kwargs)
Creates a 2D grid mesh with horizontal faces numbered first and then vertical faces. The points are skewed by
a random amount (between rand and -rand) in the X and Y directions.
physicalShape
shape

class fipy.meshes.Tri2D(dx=1.0, dy=1.0, nx=1, ny=1, _RepresentationClass=<class

‘fipy.meshes.representations.gridRepresentation._Grid2DRepresentation’>,
This class creates a mesh made out of triangles. It does this by starting with a standard Cartesian mesh (Grid2D)
and dividing each cell in that mesh (hereafter referred to as a ‘box’) into four equal parts with the dividing lines
being the diagonals.
Creates a 2D triangular mesh with horizontal faces numbered first then vertical faces, then diagonal faces.
Vertices are numbered starting with the vertices at the corners of boxes and then the vertices at the centers of
boxes. Cells on the right of boxes are numbered first, then cells on the top of boxes, then cells on the left of
boxes, then cells on the bottom of boxes. Within each of the ‘sub-categories’ in the above, the vertices, cells and
faces are numbered in the usual way.
Parameters
• dx, dy: The X and Y dimensions of each ‘box’. If dx <> dy, the line segments connecting
the cell centers will not be orthogonal to the faces.
• nx, ny: The number of boxes in the X direction and the Y direction. The total number of
boxes will be equal to nx * ny, and the total number of cells will be equal to 4 * nx * ny.
physicalShape
shape
fipy.meshes.openMSHFile(name, dimensions=None, coordDimensions=None, communica-
tor=DummyComm(), order=1, mode=’r’, background=None)
Open a Gmsh MSH file
Parameters
• filename: a string indicating gmsh output file
• dimensions: an integer indicating dimension of mesh
• order: ???
• mode: a string beginning with ‘r’ for reading and ‘w’ for writing. The file will be created if
it doesn’t exist when opened for writing; it will be truncated when opened for writing. Add
a ‘b’ to the mode for binary files.
fipy.meshes.openPOSFile(name, communicator=DummyComm(), mode=’w’)
Open a Gmsh POS post-processing file
class fipy.meshes.Gmsh2D(arg, coordDimensions=2, communicator=DummyComm(), order=1, back-
ground=None)
Construct a 2D Mesh using Gmsh
>>> radius = 5.
>>> side = 4.
>>> squaredCircle = Gmsh2D('''
... // A mesh consisting of a square inside a circle inside a circle
...
... // define the basic dimensions of the mesh
...

... cellSize = 1;
... side = %(side)g;
...
... // define the compass points of the inner circle
...
... Point(1) = {0, 0, 0, cellSize};
...
... // define the compass points of the outer circle
...
... Point(6) = {-2*radius, 0, 0, cellSize};
... Point(7) = {0, 2*radius, 0, cellSize};
... Point(8) = {2*radius, 0, 0, cellSize};
... Point(9) = {0, -2*radius, 0, cellSize};
...
...
... Point(10) = {side/2, side/2, 0, cellSize/2};
... Point(11) = {-side/2, side/2, 0, cellSize/2};
... Point(12) = {-side/2, -side/2, 0, cellSize/2};
... Point(13) = {side/2, -side/2, 0, cellSize/2};
...
... // define the inner circle
...
... Circle(1) = {2, 1, 3};
... Circle(2) = {3, 1, 4};
... Circle(3) = {4, 1, 5};
... Circle(4) = {5, 1, 2};
...
... // define the outer circle
...
... Circle(5) = {6, 1, 7};
... Circle(6) = {7, 1, 8};
... Circle(7) = {8, 1, 9};
... Circle(8) = {9, 1, 6};
...
...
... Line(9) = {10, 13};
... Line(10) = {13, 12};
... Line(11) = {12, 11};
... Line(12) = {11, 10};
...
... // define the three boundaries
...
... Line Loop(1) = {1, 2, 3, 4};
... Line Loop(2) = {5, 6, 7, 8};
... Line Loop(3) = {9, 10, 11, 12};
...
... // define the three domains
...

...
... // label the three domains
...
... // attention: if you use any "Physical" labels, you *must* label
... // all elements that correspond to FiPy Cells (Physical Surace in 2D
... // and Physical Volume in 3D) or Gmsh will not include them and FiPy
... // will not be able to include them in the Mesh.
...
... // note: if you do not use any labels, all Cells will be included.
...
... Physical Surface("Outer") = {1};
... Physical Surface("Middle") = {2};
... Physical Surface("Inner") = {3};
...
... // label the "north-west" part of the exterior boundary
...
... // note: you only need to label the Face elements
... // (Physical Line in 2D and Physical Surface in 3D) that correspond
... // to boundaries you are interested in. FiPy does not need them to
... // construct the Mesh.
...
... Physical Line("NW") = {5};
... ''' % locals())
It can be easier to specify certain domains and boundaries within Gmsh than it is to define the same domains
and boundaries with FiPy expressions.
Here we compare obtaining the same Cells and Faces using FiPy’s parametric descriptions and Gmsh’s labels.
>>> x, y = squaredCircle.cellCenters
>>> middle = ((x**2 + y**2 <= radius**2)

... & ~((x > -side/2) & (x < side/2)
... & (y > -side/2) & (y < side/2)))
>>> print (middle == squaredCircle.physicalCells["Middle"]).all()

True
>>> X, Y = squaredCircle.faceCenters
>>> NW = ((X**2 + Y**2 > (1.99*radius)**2)

... & (X**2 + Y**2 < (2.01*radius)**2)
... & (X <= 0) & (Y >= 0))
>>> print (NW == squaredCircle.physicalFaces["NW"]).all()

True
It is possible to direct Gmsh to give the mesh different densities in different locations
>>> geo = '''
... // A mesh consisting of a square
...
...
... Point(1) = {1, 1, 0, 1};
... Point(2) = {0, 1, 0, 1};
... Point(3) = {0, 0, 0, 1};
... Point(4) = {1, 0, 0, 1};
...

...
... Line(1) = {1, 2};
... Line(2) = {2, 3};
... Line(3) = {3, 4};
... Line(4) = {4, 1};
...
... // define the boundary
...
... Line Loop(1) = {1, 2, 3, 4};
...
... // define the domain
...
... '''
>>> from fipy import CellVariable, numerix
>>> std = []
>>> bkg = None
>>> for refine in range(4):
... square = Gmsh2D(geo, background=bkg)
... x, y = square.cellCenters
... bkg = CellVariable(mesh=square, value=abs(x / 4) + 0.01)
... std.append(numerix.std(numerix.sqrt(2 * square.cellVolumes) / bkg))
Check that the mesh is monotonically approaching the desired density

>>> print numerix.greater(std[:-1], std[1:]).all()
True
and that the final density is close enough to the desired density
>>> print std[-1] < 0.2
True
The initial mesh doesn’t have to be from Gmsh

>>> from fipy import Tri2D
>>> trisquare = Tri2D(nx=1, ny=1)

>>> x, y = trisquare.cellCenters
>>> bkg = CellVariable(mesh=trisquare, value=abs(x / 4) + 0.01)
>>> std1 = numerix.std(numerix.sqrt(2 * trisquare.cellVolumes) / bkg)
>>> square = Gmsh2D(geo, background=bkg)

>>> x, y = square.cellCenters
>>> bkg = CellVariable(mesh=square, value=abs(x / 4) + 0.01)
>>> std2 = numerix.std(numerix.sqrt(2 * square.cellVolumes) / bkg)
>>> print std1 > std2

True
Parameters
• order: ???

class fipy.meshes.Gmsh2DIn3DSpace(arg, communicator=DummyComm(), order=1, back-

ground=None)
Create a topologically 2D Mesh in 3D coordinates using Gmsh
Parameters
• order: ???
class fipy.meshes.Gmsh3D(arg, communicator=DummyComm(), order=1, background=None)
Create a 3D Mesh using Gmsh
Parameters
• order: ???
class fipy.meshes.GmshGrid2D(dx=1.0, dy=1.0, nx=1, ny=None, coordDimensions=2, communica-
tor=DummyComm(), order=1)
class fipy.meshes.GmshGrid3D(dx=1.0, dy=1.0, dz=1.0, nx=1, ny=None, nz=None, communica-
tor=DummyComm(), order=1)
class fipy.meshes.GmshImporter2D(arg, coordDimensions=2)
class fipy.meshes.GmshImporter2DIn3DSpace(arg)
Bases: fipy.meshes.gmshMesh.Gmsh2DIn3DSpace
class fipy.meshes.GmshImporter3D(arg)


Chapter 22
fipy.solvers package
22.1 Subpackages
22.1.1 fipy.solvers.pyAMG package
Subpackages
fipy.solvers.pyAMG.preconditioners package
Submodules
fipy.solvers.pyAMG.preconditioners.smoothedAggregationPreconditioner module
Module contents
Submodules
fipy.solvers.pyAMG.linearCGSSolver module
fipy.solvers.pyAMG.linearGMRESSolver module
fipy.solvers.pyAMG.linearGeneralSolver module
fipy.solvers.pyAMG.linearLUSolver module
fipy.solvers.pyAMG.linearPCGSolver module
fipy.solvers.pyAMG.smoothedAggregationSolver module
Module contents
22.1.2 fipy.solvers.pysparse package
Subpackages
fipy.solvers.pysparse.preconditioners package
Submodules
233
fipy.solvers.pysparse.preconditioners.jacobiPreconditioner module
class fipy.solvers.pysparse.preconditioners.jacobiPreconditioner.JacobiPreconditioner
Bases: fipy.solvers.pysparse.preconditioners.preconditioner.Preconditioner
Jacobi preconditioner for PySparse. Really just a wrapper class for pysparse.precon.jacobi.
Create a Preconditioner object.
fipy.solvers.pysparse.preconditioners.preconditioner module
class fipy.solvers.pysparse.preconditioners.preconditioner.Preconditioner
Base preconditioner class
fipy.solvers.pysparse.preconditioners.ssorPreconditioner module
class fipy.solvers.pysparse.preconditioners.ssorPreconditioner.SsorPreconditioner
SSOR preconditioner for PySparse. Really just a wrapper class for pysparse.precon.jacobi.
Module contents
class fipy.solvers.pysparse.preconditioners.JacobiPreconditioner
class fipy.solvers.pysparse.preconditioners.SsorPreconditioner
Submodules
fipy.solvers.pysparse.linearCGSSolver module
class fipy.solvers.pysparse.linearCGSSolver.LinearCGSSolver(precon=None, *args,

**kwargs)
Bases: fipy.solvers.pysparse.pysparseSolver.PysparseSolver
The LinearCGSSolver solves a linear system of equations using the conjugate gradient squared method (CGS),
a variant of the biconjugate gradient method (BiCG). CGS solves linear systems with a general non-symmetric
coefficient matrix.
The LinearCGSSolver is a wrapper class for the the PySparse itsolvers.cgs() method.
Parameters
• precon: Preconditioner to use
234 Chapter 22. fipy.solvers package

fipy.solvers.pysparse.linearGMRESSolver module
class fipy.solvers.pysparse.linearGMRESSolver.LinearGMRESSolver(precon=<fipy.solvers.pysparse.preconditio
instance>, *args,
**kwargs)
The LinearGMRESSolver solves a linear system of equations using the generalised minimal residual method
(GMRES) with Jacobi preconditioning. GMRES solves systems with a general non-symmetric coefficient ma-
trix.
The LinearGMRESSolver is a wrapper class for the the PySparse itsolvers.gmres() and precon.jacobi() methods.
Parameters
fipy.solvers.pysparse.linearJORSolver module
class fipy.solvers.pysparse.linearJORSolver.LinearJORSolver(tolerance=1e-10,
iterations=1000, relax-
ation=1.0)
The LinearJORSolver solves a linear system of equations using Jacobi over-relaxation. This method solves
systems with a general non-symmetric coefficient matrix.
The Solver class should not be invoked directly.
Parameters
• tolerance: The required error tolerance.
• iterations: The maximum number of iterative steps to perform.
• relaxation: The relaxation.
fipy.solvers.pysparse.linearLUSolver module
class fipy.solvers.pysparse.linearLUSolver.LinearLUSolver(tolerance=1e-10, iter-

ations=10, maxItera-
tions=10, precon=None)
The LinearLUSolver solves a linear system of equations using LU-factorisation. This method solves systems
with a general non-symmetric coefficient matrix using partial pivoting.
The LinearLUSolver is a wrapper class for the the PySparse superlu.factorize() method.
Creates a LinearLUSolver.
Parameters
• iterations: The number of LU decompositions to perform. For large systems a number of
iterations is generally required.
• precon: not used but maintains a common interface.
22.1. Subpackages 235

fipy.solvers.pysparse.linearPCGSolver module
class fipy.solvers.pysparse.linearPCGSolver.LinearPCGSolver(precon=<fipy.solvers.pysparse.preconditioners.ss
instance>, *args,
**kwargs)
The LinearPCGSolver solves a linear system of equations using the preconditioned conjugate gradient method
(PCG) with symmetric successive over-relaxation (SSOR) preconditioning by default. Alternatively, Jacobi
preconditioning can be specified through precon. The PCG method solves systems with a symmetric positive
definite coefficient matrix.
The LinearPCGSolver is a wrapper class for the the PySparse itsolvers.pcg() and precon.ssor() methods.
Parameters
fipy.solvers.pysparse.pysparseSolver module
class fipy.solvers.pysparse.pysparseSolver.PysparseSolver(*args, **kwargs)

Bases: fipy.solvers.pysparseMatrixSolver._PysparseMatrixSolver
The base pysparseSolver class.
Module contents
fipy.solvers.pysparse.DefaultSolver
alias of LinearPCGSolver
fipy.solvers.pysparse.DummySolver
fipy.solvers.pysparse.DefaultAsymmetricSolver
alias of LinearLUSolver
fipy.solvers.pysparse.GeneralSolver
class fipy.solvers.pysparse.LinearCGSSolver(precon=None, *args, **kwargs)
coefficient matrix.
Parameters
class fipy.solvers.pysparse.LinearPCGSolver(precon=<fipy.solvers.pysparse.preconditioners.ssorPreconditioner.SsorP
instance>, *args, **kwargs)

Parameters
class fipy.solvers.pysparse.LinearGMRESSolver(precon=<fipy.solvers.pysparse.preconditioners.jacobiPreconditioner.J
trix.
Parameters
class fipy.solvers.pysparse.LinearLUSolver(tolerance=1e-10, iterations=10, maxItera-
Parameters
class fipy.solvers.pysparse.LinearJORSolver(tolerance=1e-10, iterations=1000, relax-
ation=1.0)
Parameters
class fipy.solvers.pysparse.JacobiPreconditioner

class fipy.solvers.pysparse.SsorPreconditioner
22.1.3 fipy.solvers.scipy package
Submodules
fipy.solvers.scipy.linearBicgstabSolver module
class fipy.solvers.scipy.linearBicgstabSolver.LinearBicgstabSolver
Bases: fipy.solvers.scipy.scipyKrylovSolver._ScipyKrylovSolver
The LinearBicgstabSolver is an interface to the Bicgstab solver in Scipy, with no preconditioning by default.
Parameters
• precon: Preconditioner to use.
fipy.solvers.scipy.linearCGSSolver module
class fipy.solvers.scipy.linearCGSSolver.LinearCGSSolver(tolerance=1e-15, itera-

The LinearCGSSolver is an interface to the CGS solver in Scipy, with no preconditioning by default.
Parameters
fipy.solvers.scipy.linearGMRESSolver module
class fipy.solvers.scipy.linearGMRESSolver.LinearGMRESSolver(tolerance=1e-15,
iterations=2000,
precon=None)
The LinearGMRESSolver is an interface to the GMRES solver in Scipy, with no preconditioning by default.
Parameters

fipy.solvers.scipy.linearLUSolver module
class fipy.solvers.scipy.linearLUSolver.LinearLUSolver(tolerance=1e-10, itera-

Bases: fipy.solvers.scipy.scipySolver._ScipySolver
The LinearLUSolver solves a linear system of equations using LU-factorisation. The LinearLUSolver is a
wrapper class for the the Scipy scipy.sparse.linalg.splu moduleq.
Create a Solver object.
Parameters
• precon: Preconditioner to use. This parameter is only available for Trilinos solvers.
fipy.solvers.scipy.linearPCGSolver module
class fipy.solvers.scipy.linearPCGSolver.LinearPCGSolver(tolerance=1e-15, itera-

The LinearPCGSolver is an interface to the CG solver in Scipy, with no preconditioning by default.
Parameters
fipy.solvers.scipy.scipyKrylovSolver module
fipy.solvers.scipy.scipySolver module
Module contents
fipy.solvers.scipy.DefaultSolver
fipy.solvers.scipy.DummySolver
alias of LinearGMRESSolver
fipy.solvers.scipy.DefaultAsymmetricSolver
fipy.solvers.scipy.GeneralSolver
class fipy.solvers.scipy.LinearCGSSolver(tolerance=1e-15, iterations=2000, precon=None)
The LinearCGSSolver is an interface to the CGS solver in Scipy, with no preconditioning by default.
Parameters


class fipy.solvers.scipy.LinearGMRESSolver(tolerance=1e-15, iterations=2000, pre-
con=None)
The LinearGMRESSolver is an interface to the GMRES solver in Scipy, with no preconditioning by default.
Parameters
class fipy.solvers.scipy.LinearBicgstabSolver
The LinearBicgstabSolver is an interface to the Bicgstab solver in Scipy, with no preconditioning by default.
Parameters
class fipy.solvers.scipy.LinearLUSolver(tolerance=1e-10, iterations=1000, precon=None)
Bases: fipy.solvers.scipy.scipySolver._ScipySolver
The LinearLUSolver solves a linear system of equations using LU-factorisation. The LinearLUSolver is a
wrapper class for the the Scipy scipy.sparse.linalg.splu moduleq.
Parameters
class fipy.solvers.scipy.LinearPCGSolver(tolerance=1e-15, iterations=2000, precon=None)
The LinearPCGSolver is an interface to the CG solver in Scipy, with no preconditioning by default.
Parameters
22.1.4 fipy.solvers.trilinos package
Subpackages
fipy.solvers.trilinos.preconditioners package

Submodules
fipy.solvers.trilinos.preconditioners.domDecompPreconditioner module
fipy.solvers.trilinos.preconditioners.icPreconditioner module
fipy.solvers.trilinos.preconditioners.jacobiPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelDDMLPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelDDPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelNSSAPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelSAPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelSGSPreconditioner module
fipy.solvers.trilinos.preconditioners.multilevelSolverSmootherPreconditioner module
fipy.solvers.trilinos.preconditioners.preconditioner module
Module contents
Submodules
fipy.solvers.trilinos.linearBicgstabSolver module
fipy.solvers.trilinos.linearCGSSolver module
fipy.solvers.trilinos.linearGMRESSolver module
fipy.solvers.trilinos.linearLUSolver module
fipy.solvers.trilinos.linearPCGSolver module
fipy.solvers.trilinos.trilinosAztecOOSolver module
fipy.solvers.trilinos.trilinosMLTest module
fipy.solvers.trilinos.trilinosNonlinearSolver module
fipy.solvers.trilinos.trilinosSolver module

Module contents
22.2 Submodules
22.3 fipy.solvers.pysparseMatrixSolver module
22.4 fipy.solvers.solver module
The iterative solvers may output warnings if the solution is considered unsatisfactory. If you are not interested in these
warnings, you can invoke python with a warning filter such as:
$ python -Wignore::fipy.SolverConvergenceWarning myscript.py
If you are extremely concerned about your preconditioner for some reason, you can abort whenever it has problems
with:
$ python -Werror::fipy.PreconditionerWarning myscript.py
exception fipy.solvers.solver.SolverConvergenceWarning(solver, iter, relres)

Bases: exceptions.Warning
exception fipy.solvers.solver.MaximumIterationWarning(solver, iter, relres)
Bases: fipy.solvers.solver.SolverConvergenceWarning
exception fipy.solvers.solver.PreconditionerWarning(solver, iter, relres)
exception fipy.solvers.solver.IllConditionedPreconditionerWarning(solver, iter, rel-
res)
Bases: fipy.solvers.solver.PreconditionerWarning
exception fipy.solvers.solver.PreconditionerNotPositiveDefiniteWarning(solver,
iter,
relres)
exception fipy.solvers.solver.MatrixIllConditionedWarning(solver, iter, relres)
exception fipy.solvers.solver.StagnatedSolverWarning(solver, iter, relres)
exception fipy.solvers.solver.ScalarQuantityOutOfRangeWarning(solver, iter, relres)
class fipy.solvers.solver.Solver(tolerance=1e-10, iterations=1000, precon=None)
Bases: object
The base LinearXSolver class.

Parameters

22.5 fipy.solvers.test module
exception fipy.solvers.SolverConvergenceWarning(solver, iter, relres)

Bases: exceptions.Warning
exception fipy.solvers.MaximumIterationWarning(solver, iter, relres)
exception fipy.solvers.PreconditionerWarning(solver, iter, relres)
exception fipy.solvers.IllConditionedPreconditionerWarning(solver, iter, relres)
exception fipy.solvers.PreconditionerNotPositiveDefiniteWarning(solver, iter, relres)
exception fipy.solvers.MatrixIllConditionedWarning(solver, iter, relres)
exception fipy.solvers.StagnatedSolverWarning(solver, iter, relres)
exception fipy.solvers.ScalarQuantityOutOfRangeWarning(solver, iter, relres)
class fipy.solvers.Solver(tolerance=1e-10, iterations=1000, precon=None)
Bases: object
The base LinearXSolver class.

Parameters
fipy.solvers.DefaultSolver
fipy.solvers.DummySolver
fipy.solvers.DefaultAsymmetricSolver
fipy.solvers.GeneralSolver
22.5. fipy.solvers.test module 243

class fipy.solvers.LinearCGSSolver(precon=None, *args, **kwargs)

coefficient matrix.
Parameters
class fipy.solvers.LinearPCGSolver(precon=<fipy.solvers.pysparse.preconditioners.ssorPreconditioner.SsorPreconditioner
Parameters
class fipy.solvers.LinearGMRESSolver(precon=<fipy.solvers.pysparse.preconditioners.jacobiPreconditioner.JacobiPrecond
trix.
Parameters
class fipy.solvers.LinearLUSolver(tolerance=1e-10, iterations=10, maxIterations=10, pre-
con=None)
Parameters
class fipy.solvers.LinearJORSolver(tolerance=1e-10, iterations=1000, relaxation=1.0)


Parameters
class fipy.solvers.JacobiPreconditioner
class fipy.solvers.SsorPreconditioner


Chapter 23
fipy.steppers package
23.1 Submodules
23.2 fipy.steppers.pidStepper module
class fipy.steppers.pidStepper.PIDStepper(vardata=(), proportional=0.075, integral=0.175,

derivative=0.01)
Bases: fipy.steppers.stepper.Stepper
Adaptive stepper using a PID controller, based on:
@article{PIDpaper,
author = {A. M. P. Valli and G. F. Carey and A. L. G. A. Coutinho},
title = {Control strategies for timestep selection in finite element
simulation of incompressible flows and coupled
reaction-convection-diffusion processes},
journal = {Int. J. Numer. Meth. Fluids},
volume = 47,
year = 2005,
pages = {201-231},
}
23.3 fipy.steppers.pseudoRKQSStepper module
class fipy.steppers.pseudoRKQSStepper.PseudoRKQSStepper(vardata=(), safety=0.9,

pgrow=-0.2, pshrink=-0.25,
errcon=0.000189)
Bases: fipy.steppers.stepper.Stepper
Adaptive stepper based on the rkqs (Runge-Kutta “quality-controlled” stepper) algorithm of numerixal Recipes
in C: 2nd Edition, Section 16.2.
Not really appropriate, since we’re not doing Runge-Kutta steps in the first place, but works OK.
23.4 fipy.steppers.stepper module
class fipy.steppers.stepper.Stepper(vardata=())
static failFn(vardata, dt, *args, **kwargs)
247
step(dt, dtTry=None, dtMin=None, dtPrev=None, sweepFn=None, successFn=None, failFn=None,

*args, **kwargs)
static successFn(vardata, dt, dtPrev, elapsed, *args, **kwargs)
static sweepFn(vardata, dt, *args, **kwargs)
fipy.steppers.L1error(var, matrix, RHSvector)

Parameters
• var: The CellVariable in question.
• matrix: (ignored)
• RHSvector: (ignored)
Returns
‖var − varold ‖1
‖varold ‖1
where ‖⃗𝑥‖1 is the 𝐿1 -norm of ⃗𝑥.

fipy.steppers.L2error(var, matrix, RHSvector)
Parameters
Returns
‖var − varold ‖2
‖varold ‖2
where ‖⃗𝑥‖2 is the 𝐿2 -norm of ⃗𝑥.

fipy.steppers.LINFerror(var, matrix, RHSvector)
Parameters
Returns
‖var − varold ‖∞
‖varold ‖∞
where ‖⃗𝑥‖∞ is the 𝐿∞ -norm of ⃗𝑥.

fipy.steppers.sweepMonotonic(fn, *args, **kwargs)
Repeatedly calls fn(*args, **kwargs)() until the residual returned by fn() is no longer decreasing.
Parameters
248 Chapter 23. fipy.steppers package

• fn: The function to call

• args: The unnamed function argument list
• kwargs: The named function argument dict
Returns the final residual

250 Chapter 23. fipy.steppers package

Chapter 24
fipy.terms package
24.1 Submodules
24.2 fipy.terms.abstractBinaryTerm module
24.3 fipy.terms.abstractConvectionTerm module
24.4 fipy.terms.abstractDiffusionTerm module
24.5 fipy.terms.abstractUpwindConvectionTerm module
24.6 fipy.terms.advectionTerm module
class fipy.terms.advectionTerm.AdvectionTerm(coeff=None)
Bases: fipy.terms.firstOrderAdvectionTerm.FirstOrderAdvectionTerm
The AdvectionTerm object constructs the b vector contribution for the advection term given by
𝑢|∇𝜑|
from the advection equation given by:

𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡
The construction of the gradient magnitude term requires upwinding as in the standard FirstOrderAdvection-
Term. The higher order terms are incorperated as follows. The formula used here is given by:
[︃ ]︃1/2 [︃ ]︃1/2
∑︁ 2
∑︁ 2
𝑢𝑃 |∇𝜑|𝑃 = max (𝑢𝑃 , 0) min (𝐷𝐴𝑃 , 0) + min (𝑢𝑃 , 0) max (𝐷𝐴𝑃 , 0)
𝐴 𝐴
where,
𝜑𝐴 − 𝜑𝑃 𝑑𝐴𝑃
𝐷𝐴𝑃 = − 𝑚 (𝐿𝐴 , 𝐿𝑃 )
𝑑𝐴𝑃 2
and
𝑚 (𝑥, 𝑦) = 𝑥 if |𝑥| ≤ |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 𝑦 if |𝑥| > |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 0 if 𝑥𝑦 < 0
251
also,
𝜑𝐴𝐴 + 𝜑𝑃 − 2𝜑𝐴
𝐿𝐴 =
𝑑2𝐴𝑃
𝜑𝐴 + 𝜑𝑃 𝑃 − 2𝜑𝑃
𝐿𝑃 =
𝑑2𝐴𝑃
Here are some simple test cases for this problem:

>>> from fipy.meshes import Grid1D
>>> SparseMatrix = LinearPCGSolver()._matrixClass
>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
>>> coeff = CellVariable(mesh = mesh, value = numerix.zeros(3, 'd'))
>>> v, L, b = AdvectionTerm(0.)._buildMatrix(coeff, SparseMatrix)
>>> print numerix.allclose(b, numerix.zeros(3, 'd'), atol = 1e-10)
True
Less trivial test:

>>> coeff = CellVariable(mesh = mesh, value = numerix.arange(3))
>>> print numerix.allclose(b, numerix.array((0., -1., -1.)), atol = 1e-10)
True
Even less trivial

>>> v, L, b = AdvectionTerm(-1.)._buildMatrix(coeff, SparseMatrix)
>>> print numerix.allclose(b, numerix.array((1., 1., 0.)), atol = 1e-10)
True
Another trivial test case (more trivial than a trivial test case standing on a harpsichord singing ‘trivial test cases
are here again’)
>>> vel = numerix.array((-1, 2, -3))
>>> coeff = CellVariable(mesh = mesh, value = numerix.array((4,6,1)))
>>> v, L, b = AdvectionTerm(vel)._buildMatrix(coeff, SparseMatrix)
>>> print numerix.allclose(b, -vel * numerix.array((2, numerix.sqrt(5**2 + 2**2), 5)), atol = 1e
True
Somewhat less trivial test case:

>>> mesh = Grid2D(dx = 1., dy = 1., nx = 2, ny = 2)
>>> vel = numerix.array((3, -5, -6, -3))
>>> coeff = CellVariable(mesh = mesh, value = numerix.array((3 , 1, 6, 7)))
>>> answer = -vel * numerix.array((2, numerix.sqrt(2**2 + 6**2), 1, 0))
>>> print numerix.allclose(b, answer, atol = 1e-10)
True
For the above test cases the AdvectionTerm gives the same result as the AdvectionTerm. The following test
imposes a quadratic field. The higher order term can resolve this field correctly.
𝜑 = 𝑥2
252 Chapter 24. fipy.terms package

The returned vector b should have the value:

⃒ ⃒
⃒ 𝜕𝜑 ⃒
−|∇𝜑| = − ⃒⃒ ⃒⃒ = −2|𝑥|
𝜕𝑥
Build the test case in the following way,

>>> mesh = Grid1D(dx = 1., nx = 5)
>>> vel = 1.
>>> coeff = CellVariable(mesh = mesh, value = mesh.cellCenters[0]**2)
>>> v, L, b = __AdvectionTerm(vel)._buildMatrix(coeff, SparseMatrix)
The first order term is not accurate. The first and last element are ignored because they don’t have any neighbors
for higher order evaluation
>>> print numerix.allclose(CellVariable(mesh=mesh,
... value=b).globalValue[1:-1], -2 * mesh.cellCenters.globalValue[0][1:-1])
False
The higher order term is spot on.

True
The AdvectionTerm will also resolve a circular field with more accuracy,
)︀1/2
𝜑 = 𝑥2 + 𝑦 2
(︀

>>> vel = 1.
>>> r = numerix.sqrt(x**2 + y**2)
>>> coeff = CellVariable(mesh = mesh, value = r)
>>> v, L, b = __AdvectionTerm(1.)._buildMatrix(coeff, SparseMatrix)
>>> error = CellVariable(mesh=mesh, value=b + 1)
>>> ans = CellVariable(mesh=mesh, value=b + 1)
>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.123105625618
>>> print (error <= ans).all()
True
The maximum error is large (about 12 %) for the first order advection.
>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.0201715476598
True
The maximum error is 2 % when using a higher order contribution.
24.6. fipy.terms.advectionTerm module 253

24.7 fipy.terms.asymmetricConvectionTerm module
24.8 fipy.terms.binaryTerm module
24.9 fipy.terms.cellTerm module
class fipy.terms.cellTerm.CellTerm(coeff=1.0, var=None)

Bases: fipy.terms.nonDiffusionTerm._NonDiffusionTerm
24.10 fipy.terms.centralDiffConvectionTerm module
class fipy.terms.centralDiffConvectionTerm.CentralDifferenceConvectionTerm(coeff=1.0,
var=None)
Bases: fipy.terms.abstractConvectionTerm._AbstractConvectionTerm
This Term represents
∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the central differencing scheme. For further details
see Numerical Schemes.
Create a _AbstractConvectionTerm object.
>>> m = Grid1D(nx = 2)
>>> cv = CellVariable(mesh = m)
>>> fv = FaceVariable(mesh = m)
>>> vcv = CellVariable(mesh=m, rank=1)
>>> vfv = FaceVariable(mesh=m, rank=1)
>>> __ConvectionTerm(coeff = cv)
...
VectorCoeffError: The coefficient must be a vector value.
>>> __ConvectionTerm(coeff = fv)
...
>>> __ConvectionTerm(coeff = vcv)
__ConvectionTerm(coeff=_ArithmeticCellToFaceVariable(value=array([[ 0., 0., 0.]]), mesh=Unifor
>>> __ConvectionTerm(coeff = vfv)
__ConvectionTerm(coeff=FaceVariable(value=array([[ 0., 0., 0.]]), mesh=UniformGrid1D(dx=1.0, n
>>> __ConvectionTerm(coeff = (1,))
__ConvectionTerm(coeff=(1,))
>>> ExplicitUpwindConvectionTerm(coeff = (0,)).solve(var=cv, solver=DummySolver())
...
TransientTermError: The equation requires a TransientTerm with explicit convection.
>>> (TransientTerm(0.) - ExplicitUpwindConvectionTerm(coeff = (0,))).solve(var=cv, solver=DummyS

>>> (TransientTerm() - ExplicitUpwindConvectionTerm(coeff = 1)).solve(var=cv, solver=DummySolver

...
>>> m2 = Grid2D(nx=2, ny=1)
>>> cv2 = CellVariable(mesh=m2)
>>> vcv2 = CellVariable(mesh=m2, rank=1)
>>> vfv2 = FaceVariable(mesh=m2, rank=1)
>>> __ConvectionTerm(coeff=vcv2)
__ConvectionTerm(coeff=_ArithmeticCellToFaceVariable(value=array([[ 0., 0., 0., 0., 0., 0.,
[ 0., 0., 0., 0., 0., 0., 0.]]), mesh=UniformGrid2D(dx=1.0, nx=2, dy=1.0, ny=1)))
>>> __ConvectionTerm(coeff=vfv2)
__ConvectionTerm(coeff=FaceVariable(value=array([[ 0., 0., 0., 0., 0., 0., 0.],
>>> (TransientTerm() - ExplicitUpwindConvectionTerm(coeff = ((0,),(0,)))).solve(var=cv2, solver=
>>> (TransientTerm() - ExplicitUpwindConvectionTerm(coeff = (0,0))).solve(var=cv2, solver=DummyS
Parameters
• coeff : The Term‘s coefficient value.
24.11 fipy.terms.coupledBinaryTerm module
24.12 fipy.terms.diffusionTerm module
class fipy.terms.diffusionTerm.DiffusionTerm(coeff=(1.0, ), var=None)

Bases: fipy.terms.diffusionTermNoCorrection.DiffusionTermNoCorrection
This term represents a higher order diffusion term. The order of the term is determined by the number of coeffs,
such that:
DiffusionTerm(D1)
represents a typical 2nd-order diffusion term of the form
∇ · (𝐷1 ∇𝜑)
and:
DiffusionTerm((D1,D2))
represents a 4th-order Cahn-Hilliard term of the form
∇ · {𝐷1 ∇ [∇ · (𝐷2 ∇𝜑)]}
and so on.
class fipy.terms.diffusionTerm.DiffusionTermCorrection(coeff=(1.0, ), var=None)
Bases: fipy.terms.abstractDiffusionTerm._AbstractDiffusionTerm
class fipy.terms.diffusionTerm.DiffusionTermNoCorrection(coeff=(1.0, ), var=None)
24.11. fipy.terms.coupledBinaryTerm module 255

24.13 fipy.terms.diffusionTermCorrection module
class fipy.terms.diffusionTermCorrection.DiffusionTermCorrection(coeff=(1.0, ),
var=None)
24.14 fipy.terms.diffusionTermNoCorrection module
class fipy.terms.diffusionTermNoCorrection.DiffusionTermNoCorrection(coeff=(1.0,
),
var=None)
24.15 fipy.terms.explicitDiffusionTerm module
class fipy.terms.explicitDiffusionTerm.ExplicitDiffusionTerm(coeff=(1.0, ),
var=None)
The discretization for the ExplicitDiffusionTerm is given by
∫︁ ∑︁ 𝜑old − 𝜑old
∇ · (Γ∇𝜑)𝑑𝑉 ≃ Γ𝑓 𝐴 𝑃
𝐴𝑓
𝑉 𝑑𝐴𝑃
𝑓
where 𝜑old old

𝐴 and 𝜑𝑃 are the old values of the variable. The term is added to the RHS vector and makes no
contribution to the solution matrix.
24.16 fipy.terms.explicitSourceTerm module
24.17 fipy.terms.explicitUpwindConvectionTerm module
class fipy.terms.explicitUpwindConvectionTerm.ExplicitUpwindConvectionTerm(coeff=1.0,
var=None)
Bases: fipy.terms.abstractUpwindConvectionTerm._AbstractUpwindConvectionTerm
The discretization for this Term is given by
∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑old old

𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the upwind scheme. For further details see
Numerical Schemes.
>>> m = Grid1D(nx = 2)


...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
24.18 fipy.terms.exponentialConvectionTerm module
class fipy.terms.exponentialConvectionTerm.ExponentialConvectionTerm(coeff=1.0,
var=None)
Bases: fipy.terms.asymmetricConvectionTerm._AsymmetricConvectionTerm
∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the exponential scheme. For further details see
Numerical Schemes.
24.18. fipy.terms.exponentialConvectionTerm module 257


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
24.19 fipy.terms.faceTerm module
class fipy.terms.faceTerm.FaceTerm(coeff=1.0, var=None)


24.20 fipy.terms.firstOrderAdvectionTerm module
class fipy.terms.firstOrderAdvectionTerm.FirstOrderAdvectionTerm(coeff=None)
The FirstOrderAdvectionTerm object constructs the b vector contribution for the advection term given by
𝑢|∇𝜑|
𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡
The construction of the gradient magnitude term requires upwinding. The formula used here is given by:
[︃ (︂ )︂2 ]︃1/2 [︃ (︂ )︂2 ]︃1/2
∑︁ 𝜑𝐴 − 𝜑𝑃 ∑︁ 𝜑𝐴 − 𝜑𝑃
𝑢𝑃 |∇𝜑|𝑃 = max (𝑢𝑃 , 0) min ,0 + min (𝑢𝑃 , 0) max ,0
𝑑𝐴𝑃 𝑑𝐴𝑃
𝐴 𝐴

>>> SparseMatrix = LinearLUSolver()._matrixClass
>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
>>> var = CellVariable(value = numerix.zeros(3, 'd'), mesh = mesh)
>>> v, L, b = FirstOrderAdvectionTerm(0.)._buildMatrix(var, SparseMatrix)
True
Less trivial test:

>>> var = CellVariable(value = numerix.arange(3), mesh = mesh)
True
Even less trivial

>>> v, L, b = FirstOrderAdvectionTerm(-1.)._buildMatrix(var, SparseMatrix)
True
are here again’)
>>> var = CellVariable(value = numerix.array((4,6,1)), mesh = mesh)
>>> v, L, b = FirstOrderAdvectionTerm(vel)._buildMatrix(var, SparseMatrix)
True
24.20. fipy.terms.firstOrderAdvectionTerm module 259


>>> var = CellVariable(value = numerix.array((3 , 1, 6, 7)), mesh = mesh)
True
24.21 fipy.terms.hybridConvectionTerm module
class fipy.terms.hybridConvectionTerm.HybridConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 +(1−𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the hybrid scheme. For further details see Numerical
Schemes.
>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)


Parameters
24.22 fipy.terms.implicitDiffusionTerm module
fipy.terms.implicitDiffusionTerm.ImplicitDiffusionTerm
alias of DiffusionTerm
24.23 fipy.terms.implicitSourceTerm module
class fipy.terms.implicitSourceTerm.ImplicitSourceTerm(coeff=0.0, var=None)

Bases: fipy.terms.sourceTerm.SourceTerm
The ImplicitSourceTerm represents
∫︁
𝜑𝑆 𝑑𝑉 ≃ 𝜑𝑃 𝑆𝑃 𝑉𝑃
𝑉
where 𝑆 is the coeff value.
24.24 fipy.terms.nonDiffusionTerm module
24.25 fipy.terms.powerLawConvectionTerm module
class fipy.terms.powerLawConvectionTerm.PowerLawConvectionTerm(coeff=1.0,
var=None)
∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the power law scheme. For further details see
Numerical Schemes.
24.22. fipy.terms.implicitDiffusionTerm module 261


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
24.26 fipy.terms.residualTerm module
class fipy.terms.residualTerm.ResidualTerm(equation, underRelaxation=1.0)

Bases: fipy.terms.explicitSourceTerm._ExplicitSourceTerm
The ResidualTerm is a special form of explicit SourceTerm that adds the residual of one equation to another
equation. Useful for Newton’s method.

24.27 fipy.terms.sourceTerm module
class fipy.terms.sourceTerm.SourceTerm(coeff=0.0, var=None)

Bases: fipy.terms.cellTerm.CellTerm
24.28 fipy.terms.term module
class fipy.terms.term.Term(coeff=1.0, var=None)

Bases: object
Create a Term.
Parameters
• coeff : The coefficient for the term. A CellVariable or number. FaceVariable objects are also
acceptable for diffusion or convection terms.
RHSvector
Return the RHS vector caculated in solve() or sweep(). The cacheRHSvector() method should be called
before solve() or sweep() to cache the vector.
cacheMatrix()
Informs solve() and sweep() to cache their matrix so that getMatrix() can return the matrix.
cacheRHSvector()
Informs solve() and sweep() to cache their right hand side vector so that getRHSvector() can return it.
copy()
getDefaultSolver(var=None, solver=None, *args, **kwargs)
getMatrix(*args, **kwds)
Deprecated since version 3.0: use the matrix property instead
getRHSvector(*args, **kwds)
Deprecated since version 3.0: use the rHSvector property instead
justErrorVector(var=None, solver=None, boundaryConditions=(), dt=1.0, underRelax-
ation=None, residualFn=None)
Builds the Term‘s linear system once. This method also recalculates and returns the error as well as
applying under-relaxation.
Parameters
• var: The variable to be solved for. Provides the initial condition, the old value and holds
the solution on completion.
• solver: The iterative solver to be used to solve the linear system of equations. Defaults to
LinearPCGSolver for Pysparse and LinearLUSolver for Trilinos.
• boundaryConditions: A tuple of boundaryConditions.
• dt: The time step size.
24.27. fipy.terms.sourceTerm module 263

• underRelaxation: Usually a value between 0 and 1 or None in the case of no under-

relaxation
• residualFn: A function that takes var, matrix, and RHSvector arguments used to customize
the residual calculation.
justErrorVector returns the overlapping local value in parallel (not the non-overlapping value).
>>> from fipy.solvers import DummySolver
>>> m = Grid1D(nx=10)
>>> v = CellVariable(mesh=m)
>>> len(DiffusionTerm().justErrorVector(v, solver=DummySolver())) == m.numberOfCells
True
justResidualVector(var=None, solver=None, boundaryConditions=(), dt=None, underRelax-

ation=None, residualFn=None)
Builds the Term‘s linear system once. This method also recalculates and returns the residual as well as
Parameters
relaxation
justResidualVector returns the overlapping local value in parallel (not the non-overlapping value).
>>> m = Grid1D(nx=10)
>>> len(DiffusionTerm().justResidualVector(v)) == m.numberOfCells
True
matrix
Return the matrix caculated in solve() or sweep(). The cacheMatrix() method should be called before
solve() or sweep() to cache the matrix.
residualVectorAndNorm(var=None, solver=None, boundaryConditions=(), dt=None, underRe-
laxation=None, residualFn=None)
Builds the Term‘s linear system once. This method also recalculates and returns the residual as well as
Parameters


relaxation
solve(var=None, solver=None, boundaryConditions=(), dt=None)
Builds and solves the Term‘s linear system once. This method does not return the residual. It should be
used when the residual is not required.
Parameters
sweep(var=None, solver=None, boundaryConditions=(), dt=None, underRelaxation=None, residu-
alFn=None, cacheResidual=False, cacheError=False)
Builds and solves the Term‘s linear system once. This method also recalculates and returns the residual as
well as applying under-relaxation.
Parameters
relaxation
• residualFn: A function that takes var, matrix, and RHSvector arguments, used to cus-
tomize the residual calculation.
• cacheResidual: If True, calculate and store the residual vector ⃗𝑟 = L⃗𝑥 − ⃗𝑏 in the
residualVector member of Term
• cacheError: If True, use the residual vector ⃗𝑟 to solve L⃗𝑒 = ⃗𝑟 for the error vector ⃗𝑒
and store it in the errorVector member of Term
24.29 fipy.terms.test module
24.30 fipy.terms.transientTerm module
class fipy.terms.transientTerm.TransientTerm(coeff=1.0, var=None)

24.29. fipy.terms.test module 265

The TransientTerm represents

∫︁
𝑑𝑉 ≃
where 𝜌 is the coeff value.
The following test case verifies that variable coefficients and old coefficient values work correctly. We will solve
the following equation
𝜕𝜑2
= 𝑘.
𝜕𝑡
The analytic solution is given by
√︁
𝜑= 𝜑20 + 𝑘𝑡,
where 𝜑0 is the initial value.

>>> phi0 = 1.
>>> k = 1.
>>> dt = 1.
>>> relaxationFactor = 1.5
>>> steps = 2
>>> sweeps = 8

>>> mesh = Grid1D(nx = 1)
>>> var = CellVariable(mesh = mesh, value = phi0, hasOld = 1)
>>> from fipy.terms.transientTerm import TransientTerm
>>> from fipy.terms.implicitSourceTerm import ImplicitSourceTerm
Relaxation, given by relaxationFactor, is required for a converged solution.

>>> eq = TransientTerm(var) == ImplicitSourceTerm(-relaxationFactor) \
... + var * relaxationFactor + k
A number of sweeps at each time step are required to let the relaxation take effect.
... var.updateOld()
... eq.solve(var, dt = dt)
Compare the final result with the analytical solution.

>>> print var.allclose(numerix.sqrt(k * dt * steps + phi0**2))
1
24.31 fipy.terms.unaryTerm module
24.32 fipy.terms.upwindConvectionTerm module
class fipy.terms.upwindConvectionTerm.UpwindConvectionTerm(coeff=1.0, var=None)



∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the upwind convection scheme. For further details
>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
24.32. fipy.terms.upwindConvectionTerm module 267

24.33 fipy.terms.vanLeerConvectionTerm module
class fipy.terms.vanLeerConvectionTerm.VanLeerConvectionTerm(coeff=1.0, var=None)

Bases: fipy.terms.explicitUpwindConvectionTerm.ExplicitUpwindConvectionTerm
>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters

exception fipy.terms.ExplicitVariableError(s=’Terms with explicit Variables cannot mix with

Terms with implicit Variables.’)
exception fipy.terms.TermMultiplyError(s=’Must multiply terms by int or float.’)
exception fipy.terms.AbstractBaseClassError(s=”can’t instantiate abstract base class”)
Bases: exceptions.NotImplementedError
exception fipy.terms.VectorCoeffError(s=’The coefficient must be a vector value.’)
Bases: exceptions.TypeError
exception fipy.terms.SolutionVariableNumberError(s=’Different number of solution variables
and equations.’)
exception fipy.terms.SolutionVariableRequiredError(s=’The solution variable needs to be
specified.’)
exception fipy.terms.IncorrectSolutionVariable(s=’The solution variable is incorrect.’)
fipy.terms.ConvectionTerm
alias of PowerLawConvectionTerm
class fipy.terms.FirstOrderAdvectionTerm(coeff=None)
𝑢|∇𝜑|
𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡
[︃ (︂ )︂2 ]︃1/2 [︃ (︂ )︂2 ]︃1/2
∑︁ 𝜑𝐴 − 𝜑𝑃 ∑︁ 𝜑𝐴 − 𝜑𝑃
𝐴 𝐴

>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
True

Less trivial test:

True
Even less trivial

True
are here again’)
True

True
class fipy.terms.AdvectionTerm(coeff=None)
𝑢|∇𝜑|
𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡
[︃ ]︃1/2 [︃ ]︃1/2
∑︁ 2
∑︁ 2
𝐴 𝐴
where,
𝑑𝐴𝑃 2
and
𝑚 (𝑥, 𝑦) = 𝑥 if |𝑥| ≤ |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 𝑦 if |𝑥| > |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 0 if 𝑥𝑦 < 0

also,
𝐿𝐴 =
𝑑2𝐴𝑃
𝐿𝑃 =
𝑑2𝐴𝑃

>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
True
Less trivial test:

True
Even less trivial

True
are here again’)
True

True
𝜑 = 𝑥2


⃒ ⃒
⃒ 𝜕𝜑 ⃒
−|∇𝜑| = − ⃒⃒ ⃒⃒ = −2|𝑥|
𝜕𝑥
>>> mesh = Grid1D(dx = 1., nx = 5)
>>> vel = 1.
False

True
)︀1/2
𝜑 = 𝑥2 + 𝑦 2
(︀

>>> vel = 1.
>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.123105625618
True
>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.0201715476598
True

class fipy.terms.TransientTerm(coeff=1.0, var=None)
The TransientTerm represents

∫︁
𝑑𝑉 ≃

where 𝜌 is the coeff value.

The following test case verifies that variable coefficients and old coefficient values work correctly. We will solve
the following equation
𝜕𝜑2
= 𝑘.
𝜕𝑡
The analytic solution is given by
√︁
𝜑= 𝜑20 + 𝑘𝑡,
where 𝜑0 is the initial value.

>>> phi0 = 1.
>>> k = 1.
>>> dt = 1.
>>> relaxationFactor = 1.5
>>> steps = 2
>>> sweeps = 8

>>> var = CellVariable(mesh = mesh, value = phi0, hasOld = 1)
>>> from fipy.terms.transientTerm import TransientTerm
>>> from fipy.terms.implicitSourceTerm import ImplicitSourceTerm
Relaxation, given by relaxationFactor, is required for a converged solution.

>>> eq = TransientTerm(var) == ImplicitSourceTerm(-relaxationFactor) \
... + var * relaxationFactor + k
A number of sweeps at each time step are required to let the relaxation take effect.
... var.updateOld()
... eq.solve(var, dt = dt)
Compare the final result with the analytical solution.

>>> print var.allclose(numerix.sqrt(k * dt * steps + phi0**2))
1
class fipy.terms.DiffusionTerm(coeff=(1.0, ), var=None)

Bases: fipy.terms.diffusionTermNoCorrection.DiffusionTermNoCorrection
This term represents a higher order diffusion term. The order of the term is determined by the number of coeffs,
such that:
DiffusionTerm(D1)
represents a typical 2nd-order diffusion term of the form
∇ · (𝐷1 ∇𝜑)
and:

DiffusionTerm((D1,D2))
represents a 4th-order Cahn-Hilliard term of the form
∇ · {𝐷1 ∇ [∇ · (𝐷2 ∇𝜑)]}
and so on.
class fipy.terms.DiffusionTermCorrection(coeff=(1.0, ), var=None)
class fipy.terms.DiffusionTermNoCorrection(coeff=(1.0, ), var=None)
class fipy.terms.DiffusionTermCorrection(coeff=(1.0, ), var=None)
class fipy.terms.DiffusionTermNoCorrection(coeff=(1.0, ), var=None)
class fipy.terms.ExplicitDiffusionTerm(coeff=(1.0, ), var=None)
The discretization for the ExplicitDiffusionTerm is given by
∫︁ ∑︁ 𝜑old − 𝜑old
∇ · (Γ∇𝜑)𝑑𝑉 ≃ Γ𝑓 𝐴 𝑃
𝐴𝑓
𝑉 𝑑𝐴𝑃
𝑓
where 𝜑old old

𝐴 and 𝜑𝑃 are the old values of the variable. The term is added to the RHS vector and makes no
contribution to the solution matrix.
fipy.terms.ImplicitDiffusionTerm
alias of DiffusionTerm
class fipy.terms.ImplicitSourceTerm(coeff=0.0, var=None)
Bases: fipy.terms.sourceTerm.SourceTerm
The ImplicitSourceTerm represents
∫︁
𝜑𝑆 𝑑𝑉 ≃ 𝜑𝑃 𝑆𝑃 𝑉𝑃
𝑉
where 𝑆 is the coeff value.

class fipy.terms.ResidualTerm(equation, underRelaxation=1.0)
Bases: fipy.terms.explicitSourceTerm._ExplicitSourceTerm
The ResidualTerm is a special form of explicit SourceTerm that adds the residual of one equation to another
equation. Useful for Newton’s method.
class fipy.terms.CentralDifferenceConvectionTerm(coeff=1.0, var=None)
Bases: fipy.terms.abstractConvectionTerm._AbstractConvectionTerm
This Term represents
∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the central differencing scheme. For further details


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.ExplicitUpwindConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑old old

𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the upwind scheme. For further details see
Numerical Schemes.


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.ExponentialConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the exponential scheme. For further details see
Numerical Schemes.


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.HybridConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 +(1−𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the hybrid scheme. For further details see Numerical
Schemes.


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.PowerLawConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the power law scheme. For further details see
Numerical Schemes.


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.UpwindConvectionTerm(coeff=1.0, var=None)

∫︁ ∑︁
∇ · (⃗𝑢𝜑) 𝑑𝑉 ≃ (⃗𝑛 · ⃗𝑢)𝑓 𝜑𝑓 𝐴𝑓
𝑉 𝑓
where 𝜑𝑓 = 𝛼𝑓 𝜑𝑃 + (1 − 𝛼𝑓 )𝜑𝐴 and 𝛼𝑓 is calculated using the upwind convection scheme. For further details


>>> m = Grid1D(nx = 2)
...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.VanLeerConvectionTerm(coeff=1.0, var=None)

Bases: fipy.terms.explicitUpwindConvectionTerm.ExplicitUpwindConvectionTerm
>>> m = Grid1D(nx = 2)


...
...
...

...
>>> m2 = Grid2D(nx=2, ny=1)
Parameters
class fipy.terms.FirstOrderAdvectionTerm(coeff=None)
𝑢|∇𝜑|
𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡
[︃ (︂ )︂2 ]︃1/2 [︃ (︂ )︂2 ]︃1/2
∑︁ 𝜑𝐴 − 𝜑𝑃 ∑︁ 𝜑𝐴 − 𝜑𝑃
𝐴 𝐴


>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
True
Less trivial test:

True
Even less trivial

True
are here again’)
True

True
class fipy.terms.AdvectionTerm(coeff=None)
𝑢|∇𝜑|
𝜕𝜑
+ 𝑢|∇𝜑| = 0
𝜕𝑡

[︃ ]︃1/2 [︃ ]︃1/2
∑︁ 2
∑︁ 2
𝐴 𝐴
where,
𝑑𝐴𝑃 2
and
𝑚 (𝑥, 𝑦) = 𝑥 if |𝑥| ≤ |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 𝑦 if |𝑥| > |𝑦|∀𝑥𝑦 ≥ 0
𝑚 (𝑥, 𝑦) = 0 if 𝑥𝑦 < 0
also,
𝐿𝐴 =
𝑑2𝐴𝑃
𝐿𝑃 =
𝑑2𝐴𝑃
>>> mesh = Grid1D(dx = 1., nx = 3)
Trivial test:
True
Less trivial test:

True
Even less trivial

True
are here again’)
True


True
𝜑 = 𝑥2

⃒ ⃒
⃒ 𝜕𝜑 ⃒
−|∇𝜑| = − ⃒⃒ ⃒⃒ = −2|𝑥|
𝜕𝑥
>>> mesh = Grid1D(dx = 1., nx = 5)
>>> vel = 1.
False

True
)︀1/2
𝜑 = 𝑥2 + 𝑦 2
(︀

>>> vel = 1.
>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.123105625618
True


>>> ans[(x > 2) & (x < 8) & (y > 2) & (y < 8)] = 0.0201715476598
True


Chapter 25
fipy.tests package
25.1 Submodules
25.2 fipy.tests.doctestPlus module
fipy.tests.doctestPlus.execButNoTest(name=’__main__’)
fipy.tests.doctestPlus.register_skipper(flag, test, why, skipWarning=True)
Create a new doctest option flag for skipping tests
Parameters flag : str
Name of the option flag
test : function
A function which should return True if the test should be run
why : str
Explanation for why the test was skipped (to be used in a string “Skipped
%%(count)d doctest examples because %%(why)s”)
skipWarning : bool
Whether or not to report on tests skipped by this flag (default True)
fipy.tests.doctestPlus.report_skips()
Print out how many doctest examples were skipped due to flags
fipy.tests.doctestPlus.testmod(m=None, name=None, globs=None, verbose=None,
report=True, optionflags=0, extraglobs=None,
raise_on_error=False, exclude_empty=False)
Test examples in the given module. Return (#failures, #tests).
Largely duplicated from doctest.testmod(), but using _SelectiveDocTestParser.
Test examples in docstrings in functions and classes reachable from module m (or the current module if m is not
supplied), starting with m.__doc__.
Also test examples reachable from dict m.__test__ if it exists and is not None. m.__test__ maps names to
functions, classes and strings; function and class docstrings are tested even if the name is private; strings are
tested directly, as if they were docstrings.
Return (#failures, #tests).
See help(doctest) for an overview.
Optional keyword arg “name” gives the name of the module; by default use m.__name__.
287
Optional keyword arg “globs” gives a dict to be used as the globals when executing examples; by default, use
m.__dict__. A copy of this dict is actually used for each docstring, so that each docstring’s examples start with
a clean slate.
Optional keyword arg “extraglobs” gives a dictionary that should be merged into the globals that are used to
execute examples. By default, no extra globals are used. This is new in 2.4.
Optional keyword arg “verbose” prints lots of stuff if true, prints only failures if false; by default, it’s true iff
“-v” is in sys.argv.
Optional keyword arg “report” prints a summary at the end when true, else prints nothing at the end. In verbose
mode, the summary is detailed, else very brief (in fact, empty if all tests passed).
Optional keyword arg “optionflags” or’s together module constants, and defaults to 0. This is new in 2.3.
Possible values (see the docs for details):
DONT_ACCEPT_TRUE_FOR_1 DONT_ACCEPT_BLANKLINE NORMALIZE_WHITESPACE
ELLIPSIS SKIP IGNORE_EXCEPTION_DETAIL REPORT_UDIFF REPORT_CDIFF RE-
PORT_NDIFF REPORT_ONLY_FIRST_FAILURE
as well as FiPy’s flags
GMSH SCIPY TVTK SERIAL PARALLEL PROCESSOR_0 PROCESSOR_0_OF_2 PROCES-
SOR_1_OF_2 PROCESSOR_0_OF_3 PROCESSOR_1_OF_3 PROCESSOR_2_OF_3
Optional keyword arg “raise_on_error” raises an exception on the first unexpected exception or failure. This
allows failures to be post-mortem debugged.
25.3 fipy.tests.lateImportTest module
25.4 fipy.tests.testBase module
25.5 fipy.tests.testClass module
25.6 fipy.tests.testProgram module
unit testing scripts no chapter heading
288 Chapter 25. fipy.tests package

Chapter 26
fipy.tools package
26.1 Subpackages
26.1.1 fipy.tools.comms package
Submodules
fipy.tools.comms.commWrapper module
class fipy.tools.comms.commWrapper.CommWrapper(Epetra=None)
Bases: object
MPI Communicator wrapper
Encapsulates capabilities needed for Epetra. Some capabilities are not parallel.
Barrier()
MaxAll(vec)
MinAll(vec)
Norm2(vec)
Nproc
all(a, axis=None)
allclose(a, b, rtol=1e-05, atol=1e-08)
allequal(a, b)
allgather(sendobj=None, recvobj=None)
any(a, axis=None)
bcast(obj=None, root=0)
procID
sum(a, axis=None)
class fipy.tools.comms.commWrapper.ParallelCommWrapper(Epetra=None)
Bases: fipy.tools.comms.commWrapper.CommWrapper
MPI Communicator wrapper for parallel processes
289
fipy.tools.comms.dummyComm module
class fipy.tools.comms.dummyComm.DummyComm
Bases: fipy.tools.comms.serialCommWrapper.SerialCommWrapper
Barrier()
MaxAll(vec)
MinAll(vec)
sum(a, axis=None)
fipy.tools.comms.mpi4pyCommWrapper module
class fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper(Epetra, MPI)

MPI Communicator wrapper
Encapsulates capabilities needed for both Epetra and mpi4py.
all(a, axis=None)
allclose(a, b, rtol=1e-05, atol=1e-08)
allequal(a, b)
allgather(sendobj=None, recvobj=None)
any(a, axis=None)
bcast(obj=None, root=0)
fipy.tools.comms.serialCommWrapper module
class fipy.tools.comms.serialCommWrapper.SerialCommWrapper(Epetra=None)
Norm2(vec)
Nproc
procID
Module contents
26.1.2 fipy.tools.dimensions package
Submodules
fipy.tools.dimensions.DictWithDefault module
fipy.tools.dimensions.NumberDict module
fipy.tools.dimensions.physicalField module
Physical quantities with units.
290 Chapter 26. fipy.tools package

This module derives from Konrad Hinsen‘s PhysicalQuantity <http://dirac.cnrs-

orleans.fr/ScientificPython/ScientificPythonManual/Scientific.Physics.PhysicalQuantities-module.html>.
This module provides a data type that represents a physical quantity together with its unit. It is possible to add
and subtract these quantities if the units are compatible, and a quantity can be converted to another compatible unit.
Multiplication, subtraction, and raising to integer powers is allowed without restriction, and the result will have the
correct unit. A quantity can be raised to a non-integer power only if the result can be represented by integer powers of
the base units.
The values of physical constants are taken from the 2002 recommended values from CODATA. Other conversion
factors (e.g. for British units) come from Appendix B of NIST Special Publication 811.
Warning: We can’t guarantee for the correctness of all entries in the unit table, so use this at your own risk!
Base SI units:
m, kg, s, A, K, mol, cd, rad, sr
SI prefixes:
Y = 1e+24
Z = 1e+21
E = 1e+18
P = 1e+15
T = 1e+12
G = 1e+09
M = 1e+06
k = 1000
h = 100
da = 10
d = 0.1
c = 0.01
m = 0.001
mu = 1e-06
n = 1e-09
p = 1e-12
f = 1e-15
a = 1e-18
z = 1e-21
y = 1e-24
Units derived from SI (accepting SI prefixes):

1 Bq = 1 1/s
1 C = 1 A *s
1 degC = 1 K
1 F = 1 A**2*s**4/kg/m**2
1 Gy = 1 m**2/s**2
1 H = 1 kg*m**2/A**2/s**2
1 Hz = 1 1/s
1 J = 1 m**2*kg/s**2
1 lm = 1 sr*cd
1 lx = 1 sr*cd/m**2
1 N = 1 m*kg/s**2
1 ohm = 1 kg*m**2/A**2/s**3
1 Pa = 1 kg/s**2/m
1 S = 1 A**2*s**3/kg/m**2
1 Sv = 1 m**2/s**2
1 T = 1 kg/A/s**2

1 V = 1 kg*m**2/A/s**3
1 W = 1 m**2*kg/s**3
1 Wb = 1 kg*m**2/A/s**2
Other units that accept SI prefixes:

1 eV = 1.60217653e-19 m**2*kg/s**2
Additional units and constants:

1 acres = 4046.8564224 m**2
1 amu = 1.6605402e-27 kg
1 Ang = 1e-10 m
1 atm = 101325.0 kg/s**2/m
1 b = 1e-28 m
1 bar = 100000.0 kg/s**2/m
1 Bohr = 5.29177208115e-11 m
1 Btui = 1055.05585262 m**2*kg/s**2
1 c = 299792458.0 m/s
1 cal = 4.184 m**2*kg/s**2
1 cali = 4.1868 m**2*kg/s**2
1 cl = 1e-05 m**3
1 cup = 0.000236588256 m**3
1 d = 86400.0 s
1 deg = 0.0174532925199 rad
1 degF = 0.555555555556 K
1 degR = 0.555555555556 K
1 dl = 0.0001 m**3
1 dyn = 1e-05 m*kg/s**2
1 e = 1.60217653e-19 A*s
1 eps0 = 8.85418781762e-12 A**2*s**4/kg/m**3
1 erg = 1e-07 m**2*kg/s**2
1 floz = 2.9573532e-05 m**3
1 ft = 0.3048 m
1 g = 0.001 kg
1 galUK = 0.00454609 m**3
1 galUS = 0.003785412096 m**3
1 gn = 9.80665 m/s**2
1 Grav = 6.6742e-11 m**3/s**2/kg
1 h = 3600.0 s
1 ha = 10000.0 m**2
1 Hartree = 4.3597441768e-18 m**2*kg/s**2
1 hbar = 1.05457168236e-34 m**2*kg/s
1 hpEl = 746.0 m**2*kg/s**3
1 hplanck = 6.6260693e-34 m**2*kg/s
1 hpUK = 745.7 m**2*kg/s**3
1 inch = 0.0254 m
1 invcm = 1.98644560233e-23 m**2*kg/s**2
1 kB = 1.3806505e-23 kg*m**2/s**2/K
1 kcal = 4184.0 m**2*kg/s**2
1 kcali = 4186.8 m**2*kg/s**2
1 Ken = 1.3806505e-23 m**2*kg/s**2
1 l = 0.001 m**3
1 lb = 0.45359237 kg
1 lyr = 9.46073047258e+15 m
1 me = 9.1093826e-31 kg
1 mi = 1609.344 m
1 min = 60.0 s
1 ml = 1e-06 m**3

1 mp = 1.67262171e-27 kg
1 mu0 = 1.25663706144e-06 kg*m/A**2/s**2
1 Nav = 6.0221415e+23 1/mol
1 nmi = 1852.0 m
1 oz = 0.028349523125 kg
1 psi = 6894.75729317 kg/s**2/m
1 pt = 0.000473176512 m**3
1 qt = 0.000946353024 m**3
1 tbsp = 1.4786766e-05 m**3
1 ton = 907.18474 kg
1 Torr = 133.322368421 kg/s**2/m
1 tsp = 4.928922e-06 m**3
1 wk = 604800.0 s
1 yd = 0.9144 m
1 yr = 31536000.0 s
1 yrJul = 31557600.0 s
1 yrSid = 31558152.96 s
class fipy.tools.dimensions.physicalField.PhysicalField(value, unit=None, ar-

ray=None)
Bases: object
Physical field or quantity with units
Physical Fields can be constructed in one of two ways:
•PhysicalField(*value*, *unit*), where *value* is a number of arbitrary type and *unit* is a string contain-
ing the unit name
>>> print PhysicalField(value = 10., unit = 'm')
10.0 m
•PhysicalField(*string*), where *string* contains both the value and the unit. This form is provided to
make interactive use more convenient
>>> print PhysicalField(value = "10. m")
10.0 m
Dimensionless quantities, with a unit of 1, can be specified in several ways

>>> print PhysicalField(value = "1")
1.0 1
>>> print PhysicalField(value = 2., unit = " ")
2.0 1
>>> print PhysicalField(value = 2.)
2.0 1
Physical arrays are also possible (and are the reason this code was adapted from Konrad Hinsen‘s original
PhysicalQuantity). The value can be a Numeric array:
>>> a = numerix.array(((3.,4.),(5.,6.)))
>>> print PhysicalField(value = a, unit = "m")
[[ 3. 4.]
[ 5. 6.]] m
or a tuple:
>>> print PhysicalField(value = ((3.,4.),(5.,6.)), unit = "m")
[[ 3. 4.]
[ 5. 6.]] m

or as a single value to be applied to every element of a supplied array:

>>> print PhysicalField(value = 2., unit = "m", array = a)
[[ 2. 2.]
[ 2. 2.]] m
Every element in an array has the same unit, which is stored only once for the whole array.
add(other)
Add two physical quantities, so long as their units are compatible. The unit of the result is the unit of the
first operand.
>>> print PhysicalField(10., 'km') + PhysicalField(10., 'm')
10.01 km
>>> print PhysicalField(10., 'km') + PhysicalField(10., 'J')
...
TypeError: Incompatible units
allclose(other, atol=None, rtol=1e-08)

This function tests whether or not self and other are equal subject to the given relative and absolute
tolerances. The formula used is:
| self - other | < atol + rtol * | other |
This means essentially that both elements are small compared to atol or their difference divided by other‘s
value is small compared to rtol.
allequal(other)
This function tests whether or not self and other are exactly equal.
arccos()
Return the inverse cosine of the PhysicalField in radians
>>> print PhysicalField(0).arccos().allclose("1.57079632679 rad")
1
The input PhysicalField must be dimensionless

>>> print numerix.round_(PhysicalField("1 m").arccos(), 6)
...
arccosh()
Return the inverse hyperbolic cosine of the PhysicalField
>>> print numerix.allclose(PhysicalField(2).arccosh(),
... 1.31695789692)
1

>>> print numerix.round_(PhysicalField("1. m").arccosh(), 6)
...
arcsin()
Return the inverse sine of the PhysicalField in radians

>>> print PhysicalField(1).arcsin().allclose("1.57079632679 rad")

1

>>> print numerix.round_(PhysicalField("1 m").arcsin(), 6)
...
arctan()
Return the arctangent of the PhysicalField in radians
>>> print numerix.round_(PhysicalField(1).arctan(), 6)
0.785398

>>> print numerix.round_(PhysicalField("1 m").arctan(), 6)
...
arctan2(other)
Return the arctangent of self divided by other in radians
>>> print numerix.round_(PhysicalField(2.).arctan2(PhysicalField(5.)), 6)
0.380506
The input PhysicalField objects must be in the same dimensions

>>> print numerix.round_(PhysicalField(2.54, "cm").arctan2(PhysicalField(1., "inch")), 6)
0.785398
>>> print numerix.round_(PhysicalField(2.).arctan2(PhysicalField("5. m")), 6)

...
arctanh()
Return the inverse hyperbolic tangent of the PhysicalField
>>> print PhysicalField(0.5).arctanh()
0.549306144334

>>> print numerix.round_(PhysicalField("1 m").arctanh(), 6)
...
ceil()
Return the smallest integer greater than or equal to the PhysicalField.
>>> print PhysicalField(2.2,"m").ceil()
3.0 m
conjugate()
Return the complex conjugate of the PhysicalField.

>>> print PhysicalField(2.2 - 3j,"ohm").conjugate() == PhysicalField(2.2 + 3j,"ohm")

True
convertToUnit(unit)
Changes the unit to unit and adjusts the value such that the combination is equivalent. The new unit is by
a string containing its name. The new unit must be compatible with the previous unit of the object.
>>> e = PhysicalField('2.7 Hartree*Nav')
>>> e.convertToUnit('kcal/mol')
>>> print e
1694.27557621 kcal/mol
copy()
Make a duplicate.
>>> a = PhysicalField(1, unit = 'inch')
>>> b = a.copy()
The duplicate will not reflect changes made to the original

>>> a.convertToUnit('cm')
>>> print a
2.54 cm
>>> print b
1 inch
Likewise for arrays

>>> a = PhysicalField(numerix.array((0,1,2)), unit = 'm')
>>> b = a.copy()
>>> a[0] = 3
>>> print a
[3 1 2] m
>>> print b
[0 1 2] m
cos()
Return the cosine of the PhysicalField
>>> print numerix.round_(PhysicalField(2*numerix.pi/6,"rad").cos(), 6)
0.5
>>> print numerix.round_(PhysicalField(60.,"deg").cos(), 6)
0.5
The units of the PhysicalField must be an angle

>>> PhysicalField(60.,"m").cos()
...
cosh()
Return the hyperbolic cosine of the PhysicalField
>>> PhysicalField(0.).cosh()
1.0
The units of the PhysicalField must be dimensionless

>>> PhysicalField(60.,"m").cosh()
...
divide(other)
Divide two physical quantities. The unit of the result is the unit of the first operand divided by the unit of
the second.
>>> print PhysicalField(10., 'm') / PhysicalField(2., 's')
5.0 m/s
As a special case, if the result is dimensionless, the value is returned without units, rather than with a
dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that
cannot use units, while ensuring the quantities have the desired units
>>> print (PhysicalField(1., 'inch')
... / PhysicalField(1., 'mm'))
25.4
dot(other)
Return the dot product of self with other. The resulting unit is the product of the units of self and other.
>>> v = PhysicalField(((5.,6.),(7.,8.)), "m")
>>> print PhysicalField(((1.,2.),(3.,4.)), "m").dot(v)
[ 26. 44.] m**2
floor()
Return the largest integer less than or equal to the PhysicalField.
>>> print PhysicalField(2.2,"m").floor()
2.0 m
getNumericValue(*args, **kwds)
Deprecated since version 3.0: use the numericValue property instead
getUnit(*args, **kwds)
Deprecated since version 3.0: use the unit property instead
getsctype(default=None)
Returns the Numpy sctype of the underlying array.
>>> PhysicalField(1, 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1))
True
>>> PhysicalField(1., 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1.))
True
>>> PhysicalField((1,1.), 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array((1.,
True
inBaseUnits()
Return the quantity with all units reduced to their base SI elements.
>>> print e.inBaseUnits().allclose("7088849.01085 kg*m**2/s**2/mol")
1
inDimensionless()
Returns the numerical value of a dimensionless quantity.

>>> print PhysicalField(((2.,3.),(4.,5.))).inDimensionless()

[[ 2. 3.]
[ 4. 5.]]
It’s an error to convert a quantity with units
>>> print PhysicalField(((2.,3.),(4.,5.)),"m").inDimensionless()

...
inRadians()
Converts an angular quantity to radians and returns the numerical value.
>>> print PhysicalField(((2.,3.),(4.,5.)),"rad").inRadians()

[[ 2. 3.]
[ 4. 5.]]
>>> print PhysicalField(((2.,3.),(4.,5.)),"deg").inRadians()
[[ 0.03490659 0.05235988]
[ 0.06981317 0.08726646]]
As a special case, assumes a dimensionless quantity is already in radians.
>>> print PhysicalField(((2.,3.),(4.,5.))).inRadians()

[[ 2. 3.]
[ 4. 5.]]
It’s an error to convert a quantity with non-angular units
>>> print PhysicalField(((2.,3.),(4.,5.)),"m").inRadians()

...
inSIUnits()
Return the quantity with all units reduced to SI-compatible elements.
>>> print e.inSIUnits().allclose("7088849.01085 kg*m**2/s**2/mol")
1
inUnitsOf(*units)
Returns one or more PhysicalField objects that express the same physical quantity in different units. The
units are specified by strings containing their names. The units must be compatible with the unit of the
object. If one unit is specified, the return value is a single PhysicalField.
>>> freeze = PhysicalField('0 degC')
>>> print freeze.inUnitsOf('degF').allclose("32.0 degF")
1
If several units are specified, the return value is a tuple of PhysicalField instances with with one element per
unit such that the sum of all quantities in the tuple equals the the original quantity and all the values except
for the last one are integers. This is used to convert to irregular unit systems like hour/minute/second. The
original object will not be changed.

>>> t = PhysicalField(314159., 's')

>>> print numerix.allclose([e.allclose(v) for (e, v) in zip(t.inUnitsOf('d','h','min','s'),
... ['3.0 d', '15.0 h', '15.0 min',
... True)
1
isCompatible(unit)
itemset(value)
Assign the value of a scalar array, performing appropriate conversions.
>>> a = PhysicalField(4.,"m")
>>> a.itemset(PhysicalField("6 ft"))
>>> print a.allclose("1.8288 m")
1
>>> a = PhysicalField(((3.,4.),(5.,6.)),"m")
>>> try:
... a.itemset(PhysicalField("6 ft"))
... except IndexError:
... # NumPy 1.7 has changed the exception type
... raise ValueError("can only place a scalar for an array of size 1")
...
ValueError: can only place a scalar for an array of size 1
>>> a.itemset(PhysicalField("2 min"))
...
itemsize
log()
Return the natural logarithm of the PhysicalField
>>> print numerix.round_(PhysicalField(10).log(), 6)
2.302585

>>> print numerix.round_(PhysicalField("1. m").log(), 6)
...
log10()
Return the base-10 logarithm of the PhysicalField
>>> print numerix.round_(PhysicalField(10.).log10(), 6)
1.0

>>> print numerix.round_(PhysicalField("1. m").log10(), 6)
...
multiply(other)
Multiply two physical quantities. The unit of the result is the product of the units of the operands.

>>> print PhysicalField(10., 'N') * PhysicalField(10., 'm')

100.0 m*N
cannot use units, while ensuring the quantities have the desired units.
>>> print (PhysicalField(10., 's') * PhysicalField(2., 'Hz'))
20.0
numericValue
Return the PhysicalField without units, after conversion to base SI units.
>>> print numerix.round_(PhysicalField("1 inch").numericValue, 6)
0.0254
put(indices, values)
put is the opposite of take. The values of self at the locations specified in indices are set to the correspond-
ing value of values.
The indices can be any integer sequence object with values suitable for indexing into the flat form of self.
The values must be any sequence of values that can be converted to the typecode of self.
>>> f = PhysicalField((1.,2.,3.),"m")
>>> f.put((2,0), PhysicalField((2.,3.),"inch"))
>>> print f
[ 0.0762 2. 0.0508] m
The units of values must be compatible with self.

>>> f.put(1, PhysicalField(3,"kg"))
...
ravel()
reshape(shape)
Changes the shape of self to that specified in shape
>>> print PhysicalField((1.,2.,3.,4.),"m").reshape((2,2))
[[ 1. 2.]
[ 3. 4.]] m
The new shape must have the same size as the existing one.
...
ValueError: total size of new array must be unchanged
setUnit(*args, **kwds)
shape
Tuple of array dimensions.
sign()
Return the sign of the quantity. The unit is unchanged.

>>> from fipy.tools.numerix import sign

>>> print sign(PhysicalField(((3.,-2.),(-1.,4.)), 'm'))
[[ 1. -1.]
[-1. 1.]]
sin()
Return the sine of the PhysicalField
>>> print PhysicalField(numerix.pi/6,"rad").sin()
0.5
>>> print PhysicalField(30.,"deg").sin()
0.5

>>> PhysicalField(30.,"m").sin()
...
sinh()
Return the hyperbolic sine of the PhysicalField
>>> PhysicalField(0.).sinh()
0.0

>>> PhysicalField(60.,"m").sinh()
...
sqrt()
Return the square root of the PhysicalField
>>> print PhysicalField("100. m**2").sqrt()
10.0 m
The resulting unit must be integral

>>> print PhysicalField("100. m").sqrt()
...
TypeError: Illegal exponent
subtract(other)
Subtract two physical quantities, so long as their units are compatible. The unit of the result is the unit of
the first operand.
>>> print PhysicalField(10., 'km') - PhysicalField(10., 'm')
9.99 km
>>> print PhysicalField(10., 'km') - PhysicalField(10., 'J')
...
sum(index=0)
Returns the sum of all of the elements in self along the specified axis (first axis by default).

>>> print PhysicalField(((1.,2.),(3.,4.)), "m").sum()

[ 4. 6.] m
>>> print PhysicalField(((1.,2.),(3.,4.)), "m").sum(1)
[ 3. 7.] m
take(indices, axis=0)
Return the elements of self specified by the elements of indices. The resulting PhysicalField array has the
same units as the original.
>>> print PhysicalField((1.,2.,3.),"m").take((2,0))
[ 3. 1.] m
The optional third argument specifies the axis along which the selection occurs, and the default value (as
in the example above) is 0, the first axis.
>>> print PhysicalField(((1.,2.,3.),(4.,5.,6.)),"m").take((2,0), axis = 1)
[[ 3. 1.]
[ 6. 4.]] m
tan()
Return the tangent of the PhysicalField
>>> numerix.round_(PhysicalField(numerix.pi/4,"rad").tan(), 6)
1.0
>>> numerix.round_(PhysicalField(45,"deg").tan(), 6)
1.0

>>> PhysicalField(45.,"m").tan()
...
tanh()
Return the hyperbolic tangent of the PhysicalField
>>> print numerix.allclose(PhysicalField(1.).tanh(), 0.761594155956)
True

>>> PhysicalField(60.,"m").tanh()
...
tostring(max_line_width=75, precision=8, suppress_small=False, separator=’ ‘)

Return human-readable form of a physical quantity
>>> p = PhysicalField(value = (3., 3.14159), unit = "eV")
>>> print p.tostring(precision = 3, separator = '|')
[ 3. | 3.142] eV
unit
Return the unit object of self.
>>> PhysicalField("1 m").unit
<PhysicalUnit m>

class fipy.tools.dimensions.physicalField.PhysicalUnit(names, factor, powers, off-

set=0)
A PhysicalUnit represents the units of a PhysicalField.
This class is not generally not instantiated by users of this module, but rather it is created in the process of
constructing a PhysicalField.
Parameters
• names: the name of the unit
• factor: the multiplier between the unit and the fundamental SI unit
• powers: a nine-element list, tuple, or Numeric array representing the fundamental SI units
of [”m”, “kg”, “s”, “A”, “K”, “mol”, “cd”, “rad”, “sr”]
• offset: the displacement between the zero-point of the unit and the zero-point of the corre-
sponding fundamental SI unit.
conversionFactorTo(other)
Return the multiplication factor between two physical units
>>> a = PhysicalField("1. mm")
>>> b = PhysicalField("1. inch")
>>> print numerix.round_(b.unit.conversionFactorTo(a.unit), 6)
25.4
Units must have the same fundamental SI units

>>> c = PhysicalField("1. K")
>>> c.unit.conversionFactorTo(a.unit)
...
If units have different offsets, they must have the same factor
>>> d = PhysicalField("1. degC")
>>> c.unit.conversionFactorTo(d.unit)
1.0
>>> e = PhysicalField("1. degF")
>>> c.unit.conversionFactorTo(e.unit)
...
TypeError: Unit conversion (K to degF) cannot be expressed as a simple multiplicative factor
conversionTupleTo(other)
Return a tuple of the multiplication factor and offset between two physical units
>>> a = PhysicalField("1. K").unit
>>> b = PhysicalField("1. degF").unit
>>> [str(numerix.round_(element,6)) for element in b.conversionTupleTo(a)]
['0.555556', '459.67']
isAngle()
Returns True if the unit is an angle
>>> PhysicalField("1. deg").unit.isAngle()
1
>>> PhysicalField("1. rad").unit.isAngle()
1

>>> PhysicalField("1. inch").unit.isAngle()

0
isCompatible(other)
Returns a list of which fundamental SI units are compatible between self and other
>>> a = PhysicalField("1. mm")
>>> b = PhysicalField("1. inch")
>>> print numerix.allclose(a.unit.isCompatible(b.unit),
... [True, True, True, True, True, True, True, True, True])
True
>>> c = PhysicalField("1. K")
>>> print numerix.allclose(a.unit.isCompatible(c.unit),
... [False, True, True, True, False, True, True, True, True])
True
isDimensionless()
Returns True if the unit is dimensionless
>>> PhysicalField("1. m/m").unit.isDimensionless()
1
>>> PhysicalField("1. inch").unit.isDimensionless()
0
isDimensionlessOrAngle()
Returns True if the unit is dimensionless or an angle
>>> PhysicalField("1. m/m").unit.isDimensionlessOrAngle()
1
>>> PhysicalField("1. deg").unit.isDimensionlessOrAngle()
1
>>> PhysicalField("1. rad").unit.isDimensionlessOrAngle()
1
>>> PhysicalField("1. inch").unit.isDimensionlessOrAngle()
0
isInverseAngle()
Returns True if the 1 divided by the unit is an angle
>>> PhysicalField("1. deg**-1").unit.isInverseAngle()
1
>>> PhysicalField("1. 1/rad").unit.isInverseAngle()
1
>>> PhysicalField("1. inch").unit.isInverseAngle()
0
name()
Return the name of the unit
>>> PhysicalField("1. m").unit.name()
'm'
>>> (PhysicalField("1. m") / PhysicalField("1. s")
... / PhysicalField("1. s")).unit.name()
'm/s**2'
setName(name)
Set the name of the unit to name
>>> a = PhysicalField("1. m/s").unit
>>> a

<PhysicalUnit m/s>
>>> a.setName('meterpersecond')
>>> a
<PhysicalUnit meterpersecond>
Module contents
26.1.3 fipy.tools.performance package
Submodules
fipy.tools.performance.efficiencyTestGenerator module
fipy.tools.performance.efficiencyTestHistory module
fipy.tools.performance.efficiency_test module
class fipy.tools.performance.efficiency_test.Efficiency_test(dist)
Bases: distutils.cmd.Command
Create and initialize a new Command object. Most importantly, invokes the ‘initialize_options()’ method, which
is the real initializer and depends on the actual command being instantiated.
description = ‘run FiPy efficiency tests’
finalize_options()
initialize_options()
run()
user_options = [(‘minimumelements=’, None, ‘minimum number of elements’), (‘factor=’, None, ‘factor by which the
fipy.tools.performance.memoryLeak module
This python script is ripped from http://www.nightmare.com/medusa/memory-leaks.html

It outputs the top 100 number of outstanding references for each object.
fipy.tools.performance.memoryLogger module
class fipy.tools.performance.memoryLogger.MemoryHighWaterThread(pid, sample-

Time=1)
Bases: threading.Thread
run()
stop()
class fipy.tools.performance.memoryLogger.MemoryLogger(sampleTime=1)
start()
stop()

fipy.tools.performance.memoryUsage module
This python script is ripped from http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/286222/index_txt
Module contents
26.2 Submodules
26.3 fipy.tools.copy_script module
class fipy.tools.copy_script.Copy_script(dist)
Bases: distutils.cmd.Command
Create and initialize a new Command object. Most importantly, invokes the ‘initialize_options()’ method, which
is the real initializer and depends on the actual command being instantiated.
description = ‘copy an example script into a new editable file’
finalize_options()
initialize_options()
run()
user_options = [(‘From=’, None, ‘path and file name containing script to copy’), (‘To=’, None, ‘path and file name to s
26.4 fipy.tools.debug module
fipy.tools.debug.PRINT(label, arg=’‘, stall=True)
26.5 fipy.tools.decorators module
fipy.tools.decorators.getsetDeprecated(*args, **kwargs)
Issues a DeprecationWarning to use the appropriate property, rather than the get/set method of the same name
This function may also be used as a decorator.
Parameters func : function
The function to be deprecated.
old_name : str, optional
The name of the function to be deprecated. Default is None, in which case the name of
func is used.
new_name : str, optional
The new name for the function. Default is None, in which case the deprecation message
is that old_name is deprecated. If given, the deprecation message is that old_name is
deprecated and new_name should be used instead.
message : str, optional
Additional explanation of the deprecation. Displayed in the docstring after the warning.

Returns old_func : function

The deprecated function.
fipy.tools.decorators.mathMethodDeprecated(*args, **kwargs)
Issues a DeprecationWarning to use the appropriate ufunc from numerix, rather than the method of the same
name
This function may also be used as a decorator.
Parameters func : function
The function to be deprecated.
old_name : str, optional
The name of the function to be deprecated. Default is None, in which case the name of
func is used.
new_name : str, optional
The new name for the function. Default is None, in which case the deprecation message
is that old_name is deprecated. If given, the deprecation message is that old_name is
deprecated and new_name should be used instead.
message : str, optional
Additional explanation of the deprecation. Displayed in the docstring after the warning.
Returns old_func : function
The deprecated function.
26.6 fipy.tools.dump module
fipy.tools.dump.write(data, filename=None, extension=’‘, communicator=DummyComm())

Pickle an object and write it to a file. Wrapper for cPickle.dump().
Parameters
• data: The object to be pickled.
• filename: The name of the file to place the pickled object. If filename is None then a tempo-
rary file will be used and the file object and file name will be returned as a tuple
• extension: Used if filename is not given.
• communicator: Object with procID and Nproc attributes.
Test to check pickling and unpickling.
>>> old = Grid1D(nx = 2)
>>> f, tempfile = write(old)
>>> new = read(tempfile, f)
>>> print old.numberOfCells == new.numberOfCells
True
fipy.tools.dump.read(filename, fileobject=None, communicator=DummyComm(),

mesh_unmangle=False)
Read a pickled object from a file. Returns the unpickled object. Wrapper for cPickle.load().
Parameters
26.6. fipy.tools.dump module 307

• filename: The name of the file to unpickle the object from.

• fileobject: Used to remove temporary files
• communicator: Object with procID and Nproc attributes.
• mesh_unmangle: Correct improper pickling of non-uniform meshes (ticket:243)
26.7 fipy.tools.inline module
26.8 fipy.tools.numerix module
Replacement module for NumPy
Attention: This module should be the only place in the code where numpy is explicitly imported and you should
always import this module and not numpy in your own code. The documentation for numpy remains canonical
for all functions and classes not explicitly documented here.
The functions provided in ths module replace and augment the NumPy module. The functions work with Variables,
arrays or numbers. For example, create a Variable.
>>> from fipy.variables.variable import Variable
>>> var = Variable(value=0)
Take the tangent of such a variable. The returned value is itself a Variable.
>>> v = tan(var)
>>> v
tan(Variable(value=array(0)))
>>> print float(v)
0.0
Take the tangent of a int.

>>> tan(0)
0.0
Take the tangent of an array.

>>> print tan(array((0,0,0)))
[ 0. 0. 0.]
fipy.tools.numerix.dot(a1, a2, axis=0)

return array of vector dot-products of v1 and v2 for arrays a1 and a2 of vectors v1 and v2
We can’t use numpy.dot() on an array of vectors
Test that Variables are returned as Variables.
>>> mesh = Grid2D(nx=2, ny=1)
>>> v1 = CellVariable(mesh=mesh, value=((0,1),(2,3)), rank=1)
>>> dot(v1, v2)._variableClass
<class 'fipy.variables.cellVariable.CellVariable'>

>>> print rank(dot(v2, v1))

0
>>> print dot(v1, v2)
[ 4 10]
[ 4 10]
>>> v3 = array(((0,1),(2,3)))
>>> print type(dot(v3, v3)) is type(array(1))
1
[ 4 10]
fipy.tools.numerix.indices(dimensions, typecode=None)
indices(dimensions,typecode=None) returns an array representing a grid of indices with row-only, and column-
only variation.
>>> NUMERIX.allclose(NUMERIX.array(indices((4, 6))), NUMERIX.indices((4,6)))
1
>>> NUMERIX.allclose(NUMERIX.array(indices((4, 6, 2))), NUMERIX.indices((4, 6, 2)))
1
>>> NUMERIX.allclose(NUMERIX.array(indices((1,))), NUMERIX.indices((1,)))
1
1
fipy.tools.numerix.isclose(first, second, rtol=1e-05, atol=1e-08)

Returns which elements of first and second are equal, subect to the given relative and absolute tolerances, such
that:
|first - second| < atol + rtol * |second|
This means essentially that both elements are small compared to atol or their difference divided by second‘s
fipy.tools.numerix.allclose(first, second, rtol=1e-05, atol=1e-08)
Tests whether or not first and second are equal, subect to the given relative and absolute tolerances, such that:
fipy.tools.numerix.all(a, axis=None, out=None)
Test whether all array elements along a given axis evaluate to True.
Parameters a : array_like
Input array or object that can be converted to an array.
axis : int, optional
Axis along which an logical AND is performed. The default (axis = None) is to perform
a logical AND over a flattened input array. axis may be negative, in which case it counts
from the last to the first axis.
out : ndarray, optional
Alternative output array in which to place the result. It must have the same shape as the
expected output and the type is preserved.
26.8. fipy.tools.numerix module 309

fipy.tools.numerix.put(arr, ids, values)

The opposite of take. The values of arr at the locations specified by ids are set to the corresponding value of
values.
The following is to test improvments to puts with masked arrays. Places in the code were assuming incorrect
put behavior.
>>> maskValue = 999999
>>> arr = zeros(3, 'l')

>>> ids = MA.masked_values((2, maskValue), maskValue)
>>> values = MA.masked_values((4, maskValue), maskValue)
>>> put(arr, ids, values) ## this should work
>>> print arr
[0 0 4]
>>> arr = MA.masked_values((maskValue, 5, 10), maskValue)

>>> put(arr, ids, values)
>>> print arr ## works as expected
[-- 5 4]

>>> ids = MA.masked_values((maskValue, 2), maskValue)
>>> print arr ## should be [-- 5 --] maybe??
[-- 5 999999]
fipy.tools.numerix.rank(a)
Get the rank of sequence a (the number of dimensions, not a matrix rank) The rank of a scalar is zero.
Note: The rank of a MeshVariable is for any single element. E.g., A CellVariable containing scalars at each
cell, and defined on a 9 element Grid1D, has rank 0. If it is defined on a 3x3 Grid2D, it is still rank 0.
fipy.tools.numerix.reshape(arr, shape)
Change the shape of arr to shape, as long as the product of all the lenghts of all the axes is constant (the total
number of elements does not change).
fipy.tools.numerix.sum(arr, axis=0)
The sum of all the elements of arr along the specified axis.
fipy.tools.numerix.take(a, indices, axis=0, fill_value=None)
Selects the elements of a corresponding to indices.

values.
put behavior.
>>> arr = zeros(3, 'l')

>>> print arr
[0 0 4]

[-- 5 4]

[-- 5 999999]


fipy.tools.numerix.allclose(first, second, rtol=1e-05, atol=1e-08)
Tests whether or not first and second are equal, subect to the given relative and absolute tolerances, such that:
fipy.tools.numerix.allequal(first, second)
Returns true if every element of first is equal to the corresponding element of second.
fipy.tools.numerix.dot(a1, a2, axis=0)
return array of vector dot-products of v1 and v2 for arrays a1 and a2 of vectors v1 and v2
We can’t use numpy.dot() on an array of vectors
Test that Variables are returned as Variables.
>>> print rank(dot(v2, v1))
0
[ 4 10]
[ 4 10]
>>> v3 = array(((0,1),(2,3)))
>>> print type(dot(v3, v3)) is type(array(1))
1
[ 4 10]
fipy.tools.numerix.getShape(arr)
Return the shape of arr
>>> getShape(1)
()
>>> getShape(1.)
()
>>> getShape(Variable(1))
()
>>> getShape(Variable(1.))
()
>>> getShape(Variable(1., unit="m"))
()

>>> getShape(Variable("1 m"))

()
fipy.tools.numerix.getUnit(arr)
fipy.tools.numerix.indices(dimensions, typecode=None)
indices(dimensions,typecode=None) returns an array representing a grid of indices with row-only, and column-
only variation.
>>> NUMERIX.allclose(NUMERIX.array(indices((4, 6))), NUMERIX.indices((4,6)))
1
>>> NUMERIX.allclose(NUMERIX.array(indices((4, 6, 2))), NUMERIX.indices((4, 6, 2)))
1
1
1
fipy.tools.numerix.isclose(first, second, rtol=1e-05, atol=1e-08)

Returns which elements of first and second are equal, subect to the given relative and absolute tolerances, such
that:
fipy.tools.numerix.isFloat(arr)
fipy.tools.numerix.isInt(arr)
fipy.tools.numerix.L1norm(arr)
Parameters
• arr: The array to evaluate.
∑︀𝑛
Returns ‖arr‖1 = 𝑗=1 |arr𝑗 | is the 𝐿1 -norm of arr.
fipy.tools.numerix.L2norm(arr)
Parameters
√︁∑︀
𝑛
𝑗=1 |arr𝑗 | is the 𝐿 -norm of arr.
2 2
Returns ‖arr‖2 =
fipy.tools.numerix.LINFnorm(arr)
Parameters
∞ ∞
‖arr‖∞ =[ 𝑛
∑︀
𝑗=1 |arr𝑗 | ] =
Returns max𝑗|arr𝑗 | is the 𝐿∞ -norm of arr.
fipy.tools.numerix.nearest(data, points, max_mem=100000000.0)
find the indices of data that are closest to points
>>> m0 = Grid2D(dx=(.1, 1., 10.), dy=(.1, 1., 10.))
>>> m1 = Grid2D(nx=2, ny=2, dx=5., dy=5.)
>>> print nearest(m0.cellCenters.globalValue, m1.cellCenters.globalValue)
[4 5 7 8]

>>> print nearest(m0.cellCenters.globalValue, m1.cellCenters.globalValue, max_mem=100)

[4 5 7 8]
>>> print nearest(m0.cellCenters.globalValue, m1.cellCenters.globalValue, max_mem=10000)
[4 5 7 8]

values.
put behavior.
>>> arr = zeros(3, 'l')

>>> print arr
[0 0 4]

[-- 5 4]

[-- 5 999999]
fipy.tools.numerix.sqrtDot(a1, a2)
Return array of square roots of vector dot-products for arrays a1 and a2 of vectors v1 and v2
Usually used with v1==v2 to return magnitude of v1.
fipy.tools.numerix.tostring(arr, max_line_width=75, precision=8, suppress_small=False, sepa-
rator=’ ‘, array_output=0)
Returns a textual representation of a number or field of numbers. Each dimension is indicated by a pair of match-
ing square brackets ([]), within which each subset of the field is output. The orientation of the dimensions is as

follows: the last (rightmost) dimension is always horizontal, so that the frequent rank-1 fields use a minimum
of screen real-estate. The next-to-last dimesnion is displayed vertically if present and any earlier dimension is
displayed with additional bracket divisions.
Parameters
• max_line_width: the maximum number of characters used in a single line. Default is
sys.output_line_width or 77.
• precision: the number of digits after the decimal point. Default is sys.float_output_precision
or 8.
• suppress_small: whether small values should be suppressed (and output as 0). Default is
sys.float_output_suppress_small or false.
• separator: what character string to place between two numbers.
• array_output: Format output for an eval. Only used if arr is a Numeric array.
>>> from fipy import Variable

>>> print tostring(Variable((1,0,11.2345)), precision=1)
[ 1. 0. 11.2]
>>> print tostring(array((1,2)), precision=5)
[1 2]
>>> print tostring(array((1.12345,2.79)), precision=2)
[ 1.12 2.79]
>>> print tostring(1)
1
>>> print tostring(array(1))
1
>>> print tostring(array([1.23345]), precision=2)
[ 1.23]
>>> print tostring(array([1]), precision=2)
[1]
>>> print tostring(1.123456, precision=2)
1.12
>>> print tostring(array(1.123456), precision=3)
1.123
26.9 fipy.tools.parser module
fipy.tools.parser.parse(larg, action=None, type=None, default=None)

This is a wrapper function for the python optparse module. Unfortunately optparse does not allow command
line arguments to be ignored. See the documentation for optparse for more details. Returns the argument value.
Parameters
• larg: The argument to be parsed.
• action: store or store_true are possibilities
• type: Type of the argument. int or float are possibilities.
• default: Default value.
26.9. fipy.tools.parser module 315

26.10 fipy.tools.test module
26.11 fipy.tools.vector module
Vector utility functions that are inexplicably absent from Numeric

fipy.tools.vector.putAdd(vector, ids, additionVector)
This is a temporary replacement for Numeric.put as it was not doing what we thought it was doing.
fipy.tools.vector.prune(array, shift, start=0, axis=0)
removes elements with indices i = start + shift * n where n = 0, 1, 2, ...
>>> prune(numerix.arange(10), 3, 5)
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> prune(numerix.arange(10), 3, 2)
array([0, 1, 3, 4, 6, 7, 9])
>>> prune(numerix.arange(10), 3)
array([1, 2, 4, 5, 7, 8])
>>> prune(numerix.arange(4, 7), 3)
array([5, 6])
26.12 fipy.tools.vitals module
class fipy.tools.vitals.Vitals
Bases: xml.dom.minidom.Document
Returns XML formatted information about current FiPy environment
appendChild(child)
appendInfo(name, svnpath=None, **kwargs)
append some additional information, possibly about a project under a separate svn repository
dictToXML(d, name)
save(fname)
svn(*args)
svncmd(cmd, *args)
tupleToXML(t, name, keys=None)
class fipy.tools.PhysicalField(value, unit=None, array=None)

Bases: object
Physical field or quantity with units
Physical Fields can be constructed in one of two ways:
•PhysicalField(*value*, *unit*), where *value* is a number of arbitrary type and *unit* is a string contain-
ing the unit name

>>> print PhysicalField(value = 10., unit = 'm')

10.0 m
•PhysicalField(*string*), where *string* contains both the value and the unit. This form is provided to
make interactive use more convenient
>>> print PhysicalField(value = "10. m")
10.0 m
Dimensionless quantities, with a unit of 1, can be specified in several ways

>>> print PhysicalField(value = "1")
1.0 1
>>> print PhysicalField(value = 2., unit = " ")
2.0 1
>>> print PhysicalField(value = 2.)
2.0 1
Physical arrays are also possible (and are the reason this code was adapted from Konrad Hinsen‘s original
PhysicalQuantity). The value can be a Numeric array:
>>> a = numerix.array(((3.,4.),(5.,6.)))
>>> print PhysicalField(value = a, unit = "m")
[[ 3. 4.]
[ 5. 6.]] m
or a tuple:
>>> print PhysicalField(value = ((3.,4.),(5.,6.)), unit = "m")
[[ 3. 4.]
[ 5. 6.]] m
or as a single value to be applied to every element of a supplied array:

>>> print PhysicalField(value = 2., unit = "m", array = a)
[[ 2. 2.]
[ 2. 2.]] m
Every element in an array has the same unit, which is stored only once for the whole array.
add(other)
Add two physical quantities, so long as their units are compatible. The unit of the result is the unit of the
first operand.
>>> print PhysicalField(10., 'km') + PhysicalField(10., 'm')
10.01 km
>>> print PhysicalField(10., 'km') + PhysicalField(10., 'J')
...
allclose(other, atol=None, rtol=1e-08)

This function tests whether or not self and other are equal subject to the given relative and absolute
tolerances. The formula used is:
| self - other | < atol + rtol * | other |
This means essentially that both elements are small compared to atol or their difference divided by other‘s

allequal(other)
This function tests whether or not self and other are exactly equal.
arccos()
Return the inverse cosine of the PhysicalField in radians
>>> print PhysicalField(0).arccos().allclose("1.57079632679 rad")
1

>>> print numerix.round_(PhysicalField("1 m").arccos(), 6)
...
arccosh()
Return the inverse hyperbolic cosine of the PhysicalField
>>> print numerix.allclose(PhysicalField(2).arccosh(),
... 1.31695789692)
1

>>> print numerix.round_(PhysicalField("1. m").arccosh(), 6)
...
arcsin()
Return the inverse sine of the PhysicalField in radians
>>> print PhysicalField(1).arcsin().allclose("1.57079632679 rad")
1

>>> print numerix.round_(PhysicalField("1 m").arcsin(), 6)
...
arctan()
Return the arctangent of the PhysicalField in radians
>>> print numerix.round_(PhysicalField(1).arctan(), 6)
0.785398

>>> print numerix.round_(PhysicalField("1 m").arctan(), 6)
...
arctan2(other)
Return the arctangent of self divided by other in radians
>>> print numerix.round_(PhysicalField(2.).arctan2(PhysicalField(5.)), 6)
0.380506

The input PhysicalField objects must be in the same dimensions

>>> print numerix.round_(PhysicalField(2.54, "cm").arctan2(PhysicalField(1., "inch")), 6)
0.785398
>>> print numerix.round_(PhysicalField(2.).arctan2(PhysicalField("5. m")), 6)

...
arctanh()
Return the inverse hyperbolic tangent of the PhysicalField
>>> print PhysicalField(0.5).arctanh()
0.549306144334

>>> print numerix.round_(PhysicalField("1 m").arctanh(), 6)
...
ceil()
Return the smallest integer greater than or equal to the PhysicalField.
>>> print PhysicalField(2.2,"m").ceil()
3.0 m
conjugate()
Return the complex conjugate of the PhysicalField.
>>> print PhysicalField(2.2 - 3j,"ohm").conjugate() == PhysicalField(2.2 + 3j,"ohm")
True
convertToUnit(unit)
Changes the unit to unit and adjusts the value such that the combination is equivalent. The new unit is by
a string containing its name. The new unit must be compatible with the previous unit of the object.
>>> e.convertToUnit('kcal/mol')
>>> print e
1694.27557621 kcal/mol
copy()
Make a duplicate.
>>> a = PhysicalField(1, unit = 'inch')
>>> b = a.copy()

>>> a.convertToUnit('cm')
>>> print a
2.54 cm
>>> print b
1 inch
Likewise for arrays

>>> a = PhysicalField(numerix.array((0,1,2)), unit = 'm')

>>> b = a.copy()
>>> a[0] = 3
>>> print a
[3 1 2] m
>>> print b
[0 1 2] m
cos()
Return the cosine of the PhysicalField
>>> print numerix.round_(PhysicalField(2*numerix.pi/6,"rad").cos(), 6)
0.5
>>> print numerix.round_(PhysicalField(60.,"deg").cos(), 6)
0.5

>>> PhysicalField(60.,"m").cos()
...
cosh()
Return the hyperbolic cosine of the PhysicalField
>>> PhysicalField(0.).cosh()
1.0

>>> PhysicalField(60.,"m").cosh()
...
divide(other)
Divide two physical quantities. The unit of the result is the unit of the first operand divided by the unit of
the second.
>>> print PhysicalField(10., 'm') / PhysicalField(2., 's')
5.0 m/s
cannot use units, while ensuring the quantities have the desired units
>>> print (PhysicalField(1., 'inch')
... / PhysicalField(1., 'mm'))
25.4
dot(other)
Return the dot product of self with other. The resulting unit is the product of the units of self and other.
>>> v = PhysicalField(((5.,6.),(7.,8.)), "m")
>>> print PhysicalField(((1.,2.),(3.,4.)), "m").dot(v)
[ 26. 44.] m**2
floor()
Return the largest integer less than or equal to the PhysicalField.

>>> print PhysicalField(2.2,"m").floor()

2.0 m
>>> PhysicalField(1, 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1))
True
>>> PhysicalField(1., 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1.))
True
>>> PhysicalField((1,1.), 'm').getsctype() == numerix.NUMERIX.obj2sctype(numerix.array((1.,
True
inBaseUnits()
Return the quantity with all units reduced to their base SI elements.
1
inDimensionless()
Returns the numerical value of a dimensionless quantity.
>>> print PhysicalField(((2.,3.),(4.,5.))).inDimensionless()

[[ 2. 3.]
[ 4. 5.]]
It’s an error to convert a quantity with units
>>> print PhysicalField(((2.,3.),(4.,5.)),"m").inDimensionless()

...
inRadians()
Converts an angular quantity to radians and returns the numerical value.
>>> print PhysicalField(((2.,3.),(4.,5.)),"rad").inRadians()

[[ 2. 3.]
[ 4. 5.]]
>>> print PhysicalField(((2.,3.),(4.,5.)),"deg").inRadians()
[[ 0.03490659 0.05235988]
[ 0.06981317 0.08726646]]
As a special case, assumes a dimensionless quantity is already in radians.
>>> print PhysicalField(((2.,3.),(4.,5.))).inRadians()

[[ 2. 3.]
[ 4. 5.]]

It’s an error to convert a quantity with non-angular units
>>> print PhysicalField(((2.,3.),(4.,5.)),"m").inRadians()

...
inSIUnits()
Return the quantity with all units reduced to SI-compatible elements.
>>> print e.inSIUnits().allclose("7088849.01085 kg*m**2/s**2/mol")
1
inUnitsOf(*units)
Returns one or more PhysicalField objects that express the same physical quantity in different units. The
units are specified by strings containing their names. The units must be compatible with the unit of the
object. If one unit is specified, the return value is a single PhysicalField.
>>> freeze = PhysicalField('0 degC')
1
If several units are specified, the return value is a tuple of PhysicalField instances with with one element per
>>> t = PhysicalField(314159., 's')
... ['3.0 d', '15.0 h', '15.0 min',
... True)
1
isCompatible(unit)
itemset(value)
Assign the value of a scalar array, performing appropriate conversions.
>>> a = PhysicalField(4.,"m")
>>> a.itemset(PhysicalField("6 ft"))
>>> print a.allclose("1.8288 m")
1
>>> a = PhysicalField(((3.,4.),(5.,6.)),"m")
>>> try:
... a.itemset(PhysicalField("6 ft"))
... except IndexError:
... # NumPy 1.7 has changed the exception type
... raise ValueError("can only place a scalar for an array of size 1")
...
ValueError: can only place a scalar for an array of size 1
>>> a.itemset(PhysicalField("2 min"))
...
itemsize

log()
Return the natural logarithm of the PhysicalField
>>> print numerix.round_(PhysicalField(10).log(), 6)
2.302585

>>> print numerix.round_(PhysicalField("1. m").log(), 6)
...
log10()
Return the base-10 logarithm of the PhysicalField
>>> print numerix.round_(PhysicalField(10.).log10(), 6)
1.0

>>> print numerix.round_(PhysicalField("1. m").log10(), 6)
...
multiply(other)
Multiply two physical quantities. The unit of the result is the product of the units of the operands.
>>> print PhysicalField(10., 'N') * PhysicalField(10., 'm')
100.0 m*N
cannot use units, while ensuring the quantities have the desired units.
>>> print (PhysicalField(10., 's') * PhysicalField(2., 'Hz'))
20.0
numericValue
Return the PhysicalField without units, after conversion to base SI units.
>>> print numerix.round_(PhysicalField("1 inch").numericValue, 6)
0.0254
put(indices, values)
put is the opposite of take. The values of self at the locations specified in indices are set to the correspond-
ing value of values.
The indices can be any integer sequence object with values suitable for indexing into the flat form of self.
The values must be any sequence of values that can be converted to the typecode of self.
>>> f = PhysicalField((1.,2.,3.),"m")
>>> f.put((2,0), PhysicalField((2.,3.),"inch"))
>>> print f
[ 0.0762 2. 0.0508] m
The units of values must be compatible with self.

>>> f.put(1, PhysicalField(3,"kg"))

...
ravel()
reshape(shape)
Changes the shape of self to that specified in shape
[[ 1. 2.]
[ 3. 4.]] m
The new shape must have the same size as the existing one.
...
ValueError: total size of new array must be unchanged
shape
sign()
Return the sign of the quantity. The unit is unchanged.
>>> from fipy.tools.numerix import sign
>>> print sign(PhysicalField(((3.,-2.),(-1.,4.)), 'm'))
[[ 1. -1.]
[-1. 1.]]
sin()
Return the sine of the PhysicalField
>>> print PhysicalField(numerix.pi/6,"rad").sin()
0.5
>>> print PhysicalField(30.,"deg").sin()
0.5

>>> PhysicalField(30.,"m").sin()
...
sinh()
Return the hyperbolic sine of the PhysicalField
>>> PhysicalField(0.).sinh()
0.0

>>> PhysicalField(60.,"m").sinh()
...

sqrt()
Return the square root of the PhysicalField
>>> print PhysicalField("100. m**2").sqrt()
10.0 m
The resulting unit must be integral

>>> print PhysicalField("100. m").sqrt()
...
TypeError: Illegal exponent
subtract(other)
Subtract two physical quantities, so long as their units are compatible. The unit of the result is the unit of
the first operand.
>>> print PhysicalField(10., 'km') - PhysicalField(10., 'm')
9.99 km
>>> print PhysicalField(10., 'km') - PhysicalField(10., 'J')
...
sum(index=0)
Returns the sum of all of the elements in self along the specified axis (first axis by default).
>>> print PhysicalField(((1.,2.),(3.,4.)), "m").sum()
[ 4. 6.] m
>>> print PhysicalField(((1.,2.),(3.,4.)), "m").sum(1)
[ 3. 7.] m
take(indices, axis=0)
Return the elements of self specified by the elements of indices. The resulting PhysicalField array has the
same units as the original.
>>> print PhysicalField((1.,2.,3.),"m").take((2,0))
[ 3. 1.] m
The optional third argument specifies the axis along which the selection occurs, and the default value (as
in the example above) is 0, the first axis.
>>> print PhysicalField(((1.,2.,3.),(4.,5.,6.)),"m").take((2,0), axis = 1)
[[ 3. 1.]
[ 6. 4.]] m
tan()
Return the tangent of the PhysicalField
>>> numerix.round_(PhysicalField(numerix.pi/4,"rad").tan(), 6)
1.0
>>> numerix.round_(PhysicalField(45,"deg").tan(), 6)
1.0

>>> PhysicalField(45.,"m").tan()
...

tanh()
Return the hyperbolic tangent of the PhysicalField
>>> print numerix.allclose(PhysicalField(1.).tanh(), 0.761594155956)
True

>>> PhysicalField(60.,"m").tanh()
...

Return human-readable form of a physical quantity
>>> p = PhysicalField(value = (3., 3.14159), unit = "eV")
>>> print p.tostring(precision = 3, separator = '|')
[ 3. | 3.142] eV
unit
>>> PhysicalField("1 m").unit
<PhysicalUnit m>
class fipy.tools.Vitals
Bases: xml.dom.minidom.Document
Returns XML formatted information about current FiPy environment
appendChild(child)
appendInfo(name, svnpath=None, **kwargs)
append some additional information, possibly about a project under a separate svn repository
dictToXML(d, name)
save(fname)
svn(*args)
svncmd(cmd, *args)
tupleToXML(t, name, keys=None)

Chapter 27
fipy.variables package
27.1 Submodules
27.2 fipy.variables.addOverFacesVariable module
27.3 fipy.variables.arithmeticCellToFaceVariable module
27.4 fipy.variables.betaNoiseVariable module
class fipy.variables.betaNoiseVariable.BetaNoiseVariable(mesh, alpha, beta, name=’‘,

hasOld=0)
Bases: fipy.variables.noiseVariable.NoiseVariable
Represents a beta distribution of random numbers with the probability distribution
𝛽 𝛼 𝑒−𝛽𝑥
𝑥𝛼−1
Γ(𝛼)
∫︀ ∞
with a shape parameter 𝛼, a rate parameter 𝛽, and Γ(𝑧) = 0 𝑡𝑧−1 𝑒−𝑡 𝑑𝑡.
Seed the random module for the sake of deterministic test results.
We generate noise on a uniform cartesian mesh

>>> alpha = Variable()
>>> beta = Variable()
>>> noise = BetaNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), alpha = alpha, beta = beta)
We histogram the root-volume-weighted noise distribution

>>> from fipy.variables.histogramVariable import HistogramVariable
>>> histogram = HistogramVariable(distribution = noise, dx = 0.01, nx = 100)
and compare to a Gaussian distribution

>>> betadist = CellVariable(mesh = histogram.mesh)
>>> x = CellVariable(mesh=histogram.mesh, value=histogram.mesh.cellCenters[0])
>>> from scipy.special import gamma as Gamma
327
>>> betadist = ((Gamma(alpha + beta) / (Gamma(alpha) * Gamma(beta)))

... * x**(alpha - 1) * (1 - x)**(beta - 1))
>>> if __name__ == '__main__':

... from fipy import Viewer
... viewer = Viewer(vars=noise, datamin=0, datamax=1)
... histoplot = Viewer(vars=(histogram, betadist),
... datamin=0, datamax=1.5)
>>> from fipy.tools.numerix import arange
>>> for a in arange(0.5,5,0.5):

... alpha.value = a
... for b in arange(0.5,5,0.5):
... beta.value = b
... if __name__ == '__main__':
... import sys
... print >>sys.stderr, "alpha: %g, beta: %g" % (alpha, beta)
... viewer.plot()
... histoplot.plot()
>>> print abs(noise.faceGrad.divergence.cellVolumeAverage) < 5e-15

1
2 a = b = 0.5
0.0 0.2 0.4 0.6 0.8 1.0
Parameters
• mesh: The mesh on which to define the noise.
• alpha: The parameter 𝛼.
• beta: The parameter 𝛽.
random()
328 Chapter 27. fipy.variables package

27.5 fipy.variables.binaryOperatorVariable module
27.6 fipy.variables.cellToFaceVariable module
27.7 fipy.variables.cellVariable module
class fipy.variables.cellVariable.CellVariable(mesh, name=’‘, value=0.0, rank=None, el-

ementshape=None, unit=None, hasOld=0)
Bases: fipy.variables.meshVariable._MeshVariable
Represents the field of values of a variable on a Mesh.
A CellVariable can be pickled to persistent storage (disk) for later use:
>>> var = CellVariable(mesh = mesh, value = 1., hasOld = 1, name = 'test')

>>> var.value = (x * y)
>>> from fipy.tools import dump

>>> (f, filename) = dump.write(var, extension = '.gz')
>>> unPickledVar = dump.read(filename, f)
>>> print var.allclose(unPickledVar, atol = 1e-10, rtol = 1e-10)

1
arithmeticFaceValue
Returns a FaceVariable whose value corresponds to the arithmetic interpolation of the adjacent cells:
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1
>>> R = 2
>>> var = CellVariable(mesh = mesh, value = (L, R))
>>> faceValue = var.arithmeticFaceValue[mesh.interiorFaces.value]
>>> answer = (R - L) * (0.5 / 1.) + L
>>> print numerix.allclose(faceValue, answer, atol = 1e-10, rtol = 1e-10)
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = (R - L) * (1.0 / 3.0) + L
True
>>> mesh = Grid1D(dx = (10., 100.))

27.5. fipy.variables.binaryOperatorVariable module 329

>>> answer = (R - L) * (5.0 / 55.0) + L

True
cellVolumeAverage
Return the cell-volume-weighted average of the CellVariable:
∑︀
cells 𝜑cell 𝑉cell
< 𝜑 >vol = ∑︀
cells 𝑉cell

>>> mesh = Grid2D(nx = 3, ny = 1, dx = .5, dy = .1)
>>> var = CellVariable(value = (1, 2, 6), mesh = mesh)
>>> print var.cellVolumeAverage
3.0
constrain(value, where=None)
Constrains the CellVariable to value at a location specified by where.
>>> m = Grid1D(nx=3)
>>> v = CellVariable(mesh=m, value=m.cellCenters[0])
>>> v.constrain(0., where=m.facesLeft)
>>> v.faceGrad.constrain([1.], where=m.facesRight)
>>> print v.faceGrad
[[ 1. 1. 1. 1.]]
>>> print v.faceValue
[ 0. 1. 2. 2.5]
Changing the constraint changes the dependencies

[[-1. 1. 1. 1.]]
[ 1. 1. 2. 2.5]
Constraints can be Variable

>>> c = Variable(0.)
>>> v.constrain(c, where=m.facesLeft)
[[ 1. 1. 1. 1.]]
[ 0. 1. 2. 2.5]
>>> c.value = 1.
[[-1. 1. 1. 1.]]
[ 1. 1. 2. 2.5]
Constraints can have a Variable mask.

>>> mask = FaceVariable(mesh=m, value=m.facesLeft)
>>> v.constrain(1., where=mask)

[ 1. 0. 0. 0.]
>>> mask[:] = mask | m.facesRight
[ 1. 0. 0. 1.]
copy()
faceGrad
Return ∇𝜑 as a rank-1 FaceVariable using differencing for the normal direction(second-order gradient).
faceGradAverage
Return ∇𝜑 as a rank-1 FaceVariable using averaging for the normal direction(second-order gradient)
faceValue
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1
>>> R = 2
>>> answer = (R - L) * (0.5 / 1.) + L
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = (R - L) * (1.0 / 3.0) + L
True
>>> mesh = Grid1D(dx = (10., 100.))

>>> answer = (R - L) * (5.0 / 55.0) + L
True
gaussGrad ∑︀
Return 𝑉1𝑃 𝑓 ⃗𝑛𝜑𝑓 𝐴𝑓 as a rank-1 CellVariable (first-order gradient).
getArithmeticFaceValue(*args, **kwds)
Deprecated since version 3.0: use the arithmeticFaceValue property instead
getCellVolumeAverage(*args, **kwds)
Deprecated since version 3.0: use the cellVolumeAverage property instead
getFaceGrad(*args, **kwds)
Deprecated since version 3.0: use the faceGrad property instead
getFaceGradAverage(*args, **kwds)
Deprecated since version 3.0: use the faceGradAverage property instead
27.7. fipy.variables.cellVariable module 331

getFaceValue(*args, **kwds)
getGaussGrad(*args, **kwds)
Deprecated since version 3.0: use the gaussGrad property instead
getGrad(*args, **kwds)
Deprecated since version 3.0: use the grad property instead
getHarmonicFaceValue(*args, **kwds)
Deprecated since version 3.0: use the harmonicFaceValue property instead
getLeastSquaresGrad(*args, **kwds)
Deprecated since version 3.0: use the leastSquaresGrad property instead
getMinmodFaceValue(*args, **kwds)
Deprecated since version 3.0: use the minmodFaceValue property instead
getOld(*args, **kwds)
Deprecated since version 3.0: use the old property instead
globalValue
Concatenate and return values from all processors
When running on a single processor, the result is identical to value.
grad
Return ∇𝜑 as a rank-1 CellVariable (first-order gradient).
harmonicFaceValue
Returns a FaceVariable whose value corresponds to the harmonic interpolation of the adjacent cells:
𝜑1 𝜑2
𝜑𝑓 = 𝑑
(𝜑2 − 𝜑1 ) 𝑑𝑓122 + 𝜑1

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1
>>> R = 2
>>> faceValue = var.harmonicFaceValue[mesh.interiorFaces.value]
>>> answer = L * R / ((R - L) * (0.5 / 1.) + L)
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = L * R / ((R - L) * (1.0 / 3.0) + L)
True
>>> mesh = Grid1D(dx = (10., 100.))

>>> answer = L * R / ((R - L) * (5.0 / 55.0) + L)
True

leastSquaresGrad
Return ∇𝜑, which is determined by solving for ∇𝜑 in the following matrix equation,
∑︁ ∑︁
∇𝜑 · 𝑑2𝐴𝑃 ⃗𝑛𝐴𝑃 ⊗ ⃗𝑛𝐴𝑃 = 𝑑2𝐴𝑃 (⃗𝑛 · ∇𝜑)𝐴𝑃
𝑓 𝑓
The matrix equation is derived by minimizing the following least squares sum,
√︃∑︁
2
𝐹 (𝜑𝑥 , 𝜑𝑦 ) = (𝑑𝐴𝑃 ⃗𝑛𝐴𝑃 · ∇𝜑 − 𝑑𝐴𝑃 (⃗𝑛𝐴𝑃 · ∇𝜑)𝐴𝑃 )
𝑓
Tests
>>> from fipy import Grid2D
>>> m = Grid2D(nx=2, ny=2, dx=0.1, dy=2.0)
>>> print numerix.allclose(CellVariable(mesh=m, value=(0,1,3,6)).leastSquaresGrad.globalValu
... [[8.0, 8.0, 24.0, 24.0],
... [1.2, 2.0, 1.2, 2.0]])
True

>>> print numerix.allclose(CellVariable(mesh=Grid1D(dx=(2.0, 1.0, 0.5)),
... value=(0, 1, 2)).leastSquaresGrad.globalValue, [[0.4
True
minmodFaceValue
Returns a FaceVariable with a value that is the minimum of the absolute values of the adjacent cells. If the
values are of opposite sign then the result is zero:
⎧
⎨𝜑1 when |𝜑1 | ≤ |𝜑2 |,
⎪
𝜑𝑓 = 𝜑2 when |𝜑2 | < |𝜑1 |,
when 𝜑1𝜑2 < 0
⎪
0
⎩

>>> print CellVariable(mesh=Grid1D(nx=2), value=(1, 2)).minmodFaceValue
[1 1 2]
>>> print CellVariable(mesh=Grid1D(nx=2), value=(-1, -2)).minmodFaceValue
[-1 -1 -2]
>>> print CellVariable(mesh=Grid1D(nx=2), value=(-1, 2)).minmodFaceValue
[-1 0 2]
old
Return the values of the CellVariable from the previous solution sweep.
Combinations of CellVariable’s should also return old values.
>>> var1 = CellVariable(mesh = mesh, value = (2, 3), hasOld = 1)
>>> var2 = CellVariable(mesh = mesh, value = (3, 4))
>>> v = var1 * var2
>>> print v
[ 6 12]
>>> var1.value = ((3,2))
>>> print v
[9 8]
>>> print v.old
[ 6 12]
27.7. fipy.variables.cellVariable module 333

The following small test is to correct for a bug when the operator does not just use variables.
>>> v1 = var1 * 3
>>> print v1
[9 6]
>>> print v1.old
[6 9]
release(constraint)
Remove constraint from self
>>> c = Constraint(0., where=m.facesLeft)
>>> v.constrain(c)
[ 0. 1. 2. 2.5]
>>> v.release(constraint=c)
[ 0.5 1. 2. 2.5]
setValue(value, unit=None, where=None)

updateOld()
Set the values of the previous solution sweep to the current values.
>>> v = CellVariable(mesh=Grid1D(), hasOld=False)
>>> v.updateOld()
...
AssertionError: The updateOld method requires the CellVariable to have an old value. Set has
27.8 fipy.variables.cellVolumeAverageVariable module
27.9 fipy.variables.constant module
27.10 fipy.variables.constraintMask module
27.11 fipy.variables.coupledCellVariable module
27.12 fipy.variables.distanceVariable module
class fipy.variables.distanceVariable.DistanceVariable(mesh, name=’‘, value=0.0,

unit=None, hasOld=0)
Bases: fipy.variables.cellVariable.CellVariable
A DistanceVariable object calculates 𝜑 so it satisfies,
|∇𝜑| = 1
using the fast marching method with an initial condition defined by the zero level set. The solution can either be
first or second order.

Here we will define a few test cases. Firstly a 1D test case

>>> mesh = Grid1D(dx = .5, nx = 8, communicator=serialComm)
>>> from distanceVariable import DistanceVariable
>>> var = DistanceVariable(mesh = mesh, value = (-1., -1., -1., -1., 1., 1., 1., 1.))
>>> answer = (-1.75, -1.25, -.75, -0.25, 0.25, 0.75, 1.25, 1.75)
1
A 1D test case with very small dimensions.

>>> dx = 1e-10
>>> mesh = Grid1D(dx = dx, nx = 8, communicator=serialComm)
>>> answer = numerix.arange(8) * dx - 3.5 * dx
1
A 2D test case to test _calcTrialValue for a pathological case.

>>> dx = 1.
>>> dy = 2.
>>> mesh = Grid2D(dx = dx, dy = dy, nx = 2, ny = 3)
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1., 1., -1., 1.))
>>> vbl = -dx * dy / numerix.sqrt(dx**2 + dy**2) / 2.
>>> vbr = dx / 2
>>> vml = dy / 2.
>>> crossProd = dx * dy
>>> dsq = dx**2 + dy**2
>>> top = vbr * dx**2 + vml * dy**2
>>> sqrt = crossProd**2 *(dsq - (vbr - vml)**2)
>>> sqrt = numerix.sqrt(max(sqrt, 0))
>>> vmr = (top + sqrt) / dsq
>>> answer = (vbl, vbr, vml, vmr, vbl, vbr)
1
The extendVariable method solves the following equation for a given extensionVariable.
∇𝑢 · ∇𝜑 = 0
using the fast marching method with an initial condition defined at the zero level set.
>>> mesh = Grid2D(dx = 1., dy = 1., nx = 2, ny = 2, communicator=serialComm)
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1., 1.))
>>> extensionVar = CellVariable(mesh = mesh, value = (-1, .5, 2, -1))
>>> tmp = 1 / numerix.sqrt(2)
>>> print var.allclose((-tmp / 2, 0.5, 0.5, 0.5 + tmp))
1
>>> var.extendVariable(extensionVar, order=1)
>>> print extensionVar.allclose((1.25, .5, 2, 1.25))
27.12. fipy.variables.distanceVariable module 335

1
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1.,
... 1., 1., 1.,
... 1., 1., 1.))
>>> extensionVar = CellVariable(mesh = mesh, value = (-1., .5, -1.,
... 2., -1., -1.,
... -1., -1., -1.))
>>> v1 = 0.5 + tmp

>>> v2 = 1.5
>>> tmp1 = (v1 + v2) / 2 + numerix.sqrt(2. - (v1 - v2)**2) / 2
>>> tmp2 = tmp1 + 1 / numerix.sqrt(2)
>>> print var.allclose((-tmp / 2, 0.5, 1.5, 0.5, 0.5 + tmp,
... tmp1, 1.5, tmp1, tmp2))
1
>>> answer = (1.25, .5, .5, 2, 1.25, 0.9544, 2, 1.5456, 1.25)
>>> print extensionVar.allclose(answer, rtol = 1e-4)
1
Test case for a bug that occurs when initializing the distance variable at the interface. Currently it is assumed
that adjacent cells that are opposite sign neighbors have perpendicular normal vectors. In fact the two closest
cells could have opposite normals.
>>> mesh = Grid1D(dx = 1., nx = 3)
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., -1.))
>>> print var.allclose((-0.5, 0.5, -0.5))
1
Testing second order. This example failed with Scikit-fmm.

>>> var = DistanceVariable(mesh = mesh, value = (-1., -1., 1., 1.,
... -1., -1., 1., 1.,
... 1., 1., 1., 1.,
... 1, 1, 1, 1))
>>> answer = [-1.30473785, -0.5, 0.5, 1.49923009,
... -0.5, -0.35355339, 0.5, 1.45118446,
... 0.5, 0.5, 0.97140452, 1.76215286,
... 1.49923009, 1.45118446, 1.76215286, 2.33721352]
>>> print numerix.allclose(var, answer, rtol=1e-9)
True
** A test for a bug in both LSMLIB and Scikit-fmm **

The following test gives different result depending on whether LSMLIB or Scikit-fmm is used. There is a deeper
problem that is related to this issue. When a value becomes “known” after previously being a “trial” value it
updates its neighbors’ values. In a second order scheme the neighbors one step away also need to be updated
(if the in between cell is “known” and the far cell is a “trial” cell), but are not in either package. By luck (due
to trial values having the same value), the values calculated in Scikit-fmm for the following example are correct
although an example that didn’t work for Scikit-fmm could also be constructed.
>>> var = DistanceVariable(mesh = mesh, value = (-1., -1., -1., -1.,
... 1., 1., -1., -1.,

... 1., 1., -1., -1.,

... 1., 1., -1., -1.))
>>> answer = [-0.5, -0.58578644, -1.08578644, -1.85136395,
... 0.5, 0.29289322, -0.58578644, -1.54389939,
... 1.30473785, 0.5, -0.5, -1.5,
... 1.49547948, 0.5, -0.5, -1.5]
The 3rd and 7th element are different for LSMLIB. This is because the 15th element is not “known” when the
“trial” value for the 7th element is calculated. Scikit-fmm calculates the values in a slightly different order so
gets a seemingly better answer, but this is just chance.
True
Creates a distanceVariable object.

Parameters
• mesh: The mesh that defines the geometry of this variable.
• name: The name of the variable.
• value: The initial value.
• unit: the physical units of the variable
• hasOld: Whether the variable maintains an old value.
calcDistanceFunction(order=2)
Calculates the distanceVariable as a distance function.
Parameters
• order: The order of accuracy for the distance funtion calculation, either 1 or 2.
cellInterfaceAreas
Returns the length of the interface that crosses the cell
A simple 1D test:
>>> mesh = Grid1D(dx = 1., nx = 4)
>>> distanceVariable = DistanceVariable(mesh = mesh,
... value = (-1.5, -0.5, 0.5, 1.5))
>>> answer = CellVariable(mesh=mesh, value=(0, 0., 1., 0))
>>> print numerix.allclose(distanceVariable.cellInterfaceAreas,
... answer)
True
A 2D test case:
... value = (1.5, 0.5, 1.5,
... 0.5,-0.5, 0.5,
... 1.5, 0.5, 1.5))
>>> answer = CellVariable(mesh=mesh,
... value=(0, 1, 0, 1, 0, 1, 0, 1, 0))
27.12. fipy.variables.distanceVariable module 337

>>> print numerix.allclose(distanceVariable.cellInterfaceAreas, answer)

True
Another 2D test case:

>>> mesh = Grid2D(dx = .5, dy = .5, nx = 2, ny = 2)
... value = (-0.5, 0.5, 0.5, 1.5))
... value=(0, numerix.sqrt(2) / 4, numerix.sqrt(2) / 4, 0))
... answer)
True
Test to check that the circumfrence of a circle is, in fact, 2𝜋𝑟.

>>> mesh = Grid2D(dx = 0.05, dy = 0.05, nx = 20, ny = 20)
>>> r = 0.25
>>> rad = numerix.sqrt((x - .5)**2 + (y - .5)**2) - r
>>> distanceVariable = DistanceVariable(mesh = mesh, value = rad)
>>> print numerix.allclose(distanceVariable.cellInterfaceAreas.sum(), 1.57984690073)
1
extendVariable(extensionVariable, order=2)
Calculates the extension of extensionVariable from the zero level set.
Parameters
• extensionVariable: The variable to extend from the zero level set.
getCellInterfaceAreas(*args, **kwds)
Deprecated since version 3.0: use the cellInterfaceAreas property instead
getLSMshape()
27.13 fipy.variables.exponentialNoiseVariable module
class fipy.variables.exponentialNoiseVariable.ExponentialNoiseVariable(mesh,
mean=0.0,
name=’‘,
ha-
sOld=0)
Represents an exponential distribution of random numbers with the probability distribution
𝑥
𝜇−1 𝑒− 𝜇
with a mean parameter 𝜇.



>>> mean = Variable()
>>> noise = ExponentialNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), mean = mean)


>>> expdist = CellVariable(mesh = histogram.mesh)
>>> x = histogram.mesh.cellCenters[0]
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=(histogram, expdist),
>>> from fipy.tools.numerix import arange, exp
>>> for mu in arange(0.5,3,0.5):

... mean.value = (mu)
... expdist.value = ((1/mean)*exp(-x/mean))
... if __name__ == '__main__':
... import sys
... print >>sys.stderr, "mean: %g" % mean
... viewer.plot()

1
0.6
0.5
0.4
m = 1.5
0.3
0.2
0.1
0.0
2 4 6 8
Parameters
• mean: The mean of the distribution 𝜇.
27.13. fipy.variables.exponentialNoiseVariable module 339

random()
27.14 fipy.variables.faceGradContributionsVariable module
27.15 fipy.variables.faceGradVariable module
27.16 fipy.variables.faceVariable module
class fipy.variables.faceVariable.FaceVariable(mesh, name=’‘, value=0.0, rank=None, el-

ementshape=None, unit=None, cached=1)
Parameters
• mesh: the mesh that defines the geometry of this Variable
• name: the user-readable name of the Variable
• value: the initial value
• rank: the rank (number of dimensions) of each element of this Variable. Default: 0
• elementshape: the shape of each element of this variable Default: rank * (mesh.dim,)
• unit: the physical units of the Variable
copy()
divergence

>>> var = CellVariable(mesh=mesh, value=range(3*2))
>>> print var.faceGrad.divergence
[ 4. 3. 2. -2. -3. -4.]
getDivergence(*args, **kwds)
Deprecated since version 3.0: use the divergence property instead
globalValue
27.17 fipy.variables.gammaNoiseVariable module
class fipy.variables.gammaNoiseVariable.GammaNoiseVariable(mesh, shape, rate,

name=’‘, hasOld=0)
Represents a gamma distribution of random numbers with the probability distribution
𝑥𝛼−1
Γ(𝛼)
∫︀ ∞


>>> noise = GammaNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), shape = alpha, rate = beta)


>>> gammadist = (x**(alpha - 1) * (beta**alpha * exp(-beta * x)) / Gamma(alpha))
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=(histogram, gammadist),
>>> for shape in arange(1,8,1):

... alpha.value = shape
... for rate in arange(0.5,2.5,0.5):
... beta.value = rate
... if __name__ == '__main__':
... import sys
... viewer.plot()

1
27.17. fipy.variables.gammaNoiseVariable module 341

0.30
0.25
0.20
a=4
0.15
b=1
0.10
0.05
0.00
0 5 10 15 20
Parameters
• shape: The shape parameter, 𝛼.
• rate: The rate or inverse scale parameter, 𝛽.
random()
27.18 fipy.variables.gaussCellGradVariable module
27.19 fipy.variables.gaussianNoiseVariable module
class fipy.variables.gaussianNoiseVariable.GaussianNoiseVariable(mesh, name=’‘,

mean=0.0,
variance=1.0,
hasOld=0)
Represents a normal (Gaussian) distribution of random numbers with mean 𝜇 and variance ⟨𝜂(⃗𝑟)𝜂(⃗𝑟 ′ )⟩ = 𝜎 2 ,
which has the probability distribution
1 (𝑥 − 𝜇)2
√ exp −
𝜎 2𝜋 2𝜎 2
For example, the variance of thermal noise that is uncorrelated in space and time is often expressed as
⟨𝜂(⃗𝑟, 𝑡)𝜂(⃗𝑟 ′ , 𝑡′ )⟩ = 𝑀 𝑘𝐵 𝑇 𝛿(⃗𝑟 − ⃗𝑟 ′ )𝛿(𝑡 − 𝑡′ )
which can be obtained with:

sigmaSqrd = Mobility * kBoltzmann * Temperature / (mesh.cellVolumes * timeStep)
GaussianNoiseVariable(mesh = mesh, variance = sigmaSqrd)
Note: If the time step will change as the simulation progresses, either through use of an adaptive iterator or by
making manual changes at different stages, remember to declare timeStep as a Variable and to change its value
with its setValue() method.
>>> import sys

>>> from fipy.tools.numerix import *
>>> mean = 0.
>>> variance = 4.

We generate noise on a non-uniform cartesian mesh with cell dimensions of 𝑥2 and 𝑦 3 .

>>> mesh = Grid2D(dx = arange(0.1, 5., 0.1)**2, dy = arange(0.1, 3., 0.1)**3)
>>> volumes = CellVariable(mesh=mesh,value=mesh.cellVolumes)
>>> noise = GaussianNoiseVariable(mesh = mesh, mean = mean,
... variance = variance / volumes)

>>> histogram = HistogramVariable(distribution = noise * sqrt(volumes),
... dx = 0.1, nx = 600, offset = -30)

>>> gauss = CellVariable(mesh = histogram.mesh)
>>> gauss.value = ((1/(sqrt(variance * 2 * pi))) * exp(-(x - mean)**2 / (2 * variance)))
>>> if __name__ == '__main__':

... from fipy.viewers import Viewer
... viewer = Viewer(vars=noise,
... datamin=-5, datamax=5)
... histoplot = Viewer(vars=(histogram, gauss))

... noise.scramble()
... if __name__ == '__main__':
... viewer.plot()

1
Note that the noise exhibits larger amplitude in the small cells than in the large ones
but that the root-volume-weighted histogram is Gaussian.

0.25
0.20
0.15
s=2
0.10
0.05
0.00
-10 -5 0 5 10
Parameters
27.19. fipy.variables.gaussianNoiseVariable module 343

• mean: The mean of the noise distrubution, 𝜇.

• variance: The variance of the noise distribution, 𝜎 2 .
parallelRandom()
27.20 fipy.variables.harmonicCellToFaceVariable module
27.21 fipy.variables.histogramVariable module
class fipy.variables.histogramVariable.HistogramVariable(distribution, dx=1.0,

nx=None, offset=0.0)
Produces a histogram of the values of the supplied distribution.
Parameters
• distribution: The collection of values to sample.
• dx: the bin size
• nx: the number of bins
• offset: the position of the first bin

27.22 fipy.variables.interfaceAreaVariable module
27.23 fipy.variables.interfaceFlagVariable module
27.24 fipy.variables.leastSquaresCellGradVariable module
27.25 fipy.variables.levelSetDiffusionVariable module
27.26 fipy.variables.meshVariable module
27.27 fipy.variables.minmodCellToFaceVariable module
27.28 fipy.variables.modCellGradVariable module
27.29 fipy.variables.modCellToFaceVariable module
27.30 fipy.variables.modFaceGradVariable module
27.31 fipy.variables.modPhysicalField module
27.32 fipy.variables.modularVariable module
class fipy.variables.modularVariable.ModularVariable(mesh, name=’‘, value=0.0,

rank=None, elementshape=None,
The ModularVariable defines a variable that exisits on the circle between −𝜋 and 𝜋
The following examples show how ModularVariable works. When subtracting the answer wraps back around
the circle.
>>> pi = numerix.pi
>>> v1 = ModularVariable(mesh = mesh, value = (2*pi/3, -2*pi/3))
>>> v2 = ModularVariable(mesh = mesh, value = -2*pi/3)
>>> print numerix.allclose(v2 - v1, (2*pi/3, 0))
1
Obtaining the arithmetic face value.

>>> print numerix.allclose(v1.arithmeticFaceValue, (2*pi/3, pi, -2*pi/3))
1
Obtaining the gradient.
27.22. fipy.variables.interfaceAreaVariable module 345

>>> print numerix.allclose(v1.grad, ((pi/3, pi/3),))

1
Obtaining the gradient at the faces.

>>> print numerix.allclose(v1.faceGrad, ((0, 2*pi/3, 0),))
1
Obtaining the gradient at the faces but without modular arithmetic.

>>> print numerix.allclose(v1.faceGradNoMod, ((0, -4*pi/3, 0),))
1
arithmeticFaceValue
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12
Adjusted for a ModularVariable
faceGrad
Return ∇𝜑 as a rank-1 FaceVariable (second-order gradient). Adjusted for a ModularVariable
faceGradNoMod
Return ∇𝜑 as a rank-1 FaceVariable (second-order gradient). Not adjusted for a ModularVariable
getFaceGradNoMod(*args, **kwds)
Deprecated since version 3.0: use the faceGradNoMod property instead
grad
Return ∇𝜑 as a rank-1 CellVariable (first-order gradient). Adjusted for a ModularVariable
updateOld()
Set the values of the previous solution sweep to the current values. Test case due to bug.
>>> var = ModularVariable(mesh=mesh, value=1., hasOld=1)
>>> var.updateOld()
>>> var[:] = 2
>>> answer = CellVariable(mesh=mesh, value=1.)
>>> print var.old.allclose(answer)
True
27.33 fipy.variables.noiseVariable module
class fipy.variables.noiseVariable.NoiseVariable(mesh, name=’‘, hasOld=0)

A generic base class for sources of noise distributed over the cells of a mesh.
In the event that the noise should be conserved, use:
<Specific>NoiseVariable(...).faceGrad.divergence

The seed() and get_seed() functions of the fipy.tools.numerix.random module can be set and query the random
number generated used by all NoiseVariable objects.
copy()
Copy the value of the NoiseVariable to a static CellVariable.
parallelRandom()
random()
scramble()
Generate a new random distribution.
27.34 fipy.variables.operatorVariable module
27.35 fipy.variables.scharfetterGummelFaceVariable module
class fipy.variables.scharfetterGummelFaceVariable.ScharfetterGummelFaceVariable(var,
bound-
aryCon-
di-
tions=())
Bases: fipy.variables.cellToFaceVariable._CellToFaceVariable
27.36 fipy.variables.surfactantConvectionVariable module
class fipy.variables.surfactantConvectionVariable.SurfactantConvectionVariable(distanceVar)
Bases: fipy.variables.faceVariable.FaceVariable
Convection coefficient for the ConservativeSurfactantEquation. The coeff only has a value for a negative dis-
tanceVar.
Simple one dimensional test:
>>> mesh = Grid2D(nx = 3, ny = 1, dx = 1., dy = 1.)
>>> from fipy.variables.distanceVariable import DistanceVariable
>>> distanceVar = DistanceVariable(mesh, value = (-.5, .5, 1.5))
>>> ## answer = numerix.zeros((2, mesh.numberOfFaces),'d')
>>> answer = FaceVariable(mesh=mesh, rank=1, value=0.).globalValue
>>> answer[0,7] = -1
>>> print numerix.allclose(SurfactantConvectionVariable(distanceVar).globalValue, answer)
True
Change the dimensions:

>>> distanceVar = DistanceVariable(mesh, value = (-.25, .25, .75))
>>> answer[0,7] = -.5
True
Two dimensional example:
27.34. fipy.variables.operatorVariable module 347


>>> distanceVar = DistanceVariable(mesh, value = (-1.5, -.5, -.5, .5))
>>> answer[1,2] = -.5
>>> answer[1,3] = -1
>>> answer[0,7] = -.5
>>> answer[0,10] = -1
True
Larger grid:
>>> distanceVar = DistanceVariable(mesh, value = (1.5, .5 , 1.5,
... .5 , -.5, .5 ,
... 1.5, .5 , 1.5))
>>> answer[1,4] = .25
>>> answer[1,7] = -.25
>>> answer[0,17] = .25
>>> answer[0,18] = -.25
True
27.37 fipy.variables.surfactantVariable module
class fipy.variables.surfactantVariable.SurfactantVariable(value=0.0, dis-

tanceVar=None,
name=’surfactant vari-
able’, hasOld=False)
The SurfactantVariable maintains a conserved volumetric concentration on cells adjacent to, but in front of,
the interface. The value argument corresponds to the initial concentration of surfactant on the interface (moles
divided by area). The value held by the SurfactantVariable is actually a volume density (moles divided by
volume).
A simple 1D test:
>>> mesh = Grid1D(dx = 1., nx = 4)
... value = (-1.5, -0.5, 0.5, 941.5))
>>> surfactantVariable = SurfactantVariable(value = 1,
... distanceVar = distanceVariable)
>>> print numerix.allclose(surfactantVariable, (0, 0., 1., 0))
1
A 2D test case:
... value = (1.5, 0.5, 1.5,
... 0.5,-0.5, 0.5,
... 1.5, 0.5, 1.5))


>>> print numerix.allclose(surfactantVariable, (0, 1, 0, 1, 0, 1, 0, 1, 0))
1

... value = (-0.5, 0.5, 0.5, 1.5))
>>> print numerix.allclose(surfactantVariable,
... (0, numerix.sqrt(2), numerix.sqrt(2), 0))
1
Parameters
• distanceVar: A DistanceVariable object.
copy()
getInterfaceVar(*args, **kwds)
Deprecated since version 3.0: use the interfaceVar property instead
interfaceVar
Returns the SurfactantVariable rendered as an _InterfaceSurfactantVariable which evaluates the surfactant
concentration as an area concentration the interface rather than a volumetric concentration.
27.38 fipy.variables.test module
27.39 fipy.variables.unaryOperatorVariable module
27.40 fipy.variables.uniformNoiseVariable module
class fipy.variables.uniformNoiseVariable.UniformNoiseVariable(mesh, name=’‘,

minimum=0.0,
maximum=1.0,
hasOld=0)
Represents a uniform distribution of random numbers.
>>> noise = UniformNoiseVariable(mesh=Grid2D(nx=100, ny=100))
and histogram the noise
27.38. fipy.variables.test module 349


>>> histogram = HistogramVariable(distribution=noise, dx=0.01, nx=120, offset=-.1)
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=histogram)

... if __name__ == '__main__':
... viewer.plot()
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Parameters
• minimum: The minimum (not-inclusive) value of the distribution.
• maximum: The maximum (not-inclusive) value of the distribution.
random()
27.41 fipy.variables.variable module
class fipy.variables.variable.Variable(value=0.0, unit=None, array=None, name=’‘,

cached=1)
Bases: object
Lazily evaluated quantity with units.
Using a Variable in a mathematical expression will create an automatic dependency Variable, e.g.,
>>> a = Variable(value=3)
>>> b = 4 * a
>>> b
(Variable(value=array(3)) * 4)
>>> b()
12

Changes to the value of a Variable will automatically trigger changes in any dependent Variable objects
>>> a.setValue(5)
>>> b
>>> print b()
20
Create a Variable.
>>> Variable(value=3)
Variable(value=array(3))
>>> Variable(value=3, unit="m")
Variable(value=PhysicalField(3,'m'))
>>> Variable(value=3, unit="m", array=numerix.zeros((3,2), 'l'))
Variable(value=PhysicalField(array([[3, 3],
[3, 3],
[3, 3]]),'m'))
Parameters
• array: the storage array for the Variable
• cached: whether to cache or always recalculate the value
all(axis=None)
>>> print Variable(value=(0, 0, 1, 1)).all()

0
1
allclose(other, rtol=1e-05, atol=1e-08)
>>> var = Variable((1, 1))

>>> print var.allclose((1, 1))
1
>>> print var.allclose((1,))
1
The following test is to check that the system does not run out of memory.
>>> var = Variable(numerix.ones(10000))
>>> print var.allclose(numerix.zeros(10000, 'l'))
False
allequal(other)
any(axis=None)
27.41. fipy.variables.variable module 351

>>> print Variable(value=0).any()

0
>>> print Variable(value=(0, 0, 1, 1)).any()
1
arccos(*args, **kwds)
Deprecated since version 3.0: use numerix.arccos() instead
arccosh(*args, **kwds)
Deprecated since version 3.0: use numerix.arccosh() instead
arcsin(*args, **kwds)
Deprecated since version 3.0: use numerix.arcsin() instead
arcsinh(*args, **kwds)
Deprecated since version 3.0: use numerix.arcsinh() instead
arctan(*args, **kwds)
Deprecated since version 3.0: use numerix.arctan() instead
arctan2(*args, **kwds)
Deprecated since version 3.0: use numerix.arctan2() instead
arctanh(*args, **kwds)
Deprecated since version 3.0: use numerix.arctanh() instead
cacheMe(recursive=False)
ceil(*args, **kwds)
Deprecated since version 3.0: use numerix.ceil() instead
conjugate(*args, **kwds)
Deprecated since version 3.0: use numerix.conjugate() instead
Constrain the Variable to have a value at an index or mask location specified by where.
>>> v = Variable((0,1,2,3))
>>> v.constrain(2, numerix.array((True, False, False, False)))
>>> print v
[2 1 2 3]
>>> v[:] = 10
>>> print v
[ 2 10 10 10]
>>> v.constrain(5, numerix.array((False, False, True, False)))
>>> print v
[ 2 10 5 10]
>>> v[:] = 6
>>> print v
[2 6 5 6]
>>> v.constrain(8)
>>> print v
[8 8 8 8]
>>> v[:] = 10
>>> print v
[8 8 8 8]
>>> del v.constraints[2]
>>> print v
[ 2 10 5 10]


>>> m = Grid2D(nx=2, ny=2)
>>> x, y = m.cellCenters
>>> v = CellVariable(mesh=m, rank=1, value=(x, y))
>>> v.constrain(((0.,), (-1.,)), where=m.facesLeft)
[[ 0.5 1.5 0.5 1.5 0.5 1.5 0. 1. 1.5 0. 1. 1.5]
[ 0.5 0.5 1. 1. 1.5 1.5 -1. 0.5 0.5 -1. 1.5 1.5]]
Parameters
• value: the value of the constraint
• where: the constraint mask or index specifying the location of the constraint
constraints
copy()
Make an duplicate of the Variable
>>> b = a.copy()
>>> b

>>> a.setValue(5)
>>> b
Check that this works for arrays.

>>> a = Variable(value=numerix.array((0,1,2)))
>>> b = a.copy()
>>> b
Variable(value=array([0, 1, 2]))
>>> a[1] = 3
>>> b
cos(*args, **kwds)
Deprecated since version 3.0: use numerix.cos() instead
cosh(*args, **kwds)
Deprecated since version 3.0: use numerix.cosh() instead
dontCacheMe(recursive=False)
dot(other, opShape=None, operatorClass=None, axis=0)
exp(*args, **kwds)
Deprecated since version 3.0: use numerix.exp() instead
floor(*args, **kwds)
Deprecated since version 3.0: use numerix.floor() instead
getMag(*args, **kwds)
Deprecated since version 3.0: use the mag property instead

getName(*args, **kwds)
Deprecated since version 3.0: use the name property instead
getSubscribedVariables(*args, **kwds)
Deprecated since version 3.0: use the subscribedVariables property instead
getValue(*args, **kwds)
Deprecated since version 3.0: use the value property instead
>>> Variable(1).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1))
True
>>> Variable(1.).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1.))
True
>>> Variable((1,1.)).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array((1., 1.)))
True
inBaseUnits()
Return the value of the Variable with all units reduced to their base SI elements.
>>> e = Variable(value="2.7 Hartree*Nav")
1
inUnitsOf(*units)
Returns one or more Variable objects that express the same physical quantity in different units. The units
are specified by strings containing their names. The units must be compatible with the unit of the object.
If one unit is specified, the return value is a single Variable.
>>> freeze = Variable('0 degC')
1
If several units are specified, the return value is a tuple of Variable instances with with one element per
>>> t = Variable(value=314159., unit='s')
... ['3.0 d', '15.0 h', '15.0 min',
... True)
1
itemset(value)
itemsize
log(*args, **kwds)
Deprecated since version 3.0: use numerix.log() instead

log10(*args, **kwds)
Deprecated since version 3.0: use numerix.log10() instead
mag
max(axis=None)
min(axis=None)
name
numericValue
put(indices, value)
ravel()
release(constraint)
>>> v = Variable((0,1,2,3))
>>> v[:] = 10
>>> from fipy.boundaryConditions.constraint import Constraint
>>> c1 = Constraint(5, numerix.array((False, False, True, False)))
>>> v.constrain(c1)
>>> v[:] = 6
>>> v.constrain(8)
>>> v[:] = 10
>>> v.release(constraint=c1)
>>> print v
[ 2 10 10 10]
reshape(*args, **kwds)
Deprecated since version 3.0: use numerix.reshape() instead
setName(*args, **kwds)
Set the value of the Variable. Can take a masked array.
>>> a = Variable((1,2,3))
>>> a.setValue(5, where=(1, 0, 1))
>>> print a
[5 2 5]
>>> b = Variable((4,5,6))
>>> a.setValue(b, where=(1, 0, 1))
>>> print a
[4 2 6]
>>> print b
[4 5 6]
>>> a.value = 3
>>> print a
[3 3 3]

>>> b = numerix.array((3,4,5))
>>> a.value = b
>>> a[:] = 1
>>> print b
[3 4 5]
>>> a.setValue((4,5,6), where=(1, 0))

....
ValueError: shape mismatch: objects cannot be broadcast to a single shape
shape
>>> Variable(value=3).shape
()
>>> Variable(value=(3,)).shape
(1,)
>>> Variable(value=(3,4)).shape
(2,)
>>> Variable(value="3 m").shape

()
>>> Variable(value=(3,), unit="m").shape
(1,)
>>> Variable(value=(3,4), unit="m").shape
(2,)
sign(*args, **kwds)
Deprecated since version 3.0: use numerix.sign() instead
sin(*args, **kwds)
Deprecated since version 3.0: use numerix.sin() instead
sinh(*args, **kwds)
Deprecated since version 3.0: use numerix.sinh() instead
sqrt(*args, **kwds)
Deprecated since version 3.0: use numerix.sqrt() instead
>>> mesh= Grid1D(nx=3)

>>> var = CellVariable(mesh=mesh, value=((0., 2., 3.),), rank=1)
>>> print (var.dot(var)).sqrt()
[ 0. 2. 3.]
subscribedVariables
sum(axis=None)
take(ids, axis=0)
tan(*args, **kwds)
Deprecated since version 3.0: use numerix.tan() instead
tanh(*args, **kwds)
Deprecated since version 3.0: use numerix.tanh() instead

unit
>>> Variable(value="1 m").unit
<PhysicalUnit m>
value
“Evaluate” the Variable and return its value (longhand)
>>> print a.value
3
>>> b = a + 4
>>> b
(Variable(value=array(3)) + 4)
>>> b.value
7
class fipy.variables.Variable(value=0.0, unit=None, array=None, name=’‘, cached=1)

Bases: object
Lazily evaluated quantity with units.
Using a Variable in a mathematical expression will create an automatic dependency Variable, e.g.,
>>> b = 4 * a
>>> b
>>> b()
12
Changes to the value of a Variable will automatically trigger changes in any dependent Variable objects
>>> a.setValue(5)
>>> b
>>> print b()
20
Create a Variable.
>>> Variable(value=3)
>>> Variable(value=3, unit="m")
Variable(value=PhysicalField(3,'m'))
>>> Variable(value=3, unit="m", array=numerix.zeros((3,2), 'l'))
Variable(value=PhysicalField(array([[3, 3],
[3, 3],
[3, 3]]),'m'))
Parameters

• array: the storage array for the Variable

• cached: whether to cache or always recalculate the value
all(axis=None)

0
1
allclose(other, rtol=1e-05, atol=1e-08)
>>> var = Variable((1, 1))

>>> print var.allclose((1, 1))
1
>>> print var.allclose((1,))
1
The following test is to check that the system does not run out of memory.
>>> var = Variable(numerix.ones(10000))
>>> print var.allclose(numerix.zeros(10000, 'l'))
False
allequal(other)
any(axis=None)
>>> print Variable(value=0).any()

0
>>> print Variable(value=(0, 0, 1, 1)).any()
1
arccos(*args, **kwds)
Deprecated since version 3.0: use numerix.arccos() instead
arccosh(*args, **kwds)
Deprecated since version 3.0: use numerix.arccosh() instead
arcsin(*args, **kwds)
Deprecated since version 3.0: use numerix.arcsin() instead
arcsinh(*args, **kwds)
Deprecated since version 3.0: use numerix.arcsinh() instead
arctan(*args, **kwds)
Deprecated since version 3.0: use numerix.arctan() instead
arctan2(*args, **kwds)
Deprecated since version 3.0: use numerix.arctan2() instead
arctanh(*args, **kwds)
Deprecated since version 3.0: use numerix.arctanh() instead
cacheMe(recursive=False)

ceil(*args, **kwds)
Deprecated since version 3.0: use numerix.ceil() instead
conjugate(*args, **kwds)
Deprecated since version 3.0: use numerix.conjugate() instead
Constrain the Variable to have a value at an index or mask location specified by where.
>>> v = Variable((0,1,2,3))
>>> print v
[2 1 2 3]
>>> v[:] = 10
>>> print v
[ 2 10 10 10]
>>> v.constrain(5, numerix.array((False, False, True, False)))
>>> print v
[ 2 10 5 10]
>>> v[:] = 6
>>> print v
[2 6 5 6]
>>> v.constrain(8)
>>> print v
[8 8 8 8]
>>> v[:] = 10
>>> print v
[8 8 8 8]
>>> print v
[ 2 10 5 10]

>>> m = Grid2D(nx=2, ny=2)
>>> x, y = m.cellCenters
>>> v = CellVariable(mesh=m, rank=1, value=(x, y))
>>> v.constrain(((0.,), (-1.,)), where=m.facesLeft)
[[ 0.5 1.5 0.5 1.5 0.5 1.5 0. 1. 1.5 0. 1. 1.5]
[ 0.5 0.5 1. 1. 1.5 1.5 -1. 0.5 0.5 -1. 1.5 1.5]]
Parameters
• value: the value of the constraint
• where: the constraint mask or index specifying the location of the constraint
constraints
copy()
Make an duplicate of the Variable
>>> b = a.copy()
>>> b

>>> a.setValue(5)
>>> b
Check that this works for arrays.

>>> a = Variable(value=numerix.array((0,1,2)))
>>> b = a.copy()
>>> b
>>> a[1] = 3
>>> b
cos(*args, **kwds)
Deprecated since version 3.0: use numerix.cos() instead
cosh(*args, **kwds)
Deprecated since version 3.0: use numerix.cosh() instead
dontCacheMe(recursive=False)
dot(other, opShape=None, operatorClass=None, axis=0)
exp(*args, **kwds)
Deprecated since version 3.0: use numerix.exp() instead
floor(*args, **kwds)
Deprecated since version 3.0: use numerix.floor() instead
getMag(*args, **kwds)
Deprecated since version 3.0: use the mag property instead
getName(*args, **kwds)
getSubscribedVariables(*args, **kwds)
Deprecated since version 3.0: use the subscribedVariables property instead
getValue(*args, **kwds)
Deprecated since version 3.0: use the value property instead
>>> Variable(1).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1))
True
>>> Variable(1.).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array(1.))
True
>>> Variable((1,1.)).getsctype() == numerix.NUMERIX.obj2sctype(numerix.array((1., 1.)))
True
inBaseUnits()
Return the value of the Variable with all units reduced to their base SI elements.

>>> e = Variable(value="2.7 Hartree*Nav")

1
inUnitsOf(*units)
Returns one or more Variable objects that express the same physical quantity in different units. The units
are specified by strings containing their names. The units must be compatible with the unit of the object.
If one unit is specified, the return value is a single Variable.
>>> freeze = Variable('0 degC')
1
If several units are specified, the return value is a tuple of Variable instances with with one element per
>>> t = Variable(value=314159., unit='s')
... ['3.0 d', '15.0 h', '15.0 min',
... True)
1
itemset(value)
itemsize
log(*args, **kwds)
Deprecated since version 3.0: use numerix.log() instead
log10(*args, **kwds)
Deprecated since version 3.0: use numerix.log10() instead
mag
max(axis=None)
min(axis=None)
name
numericValue
put(indices, value)
ravel()
release(constraint)
>>> v = Variable((0,1,2,3))
>>> v[:] = 10
>>> from fipy.boundaryConditions.constraint import Constraint
>>> c1 = Constraint(5, numerix.array((False, False, True, False)))
>>> v.constrain(c1)
>>> v[:] = 6
>>> v.constrain(8)
>>> v[:] = 10
>>> v.release(constraint=c1)

>>> print v
[ 2 10 10 10]
reshape(*args, **kwds)
Deprecated since version 3.0: use numerix.reshape() instead
setName(*args, **kwds)
Set the value of the Variable. Can take a masked array.
>>> a = Variable((1,2,3))
>>> a.setValue(5, where=(1, 0, 1))
>>> print a
[5 2 5]
>>> b = Variable((4,5,6))
>>> a.setValue(b, where=(1, 0, 1))
>>> print a
[4 2 6]
>>> print b
[4 5 6]
>>> a.value = 3
>>> print a
[3 3 3]
>>> b = numerix.array((3,4,5))
>>> a.value = b
>>> a[:] = 1
>>> print b
[3 4 5]
>>> a.setValue((4,5,6), where=(1, 0))

....
ValueError: shape mismatch: objects cannot be broadcast to a single shape
shape
>>> Variable(value=3).shape
()
>>> Variable(value=(3,)).shape
(1,)
>>> Variable(value=(3,4)).shape
(2,)
>>> Variable(value="3 m").shape

()
>>> Variable(value=(3,), unit="m").shape
(1,)
>>> Variable(value=(3,4), unit="m").shape
(2,)
sign(*args, **kwds)
Deprecated since version 3.0: use numerix.sign() instead

sin(*args, **kwds)
Deprecated since version 3.0: use numerix.sin() instead
sinh(*args, **kwds)
Deprecated since version 3.0: use numerix.sinh() instead
sqrt(*args, **kwds)
Deprecated since version 3.0: use numerix.sqrt() instead
>>> mesh= Grid1D(nx=3)

>>> var = CellVariable(mesh=mesh, value=((0., 2., 3.),), rank=1)
>>> print (var.dot(var)).sqrt()
[ 0. 2. 3.]
subscribedVariables
sum(axis=None)
take(ids, axis=0)
tan(*args, **kwds)
Deprecated since version 3.0: use numerix.tan() instead
tanh(*args, **kwds)
Deprecated since version 3.0: use numerix.tanh() instead
unit
>>> Variable(value="1 m").unit
<PhysicalUnit m>
value
“Evaluate” the Variable and return its value (longhand)
>>> print a.value
3
>>> b = a + 4
>>> b
(Variable(value=array(3)) + 4)
>>> b.value
7
class fipy.variables.CellVariable(mesh, name=’‘, value=0.0, rank=None, elementshape=None,

Represents the field of values of a variable on a Mesh.
A CellVariable can be pickled to persistent storage (disk) for later use:
>>> var = CellVariable(mesh = mesh, value = 1., hasOld = 1, name = 'test')

>>> var.value = (x * y)

>>> from fipy.tools import dump

>>> (f, filename) = dump.write(var, extension = '.gz')
>>> unPickledVar = dump.read(filename, f)
>>> print var.allclose(unPickledVar, atol = 1e-10, rtol = 1e-10)

1
arithmeticFaceValue
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1
>>> R = 2
>>> answer = (R - L) * (0.5 / 1.) + L
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = (R - L) * (1.0 / 3.0) + L
True
>>> mesh = Grid1D(dx = (10., 100.))

>>> answer = (R - L) * (5.0 / 55.0) + L
True
cellVolumeAverage
Return the cell-volume-weighted average of the CellVariable:
∑︀
cells 𝜑cell 𝑉cell
< 𝜑 >vol = ∑︀
cells 𝑉cell

>>> var = CellVariable(value = (1, 2, 6), mesh = mesh)
>>> print var.cellVolumeAverage
3.0
Constrains the CellVariable to value at a location specified by where.


>>> v.faceGrad.constrain([1.], where=m.facesRight)
[[ 1. 1. 1. 1.]]
[ 0. 1. 2. 2.5]
Changing the constraint changes the dependencies

[[-1. 1. 1. 1.]]
[ 1. 1. 2. 2.5]
Constraints can be Variable

>>> c = Variable(0.)
>>> v.constrain(c, where=m.facesLeft)
[[ 1. 1. 1. 1.]]
[ 0. 1. 2. 2.5]
>>> c.value = 1.
[[-1. 1. 1. 1.]]
[ 1. 1. 2. 2.5]
Constraints can have a Variable mask.

>>> mask = FaceVariable(mesh=m, value=m.facesLeft)
>>> v.constrain(1., where=mask)
[ 1. 0. 0. 0.]
>>> mask[:] = mask | m.facesRight
[ 1. 0. 0. 1.]
copy()
faceGrad
Return ∇𝜑 as a rank-1 FaceVariable using differencing for the normal direction(second-order gradient).
faceGradAverage
Return ∇𝜑 as a rank-1 FaceVariable using averaging for the normal direction(second-order gradient)
faceValue
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1

>>> R = 2
>>> answer = (R - L) * (0.5 / 1.) + L
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = (R - L) * (1.0 / 3.0) + L
True
>>> mesh = Grid1D(dx = (10., 100.))

>>> answer = (R - L) * (5.0 / 55.0) + L
True
gaussGrad ∑︀
Return 𝑉1𝑃 𝑓 ⃗𝑛𝜑𝑓 𝐴𝑓 as a rank-1 CellVariable (first-order gradient).
getArithmeticFaceValue(*args, **kwds)
getCellVolumeAverage(*args, **kwds)
Deprecated since version 3.0: use the cellVolumeAverage property instead
getFaceGrad(*args, **kwds)
Deprecated since version 3.0: use the faceGrad property instead
getFaceGradAverage(*args, **kwds)
Deprecated since version 3.0: use the faceGradAverage property instead
getFaceValue(*args, **kwds)
getGaussGrad(*args, **kwds)
Deprecated since version 3.0: use the gaussGrad property instead
getGrad(*args, **kwds)
Deprecated since version 3.0: use the grad property instead
getHarmonicFaceValue(*args, **kwds)
Deprecated since version 3.0: use the harmonicFaceValue property instead
getLeastSquaresGrad(*args, **kwds)
Deprecated since version 3.0: use the leastSquaresGrad property instead
getMinmodFaceValue(*args, **kwds)
Deprecated since version 3.0: use the minmodFaceValue property instead
getOld(*args, **kwds)
Deprecated since version 3.0: use the old property instead
globalValue
Concatenate and return values from all processors
When running on a single processor, the result is identical to value.

grad
Return ∇𝜑 as a rank-1 CellVariable (first-order gradient).
harmonicFaceValue
Returns a FaceVariable whose value corresponds to the harmonic interpolation of the adjacent cells:
𝜑1 𝜑2
𝜑𝑓 = 𝑑
(𝜑2 − 𝜑1 ) 𝑑𝑓122 + 𝜑1

>>> mesh = Grid1D(dx = (1., 1.))
>>> L = 1
>>> R = 2
>>> answer = L * R / ((R - L) * (0.5 / 1.) + L)
True
>>> mesh = Grid1D(dx = (2., 4.))

>>> answer = L * R / ((R - L) * (1.0 / 3.0) + L)
True
>>> mesh = Grid1D(dx = (10., 100.))

>>> answer = L * R / ((R - L) * (5.0 / 55.0) + L)
True
leastSquaresGrad
Return ∇𝜑, which is determined by solving for ∇𝜑 in the following matrix equation,
∑︁ ∑︁
∇𝜑 · 𝑑2𝐴𝑃 ⃗𝑛𝐴𝑃 ⊗ ⃗𝑛𝐴𝑃 = 𝑑2𝐴𝑃 (⃗𝑛 · ∇𝜑)𝐴𝑃
𝑓 𝑓
The matrix equation is derived by minimizing the following least squares sum,
√︃∑︁
2
𝐹 (𝜑𝑥 , 𝜑𝑦 ) = (𝑑𝐴𝑃 ⃗𝑛𝐴𝑃 · ∇𝜑 − 𝑑𝐴𝑃 (⃗𝑛𝐴𝑃 · ∇𝜑)𝐴𝑃 )
𝑓
Tests
>>> m = Grid2D(nx=2, ny=2, dx=0.1, dy=2.0)
>>> print numerix.allclose(CellVariable(mesh=m, value=(0,1,3,6)).leastSquaresGrad.globalValu
... [[8.0, 8.0, 24.0, 24.0],
... [1.2, 2.0, 1.2, 2.0]])
True

>>> print numerix.allclose(CellVariable(mesh=Grid1D(dx=(2.0, 1.0, 0.5)),
... value=(0, 1, 2)).leastSquaresGrad.globalValue, [[0.4
True

minmodFaceValue
Returns a FaceVariable with a value that is the minimum of the absolute values of the adjacent cells. If the
values are of opposite sign then the result is zero:
⎧
⎨𝜑1 when |𝜑1 | ≤ |𝜑2 |,
⎪
𝜑𝑓 = 𝜑2 when |𝜑2 | < |𝜑1 |,
when 𝜑1𝜑2 < 0
⎪
0
⎩

>>> print CellVariable(mesh=Grid1D(nx=2), value=(1, 2)).minmodFaceValue
[1 1 2]
>>> print CellVariable(mesh=Grid1D(nx=2), value=(-1, -2)).minmodFaceValue
[-1 -1 -2]
>>> print CellVariable(mesh=Grid1D(nx=2), value=(-1, 2)).minmodFaceValue
[-1 0 2]
old
Return the values of the CellVariable from the previous solution sweep.
Combinations of CellVariable’s should also return old values.
>>> var1 = CellVariable(mesh = mesh, value = (2, 3), hasOld = 1)
>>> var2 = CellVariable(mesh = mesh, value = (3, 4))
>>> v = var1 * var2
>>> print v
[ 6 12]
>>> var1.value = ((3,2))
>>> print v
[9 8]
>>> print v.old
[ 6 12]
The following small test is to correct for a bug when the operator does not just use variables.
>>> v1 = var1 * 3
>>> print v1
[9 6]
>>> print v1.old
[6 9]
release(constraint)
>>> c = Constraint(0., where=m.facesLeft)
>>> v.constrain(c)
[ 0. 1. 2. 2.5]
>>> v.release(constraint=c)
[ 0.5 1. 2. 2.5]

updateOld()
Set the values of the previous solution sweep to the current values.
>>> v = CellVariable(mesh=Grid1D(), hasOld=False)
>>> v.updateOld()
...
AssertionError: The updateOld method requires the CellVariable to have an old value. Set has
class fipy.variables.FaceVariable(mesh, name=’‘, value=0.0, rank=None, elementshape=None,

unit=None, cached=1)
Parameters
• mesh: the mesh that defines the geometry of this Variable
• rank: the rank (number of dimensions) of each element of this Variable. Default: 0
• elementshape: the shape of each element of this variable Default: rank * (mesh.dim,)
copy()
divergence

>>> var = CellVariable(mesh=mesh, value=range(3*2))
>>> print var.faceGrad.divergence
[ 4. 3. 2. -2. -3. -4.]
getDivergence(*args, **kwds)
Deprecated since version 3.0: use the divergence property instead
globalValue
class fipy.variables.ScharfetterGummelFaceVariable(var, boundaryConditions=())
Bases: fipy.variables.cellToFaceVariable._CellToFaceVariable
class fipy.variables.ModularVariable(mesh, name=’‘, value=0.0, rank=None, ele-
mentshape=None, unit=None, hasOld=0)
The ModularVariable defines a variable that exisits on the circle between −𝜋 and 𝜋
The following examples show how ModularVariable works. When subtracting the answer wraps back around
the circle.
>>> pi = numerix.pi
>>> v1 = ModularVariable(mesh = mesh, value = (2*pi/3, -2*pi/3))

>>> v2 = ModularVariable(mesh = mesh, value = -2*pi/3)

>>> print numerix.allclose(v2 - v1, (2*pi/3, 0))
1
Obtaining the arithmetic face value.

>>> print numerix.allclose(v1.arithmeticFaceValue, (2*pi/3, pi, -2*pi/3))
1
Obtaining the gradient.

>>> print numerix.allclose(v1.grad, ((pi/3, pi/3),))
1
Obtaining the gradient at the faces.

>>> print numerix.allclose(v1.faceGrad, ((0, 2*pi/3, 0),))
1
Obtaining the gradient at the faces but without modular arithmetic.

>>> print numerix.allclose(v1.faceGradNoMod, ((0, -4*pi/3, 0),))
1
arithmeticFaceValue
𝑑𝑓 2
𝜑𝑓 = (𝜑1 − 𝜑2 ) + 𝜑2
𝑑12
Adjusted for a ModularVariable
faceGrad
Return ∇𝜑 as a rank-1 FaceVariable (second-order gradient). Adjusted for a ModularVariable
faceGradNoMod
Return ∇𝜑 as a rank-1 FaceVariable (second-order gradient). Not adjusted for a ModularVariable
getFaceGradNoMod(*args, **kwds)
Deprecated since version 3.0: use the faceGradNoMod property instead
grad
Return ∇𝜑 as a rank-1 CellVariable (first-order gradient). Adjusted for a ModularVariable
updateOld()
Set the values of the previous solution sweep to the current values. Test case due to bug.
>>> var = ModularVariable(mesh=mesh, value=1., hasOld=1)
>>> var.updateOld()
>>> var[:] = 2
>>> answer = CellVariable(mesh=mesh, value=1.)
>>> print var.old.allclose(answer)
True
class fipy.variables.BetaNoiseVariable(mesh, alpha, beta, name=’‘, hasOld=0)

Represents a beta distribution of random numbers with the probability distribution
𝑥𝛼−1
Γ(𝛼)

∫︀ ∞
with a shape parameter 𝛼, a rate parameter 𝛽, and Γ(𝑧) = 0
𝑡𝑧−1 𝑒−𝑡 𝑑𝑡.

>>> noise = BetaNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), alpha = alpha, beta = beta)


>>> betadist = CellVariable(mesh = histogram.mesh)
>>> betadist = ((Gamma(alpha + beta) / (Gamma(alpha) * Gamma(beta)))
... * x**(alpha - 1) * (1 - x)**(beta - 1))
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=(histogram, betadist),
>>> for a in arange(0.5,5,0.5):

... alpha.value = a
... for b in arange(0.5,5,0.5):
... beta.value = b
... if __name__ == '__main__':
... import sys
... viewer.plot()

1

2 a = b = 0.5
0.0 0.2 0.4 0.6 0.8 1.0
Parameters
• alpha: The parameter 𝛼.
• beta: The parameter 𝛽.
random()
class fipy.variables.ExponentialNoiseVariable(mesh, mean=0.0, name=’‘, hasOld=0)
Represents an exponential distribution of random numbers with the probability distribution
𝑥
𝜇−1 𝑒− 𝜇
with a mean parameter 𝜇.


>>> mean = Variable()
>>> noise = ExponentialNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), mean = mean)


>>> expdist = CellVariable(mesh = histogram.mesh)
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=(histogram, expdist),
>>> from fipy.tools.numerix import arange, exp
>>> for mu in arange(0.5,3,0.5):

... mean.value = (mu)
... expdist.value = ((1/mean)*exp(-x/mean))
... if __name__ == '__main__':
... import sys

... print >>sys.stderr, "mean: %g" % mean

... viewer.plot()

1
0.6
0.5
0.4
m = 1.5
0.3
0.2
0.1
0.0
2 4 6 8
Parameters
• mean: The mean of the distribution 𝜇.
random()
class fipy.variables.GammaNoiseVariable(mesh, shape, rate, name=’‘, hasOld=0)
Represents a gamma distribution of random numbers with the probability distribution
𝑥𝛼−1
Γ(𝛼)
∫︀ ∞

>>> noise = GammaNoiseVariable(mesh = Grid2D(nx = 100, ny = 100), shape = alpha, rate = beta)



>>> gammadist = (x**(alpha - 1) * (beta**alpha * exp(-beta * x)) / Gamma(alpha))
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=(histogram, gammadist),
>>> for shape in arange(1,8,1):

... alpha.value = shape
... for rate in arange(0.5,2.5,0.5):
... beta.value = rate
... if __name__ == '__main__':
... import sys
... viewer.plot()

1
0.30
0.25
0.20
a=4
0.15
b=1
0.10
0.05
0.00
0 5 10 15 20
Parameters
• shape: The shape parameter, 𝛼.
• rate: The rate or inverse scale parameter, 𝛽.
random()
class fipy.variables.GaussianNoiseVariable(mesh, name=’‘, mean=0.0, variance=1.0, ha-
sOld=0)

Represents a normal (Gaussian) distribution of random numbers with mean 𝜇 and variance ⟨𝜂(⃗𝑟)𝜂(⃗𝑟 ′ )⟩ = 𝜎 2 ,
which has the probability distribution
1 (𝑥 − 𝜇)2
√ exp −
𝜎 2𝜋 2𝜎 2
For example, the variance of thermal noise that is uncorrelated in space and time is often expressed as
⟨𝜂(⃗𝑟, 𝑡)𝜂(⃗𝑟 ′ , 𝑡′ )⟩ = 𝑀 𝑘𝐵 𝑇 𝛿(⃗𝑟 − ⃗𝑟 ′ )𝛿(𝑡 − 𝑡′ )
which can be obtained with:

sigmaSqrd = Mobility * kBoltzmann * Temperature / (mesh.cellVolumes * timeStep)
GaussianNoiseVariable(mesh = mesh, variance = sigmaSqrd)
Note: If the time step will change as the simulation progresses, either through use of an adaptive iterator or by
making manual changes at different stages, remember to declare timeStep as a Variable and to change its value
with its setValue() method.
>>> import sys

>>> from fipy.tools.numerix import *
>>> mean = 0.
>>> variance = 4.
We generate noise on a non-uniform cartesian mesh with cell dimensions of 𝑥2 and 𝑦 3 .

>>> mesh = Grid2D(dx = arange(0.1, 5., 0.1)**2, dy = arange(0.1, 3., 0.1)**3)
>>> volumes = CellVariable(mesh=mesh,value=mesh.cellVolumes)
>>> noise = GaussianNoiseVariable(mesh = mesh, mean = mean,
... variance = variance / volumes)

>>> histogram = HistogramVariable(distribution = noise * sqrt(volumes),
... dx = 0.1, nx = 600, offset = -30)

>>> gauss = CellVariable(mesh = histogram.mesh)
>>> gauss.value = ((1/(sqrt(variance * 2 * pi))) * exp(-(x - mean)**2 / (2 * variance)))
>>> if __name__ == '__main__':

... from fipy.viewers import Viewer
... datamin=-5, datamax=5)
... histoplot = Viewer(vars=(histogram, gauss))


... if __name__ == '__main__':
... viewer.plot()

1
Note that the noise exhibits larger amplitude in the small cells than in the large ones
but that the root-volume-weighted histogram is Gaussian.

0.25
0.20
0.15
s=2
0.10
0.05
0.00
-10 -5 0 5 10
Parameters
• mean: The mean of the noise distrubution, 𝜇.
• variance: The variance of the noise distribution, 𝜎 2 .
parallelRandom()
class fipy.variables.UniformNoiseVariable(mesh, name=’‘, minimum=0.0, maximum=1.0, ha-
sOld=0)
Represents a uniform distribution of random numbers.
>>> noise = UniformNoiseVariable(mesh=Grid2D(nx=100, ny=100))
and histogram the noise

>>> histogram = HistogramVariable(distribution=noise, dx=0.01, nx=120, offset=-.1)
>>> if __name__ == '__main__':

... histoplot = Viewer(vars=histogram)


... if __name__ == '__main__':
... viewer.plot()
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Parameters
• minimum: The minimum (not-inclusive) value of the distribution.
• maximum: The maximum (not-inclusive) value of the distribution.
random()
class fipy.variables.HistogramVariable(distribution, dx=1.0, nx=None, offset=0.0)
Produces a histogram of the values of the supplied distribution.
Parameters
• distribution: The collection of values to sample.
• dx: the bin size
• nx: the number of bins
• offset: the position of the first bin
class fipy.variables.SurfactantVariable(value=0.0, distanceVar=None, name=’surfactant
variable’, hasOld=False)
The SurfactantVariable maintains a conserved volumetric concentration on cells adjacent to, but in front of,
the interface. The value argument corresponds to the initial concentration of surfactant on the interface (moles
divided by area). The value held by the SurfactantVariable is actually a volume density (moles divided by
volume).
A simple 1D test:
>>> mesh = Grid1D(dx = 1., nx = 4)


... value = (-1.5, -0.5, 0.5, 941.5))
>>> print numerix.allclose(surfactantVariable, (0, 0., 1., 0))
1
A 2D test case:
... value = (1.5, 0.5, 1.5,
... 0.5,-0.5, 0.5,
... 1.5, 0.5, 1.5))
>>> print numerix.allclose(surfactantVariable, (0, 1, 0, 1, 0, 1, 0, 1, 0))
1

... value = (-0.5, 0.5, 0.5, 1.5))
>>> print numerix.allclose(surfactantVariable,
... (0, numerix.sqrt(2), numerix.sqrt(2), 0))
1
Parameters
• distanceVar: A DistanceVariable object.
copy()
getInterfaceVar(*args, **kwds)
Deprecated since version 3.0: use the interfaceVar property instead
interfaceVar
Returns the SurfactantVariable rendered as an _InterfaceSurfactantVariable which evaluates the surfactant
concentration as an area concentration the interface rather than a volumetric concentration.
class fipy.variables.SurfactantConvectionVariable(distanceVar)
Bases: fipy.variables.faceVariable.FaceVariable
Convection coefficient for the ConservativeSurfactantEquation. The coeff only has a value for a negative dis-
tanceVar.
Simple one dimensional test:
>>> distanceVar = DistanceVariable(mesh, value = (-.5, .5, 1.5))

>>> ## answer = numerix.zeros((2, mesh.numberOfFaces),'d')

>>> answer[0,7] = -1
True
Change the dimensions:

>>> distanceVar = DistanceVariable(mesh, value = (-.25, .25, .75))
>>> answer[0,7] = -.5
True
Two dimensional example:

>>> distanceVar = DistanceVariable(mesh, value = (-1.5, -.5, -.5, .5))
>>> answer[1,2] = -.5
>>> answer[1,3] = -1
>>> answer[0,7] = -.5
>>> answer[0,10] = -1
True
Larger grid:
>>> distanceVar = DistanceVariable(mesh, value = (1.5, .5 , 1.5,
... .5 , -.5, .5 ,
... 1.5, .5 , 1.5))
>>> answer[1,4] = .25
>>> answer[1,7] = -.25
>>> answer[0,17] = .25
>>> answer[0,18] = -.25
True
class fipy.variables.DistanceVariable(mesh, name=’‘, value=0.0, unit=None, hasOld=0)

A DistanceVariable object calculates 𝜑 so it satisfies,
|∇𝜑| = 1
using the fast marching method with an initial condition defined by the zero level set. The solution can either be
first or second order.
Here we will define a few test cases. Firstly a 1D test case
>>> mesh = Grid1D(dx = .5, nx = 8, communicator=serialComm)
>>> from distanceVariable import DistanceVariable
>>> answer = (-1.75, -1.25, -.75, -0.25, 0.25, 0.75, 1.25, 1.75)
1

A 1D test case with very small dimensions.

>>> dx = 1e-10
>>> mesh = Grid1D(dx = dx, nx = 8, communicator=serialComm)
>>> answer = numerix.arange(8) * dx - 3.5 * dx
1
A 2D test case to test _calcTrialValue for a pathological case.

>>> dx = 1.
>>> dy = 2.
>>> mesh = Grid2D(dx = dx, dy = dy, nx = 2, ny = 3)
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1., 1., -1., 1.))
>>> vbl = -dx * dy / numerix.sqrt(dx**2 + dy**2) / 2.
>>> vbr = dx / 2
>>> vml = dy / 2.
>>> crossProd = dx * dy
>>> dsq = dx**2 + dy**2
>>> top = vbr * dx**2 + vml * dy**2
>>> sqrt = crossProd**2 *(dsq - (vbr - vml)**2)
>>> sqrt = numerix.sqrt(max(sqrt, 0))
>>> vmr = (top + sqrt) / dsq
>>> answer = (vbl, vbr, vml, vmr, vbl, vbr)
1
The extendVariable method solves the following equation for a given extensionVariable.
∇𝑢 · ∇𝜑 = 0
using the fast marching method with an initial condition defined at the zero level set.
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1., 1.))
>>> extensionVar = CellVariable(mesh = mesh, value = (-1, .5, 2, -1))
>>> tmp = 1 / numerix.sqrt(2)
>>> print var.allclose((-tmp / 2, 0.5, 0.5, 0.5 + tmp))
1
>>> print extensionVar.allclose((1.25, .5, 2, 1.25))
1
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., 1.,
... 1., 1., 1.,
... 1., 1., 1.))
>>> extensionVar = CellVariable(mesh = mesh, value = (-1., .5, -1.,
... 2., -1., -1.,
... -1., -1., -1.))

>>> v1 = 0.5 + tmp

>>> v2 = 1.5
>>> tmp1 = (v1 + v2) / 2 + numerix.sqrt(2. - (v1 - v2)**2) / 2
>>> tmp2 = tmp1 + 1 / numerix.sqrt(2)
>>> print var.allclose((-tmp / 2, 0.5, 1.5, 0.5, 0.5 + tmp,
... tmp1, 1.5, tmp1, tmp2))
1
>>> answer = (1.25, .5, .5, 2, 1.25, 0.9544, 2, 1.5456, 1.25)
>>> print extensionVar.allclose(answer, rtol = 1e-4)
1
Test case for a bug that occurs when initializing the distance variable at the interface. Currently it is assumed
that adjacent cells that are opposite sign neighbors have perpendicular normal vectors. In fact the two closest
cells could have opposite normals.
>>> mesh = Grid1D(dx = 1., nx = 3)
>>> var = DistanceVariable(mesh = mesh, value = (-1., 1., -1.))
>>> print var.allclose((-0.5, 0.5, -0.5))
1
Testing second order. This example failed with Scikit-fmm.

>>> var = DistanceVariable(mesh = mesh, value = (-1., -1., 1., 1.,
... -1., -1., 1., 1.,
... 1., 1., 1., 1.,
... 1, 1, 1, 1))
>>> answer = [-1.30473785, -0.5, 0.5, 1.49923009,
... -0.5, -0.35355339, 0.5, 1.45118446,
... 0.5, 0.5, 0.97140452, 1.76215286,
... 1.49923009, 1.45118446, 1.76215286, 2.33721352]
True
** A test for a bug in both LSMLIB and Scikit-fmm **

The following test gives different result depending on whether LSMLIB or Scikit-fmm is used. There is a deeper
problem that is related to this issue. When a value becomes “known” after previously being a “trial” value it
updates its neighbors’ values. In a second order scheme the neighbors one step away also need to be updated
(if the in between cell is “known” and the far cell is a “trial” cell), but are not in either package. By luck (due
to trial values having the same value), the values calculated in Scikit-fmm for the following example are correct
although an example that didn’t work for Scikit-fmm could also be constructed.
>>> var = DistanceVariable(mesh = mesh, value = (-1., -1., -1., -1.,
... 1., 1., -1., -1.,
... 1., 1., -1., -1.,
... 1., 1., -1., -1.))
>>> answer = [-0.5, -0.58578644, -1.08578644, -1.85136395,
... 0.5, 0.29289322, -0.58578644, -1.54389939,
... 1.30473785, 0.5, -0.5, -1.5,
... 1.49547948, 0.5, -0.5, -1.5]
The 3rd and 7th element are different for LSMLIB. This is because the 15th element is not “known” when the

“trial” value for the 7th element is calculated. Scikit-fmm calculates the values in a slightly different order so
gets a seemingly better answer, but this is just chance.
True
Creates a distanceVariable object.

Parameters
• mesh: The mesh that defines the geometry of this variable.
• unit: the physical units of the variable
• hasOld: Whether the variable maintains an old value.
calcDistanceFunction(order=2)
Calculates the distanceVariable as a distance function.
Parameters
• order: The order of accuracy for the distance funtion calculation, either 1 or 2.
cellInterfaceAreas
Returns the length of the interface that crosses the cell
A simple 1D test:
>>> mesh = Grid1D(dx = 1., nx = 4)
... value = (-1.5, -0.5, 0.5, 1.5))
>>> answer = CellVariable(mesh=mesh, value=(0, 0., 1., 0))
... answer)
True
A 2D test case:
... value = (1.5, 0.5, 1.5,
... 0.5,-0.5, 0.5,
... 1.5, 0.5, 1.5))
... value=(0, 1, 0, 1, 0, 1, 0, 1, 0))
>>> print numerix.allclose(distanceVariable.cellInterfaceAreas, answer)
True

... value = (-0.5, 0.5, 0.5, 1.5))
... value=(0, numerix.sqrt(2) / 4, numerix.sqrt(2) / 4, 0))

... answer)
True
Test to check that the circumfrence of a circle is, in fact, 2𝜋𝑟.

>>> mesh = Grid2D(dx = 0.05, dy = 0.05, nx = 20, ny = 20)
>>> r = 0.25
>>> rad = numerix.sqrt((x - .5)**2 + (y - .5)**2) - r
>>> distanceVariable = DistanceVariable(mesh = mesh, value = rad)
>>> print numerix.allclose(distanceVariable.cellInterfaceAreas.sum(), 1.57984690073)
1
extendVariable(extensionVariable, order=2)
Calculates the extension of extensionVariable from the zero level set.
Parameters
• extensionVariable: The variable to extend from the zero level set.
getCellInterfaceAreas(*args, **kwds)
Deprecated since version 3.0: use the cellInterfaceAreas property instead
getLSMshape()


Chapter 28
fipy.viewers package
28.1 Subpackages
28.1.1 fipy.viewers.gistViewer package
Submodules
fipy.viewers.gistViewer.colorbar module
fipy.viewers.gistViewer.gist1DViewer module
fipy.viewers.gistViewer.gist2DViewer module
fipy.viewers.gistViewer.gistVectorViewer module
fipy.viewers.gistViewer.gistViewer module
fipy.viewers.gistViewer.test module
Module contents
fipy.viewers.gistViewer.GistViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a GistViewer.
The GistViewer factory will search the module tree and return an instance of the first GistViewer it finds of the
correct dimension.
Parameters
vars a CellVariable or tuple of CellVariable objects to plot
title displayed at the top of the Viewer window
limits [dict] a (deprecated) alternative to limit keyword arguments
xmin, xmax, ymin, ymax, datamin, datamax displayed range of data. A 1D Viewer will only
use xmin and xmax, a 2D viewer will also use ymin and ymax. All viewers will use datamin
and datamax. Any limit set to a (default) value of None will autoscale.
385
28.1.2 fipy.viewers.gnuplotViewer package
Submodules
fipy.viewers.gnuplotViewer.gnuplot1DViewer module
fipy.viewers.gnuplotViewer.gnuplot2DViewer module
fipy.viewers.gnuplotViewer.gnuplotViewer module
fipy.viewers.gnuplotViewer.test module
Module contents
fipy.viewers.gnuplotViewer.GnuplotViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a GnuplotViewer.
The GnuplotViewer factory will search the module tree and return an instance of the first GnuplotViewer it finds
of the correct dimension.
Parameters
28.1.3 fipy.viewers.matplotlibViewer package
Submodules
fipy.viewers.matplotlibViewer.matplotlib1DViewer module
class fipy.viewers.matplotlibViewer.matplotlib1DViewer.Matplotlib1DViewer(vars,
ti-
tle=None,
xlog=False,
ylog=False,
lim-
its={},
leg-
end=’upper
left’,
axes=None,
**kwlim-
its)
Bases: fipy.viewers.matplotlibViewer.matplotlibViewer.AbstractMatplotlibViewer
Displays a y vs. x plot of one or more 1D CellVariable objects using Matplotlib.
386 Chapter 28. fipy.viewers package


>>> mesh = Grid1D(nx=100)
>>> x, = mesh.cellCenters
>>> xVar = CellVariable(mesh=mesh, name="x", value=x)
>>> k = Variable(name="k", value=0.)
>>> viewer = Matplotlib1DViewer(vars=(numerix.sin(k * xVar), numerix.cos(k * xVar / numerix.pi))
... limits={'xmin': 10, 'xmax': 90},
... datamin=-0.9, datamax=2.0,
... title="Matplotlib1DViewer test")
>>> for kval in numerix.arange(0,0.3,0.03):
... k.setValue(kval)
... viewer.plot()
>>> viewer._promptForOpinion()
Parameters
xlog log scaling of x axis if True
ylog log scaling of y axis if True
xmin, xmax, datamin, datamax displayed range of data. Any limit set to a (default) value of
None will autoscale. (ymin and ymax are synonyms for datamin and datamax).
legend place a legend at the specified position, if not None
axes if not None, vars will be plotted into this Matplotlib Axes object
log
logarithmic data scaling
fipy.viewers.matplotlibViewer.matplotlib2DContourViewer module
fipy.viewers.matplotlibViewer.matplotlib2DGridContourViewer module
class fipy.viewers.matplotlibViewer.matplotlib2DGridContourViewer.Matplotlib2DGridContourView
Bases: fipy.viewers.matplotlibViewer.matplotlib2DViewer.AbstractMatplotlib2DViewer
Displays a contour plot of a 2D CellVariable object.
The Matplotlib2DGridContourViewer plots a 2D CellVariable using Matplotlib.


>>> mesh = Grid2D(nx=50, ny=100, dx=0.1, dy=0.01)
>>> xyVar = CellVariable(mesh=mesh, name="x y", value=x * y)
>>> viewer = Matplotlib2DGridContourViewer(vars=numerix.sin(k * xyVar),
... limits={'ymin': 0.1, 'ymax': 0.9},
... title="Matplotlib2DGridContourViewer test")
>>> for kval in range(10):
... viewer.plot()
Creates a Matplotlib2DViewer.
Parameters
vars a CellVariable object.
xmin, xmax, ymin, ymax, datamin, datamax displayed range of data. Any limit set to a (de-
fault) value of None will autoscale.
cmap the colormap. Defaults to matplotlib.cm.jet
colorbar plot a colorbar in specified orientation if not None
figaspect desired aspect ratio of figure. If arg is a number, use that aspect ratio. If arg is ‘auto’,
the aspect ratio will be determined from the Variable’s mesh.
fipy.viewers.matplotlibViewer.matplotlib2DGridViewer module
class fipy.viewers.matplotlibViewer.matplotlib2DGridViewer.Matplotlib2DGridViewer(vars,
ti-
tle=None,
lim-
its={},
cmap=None,
col-
or-
bar=’vertical’,
axes=None,
fi-
gaspect=’auto’,
**kwlim-
its)
Displays an image plot of a 2D CellVariable object using Matplotlib.


>>> viewer = Matplotlib2DGridViewer(vars=numerix.sin(k * xyVar),
... limits={'ymin': 0.1, 'ymax': 0.9},
... title="Matplotlib2DGridViewer test")
... viewer.plot()
Creates a Matplotlib2DGridViewer.
Parameters
vars A CellVariable object.
cmap The colormap. Defaults to matplotlib.cm.jet
fipy.viewers.matplotlibViewer.matplotlib2DViewer module
class fipy.viewers.matplotlibViewer.matplotlib2DViewer.Matplotlib2DViewer(vars,
ti-
tle=None,
lim-
its={},
cmap=None,
col-
or-
bar=’vertical’,
axes=None,
fi-
gaspect=’auto’,
**kwlim-
its)
The Matplotlib2DViewer plots a 2D CellVariable using Matplotlib.
>>> mesh = (Grid2D(nx=5, ny=10, dx=0.1, dy=0.1)
... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))


>>> viewer = Matplotlib2DViewer(vars=numerix.sin(k * xyVar),
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer module
class fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer.MatplotlibSparseMatrixViewer
plot(matrix, RHSvector, log=’auto’)

fipy.viewers.matplotlibViewer.matplotlibStreamViewer module
class fipy.viewers.matplotlibViewer.matplotlibStreamViewer.MatplotlibStreamViewer(vars,
ti-
tle=None,
log=False,
lim-
its={},
axes=None,
fi-
gaspect=’auto’,
den-
sity=1,
linewidth=None
color=None,
cmap=None,
norm=None,
ar-
row-
size=1,
arrowstyle=’-
|>’,
min-
length=0.1,
**kwlim-
its)
Displays a stream plot of a 2D rank-1 CellVariable or FaceVariable object using Matplotlib
One issue is that this Viewer relies on scipy.interpolate.griddata, which interpolates on the convex hull of the
data. The results is that streams are plotted across any concavities in the mesh.
Another issue is that it does not seem possible to remove the streams without calling cla(), which means that
different set of streams cannot be overlaid.
>>> viewer = MatplotlibStreamViewer(vars=numerix.sin(k * xyVar).grad,
... title="MatplotlibStreamViewer test")
>>> for kval in numerix.arange(1, 10):
... viewer.plot()
>>> viewer = MatplotlibStreamViewer(vars=numerix.sin(k * xyVar).faceGrad,

... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)

... + ((0.5,), (0.2,))))

... viewer.plot()

... viewer.plot()
Creates a MatplotlibStreamViewer.
Parameters
vars a rank-1 CellVariable or FaceVariable object.
log if True, arrow length goes at the base-10 logarithm of the magnitude
density [float or 2-tuple] Controls the closeness of streamlines. When density = 1, the domain
is divided into a 25x25 grid—density linearly scales this grid. Each cell in the grid can have,
at most, one traversing streamline. For different densities in each direction, use [density_x,
density_y].
linewidth [Numeric or rank-0 MeshVariable] vary linewidth when given a CellVariable or Face-
Variable of same type as vars.
color [matplotlib color code, or rank-0 MeshVariable] Streamline color. When given an array
with the type as vars, color values are converted to colors using cmap.
cmap [Colormap] Colormap used to plot streamlines and arrows. Only necessary when using
an MeshVariable input for color.
norm [Normalize] Normalize object used to scale luminance data to 0, 1. If None, stretch
(min, max) to (0, 1). Only necessary when color is an MeshVariable.
arrowsize [float] Factor scale arrow size.
arrowstyle [str] Arrow style specification. See FancyArrowPatch.
minlength [float] Minimum length of streamline in axes coordinates.

fipy.viewers.matplotlibViewer.matplotlibVectorViewer module
class fipy.viewers.matplotlibViewer.matplotlibVectorViewer.MatplotlibVectorViewer(vars,
ti-
tle=None,
scale=None,
spar-
sity=None,
log=False,
lim-
its={},
axes=None,
fi-
gaspect=’auto’,
**kwlim-
its)
Displays a vector plot of a 2D rank-1 CellVariable or FaceVariable object using Matplotlib
>>> viewer = MatplotlibVectorViewer(vars=numerix.sin(k * xyVar).grad,
... title="MatplotlibVectorViewer test")
... viewer.plot()
>>> viewer = MatplotlibVectorViewer(vars=numerix.sin(k * xyVar).faceGrad,

... viewer.plot()
>>> for sparsity in numerix.arange(5000, 0, -500):

... viewer.quiver(sparsity=sparsity)
... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... viewer.plot()


... viewer.plot()
Parameters
scale if not None, scale all arrow lengths by this value
sparsity if not None, then this number of arrows will be randomly chosen (weighted by the cell
volume or face area)
quiver(sparsity=None, scale=None)
fipy.viewers.matplotlibViewer.matplotlibViewer module
class fipy.viewers.matplotlibViewer.matplotlibViewer.AbstractMatplotlibViewer(vars,
ti-
tle=None,
fi-
gaspect=1.0,
cmap=None,
col-
or-
bar=None,
axes=None,
log=False,
**kwlim-
its)
Bases: fipy.viewers.viewer.AbstractViewer
The AbstractMatplotlibViewer is the base class for the viewers that use the Matplotlib python plotting package.
Create a AbstractMatplotlibViewer.
Parameters


figaspect desired aspect ratio of figure. If arg is a number, use that aspect ratio. If arg is an array,
figaspect will determine the width and height for a figure that would fit array preserving
aspect ratio.
log whether to logarithmically scale the data
figaspect(figaspect)
log
plot(filename=None)
fipy.viewers.matplotlibViewer.test module
Module contents
fipy.viewers.matplotlibViewer.MatplotlibViewer(vars, title=None, limits={},

cmap=None, colorbar=’vertical’,
axes=None, **kwlimits)
Generic function for creating a MatplotlibViewer.
The MatplotlibViewer factory will search the module tree and return an instance of the first MatplotlibViewer it
finds of the correct dimension and rank.
Parameters
It is possible to view different Variables against different Matplotlib Axes
>>> from matplotlib import pylab

>>> pylab.ion()
>>> fig = pylab.figure()
>>> ax1 = pylab.subplot((221))

>>> mesh1 = Grid1D(nx=100)

>>> x, = mesh1.cellCenters
>>> xVar = CellVariable(mesh=mesh1, name="x", value=x)
>>> viewer1 = MatplotlibViewer(vars=(numerix.sin(0.1 * k * xVar), numerix.cos(0.1 * k * xVar / n
... limits={'xmin': 10, 'xmax': 90},
... title="Grid1D test",
... axes=ax1,
... legend=None)
>>> mesh2 = Grid2D(nx=50, ny=100, dx=0.1, dy=0.01)

>>> x, y = mesh2.cellCenters
>>> xyVar = CellVariable(mesh=mesh2, name="x y", value=x * y)
>>> viewer2 = MatplotlibViewer(vars=numerix.sin(k * xyVar),
... limits={'ymin': 0.1, 'ymax': 0.9},
... axes=ax2,
... colorbar=None)
>>> mesh3 = (Grid2D(nx=5, ny=10, dx=0.1, dy=0.1)

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... title="Irregular 2D test",
... axes=ax3,
... cmap = pylab.cm.OrRd)
>>> viewer = MultiViewer(viewers=(viewer1, viewer2, viewer3))

... viewer.plot()
class fipy.viewers.matplotlibViewer.Matplotlib1DViewer(vars, title=None, xlog=False,

ylog=False, limits={}, leg-
end=’upper left’, axes=None,
**kwlimits)


... limits={'xmin': 10, 'xmax': 90},
... viewer.plot()
Parameters
log
class fipy.viewers.matplotlibViewer.Matplotlib2DGridViewer(vars, title=None, lim-
its={}, cmap=None,
colorbar=’vertical’,
axes=None, fi-
gaspect=’auto’, **kwlim-
its)
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters


class fipy.viewers.matplotlibViewer.Matplotlib2DGridContourViewer(vars, ti-
tle=None,
limits={},
cmap=None,
color-
bar=’vertical’,
axes=None, fi-
gaspect=’auto’,
**kwlimits)
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters

class fipy.viewers.matplotlibViewer.Matplotlib2DViewer(vars, title=None, lim-
its={}, cmap=None, color-
bar=’vertical’, axes=None,
figaspect=’auto’, **kwlimits)
... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
class fipy.viewers.matplotlibViewer.MatplotlibVectorViewer(vars, title=None,
scale=None, spar-
sity=None, log=False,
limits={}, axes=None, fi-
gaspect=’auto’, **kwlim-
its)

... viewer.plot()

... viewer.plot()

... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... viewer.plot()

... viewer.plot()
Parameters

class fipy.viewers.matplotlibViewer.MatplotlibStreamViewer(vars, title=None,
log=False, lim-
its={}, axes=None,
figaspect=’auto’, den-
sity=1, linewidth=None,
color=None,
cmap=None,
norm=None, arrow-
size=1, arrowstyle=’-|>’,
minlength=0.1, **kwlim-
its)
... viewer.plot()

... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))

... viewer.plot()

... viewer.plot()
Parameters
density_y].
28.1.4 fipy.viewers.mayaviViewer package
Submodules

fipy.viewers.mayaviViewer.mayaviClient module
class fipy.viewers.mayaviViewer.mayaviClient.MayaviClient(vars, title=None, dae-

mon_file=None, fps=1.0,
**kwlimits)
The MayaviClient uses the Mayavi python plotting package.
>>> viewer = MayaviClient(vars=(numerix.sin(k * xVar), numerix.cos(k * xVar / numerix.pi)),
... limits={'xmin': 10, 'xmax': 90},
... title="MayaviClient test")
... viewer.plot()

>>> viewer = MayaviClient(vars=numerix.sin(k * xyVar),
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

>>> mesh = Grid3D(nx=50, ny=100, nz=10, dx=0.1, dy=0.01, dz=0.1)
>>> x, y, z = mesh.cellCenters
>>> xyzVar = CellVariable(mesh=mesh, name=r"x y z", value=x * y * z)
>>> viewer = MayaviClient(vars=numerix.sin(k * xyzVar),

... limits={'ymin': 0.1, 'ymax': 0.9},

... viewer.plot()
Create a MayaviClient.
Parameters
xmin, xmax, ymin, ymax, zmin, zmax, datamin, datamax displayed range of data. A 1D
Viewer will only use xmin and xmax, a 2D viewer will also use ymin and ymax, and so
on. All viewers will use datamin and datamax. Any limit set to a (default) value of None
will autoscale.
daemon_file the path to the script to run the separate MayaVi viewer process. Defaults to
“fipy/viewers/mayaviViewer/mayaviDaemon.py”
fps frames per second to attempt to display
plot(filename=None)
fipy.viewers.mayaviViewer.mayaviDaemon module
fipy.viewers.mayaviViewer.test module
Module contents
class fipy.viewers.mayaviViewer.MayaviClient(vars, title=None, daemon_file=None, fps=1.0,

**kwlimits)
... limits={'xmin': 10, 'xmax': 90},
... viewer.plot()



... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
will autoscale.
plot(filename=None)

28.1.5 fipy.viewers.vtkViewer package
Submodules
fipy.viewers.vtkViewer.test module
fipy.viewers.vtkViewer.vtkCellViewer module
class fipy.viewers.vtkViewer.vtkCellViewer.VTKCellViewer(vars, title=None, limits={},

**kwlimits)
Bases: fipy.viewers.vtkViewer.vtkViewer.VTKViewer
Renders CellVariable data in VTK format
Creates a VTKViewer
Parameters
vars a _MeshVariable or a tuple of them
xmin, xmax, ymin, ymax, zmin, zmax, datamin, datamax displayed range of data. Any limit
set to a (default) value of None will autoscale.
fipy.viewers.vtkViewer.vtkFaceViewer module
class fipy.viewers.vtkViewer.vtkFaceViewer.VTKFaceViewer(vars, title=None, limits={},

**kwlimits)
Renders _MeshVariable data in VTK format
Creates a VTKViewer
Parameters
fipy.viewers.vtkViewer.vtkViewer module
class fipy.viewers.vtkViewer.vtkViewer.VTKViewer(vars, title=None, limits={}, **kwlimits)

Creates a VTKViewer
Parameters


plot(filename=None)
Module contents
fipy.viewers.vtkViewer.VTKViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a VTKViewer.
The VTKViewer factory will search the module tree and return an instance of the first VTKViewer it finds of the
correct dimension and rank.
Parameters
vars a _MeshVariable or tuple of _MeshVariable objects to plot
will autoscale.
class fipy.viewers.vtkViewer.VTKCellViewer(vars, title=None, limits={}, **kwlimits)
Creates a VTKViewer
Parameters
class fipy.viewers.vtkViewer.VTKFaceViewer(vars, title=None, limits={}, **kwlimits)
Creates a VTKViewer
Parameters

28.2 Submodules
28.3 fipy.viewers.multiViewer module
class fipy.viewers.multiViewer.MultiViewer(viewers)
Treat a collection of different viewers (such for different 2D plots or 1D plots with different axes) as a single
viewer that will plot() all subviewers simultaneously.
Parameters
viewers [list] the viewers to bind together
getViewers(*args, **kwds)
Deprecated since version 3.0: use the viewers property instead
plot()
setLimits(limits={}, **kwlimits)
28.4 fipy.viewers.test module
Test implementation of the viewers
28.5 fipy.viewers.testinteractive module
Interactively test the viewers
28.6 fipy.viewers.tsvViewer module
class fipy.viewers.tsvViewer.TSVViewer(vars, title=None, limits={}, **kwlimits)

“Views” one or more variables in tab-separated-value format.
Output is a list of coordinates and variable values at each cell center.
File contents will be, e.g.:
title
x y ... var0 var2 ...
0.0 0.0 ... 3.14 1.41 ...
1.0 0.0 ... 2.72 0.866 ...
:
:

Creates a TSVViewer.
Any cell centers that lie outside the limits provided will not be included. Any values that lie outside the datamin
or datamax will be replaced with nan.
All variables must have the same mesh.
It tries to do something reasonable with rank-1 CellVariable and FaceVariable objects.
Parameters
vars a CellVariable, a FaceVariable, a tuple of CellVariable objects, or a tuple of FaceVariable
objects to plot
plot(filename=None)
“plot” the coordinates and values of the variables to filename. If filename is not provided, “plots” to stdout.
>>> m = Grid1D(nx = 3, dx = 0.4)
>>> v = CellVariable(mesh = m, name = "var", value = (0, 2, 5))
>>> TSVViewer(vars = (v, v.grad)).plot()
x var var_gauss_grad_x
0.2 0 2.5
0.6 2 6.25
1 5 3.75

>>> m = Grid2D(nx = 2, dx = .1, ny = 2, dy = 0.3)
>>> v = CellVariable(mesh = m, name = "var", value = (0, 2, -2, 5))
x y var var_gauss_grad_x var_gauss_grad_y
0.05 0.15 0 10 -3.33333333333333
0.15 0.15 2 10 5
0.05 0.45 -2 35 -3.33333333333333
0.15 0.45 5 35 5
Parameters
filename If not None, the name of a file to save the image into.
28.7 fipy.viewers.viewer module
class fipy.viewers.viewer.AbstractViewer(vars, title=None, **kwlimits)

Bases: object
Create a AbstractViewer object.

Parameters
28.7. fipy.viewers.viewer module 409


will autoscale.
getVars(*args, **kwds)
Get the Variables
Deprecated since version 3.0: use the vars property instead
Deprecated since version Use: fipy.viewers.viewer.AbstractViewer.vars in-
stead
plot(filename=None)
Update the display of the viewed variables.
Parameters
plotMesh(filename=None)
Display a representation of the mesh
Parameters
Update the limits.
Parameters
Viewer will only use xmin and xmax, a 2D viewer will also use ymin and ymax, and so on.
All viewers will use datamin and datamax. Any limit set to a (default) value of None will
autoscale.
fipy.viewers.GistViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a GistViewer.
The GistViewer factory will search the module tree and return an instance of the first GistViewer it finds of the
correct dimension.
Parameters

fipy.viewers.GnuplotViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a GnuplotViewer.
The GnuplotViewer factory will search the module tree and return an instance of the first GnuplotViewer it finds
of the correct dimension.
Parameters
fipy.viewers.MatplotlibViewer(vars, title=None, limits={}, cmap=None, colorbar=’vertical’,
axes=None, **kwlimits)
Generic function for creating a MatplotlibViewer.
The MatplotlibViewer factory will search the module tree and return an instance of the first MatplotlibViewer it
finds of the correct dimension and rank.
Parameters
It is possible to view different Variables against different Matplotlib Axes
>>> from matplotlib import pylab
>>> pylab.ion()
>>> fig = pylab.figure()

>>> mesh1 = Grid1D(nx=100)

>>> x, = mesh1.cellCenters
>>> xVar = CellVariable(mesh=mesh1, name="x", value=x)
>>> viewer1 = MatplotlibViewer(vars=(numerix.sin(0.1 * k * xVar), numerix.cos(0.1 * k * xVar / n
... limits={'xmin': 10, 'xmax': 90},

... axes=ax1,
... legend=None)
>>> mesh2 = Grid2D(nx=50, ny=100, dx=0.1, dy=0.01)

... limits={'ymin': 0.1, 'ymax': 0.9},
... axes=ax2,
... colorbar=None)
>>> mesh3 = (Grid2D(nx=5, ny=10, dx=0.1, dy=0.1)

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... title="Irregular 2D test",
... axes=ax3,
... cmap = pylab.cm.OrRd)
>>> viewer = MultiViewer(viewers=(viewer1, viewer2, viewer3))

... viewer.plot()
class fipy.viewers.Matplotlib1DViewer(vars, title=None, xlog=False, ylog=False, limits={}, leg-

end=’upper left’, axes=None, **kwlimits)
... limits={'xmin': 10, 'xmax': 90},
... viewer.plot()
Parameters


log
class fipy.viewers.Matplotlib2DGridViewer(vars, title=None, limits={}, cmap=None, col-
orbar=’vertical’, axes=None, figaspect=’auto’,
**kwlimits)
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
class fipy.viewers.Matplotlib2DGridContourViewer(vars, title=None, limits={}, cmap=None,
colorbar=’vertical’, axes=None, fi-
gaspect=’auto’, **kwlimits)


... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
class fipy.viewers.Matplotlib2DViewer(vars, title=None, limits={}, cmap=None, color-
bar=’vertical’, axes=None, figaspect=’auto’, **kwlim-
its)
... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

Parameters
class fipy.viewers.MatplotlibVectorViewer(vars, title=None, scale=None, sparsity=None,
log=False, limits={}, axes=None, fi-
gaspect=’auto’, **kwlimits)
... viewer.plot()

... viewer.plot()

... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))

... viewer.plot()

... viewer.plot()
Parameters
class fipy.viewers.MatplotlibStreamViewer(vars, title=None, log=False, limits={}, axes=None,
figaspect=’auto’, density=1, linewidth=None,
color=None, cmap=None, norm=None, ar-
rowsize=1, arrowstyle=’-|>’, minlength=0.1,
**kwlimits)
... viewer.plot()


... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... viewer.plot()

... viewer.plot()
Parameters
density_y].

class fipy.viewers.MayaviClient(vars, title=None, daemon_file=None, fps=1.0, **kwlimits)
... limits={'xmin': 10, 'xmax': 90},
... viewer.plot()

... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()

... + (Tri2D(nx=5, ny=5, dx=0.1, dy=0.1)
... + ((0.5,), (0.2,))))
... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()


... limits={'ymin': 0.1, 'ymax': 0.9},
... viewer.plot()
Parameters
will autoscale.
plot(filename=None)
class fipy.viewers.MultiViewer(viewers)
Treat a collection of different viewers (such for different 2D plots or 1D plots with different axes) as a single
viewer that will plot() all subviewers simultaneously.
Parameters
viewers [list] the viewers to bind together
getViewers(*args, **kwds)
Deprecated since version 3.0: use the viewers property instead
plot()
class fipy.viewers.TSVViewer(vars, title=None, limits={}, **kwlimits)
“Views” one or more variables in tab-separated-value format.
Output is a list of coordinates and variable values at each cell center.
File contents will be, e.g.:
title
x y ... var0 var2 ...
0.0 0.0 ... 3.14 1.41 ...
1.0 0.0 ... 2.72 0.866 ...
:
:

Creates a TSVViewer.
Any cell centers that lie outside the limits provided will not be included. Any values that lie outside the datamin
or datamax will be replaced with nan.
All variables must have the same mesh.
It tries to do something reasonable with rank-1 CellVariable and FaceVariable objects.
Parameters
vars a CellVariable, a FaceVariable, a tuple of CellVariable objects, or a tuple of FaceVariable
objects to plot
plot(filename=None)
“plot” the coordinates and values of the variables to filename. If filename is not provided, “plots” to stdout.
>>> m = Grid1D(nx = 3, dx = 0.4)
>>> v = CellVariable(mesh = m, name = "var", value = (0, 2, 5))
x var var_gauss_grad_x
0.2 0 2.5
0.6 2 6.25
1 5 3.75

>>> m = Grid2D(nx = 2, dx = .1, ny = 2, dy = 0.3)
>>> v = CellVariable(mesh = m, name = "var", value = (0, 2, -2, 5))
x y var var_gauss_grad_x var_gauss_grad_y
0.05 0.15 0 10 -3.33333333333333
0.15 0.15 2 10 5
0.05 0.45 -2 35 -3.33333333333333
0.15 0.45 5 35 5
Parameters
fipy.viewers.VTKViewer(vars, title=None, limits={}, **kwlimits)

Generic function for creating a VTKViewer.
The VTKViewer factory will search the module tree and return an instance of the first VTKViewer it finds of the
correct dimension and rank.
Parameters
vars a _MeshVariable or tuple of _MeshVariable objects to plot

will autoscale.
class fipy.viewers.VTKCellViewer(vars, title=None, limits={}, **kwlimits)
Creates a VTKViewer
Parameters
class fipy.viewers.VTKFaceViewer(vars, title=None, limits={}, **kwlimits)
Creates a VTKViewer
Parameters
exception fipy.viewers.MeshDimensionError
Bases: exceptions.IndexError
class fipy.viewers.DummyViewer(vars, title=None, **kwlimits)
Create a AbstractViewer object.
Parameters
will autoscale.
plot(filename=None)
fipy.viewers.Viewer(vars, title=None, limits={}, FIPY_VIEWER=None, **kwlimits)
Generic function for creating a Viewer.

The Viewer factory will search the module tree and return an instance of the first Viewer it finds that supports the
dimensions of vars. Setting the ‘FIPY_VIEWER‘ environment variable to either ‘gist‘, ‘gnuplot‘, ‘matplotlib‘,
‘tsv‘, or ‘vtk‘ will specify the viewer.
The kwlimits or limits parameters can be used to constrain the view. For example:
Viewer(vars=some1Dvar, xmin=0.5, xmax=None, datamax=3)
or:
Viewer(vars=some1Dvar,
limits={'xmin': 0.5, 'xmax': None, 'datamax': 3})
will return a viewer that displays a line plot from an x value of 0.5 up to the largest x value in the dataset. The
data values will be truncated at an upper value of 3, but will have no lower limit.
Parameters
FIPY_VIEWER a specific viewer to attempt (possibly multiple times for multiple variables)
will autoscale.

Bibliography
[1] Greg Ward. Installing Python Modules. URL: http://docs.python.org/inst/.

[2] W. J. Boettinger, J. A. Warren, C. Beckermann, and A. Karma. Phase-field simulation of solidification. Annual
Review of Materials Research, 32:163–194, 2002. doi:10.1146/annurev.matsci.32.101901.155803.
[3] L. Q. Chen. Phase-field mdoels for microstructure evolution. Annual Review of Materials Research, 32:113–140,
2002. doi:10.1146/annurev.matsci.32.112001.132041.
[4] G. B. McFadden. Phase-field models of solidification. Contemporary Mathematics, 306:107–145, 2002.
[5] David M Saylor, Jonathan E Guyer, Daniel Wheeler, and James A Warren. Predicting microstruc-
ture development during casting of drug-eluting coatings. Acta Biomaterialia, 7(2):604–613, Jan 2011.
doi:10.1016/j.actbio.2010.09.019.
[6] Daniel Wheeler, James A. Warren, and William J. Boettinger. Modeling the early stages of reactive wetting.
Physical Review E, 82(5):051601, Nov 2010. doi:10.1103/PhysRevE.82.051601.
[7] C. M Hangarter, B. H Hamadani, J. E Guyer, H Xu, R Need, and D Josell. Three dimensionally structured inter-
digitated back contact thin film heterojunction solar cells. Journal of Applied Physics, 109(7):073514, Jan 2011.
doi:10.1063/1.3561487.
[8] D. Josell, D. Wheeler, W. H. Huber, and T. P. Moffat. Superconformal electrodeposition in submicron features.
Physical Review Letters, 87(1):016102, 2001. doi:10.1103/PhysRevLett.87.016102.
[9] James A. Warren, Ryo Kobayashi, Alexander E. Lobkovsky, and W. Craig Carter. Extending phase field mod-
els of solidification to polycrystalline materials. Acta Materialia, 51(20):6035–6058, 2003. doi:10.1016/S1359-
6454(03)00388-4.
[10] Scott Chacon. Pro Git. Apress, 2009. URL: http://git-scm.com/book.
[11] Guido van Rossum. Python Tutorial. URL: http://docs.python.org/tut/.
[12] Mark Pilgrim. Dive Into Python. Apress, 2004. URL: http://diveintopython.org.
[13] T. N. Croft. Unstructured Mesh - Finite Volume Algorithms for Swirling, Turbulent Reacting Flows. PhD thesis,
University of Greenwich, 1998. URL: http://www.gre.ac.uk/~ct02/research/thesis/main.html.
[14] John W. Cahn and John E. Hilliard. Free energy of a nonuniform system. I. Interfacial free energy. Journal of
Chemical Physics, 28(2):258–267, 1958.
[15] John W. Cahn. Free energy of a nonuniform system. II. Thermodynamic basis. Journal of Chemical Physics,
30(5):1121–1124, 1959.
[16] John W. Cahn and John E. Hilliard. Free energy of a nonuniform system. III. Nucleation in a two-component
incompressible fluid. Journal of Chemical Physics, 31(3):688–699, 1959.
423
[17] K. R Elder, K Thornton, and J. J Hoyt. The kirkendall effect in the phase field crystal model. Philosophical
Magagazine, 91(1):151–164, Jan 2011. doi:10.1080/14786435.2010.506427.
[18] S. V. Patankar. Numerical Heat Transfer and Fluid Flow. Taylor and Francis, 1980.
[19] H. K. Versteeg and W. Malalasekera. An Introduction to Computational Fluid Dynamics. Longman Scientific and
Technical, 1995.
[20] C. Mattiussi. An analysis of finite volume, finite element, and finite difference methods using
some concepts from algebraic topology. Journal of Computational Physics, 133:289–309, 1997. URL:
http://lis.epfl.ch/publications/JCP1997.pdf.
[21] J. A. Sethian. Level Set Methods and Fast Marching Methods. Cambridge University Press, 1996.
[22] William H. Press, Saul A. Teukolsky, William T. Vetterling, and Brian P. Flannery. Numerical Recipes in C: the
Art of Scientific Computing. Cambridge University Press, 2nd edition, 1999.
[23] D. Wheeler, D. Josell, and T. P. Moffat. Modeling superconformal electrodeposition using the level set method.
Journal of The Electrochemical Society, 150(5):C302–C310, 2003. doi:10.1149/1.1562598.
[24] D. Josell, D. Wheeler, and T. P. Moffat. Gold superfill in submicrometer trenches: experiment and prediction.
Journal of The Electrochemical Society, 153(1):C11–C18, 2006. doi:10.1149/1.2128765.
[25] T. P. Moffat, D. Wheeler, S. K. Kim, and D. Josell. Curvature enhanced adsorbate coverage model for electrode-
position. Journal of The Electrochemical Society, 153(2):C127–C132, 2006. doi:10.1149/1.2165580.
[26] A. A. Wheeler, W. J. Boettinger, and G. B. McFadden. Phase-field model for isothermal phase transitions in
binary alloys. Physical Review A, 45(10):7424–7439, 1992.
[27] J. A. Warren and W. J. Boettinger. Prediction of dendritic growth and microsegregation in a binary alloy using
the phase field method. Acta Metallurgica et Materialia, 43(2):689–703, 1995.
[28] J. E. Guyer, W. J. Boettinger, J. A. Warren, and G. B. McFadden. Phase field modeling of electrochemistry I:
Equilibrium. Physical Review E, 69:021603, 2004. arXiv:cond-mat/0308173, doi:10.1103/PhysRevE.69.021603.
[29] J. E. Guyer, W. J. Boettinger, J. A. Warren, and G. B. McFadden. Phase field modeling of electrochemistry II:
Kinetics. Physical Review E, 69:021604, 2004. arXiv:cond-mat/0308179, doi:10.1103/PhysRevE.69.021604.
[30] Guido van Rossum. Python Reference Manual. URL: http://docs.python.org/ref/.
[31] Ben Collins-Sussman, Brian W. Fitzpatrick, and C. Michael Pilato. Version Control with Subversion. O’Reilly
Media, 2004. URL: http://svnbook.red-bean.com.
[32] T. P. Moffat, D. Wheeler, and D. Josell. Superfilling and the curvature enhanced acceler-
ator coverage mechanism. The Electrochemical Society, Interface, 13(4):46–52, 2004. URL:
http://www.electrochem.org/publications/interface/winter2004/IF12-04-Pg46.pdf.
[33] J. H. Ferziger and M. Perić. Computational Methods for Fluid Dynamics. Springer, 1996.
424 Bibliography
Python Module Index
e 173
examples.cahnHilliard.mesh2DCoupled, 161 examples.updating.update0_1to1_0, 184
examples.cahnHilliard.sphere, 164 examples.updating.update1_0to2_0, 180
examples.convection.exponential1D.mesh1D,examples.updating.update2_0to3_0, 179
89
f
examples.convection.exponential1DSource.mesh1D,
90 fipy.boundaryConditions, 196
examples.convection.robin, 91 fipy.boundaryConditions.boundaryCondition,
examples.convection.source, 93 195
examples.diffusion.anisotropy, 87 fipy.boundaryConditions.constraint, 195
examples.diffusion.circle, 76 fipy.boundaryConditions.fixedFlux, 195
examples.diffusion.coupled, 72 fipy.boundaryConditions.fixedValue, 196
examples.diffusion.electrostatics, 81 fipy.boundaryConditions.nthOrderBoundaryCondition,
examples.diffusion.mesh1D, 55 196
examples.diffusion.mesh20x20, 74 fipy.boundaryConditions.test, 196
fipy.matrices, 199
examples.diffusion.nthOrder.input4thOrder1D,
85 fipy.matrices.offsetSparseMatrix, 199
examples.flow.stokesCavity, 167 fipy.matrices.pysparseMatrix, 199
examples.levelSet.advection.circle, 142 fipy.matrices.scipyMatrix, 199
examples.levelSet.advection.mesh1D, 141 fipy.matrices.sparseMatrix, 199
fipy.matrices.test, 199
examples.levelSet.distanceFunction.circle,
140 fipy.meshes, 223
fipy.meshes.abstractMesh, 203
examples.levelSet.distanceFunction.mesh1D,
139 fipy.meshes.builders, 201
examples.levelSet.electroChem.gold, 148 fipy.meshes.builders.abstractGridBuilder,
examples.levelSet.electroChem.howToWriteAScript,201
153 fipy.meshes.builders.grid1DBuilder, 201
examples.levelSet.electroChem.leveler, fipy.meshes.builders.grid2DBuilder, 201
149 fipy.meshes.builders.grid3DBuilder, 201
fipy.meshes.builders.periodicGrid1DBuilder,
examples.levelSet.electroChem.simpleTrenchSystem,
145 201
examples.phase.anisotropy, 118 fipy.meshes.builders.utilityClasses, 201
examples.phase.binaryCoupled, 103 fipy.meshes.cylindricalNonUniformGrid1D,
examples.phase.impingement.mesh20x20, 207
124 fipy.meshes.cylindricalNonUniformGrid2D,
examples.phase.impingement.mesh40x1, 121 208
examples.phase.polyxtal, 127 fipy.meshes.cylindricalUniformGrid1D,
examples.phase.polyxtalCoupled, 133 208
examples.phase.quaternary, 112 fipy.meshes.cylindricalUniformGrid2D,
examples.phase.simple, 95 209
examples.reactiveWetting.liquidVapor1D, fipy.meshes.factoryMeshes, 209
425
fipy.meshes.gmshMesh, 211 fipy.solvers.pysparse.linearGMRESSolver,

fipy.meshes.grid1D, 216 235
fipy.meshes.grid2D, 216 fipy.solvers.pysparse.linearJORSolver,
fipy.meshes.grid3D, 216 235
fipy.meshes.mesh, 216 fipy.solvers.pysparse.linearLUSolver,
fipy.meshes.mesh1D, 216 235
fipy.meshes.mesh2D, 216 fipy.solvers.pysparse.linearPCGSolver,
fipy.meshes.nonUniformGrid1D, 217 236
fipy.meshes.nonUniformGrid2D, 217 fipy.solvers.pysparse.preconditioners,
fipy.meshes.nonUniformGrid3D, 218 234
fipy.meshes.numMesh, 202 fipy.solvers.pysparse.preconditioners.jacobiPrecond
fipy.meshes.numMesh.cylindricalUniformGrid1D, 234
201 fipy.solvers.pysparse.preconditioners.preconditione
fipy.meshes.numMesh.cylindricalUniformGrid2D, 234
201 fipy.solvers.pysparse.preconditioners.ssorPrecondit
fipy.meshes.numMesh.deprecatedWarning, 234
201 fipy.solvers.pysparse.pysparseSolver,
fipy.meshes.numMesh.gmshImport, 202 236
fipy.meshes.numMesh.grid1D, 202 fipy.solvers.pysparseMatrixSolver, 242
fipy.meshes.numMesh.grid2D, 202 fipy.solvers.scipy, 239
fipy.meshes.numMesh.grid3D, 202 fipy.solvers.scipy.linearBicgstabSolver,
fipy.meshes.numMesh.periodicGrid1D, 202 238
fipy.meshes.numMesh.periodicGrid2D, 202 fipy.solvers.scipy.linearCGSSolver, 238
fipy.meshes.numMesh.skewedGrid2D, 202 fipy.solvers.scipy.linearGMRESSolver,
fipy.meshes.numMesh.tri2D, 202 238
fipy.meshes.numMesh.uniformGrid1D, 202 fipy.solvers.scipy.linearLUSolver, 239
fipy.meshes.numMesh.uniformGrid2D, 202 fipy.solvers.scipy.linearPCGSolver, 239
fipy.meshes.numMesh.uniformGrid3D, 202 fipy.solvers.scipy.scipyKrylovSolver,
fipy.meshes.periodicGrid1D, 218 239
fipy.meshes.periodicGrid2D, 219 fipy.solvers.scipy.scipySolver, 239
fipy.meshes.representations, 202 fipy.solvers.solver, 242
fipy.solvers.test, 243
fipy.meshes.representations.abstractRepresentation,
202 fipy.steppers, 248
fipy.steppers.pidStepper, 247
fipy.meshes.representations.gridRepresentation,
202 fipy.steppers.pseudoRKQSStepper, 247
fipy.steppers.stepper, 247
fipy.meshes.representations.meshRepresentation,
202 fipy.terms, 269
fipy.meshes.skewedGrid2D, 220 fipy.terms.abstractBinaryTerm, 251
fipy.meshes.test, 220 fipy.terms.abstractConvectionTerm, 251
fipy.meshes.topologies, 203 fipy.terms.abstractDiffusionTerm, 251
fipy.meshes.topologies.abstractTopology, fipy.terms.abstractUpwindConvectionTerm,
202 251
fipy.meshes.topologies.gridTopology, 202 fipy.terms.advectionTerm, 251
fipy.meshes.topologies.meshTopology, 202 fipy.terms.asymmetricConvectionTerm, 254
fipy.meshes.tri2D, 221 fipy.terms.binaryTerm, 254
fipy.meshes.uniformGrid, 221 fipy.terms.cellTerm, 254
fipy.meshes.uniformGrid1D, 221 fipy.terms.centralDiffConvectionTerm,
fipy.meshes.uniformGrid2D, 222 254
fipy.meshes.uniformGrid3D, 222 fipy.terms.coupledBinaryTerm, 255
fipy.solvers, 243 fipy.terms.diffusionTerm, 255
fipy.solvers.pysparse, 236 fipy.terms.diffusionTermCorrection, 256
fipy.solvers.pysparse.linearCGSSolver, fipy.terms.diffusionTermNoCorrection,
234 256
fipy.terms.explicitDiffusionTerm, 256
426 Python Module Index

fipy.terms.explicitSourceTerm, 256 fipy.tools.vitals, 316

fipy.terms.explicitUpwindConvectionTerm, fipy.variables, 357
256 fipy.variables.addOverFacesVariable, 327
fipy.terms.exponentialConvectionTerm, fipy.variables.arithmeticCellToFaceVariable,
257 327
fipy.terms.faceTerm, 258 fipy.variables.betaNoiseVariable, 327
fipy.terms.firstOrderAdvectionTerm, 259 fipy.variables.binaryOperatorVariable,
fipy.terms.hybridConvectionTerm, 260 329
fipy.terms.implicitDiffusionTerm, 261 fipy.variables.cellToFaceVariable, 329
fipy.terms.implicitSourceTerm, 261 fipy.variables.cellVariable, 329
fipy.terms.nonDiffusionTerm, 261 fipy.variables.cellVolumeAverageVariable,
fipy.terms.powerLawConvectionTerm, 261 334
fipy.terms.residualTerm, 262 fipy.variables.constant, 334
fipy.terms.sourceTerm, 263 fipy.variables.constraintMask, 334
fipy.terms.term, 263 fipy.variables.coupledCellVariable, 334
fipy.terms.test, 265 fipy.variables.distanceVariable, 334
fipy.terms.transientTerm, 265 fipy.variables.exponentialNoiseVariable,
fipy.terms.unaryTerm, 266 338
fipy.terms.upwindConvectionTerm, 266 fipy.variables.faceGradContributionsVariable,
fipy.terms.vanLeerConvectionTerm, 268 340
fipy.tests, 288 fipy.variables.faceGradVariable, 340
fipy.tests.doctestPlus, 287 fipy.variables.faceVariable, 340
fipy.tests.lateImportTest, 288 fipy.variables.gammaNoiseVariable, 340
fipy.tests.testBase, 288 fipy.variables.gaussCellGradVariable,
fipy.tests.testClass, 288 342
fipy.tests.testProgram, 288 fipy.variables.gaussianNoiseVariable,
fipy.tools, 316 342
fipy.tools.comms, 290 fipy.variables.harmonicCellToFaceVariable,
fipy.tools.comms.commWrapper, 289 344
fipy.tools.comms.dummyComm, 290 fipy.variables.histogramVariable, 344
fipy.tools.comms.mpi4pyCommWrapper, 290 fipy.variables.interfaceAreaVariable,
fipy.tools.comms.serialCommWrapper, 290 345
fipy.tools.copy_script, 306 fipy.variables.interfaceFlagVariable,
fipy.tools.debug, 306 345
fipy.tools.decorators, 306 fipy.variables.leastSquaresCellGradVariable,
fipy.tools.dimensions, 305 345
fipy.tools.dimensions.DictWithDefault, fipy.variables.levelSetDiffusionVariable,
290 345
fipy.tools.dimensions.NumberDict, 290 fipy.variables.meshVariable, 345
fipy.tools.dimensions.physicalField, 290 fipy.variables.minmodCellToFaceVariable,
fipy.tools.dump, 307 345
fipy.tools.inline, 308 fipy.variables.modCellGradVariable, 345
fipy.tools.numerix, 308 fipy.variables.modCellToFaceVariable,
fipy.tools.parser, 315 345
fipy.tools.performance, 306 fipy.variables.modFaceGradVariable, 345
fipy.tools.performance.efficiency_test, fipy.variables.modPhysicalField, 345
305 fipy.variables.modularVariable, 345
fipy.variables.noiseVariable, 346
fipy.tools.performance.efficiencyTestGenerator,
305 fipy.variables.operatorVariable, 347
fipy.tools.performance.memoryLeak, 305 fipy.variables.scharfetterGummelFaceVariable,
fipy.tools.performance.memoryLogger, 305 347
fipy.tools.performance.memoryUsage, 306 fipy.variables.surfactantConvectionVariable,
fipy.tools.test, 316 347
fipy.tools.vector, 316 fipy.variables.surfactantVariable, 348
Python Module Index 427

fipy.variables.test, 349 fipy.viewers.vtkViewer.vtkFaceViewer,

fipy.variables.unaryOperatorVariable, 406
349 fipy.viewers.vtkViewer.vtkViewer, 406
fipy.variables.uniformNoiseVariable, 349
fipy.variables.variable, 350 p
fipy.viewers, 410 package.subpackage, 193
fipy.viewers.gistViewer, 385 package.subpackage.base, 191
fipy.viewers.gistViewer.gist1DViewer, package.subpackage.object, 192
385
fipy.viewers.gistViewer.gist2DViewer,
385
fipy.viewers.gistViewer.gistVectorViewer,
385
fipy.viewers.gistViewer.gistViewer, 385
fipy.viewers.gistViewer.test, 385
fipy.viewers.gnuplotViewer, 386
fipy.viewers.gnuplotViewer.gnuplot1DViewer,
386
fipy.viewers.gnuplotViewer.gnuplot2DViewer,
386
fipy.viewers.gnuplotViewer.gnuplotViewer,
386
fipy.viewers.gnuplotViewer.test, 386
fipy.viewers.matplotlibViewer, 395
fipy.viewers.matplotlibViewer.matplotlib1DViewer,
386
fipy.viewers.matplotlibViewer.matplotlib2DGridContourViewer,
387
fipy.viewers.matplotlibViewer.matplotlib2DGridViewer,
388
fipy.viewers.matplotlibViewer.matplotlib2DViewer,
389
fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer,
390
fipy.viewers.matplotlibViewer.matplotlibStreamViewer,
391
fipy.viewers.matplotlibViewer.matplotlibVectorViewer,
393
fipy.viewers.matplotlibViewer.matplotlibViewer,
394
fipy.viewers.matplotlibViewer.test, 395
fipy.viewers.mayaviViewer, 404
fipy.viewers.mayaviViewer.mayaviClient,
403
fipy.viewers.mayaviViewer.test, 404
fipy.viewers.multiViewer, 408
fipy.viewers.test, 408
fipy.viewers.testinteractive, 408
fipy.viewers.tsvViewer, 408
fipy.viewers.viewer, 409
fipy.viewers.vtkViewer, 407
fipy.viewers.vtkViewer.test, 406
fipy.viewers.vtkViewer.vtkCellViewer,
406
428 Python Module Index

Index
Symbols allclose() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper

–inline method), 290
command line option, 20 allclose() (fipy.tools.dimensions.physicalField.PhysicalField
–lsmlib method), 294
command line option, 21 allclose() (fipy.tools.PhysicalField method), 317
–no-pysparse allclose() (fipy.variables.Variable method), 358
command line option, 20 allclose() (fipy.variables.variable.Variable method), 351
–pyamg allclose() (in module fipy.tools.numerix), 309, 312
command line option, 21 allequal() (fipy.tools.comms.commWrapper.CommWrapper
–pysparse method), 289
command line option, 20 allequal() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper
–scipy method), 290
command line option, 20 allequal() (fipy.tools.dimensions.physicalField.PhysicalField
–skfmm method), 294
command line option, 21 allequal() (fipy.tools.PhysicalField method), 317
–trilinos allequal() (fipy.variables.Variable method), 358
command line option, 20 allequal() (fipy.variables.variable.Variable method), 351
:math:‘\pi‘, 119, 123, 125 allequal() (in module fipy.tools.numerix), 312
:module: viewers, 116 allgather() (fipy.tools.comms.commWrapper.CommWrapper
method), 289
A allgather() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrappe
AbstractBaseClassError, 269 method), 290
AbstractMatplotlibViewer (class in any() (fipy.tools.comms.commWrapper.CommWrapper
fipy.viewers.matplotlibViewer.matplotlibViewer), method), 289
394 any() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper
AbstractMesh (class in fipy.meshes.abstractMesh), 203 method), 290
AbstractViewer (class in fipy.viewers.viewer), 409 any() (fipy.variables.Variable method), 358
add() (fipy.tools.dimensions.physicalField.PhysicalField any() (fipy.variables.variable.Variable method), 351
method), 294 appendChild() (fipy.tools.Vitals method), 326
add() (fipy.tools.PhysicalField method), 317 appendChild() (fipy.tools.vitals.Vitals method), 316
AdvectionTerm, 157 appendInfo() (fipy.tools.Vitals method), 326
AdvectionTerm (class in fipy.terms), 270, 282 appendInfo() (fipy.tools.vitals.Vitals method), 316
AdvectionTerm (class in fipy.terms.advectionTerm), 251 arccos() (fipy.tools.dimensions.physicalField.PhysicalField
all() (fipy.tools.comms.commWrapper.CommWrapper method), 294
method), 289 arccos() (fipy.tools.PhysicalField method), 318
all() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper arccos() (fipy.variables.Variable method), 358
method), 290 arccos() (fipy.variables.variable.Variable method), 352
all() (fipy.variables.Variable method), 358 arccosh() (fipy.tools.dimensions.physicalField.PhysicalField
all() (fipy.variables.variable.Variable method), 351 method), 294
all() (in module fipy.tools.numerix), 309–311 arccosh() (fipy.tools.PhysicalField method), 318
allclose() (fipy.tools.comms.commWrapper.CommWrapper arccosh() (fipy.variables.Variable method), 358
method), 289 arccosh() (fipy.variables.variable.Variable method), 352
429
arcsin() (fipy.tools.dimensions.physicalField.PhysicalField C
method), 294 cacheMatrix, 169
arcsin() (fipy.tools.PhysicalField method), 318 cacheMatrix() (fipy.terms.term.Term method), 263
arcsin() (fipy.variables.Variable method), 358 cacheMe() (fipy.variables.Variable method), 358
arcsin() (fipy.variables.variable.Variable method), 352 cacheMe() (fipy.variables.variable.Variable method), 352
arcsinh() (fipy.variables.Variable method), 358 cacheRHSvector, 169
arcsinh() (fipy.variables.variable.Variable method), 352 cacheRHSvector() (fipy.terms.term.Term method), 263
arctan, 119 calcDistanceFunction() (fipy.variables.DistanceVariable
arctan() (fipy.tools.dimensions.physicalField.PhysicalField method), 382
method), 295 calcDistanceFunction() (fipy.variables.distanceVariable.DistanceVariable
arctan() (fipy.tools.PhysicalField method), 318 method), 337
arctan() (fipy.variables.Variable method), 358 ceil() (fipy.tools.dimensions.physicalField.PhysicalField
arctan() (fipy.variables.variable.Variable method), 352 method), 295
arctan2, 119 ceil() (fipy.tools.PhysicalField method), 319
arctan2() (fipy.tools.dimensions.physicalField.PhysicalField ceil() (fipy.variables.Variable method), 358
method), 295 ceil() (fipy.variables.variable.Variable method), 352
arctan2() (fipy.tools.PhysicalField method), 318 cellCenters (fipy.meshes.abstractMesh.AbstractMesh at-
arctan2() (fipy.variables.Variable method), 358 tribute), 203
arctan2() (fipy.variables.variable.Variable method), 352 cellCenters (fipy.meshes.PeriodicGrid1D attribute), 225
arctanh() (fipy.tools.dimensions.physicalField.PhysicalField cellCenters (fipy.meshes.periodicGrid1D.PeriodicGrid1D
method), 295 attribute), 219
arctanh() (fipy.tools.PhysicalField method), 319 cellDistanceVectors (fipy.meshes.abstractMesh.AbstractMesh
arctanh() (fipy.variables.Variable method), 358 attribute), 203
arctanh() (fipy.variables.variable.Variable method), 352 cellFaceIDs (fipy.meshes.abstractMesh.AbstractMesh at-
arithmeticFaceValue (fipy.variables.CellVariable attribute), 203
tribute), 364 cellInterfaceAreas (fipy.variables.DistanceVariable
arithmeticFaceValue (fipy.variables.cellVariable.CellVariable attribute), 382
attribute), 329 cellInterfaceAreas (fipy.variables.distanceVariable.DistanceVariable
arithmeticFaceValue (fipy.variables.ModularVariable at- attribute), 337
tribute), 370 CellTerm (class in fipy.terms.cellTerm), 254
arithmeticFaceValue (fipy.variables.modularVariable.ModularVariable
cellToFaceDistanceVectors
attribute), 346 (fipy.meshes.abstractMesh.AbstractMesh
array, 108 attribute), 203
aspect2D (fipy.meshes.abstractMesh.AbstractMesh at- CellVariable, 86, 91, 96, 104, 112, 122, 125, 155, 168,
tribute), 203 185, 187
CellVariable (class in fipy.variables), 363
B CellVariable (class in fipy.variables.cellVariable), 329
Barrier() (fipy.tools.comms.commWrapper.CommWrapper cellVolumeAverage (fipy.variables.CellVariable at-
method), 289 tribute), 364
Barrier() (fipy.tools.comms.dummyComm.DummyComm cellVolumeAverage (fipy.variables.cellVariable.CellVariable
method), 290 attribute), 330
Base (class in package.subpackage.base), 191 cellVolumes (fipy.meshes.abstractMesh.AbstractMesh at-
bcast() (fipy.tools.comms.commWrapper.CommWrapper tribute), 203
method), 289 cellVolumes (fipy.meshes.cylindricalUniformGrid1D.CylindricalUniformGr
bcast() (fipy.tools.comms.mpi4pyCommWrapper.Mpi4pyCommWrapper attribute), 208
method), 290 cellVolumes (fipy.meshes.cylindricalUniformGrid2D.CylindricalUniformGr
BetaNoiseVariable (class in fipy.variables), 370 attribute), 209
BetaNoiseVariable (class in CentralDifferenceConvectionTerm (class in fipy.terms),
fipy.variables.betaNoiseVariable), 327 274
BoundaryCondition (class in CentralDifferenceConvectionTerm (class in
fipy.boundaryConditions.boundaryCondition), fipy.terms.centralDiffConvectionTerm), 254
195 command line option
Buildbot, 49 –inline, 20
–lsmlib, 21
430 Index
–no-pysparse, 20 cosh() (fipy.tools.dimensions.physicalField.PhysicalField

–pyamg, 21 method), 296
–pysparse, 20 cosh() (fipy.tools.PhysicalField method), 320
–scipy, 20 cosh() (fipy.variables.Variable method), 360
–skfmm, 21 cosh() (fipy.variables.variable.Variable method), 353
–trilinos, 20 CylindricalGrid1D() (in module fipy.meshes), 224
CommWrapper (class in CylindricalGrid1D() (in module
fipy.tools.comms.commWrapper), 289 fipy.meshes.factoryMeshes), 211
conjugate() (fipy.tools.dimensions.physicalField.PhysicalFieldCylindricalGrid2D() (in module fipy.meshes), 224
method), 295 CylindricalGrid2D() (in module
conjugate() (fipy.tools.PhysicalField method), 319 fipy.meshes.factoryMeshes), 210
conjugate() (fipy.variables.Variable method), 359 CylindricalNonUniformGrid1D (class in
conjugate() (fipy.variables.variable.Variable method), 352 fipy.meshes.cylindricalNonUniformGrid1D),
constrain() (fipy.variables.CellVariable method), 364 207
constrain() (fipy.variables.cellVariable.CellVariable CylindricalNonUniformGrid2D (class in
method), 330 fipy.meshes.cylindricalNonUniformGrid2D),
constrain() (fipy.variables.Variable method), 359 208
constrain() (fipy.variables.variable.Variable method), 352 CylindricalUniformGrid1D (class in
Constraint (class in fipy.boundaryConditions), 196 fipy.meshes.cylindricalUniformGrid1D),
Constraint (class in fipy.boundaryConditions.constraint), 208
195 CylindricalUniformGrid2D (class in
constraints (fipy.variables.Variable attribute), 359 fipy.meshes.cylindricalUniformGrid2D),
constraints (fipy.variables.variable.Variable attribute), 353 209
ConvectionTerm (in module fipy.terms), 269
D
conversionFactorTo() (fipy.tools.dimensions.physicalField.PhysicalUnit
method), 303 DefaultAsymmetricSolver (in module fipy.solvers), 243
conversionTupleTo() (fipy.tools.dimensions.physicalField.PhysicalUnit
DefaultAsymmetricSolver (in module
method), 303 fipy.solvers.pysparse), 236
convertToUnit() (fipy.tools.dimensions.physicalField.PhysicalField
DefaultAsymmetricSolver (in module fipy.solvers.scipy),
method), 296 239
convertToUnit() (fipy.tools.PhysicalField method), 319 DefaultSolver (in module fipy.solvers), 243
copy() (fipy.terms.term.Term method), 263 DefaultSolver (in module fipy.solvers.pysparse), 236
copy() (fipy.tools.dimensions.physicalField.PhysicalField DefaultSolver (in module fipy.solvers.scipy), 239
method), 296 description (fipy.tools.copy_script.Copy_script attribute),
copy() (fipy.tools.PhysicalField method), 319 306
copy() (fipy.variables.CellVariable method), 365 description (fipy.tools.performance.efficiency_test.Efficiency_test
copy() (fipy.variables.cellVariable.CellVariable method), attribute), 305
331 dictToXML() (fipy.tools.Vitals method), 326
copy() (fipy.variables.FaceVariable method), 369 dictToXML() (fipy.tools.vitals.Vitals method), 316
copy() (fipy.variables.faceVariable.FaceVariable method), DiffusionTerm (class in fipy.terms), 273
340 DiffusionTerm (class in fipy.terms.diffusionTerm), 255
copy() (fipy.variables.noiseVariable.NoiseVariable DiffusionTermCorrection (class in fipy.terms), 274
method), 347 DiffusionTermCorrection (class in
copy() (fipy.variables.SurfactantVariable method), 378 fipy.terms.diffusionTerm), 255
copy() (fipy.variables.surfactantVariable.SurfactantVariable DiffusionTermCorrection (class in
method), 349 fipy.terms.diffusionTermCorrection), 256
copy() (fipy.variables.Variable method), 359 DiffusionTermNoCorrection (class in fipy.terms), 274
copy() (fipy.variables.variable.Variable method), 353 DiffusionTermNoCorrection (class in
Copy_script (class in fipy.tools.copy_script), 306 fipy.terms.diffusionTerm), 255
cos() (fipy.tools.dimensions.physicalField.PhysicalField DiffusionTermNoCorrection (class in
method), 296 fipy.terms.diffusionTermNoCorrection), 256
cos() (fipy.tools.PhysicalField method), 320 DistanceVariable, 155
cos() (fipy.variables.Variable method), 360 DistanceVariable (class in fipy.variables), 379
cos() (fipy.variables.variable.Variable method), 353
Index 431
DistanceVariable (class inexamples.diffusion.nthOrder.input4thOrder1D (module),

fipy.variables.distanceVariable), 334 85
divergence (fipy.variables.FaceVariable attribute), 369 examples.flow.stokesCavity (module), 167
divergence (fipy.variables.faceVariable.FaceVariable at- examples.levelSet.advection.circle (module), 142
tribute), 340 examples.levelSet.advection.mesh1D (module), 141
divide() (fipy.tools.dimensions.physicalField.PhysicalField
examples.levelSet.distanceFunction.circle (module), 140
method), 297 examples.levelSet.distanceFunction.mesh1D (module),
divide() (fipy.tools.PhysicalField method), 320 139
dontCacheMe() (fipy.variables.Variable method), 360 examples.levelSet.electroChem.gold (module), 148
dontCacheMe() (fipy.variables.variable.Variable method), examples.levelSet.electroChem.howToWriteAScript
353 (module), 153
dot() (fipy.tools.dimensions.physicalField.PhysicalField examples.levelSet.electroChem.leveler (module), 149
method), 297 examples.levelSet.electroChem.simpleTrenchSystem
dot() (fipy.tools.PhysicalField method), 320 (module), 145
dot() (fipy.variables.Variable method), 360 examples.phase.anisotropy (module), 118
dot() (fipy.variables.variable.Variable method), 353 examples.phase.binaryCoupled (module), 103
dot() (in module fipy.tools.numerix), 308, 312 examples.phase.impingement.mesh20x20 (module), 124
DummyComm (class inexamples.phase.impingement.mesh40x1 (module), 121
fipy.tools.comms.dummyComm), 290 examples.phase.polyxtal (module), 127
DummySolver (in module fipy.solvers), 243 examples.phase.polyxtalCoupled (module), 133
DummySolver (in module fipy.solvers.pysparse), 236 examples.phase.quaternary (module), 112
DummySolver (in module fipy.solvers.scipy), 239 examples.phase.simple (module), 95
DummyViewer (class in fipy.viewers), 421 examples.reactiveWetting.liquidVapor1D (module), 173
DYLD_LIBRARY_PATH, 15 examples.updating.update0_1to1_0 (module), 184
examples.updating.update1_0to2_0 (module), 180
E examples.updating.update2_0to3_0 (module), 179
Efficiency_test (class in execButNoTest() (in module fipy.tests.doctestPlus), 287
fipy.tools.performance.efficiency_test), 305 exp, 90, 91, 108, 123, 126, 154
environment variable exp() (fipy.variables.Variable method), 360
DYLD_LIBRARY_PATH, 15 exp() (fipy.variables.variable.Variable method), 353
FIPY_DISPLAY_MATRIX, 21 ExplicitDiffusionTerm, 56, 123, 125
FIPY_INCLUDE_NUMERIX_ALL, 21, 179 ExplicitDiffusionTerm (class in fipy.terms), 274
FIPY_INLINE, 21 ExplicitDiffusionTerm (class in
FIPY_INLINE_COMMENT, 21 fipy.terms.explicitDiffusionTerm), 256
FIPY_SOLVERS, 13, 20, 21 ExplicitUpwindConvectionTerm (class in fipy.terms), 275
FIPY_VERBOSE_SOLVER, 21 ExplicitUpwindConvectionTerm (class in
FIPY_VIEWER, 21 fipy.terms.explicitUpwindConvectionTerm),
LD_LIBRARY_PATH, 15 256
PYTHONPATH, 12, 15 ExplicitVariableError, 269
examples.cahnHilliard.mesh2DCoupled (module), 161 ExponentialConvectionTerm, 185
examples.cahnHilliard.sphere (module), 164 ExponentialConvectionTerm (class in fipy.terms), 276
examples.convection.exponential1D.mesh1D (module), ExponentialConvectionTerm (class in
89 fipy.terms.exponentialConvectionTerm),
examples.convection.exponential1DSource.mesh1D 257
(module), 90 ExponentialNoiseVariable (class in fipy.variables), 372
examples.convection.robin (module), 91 ExponentialNoiseVariable (class in
examples.convection.source (module), 93 fipy.variables.exponentialNoiseVariable),
examples.diffusion.anisotropy (module), 87 338
examples.diffusion.circle (module), 76 extendVariable() (fipy.variables.DistanceVariable
examples.diffusion.coupled (module), 72 method), 383
examples.diffusion.electrostatics (module), 81 extendVariable() (fipy.variables.distanceVariable.DistanceVariable
examples.diffusion.mesh1D (module), 55 method), 338
examples.diffusion.mesh20x20 (module), 74 extents (fipy.meshes.abstractMesh.AbstractMesh at-
tribute), 203
432 Index
exteriorFaces (fipy.meshes.abstractMesh.AbstractMesh faceValue (fipy.variables.cellVariable.CellVariable

attribute), 203 attribute), 331
exteriorFaces (fipy.meshes.uniformGrid1D.UniformGrid1D FaceVariable, 62
attribute), 222 FaceVariable (class in fipy.variables), 369
extrude() (fipy.meshes.mesh2D.Mesh2D method), 216 FaceVariable (class in fipy.variables.faceVariable), 340
faceVertexIDs (fipy.meshes.uniformGrid2D.UniformGrid2D
F attribute), 222
faceCellIDs (fipy.meshes.uniformGrid1D.UniformGrid1D faceVertexIDs (fipy.meshes.uniformGrid3D.UniformGrid3D
faceCellIDs (fipy.meshes.uniformGrid2D.UniformGrid2D failFn() (fipy.steppers.stepper.Stepper static method), 247
attribute), 222 figaspect() (fipy.viewers.matplotlibViewer.matplotlibViewer.AbstractMatplo
faceCellIDs (fipy.meshes.uniformGrid3D.UniformGrid3D method), 395
attribute), 223 finalize_options() (fipy.tools.copy_script.Copy_script
faceCenters (fipy.meshes.abstractMesh.AbstractMesh at- method), 306
tribute), 203 finalize_options() (fipy.tools.performance.efficiency_test.Efficiency_test
faceGrad (fipy.variables.CellVariable attribute), 365 method), 305
faceGrad (fipy.variables.cellVariable.CellVariable at- FiPy, 49
tribute), 331 fipy.boundaryConditions (module), 196
faceGrad (fipy.variables.ModularVariable attribute), 370 fipy.boundaryConditions.boundaryCondition (module),
faceGrad (fipy.variables.modularVariable.ModularVariable 195
attribute), 346 fipy.boundaryConditions.constraint (module), 195
faceGradAverage (fipy.variables.CellVariable attribute), fipy.boundaryConditions.fixedFlux (module), 195
365 fipy.boundaryConditions.fixedValue (module), 196
faceGradAverage (fipy.variables.cellVariable.CellVariable fipy.boundaryConditions.nthOrderBoundaryCondition
attribute), 331 (module), 196
faceGradNoMod (fipy.variables.ModularVariable at- fipy.boundaryConditions.test (module), 196
tribute), 370 fipy.matrices (module), 199
fipy.matrices.offsetSparseMatrix (module), 199
faceGradNoMod (fipy.variables.modularVariable.ModularVariable
attribute), 346 fipy.matrices.pysparseMatrix (module), 199
faceNormals (fipy.meshes.uniformGrid1D.UniformGrid1D fipy.matrices.scipyMatrix (module), 199
attribute), 222 fipy.matrices.sparseMatrix (module), 199
faceNormals (fipy.meshes.uniformGrid2D.UniformGrid2D fipy.matrices.test (module), 199
attribute), 222 fipy.meshes (module), 223
faceNormals (fipy.meshes.uniformGrid3D.UniformGrid3D fipy.meshes.abstractMesh (module), 203
attribute), 223 fipy.meshes.builders (module), 201
facesBack (fipy.meshes.abstractMesh.AbstractMesh at- fipy.meshes.builders.abstractGridBuilder (module), 201
tribute), 203 fipy.meshes.builders.grid1DBuilder (module), 201
facesBottom (fipy.meshes.abstractMesh.AbstractMesh at- fipy.meshes.builders.grid2DBuilder (module), 201
tribute), 203 fipy.meshes.builders.grid3DBuilder (module), 201
facesDown (fipy.meshes.abstractMesh.AbstractMesh at- fipy.meshes.builders.periodicGrid1DBuilder (module),
tribute), 203 201
facesFront (fipy.meshes.abstractMesh.AbstractMesh at- fipy.meshes.builders.utilityClasses (module), 201
tribute), 204 fipy.meshes.cylindricalNonUniformGrid1D (module),
facesLeft (fipy.meshes.abstractMesh.AbstractMesh at- 207
tribute), 204 fipy.meshes.cylindricalNonUniformGrid2D (module),
facesRight (fipy.meshes.abstractMesh.AbstractMesh at- 208
tribute), 204 fipy.meshes.cylindricalUniformGrid1D (module), 208
facesTop (fipy.meshes.abstractMesh.AbstractMesh fipy.meshes.cylindricalUniformGrid2D (module), 209
attribute), 204 fipy.meshes.factoryMeshes (module), 209
facesUp (fipy.meshes.abstractMesh.AbstractMesh at- fipy.meshes.gmshMesh (module), 211
tribute), 204 fipy.meshes.grid1D (module), 216
FaceTerm (class in fipy.terms.faceTerm), 258 fipy.meshes.grid2D (module), 216
faceValue (fipy.variables.CellVariable attribute), 365 fipy.meshes.grid3D (module), 216
fipy.meshes.mesh (module), 216
Index 433
fipy.meshes.mesh1D (module), 216 fipy.solvers.pysparse.preconditioners.ssorPreconditioner

fipy.meshes.mesh2D (module), 216 (module), 234
fipy.meshes.nonUniformGrid1D (module), 217 fipy.solvers.pysparse.pysparseSolver (module), 236
fipy.meshes.nonUniformGrid2D (module), 217 fipy.solvers.pysparseMatrixSolver (module), 242
fipy.meshes.nonUniformGrid3D (module), 218 fipy.solvers.scipy (module), 239
fipy.meshes.numMesh (module), 202 fipy.solvers.scipy.linearBicgstabSolver (module), 238
fipy.meshes.numMesh.cylindricalUniformGrid1D (mod- fipy.solvers.scipy.linearCGSSolver (module), 238
ule), 201 fipy.solvers.scipy.linearGMRESSolver (module), 238
fipy.meshes.numMesh.cylindricalUniformGrid2D (mod- fipy.solvers.scipy.linearLUSolver (module), 239
ule), 201 fipy.solvers.scipy.linearPCGSolver (module), 239
fipy.meshes.numMesh.deprecatedWarning (module), 201 fipy.solvers.scipy.scipyKrylovSolver (module), 239
fipy.meshes.numMesh.gmshImport (module), 202 fipy.solvers.scipy.scipySolver (module), 239
fipy.meshes.numMesh.grid1D (module), 202 fipy.solvers.solver (module), 242
fipy.meshes.numMesh.grid2D (module), 202 fipy.solvers.test (module), 243
fipy.meshes.numMesh.grid3D (module), 202 fipy.steppers (module), 248
fipy.meshes.numMesh.periodicGrid1D (module), 202 fipy.steppers.pidStepper (module), 247
fipy.meshes.numMesh.periodicGrid2D (module), 202 fipy.steppers.pseudoRKQSStepper (module), 247
fipy.meshes.numMesh.skewedGrid2D (module), 202 fipy.steppers.stepper (module), 247
fipy.meshes.numMesh.tri2D (module), 202 fipy.terms (module), 269
fipy.meshes.numMesh.uniformGrid1D (module), 202 fipy.terms.abstractBinaryTerm (module), 251
fipy.meshes.numMesh.uniformGrid2D (module), 202 fipy.terms.abstractConvectionTerm (module), 251
fipy.meshes.numMesh.uniformGrid3D (module), 202 fipy.terms.abstractDiffusionTerm (module), 251
fipy.meshes.periodicGrid1D (module), 218 fipy.terms.abstractUpwindConvectionTerm (module),
fipy.meshes.periodicGrid2D (module), 219 251
fipy.meshes.representations (module), 202 fipy.terms.advectionTerm (module), 251
fipy.meshes.representations.abstractRepresentation (mod- fipy.terms.asymmetricConvectionTerm (module), 254
ule), 202 fipy.terms.binaryTerm (module), 254
fipy.meshes.representations.gridRepresentation (module), fipy.terms.cellTerm (module), 254
202 fipy.terms.centralDiffConvectionTerm (module), 254
fipy.meshes.representations.meshRepresentation (mod- fipy.terms.coupledBinaryTerm (module), 255
ule), 202 fipy.terms.diffusionTerm (module), 255
fipy.meshes.skewedGrid2D (module), 220 fipy.terms.diffusionTermCorrection (module), 256
fipy.meshes.test (module), 220 fipy.terms.diffusionTermNoCorrection (module), 256
fipy.meshes.topologies (module), 203 fipy.terms.explicitDiffusionTerm (module), 256
fipy.meshes.topologies.abstractTopology (module), 202 fipy.terms.explicitSourceTerm (module), 256
fipy.meshes.topologies.gridTopology (module), 202 fipy.terms.explicitUpwindConvectionTerm (module), 256
fipy.meshes.topologies.meshTopology (module), 202 fipy.terms.exponentialConvectionTerm (module), 257
fipy.meshes.tri2D (module), 221 fipy.terms.faceTerm (module), 258
fipy.meshes.uniformGrid (module), 221 fipy.terms.firstOrderAdvectionTerm (module), 259
fipy.meshes.uniformGrid1D (module), 221 fipy.terms.hybridConvectionTerm (module), 260
fipy.meshes.uniformGrid2D (module), 222 fipy.terms.implicitDiffusionTerm (module), 261
fipy.meshes.uniformGrid3D (module), 222 fipy.terms.implicitSourceTerm (module), 261
fipy.solvers (module), 243 fipy.terms.nonDiffusionTerm (module), 261
fipy.solvers.pysparse (module), 236 fipy.terms.powerLawConvectionTerm (module), 261
fipy.solvers.pysparse.linearCGSSolver (module), 234 fipy.terms.residualTerm (module), 262
fipy.solvers.pysparse.linearGMRESSolver (module), 235 fipy.terms.sourceTerm (module), 263
fipy.solvers.pysparse.linearJORSolver (module), 235 fipy.terms.term (module), 263
fipy.solvers.pysparse.linearLUSolver (module), 235 fipy.terms.test (module), 265
fipy.solvers.pysparse.linearPCGSolver (module), 236 fipy.terms.transientTerm (module), 265
fipy.solvers.pysparse.preconditioners (module), 234 fipy.terms.unaryTerm (module), 266
fipy.solvers.pysparse.preconditioners.jacobiPreconditioner fipy.terms.upwindConvectionTerm (module), 266
(module), 234 fipy.terms.vanLeerConvectionTerm (module), 268
fipy.solvers.pysparse.preconditioners.preconditioner fipy.tests (module), 288
(module), 234 fipy.tests.doctestPlus (module), 287
434 Index
fipy.tests.lateImportTest (module), 288 fipy.variables.faceVariable (module), 340

fipy.tests.testBase (module), 288 fipy.variables.gammaNoiseVariable (module), 340
fipy.tests.testClass (module), 288 fipy.variables.gaussCellGradVariable (module), 342
fipy.tests.testProgram (module), 288 fipy.variables.gaussianNoiseVariable (module), 342
fipy.tools (module), 316 fipy.variables.harmonicCellToFaceVariable (module),
fipy.tools.comms (module), 290 344
fipy.tools.comms.commWrapper (module), 289 fipy.variables.histogramVariable (module), 344
fipy.tools.comms.dummyComm (module), 290 fipy.variables.interfaceAreaVariable (module), 345
fipy.tools.comms.mpi4pyCommWrapper (module), 290 fipy.variables.interfaceFlagVariable (module), 345
fipy.tools.comms.serialCommWrapper (module), 290 fipy.variables.leastSquaresCellGradVariable (module),
fipy.tools.copy_script (module), 306 345
fipy.tools.debug (module), 306 fipy.variables.levelSetDiffusionVariable (module), 345
fipy.tools.decorators (module), 306 fipy.variables.meshVariable (module), 345
fipy.tools.dimensions (module), 305 fipy.variables.minmodCellToFaceVariable (module), 345
fipy.tools.dimensions.DictWithDefault (module), 290 fipy.variables.modCellGradVariable (module), 345
fipy.tools.dimensions.NumberDict (module), 290 fipy.variables.modCellToFaceVariable (module), 345
fipy.tools.dimensions.physicalField (module), 290 fipy.variables.modFaceGradVariable (module), 345
fipy.tools.dump fipy.variables.modPhysicalField (module), 345
module, 127 fipy.variables.modularVariable (module), 345
fipy.tools.dump (module), 307 fipy.variables.noiseVariable (module), 346
fipy.tools.inline (module), 308 fipy.variables.operatorVariable (module), 347
fipy.tools.numerix (module), 308 fipy.variables.scharfetterGummelFaceVariable (module),
fipy.tools.parser 347
module, 124, 154 fipy.variables.surfactantConvectionVariable (module),
fipy.tools.parser (module), 315 347
fipy.tools.performance (module), 306 fipy.variables.surfactantVariable (module), 348
fipy.tools.performance.efficiency_test (module), 305 fipy.variables.test (module), 349
fipy.tools.performance.efficiencyTestGenerator (module), fipy.variables.unaryOperatorVariable (module), 349
305 fipy.variables.uniformNoiseVariable (module), 349
fipy.tools.performance.memoryLeak (module), 305 fipy.variables.variable (module), 350
fipy.tools.performance.memoryLogger (module), 305 fipy.viewers, 187
fipy.tools.performance.memoryUsage (module), 306 module, 56, 74, 86, 90, 91, 96, 109, 123, 126, 169
fipy.tools.test (module), 316 fipy.viewers (module), 410
fipy.tools.vector (module), 316 fipy.viewers.gistViewer (module), 385
fipy.tools.vitals (module), 316 fipy.viewers.gistViewer.gist1DViewer (module), 385
fipy.variables (module), 357 fipy.viewers.gistViewer.gist2DViewer (module), 385
fipy.variables.addOverFacesVariable (module), 327 fipy.viewers.gistViewer.gistVectorViewer (module), 385
fipy.variables.arithmeticCellToFaceVariable (module), fipy.viewers.gistViewer.gistViewer (module), 385
327 fipy.viewers.gistViewer.test (module), 385
fipy.variables.betaNoiseVariable (module), 327 fipy.viewers.gnuplotViewer (module), 386
fipy.variables.binaryOperatorVariable (module), 329 fipy.viewers.gnuplotViewer.gnuplot1DViewer (module),
fipy.variables.cellToFaceVariable (module), 329 386
fipy.variables.cellVariable (module), 329 fipy.viewers.gnuplotViewer.gnuplot2DViewer (module),
fipy.variables.cellVariable.CellVariable 386
object, 77 fipy.viewers.gnuplotViewer.gnuplotViewer (module), 386
fipy.variables.cellVolumeAverageVariable (module), 334 fipy.viewers.gnuplotViewer.test (module), 386
fipy.variables.constant (module), 334 fipy.viewers.matplotlibViewer (module), 395
fipy.variables.constraintMask (module), 334 fipy.viewers.matplotlibViewer.matplotlib1DViewer
fipy.variables.coupledCellVariable (module), 334 (module), 386
fipy.variables.distanceVariable (module), 334 fipy.viewers.matplotlibViewer.matplotlib2DGridContourViewer
fipy.variables.exponentialNoiseVariable (module), 338 (module), 387
fipy.variables.faceGradContributionsVariable (module), fipy.viewers.matplotlibViewer.matplotlib2DGridViewer
340 (module), 388
fipy.variables.faceGradVariable (module), 340
Index 435
fipy.viewers.matplotlibViewer.matplotlib2DViewer GaussianNoiseVariable (class in

(module), 389 fipy.variables.gaussianNoiseVariable), 342
fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer GeneralSolver (in module fipy.solvers), 243
(module), 390 GeneralSolver (in module fipy.solvers.pysparse), 236
fipy.viewers.matplotlibViewer.matplotlibStreamViewer GeneralSolver (in module fipy.solvers.scipy), 239
(module), 391 getArithmeticFaceValue() (fipy.variables.CellVariable
fipy.viewers.matplotlibViewer.matplotlibVectorViewer method), 366
(module), 393 getArithmeticFaceValue()
fipy.viewers.matplotlibViewer.matplotlibViewer (mod- (fipy.variables.cellVariable.CellVariable
ule), 394 method), 331
fipy.viewers.matplotlibViewer.test (module), 395 getCellCenters() (fipy.meshes.abstractMesh.AbstractMesh
fipy.viewers.mayaviViewer (module), 404 method), 205
fipy.viewers.mayaviViewer.mayaviClient (module), 403 getCellInterfaceAreas() (fipy.variables.DistanceVariable
fipy.viewers.mayaviViewer.test (module), 404 method), 383
fipy.viewers.multiViewer (module), 408 getCellInterfaceAreas() (fipy.variables.distanceVariable.DistanceVariable
fipy.viewers.test (module), 408 method), 338
fipy.viewers.testinteractive (module), 408 getCellVolumeAverage() (fipy.variables.CellVariable
fipy.viewers.tsvViewer (module), 408 method), 366
fipy.viewers.tsvViewer.TSVViewer getCellVolumeAverage() (fipy.variables.cellVariable.CellVariable
object, 79 method), 331
fipy.viewers.viewer (module), 409 getCellVolumes() (fipy.meshes.abstractMesh.AbstractMesh
fipy.viewers.vtkViewer (module), 407 method), 205
fipy.viewers.vtkViewer.test (module), 406 getDefaultSolver() (fipy.terms.term.Term method), 263
fipy.viewers.vtkViewer.vtkCellViewer (module), 406 getDim() (fipy.meshes.abstractMesh.AbstractMesh
fipy.viewers.vtkViewer.vtkFaceViewer (module), 406 method), 205
fipy.viewers.vtkViewer.vtkViewer (module), 406 getDivergence() (fipy.variables.FaceVariable method),
FIPY_INCLUDE_NUMERIX_ALL, 179 369
FIPY_SOLVERS, 13, 20 getDivergence() (fipy.variables.faceVariable.FaceVariable
FirstOrderAdvectionTerm (class in fipy.terms), 269, 281 method), 340
FirstOrderAdvectionTerm (class in getExteriorFaces() (fipy.meshes.abstractMesh.AbstractMesh
fipy.terms.firstOrderAdvectionTerm), 259 method), 205
FixedFlux, 184 getFaceCellIDs() (fipy.meshes.abstractMesh.AbstractMesh
FixedFlux (class in fipy.boundaryConditions), 196 method), 205
FixedFlux (class in fipy.boundaryConditions.fixedFlux), getFaceCenters() (fipy.meshes.abstractMesh.AbstractMesh
195 method), 205
FixedValue, 184, 186 getFaceGrad() (fipy.variables.CellVariable method), 366
FixedValue (class in fipy.boundaryConditions), 197 getFaceGrad() (fipy.variables.cellVariable.CellVariable
FixedValue (class in fipy.boundaryConditions.fixedValue), method), 331
196 getFaceGradAverage() (fipy.variables.CellVariable
floor() (fipy.tools.dimensions.physicalField.PhysicalField method), 366
method), 297 getFaceGradAverage() (fipy.variables.cellVariable.CellVariable
floor() (fipy.tools.PhysicalField method), 320 method), 331
floor() (fipy.variables.Variable method), 360 getFaceGradNoMod() (fipy.variables.ModularVariable
floor() (fipy.variables.variable.Variable method), 353 method), 370
getFaceGradNoMod() (fipy.variables.modularVariable.ModularVariable
G method), 346
GammaNoiseVariable (class in fipy.variables), 373 getFacesBack() (fipy.meshes.abstractMesh.AbstractMesh
GammaNoiseVariable (class in method), 205
fipy.variables.gammaNoiseVariable), 340 getFacesBottom() (fipy.meshes.abstractMesh.AbstractMesh
gaussGrad (fipy.variables.CellVariable attribute), 366 method), 205
gaussGrad (fipy.variables.cellVariable.CellVariable at- getFacesDown() (fipy.meshes.abstractMesh.AbstractMesh
tribute), 331 method), 205
GaussianNoiseVariable (class in fipy.variables), 374 getFacesFront() (fipy.meshes.abstractMesh.AbstractMesh
method), 205
436 Index
getFacesLeft() (fipy.meshes.abstractMesh.AbstractMesh getNumericValue() (fipy.tools.dimensions.physicalField.PhysicalField

method), 205 method), 297
getFacesRight() (fipy.meshes.abstractMesh.AbstractMesh getNumericValue() (fipy.tools.PhysicalField method),
method), 205 321
getFacesTop() (fipy.meshes.abstractMesh.AbstractMesh getNumericValue() (fipy.variables.Variable method), 360
method), 205 getNumericValue() (fipy.variables.variable.Variable
getFacesUp() (fipy.meshes.abstractMesh.AbstractMesh method), 354
method), 205 getOld() (fipy.variables.CellVariable method), 366
getFaceValue() (fipy.variables.CellVariable method), 366 getOld() (fipy.variables.cellVariable.CellVariable
getFaceValue() (fipy.variables.cellVariable.CellVariable method), 332
method), 331 getPhysicalShape() (fipy.meshes.abstractMesh.AbstractMesh
getGaussGrad() (fipy.variables.CellVariable method), 366 method), 206
getGaussGrad() (fipy.variables.cellVariable.CellVariable getRHSvector, 169
method), 332 getRHSvector() (fipy.terms.term.Term method), 263
getGrad() (fipy.variables.CellVariable method), 366 getScale() (fipy.meshes.abstractMesh.AbstractMesh
getGrad() (fipy.variables.cellVariable.CellVariable method), 206
method), 332 getsctype() (fipy.tools.dimensions.physicalField.PhysicalField
getHarmonicFaceValue() (fipy.variables.CellVariable method), 297
method), 366 getsctype() (fipy.tools.PhysicalField method), 321
getHarmonicFaceValue() getsctype() (fipy.variables.Variable method), 360
(fipy.variables.cellVariable.CellVariable getsctype() (fipy.variables.variable.Variable method), 354
method), 332 getsetDeprecated() (in module fipy.tools.decorators), 306
getInterfaceVar() (fipy.variables.SurfactantVariable getShape() (fipy.meshes.abstractMesh.AbstractMesh
getInterfaceVar() (fipy.variables.surfactantVariable.SurfactantVariable
getShape() (fipy.tools.dimensions.physicalField.PhysicalField
getInteriorFaceCellIDs() (fipy.meshes.abstractMesh.AbstractMesh
getShape() (fipy.tools.PhysicalField method), 321
method), 205 getShape() (fipy.variables.Variable method), 360
getInteriorFaceIDs() (fipy.meshes.abstractMesh.AbstractMesh getShape() (fipy.variables.variable.Variable method), 354
method), 205 getShape() (in module fipy.tools.numerix), 312
getInteriorFaces() (fipy.meshes.abstractMesh.AbstractMesh getSubscribedVariables() (fipy.variables.Variable
getLeastSquaresGrad() (fipy.variables.CellVariable getSubscribedVariables() (fipy.variables.variable.Variable
getLeastSquaresGrad() (fipy.variables.cellVariable.CellVariable
getUnit() (fipy.tools.dimensions.physicalField.PhysicalField
getLSMshape() (fipy.variables.DistanceVariable method), getUnit() (fipy.tools.PhysicalField method), 321
383 getUnit() (fipy.variables.Variable method), 360
getLSMshape() (fipy.variables.distanceVariable.DistanceVariable
getUnit() (fipy.variables.variable.Variable method), 354
method), 338 getUnit() (in module fipy.tools.numerix), 313
getMag() (fipy.variables.Variable method), 360 getValue() (fipy.variables.Variable method), 360
getMag() (fipy.variables.variable.Variable method), 353 getValue() (fipy.variables.variable.Variable method), 354
getMatrix, 169 getVars() (fipy.viewers.viewer.AbstractViewer method),
getMatrix() (fipy.terms.term.Term method), 263 410
getMinmodFaceValue() (fipy.variables.CellVariable getVertexCoords() (fipy.meshes.abstractMesh.AbstractMesh
getMinmodFaceValue() (fipy.variables.cellVariable.CellVariablegetViewers() (fipy.viewers.MultiViewer method), 419
method), 332 getViewers() (fipy.viewers.multiViewer.MultiViewer
getName() (fipy.variables.Variable method), 360 method), 408
getName() (fipy.variables.variable.Variable method), 353 Gist1DViewer, 187
getNearestCell() (fipy.meshes.abstractMesh.AbstractMesh GistViewer() (in module fipy.viewers), 410
method), 206 GistViewer() (in module fipy.viewers.gistViewer), 385
getNumberOfCells() (fipy.meshes.abstractMesh.AbstractMesh globalValue (fipy.variables.CellVariable attribute), 366
method), 206
Index 437
globalValue (fipy.variables.cellVariable.CellVariable at- HybridConvectionTerm (class in

tribute), 332 fipy.terms.hybridConvectionTerm), 260
globalValue (fipy.variables.FaceVariable attribute), 369
globalValue (fipy.variables.faceVariable.FaceVariable at- I
tribute), 340 IllConditionedPreconditionerWarning, 242, 243
Gmsh, 49 ImplicitDiffusionTerm (in module fipy.terms), 274
gmsh, 146, 148, 150 ImplicitDiffusionTerm (in module
Gmsh2D (class in fipy.meshes), 227 fipy.terms.implicitDiffusionTerm), 261
Gmsh2D (class in fipy.meshes.gmshMesh), 211 ImplicitSourceTerm, 99, 123, 125
Gmsh2DIn3DSpace (class in fipy.meshes), 231 ImplicitSourceTerm (class in fipy.terms), 274
Gmsh2DIn3DSpace (class in fipy.meshes.gmshMesh), ImplicitSourceTerm (class in
215 fipy.terms.implicitSourceTerm), 261
Gmsh3D (class in fipy.meshes), 231 inBaseUnits() (fipy.tools.dimensions.physicalField.PhysicalField
Gmsh3D (class in fipy.meshes.gmshMesh), 215 method), 297
GmshGrid2D (class in fipy.meshes), 231 inBaseUnits() (fipy.tools.PhysicalField method), 321
GmshGrid2D (class in fipy.meshes.gmshMesh), 215 inBaseUnits() (fipy.variables.Variable method), 360
GmshGrid3D (class in fipy.meshes), 231 inBaseUnits() (fipy.variables.variable.Variable method),
GmshGrid3D (class in fipy.meshes.gmshMesh), 215 354
GmshImporter2D (class in fipy.meshes), 231 IncorrectSolutionVariable, 269
GmshImporter2D (class in fipy.meshes.gmshMesh), 215 indices() (in module fipy.tools.numerix), 309, 313
GmshImporter2DIn3DSpace (class in fipy.meshes), 231 inDimensionless() (fipy.tools.dimensions.physicalField.PhysicalField
GmshImporter2DIn3DSpace (class in method), 297
fipy.meshes.gmshMesh), 215 inDimensionless() (fipy.tools.PhysicalField method), 321
GmshImporter3D (class in fipy.meshes), 231 initialize_options() (fipy.tools.copy_script.Copy_script
GmshImporter3D (class in fipy.meshes.gmshMesh), 215 method), 306
GnuplotViewer() (in module fipy.viewers), 410 initialize_options() (fipy.tools.performance.efficiency_test.Efficiency_test
GnuplotViewer() (in module method), 305
fipy.viewers.gnuplotViewer), 386 inRadians() (fipy.tools.dimensions.physicalField.PhysicalField
grad (fipy.variables.CellVariable attribute), 366 method), 298
grad (fipy.variables.cellVariable.CellVariable attribute), inRadians() (fipy.tools.PhysicalField method), 321
332 inSIUnits() (fipy.tools.dimensions.physicalField.PhysicalField
grad (fipy.variables.ModularVariable attribute), 370 method), 298
grad (fipy.variables.modularVariable.ModularVariable at- inSIUnits() (fipy.tools.PhysicalField method), 322
tribute), 346 interfaceVar (fipy.variables.SurfactantVariable attribute),
Grid1D, 85, 89, 90, 96, 103, 112, 121, 142, 186 378
Grid1D() (in module fipy.meshes), 224 interfaceVar (fipy.variables.surfactantVariable.SurfactantVariable
Grid1D() (in module fipy.meshes.factoryMeshes), 210 attribute), 349
Grid2D, 74, 124, 139, 140, 155, 168, 184 interiorFaceCellIDs (fipy.meshes.abstractMesh.AbstractMesh
Grid2D() (in module fipy.meshes), 223 attribute), 206
Grid2D() (in module fipy.meshes.factoryMeshes), 209 interiorFaceIDs (fipy.meshes.abstractMesh.AbstractMesh
Grid2DGistViewer, 186 attribute), 206
Grid3D() (in module fipy.meshes), 223 interiorFaces (fipy.meshes.abstractMesh.AbstractMesh
Grid3D() (in module fipy.meshes.factoryMeshes), 209 attribute), 206
inUnitsOf() (fipy.tools.dimensions.physicalField.PhysicalField
H method), 298
harmonicFaceValue (fipy.variables.CellVariable at- inUnitsOf() (fipy.tools.PhysicalField method), 322
tribute), 367 inUnitsOf() (fipy.variables.Variable method), 361
harmonicFaceValue (fipy.variables.cellVariable.CellVariableinUnitsOf() (fipy.variables.variable.Variable method),
attribute), 332 354
HistogramVariable (class in fipy.variables), 377 IPython, 49
HistogramVariable (class in isAngle() (fipy.tools.dimensions.physicalField.PhysicalUnit
fipy.variables.histogramVariable), 344 method), 303
HybridConvectionTerm (class in fipy.terms), 277 isclose() (in module fipy.tools.numerix), 309, 313
438 Index
isCompatible() (fipy.tools.dimensions.physicalField.PhysicalField
LinearCGSSolver (class in fipy.solvers.scipy), 239
method), 299 LinearCGSSolver (class in
isCompatible() (fipy.tools.dimensions.physicalField.PhysicalUnit fipy.solvers.scipy.linearCGSSolver), 238
method), 304 LinearGMRESSolver (class in fipy.solvers), 244
isCompatible() (fipy.tools.PhysicalField method), 322 LinearGMRESSolver (class in fipy.solvers.pysparse), 237
isDimensionless() (fipy.tools.dimensions.physicalField.PhysicalUnit
LinearGMRESSolver (class in
method), 304 fipy.solvers.pysparse.linearGMRESSolver),
isDimensionlessOrAngle() 235
(fipy.tools.dimensions.physicalField.PhysicalUnit LinearGMRESSolver (class in fipy.solvers.scipy), 240
method), 304 LinearGMRESSolver (class in
isFloat() (in module fipy.tools.numerix), 313 fipy.solvers.scipy.linearGMRESSolver), 238
isInt() (in module fipy.tools.numerix), 313 LinearJORSolver (class in fipy.solvers), 244
isInverseAngle() (fipy.tools.dimensions.physicalField.PhysicalUnit
LinearJORSolver (class in fipy.solvers.pysparse), 237
method), 304 LinearJORSolver (class in
itemset() (fipy.tools.dimensions.physicalField.PhysicalField fipy.solvers.pysparse.linearJORSolver), 235
method), 299 LinearLUSolver, 109, 185
itemset() (fipy.tools.PhysicalField method), 322 LinearLUSolver (class in fipy.solvers), 244
itemset() (fipy.variables.Variable method), 361 LinearLUSolver (class in fipy.solvers.pysparse), 237
itemset() (fipy.variables.variable.Variable method), 354 LinearLUSolver (class in
itemsize (fipy.tools.dimensions.physicalField.PhysicalField fipy.solvers.pysparse.linearLUSolver), 235
attribute), 299 LinearLUSolver (class in fipy.solvers.scipy), 240
itemsize (fipy.tools.PhysicalField attribute), 322 LinearLUSolver (class in
itemsize (fipy.variables.Variable attribute), 361 fipy.solvers.scipy.linearLUSolver), 239
itemsize (fipy.variables.variable.Variable attribute), 354 LinearPCGSolver (class in fipy.solvers), 244
Iterator, 185 LinearPCGSolver (class in fipy.solvers.pysparse), 236
LinearPCGSolver (class in
J fipy.solvers.pysparse.linearPCGSolver), 236
JacobiPreconditioner (class in fipy.solvers), 245 LinearPCGSolver (class in fipy.solvers.scipy), 240
JacobiPreconditioner (class in fipy.solvers.pysparse), 237 LinearPCGSolver (class in
JacobiPreconditioner (class in fipy.solvers.scipy.linearPCGSolver), 239
fipy.solvers.pysparse.preconditioners), 234 LINFerror() (in module fipy.steppers), 248
JacobiPreconditioner (class in LINFnorm() (in module fipy.tools.numerix), 313
loadtxt, 124, 126, 159
fipy.solvers.pysparse.preconditioners.jacobiPreconditioner),
234 log, 108, 114
justErrorVector() (fipy.terms.term.Term method), 263 log (fipy.viewers.Matplotlib1DViewer attribute), 413
justResidualVector() (fipy.terms.term.Term method), 264 log (fipy.viewers.matplotlibViewer.Matplotlib1DViewer
attribute), 397
L log (fipy.viewers.matplotlibViewer.matplotlib1DViewer.Matplotlib1DViewe
L1error() (in module fipy.steppers), 248 attribute), 387
L1norm() (in module fipy.tools.numerix), 313 log (fipy.viewers.matplotlibViewer.matplotlibViewer.AbstractMatplotlibView
L2error() (in module fipy.steppers), 248 attribute), 395
L2norm() (in module fipy.tools.numerix), 313 log() (fipy.tools.dimensions.physicalField.PhysicalField
LD_LIBRARY_PATH, 15 method), 299
leastSquaresGrad (fipy.variables.CellVariable attribute), log() (fipy.tools.PhysicalField method), 322
367 log() (fipy.variables.Variable method), 361
leastSquaresGrad (fipy.variables.cellVariable.CellVariable log() (fipy.variables.variable.Variable method), 354
attribute), 332 log10() (fipy.tools.dimensions.physicalField.PhysicalField
LinearBicgstabSolver (class in fipy.solvers.scipy), 240 method), 299
LinearBicgstabSolver (class in log10() (fipy.tools.PhysicalField method), 323
fipy.solvers.scipy.linearBicgstabSolver), 238 log10() (fipy.variables.Variable method), 361
LinearCGSSolver (class in fipy.solvers), 243 log10() (fipy.variables.variable.Variable method), 354
LinearCGSSolver (class in fipy.solvers.pysparse), 236
LinearCGSSolver (class in M
fipy.solvers.pysparse.linearCGSSolver), 234 mag (fipy.variables.Variable attribute), 361
Index 439
mag (fipy.variables.variable.Variable attribute), 355 method), 290

mathMethodDeprecated() (in module MaximumIterationWarning, 242, 243
fipy.tools.decorators), 307 MayaVi, 49
Matplotlib, 49 Mayavi, 49
Matplotlib1DViewer (class in fipy.viewers), 412 MayaviClient (class in fipy.viewers), 418
Matplotlib1DViewer (class in MayaviClient (class in fipy.viewers.mayaviViewer), 404
fipy.viewers.matplotlibViewer), 396 MayaviClient (class in
Matplotlib1DViewer (class in fipy.viewers.mayaviViewer.mayaviClient),
fipy.viewers.matplotlibViewer.matplotlib1DViewer), 403
386 MayaviSurfactantViewer, 158
Matplotlib2DGridContourViewer (class in fipy.viewers), MemoryHighWaterThread (class in
413 fipy.tools.performance.memoryLogger), 305
Matplotlib2DGridContourViewer (class in MemoryLogger (class in
fipy.viewers.matplotlibViewer), 398 fipy.tools.performance.memoryLogger), 305
Matplotlib2DGridContourViewer (class in Mesh (class in fipy.meshes.mesh), 216
fipy.viewers.matplotlibViewer.matplotlib2DGridContourViewer),
Mesh1D (class in fipy.meshes.mesh1D), 216
387 Mesh2D (class in fipy.meshes.mesh2D), 216
Matplotlib2DGridViewer (class in fipy.viewers), 413 MeshAdditionError, 216
Matplotlib2DGridViewer (class in MeshDimensionError, 421
fipy.viewers.matplotlibViewer), 397 method1() (package.subpackage.base.Base method), 192
Matplotlib2DGridViewer (class in method2() (package.subpackage.base.Base method), 192
fipy.viewers.matplotlibViewer.matplotlib2DGridViewer),
method2() (package.subpackage.object.Object method),
388 193
Matplotlib2DViewer (class in fipy.viewers), 414 min() (fipy.variables.Variable method), 361
Matplotlib2DViewer (class in min() (fipy.variables.variable.Variable method), 355
fipy.viewers.matplotlibViewer), 399 MinAll() (fipy.tools.comms.commWrapper.CommWrapper
Matplotlib2DViewer (class in method), 289
fipy.viewers.matplotlibViewer.matplotlib2DViewer),
MinAll() (fipy.tools.comms.dummyComm.DummyComm
389 method), 290
MatplotlibSparseMatrixViewer (class in minmodFaceValue (fipy.variables.CellVariable attribute),
fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer),367
390 minmodFaceValue (fipy.variables.cellVariable.CellVariable
MatplotlibStreamViewer (class in fipy.viewers), 416 attribute), 333
MatplotlibStreamViewer (class in ModularVariable, 125
fipy.viewers.matplotlibViewer), 401 ModularVariable (class in fipy.variables), 369
MatplotlibStreamViewer (class in ModularVariable (class in
fipy.viewers.matplotlibViewer.matplotlibStreamViewer), fipy.variables.modularVariable), 345
391 module
MatplotlibVectorViewer (class in fipy.viewers), 415 fipy.tools.dump, 127
MatplotlibVectorViewer (class in fipy.tools.parser, 124, 154
fipy.viewers.matplotlibViewer), 399 fipy.viewers, 56, 74, 86, 90, 91, 96, 109, 123, 126,
MatplotlibVectorViewer (class in 169
fipy.viewers.matplotlibViewer.matplotlibVectorViewer),scipy, 80
393 Mpi4pyCommWrapper (class in
MatplotlibViewer() (in module fipy.viewers), 411 fipy.tools.comms.mpi4pyCommWrapper),
MatplotlibViewer() (in module 290
fipy.viewers.matplotlibViewer), 395 multiply() (fipy.tools.dimensions.physicalField.PhysicalField
matrix (fipy.terms.term.Term attribute), 264 method), 299
MatrixIllConditionedWarning, 242, 243 multiply() (fipy.tools.PhysicalField method), 323
max() (fipy.variables.Variable method), 361 MultiViewer (class in fipy.viewers), 419
max() (fipy.variables.variable.Variable method), 355 MultiViewer (class in fipy.viewers.multiViewer), 408
MaxAll() (fipy.tools.comms.commWrapper.CommWrapper
method), 289 N
MaxAll() (fipy.tools.comms.dummyComm.DummyComm name (fipy.variables.Variable attribute), 361
440 Index
name (fipy.variables.variable.Variable attribute), 355 P

name() (fipy.tools.dimensions.physicalField.PhysicalUnit package.subpackage (module), 193
method), 304 package.subpackage.base (module), 191
nearest() (in module fipy.tools.numerix), 313 package.subpackage.object (module), 192
NoiseVariable (class in fipy.variables.noiseVariable), 346 ParallelCommWrapper (class in
NonUniformGrid1D (class in fipy.tools.comms.commWrapper), 289
fipy.meshes.nonUniformGrid1D), 217 parallelRandom() (fipy.variables.GaussianNoiseVariable
NonUniformGrid2D (class in method), 376
fipy.meshes.nonUniformGrid2D), 217 parallelRandom() (fipy.variables.gaussianNoiseVariable.GaussianNoiseVaria
NonUniformGrid3D (class in method), 344
fipy.meshes.nonUniformGrid3D), 218 parallelRandom() (fipy.variables.noiseVariable.NoiseVariable
Norm2() (fipy.tools.comms.commWrapper.CommWrapper method), 347
method), 289 parse() (in module fipy.tools.parser), 315
Norm2() (fipy.tools.comms.serialCommWrapper.SerialCommWrapper
PeriodicGrid1D (class in fipy.meshes), 225
method), 290 PeriodicGrid1D (class in fipy.meshes.periodicGrid1D),
Nproc (fipy.tools.comms.commWrapper.CommWrapper 218
attribute), 289 PeriodicGrid2D (class in fipy.meshes), 225
Nproc (fipy.tools.comms.serialCommWrapper.SerialCommWrapper PeriodicGrid2D (class in fipy.meshes.periodicGrid2D),
attribute), 290 219
NthOrderBoundaryCondition, 86 PeriodicGrid2DLeftRight (class in fipy.meshes), 226
NthOrderBoundaryCondition (class in PeriodicGrid2DLeftRight (class in
fipy.boundaryConditions), 197 fipy.meshes.periodicGrid2D), 220
NthOrderBoundaryCondition (class in PeriodicGrid2DTopBottom (class in fipy.meshes), 226
fipy.boundaryConditions.nthOrderBoundaryCondition),
PeriodicGrid2DTopBottom (class in
196 fipy.meshes.periodicGrid2D), 220
numarray, 49 PhysicalField (class in fipy.tools), 316
Numeric, 49 PhysicalField (class in
numericValue (fipy.tools.dimensions.physicalField.PhysicalField fipy.tools.dimensions.physicalField), 293
attribute), 300 physicalShape (fipy.meshes.SkewedGrid2D attribute),
numericValue (fipy.tools.PhysicalField attribute), 323 226
numericValue (fipy.variables.Variable attribute), 361 physicalShape (fipy.meshes.skewedGrid2D.SkewedGrid2D
numericValue (fipy.variables.variable.Variable attribute), attribute), 220
355 physicalShape (fipy.meshes.Tri2D attribute), 227
numerix, 186 physicalShape (fipy.meshes.tri2D.Tri2D attribute), 221
numMeshDeprecated() (in module PhysicalUnit (class in
fipy.meshes.numMesh.deprecatedWarning), fipy.tools.dimensions.physicalField), 302
201 pi, 119, 123, 125
NumPy, 49 PIDStepper (class in fipy.steppers.pidStepper), 247
pip, 49
O plot() (fipy.viewers.DummyViewer method), 421
object plot() (fipy.viewers.matplotlibViewer.matplotlibSparseMatrixViewer.Matplo
fipy.variables.cellVariable.CellVariable, 77 method), 390
fipy.viewers.tsvViewer.TSVViewer, 79 plot() (fipy.viewers.matplotlibViewer.matplotlibViewer.AbstractMatplotlibV
Object (class in package.subpackage.object), 192 method), 395
OffsetSparseMatrix() (in module plot() (fipy.viewers.MayaviClient method), 419
fipy.matrices.offsetSparseMatrix), 199 plot() (fipy.viewers.mayaviViewer.MayaviClient
old (fipy.variables.CellVariable attribute), 368 method), 405
old (fipy.variables.cellVariable.CellVariable attribute), plot() (fipy.viewers.mayaviViewer.mayaviClient.MayaviClient
333 method), 404
openMSHFile() (in module fipy.meshes), 227 plot() (fipy.viewers.MultiViewer method), 419
openMSHFile() (in module fipy.meshes.gmshMesh), 211 plot() (fipy.viewers.multiViewer.MultiViewer method),
openPOSFile() (in module fipy.meshes), 227 408
openPOSFile() (in module fipy.meshes.gmshMesh), 211 plot() (fipy.viewers.TSVViewer method), 420
plot() (fipy.viewers.tsvViewer.TSVViewer method), 409
Index 441
plot() (fipy.viewers.viewer.AbstractViewer method), 410 random() (fipy.variables.betaNoiseVariable.BetaNoiseVariable

plot() (fipy.viewers.vtkViewer.vtkViewer.VTKViewer method), 328
method), 407 random() (fipy.variables.ExponentialNoiseVariable
plotMesh() (fipy.viewers.viewer.AbstractViewer method), method), 373
410 random() (fipy.variables.exponentialNoiseVariable.ExponentialNoiseVariabl
PowerLawConvectionTerm, 115 method), 339
PowerLawConvectionTerm (class in fipy.terms), 278 random() (fipy.variables.GammaNoiseVariable method),
PowerLawConvectionTerm (class in 374
fipy.terms.powerLawConvectionTerm), 261 random() (fipy.variables.gammaNoiseVariable.GammaNoiseVariable
Preconditioner (class in method), 342
fipy.solvers.pysparse.preconditioners.preconditioner),
random() (fipy.variables.noiseVariable.NoiseVariable
234 method), 347
PreconditionerNotPositiveDefiniteWarning, 242, 243 random() (fipy.variables.UniformNoiseVariable method),
PreconditionerWarning, 242, 243 377
PRINT() (in module fipy.tools.debug), 306 random() (fipy.variables.uniformNoiseVariable.UniformNoiseVariable
procID (fipy.tools.comms.commWrapper.CommWrapper method), 350
attribute), 289 rank() (in module fipy.tools.numerix), 310, 311, 314
procID (fipy.tools.comms.serialCommWrapper.SerialCommWrapper ravel() (fipy.tools.dimensions.physicalField.PhysicalField
attribute), 290 method), 300
prune() (in module fipy.tools.vector), 316 ravel() (fipy.tools.PhysicalField method), 324
PseudoRKQSStepper (class in ravel() (fipy.variables.Variable method), 361
fipy.steppers.pseudoRKQSStepper), 247 ravel() (fipy.variables.variable.Variable method), 355
put() (fipy.tools.dimensions.physicalField.PhysicalField read() (in module fipy.tools.dump), 307
method), 300 register_skipper() (in module fipy.tests.doctestPlus), 287
put() (fipy.tools.PhysicalField method), 323 release() (fipy.variables.CellVariable method), 368
put() (fipy.variables.Variable method), 361 release() (fipy.variables.cellVariable.CellVariable
put() (fipy.variables.variable.Variable method), 355 method), 334
put() (in module fipy.tools.numerix), 309, 311, 314 release() (fipy.variables.Variable method), 361
putAdd() (in module fipy.tools.vector), 316 release() (fipy.variables.variable.Variable method), 355
PyAMG, 49 report_skips() (in module fipy.tests.doctestPlus), 287
PyPI, 49 reshape() (fipy.tools.dimensions.physicalField.PhysicalField
Pyrex, 49 method), 300
PySparse, 49 reshape() (fipy.tools.PhysicalField method), 324
PysparseSolver (class in reshape() (fipy.variables.Variable method), 362
fipy.solvers.pysparse.pysparseSolver), 236 reshape() (fipy.variables.variable.Variable method), 355
Python, 49 reshape() (in module fipy.tools.numerix), 310, 311, 314
Python 3, 49 ResidualTerm (class in fipy.terms), 274
Python Enhancement Proposals ResidualTerm (class in fipy.terms.residualTerm), 262
PEP 3000, 49 residualVectorAndNorm() (fipy.terms.term.Term
PYTHONPATH, 12, 15 method), 264
PyTrilinos, 49 RHSvector (fipy.terms.term.Term attribute), 263
PyxViewer, 49 run() (fipy.tools.copy_script.Copy_script method), 306
run() (fipy.tools.performance.efficiency_test.Efficiency_test
Q method), 305
quiver() (fipy.viewers.MatplotlibVectorViewer method), run() (fipy.tools.performance.memoryLogger.MemoryHighWaterThread
416 method), 305
runGold,
quiver() (fipy.viewers.matplotlibViewer.MatplotlibVectorViewer 148
method), 401 runLeveler, 150
runSimpleTrenchSystem, 145
quiver() (fipy.viewers.matplotlibViewer.matplotlibVectorViewer.MatplotlibVectorViewer
method), 394
S
R save() (fipy.tools.Vitals method), 326
random() (fipy.variables.BetaNoiseVariable method), 372 save() (fipy.tools.vitals.Vitals method), 316
ScalarQuantityOutOfRangeWarning, 242, 243
442 Index
scale (fipy.meshes.abstractMesh.AbstractMesh attribute), shape (fipy.meshes.Tri2D attribute), 227

206 shape (fipy.meshes.tri2D.Tri2D attribute), 221
scaledCellDistances (fipy.meshes.abstractMesh.AbstractMesh shape (fipy.tools.dimensions.physicalField.PhysicalField
scaledCellToCellDistances shape (fipy.tools.PhysicalField attribute), 324
(fipy.meshes.abstractMesh.AbstractMesh shape (fipy.variables.Variable attribute), 362
attribute), 206 shape (fipy.variables.variable.Variable attribute), 356
scaledCellVolumes (fipy.meshes.abstractMesh.AbstractMeshsign() (fipy.tools.dimensions.physicalField.PhysicalField
scaledFaceAreas (fipy.meshes.abstractMesh.AbstractMesh sign() (fipy.tools.PhysicalField method), 324
attribute), 206 sign() (fipy.variables.Variable method), 362
scaledFaceToCellDistances sign() (fipy.variables.variable.Variable method), 356
(fipy.meshes.abstractMesh.AbstractMesh sin() (fipy.tools.dimensions.physicalField.PhysicalField
ScharfetterGummelFaceVariable (class in fipy.variables), sin() (fipy.tools.PhysicalField method), 324
369 sin() (fipy.variables.Variable method), 363
ScharfetterGummelFaceVariable (class in sin() (fipy.variables.variable.Variable method), 356
fipy.variables.scharfetterGummelFaceVariable), sinh() (fipy.tools.dimensions.physicalField.PhysicalField
347 method), 301
ScientificPython, 49 sinh() (fipy.tools.PhysicalField method), 324
SciPy, 50, 102, 108 sinh() (fipy.variables.Variable method), 363
scipy sinh() (fipy.variables.variable.Variable method), 356
module, 80 SkewedGrid2D (class in fipy.meshes), 226
scramble() (fipy.variables.noiseVariable.NoiseVariable SkewedGrid2D (class in fipy.meshes.skewedGrid2D),
method), 347 220
SerialCommWrapper (class in SolutionVariableNumberError, 269
fipy.tools.comms.serialCommWrapper), 290 SolutionVariableRequiredError, 269
setLimits() (fipy.viewers.MultiViewer method), 419 solve, 109
setLimits() (fipy.viewers.multiViewer.MultiViewer solve() (fipy.terms.term.Term method), 265
method), 408 Solver (class in fipy.solvers), 243
setLimits() (fipy.viewers.viewer.AbstractViewer method), Solver (class in fipy.solvers.solver), 242
410 SolverConvergenceWarning, 242, 243
setName() (fipy.tools.dimensions.physicalField.PhysicalUnitSourceTerm (class in fipy.terms.sourceTerm), 263
method), 304 Sphinx, 50
setName() (fipy.variables.Variable method), 362 sqrt, 97, 124, 154
setName() (fipy.variables.variable.Variable method), 355 arcsin
setScale() (fipy.meshes.abstractMesh.AbstractMesh cos, 80
method), 206 sqrt() (fipy.tools.dimensions.physicalField.PhysicalField
setUnit() (fipy.tools.dimensions.physicalField.PhysicalField method), 301
method), 300 sqrt() (fipy.tools.PhysicalField method), 324
setUnit() (fipy.tools.PhysicalField method), 324 sqrt() (fipy.variables.Variable method), 363
setUnit() (fipy.variables.Variable method), 362 sqrt() (fipy.variables.variable.Variable method), 356
setUnit() (fipy.variables.variable.Variable method), 355 sqrtDot() (in module fipy.tools.numerix), 314
setValue() (fipy.variables.CellVariable method), 368 SsorPreconditioner (class in fipy.solvers), 245
setValue() (fipy.variables.cellVariable.CellVariable SsorPreconditioner (class in fipy.solvers.pysparse), 238
method), 334 SsorPreconditioner (class in
setValue() (fipy.variables.FaceVariable method), 369 fipy.solvers.pysparse.preconditioners), 234
setValue() (fipy.variables.faceVariable.FaceVariable SsorPreconditioner (class in
method), 340 fipy.solvers.pysparse.preconditioners.ssorPreconditioner),
setValue() (fipy.variables.Variable method), 362 234
setValue() (fipy.variables.variable.Variable method), 355 StagnatedSolverWarning, 242, 243
shape (fipy.meshes.SkewedGrid2D attribute), 226 start() (fipy.tools.performance.memoryLogger.MemoryLogger
shape (fipy.meshes.skewedGrid2D.SkewedGrid2D method), 305
attribute), 220 SteadyConvectionDiffusionScEquation, 185
Index 443
step() (fipy.steppers.stepper.Stepper method), 247 tan() (fipy.tools.PhysicalField method), 325

Stepper (class in fipy.steppers.stepper), 247 tan() (fipy.variables.Variable method), 363
stop() (fipy.tools.performance.memoryLogger.MemoryHighWaterThread
tan() (fipy.variables.variable.Variable method), 356
method), 305 tanh, 97
stop() (fipy.tools.performance.memoryLogger.MemoryLogger tanh() (fipy.tools.dimensions.physicalField.PhysicalField
subscribedVariables (fipy.variables.Variable attribute), tanh() (fipy.tools.PhysicalField method), 325
363 tanh() (fipy.variables.Variable method), 363
subscribedVariables (fipy.variables.variable.Variable at- tanh() (fipy.variables.variable.Variable method), 356
tribute), 356 Term (class in fipy.terms.term), 263
subtract() (fipy.tools.dimensions.physicalField.PhysicalFieldTermMultiplyError, 269
method), 301 testmod() (in module fipy.tests.doctestPlus), 287
subtract() (fipy.tools.PhysicalField method), 325 tostring() (fipy.tools.dimensions.physicalField.PhysicalField
successFn() (fipy.steppers.stepper.Stepper static method), method), 302
248 tostring() (fipy.tools.PhysicalField method), 326
sum() (fipy.tools.comms.commWrapper.CommWrapper tostring() (fipy.variables.Variable method), 363
method), 289 tostring() (fipy.variables.variable.Variable method), 356
sum() (fipy.tools.comms.dummyComm.DummyComm tostring() (in module fipy.tools.numerix), 314
method), 290 TransientTerm, 56, 98, 123, 125
sum() (fipy.tools.dimensions.physicalField.PhysicalField TransientTerm (class in fipy.terms), 272
method), 301 TransientTerm (class in fipy.terms.transientTerm), 265
sum() (fipy.tools.PhysicalField method), 325 Tri2D (class in fipy.meshes), 226
sum() (fipy.variables.Variable method), 363 Tri2D (class in fipy.meshes.tri2D), 221
sum() (fipy.variables.variable.Variable method), 356 Trilinos, 50
sum() (in module fipy.tools.numerix), 310, 311, 314 TSVViewer (class in fipy.viewers), 419
SurfactantConvectionVariable (class in fipy.variables), TSVViewer (class in fipy.viewers.tsvViewer), 408
378 tupleToXML() (fipy.tools.Vitals method), 326
SurfactantConvectionVariable (class in tupleToXML() (fipy.tools.vitals.Vitals method), 316
fipy.variables.surfactantConvectionVariable),
347 U
SurfactantVariable, 155 UniformGrid (class in fipy.meshes.uniformGrid), 221
SurfactantVariable (class in fipy.variables), 377 UniformGrid1D (class in fipy.meshes.uniformGrid1D),
SurfactantVariable (class in 221
fipy.variables.surfactantVariable), 348 UniformGrid2D (class in fipy.meshes.uniformGrid2D),
svn() (fipy.tools.Vitals method), 326 222
svn() (fipy.tools.vitals.Vitals method), 316 UniformGrid3D (class in fipy.meshes.uniformGrid3D),
svncmd() (fipy.tools.Vitals method), 326 222
svncmd() (fipy.tools.vitals.Vitals method), 316 UniformNoiseVariable (class in fipy.variables), 376
sweep, 102, 109, 169 UniformNoiseVariable (class in
sweep() (fipy.terms.term.Term method), 265 fipy.variables.uniformNoiseVariable), 349
sweepFn() (fipy.steppers.stepper.Stepper static method), unit (fipy.tools.dimensions.physicalField.PhysicalField
248 attribute), 302
sweepMonotonic() (in module fipy.steppers), 248 unit (fipy.tools.PhysicalField attribute), 326
unit (fipy.variables.Variable attribute), 363
T unit (fipy.variables.variable.Variable attribute), 356
take() (fipy.tools.dimensions.physicalField.PhysicalField updateOld() (fipy.variables.CellVariable method), 368
method), 302 updateOld() (fipy.variables.cellVariable.CellVariable
take() (fipy.tools.PhysicalField method), 325 method), 334
take() (fipy.variables.Variable method), 363 updateOld() (fipy.variables.ModularVariable method),
take() (fipy.variables.variable.Variable method), 356 370
take() (in module fipy.tools.numerix), 310, 311, 314 updateOld() (fipy.variables.modularVariable.ModularVariable
tan, 119 method), 346
tan() (fipy.tools.dimensions.physicalField.PhysicalField UpwindConvectionTerm (class in fipy.terms), 279
method), 302
444 Index
UpwindConvectionTerm (class in
fipy.terms.upwindConvectionTerm), 266
user_options (fipy.tools.copy_script.Copy_script at-
tribute), 306
user_options (fipy.tools.performance.efficiency_test.Efficiency_test
attribute), 305
V
value (fipy.variables.Variable attribute), 363
value (fipy.variables.variable.Variable attribute), 357
VanLeerConvectionTerm (class in fipy.terms), 280
VanLeerConvectionTerm (class in
fipy.terms.vanLeerConvectionTerm), 268
Variable, 101, 104
Variable (class in fipy.variables), 357
Variable (class in fipy.variables.variable), 350
VectorCoeffError, 269
vertexCoords (fipy.meshes.uniformGrid1D.UniformGrid1D
attribute), 222
attribute), 222
attribute), 223
Viewer() (in module fipy.viewers), 421
Vitals (class in fipy.tools), 326
Vitals (class in fipy.tools.vitals), 316
VTKCellDataSet (fipy.meshes.abstractMesh.AbstractMesh
attribute), 203
VTKCellViewer (class in fipy.viewers), 421
VTKCellViewer (class in fipy.viewers.vtkViewer), 407
VTKCellViewer (class in
fipy.viewers.vtkViewer.vtkCellViewer), 406
VTKFaceDataSet (fipy.meshes.abstractMesh.AbstractMesh
attribute), 203
VTKFaceViewer (class in fipy.viewers), 421
VTKFaceViewer (class in fipy.viewers.vtkViewer), 407
VTKFaceViewer (class in
fipy.viewers.vtkViewer.vtkFaceViewer), 406
VTKViewer (class in fipy.viewers.vtkViewer.vtkViewer),
406
VTKViewer() (in module fipy.viewers), 420
VTKViewer() (in module fipy.viewers.vtkViewer), 407
W
write() (in module fipy.tools.dump), 307
X
x (fipy.meshes.abstractMesh.AbstractMesh attribute), 206
Y
y (fipy.meshes.abstractMesh.AbstractMesh attribute), 206
Z
z (fipy.meshes.abstractMesh.AbstractMesh attribute), 206
Index 445

Fipy-3 1 3

Uploaded by

Copyright:

Available Formats

Fipy-3 1 3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fipy-3 1 3

Uploaded by

Copyright:

Available Formats

𝜙

Materials Science and Engineering Division

January 17, 2017

6 Theoretical and Numerical Background 29

7 Design and Implementation 37

8 Frequently Asked Questions 41

12 Phase Field Examples 95

13 Level Set Examples 139

14 Cahn Hilliard Examples 161

15 Fluid Flow Examples 167

16 Reactive Wetting Examples 173

17 Updating FiPy 179

III fipy Package Documentation 189

19 fipy.boundaryConditions package 195

20 fipy.matrices package 199

21 fipy.meshes package 201

22 fipy.solvers package 233

23 fipy.steppers package 247

24 fipy.terms package 251

25 fipy.tests package 287

26 fipy.tools package 289

27 fipy.variables package 327

28 fipy.viewers package 385

Python Module Index 425

1.1 Even if you don’t read manuals...

1.2 What’s new in version 3.1.3?

The significant changes since version 3.0 are:

143 "Trying to ""solve"" an integer `CellVariable` should raise an error"

1.3 Download and Installation

1.4.1 Mailing List

1.3. Download and Installation 5

1.5 Conventions and Notation

will correspond to, e.g.:

Note: to indicate something that may be of interest

Warning: to indicate something that could cause serious problems.

2.1 Recommended Method

$ conda create --name MYFIPYENV --channel guyer --channel conda-forge fipy

• enable this new environment with:

$ source activate MYFIPYENV

• move on to Using FiPy.

2.2 Installing Python Packages

• Download the packages manually, unpack and run:

Further information about each setup.py script is available by typing:

2.3 Obtaining FiPy

2.4 Installing FiPy

or by unpacking the archive and typing:

$ python setup.py install

2.5 Required Packages

Enthought Python Distribution

2.5. Required Packages 9

2.6 Optional Packages

2.7 Level Set Packages

2.8 Platform-Specific Instructions

2.8.1 Mac OS X Installation

2.8.2 Windows Installation

2.8.3 Ubuntu/Debian Installation

2.8. Platform-Specific Instructions 11

$ VERSION=x.y-z # choose the version you want

2.8.4 Miscellaneous Build Recipes