Lumped Parameter Model Analysis

Lumped Parameter
Model Analysis
CAPE Centre, The University of Queensland

Hungarian Academy of Sciences
PROCESS MODELLING
AND MODEL ANALYSIS
C5B
1
Key Issues to Consider
Must analyze before solving
Degrees of freedom
Index of the differential-algebraic system
Stability of the model

PROCESS MODELLING
AND MODEL ANALYSIS
Degrees of Freedom
Definition:
N DF NU N E
- NU = number of unknowns (variables)
- NE = number of independent equations

PROCESS MODELLING
AND MODEL ANALYSIS
Example DoF analysis

P
z
P
P
F
Model
dz ( F1 F2 )
dt
A
F1 CV P1 P2 0
F2 CV P2 P3 0
P2 P0 gz 0
DOF analysis
State variable z
Algebraic variables F1 , F2 , P0 , P1 , P 2 , P3
Parameters C V , A,
Constants g
Equations 4
N DF NU N E 7 4 3
PROCESS MODELLING
AND MODEL ANALYSIS
Different specifications (DoF selection)

P
Degrees of freedom = 3
z
P
Specification 1
S1 P0 , P1 , P3
g1
g2
g3
F1
x
F2
x
P2
x
x
x

Specification 2
S 2 P0 , P1 , P2
g1
g2
g3
F1
x
F2
P3
PROCESS MODELLING
AND MODEL ANALYSIS
High Index DAEs

Only
for Differential-algebraic equations

Pure ODE systems are index 0
Many index 1 DAE solvers
Few higher index solvers
Ensure index 1 models are used

PROCESS MODELLING
AND MODEL ANALYSIS
DAE Index Definition

The
index of the DAE system is the

number of times the algebraic sub-system
must be differentiated to give a set of
ODEs.

PROCESS MODELLING
AND MODEL ANALYSIS
General DAE system

General system
dy
f ( y, z , t )
dt
0 g ( y, z , t )
Differentiate the algebraic equations

dy
dz
gz
dt
dt
dz
0 gy f gz
dt
dz
1
gz g y f
iff
dt
0 gy

g z full rank
PROCESS MODELLING
AND MODEL ANALYSIS
High index example

Case 1
y 1 y1 y2 z1
y 2 y1 y2 z1
0 y1 2 y2 z1
Differentiate once
z1 y 1 2 y 2
z1 3 y1 y2 z1
index 1 model !
Case 2
y 1 y1 y2 z1
y 2 y1 y2 z1
0 y1 2 y2
Differentiate once
0 y 1 2 y 2
0 3 y1 y2 z1
and again ...
0 3 y 1 y 2 z1
z1 2 y1 4 y2 4 z1
index 2 model !

PROCESS MODELLING
AND MODEL ANALYSIS
Stability of the Model

Indicated
by the dynamic modes

- time constants of the system
- related to physico-chemical phenomena

fluid flow (generally fast)
mass transfer rates (can be slow)
heat transfer (usually slow)
reaction kinetics (very fast to very slow)

PROCESS MODELLING
AND MODEL ANALYSIS
10
Linear model analysis

Linear equation system
y Ay
y1 2000 999.75
1
1
y2
y1 1000.25
0
y2
Solution given by
y1 (t ) 1.449e 0.5t 0.499e 2005t 1

y2 (t ) 2.999e 0.5t 0.0025e 2005t 1
Slow mode (component)
Fast mode (component)
PROCESS MODELLING
AND MODEL ANALYSIS
11
General Linear System Analysis

dx
Ax,
x(0) x 0
dt
x(t ) Z exp(t )
n
xi (t ) Z ij e
where :
Z ij Vij V 1
k 1
Solution to the equations

Individual component
solution to the equations
note the eigenvalues !
note the eigenvectors !
jt
j 1
Linear set of equations
0
x
jk k

The eigenvalues
contain the stability
information
PROCESS MODELLING
AND MODEL ANALYSIS
12
Computing Eigenvalues in Matlab

Example of MATLAB use
A=[-2000 999.75;1 -1];

b=eig(A)
MATLAB function:
eig(A)
A= state matrix
b=
1.0e+003 *
Two eigenvalues
(-2000.5, -0.0005)
-2.0005
-0.0005
PROCESS MODELLING
AND MODEL ANALYSIS
13
Eigenvalue Analysis
Dynamic modes determined by eigenvalues of

the linear system
Nonlinear problems must be linearised first to

get Jacobian J(y,t) on the trajectory
Eigenvalues of Jacobian computed using

MATLAB eig(A) or similar function.

PROCESS MODELLING
AND MODEL ANALYSIS
14
Behaviour of Linear Systems
Phase plane analysis illustrates motion of system
Second order systems plot
Typical second order system (cf. valve actuator) :
y1 vs y2
y ay by 0
with
a
a
1

a
a
2
2
or
y vs y
4b / 2
4b / 2
PROCESS MODELLING
AND MODEL ANALYSIS
15
Phase Plane Analysis

Several
cases dependent on eigenvalues
1 and 2 both real and same sign

1 and 2 both real but opposite sign
1 and 2 complex conjugates with non - zero real parts
1 and 2 complex conjugates with zero real parts

PROCESS MODELLING
AND MODEL ANALYSIS
16
Phase Plane Diagrams

PROCESS MODELLING
AND MODEL ANALYSIS
17
General Nonlinear Model

Linearize
at some point
y f ( y, t )
Obtain
linear model
Solution
is
y A( y g (t )) f (t , g (t ))
y (t ) c e v g (t )
i t
i 1

PROCESS MODELLING
AND MODEL ANALYSIS
18
Stability Cases
Stable model
0 for all i
i
Unstable model
Some will be 0
i
Ultra-stable (stiff)
0
i
Some " small"

i
Some " large"

i

PROCESS MODELLING
AND MODEL ANALYSIS
19
Stability implications for

numerical solution (Chapter 6)
The problem determines the choice of numerical

method
Stiff problems need special attention
All simple explicit methods have limitations on
allowable step lengths
method stability bound

h
max
i max

PROCESS MODELLING
AND MODEL ANALYSIS
20

Lumped Parameter Model Analysis

Uploaded by

Copyright:

Available Formats

Lumped Parameter Model Analysis

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

Lumped Parameter Model Analysis

Uploaded by

Copyright:

Available Formats

Lumped Parameter

CAPE Centre, The University of Queensland

Key Issues to Consider

Must analyze before solving

Index of the differential-algebraic system

Stability of the model

CAPE Centre, The University of Queensland

CAPE Centre, The University of Queensland

Example DoF analysis

Different specifications (DoF selection)

CAPE Centre, The University of Queensland

High Index DAEs

for Differential-algebraic equations

CAPE Centre, The University of Queensland

DAE Index Definition

index of the DAE system is the

CAPE Centre, The University of Queensland

General DAE system

Differentiate the algebraic equations

CAPE Centre, The University of Queensland

High index example

CAPE Centre, The University of Queensland

Stability of the Model

by the dynamic modes

- related to physico-chemical phenomena

CAPE Centre, The University of Queensland

Linear model analysis

y1 (t ) 1.449e 0.5t 0.499e 2005t 1

Fast mode (component)

General Linear System Analysis

Solution to the equations

Linear set of equations

CAPE Centre, The University of Queensland

Computing Eigenvalues in Matlab

A=[-2000 999.75;1 -1];

Dynamic modes determined by eigenvalues of

Nonlinear problems must be linearised first to

Eigenvalues of Jacobian computed using

CAPE Centre, The University of Queensland

Behaviour of Linear Systems

Phase plane analysis illustrates motion of system

Second order systems plot

Typical second order system (cf. valve actuator) :

CAPE Centre, The University of Queensland

Phase Plane Analysis

cases dependent on eigenvalues

1 and 2 both real and same sign

CAPE Centre, The University of Queensland

Phase Plane Diagrams

CAPE Centre, The University of Queensland

General Nonlinear Model

CAPE Centre, The University of Queensland

Some " small"

Some " large"

CAPE Centre, The University of Queensland

Stability implications for

The problem determines the choice of numerical

method stability bound

CAPE Centre, The University of Queensland