Die Casting
Die Casting
Die Casting
Genick BarMeir, Ph. D. 7449 North Washtenaw Ave Chicago, IL 60645 email:barmeir at gmail dot com
Copyright 2009, 2008, 2007, and 1999 by Genick Bar-Meir See the le copying.fdl or copyright.tex for copying conditions. Version (0.1.3 November 9, 2012)
CONTENTS
Nomenclature GNU Free Documentation License . . . . . . . . . . . . . . . . 1. APPLICABILITY AND DEFINITIONS . . . . . . . . . 2. VERBATIM COPYING . . . . . . . . . . . . . . . . . . 3. COPYING IN QUANTITY . . . . . . . . . . . . . . . . 4. MODIFICATIONS . . . . . . . . . . . . . . . . . . . . 5. COMBINING DOCUMENTS . . . . . . . . . . . . . . 6. COLLECTIONS OF DOCUMENTS . . . . . . . . . . . 7. AGGREGATION WITH INDEPENDENT WORKS . . . 8. TRANSLATION . . . . . . . . . . . . . . . . . . . . . 9. TERMINATION . . . . . . . . . . . . . . . . . . . . . 10. FUTURE REVISIONS OF THIS LICENSE . . . . . . . ADDENDUM: How to use this License for your documents CONTRIBUTORS LIST How to contribute to this book . . . . . . . . . . . . . Credits . . . . . . . . . . . . . . . . . . . . . . . . . . Steven from artofproblemsolving.com . . . . . . . Tousher Yang . . . . . . . . . . . . . . . . . . . . Steve Spurgeon . . . . . . . . . . . . . . . . . . . Irene Tan . . . . . . . . . . . . . . . . . . . . . . Your name here . . . . . . . . . . . . . . . . . . . Typo corrections and other minor contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
xv xix xx xxi xxi xxii xxiv xxiv xxv xxv xxv xxv xxvi xxvii xxvii xxvii xxvii xxviii xxviii xxviii xxviii xxviii
Prologue For The POTTO Project xxxi Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxi Why Volunteer? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxii
iii
iv
Prologue For This Book Version 0.1 January 12, 2009 . pages 213 size 1.5M . . . Version 0.0.3 October 9, 1999 pages 178 size 3.2M . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 Introduction 1.1 The Importance of Reducing Production Costs . . . 1.2 Designed/Undesigned Scrap/Cost . . . . . . . . . . 1.3 Linking the Production Cost to the Product Design 1.4 Historical Background . . . . . . . . . . . . . . . . 1.5 Numerical Simulations . . . . . . . . . . . . . . . . 1.6 Integral Models . . . . . . . . . . . . . . . . . . 1.7 Summary . . . . . . . . . . . . . . . . . . . . . . . 2 Basic Fluid Mechanics 2.1 Introduction . . . . . . . . . . . . . . . . 2.2 What is uid? Shear stress . . . . . . . . 2.2.1 What is Fluid? . . . . . . . . . . . 2.2.2 What is Shear Stress? . . . . . . . 2.3 Thermodynamics and mechanics concepts 2.3.1 Thermodynamics . . . . . . . . . 2.3.2 Basic Denitions . . . . . . . . . 2.3.3 Momentum Equation . . . . . . . 2.3.4 Compressible ow . . . . . . . . . 2.3.5 Speed of Sound . . . . . . . . . . 2.3.6 Choked Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 Dimensional Analysis 3.0.7 How The Dimensional Analysis Work . . . . . . . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 The Die Casting Process Stages . . . . . . . . . . . . . . 3.2.1 Filling the Shot Sleeve . . . . . . . . . . . . . . . 3.2.2 Plunger Slow Moving Part . . . . . . . . . . . . . 3.2.3 Runner system . . . . . . . . . . . . . . . . . . . . 3.2.4 Die Cavity . . . . . . . . . . . . . . . . . . . . . . 3.2.5 Intensication Period . . . . . . . . . . . . . . . . 3.3 Special Topics . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Is the Flow in Die Casting Turbulent? . . . . . . . 3.3.2 Dissipation eect on the temperature rise . . . . . 3.3.3 Gravity eects . . . . . . . . . . . . . . . . . . . . 3.4 Estimates of the time scales in die casting . . . . . . . . . 3.4.1 Utilizing semi dimensional analysis for characteristic
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . time
CONTENTS 3.5 3.4.2 The ratios of various time scales Similarity applied to Die cavity . . . . . 3.5.1 Governing equations . . . . . . . 3.5.2 Design of Experiments . . . . . Summary of dimensionless numbers . . . Summary . . . . . . . . . . . . . . . . . Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v 54 55 56 59 59 61 61 63 63 63 64 66 66 66 67 67 67 67 68 69 69 71 71 73 73 73 74 75
4 Fundamentals of Pipe Flow 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . 4.2 Universality of the loss coecients . . . . . . . . . . 4.3 A simple ow in a straight conduit . . . . . . . . . . 4.3.1 Examples of the calculations . . . . . . . . . 4.4 Typical Components in the Runner andVent Systems 4.4.1 bend . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Y connection . . . . . . . . . . . . . . . . . 4.4.3 Expansion/Contraction . . . . . . . . . . . . 4.5 Putting it all to Together . . . . . . . . . . . . . . . 4.5.1 Series Connection . . . . . . . . . . . . . . . 4.5.2 Parallel Connection . . . . . . . . . . . . . .
5 Flow in Open Channels 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Typical diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Hydraulic Jump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Runner Design 6.1 Introduction . . . . . . . 6.1.1 Backward Design 6.1.2 Connecting runner 6.1.3 Resistance . . . . 7 pQ2 7.1 7.2 7.3 . . . . . . . . . . . . segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.4
7.5
Diagram Calculations Introduction . . . . . . . . . . . . . . . The common pQ2 diagram . . . . . The validity of the common diagram . 7.3.1 Is the Common Model Valid? 7.3.2 Are the Trends Reasonable? . . 7.3.3 Variations of the Gate area, A3 . The reformed pQ2 diagram . . . . . . 7.4.1 The reform model . . . . . . . . 7.4.2 Examining the solution . . . . . 7.4.3 Poor design eects . . . . . . . 7.4.4 Transient eects . . . . . . . . . Design Process . . . . . . . . . . . . . .
CONTENTS The Intensication Consideration . . . . . . . . . . . . . . . . . . . . . 104 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 107 107 108 110 111 111 112 112 112 112 113 113 113 115 116 116 117 117 118 118 118 119 121 123 126 126 127 131 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 133 133 134 135 135 136 137
8 Critical Slow Plunger Velocity 8.1 Introduction . . . . . . . . . . . . . . . 8.2 The common models . . . . . . . . . 8.2.1 Garbers model . . . . . . . . . 8.2.2 Brevicks Model . . . . . . . . . 8.2.3 Brevicks circular model . . . . . 8.2.4 Millers square model . . . . . . 8.3 The validity of the common models . 8.3.1 Garbers model . . . . . . . . . 8.3.2 Brevicks models . . . . . . . . . 8.3.3 Millers model . . . . . . . . . . 8.3.4 EKKs model (numerical model) 8.4 The Reformed Model . . . . . . . . . . 8.4.1 The reformed model . . . . . . . 8.4.2 Design process . . . . . . . . . . 8.5 Summary . . . . . . . . . . . . . . . . . 8.6 Questions . . . . . . . . . . . . . . . . 9 Venting System Design 9.1 Introduction . . . . . . . 9.2 The common models . 9.2.1 Early (etc.) model 9.2.2 Millers model . . 9.3 General Discussion . . . . 9.4 The Analysis . . . . . . . 9.5 Results and Discussion . 9.6 Summary . . . . . . . . . 9.7 Questions . . . . . . . . 10 Density change eects 11 Clamping Force Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12 Analysis of Die Casting Economy 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . 12.2 The common model, Millers approach . . . . . . 12.3 The validity of Millers price model . . . . . . . . . 12.4 The combined Cost of the Controlled Components 12.5 Die Casting Machine Capital Costs . . . . . . . . . 12.6 Operational Cost of the Die Casting Machine . . . 12.7 Runner Cost (Scrap Cost) . . . . . . . . . . . . . .
CONTENTS 12.8 Startup and Mold Manufacturing Cost 12.9 Personnel Cost . . . . . . . . . . . . . . 12.10Uncontrolled components . . . . . . . . 12.11Minimizing Cost of Single Operation . . 12.12Introduction to Economics . . . . . . . 12.13Summary . . . . . . . . . . . . . . . . . 12.14Question . . . . . . . . . . . . . . . . . A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vii 139 140 140 141 144 145 145 147 147 148 149 152 153 156 161 162 162 163 167 171 177 178 180 182 183 184 186 189 190 191 192 195 201 217
Fanno Flow A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.2 Fanno Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3 NonDimensionalization of the Equations . . . . . . . . . . . . . A.4 The Mechanics and Why the Flow is Choked? . . . . . . . . . . . A.5 The Working Equations . . . . . . . . . . . . . . . . . . . . . . . A.6 Examples of Fanno Flow . . . . . . . . . . . . . . . . . . . . . . A.7 Supersonic Branch . . . . . . . . . . . . . . . . . . . . . . . . . . A.8 Maximum Length for the Supersonic Flow . . . . . . . . . . . . . A.9 Working Conditions . . . . . . . . . . . . . . . . . . . . . . . . . A.9.1 Variations of The Tube Length ( 4f L ) Eects . . . . . . . D A.9.2 The Pressure Ratio, P2 , eects . . . . . . . . . . . . . . . P1 A.9.3 Entrance Mach number, M1 , eects . . . . . . . . . . . . A.10 Practical Examples for Subsonic Flow . . . . . . . . . . . . . . . A.10.1 Subsonic Fanno Flow for Given 4f L and Pressure Ratio . D A.10.2 Subsonic Fanno Flow for a Given M1 and Pressure Ratio . A.11 The Approximation of the Fanno Flow by Isothermal Flow . . . . A.12 More Examples of Fanno Flow . . . . . . . . . . . . . . . . . . . A.13 The Table for Fanno Flow . . . . . . . . . . . . . . . . . . . . . A.14 Appendix Reynolds Number Eects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
B What The Establishments Scientists Say B.1 Summary of Referee positions . . . . . . B.2 Referee 1 (from hand written notes) . . B.3 Referee 2 . . . . . . . . . . . . . . . . . B.4 Referee 3 . . . . . . . . . . . . . . . . .
viii
CONTENTS
LIST OF FIGURES
1.1 1.2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15
The prots as a function of the amount of the scrap . . . . . . . . . . Increase of prots as reduction of scrap reduction. . . . . . . . . . . . . The velocity distribution in Couette ow . . . . . . . . . . . . . . . The deformation of uid due to shear stress as progression of time. A very slow moving piston in a still gas. . . . . . . . . . . . . . . . Stationary sound wave and gas moves relative to the pulse. . . . . . Gas ow through a convergingdiverging nozzle. . . . . . . . . . . . The stagnation properties as a function of the Mach number, k=1.4 Various ratios as a function of Mach number for isothermal Nozzle . Rod into the hole example . . . . . . . . . . . . . . . . . . . . . Hydraulic jump in the shot sleeve. . . . . . . . . . . . . . . . . . Filling of the shot sleeve. . . . . . . . . . . . . . . . . . . . . . . Heat transfer processes in the shot sleeve. . . . . . . . . . . . . . Solidication of the shot sleeve time estimates. . . . . . . . . . . Entrance of liquid metal to the runner. . . . . . . . . . . . . . . . Flow in runner when during pressurizing process. . . . . . . . . . . Typical ow pattern in die casting, jet entering into empty cavity. Transition to turbulent ow in instantaneous ow after Wygnanski Flow pattern in the shot sleeve . . . . . . . . . . . . . . . . . . . Two streams of uids into a medium. . . . . . . . . . . . . . . . Schematic of heat transfer processes in the die. . . . . . . . . . . The oscillating manometer for the example 3.1. . . . . . . . . . . Mass Balance on the lest side of the manometer . . . . . . . . . . Rigid body brought into rest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 3 11 13 21 21 23 24 28 33 34 35 36 37 39 39 40 42 42 44 47 50 51 54
ix
x 4.1 4.2 4.3 4.4 4.5 5.1 5.2 6.1 6.2 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13 7.14 7.15 8.1 8.2 8.3 8.4 9.1 9.2 9.3 The results for the ow in a pipe with orice. General simple conduit description. . . . . . . General simple conduit description. . . . . . . A sketch of the bend in die casting. . . . . . A parallel connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 64 65 66 68 69 70 74 75
Equilibrium of Forces in an open channel. . . . . . . . . . . . . . . . . Specic Energy and momentum Curves. . . . . . . . . . . . . . . . . . A geometry of runner connection. . . . . . . . . . . . . . . . . . . . . y connection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic of typical die casting machine. . . . . . . . . . . . A typical trace on a cold chamber machine . . . . . . . . . . pQ2 diagram typical characteristics. . . . . . . . . . . . . . . P as A3 to be relocated . . . . . . . . . . . . . . . . . . . . . Presure of die casting machine. . . . . . . . . . . . . . . . . . P1 as a function of Pmax . . . . . . . . . . . . . . . . . . . . KF as a function of gate area, A3 . . . . . . . . . . . . . . . Die casting characteristics. . . . . . . . . . . . . . . . . . . . Various die casting machine performances . . . . . . . . . . . Reduced pressure performances as a function of Ozer number. Schematic of the plunger and piston balance forces . . . . . . Metal pressure at the plunger tip. . . . . . . . . . . . . . . . . Hydralic piston schematic . . . . . . . . . . . . . . . . . . . . The gate velocity, U3 as a function of the plunger area, A1 . . The reduced power as a function of the normalized ow rate. . A schematic of wave formation in stationary coordinates The two kinds in the sleeve. . . . . . . . . . . . . . . . . A schematic of the wave with moving coordinates . . . . The Froude number as a function of the relative height. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The relative shrinkage porosity as a function of the casting thickness. . 117 A simplied model for the venting system. . . . . . . . . . . . . . . . . 122 The pressure ratios for air and vacuum venting at end. . . . . . . . . . 124
10.1 The control volume of the phase change. . . . . . . . . . . . . . . . . . 127 12.1 Production cost as a function of the runner hydraulic diameter. . . . . . 134 12.2 The reduced power as a function of the normalized ow rate. . . . . . . 137 12.3 Supply and Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 A.1 Control volume of the gas ow in a constant cross section . . . . . . . 147 A.2 Various parameters in Fanno ow as a function of Mach number . . . . 156 A.3 Schematic of Example (A.1) . . . . . . . . . . . . . . . . . . . . . . . 156
LIST OF FIGURES A.4 A.5 A.6 A.7 A.8 A.9 A.10 A.11 A.12 A.13 A.14 A.15 A.16 A.17 A.18 A.19 A.20 A.21 A.22 The schematic of Example (A.2) . . . . . . . . . . . . . . . . The maximum length as a function of specic heat, k . . . . . The eects of increase of 4f L on the Fanno line . . . . . . . D The development properties in of converging nozzle . . . . . . Min and m as a function of the 4f L . . . . . . . . . . . . . . D M1 as a function M2 for various 4f L . . . . . . . . . . . . . D M1 as a function M2 . . . . . . . . . . . . . . . . . . . . . . The pressure distribution as a function of 4f L for a short 4f L D D The pressure distribution as a function of 4f L for a long 4f L D D The eects of pressure variations on Mach number prole . . Mach number as a function of 4f L when the total 4f L = 0.3 D D Schematic of a long tube in supersonic branch . . . . . . . The extra tube length as a function of the shock location . . The maximum entrance Mach number as a function of 4f L . D Unchoked ow showing the hypothetical full tube . . . . . The results of the algorithm showing the conversion rate. . . Solution to a missing diameter . . . . . . . . . . . . . . . . . M1 as a function of 4f L comparison with Isothermal Flow . . D Moody diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xi 158 163 164 165 165 167 168 169 170 171 172 173 174 175 178 179 182 183 186
xii
LIST OF FIGURES
LIST OF TABLES
1 1 2.1
Books Under Potto Project . . . . . . . . . . . . . . . . . . . . . . . . xxxiii continue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxiv Properties of Various Ideal Gases [300K] . . . . . . . . . . . . . . . . . 19
A.1 Fanno Flow Standard basic Table . . . . . . . . . . . . . . . . . . . . 184 A.1 continue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 A.1 continue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
xiii
xiv
LIST OF TABLES
NOMENCLATURE
Universal gas constant, see equation (2.36), page 18 Units length., see equation (2.11), page 14
B Bf c Cp Cv EU Eu Ei H h k M n P
Density of the uid, see equation (2.55), page 22 bulk modulus, see equation (2.62), page 22 Body force, see equation (2.19), page 15 Speed of sound, see equation (2.55), page 22 Specic pressure heat, see equation (2.33), page 17 Specic volume heat, see equation (2.32), page 17 Internal energy, see equation (2.13), page 15 Internal Energy per unit mass, see equation (2.16), page 15 System energy at state i, see equation (2.12), page 14 Enthalpy, see equation (2.28), page 17 Specic enthalpy, see equation (2.28), page 17 the ratio of the specic heats, see equation (2.34), page 18 Mach number, see equation (2.64), page 23 The poletropic coecient, see equation (2.60), page 22 Pressure, see equation (2.57), page 22
xv
xvi q Q12 R S U w W12 Energy per unit mass, see equation (2.16), page 15
LIST OF TABLES
The energy transferred to the system between state 1 and state 2, see equation (2.12), page 14 Specic gas constant, see equation (2.37), page 18 Entropy of the system, see equation (2.23), page 16 velocity , see equation (2.14), page 15 Work per unit mass, see equation (2.16), page 15 The work done by the system between state 1 and state 2, see equation (2.12), page 14
Version 0.1.3
Nov 8, 2012 (1.9M 265 pages)
Improvements to some of the gures of dimensional analysis chapter (utilizing blender). Add an analysis of the minimum cost ordering supply. The minimum cost ordering refers to the analysis dealing with the minimum cost achieved by nding the optimum number of ordering.
Version 0.1.2
April 1, 2009 (1.9M 263 pages)
Irene Tan provided many English corrections to the dimensional analysis chapter.
Version 0.1.1
Feb 8, 2009 (1.9M 261 pages)
Add Steve Spurgeon (from Dynacast England) corrections to pQ2 diagram. Minor English corrections to pQ2 diagram chapter (unnished). Fix some gures and captions issues. Move to potto style le.
xvii
xviii
LIST OF TABLES
Version 0.1
Jan 6, 2009 (1.6M 213 pages)
Change to modern Potto format. English corrections Finish some examples in Dimensionless Chapter (manometer etc)
Version 0.0.3
Nov 1, 1999 (3.1 M 178 pages)
Initial book of Potto project. Start of economy, dimensional analysis, pQ2 diagram chapters.
This document published Modied FDL. The change of the license is to prevent from situations where the author has to buy his own book. The Potto Project License isnt long apply to this document and associated docoments.
Preamble
The purpose of this License is to make a manual, textbook, or other functional and useful document free in the sense of freedom: to assure everyone the eective freedom to copy and redistribute it, with or without modifying it, either commercially or noncommercially. Secondarily, this License preserves for the author and
xix
xx
LIST OF TABLES
publisher a way to get credit for their work, while not being considered responsible for modications made by others. This License is a kind of copyleft, which means that derivative works of the document must themselves be free in the same sense. It complements the GNU General Public License, which is a copyleft license designed for free software. We have designed this License in order to use it for manuals for free software, because free software needs free documentation: a free program should come with manuals providing the same freedoms that the software does. But this License is not limited to software manuals; it can be used for any textual work, regardless of subject matter or whether it is published as a printed book. We recommend this License principally for works whose purpose is instruction or reference.
xxi
images composed of pixels) generic paint programs or (for drawings) some widely available drawing editor, and that is suitable for input to text formatters or for automatic translation to a variety of formats suitable for input to text formatters. A copy made in an otherwise Transparent le format whose markup, or absence of markup, has been arranged to thwart or discourage subsequent modication by readers is not Transparent. An image format is not Transparent if used for any substantial amount of text. A copy that is not Transparent is called Opaque. Examples of suitable formats for Transparent copies include plain ASCII without markup, Texinfo input format, LaTeX input format, SGML or XML using a publicly available DTD, and standard-conforming simple HTML, PostScript or PDF designed for human modication. Examples of transparent image formats include PNG, XCF and JPG. Opaque formats include proprietary formats that can be read and edited only by proprietary word processors, SGML or XML for which the DTD and/or processing tools are not generally available, and the machine-generated HTML, PostScript or PDF produced by some word processors for output purposes only. The Title Page means, for a printed book, the title page itself, plus such following pages as are needed to hold, legibly, the material this License requires to appear in the title page. For works in formats which do not have any title page as such, Title Page means the text near the most prominent appearance of the works title, preceding the beginning of the body of the text. A section Entitled XYZ means a named subunit of the Document whose title either is precisely XYZ or contains XYZ in parentheses following text that translates XYZ in another language. (Here XYZ stands for a specic section name mentioned below, such as Acknowledgements, Dedications, Endorsements, or History.) To Preserve the Title of such a section when you modify the Document means that it remains a section Entitled XYZ according to this denition. The Document may include Warranty Disclaimers next to the notice which states that this License applies to the Document. These Warranty Disclaimers are considered to be included by reference in this License, but only as regards disclaiming warranties: any other implication that these Warranty Disclaimers may have is void and has no eect on the meaning of this License.
2. VERBATIM COPYING
You may copy and distribute the Document in any medium, either commercially or noncommercially, provided that this License, the copyright notices, and the license notice saying this License applies to the Document are reproduced in all copies, and that you add no other conditions whatsoever to those of this License. You may not use technical measures to obstruct or control the reading or further copying of the copies you make or distribute. However, you may accept compensation in exchange for copies. If you distribute a large enough number of copies you must also follow the conditions in section 3. You may also lend copies, under the same conditions stated above, and you may publicly display copies.
3. COPYING IN QUANTITY
xxii
LIST OF TABLES
If you publish printed copies (or copies in media that commonly have printed covers) of the Document, numbering more than 100, and the Documents license notice requires Cover Texts, you must enclose the copies in covers that carry, clearly and legibly, all these Cover Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on the back cover. Both covers must also clearly and legibly identify you as the publisher of these copies. The front cover must present the full title with all words of the title equally prominent and visible. You may add other material on the covers in addition. Copying with changes limited to the covers, as long as they preserve the title of the Document and satisfy these conditions, can be treated as verbatim copying in other respects. If the required texts for either cover are too voluminous to t legibly, you should put the rst ones listed (as many as t reasonably) on the actual cover, and continue the rest onto adjacent pages. If you publish or distribute Opaque copies of the Document numbering more than 100, you must either include a machine-readable Transparent copy along with each Opaque copy, or state in or with each Opaque copy a computer-network location from which the general network-using public has access to download using public-standard network protocols a complete Transparent copy of the Document, free of added material. If you use the latter option, you must take reasonably prudent steps, when you begin distribution of Opaque copies in quantity, to ensure that this Transparent copy will remain thus accessible at the stated location until at least one year after the last time you distribute an Opaque copy (directly or through your agents or retailers) of that edition to the public. It is requested, but not required, that you contact the authors of the Document well before redistributing any large number of copies, to give them a chance to provide you with an updated version of the Document.
4. MODIFICATIONS
You may copy and distribute a Modied Version of the Document under the conditions of sections 2 and 3 above, provided that you release the Modied Version under precisely this License, with the Modied Version lling the role of the Document, thus licensing distribution and modication of the Modied Version to whoever possesses a copy of it. In addition, you must do these things in the Modied Version: A. Use in the Title Page (and on the covers, if any) a title distinct from that of the Document, and from those of previous versions (which should, if there were any, be listed in the History section of the Document). You may use the same title as a previous version if the original publisher of that version gives permission. B. List on the Title Page, as authors, one or more persons or entities responsible for authorship of the modications in the Modied Version, together with at least ve of the principal authors of the Document (all of its principal authors, if it has fewer than ve), unless they release you from this requirement. C. State on the Title page the name of the publisher of the Modied Version, as the publisher.
GNU FREE DOCUMENTATION LICENSE D. Preserve all the copyright notices of the Document.
xxiii
E. Add an appropriate copyright notice for your modications adjacent to the other copyright notices. F. Include, immediately after the copyright notices, a license notice giving the public permission to use the Modied Version under the terms of this License, in the form shown in the Addendum below. G. Preserve in that license notice the full lists of Invariant Sections and required Cover Texts given in the Documents license notice. H. Include an unaltered copy of this License. I. Preserve the section Entitled History, Preserve its Title, and add to it an item stating at least the title, year, new authors, and publisher of the Modied Version as given on the Title Page. If there is no section Entitled History in the Document, create one stating the title, year, authors, and publisher of the Document as given on its Title Page, then add an item describing the Modied Version as stated in the previous sentence. J. Preserve the network location, if any, given in the Document for public access to a Transparent copy of the Document, and likewise the network locations given in the Document for previous versions it was based on. These may be placed in the History section. You may omit a network location for a work that was published at least four years before the Document itself, or if the original publisher of the version it refers to gives permission. K. For any section Entitled Acknowledgements or Dedications, Preserve the Title of the section, and preserve in the section all the substance and tone of each of the contributor acknowledgements and/or dedications given therein. L. Preserve all the Invariant Sections of the Document, unaltered in their text and in their titles. Section numbers or the equivalent are not considered part of the section titles. M. Delete any section Entitled Endorsements. Such a section may not be included in the Modied Version. N. Do not retitle any existing section to be Entitled Endorsements or to conict in title with any Invariant Section. O. Preserve any Warranty Disclaimers. If the Modied Version includes new front-matter sections or appendices that qualify as Secondary Sections and contain no material copied from the Document, you may at your option designate some or all of these sections as invariant. To do this, add their titles to the list of Invariant Sections in the Modied Versions license notice. These titles must be distinct from any other section titles.
xxiv
LIST OF TABLES
You may add a section Entitled Endorsements, provided it contains nothing but endorsements of your Modied Version by various partiesfor example, statements of peer review or that the text has been approved by an organization as the authoritative denition of a standard. You may add a passage of up to ve words as a Front-Cover Text, and a passage of up to 25 words as a Back-Cover Text, to the end of the list of Cover Texts in the Modied Version. Only one passage of Front-Cover Text and one of Back-Cover Text may be added by (or through arrangements made by) any one entity. If the Document already includes a cover text for the same cover, previously added by you or by arrangement made by the same entity you are acting on behalf of, you may not add another; but you may replace the old one, on explicit permission from the previous publisher that added the old one. The author(s) and publisher(s) of the Document do not by this License give permission to use their names for publicity for or to assert or imply endorsement of any Modied Version.
5. COMBINING DOCUMENTS
You may combine the Document with other documents released under this License, under the terms dened in section 4 above for modied versions, provided that you include in the combination all of the Invariant Sections of all of the original documents, unmodied, and list them all as Invariant Sections of your combined work in its license notice, and that you preserve all their Warranty Disclaimers. The combined work need only contain one copy of this License, and multiple identical Invariant Sections may be replaced with a single copy. If there are multiple Invariant Sections with the same name but dierent contents, make the title of each such section unique by adding at the end of it, in parentheses, the name of the original author or publisher of that section if known, or else a unique number. Make the same adjustment to the section titles in the list of Invariant Sections in the license notice of the combined work. In the combination, you must combine any sections Entitled History in the various original documents, forming one section Entitled History; likewise combine any sections Entitled Acknowledgements, and any sections Entitled Dedications. You must delete all sections Entitled Endorsements.
6. COLLECTIONS OF DOCUMENTS
You may make a collection consisting of the Document and other documents released under this License, and replace the individual copies of this License in the various documents with a single copy that is included in the collection, provided that you follow the rules of this License for verbatim copying of each of the documents in all other respects. You may extract a single document from such a collection, and distribute it individually under this License, provided you insert a copy of this License into the extracted document, and follow this License in all other respects regarding verbatim copying of that document.
xxv
8. TRANSLATION
Translation is considered a kind of modication, so you may distribute translations of the Document under the terms of section 4. Replacing Invariant Sections with translations requires special permission from their copyright holders, but you may include translations of some or all Invariant Sections in addition to the original versions of these Invariant Sections. You may include a translation of this License, and all the license notices in the Document, and any Warranty Disclaimers, provided that you also include the original English version of this License and the original versions of those notices and disclaimers. In case of a disagreement between the translation and the original version of this License or a notice or disclaimer, the original version will prevail. If a section in the Document is Entitled Acknowledgements, Dedications, or History, the requirement (section 4) to Preserve its Title (section 1) will typically require changing the actual title.
9. TERMINATION
You may not copy, modify, sublicense, or distribute the Document except as expressly provided for under this License. Any other attempt to copy, modify, sublicense or distribute the Document is void, and will automatically terminate your rights under this License. However, parties who have received copies, or rights, from you under this License will not have their licenses terminated so long as such parties remain in full compliance.
xxvi
LIST OF TABLES
version applies to it, you have the option of following the terms and conditions either of that specied version or of any later version that has been published (not as a draft) by the Free Software Foundation. If the Document does not specify a version number of this License, you may choose any version ever published (not as a draft) by the Free Software Foundation.
CONTRIBUTORS LIST
How to contribute to this book
As a copylefted work, this book is open to revisions and expansions by any interested parties. The only catch is that credit must be given where credit is due. This is a copyrighted work: it is not in the public domain! If you wish to cite portions of this book in a work of your own, you must follow the same guidelines as for any other GDL copyrighted work.
Credits
All entries have been arranged in alphabetical order of surname (hopefully). Major contributions are listed by individual name with some detail on the nature of the contribution(s), date, contact info, etc. Minor contributions (typo corrections, etc.) are listed by name only for reasons of brevity. Please understand that when I classify a contribution as minor, it is in no way inferior to the eort or value of a major contribution, just smaller in the sense of less text changed. Any and all contributions are gratefully accepted. I am indebted to all those who have given freely of their own knowledge, time, and resources to make this a better book!
Date(s) of contribution(s): 1999 to present Nature of contribution: Original author. Contact at: barmeir at gmail.com
xxvii
xxviii
CONTRIBUTORS LIST
Tousher Yang
Date(s) of contribution(s): Mat 2008 Nature of contribution: Major review of dimensional analysis and intro chapters.
Steve Spurgeon
Date(s) of contribution(s): November 200x Nature of contribution: Correction to pQ2 diagram derivations.
Irene Tan
Date(s) of contribution(s): January, 2009 Nature of contribution: Repair of dimensional analysis chapter.
Genick Bar-Meir holds a Ph.D. in Mechanical Engineering from University of Minnesota and a Master in Fluid Mechanics from Tel Aviv University. Dr. Bar-Meir was the last student of the late Dr. R.G.E. Eckert. Much of his time has been spend doing research in the eld of heat and mass transfer (related to renewal energy issues) and this includes uid mechanics related to manufacturing processes and design. Currently, he spends time writing books (there are already three very popular books) and softwares for the POTTO project (see Potto Prologue). The author enjoys to encourage his students to understand the material beyond the basic requirements of exams. Bar-Meirs books are used by hundred of thousands of peoples. His book on compressible is the most popular and preferred by practitioners and students. His books books are used in many universities like Purdue, Caltech, Queens University in Canada, and Singapore. One reason that his books are so popular is that they contain up to date material much of it original work by Bar-Meir. In his early part of his professional life, Bar-Meir was mainly interested in elegant models whether they have or not a practical applicability. Now, this authors views had changed and the virtue of the practical part of any model becomes the essential part of his ideas, books and software. He developed models for Mass Transfer in high concentration that became a building blocks for many other models. These models are based on analytical solution to a family of equations1 . As the change in the view occurred, Bar-Meir developed models that explained several manufacturing processes such the rapid evacuation of gas from containers, the critical piston velocity in a partially lled chamber (related to hydraulic jump), application of supply and demand to rapid change power system and etc. All the models have practical applicability. These models have been extended by several research groups (needless to say with large research grants). For example, the Spanish Comision Interministerial provides grants TAP97-0489 and PB98-0007, and the CICYT and the European Commission provides
1 Where
the mathematicians were able only to prove that the solution exists.
xxix
xxx
CONTRIBUTORS LIST
1FD97-2333 grants for minor aspects of that models. Moreover, the authors models were used in numerical works, in GM, British industry, Spain, and Canada. In the area of compressible ow, it was commonly believed and taught that there is only weak and strong shock and it is continue by PrandtlMeyer function. Bar Meir discovered the analytical solution for oblique shock and showed that there is a quiet buer between the oblique shock and PrandtlMeyer. He also build analytical solution to several moving shock cases. He described and categorized the lling and evacuating of chamber by compressible uid in which he also found analytical solutions to cases where the working uid was ideal gas. The common explanation to PrandtlMeyer function shows that ow can turn in a sharp corner. Engineers have constructed design that based on this conclusion. Bar-Meir demonstrated that common PrandtlMeyer explanation violates the conservation of mass and therefor the turn must be around a nite radius. The authors explanations on missing diameter and other issues in fanno ow and naughty professors question are used in the industry. In his book Basics of Fluid Mechanics, Bar-Meir demonstrated that uids must have wavy surface when two dierent materials ow together. All the previous models for the ooding phenomenon did not have a physical explanation to the dryness. He built a model to explain the ooding problem (two phase ow) based on the physics. He also constructed and explained many new categories for two ow regimes. The author lives with his wife and three children. A past project of his was building a four stories house, practically from scratch. While he writes his programs and does other computer chores, he often feels clueless about computers and programing. While he is known to look like he knows about many things, the author just know to learn quickly. The author spent years working on the sea (ships) as a engine sea ocer but now the author prefers to remain on solid ground.
xxxi
xxxii
purpose than practicing the solving of problems and exercises. One can be successful when one solves as many problems as possible. To reach this possibility the collective book idea was created/adapted. While one can be as creative as possible, there are always others who can see new aspects of or add to the material. The collective material is much richer than any single person can create by himself. The following example explains this point: The army ant is a kind of carnivorous ant that lives and hunts in the tropics, hunting animals that are even up to a hundred kilograms in weight. The secret of the ants power lies in their collective intelligence. While a single ant is not intelligent enough to attack and hunt large prey, the collective power of their networking creates an extremely powerful intelligence to carry out this attack4 . When an insect which is blind can be so powerful by networking, So can we in creating textbooks by this powerful tool.
Why Volunteer?
Why would someone volunteer to be an author or organizer of such a book? This is the rst question the undersigned was asked. The answer varies from individual to individual. It is hoped that because of the open nature of these books, they will become the most popular books and the most read books in their respected eld. For example, the books on compressible ow and die casting became the most popular books in their respective area. In a way, the popularity of the books should be one of the incentives for potential contributors. The desire to be an author of a wellknown book (at least in his/her profession) will convince some to put forth the eort. For some authors, the reason is the pure fun of writing and organizing educational material. Experience has shown that in explaining to others any given subject, one also begins to better understand the material. Thus, contributing to these books will help one to understand the material better. For others, the writing of or contributing to this kind of books will serve as a social function. The social function can have at least two components. One component is to come to know and socialize with many in the profession. For others the social part is as simple as a desire to reduce the price of college textbooks, especially for family members or relatives and those students lacking funds. For some contributors/authors, in the course of their teaching they have found that the textbook they were using contains sections that can be improved or that are not as good as their own notes. In these cases, they now have an opportunity to put their notes to use for others. Whatever the reasons, the undersigned believes that personal intentions are appropriate and are the authors/organizers private aair. If a contributor of a section in such a book can be easily identied, then that contributor will be the copyright holder of that specic section (even within question/answer sections). The books contributors names could be written by their sections. It is not just for experts to contribute, but also students who happened to be doing their homework. The students contributions can be done by adding a question and perhaps the solution. Thus, this method is expected to accelerate the creation of
4 see also in Franks, Nigel R.; Army Ants: A Collective Intelligence, American Scientist, 77:139, 1989 (see for information http://www.ex.ac.uk/bugclub/raiders.html)
xxxiii
these high quality books. These books are written in a similar manner to the open source software process. Someone has to write the skeleton and hopefully others will add esh and skin. In this process, chapters or sections can be added after the skeleton has been written. It is also hoped that others will contribute to the question and answer sections in the book. But more than that, other books contain data5 which can be typeset in A LTEX. These data (tables, graphs and etc.) can be redone by anyone who has the time to do it. Thus, the contributions to books can be done by many who are not experts. Additionally, contributions can be made from any part of the world by those who wish to translate the book. It is hoped that the books will be error-free. Nevertheless, some errors are possible and expected. Even if not complete, better discussions or better explanations are all welcome to these books. These books are intended to be continuous in the sense that there will be someone who will maintain and improve the books with time (the organizer(s)). These books should be considered more as a project than to t the traditional denition of plain books. Thus, the traditional role of author will be replaced by an organizer who will be the one to compile the book. The organizer of the book in some instances will be the main author of the work, while in other cases only the gate keeper. This may merely be the person who decides what will go into the book and what will not (gate keeper). Unlike a regular book, these works will have a version number because they are alive and continuously evolving.
Pr og re
Project Name
Remarks
Version
Based on Eckert
Number DownLoads
ss
xxxiv
Mechanics Open Channel Flow Statics Strength of Material Thermodynamics Two/Multi ow phases
Pr og
Project Name
Remarks
Version
rst chapter
Number DownLoads -
re s
TelAvivnotes
0.0.0
The mature stage of a chapter is when all or nearly all the sections are in a mature stage and have a mature bibliography as well as numerous examples for every section. The mature stage of a section is when all of the topics in the section are written, and all of the examples and data (tables, gures, etc.) are already presented. While some terms are dened in a relatively clear fashion, other denitions give merely a hint on the status. But such a thing is hard to dene and should be enough for this stage. The idea that a book can be created as a project has mushroomed from the open source software concept, but it has roots in the way science progresses. However, traditionally books have been improved by the same author(s), a process in which books have a new version every a few years. There are book(s) that have continued after their author passed away, i.e., the Boundary Layer Theory originated6 by Hermann Schlichting but continues to this day. However, projects such as the Linux Documentation project
6 Originally authored by Dr. Schlichting, who passed way some years ago. A new version is created every several years.
xxxv
demonstrated that books can be written as the cooperative eort of many individuals, many of whom volunteered to help. Writing a textbook is comprised of many aspects, which include the actual writing of the text, writing examples, creating diagrams and gures, and writing the A LTEX macros7 which will put the text into an attractive format. These chores can be done independently from each other and by more than one individual. Again, because of the open nature of this project, pieces of material and data can be used by dierent books.
7 One can only expect that open source and readable format will be used for this project. But more A than that, only LTEX, and perhaps tro, have the ability to produce the quality that one expects for A these writings. The text processes, especially LTEX, are the only ones which have a cross platform ability to produce macros and a uniform feel and quality. Word processors, such as OpenOce, Abiword, and Microsoft Word software, are not appropriate for these projects. Further, any text that is produced by Microsoft and kept in Microsoft format are against the spirit of this project In that they force spending money on Microsoft software.
xxxvi
xxxvii
xxxviii
xxxix
Is it only an accident that both the quality of the typesetting of papers in die casting congress and their technical content quality is so low? I believe there is a connection. A All the major magazines of the the scientic world using TEX or LTEX, why? Because it is very easy to use and transfer (via the Internet) and, more importantly, because it produces high quality documents. NADCA continued to produce text on a low quality word processor. Look for yourself; every transaction is ugly. Linux has liberated the world from the occupation and the control of Microsoft OS. We hope to liberate the NADCA Transaction from such a poor quality word processor. TEX and all (the good ones) supporting programs are free and available every where on the web. There is no reason not to do it. Please join me in improving A NADCAs Transaction by supporting the use of LTEX by NADCA.
Will I Be in Trouble?
Initial part
Many people have said I will be in trouble because I am telling the truth. Those with a vested interest in the status quo (North American Die Casting Association, and thus research that this author exposed there poor and or erroneous work). will try to use their power to destroy me. In response, I challenge my opponents to show that they are right. If they can do that, I will stand wherever they want and say that I am wrong and they are right. However, if they cannot prove their models and practices are based on solid scientic principles, nor nd errors with my models (and I do not mean typos and English mistakes), then they should accept my results and help the diecasting industry prosper. People have also suggested that I get life insurance and/or good lawyer because my opponents are very serious and mean business; the careers of several individuals are in jeopardy because of the truths I have exposed. If something does happen to me, then you, the reader, should punish them by supporting science and engineering and promoting the diecasting industry. By doing so, you prevent them from manipulating the industry and gaining additional wealth. For the sake of my family, I have, in fact, taken out a life insurance policy. If something does happen to me, please send a thank you and work well done card to my family.
xl
This book started because I was frustrated with the system that promote erroneous research. Then, I realized that the book cannot be stolen if it under open content. The die casting process is interesting enough to insert my contributions. I have found that works or model in this area are lack of serious scientic principles. I have started to write class notes to my clients and I add my research work to create this book. During the writing I add the material on economy which I felt is missing piece of knowledge in the die casting engineering world. Of course, this book was written on Linux (Micro$oftLess book). This book was written using the vim editor for editing (sorry never was able to be comfortable with emacs). The graphics were done by TGIF, the best graphic program that this author experienced so far. The gures were done by grap but will be modied to gle. The spell checking was done by ispell, and hope to nd a way to use gaspell, a program that currently cannot be used on new Linux systems. The gure in cover page was created by Genick Bar-Meir, and is copylefted by him.
xli
xlii
Abstract
Die-casting engineers have to compete not only with other die-casting companies, but also against other industries such as plastics, and composite materials. Clearly, the black art approach, which has been an inseparable part of the engineers tools, is in need of being replaced by a scientic approach. Excuses that science has not and never will work need to be replaced with science does work. All technologies developed in recent years are described in a clear, simple manner in this book. All the errors of the old models and the violations of physical laws are shown. For example, the common pQ2 diagram violates many physical laws, such as the rst and second laws of thermodynamics. Furthermore, the common pQ2 diagram produces trends that are the opposite of reality, which are described in this book. The die casting engineers job is to produce maximum prots for the company. In order to achieve this aim, the engineer must design high quality products at a minimum cost. Thus, understanding the economics of the die casting design and process are essential. These are described in mathematical form for the rst time in this volume. Many new concepts and ideas are also introduced. For instance, how to minimize the scrap/cost due to the runner system, and what size of die casting machine is appropriate for a specic project. The die-casting industry is undergoing a revolution, and this book is part of it. One reason (if one reason can describe the situation) companies such as Doehler Jorvis (the biggest die caster in the world) and Shelby are going bankrupt is that they do not know how to calculate and reduce their production costs. It is my hope that die-casters will turn such situations around by using the technologies presented in this book. I believe this is the only way to keep the die casting professionals and the industry itself, from being left in the dust.
xliii
xliv
Preface
"In the beginning, the POTTO project was without form, and void; and emptiness was upon the face of the bits and files. And the Fingers of the Author moved upon the face of the keyboard. And the Author said, Let there be words, and there were words." 9 .
This book, Fundamentals of Die Casting Design, describes the fundamentals of die casting process design and economics for engineers and others. This book is designed to ll the gap and the missing book on economy and scientic principles of die casting. It is hoped that the book could be used as a reference book for people who have at least some basics knowledge of science areas such as calculus, physics, etc. It has to realized the some material is very advance and required knowledge of uid mechanics particularly compressible ow and open channel ow. This authors popular book on compressible ow should provide the introductory in that area. The readers reactions to this book and the usage of the book as a textbook suggested that the chapter which deals with economy should be expand. In the following versions this area will strength and expended. The structure of this book is such that many of the chapters could be usable independently. For example, if you need information about, say, economy of the large scale productions, you can read just chapter (12). I hope this makes the book easier to use as a reference manual. However, this manuscript is rst and foremost a textbook, and secondly a reference manual only as a lucky coincidence. I have tried to describe why the theories are the way they are, rather than just listing seven easy steps for each task. This means that a lot of information is presented which is not necessary for everyone. These explanations have been marked as such and can be skipped. Reading everything will, naturally, increase your understanding of the many aspects of uid mechanics.
9 To
the power and glory of the mighty God. This book is only to explain his power.
xlv
xlvi
This book is written and maintained on a volunteer basis. Like all volunteer work, there is a limit on how much eort I was able to put into the book and its organization. Moreover, due to the fact that English is my third language and time limitations, the explanations are not as good as if I had a few years to perfect them. Nevertheless, I believe professionals working in many engineering elds will benet from this information. This book contains many worked examples, which can be very useful for many. I have left some issues which have unsatisfactory explanations in the book, marked with a Mata mark. I hope to improve or to add to these areas in the near future. Furthermore, I hope that many others will participate of this project and will contribute to this book (even small contributions such as providing examples or editing mistakes are needed). I have tried to make this text of the highest quality possible and am interested in your comments and ideas on how to make it better. Incorrect language, errors, ideas for new areas to cover, rewritten sections, more fundamental material, more mathematics (or less mathematics); I am interested in it all. I am particularly interested in the best arrangement of the book. If you want to be involved in the editing, graphic design, or proofreading, please drop me a line. You may contact me via Email at barmeir at gmail dot com. Naturally, this book contains material that never was published before (sorry cannot avoid it). This material never went through a close content review. While close content peer review and publication in a professional publication is excellent idea in theory. In practice, this process leaves a large room to blockage of novel ideas and plagiarism. For example, Brevick from Ohio State is one the individual who attempt to block this author idea on pQ2 diagram. If you would like to critic to my new ideas please send me your comment(s). However, please do not hide your identity, it will cloud your motives. Several people have helped me with this book, directly or indirectly. I would like to especially thank to my adviser, Dr. E. R. G. Eckert, whose work was the inspiration for this book. I also would like to thank to Jannie McRotien (Open Channel Flow chapter) and Tousher Yang for their advices, ideas, and assistance. A I encourage anyone with a penchant for writing, editing, graphic ability, LTEX knowledge, and material knowledge and a desire to provide open content textbooks and to improve them to join me in this project. If you have Internet e-mail access, you can contact me at [email protected].
CHAPTER 1 Introduction
In the recent years, many die casting companies have gone bankrupt (DoehlerJarvis and Shelby to name a few) and many other die casting companies have been sold (St. Paul Metalcraft, Tool Products, OMC etc.). What is/are the reason/s for this situation? Some blame poor management. Others blame bad customers (which is mostly the automobile industry). Perhaps there is something to these claims. Nevertheless, one can see that the underlying reasons are the missing knowledge of how to calculate if there are prots for a production line and how to design, so that costs will be minimized. To demonstrate how the absurd situation is the fact that there is not even one company today that can calculate the actual price of any product that they are producing. Moreover, if a company is able to produce a specic product, no one in that company looks at the redesign (mold or process) in order to reduce the cost systematically. In order to compete with other industries and other companies, the die casting industry must reduce the cost as much as possible (20% to 40%) and lead time signicantly (by 1/2 or more). To achieve these goals, the engineer must learn to connect mold design to the cost of production (charged to the customer) and to use the correct scientic principals involved in the die casting process to reduce/eliminate the guess work. This book is part of the revolution in die casting by which science is replacing the black art of design. For the rst time, a link between the cost and the design is spelled out. Many new concepts, based on scientic principles, are introduced. The old models, which was plagued by the die casting industry for many decades, are analyzed, their errors are explained and the old models are superseded. Science is good, but it is not useful in the oor of our plant!! George Reed, the former president of SDCE, in 1999 announced in a meeting in the local chapter (16) of NADCA. He does not believe that there is A relationship between science and what he does with the die casting machine. He said that because he does not follow NADCA recommendations, he achieves good castings. For instance, he
CHAPTER 1. INTRODUCTION
stated that the common and NADCA supported, recommendation in order to increase the gate velocity, plunger diameter needs to be decreased. He said that because he does not follow this recommendation, and/or others, that is the reason his succeeds in obtaining good castings. He is right and wrong. He is right not to follow the NADCA recommendations since they violate many basic scientic principles. One should expect that models violating scientic principles would produce unrealistic results. When such results occur, this should actually strengthen the idea that science has validity. The fact that models which appear in books today are violating scientic principals and therefore do not work should actually convince him, and others, that science does have validity. Mr. Reed is right (in certain ranges) to increase the diameter in order to increase the gate velocity as will be covered in Chapter 7. The above example is but one Profits of many of models that are errant and % in need of correction. To this date, the 23 author has not found so much as a sin0 gle commonly used model that has been correct in its conclusions, trends, and/or assumptions. The wrong models/methods that have plagued the in~65-75 78 80 dustry are: 1) critical slow plunger velocity, 2) pQ2 diagram, 3) plunger diameter calculations, 4) runner system design, 5) vent system design, etc These incorrect Fig. -1.1. The prots as a function of the amount models are the reasons that science of the scrap. does not work. The models presented in this book are here for the purpose of answering the questions of design in a scientic manner which will result in reduction of costs and increased product quality. Once the reasons to why science does not work are clear, one should learn the correct models for improving quality, reducing lead time and reducing production cost. The main underlying reason people are in the die casting business is to make money. One has to use science to examine what the components of production cost/scrap are and how to minimize or eliminate each of them to increase protability. The underlying purpose of this book is to help the die caster to achieve this target.
23 20 maximum profits invesment cost breakeven point no scrap 0 100 scrap 100 0
has to exist somewhere between these two extremes. Typically, for the die casting industry, the breakeven point lies within the range of 55%75% product (or 25%-35% scrap). Typical prots in the die casting industry are or should be about 20%. When the prots fails below 15% or typical prot in the stock exchange then the production should stop. From Figure 1.1 it can be noticed that relative change in prots% = new product percent BEP 1 100 old product percent BEP (1.1)
Example 1.1: What would be the eect on the prots of a small change (2%) in a amount of scrap for a job with 22% scrap (78% product) and with breakeven point of 65%? Solution 80 65 1 100 = 15.3% 78 65 A reduction of 2% in a amount of the scrap to be 20% (80% product) results in increase of more than 15.3% in the prots.
End Solution
This is a very substantial dierence. Therefore, a much bigger reduction in scrap will result in much, much bigger prots.
Scrap Cost
400
BEP = 50.0
Scrap Cost
BEP = 50.0
140 350
BEP = 55.0 BEP = 55.0
120 300
BEP = 60.0
Incrase of prots
Incrase of prots
100
BEP = 60.0
250
BEP = 65.0
200
BEP = 70.0
150
60
BEP = 70.0
100
40
50
20
Scrap percent
December 8, 2008 December 8, 2008
Scrap percent
Fig a.
Fig b.
Fig. -1.2. The left graph depicts the increase of prots as reduction of the scrap for 10% + BEP. The right graph depicts same for for 20% + BEP.
To analysis this point further Figure intro:g:scrapCostBEP is built for two old scrap values, 10% more than the BEP on the left and 20% more than BEP on the right. The two gures (left and right) in Figure intro:g:scrapCostBEP demonstrate that The higher BEP the change in reduction of scrap is more important. The lower the old scrap point is the more important the reduction is.
CHAPTER 1. INTRODUCTION
is believed that better pricing results from such a practice. In todays practice, even after the project is nished, no one calculates the actual cost of production, let alone calculating the actual prots. The consequences of such a practice are clear: it results in no push for better design and with no idea which jobs make prots and which do not. Furthermore, considerable nancial cost is incurred which could easily be eliminated. Several chapters in this book are dedicated to linking the design to the cost (end-price).
CHAPTER 1. INTRODUCTION
As with all the previous models they missed the major points of the calculations. As will be shown in Chapter 7, the Australians model produce incorrect results and predict trends opposite to reality. This model took root in die casting industry for the last 25 years. Yet, one can only wonder why this well established method (the supply and demand theory which was build by Fanno (the brother of other famous Fanno from Fanno ow), which was introduced into uid mechanics in the early of this century, reached the die casting only in the late 70s and was then erroneously implemented. This methods now properly build for the rst time for the die casting industry in this book. Until the 1980s there was no model that assisted the understanding air entrapment in the shot sleeve. Garber described the hydraulic jump in the shot sleeve and called it the wave, probably because he was not familiar with this research area. He also developed the erroneous model which took root in the industry in spite the fact that it never works. One can only wonder why any die casting institutes/associations have not published this fact. Moreover, NADCA and other institutes continue to funnel large sums of money to the researchers (for example, Brevick from Ohio State) who used Garbers model even after they knew that Garbers model was totally wrong. The turning point of the understanding was when Prof. Eckert, the father of modern heat transfer, introduced the dimensional analysis applied to the die casting process. This established a scientic approach which provided an uniform schemata for uniting experimental work with the actual situations in the die casting process. Dimensional analysis demonstrates that the uid mechanics processes, such as lling of the cavity with liquid metal and evacuation/extraction of the air from the mold, can be dealt when the heat transfer is assumed to be negligible. However, the uid mechanics has to be taken into account in the calculations of the heat transfer process (the solidication process). This proved an excellent opportunity for simple models to predict the many parameters in the die casting process, which will be discussed later in this book. Here, two examples of new ideas that mushroomed in the inspiration of prof. Eckerts work. It has been shown that [5] the net eect of the reactions is negligible. This fact is contradictory to what was believed at that stage. The development of the critical vent area concept provided the major guidance for 1) the designs to the venting system, and 2) criterion when the vacuum system needs to be used. In this book, many of the new concepts and models, such as economy of the runner design, plunger diameter calculations, minimum runner design, etc, are described for the rst time.
problem and to account for atomization of the liquid metal. The atomization model in the last model was based on the mass transfer coecient. This model atomization is not appropriate. Clearly, this model is in waiting to be replaced by a realistic model to describe the mass transfer1 . The Enthalpy method was further exploded by Swaminathan and Voller and others to study the lling and solidication problem. While numerical simulation looks very promising, all the methods (nite dierence, nite elements, or boundary elements etc) 2 suer from several major drawbacks which prevents them from yielding reasonable results.
There is no theory (model) that explains the heat transfer between the mold walls and the liquid metal. The lubricant sprayed on the mold change the characteristic of the heat transfer. The dierence in the density between the liquid phase and solid phase creates a gap during the solidication process between the mold and the ingate which depends on the geometry. For example, Osborne et al showed that a commercial software (MAGMA) required ddling with the heat transfer coecient to get the numerical simulation match the experimental results3 . As it was mentioned earlier, it is not clear when the liquid metal ows as a spray and when it ows as continuous liquid. Experimental work has demonstrated that the ow, for a large part of the lling time, is atomized [4]. The pressure in the mold cavity in all the commercial codes are calculated without taking into account the resistance to the air ow out. Thus, builtup pressure in the cavity is poorly estimated, or even not realistic, and therefore the characteristic ow of the liquid metal in the mold cavity is poorly estimated as well. The ow in all the simulations is assumed to be turbulent ow. However, time and space are required to achieved a fully turbulent ow. For example, if the ow at the entrance to a pipe with the typical conditions in die casting is laminar (actually it is a plug ow) it will take a runner with a length of about 10[m] to achieved fully developed ow. With this in mind, clearly some part of the ow is laminar. Additionally, the solidication process is faster compared to the dissipation process in the initial stage, so it is also a factor in changing the ow from a turbulent (in case the ow is turbulent) to a laminar ow. The liquid metal velocity at the entrance to the runner is assumed for the numerical simulation and not calculated. In reality this velocity has to be calculated utilizing the pQ2 diagram. If turbulence exists in the ow eld, what is the model that describes it adequately? Clearly, model such k are based on isentropic homogeneous with mild change in the properties cannot describe situations where the ow changes into two-phase ow (solid-liquid ow) etc.
nds that it is the easiest to critic ones own work or where he/she was involved. or academic versions. 3 Actually, they attempted to prove that the software is working very well. However, the fact that coecient need to eld is excellent proof why this work is meaningless.
2 Commercial 1 One
CHAPTER 1. INTRODUCTION
The heat extracted from the die is done by cooling liquid (oil or water). In most models (all the commercial models) the mechanism is assumed to be by regular cooling. In actuality, some part of the heat is removed by boiling heat transfer. The governing equations in all the numerical models, that I am aware of, neglect the dissipation term in during the solidication. The dissipation term is the most important term in that case.
One wonders how, with unknown ow pattern (or correct ow pattern), unrealistic pressure in the mold, wrong heat removal mechanism (cooling method), erroneous governing equation in the solidication phase, and inappropriate heat transfer coecient, a simulation could produce any realistic results. Clearly, much work is need to be done in these areas before any realistic results should be expected from any numerical simulation. Furthermore, to demonstrate this point, there are numerical studies that assume that the ow is turbulent, continuous, no air exist (or no air leaving the cavity) and proves with their experiments that their model simulate reality [23]. On the other hand, other numerical studies assumed that the ow does not have any eect on the solidication and of course have their experiments to support this claim [11]. Clearly, this contradiction suggest several options:
Both of the them are right and the model itself does not matter. One is right and the other one is wrong. Both of them are wrong.
The third research we mentioned here is an example where the calculations can be shown to be totally wrong and yet the researchers have experimental proofs to back them up. Viswanathan et al studied a noble process in which the liquid metal is poured into the cavity and direct pressure is applied to the cavity. In their calculations the authors assumed that metal enter to the cavity and ll the whole entrance (gate) to the cavity. Based on this assumption their model predict defects in certain geometry. A critical examination of this model present the following. The assumption of no air ow out by the authors (was explained privately that air amount is a small and therefore not important) is very critical as will be shown here. The volumetric air ow rate into the cavity has to be on average equal to liquid metal ow rate (conservation of volume for constant density). Hence, air velocity has to be approximately innite to achieve zero vent area. Conversely, if the assumption that the air ows in the same velocity as the liquid entering the cavity, liquid metal ow area is a half what is assume in the researchers model. In realty, the ow of the liquid metal is in the two phase region and in this case, it is like turning a bottle full of water over and liquid inside ows as blobs 4 . More information can be found on reversible ow in this author book in Potto series of Basics of Fluid Mechanics. In this case the whole calculations do not have much to do with reality since the velocity is not continuous and dierent from what was calculated.
4 Try
it your self! ll a bottle and turn it upside and see what happens.
Another example of such study is the model of the ow in the shot sleeve by Backer and Sant from EKK [2]5 . The researchers assumed that the ow is turbulent and they justied it because they calculated and found a jet with extreme velocity. Unfortunately, all the experimental evidence demonstrate that there is no such jet [24]. It seems that this jet results from the poor boundary and initial conditions6 . In the presentation, the researchers also stated that results they obtained for laminar and turbulent ow were the same7 while a simple analysis can demonstrate the dierence is very large. Also, one can wonder how liquid with zero velocity to be turbulent. With these results one can wonder if the code is of any value or the implementation is at fault. The bizarre belief that the numerical simulations are a panacea to all the design problem is very popular in the die casting industry. Any model has to describe and account for the physical situation in order to be useful. Experimental evidence which is supporting wrong models as a real evidence is nonsense. Clearly some wrong must be there. For example, see the paper by Murray and colleague in which they use the fact that two unknown companies (somewhere in the outer space maybe?) were using their model to claim that it is correct.. A proper way can be done by numerical calculations based on real physics principles which produce realistic results. Until that point come, the reader should be suspicious about any numerical model and its supporting evidence.8
10
CHAPTER 1. INTRODUCTION
It turned out that many of the design parameters in die casting have a critical point above which good castings are produced and below which poor castings are produced. Furthermore, much above and just above the critical point do not change much the quality but costs much more. This fact is where the economical concepts plays a signicant role. Using these concepts, one can increase the protability signicantly, and obtain very good quality casting and reduce the leading time. Additionally, the main cost components like machine cost and other are analyzed which have to be taken into considerations when one chooses to design the process will be discussed in the Chapter 12. Porosity can be divided into two main categories; shrinkage porosity and gas/air entrainment. The porosity due to entrapped gases constitutes a large part of the total porosity. The creation of gas/air entrainment can be attributed to at least four categories: lubricant evaporation (and reaction processes9 ), vent locations (last place to be lled), mixing processes, and vent/gate area. The eects of lubricant evaporation have been found to be insignicant. The vent location(s) can be considered partially solved since only qualitative explanation exist. The mixing mechanisms are divided into two zones: the mold, and the shot sleeve. Some mixing processes have been investigated and can be considered solved. The requirement on the vent/gate areas is discussed in Chapter 9. When the mixing processes are very signicant in the mold, other methods are used and they include: evacuating the cavities (vacuum venting), Pore Free Technique (in zinc and aluminum casting) and squeeze casting. The rst two techniques are used to extract the gases/air from the shot sleeve and die cavity before the gases have the opportunity to mix with the liquid metal. The squeeze casting is used to increase the capillary forces and therefore, to minimize the mixing processes. All these solutions are cumbersome and more expensive and should be avoided if possible. The mixing processes in the runners, where the liquid metal ows vertically against gravity in relatively large conduit, are considered to be insignicant10 . The enhanced air entrainment in the shot sleeve is attributed to operational conditions for which a blockage of the gate by a liquid metal wave occurs before the air is exhausted. Consequently, the residual air is forced to be mixed into the liquid metal in the shot sleeve. With Bar-Meirs formula, one can calculate the correct critical slow plunger velocity and this will be discussed in Chapter 8.
1.7 Summary
It is an exciting time in the die casing industry because for the rst time, an engineer can start using real science in designing the runner/mold and the die casting process. Many new models have been build and many old techniques mistake have been removed. It is the new revolution in the die casting industry.
9 Some researchers view the chemical reactions (e.g. release of nitrogen during solidication process) as category by itself. 10 Some work has been carried out and hopefully will be published soon. And inside, in the book Basic of Fluid Mechanics in the two phase chapter some inside was developed.
11
12
2.2.1
What is Fluid?
The uid is mainly divided into two categories: liquids and gases. The main dierence between the liquids and gases state is that gas will occupy the whole volume while liquids has an almost xed volume. This dierence can be, for most practical purposes considered, sharp even though in reality this dierence isnt sharp. The dierence between a gas phase to a liquid phase above the critical point are practically minor. But below the critical point, the change of water pressure by 1000% only change the volume by less than 1 percent. For example, a change in the volume by more than 5% will require tens of thousands percent change of the pressure. So, if the change of pressure is signicantly less than that, then the change of volume is at best 5%. Hence, the pressure will not aect the volume. In gaseous phase, any change in pressure directly aects the volume. The gas lls the volume and liquid cannot. Gas has no free interface/surface (since it does ll the entire volume).
2.2.2
The shear stress is part of the pressure tensor. However, here it will be treated as a separate issue. In solid mechanics, the shear stress is considered as the ratio of the force acting on area in the direction of the forces perpendicular to area. Dierent from solid, uid cannot pull directly but through a solid surface. Consider liquid that undergoes a shear stress between a short distance of two plates as shown in Figure (2.2). The upper plate velocity generally will be U = f (A, F, h) (2.1)
Where A is the area, the F denotes the force, h is the distance between the plates. From solid mechanics study, it was shown that when the force per area increases, the velocity of the plate increases also. Experiments show that the increase of height will increase the velocity up to a certain range. Consider moving the plate with a zero lubricant (h 0) (results in large force) or a large amount of lubricant (smaller force). In this discussion, the aim is to develop dierential equation, thus the small distance analysis is applicable. For cases where the dependency is linear, the following can be written U hF A (2.2)
2.2. WHAT IS FLUID? SHEAR STRESS Equations (2.2) can be rearranged to be U F h A Shear stress was dened as xy = F A
13
(2.3)
(2.4)
From equations (2.3) and (2.4) it follows that ratio of the velocity to height is proportional to shear stress. Hence, applying the coecient to obtain a new equality as xy = U h (2.5)
Where is called the absolute viscosity or dynamic viscosity. In steady state, the distance the upper plate moves after small amount of t0 < time, t is d = U t (2.6)
t1 < t2 <
t3
From gure (2.2) it can be noticed that for a small angle, the regular approximation provides
geometry
d = U t =
(2.7)
Fig. -2.2. The deformation of uid due to shear stress as progression of time.
If the velocity prole is linear between the plate (it will be shown later that it is consistent with derivations of velocity), then it can be written for small angle that dU = t dy (2.10)
Materials which obey equation (2.9) are referred to as Newtonian uid. For liquid metal used in the die casting industry, this property should be considered as Newtonian uid.
14
2.3.2
Basic Denitions
The following basic denitions are common to thermodynamics and will be used in this book. Work In mechanics, the work was dened as mechanical work = Fd = P dV (2.11)
This denition can be expanded to include two issues. The rst issue that must be addressed, that work done on the surroundings by the system boundaries similarly is positive. Two, there is a transfer of energy so that its eect can cause work. It must be noted that electrical current is a work while heat transfer isnt. System This term will be used in this book and it is dened as a continuous (at least partially) xed quantity of matter (neglecting Einsteins law eects). For almost all engineering purposes this law is reduced to two separate laws: mass conservation and energy conservation. Our system can receive energy, work, etc as long as the mass remains constant the denition is not broken. Thermodynamics First Law This law refers to conservation of energy in a non accelerating system. Since all the systems can be calculated in a non accelerating system, the conservation is applied to all systems. The statement describing the law is the following: Q12 W12 = E2 E1 (2.12)
The system energy is a state property. From the rst law it directly implies that for process without heat transfer (adiabatic process) the following is true W12 = E1 E2 (2.13)
Interesting results of equation (2.13) is that the way the work is done and/or intermediate states are irrelevant to nal results. The internal energy is the energy that depends on the other properties of the system. Example: for pure/homogeneous and
15
simple gases it depends on two properties like temperature and pressure. The internal energy is denoted in this book as EU and it will be treated as a state property. The system potential energy is dependent upon the body force. A common body force is gravity. For such body force, the potential energy is mgz where g is the gravity force (acceleration), m is the mass and the z is the vertical height from a datum. The kinetic energy is K.E. = mU 2 2 (2.14)
Thus the energy equation can be written as mU1 2 mU2 2 + m g z1 +EU 1 + Q = + m g z2 +EU 2 + W (2.15) 2 2 where Bf is a body force. For the unit mass of the system equation (2.15) is transformed into U2 2 U1 2 + gz1 + Eu 1 + q = + gz2 + Eu 2 + w (2.16) 2 2 where q is the energy per unit mass and w is the work per unit mass. The new internal energy, Eu , is the internal energy per unit mass. Since the above equations are true between arbitrary points, choosing any point in time will make it correct. Thus, dierentiating the energy equation with respect to time yields the rate of change energy equation. The rate of change of the energy transfer is DQ =Q Dt (2.17)
Bf Bf
In the same manner, the work change rate transferred through the boundaries of the system is DW =W Dt Since the system is with a xed mass, the rate energy equation is D EU DU Dgz QW = + mU +m Dt Dt Dt (2.19) (2.18)
For the case were the body force, Bf = g, is constant with time like in the case of gravity equation (2.19) reduced to D EU DU Dz QW = + mU + mg Dt Dt Dt
(2.20)
16
The time derivative operator, D/Dt is used instead of the common notation because it refers to system property derivative. Thermodynamics Second Law There are several denitions of the second law. No matter which denition is used to describe the second law it will end in a mathematical form. The most common mathematical form is Clausius inequality which state that Q 0 T (2.21)
The integration symbol with the circle represent integral of cycle (therefore circle) of system which returns to the same condition. If there is no lost, it is referred as a reversible process and the inequality change to equality. Q =0 T (2.22)
The last integral can go though several states. These states are independent of the path the system goes through. Hence, the integral is independent of the path. This observation leads to the denition of entropy and designated as S and the derivative of entropy is ds Q T rev
2
(2.23)
S2 S1 =
1
Q = T rev
dS
1
(2.24)
One of the conclusions that can be drawn from this analysis is for reversible and adiabatic process dS = 0. Thus, the process in which it is reversible and adiabatic, the entropy remains constant and referred to as isentropic process. It can be noted that there is a possibility that a process can be irreversible and the right amount of heat transfer to have zero change entropy change. Thus, the reverse conclusion that zero change of entropy leads to reversible process, isnt correct. For reversible process equation (2.22) can be written as Q = T dS and the work that the system is doing on the surroundings is W = P dV Substituting equations (2.25) (2.26) into (2.20) results in T dS = d EU + P dV (2.27) (2.26) (2.25)
17
Even though the derivation of the above equations were done assuming that there is no change of kinetic or potential energy, it still remains valid for all situations. Furthermore, it can be shown that it is valid for reversible and irreversible processes. Enthalpy It is a common practice to dene a new property, which is the combination of already dened properties, the enthalpy of the system. H = EU + P V The specic enthalpy is enthalpy per unit mass and denoted as, h. Or in a dierential form as dH = dEU + dP V + P dV Combining equations (2.28) the (2.27) yields T dS = dH V dP (2.30) (2.29) (2.28)
For isentropic process, equation (2.27) is reduced to dH = V dP . The equation (2.27) in mass unit is T ds = du + P dv = dh dP (2.31)
when the density enters through the relationship of = 1/v. Specic Heats The change of internal energy and enthalpy requires new denitions. The rst change of the internal energy and it is dened as the following Eu (2.32) T And since the change of the enthalpy involve some kind of work, it is dened as Cv
Cp
h T
(2.33)
The ratio between the specic pressure heat and the specic volume heat is called the ratio of the specic heats and it is denoted as, k. k Cp Cv (2.34)
18
For liquid metal used in die casting, the ratio of the specic heats is bite higher than one (1) and therefore the dierence between them is almost zero and therefore referred as C. Equation of state Equation of state is a relation between state variables. Normally the relationship of temperature, pressure, and specic volume dene the equation of state for gases. The simplest equation of state referred to as ideal gas and it is dened as P = RT (2.35)
Application of Avogadros law, that all gases at the same pressures and temperatures have the same number of molecules per unit of volume, allows the calculation of a universal gas constant. This constant to match the standard units results in R = 8.3145 kj kmol K (2.36)
The specic constants for select gas at 300K is provided in table 2.1. From equation (2.35) of state for perfect gas it follows d(P v) = RdT For perfect gas dh = dEu + d(P v) = dEu + d(RT ) = f (T ) (only) From the denition of enthalpy it follows that d(P v) = dh dEu (2.40) (2.39) (2.38)
Utilizing equation (2.38) and substituting into equation (2.40) and dividing by dT yields Cp Cv = R This relationship is valid only for ideal/perfect gases. The ratio of the specic heats can be expressed in several forms as R k1 (2.41)
Cv =
(2.42)
19
Gas
Molecular Weight 28.970 39.948 58.124 44.01 28.01 30.07 28.054 4.003 2.016 16.04 20.183 28.013 114.230 31.999 44.097 18.015
kj KgK
Cv
kj KgK
CP
kj KgK
Air Argon Butane Carbon Dioxide Carbon Monoxide Ethane Ethylene Helium Hydrogen Methane Neon Nitrogen Octane Oxygen Propane Steam
0.28700 0.20813 0.14304 0.18892 0.29683 0.27650 0.29637 2.07703 4.12418 0.51835 0.41195 0.29680 0.07279 0.25983 0.18855 0.48152
1.0035 0.5203 1.7164 0.8418 1.0413 1.7662 1.5482 5.1926 14.2091 2.2537 1.0299 1.0416 1.7113 0.9216 1.6794 1.8723
0.7165 0.3122 1.5734 0.6529 0.7445 1.4897 1.2518 3.1156 10.0849 1.7354 0.6179 0.7448 1.6385 0.6618 1.4909 1.4108
1.400 1.400 1.091 1.289 1.400 1.186 1.237 1.667 1.409 1.299 1.667 1.400 1.044 1.393 1.327 1.327
Cp =
kR k1
(2.43)
The specic heats ratio, k value ranges from unity to about 1.667. These values depend on the molecular degrees of freedom (more explanation can be obtained in Van Wylen F. of Classical thermodynamics.) The values of several gases can be approximated as ideal gas and are provided in Table (2.1). The entropy for ideal gas can be simplied as the following
2
s2 s1 =
1
dh dP T T
(2.44)
20
s2 s1 =
1
Cp
dT T
2 1
R dP T2 P2 = Cp ln R ln P T1 P1
(2.45)
Or using specic heats ratio equation (2.45) transformed into s2 s1 k T2 P2 = ln ln R k 1 T1 P1 For isentropic process, s = 0, the following is obtained T2 ln = ln T1 P2 P1
k1 k
(2.46)
(2.47)
There are several famous identities that results from equation (2.47) as
T2 = T1
P2 P1
k1 k
P2 P1
k1
(2.48)
The ideal gas model is a simplied version of the real behavior of real gas. The real gas has a correction factor to account for the deviations from the ideal gas model. This correction factor is referred to as the compressibility factor and dened as Z= PV RT (2.49)
Control Volume The control volume was introduced by L. Euler1 In the control volume (c.v) the focus is on specic volume which mass can enter and leave. The simplest c.v. is when the boundaries are xed and it is referred to as the Nondeformable c.v.. The conservation of mass to such system can be reasonably approximated by d dt dV =
Vc.v. Sc.v.
Vrn dA
(2.50)
This equation states the change in the volume came from the dierence of masses being added through the boundary. put two examples of simple for mass conservation. For deformable c.v. d dt dV =
Vc.v. Vc.v.
d dV + dt
Vrn dA
Sc.v.
(2.51)
1 A blind man known as the master of calculus, made his living by being a tutor, can you imagine he had eleven kids: where he had the time and energy to develop all the great theory and mathematics.
21
2.3.3
Momentum Equation
The second Newton law of motion is written mathematically as D mV Dt This explanation, of course, for uid particles can be written as F = F = D Dt V dV
Vsys
(2.52)
(2.53)
V Vrn dA
(2.54)
2.3.4
Compressible ow
This material is extensive and requires a semester for student to have good undersound wave standing of this complex material. Yet dU velocity=dU c to give very minimal information is seems P+dP P +d to to be essential to the understanding of the venting design. The summary material here is derived from the book Fundamentals of Compressible Flow Mechanics. Fig. -2.3. A very slow moving piston in a still
gas.
2.3.5
Speed of Sound
Control volume around the sound wave
The speed of sound is a very important parameter in the die casting process because c-dU c it eects and explains the choking in the P+dP die casting process. What is the speed of P +d the small disturbance +as it travels in a quiet medium? This velocity is referred to as the speed of sound. To answer this Fig. -2.4. Stationary sound wave and gas moves question, consider a piston moving from relative to the pulse. the left to the right at a relatively small velocity (see Figure 2.3). The information that the piston is moving passes thorough a single pressure pulse. It is assumed that if the velocity of the piston is innitesimally small, the pulse will be innitesimally small. Thus, the pressure and density can be assumed to be continuous. It is convenient to look at a control volume which is attached to a pressure pulse. Applying the mass balance yields c = ( + d)(c dU ) (2.55)
22
From the energy equation (Bernoullis equation), assuming isentropic ow and neglecting the gravity results (c dU )2 c2 dP + =0 2 neglecting second term (dU 2 ) yield cdU + dP =0 (2.58) (2.57)
Substituting the expression for dU from equation (2.56) into equation (2.58) yields d dP dP = = c2 = (2.59) d It is shown in the book Fundamentals of Compressible Fluid Mechanics that relationship between n, Z and k is c2
k
Cp n= Cv
z+T z+T
z T z T P
(2.60)
Note that n approaches k when z 1 and when z is constant. The speed of sound for a real gas can be obtained in similar manner as for an ideal gas dP = nzRT d (2.61)
Speed of Sound in Almost Incompressible Liquid Even liquid metal normally is assumed to be incompressible but in reality it has a small and important compressible aspect. The ratio of the change in the fractional volume to pressure or compression is referred to as the bulk modulus of the material. The mathematical denition of bulk modulus is as follows dP B= (2.62) d In physical terms it can be written as c= elastic property = inertial property B (2.63)
In summary, the speed of sound in liquid metals is about 5 times faster than the speed of sound in gases in the chamber.
23
2.3.6
Choked Flow
In this section a discussion on a steady state ow through a smooth and continuous area ow rate is presented which include the ow through a converging diverging nozzle. The isentropic ow models are important because of two main reasons: Stagnation State for Ideal Gas Model
Fig. -2.5.
PB = P0 P P0
Subsonic M <1
Supersonic
M >1
distance, x
It is assumed that the ow is one (gas) through a convergingdiverging nozzle. dimensional. Figure (2.5) describes a gas ow through a convergingdiverging nozzle. It has been found that a theoretical state known as the stagnation state is very useful in which the ow is brought into a complete motionless condition in isentropic process without other forces (e.g. gravity force). Several properties can be represented by this theoretical process which include temperature, pressure, and density etc and denoted by the subscript 0. A dimensionless velocity and it is referred as Mach number for the ratio of velocity to speed of sound as M The temperature ratio reads T0 k1 2 =1+ M T 2 (2.65) U c (2.64)
The ratio of stagnation pressure to the static pressure can be expressed as the function of the temperature ratio because of the isentropic relationship as P0 = P T0 T
k k1
k1 2 1+ M 2
k k1
(2.66)
k1 2 1+ M 2
1 k1
(2.67)
A new useful denition is introduced for the case when M = 1 and denoted by superscript . The special case of ratio of the star values to stagnation values are dependent only on the heat ratio as the following: c 2 2 T = 2 = T0 c0 k+1 (2.68)
24 and P = P0
2 k+1
k k1
(2.69)
= 0
2 k+1
1 k1
(2.70)
Fig. -2.6. The stagnation properties as a function of the Mach number, k=1.4
The denition of the star Mach is ratio of the velocity and star speed of sound at M = 1. The ow in a convergingdiverging nozzle has two models: First is isentropic and adiabatic model. Second is isentropic and isothermal model. Clearly, the stagnation temperature, T0 , is constant through the adiabatic ow because there isnt heat transfer. Therefore, the stagnation pressure is also constant through the ow because of the isentropic ow. Conversely, in mathematical terms, equation (2.65) and equation (2.66) are the same. If the right hand side is constant for one variable, it is constant for the other. In the same argument, the stagnation density is constant through the ow. Thus, knowing the Mach number or the temperature will provide all that is needed to nd the
25
other properties. The only properties that need to be connected are the cross section area and the Mach number. Examination of the relation between properties can then be carried out. The Properties in the Adiabatic Nozzle When there is no external work and heat transfer, the energy equation, reads dh + U dU = 0 (2.71)
Dierentiation of continuity equation, AU = m = constant, and dividing by the continuity equation reads d dA dU + + =0 A U The thermodynamic relationship between the properties can be expressed as T ds = dh dP (2.73) (2.72)
For isentropic process ds 0 and combining equations (2.71) with (2.73) yields dP + U dU = 0 (2.74)
Dierentiation of the equation state (perfect gas), P = RT , and dividing the results by the equation of state (RT ) yields dP d dT = + P T (2.75)
Obtaining an expression for dU/U from the mass balance equation (2.72) and using it in equation (2.74) reads
dU U
dP dA d U2 + =0 A
(2.76)
Rearranging equation (2.76) so that the density, , can be replaced by the static pressure, dP/ yields dP = U2 dA d dP + A dP
1 c2
(2.77)
dA d dP = U2 + dP A
26
Recalling that dP/d = c2 and substitute the speed of sound into equation (2.77) to obtain dP Or in a dimensionless form dP dA 1 M2 = U2 A (2.79) 1 U c
2
= U2
dA A
(2.78)
Equation (2.79) is a dierential equation for the pressure as a function of the cross section area. It is convenient to rearrange equation (2.79) to obtain a variables separation form of dP = U 2 dA A 1 M2 (2.80)
Before going further in the mathematical derivation it is worth while to look at the physical meaning of equation (2.80). The term U 2 /A is always positive (because all the three terms can be only positive). Now, it can be observed that dP can be positive or negative depending on the dA and Mach number. The meaning of the sign change for the pressure dierential is that the pressure can increase or decrease. It can be observed that the critical Mach number is one. If the Mach number is larger than one than dP has opposite sign of dA. If Mach number is smaller than one dP and dA have the same sign. For the subsonic branch M < 1 the term 1/(1 M 2 ) is positive hence dA > 0 = dP > 0 dA < 0 = dP < 0 From these observations the trends are similar to those in incompressible uid. An increase in area results in an increase of the static pressure (converting the dynamic pressure to a static pressure). Conversely, if the area decreases (as a function of x) the pressure decreases. Note that the pressure decrease is larger in compressible ow compared to incompressible ow. For the supersonic branch M > 1, the phenomenon is dierent. For M > 1 the term 1/1 M 2 is negative and change the character of the equation. dA > 0 dP < 0 dA < 0 dP > 0 This behavior is opposite to incompressible ow behavior. For the special case of M = 1 (sonic ow) the value of the term 1 M 2 = 0 thus mathematically dP or dA = 0. Since physically dP can increase only in a nite amount it must be that dA = 0.It must also be noted that when M = 1 occurs only when dA = 0. However, the opposite, not necessarily means that when dA = 0 that M = 1. In that case, it is possible that dM = 0 thus the diverging side is in the subsonic branch and the ow isnt choked.
27
In this section, the other extreme case model where the heat transfer to the gas is perfect, (e.g. Eckert number combination is very small) is presented. Again in reality the heat transfer is somewhere in between the two extremes. So, knowing the two limits provides a tool to examine where the reality should be expected. The perfect gas model is again assumed. In isothermal process the perfect gas model reads P = RT dP = dRT (2.81)
Integration of equation (2.82) yields the Bernoullis equation for ideal gas in isothermal process which reads U2 2 U1 2 P2 + RT ln =0 2 P1 Then the stagnation velocity is U= 2RT ln P P0 (2.84) (2.83)
eM1 2 eM2
k 2
(2.85)
As opposed to the adiabatic case (T0 = constant) in the isothermal ow the stagnation temperature ratio can be expressed 1 ! T01 T1 1 + = T02 T 2 1+
2 k1 2 M1 2 k1 2 M2
1+ 1+
2 k1 2 M1 2 k1 2 M2
(2.86)
eM2 eM1 2
k 2
(2.87)
2 The one dimensional momentum equation for steady state is U dU/dx = dP/dx+0(other eects) which are neglected here.
28
eM1 = eM1 2
k 2
(2.88)
The critical point, at this stage, is unknown (at what Mach number the nozzle is choked is unknown) so there are two possibilities: the choking point or M = 1 to normalize the equation. Here the critical point dened as the point whereM = 1 so results can be compared to the adiabatic case and denoted by star. Again it has to be emphasized that this critical point is not really related to physical critical point but it is only an arbitrary denition. The true critical point is when ow is choked and the relationship between two will be presented. The critical pressure ratio can be obtained from (2.85) to read
(1M 2 )k P = =e 2 P
(2.89)
Isothermal Nozzle
k=14 4 3.5 P/P * A/A * P0 / P0 T 0 / T0 2.5 2 T/T
* * *
(2.90)
1.5 1 0.5 0
0.5
1.5
2 M
2.5
3.5
(2.91)
Finally, the critical stagnation pressure Fig. -2.7. Various ratios as a function of Mach number for isothermal Nozzle reads
(1M 2 )k P0 2 =e P0
2 1 + k1 M1 2 2 k+1
k k1
(2.92)
2 k+1
k k1
(2.93)
For specic heats ratio of k = 1.4, this maximum value is about two. It can be noted that the stagnation pressure is monotonically reduced during this process. Of course in isothermal process T = T . All these equations are plotted in Figure (2.7). From the Figure 2.7 it can be observed that minimum of the curve A/A isnt on M = 1. The minimum of the curve is when area is minimum and at the point where
29
the ow is choked. It should be noted that the stagnation temperature is not constant as in the adiabatic case and the critical point is the only one constant. The mathematical procedure to nd the minimum is simply taking the derivative and equating to zero as the following d
A A
kM 2 e
k(M 2 1) 2
k(M 2 1) 2
M2
=0
(2.94)
kM 2 1 = 0
1 M= k
(2.95)
30
The shear, S, at the ingate is determined by the average velocity, U, of the liquid and by the ingate thickness, t. Dimensional analysis shows that is directly proportional to (U/ ). The constant of proportionality is dicult to determine, . . .1 Murray, CSIRO Australia
One of the important tools to understand the die casting process is dimensional analysis. Fifty years ago, this method transformed the uid mechanics/heat transfer into a uniform understanding. This book attempts to introduce to the die casting industry this established method2 . Experimental studies will be expanded/generalized as it was done in convective heat transfer. It is hoped that as a result, separate sections for aluminum, zinc, and magnesium will not exist anymore in die casting conferences. This chapter is based partially on Dr. Eckerts book, notes, and the article on dimensional analysis applied to die casting. Several conclusions are derived from this analysis and they will be presented throughout this chapter. This material can bring great benet to researchers who want to built their research on a solid foundation. For those who are dealing with the numerical research/calculation, it is useful to learn when some parameters should be taken into account and why.
1 Citing The Design of feed systems for thin walled zinc high pressure die castings, Metallurgical and materials transactions B Vol. 27B, February 1996, pp. 115118. This excerpt is an excellent example of poor research and poor understanding. This unknown constant is called viscosity (see Basics of Fluid Mechanics in Potto series. Here, a discussion on some specic mistakes were presented in that paper (which are numerous). Dimensional analysis is a tool which can take cluttered and meaningless paper such as the above and turn them into something with real value. As proof of their model, the researchers have mentioned two unknown companies that their model is working. What a nice proof! Are the physics laws really dierent in Australia? 2 Actually, Prof. E.R.G. Eckert introduced the dimensional analysis to the die casting long before. The author is his zealous disciple, all the credit should go to Eckert. Of course, all the mistakes are the authors and none of Dr. Eckerts. All the typos in Eckerts paper were this authors responsibility for which he apologizes.
31
32
3.0.7
In dimensional analysis, the number of the eecting parameters is reduced to a minimum by replacing the dimensional parameters by dimensionless parameters. Some researchers point out that the chief advantage of this analysis is to obtain experimental results with a minimum amount of labor, results in a form having maximum utility [18, pp. 395]. The dimensional analysis has several other advantages which include; 1)increase of understanding, 2) knowing what is important, and 3) compacting the presentation3 . The advantage of compact of presentation allows one to see the big picture with minimal eort. Dimensionless parameters are parameters which represent a ratio which does not have a physical dimension. The experimental study assists to solve problems when the solution of the governing equation cannot be obtained. To achieve this, experiments are designed to be similar to the situations which need to be solved or simulated. The base for this concept is mathematical. Two dierent sets of phenomena will produce a similar result if the governing dierential equations with boundaries conditions are similar. The actual experiments are dicult to carry out in many cases. Thus, design experiments with the same governing dierential equations as the actual phenomenon is the solution. This similarity does not necessarily mean that the experiments have to be carried exactly as studied phenomena. It is enough that the main dimensionless parameters are similar, since the minor dimensional parameters, in many cases, are insignicant. For example, a change in Reynolds number is insignicant since a change in Reynolds number in a large range does not aect the friction factor. An example of the similarity applied to the die cavity is given in the section 3.5. Researchers in casting in general and die casting in particular do not utilize this method. For example, after the Russians [6] introduced the water analogy method (in casting) in the 40s all the experiments such as Wallace, CSIRO, etc. conducted poorly designed experiments. For example, Wallace record the Reynolds and Froude number without attempting to match the governing equations. Another example is the experimental study of Gravity Tiled Die Casting (low pressure die casting) performed by Nguyens group in 1986 comparing two parameters and We. Flow of free Re falling, the velocity is a function of the height (U gH). Hence, the equation Remodel = Reactual should lead only to Hmodel Hactual and not to any function of Umodel /Uactual . The value of Umodel /Uactual is actually constant for the same height ratio. The Wallace experiments with Reynolds number matching does not lead to matching of similar governing equations. Many other important parameters which control the governing equations are not simulated [26]. The governing equations in these cases include several other important parameters which have not been controlled
3 The importance of compact presentation is attributed to Prof. M. Bentwitch who was mentor to many including the author during his masters studies.
3.1. INTRODUCTION
33
or even measured, monitored, and simulated4 . Moreover, the Re number is controlled by the ow rate and the characteristics of the ladle opening and not as in the pressurized pipe ow as the authors assumed.
3.1 Introduction
Lets take a trivial example of tting a rode into a circular hole (see Figure 3.1). To solve this problem, it is required to know two parameters; 1) the diameter of the rode and 2) the diameter of the hole. Actually, it is required to have only one parameter, the ratio of the rode diameter to the hole diameter. The ratio is a dimensionless number and with this Fig. -3.1. Rod into the hole example number one can say that for a ratio larger than one, the rode will not enter the hole; and ratio smaller than one, the rod is too small. Only when the ratio is equal to one, the rode is said to be t. This allows one to draw the situation by using only one coordinate. Furthermore, if one wants to deal with tolerances, the dimensional analysis can easily be extended to say that when the ratio is equal from 0.99 to 1.0 the rode is tting, and etc. If one were to use the two diameters description, he will need more than this simple sentence to describe it. In the preceding simplistic example, the advantages are minimal. In many real problems, including the die casting process, this approach can remove clattered views and put the problem into focus. It also helps to use information from dierent problems to a similar situation. Throughout this book the reader will notice that the systems/equations are converted to a dimensionless form to augment understanding.
34
on each process. On one hand, in processes 1 to 3, it is desirable to have a minimum heat transfer/solidication to take place for obvious reasons. On the other hand, in the rest of the processes, the solidication is the major concern. In die casting, the information and conditions do not travel upstream. For example, the turbulence does not travel from some point at the cavity to the runner and ofcourse, to the shot sleeve. This kind of relationship is customarily denoted as a parabolic process (because in mathematics the dierential equations describe these kind of cases as parabolic). To a larger extent it is true in die casting. The pressure in the cavity does not aect the ow in the sleeve or the runner if the vent system is well designed. In other words, the design of the pQ2 diagram is not controlled by downstream conditions. Another example, the critical slow plunger velocity is not aected by the air/gas ow/pressure in the cavity. In general, the turbulence generated downstream does not travel upstream in this process. One has to restrict this characterization to some points. One point is particularly mentioned here: The poor design of the vent system aects the pressure in the cavity and therefore the eects do travel down stream. For example, the pQ2 diagram calculations are aected by poor vent system design.
3.2.1
The ow from the ladle to the shot sleeve did not receive much attention in the die casting research5 because it is believed that it does not play a signicant role. For low pressure die casting, the ow of liquid metal from the ladle through channel(s) to the die cavity plays an important role6 . The importance of the understanding of this process can show us how to minimize the heat transfer, layer created on the sleeve (solidication layer), and sleeve protection from; a) erosion b) plunger problem. The jet itself has no smooth surface and two kinds of Fig. -3.2. Hydraulic jump in the instability occurs. The rst instability is of Bernoullis shot sleeve. eect and second eect is Bar-Meirs eect that boundary conditions cannot be satised for two phase ow. Yet, for die casting process, these two eects (see Figure 3.2 do not change the global ow in the sleeve. At rst, the hydraulic jump is created when the liquid metal enters the sleeve. The typical time scale for hydraulic jump creation is almost instant and extremely short as can be shown by the characteristic methods. As the liquid metal level in the sleeve rises, the location of the jump moves closer to the impinging center. At a certain point, the liquid depth level is over the critical depth level and the hydraulic jump disappears. The critical depends on the liquid properties and the ratio of impinging momentum or velocity to the hydraulic static pressure. The impinging momentum impact is proportional to U 2 r2 and hydraulic pressure is proportional
Bar-Meirs instability Hydraulic Jump
few papers ( 0) can be found dealing with this aspect. elementary estimates of uid mechanics and heat transfer were made by the author and hopefully will be added to this book.
6 Some
5 Very
35
to g h 2 r h. Where r is the radius of the impinging jet and h is the height of the liquid metal in the sleeve. The above statement leads to Ucritical g h2 r (3.1)
where hL is the distance of the ladle to the height of the liquid metal in the sleeve. The height where the hydraulic shock will not exist is hcritical r hL (3.3)
This analysis suggests that decreasing the ladle height and/or reducing less mass ow rate (the radius of the jet) result in small critical height. The air entrainment during that time will be discussed in the book Basic of Fluid Mechanics in the MultiPhase ow chapter. At this stage, air bubbles are entrained in the liquid metal which augment the heat transfer. At present, there is an extremely limited knowledge about the heat transfer during this part of the process, and of course less about how to minimize it. However, this analysis suggests that minimizing the ladle height is one of the ways to reduce it. The heat transfer from liquid metal to the surroundings is aected by the velocity and the ow patterns since the initial stage mechanism of heat transfer is changed from a dominated natural convection to a later H <H> stage dominated force convection. In addition, the liquid metal jet surface is also aected by heat transfer to some degree by change in the properties.
hydralic jump
bubles
air entrainment
Heat Transferred to the Jet The estimate on heat transfer requires some information on jet dynamics. Fig. -3.3. Filling of the shot sleeve. There are two eects that must be addressed; one the average radius and the uctuation of the radius. As rst approximation, the average jet radius changes due to the velocity change. For laminar ow, (for simplicity assume plug ow) the veloc ity function is x where x is the distance from the ladle. For constant ow rate, neglecting the change of density, the radius will change as r 1/ 4 x. Note that this relationship is not valid when it is very near the ladle proximity (r/x 0). The heat transfer increases as a function of x for these two reasons. The second eect is jet radius uctuations. Consider this, the jet leaves the ladle in a plug ow. Due to air friction, the shear stress changes the velocity prole to
36
parabolic. For simple assumption of steady state(it is not steady state), the momentum equation which governs the liquid metal is assume uz t
0
constant uz = 1 + ur r r r
uz r
(3.4)
Equation (3.4) is in simplied equation form for the gas and liquid phases. Thus, there are two equations that needs to be satised simultaneously; one for the gas side and one for the liquid side. Even neglecting several terms for this discussion, it clear that both equations are second order dierential equations which have dierent boundary conditions. Any second order dierential equation requires two dierent boundary conditions. Requirement to satisfy additional boundary condition can be achieved. Thus from physical point of view, second order dierential equation which needs to satisfy three boundary conditions is not possible, Thus there must be some wrong either with the governing equation or with the boundary conditions. In this case, the two governing equations must satisfy ve (5) dierent boundary conditions. These boundary conditions are as follows: 1) summitry at r = 0, 2) identical liquid metal and air velocities at the interface, 3) identical shear stress at the interface, 4) zero velocity at innity for the air, and 5) zero shear stress for the air at the innity. These requirements cannot be satised if the interface between the liquid metal and the gas is a straight line. The heat transfer to the sleeve in the impinging area is signicant but at present only very limited knowledge is available due to complexity.
3.2.2
Fluid Mechanics The main point is the estimate for Shot energy dissipation. The dissipaheat transfer Sleeve to the air 2 tion is proportional to < U > process 1 L. Where the strange velocity, < U > is averaged kinetic velocity provided by jet. This kinetic energy is at most the same as potential energy of liquid metal in heat transfer the ladle. The potential energy in to the sleeve Solidification process 2 the ladle is < H > m g where layer < H > is averaged height see Figure 3.3. The averaged velocity Fig. -3.4. Heat transfer processes in the shot sleeve. in the shot sleeve is 2 g < H > 2 The rate energy dissipation can be estimated as <U > R L as L is the length of R shot sleeve. The shear stress is assumed to occur equally in the volume of liquid metal in the sleeve. This assumption of shear stress grossly under estimates the dissipation.
Temperature
37
The actual dissipation is larger due to the larger velocity gradients. The estimated time is then Heat Transfer In this section, the solidication eects are examined. One of the assumptions in the analysis of the critical slow plunger velocity is that the solidication process does not play an important role (see Figure 3.4). The typical time for heat to penetrate a typical layer in air/gas phase is in the order of minutes. Moreover, the density of the air/gas is 3 order magnitude smaller than liquid metal. Hence, most of the resistance to heat transfer is in the gas phase. Additionally, it has been shown that the liquid metal surface is continuously replaced by slabs of material below the surface which is known in scientic literature as the renewal surface theory. Thus, the main heat transfer mechanism is through the liquid metal to the sleeve. The heat transfer rate for a very thin solidied layer can be approximated as Q klm T r L Ls r L t r (3.5)
Where Ls is the latent heat, klm is the thermal conductivity of liquid metal and t is the thickness of the solidication layer. Equation (3.6) results in t klm T r Ls r 2 (3.6)
The value for this die casting process in minutes is in the range of 0.01-0.001 after the thickness reaches to 1-2 [mm]. The relative thickness liquid metal further decreases as the inverse of the square solidied layer increases. If the solidication is less than shot sleeve (steel) one percent of the radius, the speed will be very small compared to the speed of the plunger. If the insulation solidication occur as a mushy zone then the heat transfer is reduced further and it is even lower than Fig. -3.5. Solidication of the shot this estimate and R 1). Therefore, the heat sleeve time estimates. transfer from the liquid metal surface to the air, as shown in Figure 3.4 (mark as process 1), acts as an insulator to the liquid metal. The governing equation in the sleeve is
y
(liquid) (solid) Temperature
d cp d
T = kd t
2T y 2
(3.7)
where the subscript d denotes the properties of the sleeve material. Boundary condition between the sleeve and the air/gas is T n =0
y=0
(3.8)
38
Where n represents the perpendicular direction to the die. Boundary conditions between the liquid metal (solid) and sleeve ksteel T y = kAL
y=l
T y
(3.9)
y=l
The governing equation for the liquid metal (solid phase) lm cp lm T = klm t 2T y 2 (3.10)
where lm denotes the properties of the liquid metal. The dissipation and the velocity are neglected due to the change of density and natural convection. Boundary condition between the phases of the liquid metal is given by vs s hsf = kl hsf s vn k T y ks
y=l+
T y
k
y=l+
(Tl Ts ) y
(3.11)
y=l+
the heat of solidication liquid metal density at the solid phase velocity of the liquid/solid interface conductivity
Neglecting the natural convection and density change, the governing equation in the liquid phase is l cp l T 2T = kl 2 t y (3.12)
The dissipation function can be assumed to be negligible in this case. There are three dierent periods in heat transfer; 1. lling the shot sleeve 2. during the quieting time, and 3. during the plunger movement. In the rst period, heat transfer is relatively very large (major solidication). At present, there is not much known about the uid mechanics not to say much about the solidication process/heat transfer in uid mechanics. The second period can be simplied and analyzed as known initial velocity prole. A simplied assumption can be made considering the fact that P r number is very small (large thermal boundary layer compared to uid mechanics boundary layer). Additionally, it can be assumed that the natural convection eects are marginal. In the last period, the heat transfer is composed from two zones: 1) behind the jump and 2) ahead of the jump. The heat transfer ahead of the jump is the same as in the second period; while the heat transfer behind the jump is like heat transfer into a plug ow for low P r number. The heat transfer in such cases have been studied in the past7 .
7 The
reader can refer, for example, to the book Heat and Mass Transfer by Eckert and Drake.
39
3.2.3
Runner system
The ow in the runner system has to be divided into sections; 1) ow with free surface 2) lling the cavity liquid when the ow is pressurized (see Figmetal ures 3.6 and 3.7). In the rst section the gravity aects the air entrainment. The dominant parameters in this case air streaks are Weber number, W e and Reynolds number, Re. This phenomenon determines how much metal has to be ushed out. It is well known that the Fig. -3.6. Entrance of liquid metal to the runner. liquid interface cannot be a straight line. Above certain velocity (typical to die casting, high Re number) air leaves streaks of air/gas slabs behind the front line as shown in Figure 3.6. These streaks create a low heat transfer zone at the head of the jet and increases its velocity. The air entrainment created in this case is supposed to be ushed out through the vent system in a proper process design. Unfortunately, at present very little is known about this issue especially the geometry typical to die casting. Gravity Limited in Runner system In the second phase, the ow in the runner system is pressurized. The typical velocity is large of the range of 10-15 [m/sec]. The typical runner length is in order of 0.1[m]. The velocity due to gravity is 2.5[m/sec]. The F r number assumes the value 102 for which gravity play a limited role. The converging nozzle such as the transition into runner system (which a good die casting engineer should design) tends to reduce the turbulence, if turbulence exists, and can even eliminate it. In that view, the liquid metal enters the runner system as a laminar ow (actually close to a plug ow). For a duct with a typical dimension of 10 [mm] and a mean Fig. -3.7. Flow in runner when during pressurvelocity, U = 10[m/sec], (during the sec- izing process. ond stage), for aluminum die casting, the Reynolds number is: Ub Re = 5 107 which is a supercritical ow. However, the ow is probably laminar ow due to the short time. Another look at turbulence issue: The boundary layer is a function of the time (during the lling period) is of order
pressure
= 12t
40
The boundary layer in this case can be estimated as8 the time of the rst phase. Anyhow, utilizing the time of 0.01[sec] the viscosity of aluminum in the boundary layer is of the thickness of 0.25[mm] which indicates that ow is laminar.
3.2.4
Die Cavity
All the numerical simulations of die lling are done almost exclusively by assuming that the ow is turbulent and continuous (no two phase ow). In the section 3.3.1 a question about the question whether existence of turbulence is discussed and if so what kind of model is appropriate. Thus, the validity of these numerical models is examined. The liquid metal enters the cavity as a noncontinuous ow. According to some researchers, it is preferred that the ow will be atomized (spray). While there is a considerable literature about many geometries none available to typical die casting congurations9 . The ow can be atomized as either in laminar or turbulent region. The experiments by the author and by others, showed that the ow turns into spray in many cases ( See Figures 3.8).
Fig a.
Flow as a jet.
Fig b.
Flow as a spray.
Fig. -3.8. Typical ow pattern in die casting, jet entering into empty cavity.
In the section 3.4.1 it was shown that the time for atomization is very fast compared with any other process (lling time scale and, of course, the conduction heat transfer or solidication time scales). Atomization requires two streams with a signicant velocity dierence; stronger surface tension forces against the maintaining stability forces. Numerous experimental studies have shown that better castings are obtained when the injected velocity is above a certain value. This fact alone is enough to convince researchers that the preferred ow pattern is a spray ow. Yet, only a very small number of numerical models exist assuming spray ow and are used for die casting (for example, the paper by Hu at el [22].). Experimental work commonly cited as a proof
during the ow in the runner system, no lling of the cavity can just wonder who were the opposition to this research? Perhaps one of the referees as in the Appendix B for the all clues that have been received.
9 One 8 only
41
of turbulence was conducted in the mid 60s [30] utilizing water analogy10 . The white spats they observed in their experiments are atomization of the water. Because these experiments were poorly conducted (no similarity to die casting process) the observation/information from these studies is very limited. Yet with this limitation in mind, one can conclude that the spray ow does exist. Experiments by Fondse et al [16] show that atomization is larger in laminar ow compared to a turbulent ow in a certain range. This fact further creates confusion of what is the critical velocity needed in die casting. Since the experiments which measure the critical velocity were poorly conducted, no reliable information is available on what is the ow pattern and what is the critical velocity11 .
3.2.5
Intensication Period
The two main concerns in this phase is to extract heat from the die and to solidify the liquid metal as aptly as possible to obtain the nal shape. Thus, two operational parameters are important; one the (minimum) time for the intensication and two the pressure of the intensication (the clamping force). These two operational parameters can improve casting design to obtain good product. The main resistance to the heat ow is in the die and the cooling liquid (oil or water based solution). In some parts of the process, the heat is transformed to the cooling liquid via the boiling mechanism. However, the characteristic of boiling heat transfer time to achieve a steady state is larger than the whole process and the typical equations (steady state) for the preferred situation (heat transfer only in the rst mode) are not accurate. When there is very limited understanding of so many aspects of the process, the eects of each process on other processes are also cluttered.
Is the ow in the shot sleeve turbulent as the EKK sale engineers claim? These sale engineers did not present any evidence or analysis for such claims. For a simple anal ysis, the initial part of the shot sleeve lling, the liquid metal goes through a hy draulic jump. The ow after the hydraulic jump is very slow because the increase of the ratio of cross section areas. For example, casting of the 1[kg] from height Time [sec] of 0.2[m] to a shot sleeve of 0.1[m] cre ates a velocity in shot sleeve of 2[m/sec] Fig. -3.9. Transition to turbulent ow in circuwhich results after the hydraulic jump to lar pipe for instantaneous ow after Wygnanski be with velocity about 0.01[m/sec]. The and others by interpolation. Reynolds number for this velocity is 104 and Froude number of about 10. After the jump the Froude number is reduced and the ow is turbulent. However, by the time the hydraulic jump vanishes, the ow turns into laminar ow and no change (waviness) in the surface can be observed. It can be noticed that the time scale for the dissipation is about the same scale as the time for the operation of the next stage. Figure 3.9 exhibits the transition to a turbulent ow for instantaneous starting ow in a circular pipe. The abscissa represents time and the yaxis represents the Re number at which transition to turbulence occurs. The points on the graphs show the transition to a turbulence. This gure demonstrates that a large time is required to turn the ow pattern to turbulent which is measured in several seconds. Fig. -3.10. Flow pattern in the shot sleeve. The gure demonstrates that the transition does not occur below a certain critical Re number (known as the critical Re number for steady state). It also shows that a considerable time has elapsed before transition to turbulence occurs even for a relatively large Reynolds number. The geometry in die casting however is dierent and therefore it is expected that the transition occurs at dierent times. Our present knowledge of this area is very limited. Yet, a similar transition delay is expected to occur after the instantaneous startup which probably will be measured in seconds. The ow in die casting in many situations is very short (in order of milliseconds) and therefore it is expected that the transition to a turbulent ow does not occur. After the liquid metal is poured, it is normally repose for sometime in a range of 10 seconds. This fact is known in the scientic literature as the quieting time for which the existed turbulence (if exist) is reduced and after enough time (measured in seconds)
12.00 11.00 10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00
Re10^3
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
Transition zone
43
is illuminated. Hence, the turbulence, which was created during the lling process of the shot sleeve, disappear due to viscous dissipation. The question is, whether the ow in the duration of the slow plunger velocity turbulent (see Figure 3.10) can be examined. Clearly, the ow in the substrate (a head of the wave) is still (almost zero velocity) and therefore the turbulence does not exist. The Re number behind the wave is above the critical Re number (which is in the range of 20003000). The typical time for the wave to travel to the end of the shot sleeve is in the range of a 100 second. At present there are no experiments on the ow behind the wave12 . The estimation can be done by looking at what is known in the literature about the transition to turbulence in instantaneous starting pipe ow. It has been shown [32] that the ow changes from laminar ow to turbulent ow in an abrupt manner for a ow with supercritical Re number. A typical velocity of the propagating front (transition between laminar to turbulent) is about the same velocity as the mean velocity of the ow. Hence, it is reasonable to assume that the turbulence is conned to a small zone in the wave front since the wave is traveling in a faster velocity than the mean velocity. Note that the thickness of the transition layer is a monotone increase function of time (traveling distance). The Re number in the shot sleeve based on the diameter is in a range of 104 which means that the boundary layer has not developed much. Therefore, the ow can be assumed as almost a plug ow with the exception of the front region. A Note on Numerical Simulations The most common model for turbulence that is used in the die casting industry for simulating the ow in cavity is k . This model is based on several assumptions 1. isentropic homogeneous turbulence, 2. constant material properties (or a mild change of the properties), 3. continuous medium (only liquid (or gas), no mixing of the gas, liquid and solid whatsoever), and 4. the dissipation does not play a signicant role (transition to laminar ow). The k model is considered reasonable for the cases where these assumptions are not far from reality. It has been shown, and should be expected, that in cases where assumptions are far from reality, the k model produces erroneous results. Clearly, if we cannot determine whether the ow is turbulent and in what zone, the assumption of isentropic homogeneous turbulence is very questionable. Furthermore, if the change to turbulence just occurred, one cannot expect the turbulence to have sucient time to become isentropic homogeneous. As if this is not enough complication, consider the eects of properties variations as a result of temperature change. Large variations of the properties such as the viscosity have been observed in many alloys especially in the mushy zone.
12 It has to be said that similar situations are found in two phase ow but they are dierent by the fact the ow in two phase ow is a sinusoidal in some respects.
44
While the assumption of the continuous medium is semi reasonable in the shot sleeve and runner, it is far from reality in the die cavity. As discussed previously, the ow is atomized and it is expected to have a large fraction of the air in the liquid metal and conversely some liquid metal drops in the air/gas phase. In such cases, the isentropic homogeneous assumption is very dubious. For these reasons the assumption of k model seems unreasonable unless good experiments can show that the choice of the turbulence model does not matter in the calculation. The question whether the ow in die cavity is turbulent or laminar is secondary. Since the two phase ow eects have to be considered such as atomization, air/gas entrainment etc. to describe the real ow in the cavity. Additional note on numerical simulation The solution of momentum equation for certain situations may lead to unstable solution. Such case is the case of two jets with dierent velocity ow into a medium and they are adjoined (see Figure 3.11). The solution of such ow can show that the velocity eld can be an unstable Fig. -3.11. Two streams of uids into a medium. solution for which the ow moderately changes to become like wave ow. However, in many cases this ow can turn out to be full with vortexes and such. The reason that this happened is the introduction of instabilities. Numerical calculations intrinsically are introducing instabilities because of truncation of the calculations. In many cases, these truncations results in overshooting or undershooting of the nature instability. In cases where the ow is unstable, a careful study is required to make sure that the solution did not produce an unrealistic solution for larger or smaller than reality introduced instabilities. An excellent example of such poor understating is a work made in EKK company [2]. In that work, the ow in the shot sleeve was analyzed. The nature of the ow is two dimensional which can be seen by all the photos taken by numerous people (staring from the 50s). The presenter of that work explained that they have used 3D calculations because they want to study the instabilities perpendicular to the ow direction. The numerical instability in this case is larger than real instabilities and therefore, the numerical results show phenomena does not exist in reality.
U1 U2 U1 U2
Reverse transition from turbulent ow to laminar ow After lling the die cavity, during the solidication process and intensication, the attained turbulence (if exist) is reduced and probably eliminated, i.e. the ow is laminar in a large portion of the solidication process. At present we dont comprehend when the transition point/criteria occurs and we must resort to experiments. It is a hope
45
that some real good experiments using the similarity technique, outlined in this book, will be performed. So more knowledge can be gained and hopefully will appear in this book.
3.3.2
The large velocities of the liquid metal (particularly at the runner) theoretically can increase the liquid metal temperature. To study this phenomenon, compare the of maximum eect of all the kinetic energy that is transformed into thermal energy. U2 = cp T 2 This equation leads to the denition of Eckert number Ec = U2 cp T (3.14) (3.13)
When Ec number is very large it means that the dissipation plays a signicant role and conversely when Ec number is small the dissipation eects are minimal. In die casting, Eckert number, Ec, is very small therefore the thermal dissipation is very small and can be ignored.
3.3.3
Gravity eects
The gravity has a large eect only when the gravity force is large relatively to other forces. A typical velocity range generated by gravity is the same as for an object falling through the air. The air eects can be neglected since the air density is very small compared with liquid metal density. The momentum is the other dominate force in the lling of the cavity. Thus, the ratio of the momentum force to the gravity force, also known as Froude number, determines if the gravity eects are important. The Froude number is dened here as Fr = U2 g (3.15)
Where U is the velocity, is the characteristic length g is the gravity force. For example, the characteristic pouring length is in order of 0.1[m], in extreme cases the velocity can reach 1.6[m] with characteristic time of 0.1[sec]. The author is not aware of experiments to verify the ow pattern in such cases (low P r number due to solidication eect)13 Yet, it is reasonable to assume that the liquid metal in such a case, ow in laminar regimes even though the Re number is relatively large ( 104 ) because of the short time and the short distance. The Re number is dened by the ow rate and the thickness of the exiting typical dimension. Note, the velocity reached its maximum value just before impinging on the sleeve surface.
13 It
46
The gravity has dominate eects on the ow in the shot sleeve since the typical value of the Froude number in that case (especially during the slow plunger velocity period) is in the range of one(1). Clearly, any analysis of the ow has to take into consideration the gravity (see Chapter 8).
lling time
The characteristic time for lling a die cavity is determined by tf L U (3.16)
Where L denotes the characteristic length of the die and U denotes the average lling velocity, determined by the pQ2 diagram, in most practical cases this time typically is in order of 5100 [millisecond]. Note, this time is not the actual lling time but related to it.
Atomization time
The characteristic time for atomization for a low Re number (large viscosity) is given by ta viscosity = (3.17)
where is the kinematic viscosity, is the surface tension, and is the thickness of the gate. The characteristic time for atomization for large Re number is given by ta momentum = 2U (3.18)
The results obtained from these equations are dierent and the actual atomization time in die casting has to be between these two values. Conduction time (die mold)
3.4. ESTIMATES OF THE TIME SCALES IN DIE CASTING The governing equation for the heat transfer for the die reads Td d cp d = kd t 2 Td 2 Td 2 Td + + x2 y 2 z 2
L
47
(3.19)
cooling liquid
To obtain the characteristic time we dimensionlessed the governing equation and present it with a group of constants that determine value of the characteristic time by setting it to unity. Denoting the following variables as
Fig. -3.12. Schematic of heat transfer processes in the die.
td =
t tc d
xd =
x L
yd = L
y L
zd =
z L
d =
T TB TM TB
(3.20)
the characteristic path of the heat transfer from the die inner surface to the cooling channels
subscript B boiling temperature of cooling liquid M liquid metal melting temperature With these denitions, equation (3.19) is transformed to tc d d = d2 t L 2 d 2 d 2 d + + x 2 y 2 z 2 (3.21)
48
where tc s is the characteristic time for conduction process and, , denotes the main path of the heat conduction process die cavity. With these denitions, similarly as was done before the characteristic time is given by
2
tc s
(3.24)
Note again that s has to be taken for properties of the liquid metal in the solid phase. Also note that the solidied length, , changes during the process and discussing the case where the whole die is solidied is not of interest. Initially the thickness, = 0 (or very small). The characteristic time for very thin layers is very small, tc s 0. As the solidied layer increases the characteristic time also increases. However, the temperature prole is almost established (if other processes were to remain in the same conditions). Similar situations can be found when a semi innite slab undergoes solidication with T changes as well as results of increase in the resistance. For the foregoing reasons the characteristic time is very small.
Solidication time
Millers approach Following Eckerts work, Miller and his student [20] altered the calculations14 and based the assumption that the conduction heat transfer characteristic time in die (liquid metal in solid phase) is the same order magnitude as the solidication time. This assumption leads them to conclude that the main resistance to the solidication is in the interface between the die and mold 15 . Hence they conclude that the solidied front moves according to the following hsl vn = hT (3.25)
Where here h is the innovative heat transfer coecient between solid and solid16 and vn is front velocity. Then the lling time is given by the equation ts = hsl hT (3.26)
14 Miller and his student calculate the typical forces required for clamping. The calculations of Miller has shown an interesting phenomenon in which small casting (2[kg]) requires a larger force than heavier casting (20[kg])?! Check it out in their paper, page 43 in NADCA Transaction 1997! If the results extrapolated (not to much) to about 50[kg] casting, no force will be required for clamping. Furthermore, the force for 20 [kg] casting was calculated to be in the range of 4000[N ]. In reality, this kind of casting will be made on 1000 [ton] machine or more (3 order of magnitude larger than Miller calculation suggested). The typical required force should be determined by the plunger force and the machine parts transient characteristics etc. Guess, who sponsored this research and how much it cost! 15 An example how to do poor research. These kind of research works are found abundantly in Dr. Miller and Dr. J. Brevick from Ohio State Univerity. These works when examined show contractions with the logic and the rest of the world of established science. 16 This coecient is commonly used either between solid and liquid, or to represent the resistance between two solids. It is hoped that Miller and coworkers refer that this coecient to represent the resistance between the two solids since it is a minor factor and does not determine the characteristic time.
49
where designates the half die thickness. As a corollary conclusion one can arrive from this construction is that the lling time is linearly proportional to the die thickness since hsl /hT is essentially constant (according to Miller). This interesting conclusion contradicts all the previous research about solidication problem (also known as the Stefan problem). That is if h is zero the time is zero also. The author is not aware of any solidication problem to show similar results. Of course, Miller has all the experimental evidence to back it up! Present approach Heat balance at the liquid-solid interface yields (Tl Ts ) (3.27) n where n is the direction perpendicular to the surface and has to be taken at the solid phase see Appendix 10. Additionally note that in many alloys, the density changes during the solidication and is substantial which has a signicant eect on the moving of the liquid/solid front. It can be noticed that at the die interface ks T/n kd T/n = (opposite to Miller) and further it can be assumed that temperature gradient in the liquid side, T/n 0 , is negligible compared to other uxes. Hence, the speed of the solid/liquid front moves s hsf vn = k vn = k Ts Tl kTM B s hsl n s hsl L (3.28)
Notice the dierence to equation (3.26) The main resistance to the heat transfer from the die to the mold (cooling liquid) is in the die mold. Hence, the characteristic heat transfer from the mold is proportional to TM B /L17 . The characteristic temperature dierence is between the melting temperature and the boiling temperature. The time scale for the front can be estimated by ts = = s hsl
2 L
vs
kd TM B
(3.29)
Note that the solidication time isnt a linear function of the die thickness, , but a function of 2 18 .
Dissipation Time
Examples of how dissipation is governing the ow can be found abundantly in nature. l l l l +u +v +w t x y z = l 2 l 2 l 2 l + + 2 2 x y z 2 + (3.30)
17 The estimate can be improved by converting the resistances of the die to be represented by die length and the same for the other resistance into the cooling liquid i.e. 1/ho + L/k + CD ots + 1/hi . 18 L can be represented by for example, see more simplied assumption leads to pure = 2 .
50
+
2
v y w y
+ +
v y v z
2
+ 2 3
v x
u y
+
2
u v w + + x y z
(3.31)
Since the dissipation characteristic time isnt commonly studied in regular uid mechanics, we rst introduce two classical examples of dissipation problems. First problem deals with the oscillating manometer and second problem focuses on the rigid body brought to a rest in a thin cylinder.
D air air H equilibrioum level H
Fig a.
Fig b.
Mass, spring
Oscillating manometer
Example 3.1: A liquid in manometer is disturbed from a rest by a distance of H0 . Assume that the ow is laminar and neglected secondary ows. Describe H(t) as a function of time. Dened 3 cases: 1)under damping, 2) critical damping, and 3) over damping. Discuss the physical signicance of the critical damping. Compute the critical radius to create the critical damping. For simplicity assume that liquid is incompressible and the velocity prole is parabolic. Solution The conservation of the mechanical energy can be written as d dt (total of inflow of) + potential energy
rate of increase of kinetic and potential energy in system total net rate of surroundings work on the system inflow = (total of energy of) + (total of inflow of) + kinetic potential energy total rate mechanical energy dissipated because viscosity
(3.32)
51
The chosen system is the liquid in the manometer. There is no ow in or out of the liquid of the manometer, and thus, terms that deal with ow in or out are canceled. It is assumed that the surface at the interface is straight without end eects like surface tension. This system is unsteady and therefore the velocity prole is function of the time and space. In order to demonstrate the way the energy dissipation is calculated it is assumed the velocity is function of the radius and time but separated. This assumption is wrong and cannot be used for real calculations because the real velocity prole is not separated and can have positive and negative velocities. It is common to assume that velocity prole is parabolic which is for the case where steady state is obtained. This assumption can used as a limiting case and the velocity prole is r U (r, t) = U (r) = U0 (t) 1 R
2
U0
R
(3.33)
H(t)
U0 1
r 2 R
V = H R2
where R the radius of the manometer. The velocity at the center is a function of time but independent of the Length. It can be noticed that this equation dim:eq:velocityH Fig. -3.14. Mass Balance to determine the is problematic because it breaks the assump- relationship between the U0 and the Height, H. tion of the straight line of the interface. The relationship between the velocity at the center, U0 to the height, H(t) can be obtained from mass conservation on left side of the manometer (see Figure 3.14) is d H R2 = dt
R 0
U0
r 1 R
dA
2 r dr
(3.34)
Equation (3.34) relates H(t) to the center velocity, U0 , and the integration results in dH U0 = dt 2 (3.35)
Note that H(t) isnt a function of the radius, R. This relationship (3.35) is based on the denition that U0 is positive for the liquid owing to right and therefore the height decreases. The total kinetic energy in the tube is then
L R 0
Kk =
0
U0 2 2
r 1 R
2 2
dA
2 r dr d =
L U0 2 R 2 6
(3.36)
where L is the total length (from one interface to another) and d is a coordinate running along the axis of the manometer neglecting the curvature of the U shape. It can be noticed that L is constant for incompressible ow. It can be observed that the disturbance of the manometer creates a potential energy which can be measured from a datum at the maximum lower point. The maximum potential energy is obtained when H is either maximum or minimum. The maximum kinetic energy is obtained when H
52
is zero. Thus, at maximum height, H0 the velocity is zero. The total potential of the system is then left side
H0 H dV
right side
2 H0 +H 0 dV
Kp =
0
( g ) R d +
( g ) R2 d = H0 2 + H 2 g R2 (3.37)
The last term to be evaluated is the viscosity dissipation. Based on the assumptions in the example, the velocity prole is function only of the radius thus the only gradient of the velocity is in the r direction. Hence
R
Ed = = L
0
dU dr
dA
2 r dr
(3.38)
4 r2 U0 2 R4
(3.39)
Ed = 2 L
0
4 r2 U0 2 r dr = 2 L R2 U0 2 R2 R R
(3.40)
The work done on system is neglected by surroundings via the pressure at the two interfaces because the pressure is assumed to be identical. Equation (3.32) is transformed, in this case, into d (Kk + Kp ) = Ed dt The kinetic energy derivative with respect to time (using equation (3.35)) is d Kk d = dt dt L U0 2 R 2 6 = L R2 U0 4 L R2 d H d2 H 2 U0 = 6 dt 3 dt dt2 (3.42) (3.41)
Substituting equations (3.43), (3.42) and (3.38) into equation (3.41) results in 4 L R 2 d H d2 H dH + 2H g R 2 + 2 L U0 2 = 0 2 3 dt dt dt (3.44)
3.4. ESTIMATES OF THE TIME SCALES IN DIE CASTING Equation (3.44) can be simplied using the identity of (3.35) to be d2 H 6 dH 3g + + H=0 2 2 dt dt R 2L
53
(3.45)
This equation is similar to the case mass tied to a spring with damping. This equation is similar to RLC circuit19 . The common method is to assume that the solution of the form of A e t where the value of A and will be such determined from the equation. When substituting the guessed function into result that having two possible solution which are
6R2 6 R2 2
6g L
(3.46)
H = A e1 t + A e2 t H = Ae
t
= =
1 = 2 1 = 2 = (3.47)
+ Ae
The constant A1 and A2 are to be determined from the initial conditions. The value under the square root determine the kind of motion. If the value is positive then the system is overdamped and the liquid height will slowly move the equilibrium point. If the value in square is zero then the system is referred to as critically damped and height will move rapidly to the equilibrium point. If the value is the square root is negative then the solution becomes a combination of sinuous and cosines. In the last case the height will oscillate with decreasing size of the oscillation. The critical radius is then Rc =
4
6 2 L g 2
(3.48)
It can be observed that this analysis is only the lower limit since the velocity prole is much more complex. Thus, the dissipation is much more signicant.
End Solution
Example 3.2: A thin (t/D 1) cylinder full with liquid is rotating in a velocity, . The rigid body is brought to a stop. Assuming no secondary ows (Bernards cell, etc.), describe the ow as a function of time. Utilize the ratio 1 t/D. d2 X + dt2
2
dX +X =0 dt
(3.49)
Discuss the case of rapid damping, and the case of the characteristic damping
19 An electrical circuit consisting of a resistor (R), an inductor (L), and a capacitor (C), connected in series or in parallel.
54
Solution
End Solution
These examples illustrate that the characteristic time of dissipation can be assessed by (du/dy )2 thus given by 2 /. Note the analogy between ts and tdiss , for which 2 appears in both of them, the characteristic length, , appears as the typical die thickness.
3.4.2
The ratio of several time ratios can be examined for typical die casting operations. The ratio of solidication time to the lling time tf Lkd TM B Ste = ts P r Re U s hsl L where Re Ste Reynolds number Stefan number
cp U lm
lm
lm s
kd klm
L L
(3.50)
TM B
hsl
the discussion is augmented on the importance of equation (3.50). The ratio is extremely important since it actually denes the required lling time. tf = C lm s kd klm L L Ste P r Re (3.51)
55
At the moment, the constant, C, is unknown and its value has to come out from experiments. Furthermore, the constant is not really a constant and is a very mild function of the geometry. Note that this equation is also dierent from all the previously proposed lling time equations, since it takes into account solidication and lling process20 . The ratio of liquid metal conduction characteristic time to characteristic lling time is given by tc d U L2 U L2 L2 = = Re P r tf L L L (3.52)
The solidication characteristic time to conduction characteristic time is given by s hsl Ld ts 1 = 2 tc Ste kd TM B L s d cp lm cp d (3.53)
The ratio of the lling time and atomization is ta viscosity U = Ca tf L L 6 108 (3.54)
Note that , in this case, is the thickness of the gate and not of the die cavity. ta momentum 2U 2 = We tf L L 0.184 (3.55)
which means that if atomization occurs, it will be very fast compared to the lling process. The ratio of the dissipation time to solidication time is given by
2 tdiss kd TM B = ts lm s hsl L
Ste Pr
kd klm
lm s
100
(3.56)
this equation yields typical values for many situations in the range of 100 indicating that the solidication process is as fast as the dissipation. It has to be noted that when the solidication progress, the die thickness decreases. The ratio, /L, reduced as well. As a result, the last stage of the solidication can be considered as a pure conduction problem as was done by the English group.
56
3.5.1
Governing equations
The lling of the mold cavity can be divided into two periods. In the rst period (only uid mechanics; minimum heat transfer/solidication) and the second period in which the solidication and dissipation occur. This discussion deals with how to conduct experiments in die casting21 . It has to be stressed that the conditions downstream have to be understood prior to the experiment with the die lling. The liquid metal velocity prole and ow pattern are still poorly understood at this stage. However, in this discussion we will assume that they are known or understood to same degree22 . The governing equations are given in the preceding sections and now the boundary conditions will be discussed. The boundary condition at the solid interface for the gas/air and for the liquid metal are assumed to be noslip condition which reads ug = vg = wg = ulm = vlm = wlm = 0 (3.57)
where the subscript g is used to indicate the gas phase. It is noteworthy to mention that this can be applied to the case where liquid metal is mixed with air/gas and both are touching the surface. At the interface between the liquid metal and gas/air, the pressure jump is expressed as (3.58) p r1 + r2 where r1 and r2 are the principal radii of the free surface curvature, and, , is the surface tension between the gas and the liquid metal. The surface geometry is determined by several factors which include the liquid movement23 instabilities etc. Now on the dicult parts, the velocity at gate has to be determined from the pQ2 diagram or previous studies on the runner and shot sleeve. The diculties arise due to the fact that we cannot assign a specic constant velocity and assume only liquid ow out. It has to be realized that due to the mixing processes in the shot sleeve and the runner (especially in a poor design process and runner system, now commonly used in the industry), some portions at the beginning of the process have a signicant part which contains air/gas. There are several possibilities that the conditions can be prescribed. The rst possibility is to describe the pressure variation at the entrance. The second possibility is to describe the velocity variation (as a function of time). The velocity is reduced during the lling of the cavity and is a function of the cavity geometry. The change in the velocity is sharp in the initial part of the lling due to the change from a free jet to an immersed jet. The pressure varies also at the entrance, however, the variations are more mild. Thus, it is a better possibility24 to consider the pressure prescription. The simplest assumption is constant pressure P = P0 = 1 U0 2 2 (3.59)
21 Only minimal time and eorts was provided how to conduct experiments on the lling of the die. In the future, other zones and dierent processes will be discussed. 22 Again the die casting process is a parabolic process. 23 Note, the liquid surface cannot be straight, for unsteady state, because it results in no pressure gradient and therefore no movement. 24 At this only an intelligent guess is possible.
3.5. SIMILARITY APPLIED TO DIE CAVITY We also assume that the air/gas obeys the ideal gas model. g = P RT
57
(3.60)
where R is the air/gas constant and T is gas/air temperature. The previous assumption of negligible heat transfer must be inserted and further it has to be assumed that the process is polytropic25 . The dimensionless gas density is dened as = = 0 P0 P
1 n
(3.61)
The subscript 0 denotes the atmospheric condition. The air/gas ow rate out the cavity is assumed to behave according to the model in Chapter 9. Thus, the knowledge of the vent relative area and 4f L are important D parameters. For cases where the vent is well designed (vent area is near the critical area or above the density, g can be determined as was done by [5]). To study the controlling parameters, the equations are dimensionlessed. The mass conservation for the liquid metal becomes lm lm u lm lm v + + t x y
lm
lm w z
lm
=0
(3.62)
where x = x , y = y/ , z = z/ , u = u/U0 , v = v/U0 , w = w/U0 and the dimensionless time is dened as t = tU0 , where U0 = 2P0 /. Equation (3.62) can be similar under the assumption of constant density to read u lm v lm w lm + + =0 x y z (3.63)
Please note that this simplication can be used for the gas phase. The momentum equation for the liquid metal in the x-coordinate assuming constant density and no body forces reads lm u lm lm u lm lm u lm lm u lm +u +v +w = t x y z p lm 1 + x Re 2 ulm 2v 2w + + x 2 ylm 2 zlm 2 (3.64)
where Re = U0 /lm and p = p/P0 . The gas phase continuity equation reads g ug g vg g wg g + + + =0 t x y z
25 There
(3.65)
are several possibilities, this option is chosen only to obtain the main controlling parameters.
58
p g lm g0 1 + x g lm Re
2 ug x 2
2 wg z
2
(3.66)
Note that in this equation, additional terms were added, (lm /g )(g0 /lm ). The no-slip conditions are converted to: ug = vg = wg = ulm = vlm = wlm = 0 The surface between the liquid metal and the air satisfy 1 p (r1 + r2 ) = We where the p , r1 , and r2 are dened as r1 = r1 / r2 = r2 / The solution to equations has the form of u v w p = = = = fu x , y , z , Re, W e, fv fw fp A 4f L g lm , D , n, , Ac lm g g lm A 4f L x , y , z , Re, W e, , D , n, , Ac lm g A 4f L g lm x , y , z , Re, W e, , , n, , Ac D lm g g lm A 4f L , n, x , y , z , Re, W e, , , Ac D lm g (3.67)
(3.68)
(3.69)
If it will be found that equation (3.66) can be approximated27 by p g u g u g u g u g +u +v +w t x y z x then the solution is reduced to u v w p = = = = fu x , y , z , Re, W e, fv fw fp A 4f L , ,n Ac D A 4f L x , y , z , Re, W e, , ,n Ac D A 4f L x , y , z , Re, W e, , ,n Ac D A 4f L x , y , z , Re, W e, , ,n Ac D (3.70)
(3.71)
26 In writing this equation, it is assumed that viscosity of the air is independent of pressure and temperature. 27 This topic is controversial in the area of two phase ow.
59
At this stage, it is not known if it is the case and if it has to come out from the experiments. The density ratio can play a role because two phase ow characteristic is a major part of the lling process.
3.5.2
Design of Experiments
28
Under Construction
Reynolds number represents the ratio of the momentum forces to the viscous forces. In die casting, Reynolds number plays a signicant role which determines the ow pattern in the runner and the vent system. The discharge coecient, CD , is used in the pQ2 diagram is determined largely by the Re number through the value of friction coecient, f, inside the runner. Eckert number Ec = 1/2U 2 inertial energy = 1/2cp T thermal energy
Eckert number determines if the role of the momentum energy transferred to thermal energy is signicant. Brinkman number Br = heat production by viscous dissipation U 2 / 2 = 2 kT / heat transfer transport by conduction
Brinkman number is a measure of the importance of the viscous heating relative the conductive heat transfer. This number is important in cases where large velocity change occurs over short distances such as lubricant ow (perhaps, the ow in the gate). In die casting, this number has small values indicating that practically the viscous heating is not important.
28 See
60 Mach number Ma =
U
p
For ideal gas (good assumption for the mixture of the gas leaving the cavity). It becomes U characteristic velocity M = = gas sound velocity RT Mach number determines the characteristic of ow in the vent system where the air/gas velocity is reaching to the speed of sound. The air is chocked at the vent exit and in some cases other locations as well for vacuum venting. In atmospheric venting the ow is not chocked for large portion of the process. Moreover, the ow, in well design vent system, is not chocked. Yet the air velocity is large enough so that the Mach number has to be taken into account for reasonable calculation of the CD . Ozer number Oz =
CD 2 Pmax 2 Qmax A3
A3 Qmax
CD 2
One of the most important number in the pQ2 diagram calculation is Ozer number. This number represents how good the runner is designed. Froude number Fr = U 2 / inertial forces = g gravity forces
F r number represent the ratio of the gravity forces to the momentum forces. It is very important in determining the critical slow plunger velocity. This number is determined by the height of the liquid metal in the shot sleeve. The Froude number does not play a signicant role in the lling of the cavity. Capillary number Ca = inertial forces U 2 / = g gravity forces
capillary number (Ca) determine when the ow during the lling of the cavity is atomized or is continuous ow (for relatively low Re number).
3.7. SUMMARY Weber number We = 1/2U 2 inertial forces = 1/2/ surface forces
61
W e number is the other parameter that govern the ow pattern in the die. The ow in die casting is atomized and, therefore, W e with combinations of the gate design also determine the drops sizes and distribution. Critical vent area Ac = V (0) ctmax mmax
The critical area is the area for which the air/gas is well vented.
3.7 Summary
The dimensional analysis demonstrates that the uid mechanics process, such as the lling of the cavity with liquid metal and evacuation/extraction of the air from the mold, can be dealt with when heat transfer is neglected. This provides an excellent opportunity for simple models to predict many parameters in the die casting process. It is recommended for interested readers to read Eckerts book Analysis of Heat and Mass transfer to have better and more general understanding of this topic.
3.8 Questions
Under construction
62
4.1 Introduction
The die casting engineer encounters many aspects of network ow. For example, the liquid metal ows in the runner is a network ow. The ow of the air and other gases out of the mold through the vent system is also another example network ow. The pQ2 diagram also requires intimate knowledge of the network ow. However, most die casting engineers/researchers are unfamiliar with uid mechanics and furthermore have a limited knowledge and understanding of the network ow. Therefore, this chapter is dedicated to describe a brief introduction to a ow in a network. It is assumed that the reader does not have extensive background in uid mechanics. However, it is assumed that the reader is familiar with the basic concepts such as pressure and force, work, power. More comprehensive coverage can be found in books dedicated to uid mechanics and pipe ow (network for pipe). First a discussion on the relevancy of the data found for other liquids to the die casting process is presented. Later a simple ow in a straight pipe/conduit is analyzed. Dierent components which can appear in network are discussed. Lastly, connection of the components in series and parallel are presented.
63
64
1 Head Loss [meter] 0.1 0.01 0.001 0.0001 1e05 1e06
0.00 0.00 0.50 1.00 0.02 0.06 0.12 0.01 0.03 0.25
1e07
100
1000
Reynolds Number
10000
Velocity[m/sec]
Fig a.
Friction of orice as
Fig b.
The collapsed
a function velocity.
iments is a ow of dierent liquids in a pipe with an orice (see Figure 4.1). Dierent liquids create signicant head loss for the same velocity. Moreover, the dierences for the dierent liquids are so signicant that the similarity is unclear as shown in Figure ??. As the results of the past geniuses work, it can be shown that when results are normalized by Reynolds number (Re) instated of the velocity and when the head loss is replaced by the loss coecient, UH g one obtains that all the lines are collapsed on to a 2 /2 single line as shown in Figure 4.1b. This result indicates that the experimental results obtained for one liquid can be used for another liquid metal provided the other liquid is a Newtonian liquid1 . Researchers shown that the liquid metal behaves as Newtonian liquid if the temperature is above the mushy zone temperature. This example is not correct only for this spesic geometry but is correct for all the cases where the results are collapsed into a single line. The parameters which control the problem are found when the results are collapsed into a single line. It was found that the resistance to the ow for many components can be calculated (or extracted from experimental data) by knowing the Re number and the geometry of the component. In a way you can think about it as a prof of the dimensional analysis (presented in Chapter 3).
1 Newtonian
65
to the ow. The relationship between the pressure dierence, the ow rate and the resistance to the ow is given by the experimental equation (4.1). This equation is used because it works2 . The pressure dierence determined by the geometrical parameters and the experimental data which expressed by f 3 which can be obtained from Moodys diagram. P = f L U2 L U2 ; H = f D 2 D 2g (4.1)
Note, head is energy per unit weight of uid (i.e. Force x Length/Weight = Length) and it has units of length. Thus, the relationship between the Head (loss) and the pressure (loss) is P = H g (4.2)
The resistance coecient for circular conduit can be dened as L KF = f (4.3) Fig. -4.3. General simple conduit description. D This equation is written for a constant density ow and a constant cross section. The ow rate is expressed as Q=UA (4.4) The cross sectional area of circular is A = r2 = D2 /4, using equation (4.4) and substituting it into equation (4.1) yields 16 L 2 Q (4.5) 2 D3 The equation (4.5) shows that the required pressure dierence, P , is a function of 1/D3 which demonstrates the tremendous eect the diameter has on the ow rate. The length, on the other hand, has mush less signicant eect on the ow rate. The power which requires to drive this ow is give by P = f P = QP (4.6)
These equations are very important in the understanding the economy of runner design, and will be studied in Chapter 12 in more details. The power in terms of the geometrical parameters and the ow rate is given P = {
2 Actually
L Q D
(4.7)
there are more reasons but they are out of the scope of this book this stage, we use dierent denition than one used in Chapter A. The dierence is by a factor of 4. Eventially we will adapt one system for the book.
3 At
66
4.3.1
Example 4.1: calculate the pressure loss (dierence) for a circular cross section pipe for driving aluminum liquid metal at velocity of 10[m/sec] for a pipe length of 0.5 [m] (like a medium quality runner) with diameter of 5[mm] 10[mm] and 15[mm] Solution
This is example 4.3.1 Example 4.2: calculate the power required for the above example Solution
4.4.1
bend
The resistance in the bend is created because a change in the momentum and the ow pattern. Engineers normally convert the bend to equivalent conduit length. This conversion produces adequate results in same cases while in other it might introduce larger error. The knowledge of this accuracy of this conversion is very limited because limited study have been carry out for the characteristic of ows in die casting. From the limited information the author of this book gadered it seem that it is reasonable to carry this conversion for the calculations of liquid metal ow resistance while in the air/liquid metal mixture it far from adequate. Moreover, hole of our knowledge of the gas ow in vent system are far more large. Nevertheless, for the engineering purpose at this stage it seem that some of the errors will cancel each other and the end result will be much better. The schematic of a bend commonly used in die casting is shown in Figure ??. The resistance of the bend is a function of several parameters: angle, , radius, R and the geometry before and after the
cross section
67
bend. Commonly, the runner is made with the same geometry before and after the bend. Moreover, we will assume in this discussion that downstream and upstream do not inuence that ow in the ow. This assumption is valid when there is no other bend or other change in the ow nearby. In cases that such a change(s) exists more complicated analysis is required. In the light of the for going discussion, we left with two parameters that control the resistance, the angle, , and the radius, R As larger the angle is larger the resistance will be. In the practice today, probably because the way the North American Die Casting Association teaching, excessive angle can be found through the industry. It is recommended never to exceed the straight angle (900 ). Figure ?? made from a data taken from several sources. From the Figure it is clear that optimum radius should be around 3.
4.4.2
Y connection
picture of Y connection
4.4.3
Expansion/Contraction
One of the undisirable element is the runner system is sudden change in the conduict area. In some instance they are inevodeble. We will disscuss how to design and what are the better design options which availble for the engineer.
4.5.1
Series Connection
The ow rate in dierent locations is a function of the temperature. Eckert [13] demonstrated that the heat transfer is insignicant in the duration of the lling of the cavity, and therefore the temperature of the liquid metal can be assumed almost constant during the lling period (which in most cases is much less 100 milliseconds). As such, the solidication is insignicant (the liquid metal density changes less than 0.1% in the runner); therefore, the volumetric ow rate can be assumed constant: Q1 = Q2 = Q3 = Qi (4.8)
68
CHAPTER 4. FUNDAMENTALS OF PIPE FLOW Clearly, the pressure in the points is dierent and P1 = P2 = P3 = (4.9)
However the total pressure loss is composed of from all the small pressure loss P1 Pend = (P1 P2 ) + (P2 P3 ) + Every single pressure loss can be written as Pi1 Pi = Ki U2 2 (4.11) (4.10)
There is also resistance due to parallel connection i.e. y connections, y splits and manifolds etc. First, lets look at the series connection. (see Figure ??). where: Kbend L f the resistance in the bend length of the duct (vent), friction factor, and
4.5.2
Parallel Connection
An example of the resistance of parallel connection (see Figure ??). The pressure at point 1 is the same for two branches however the total ow rate is the combination Qtotal = Qi + Qj between two branches and the loss in the junction is calculated as To add a gure and check if the old one is good
Fig. -4.5. A parallel connection
(4.12)
5.1 Introduction
Fl ow
One of the branches of the uid mechanics discussed in Chapter 2. Here we expand this issue further because it is give the basic understanding to the wave phenomenon. There are numerous books that dealing with open channel ow and the interested reader can broader his/her knowledge by reading book such as OpenChannel Hydraulics by Ven Te Chow (New York: McGraw-Hill Book Company, Inc. Fig. -5.1. Equilibrium of Forces in an open 1959). Here a basic concepts for the non- channel. Fluid Mechanics Engineers are given. The ow in open channel ow in steady state is balanced by between the gravity forces and mostly by the friction at the channel bed. As one might expect, the friction factor for open channel ow has similar behavior to to one of the pipe ow with transition from laminar ow to the turbulent at about Re 103 . Nevertheless, the open channel ow has several respects the cross section are variable, the surface is at almost constant pressure and the gravity force are important. The ow of a liquid in a channel can be characterized by the specic energy that is associated with it. This specic energy is comprised of two components: the hydrostatic pressure and the liquid velocity1 .
y x d
1 The
69
70
CHAPTER 5. FLOW IN OPEN CHANNELS The energy at any point of height in a rectangular channel is e= U2 P + +z 2g (5.1)
why? explain
and, since
+ z = y for any point in the cross section (free surface), e=y+ U2 2g (5.2)
where: e y g U specic energy per unit height of the liquid in the channel acceleration of gravity average velocity of the liquid
If the velocity of the liquid is increased, the height, y, has to change to keep the same ow rate Q = q b = b y U . For a specic ow rate and cross section, there are many combinations of velocity and height. Plotting these points on a diagram, with the ycoordinate as the height and the xcoordinate as the specic energy, e, creates a parabola on a graph. This line is known as the specic energy curve. Several conclusions can be drawn from Figure ??. First, there is a minimum Fig. -5.2. Specic Energy and momentum energy at a specic height known as the Curves. critical height. Second, the energy increases with a decrease in the height when the liquid height is below the critical height. In this case, the main contribution to the energy is due to the increase in the velocity. This ow is known as the supercritical ow. Third, when the height is above the critical height, the energy increases again. This ow is known as the subcritical ow, and the energy increase is due to the hydrostatic pressure component. The minimum point of energy curve happens to be at U = g yc (5.3)
Y EL Y2
specific energy line
Y3
YC
Y1
e, f
q2 g
(5.4)
5.2. TYPICAL DIAGRAMS The minimum thrust also happens to be at the same point U = can dene the dimensionless number as: g yh Fr = 2 U
Dividing the velocity by gy provides one with the ability to check if the ow is above or below critical velocity. This quantity is very important, and its signicance can be studied from many books on uid mechanics. The gravity eects are measured by the Froude number which is dened by equation (5.6).
72
6.1 Introduction
In this chapter the design and the dierent relationship between runner segments are studied herein. The rst step in runner design is to divide the mold into several logical sections. The volume of every section has to be calculated. Then the design has to ensure that the gate velocity and the lling time of every section to be as recommend by experimental results. At this stage there is no known reliable theory/model known to the author to predict these values. The values are based heavily on semi-reliable experiments. The Backward Design is discussed. The reader with knowledge in electrical engineering (electrical circles) will notice in some similarities. However, hydraulic circuits are more complex. Part of the expressions are simplied to have analytical expressions. Yet, in actuality all the terms should be taken into considerations and commercial software such DiePerfectshould be used.
6.1.1
Backward Design
Suppose that we have n sections with n gates. We know that volume to be delivered at gate i and is denoted by vi.The gate velocity has to be in a known range. The lling time has to be in a known function and we recommend to use Eckert/BarMeirs formula. For this discussion it is assumed that the lling has to be in known range and the ow rate can be calculated by Qi = Vi ti 73 (6.1)
(6.2)
Armed with this knowledge, one can start design the runner system.
6.1.2
Design of connected runner segments have insure that the ow rate at each segment has to be designed one. In Figure ??a branches i and j are connected to branch at point K. The pressure drop (dierence) on branches i and j has to be the same since the pressure in the mold cavity is the same for both segments. The sum of the ow rates for both branch has to be equal to ow rate in branch Q = Qi + Qj = Qj = Q Qi (6.3) The ow rate in every branch is related to the pressure dierence by Qi = P Ri
Mold
i K j
b Biscuit
(6.4)
Where the subscript i in this case also means any branch e.g. i, j and so on. For example, one can write for branch j Oj = DeltaP Rj (6.5)
Utilizing the mass conservation for point K in which Qkappa = Oi + Oj and the fact that the pressure dierence, DeltaP , is the same thus we can write Qk = DeltaP DeltaP Ri Rj + = DeltaP Ri Rj Ri + Rj (6.6)
Lets further manipulate the equations to get some more important relationships. Using equation (??) and equation (??) DeltaP i = DeltaP j = Oi Ri = Oj Rj (6.8)
6.1. INTRODUCTION
75
The ow rate in a branch j can be related to ow rate in branch i and corresponding resistances Oj = Ri Oi Rj (6.9)
Using equation (??) and equation (??) one can obtain Oi Rj = Qk Rj + Ri Solving for the resistance ratio since the ow rate is known Ri Oi = 1 Rj Qk (6.11) (6.10)
6.1.3
Resistance
1 2
What does the resistance include? How to achieve resistance ratio in the previous equation (??) will be discussed herein further. The total resistance reads
(6.12)
The contraction resistance, Rcontraction , is the due the contraction of the gate. The exit resistance, Rexit , is due to residence of the liquid metal in mold cavity. Or in other words, the exit resistance is due the lost of energy of immersed jet. The angle resistance, R is due to the change of direction. The Rki is the resistance due to ow in the branch on branch i. The geometry resistance Rgeometry , is due to who rounded the connection. DeltaP L U2 =f HD 2 since Ui =
Oi A
(6.13)
(6.14)
(6.15)
fi = fj = f
(C)f
L Oi 2 L Oj 2 = (C)f HD 3 i 2 HD 3 j 2
(6.17)
HD i = HD j
Oi Oj
2 3
(6.18)
Comparison between scrap between (multi-lines) two lines to one line rst nd the diameter equivalent to two lines L Qk 2 L (Oi + Oj )2 3 = (C)f 2 2 HD k HD 3 k
DeltaP = (C)f
(6.19)
Oi 2 = subtitling in to
DeltaP 2 HD i3 f L
(6.20)
HD k =
(HD 3 + HD 3 i j
(6.21)
Now we know the relationship between the hydraulic radius. Let see what is the scrap dierence between them. put drawing of the trapezoid let scrap denoted by converting the equation i constL
3/2
(6.22)
6.1. INTRODUCTION
77
i + j (HD i + 2 + HD j 2 ) = k HD k 2 and now lets write HD k in term of the two other (HD 2 + HD 2 ) i j HD 3 + HD 3 i j
2/3
(6.23)
(6.24)
78
7.1 Introduction
The pQ2 diagram is the most common gate mold calculation, if any at all, are used by most 3 die casting engineers. The importance of gate this diagram can be demonstrated by the runner fact that tens of millions of dollars have been invested by NADCA, NSF, and other 2 major institutes here and abroad in the Shot sleeve 1 pQ2 diagram research. The pQ2 diagram is one of the manifestations of supply and demand theory which was developed by Alfred Marshall (18421924) in the turn of Fig. -7.1. Schematic of typical die casting mathe century. It was rst introduced to the chine. die casting industry in the late70s [12]. In this diagram, an engineer insures that die casting machine ability can fulll the die mold design requirements; the liquid metal is injected at the right velocity range and the lling time is small enough to prevent premature freezing. One can, with the help of the pQ2 diagram, and by utilizing experimental values for desired lling time and gate velocities improve the quality of the casting. In the die casting process (see Figure 7.1), a liquid metal is poured into the shot sleeve where it is propelled by the plunger through the runner and the gate into the mold. The gate thickness is very narrow compared with the averaged mold thickness and the runner thickness to insure that breakage point of the scrap occurs at that gate location. A solution of increasing the discharge coecient, CD , (larger conduits) results 79
80
in a larger scrap. A careful design of the runner and the gate is required. First, the common pQ2 diagram1 is introduced. The errors of this model are analyzed. Later, the reformed model is described. Eects of dierent variables is studied and questions for students are given in the end of the chapter.
7.2
The injection phase is (normally) separated into three main stages which are: slow part, fast part and the intensication (see Figure 7.2). In the slow part the plunger moves in the critical velocity to prevent wave formation and therefore expels maximum air/gas before the liquid metal enters the cavity. In the fast part the cavity supposed to be lled in such way to prevent premature freezing and to obtain the right lling pattern. The intensication part is to ll the cavity with additional material to compensate for the shrinkage porosity during the solidication process. The pQ2 diagram deals with the second part of the lling phase. In the pQ2 diagram, the solution is determined by nding the intersecting point of the runner/mold characteristic line with the pump (die casting machine) characteristic line. The intersecting point sometime refereed to as the operational point. The machine characteristic line is assumed to be understood to some degree and it requires nding experimentally two Time coecients. The runner/mold characteristic line requires knowledge on the eciency/discharge coecient, CD , thus it Fig. -7.2. A typical trace on a cold chamber is an essential parameter in the calcula- machine tions. Until now, CD has been evaluated either experimentally, to be assigned to specic runner, or by the liquid metal properties (CD ) [9] which is de facto the method used today and refereed herein as the common pQ2 diagram2 . Furthermore, CD is assumed constant regardless to any change in any of the machine/operation parameters during the calculation. The experimental approach is arduous and expensive, requiring the building of the actual mold for each attempt with average cost of $5,000$10,000 and is rarely used in the industry3 . A short discussion about this issue is presented in the Appendix B comments to referee 2. Herein the common model (constant CD ) is constructed. The assumptions made in the construction of the model as following
plunger location liquid metal reaches to the venting system starging filling the cavity liquid metal pressure at the plunger tip or the hydralic pressure cavity filling
this model is described in NADCAs books method has been suggested in the literature in which the CD is evaluated based on the volume to be lled [10]. The author does not know of anyone who use this method and therefore is not discussed in this book. Nevertheless, this method is as good as the common method. 3 if you now of anyone who use this technique please tell me about it.
2 Another
1 as
7.2.
81
1. CD assumed to be constant and depends only the metal. For example, NADCA recommend dierent values for aluminum, zinc and magnesium alloys. 2. Many terms in Bernoullis equation can be neglected. 3. The liquid metal is reached to gate. 4. No air/gas is present in the liquid metal. 5. No solidication occurs during the lling. 6. The main resistance to the metal ow is in the runner. 7. A linear relationship between the pressure, P1 and ow rate (squared), Q2 . According to the last assumption, the liquid metal pressure at the plunger tip, P1 , can be written as P1 = Pmax 1 Where: P1 Q max Qmax the pressure at the plunger tip the ow rate maximum pressure which can be attained by the die casting machine in the shot sleeve maximum ow rate which can be attained in the shot sleeve Q Qmax
2
(7.1)
The Pmax and Qmax values to be determined for every set of the die casting machine and the shot sleeve. The Pmax value can be calculated using a static force balance. The determination of Qmax value is done by measuring the velocity of the plunger when the shot sleeve is empty. The maximum velocity combined with the shot sleeve crosssectional area yield the maximum ow rate, Qi = A Ui (7.2)
where i represent any possible subscription e.g. i = max Thus, the rst line can be drawn on pQ2 diagram as it shown by the line denoted as 1 in Figure ??. The line starts from a higher pressure (Pmax ) to a maximum ow rate (squared). A new combination of the same die casting machine and a dierent plunger diameter creates a dierent line. A smaller plunger diameter has a larger maximum pressure (Pmax ) and dierent maximum ow rate as shown by the line denoted as 2. The maximum ow rate is a function of the maximum plunger velocity and the plunger diameter (area). The plunger area is a obvious function of the plunger diameter, A = D2 /4. However, the maximum plunger velocity is a farmore complex function. The force that can be extracted from a die casting machine is essentially the same for dierent plunger diameters. The change in the resistance as results of changing the
82
plunger (diameter) depends on the conditions of the plunger. The dry friction will be same what dierent due to change plunger weight, even if the plunger conditions where the same. Yet, some researchers claim that plunger velocity is almost invariant in regard to the plunger diameter4 . Nevertheless, this piece of information has no bearing on the derivation in this model or reformed one, since we do not use it. Example 7.1: Prove that the maximum ow rate, Qmax is reduced and that Qmax 1/DP 2 (see Figure ??). if Umax is a constant
Pmax
Pmax
1 DP
2
Q max DP
2
Q=
2 A 3 CD
Pmax
Pmax Q max D1
Qmax
Fig b. Fig a.
The common pQ2 version.
Fig. -7.3. The left graph depicts the common version. The right graph depicts Pmax and Qmax as a function of the plunger diameter according to common model.
pQ2
A simplied force balance on the rode yields (see more details in section 7.11 page 97) Pmax = PB DB D1
2
PB DB 2 D1 2
(7.3)
where subscript B denotes the actuator. What is the pressure at the plunger tip when the pressure at the actuator is 10 [bars] with diameter of 0.1[m] and with a plunger diameter, D1 , of 0.05[m]? Substituting the data into equation (7.3) yields P1 = 10 0.1 0.05
2
= 4.0[M P a]
4 More research is need on this aspect. ++ read the comment made by referee II to the paper on pQ2 on page 192.
7.3. THE VALIDITY OF THE COMMON DIAGRAM In the common pQ2 diagram CD is dened as CD = 1 = constant 1 + KF
83
(7.4)
Note, therefore KF is also dened as a constant for every metal5 . Utilizing Bernoullis equation6 . U3 = CD 2P1 (7.5)
The ow rate in dierent locations is a function of the temperature. However, Eckert demonstrated that the heat transfer is insignicant in the duration of the lling of the cavity, and therefore the temperature of the liquid metal can be assumed almost constant during the lling period (which in most cases is much less 100 milliseconds). As such, the solidication is insignicant (the liquid metal density changes less than 0.1% in the runner); therefore, the volumetric ow rate can be assumed constant:
7
Q1 = Q2 = Q3 = Q
(7.7)
Hence, we have two equations (7.1) and (7.6) with two unknowns (Q and P1 ) for which the solution is P1 = 1 Pmax 2CD 2 Pmax A3 2 Qmax 2 (7.8)
insert a discussion in regards to the trends
insert the calculation with redU3 dP1 spect to and odA1 dA1
author would like to learn who cameout with this clever idea. more details see section 7.4 page 87. 7 read more about it in Chapter 3.
84
7.3.1
Is the mass balance really satised in the common model? Lets examine this point. Equation (7.7) states that the mass (volume, under constant density) balance is exist. A1 U1 = A3 U3 (7.9)
So, what is the condition on CD to satisfy this condition? Can CD be a constant as stated in assumption 1? To study this point let derive an expression for CD . Utilizing yields equation (7.5)
A1 U1 = A3 CD
2 P1
(7.10)
From the machine characteristic, equation (7.1), it can be shown that U1 = Umax Pmax P1 Pmax (7.11)
A3
Example 7.2: Derive equation 7.11. Start with machine characteristic equation (7.1)
Substituting equation (7.11) into equation (7.10) yield, A1 Umax Pmax P1 = A3 CD Pmax P1 (7.12)
It can be shown that equation (7.12) can be transformed into CD A1 Umax = A3 2 P1 Pmax
Pmax P1
(7.13)
Example 7.3: Find the relationship between CD and Ozer number that satisfy equation (7.13) According to the common model Umax , and Pmax are independent of the gate area, A3 . The term A3 other parameters).
P1 Pmax P1
7.3. THE VALIDITY OF THE COMMON DIAGRAM Example 7.4: nd the relationship between A3 Solution under construction
End Solution
85
P1 Pmax P1
and A3
P1
P
Q 1 DP
2
DP
1 DP P1 Q u3 D1
A1
Fig a.
P as A1 to be relocated
Fig b.
Fig. -7.5. Pressure at the plunger tip, P1 , the ow rate, Q, and the gate velocity, U3 as a function of plunger diameter , A1 .
Example 7.5: A3 what other parameters that CD depend on which do not provide the possibility CD = constant? To maintain the mass balance CD must be a function at least of the gate area, A3 . Since the common pQ2 diagram assumes that CD is a constant it diametrically opposite the mass conservation principle. Moreover, in the common model, a major assumption is that the value of CD depends on the metal, therefore, the mass balance is probably never achieved in many cases. This violation demonstrates, once for all, that the common pQ2 diagram is erroneous. Solution under construction.
End Solution
Use the information from example ?? and check what happened to the ow rate at two location ( 1) gate 2) plunger tip) when discharge coecient is varied CD = 0.40.9
86
7.3.2
Now second part, are the trends predicted by the common model are presumable (correct)? To examine that, we vary the plunger diameter, (A1 or D1 ) and the gate area, A3 to see if any violation of the physics laws occurs as results. The comparison between the real trends and the common trends is discussed in the following section. Plunger area/diameter variation First, the eect of plunger diameter size variation is examined. In section 7.2 it was shown that Pmax 1/D1 2 . Equation (7.8) demonstrates that P1 increases with an increase of Pmax . It also demonstrates that the value of P never can exceed P1 Pmax =
max
Qmax 2 CD A3
(7.14)
The value Pmax can attained is an innite value (according to the common model) therefore P1 is innite as well. The gate velocity, U3 , increases as the plunger diameter decreases as shown in Figure ??. Armed with this knowledge now, several cases can be examined if the trends are realistic. Gate area variation Energy conservation (power supply machine characteristic) Lets assume P that mass conservation is fullled, and, hence the plunger velocity can approach Q 2A C innity, U1 when D1 0 (under constant Qmax ). The hydraulic piston also has to move with the same velocity, U1 . Yet, according to the machine charP acteristic the driving pressure, approaches zero (PB 1 PB 2 ) 0. Therefore, the energy supply to the system is approachFig. -7.6. P1 as a function of Pmax . ing zero. Yet, energy obtained from the system is innite since jet is inject in innite velocity and nite ow rate. This cannot exist in our world or perhaps one can proof the opposite.
1 3 2 max 2 2 D max
Energy conservation (power supply) Assuming that the mass balance requirement is obtained, the pressure at plunger tip, P1 and gate velocity, U3 , increase (and can reach innity,(when P1 then U3 ) when the plunger diameter is reduced. Therefore, the energy supply to the system has to be innity (assuming a constant energy dissipation, actually the dissipation increases with plunger diameter in most ranges) However, the energy supply to the system (c.v. only the liquid metal) system would be PB 1 AB 1 U1 (nite amount) and the energy the system provide plus would be innity (innite gate velocity) plus dissipation.
87
Energy conservation (dissipation problem) A dierent way to look at this situation is check what happen to physical quantities. For example, the resistance to the liquid metal ow increases when the gate velocity velocity is increased. As smaller the plunger diameter the larger the gate velocity and the larger the resistance. However, the energy supply to the system has a maximum ability. Hence, this trend from this respect is unrealistic. Mass conservation (strike) According to the common model, the gate velocity decreases when the plunger diameter increases. Conversely, the gate velocity increases when the plunger diameter decreases8 . According to equation (7.2) the liquid metal ow rate at the gate increases as well. However, according to the common pQ2 diagram, the plunger can move only in a nite velocity lets say in the extreme case Umax 9 . Therefore, the ow rate at the plunger tip decreases. Clearly, these diametrically opposing trends cannot coexist. Either the common pQ2 diagram wrong or the mass balance concept is wrong, take your pick. Mass conservation (hydraulic pump): The mass balance also has to exist in hydraulic pump (obviously). If the plunger velocity have to be innite to maintain mass balance in the metal side, the mass ow rate at the hydraulic side of the rode also have to be innite. However, the pump has maximum capacity for ow rate. Hence, mass balance can be obtained.
to put table with dierent trends as a function of A3 and may be with a gure.
7.3.3
under construction
88
7.4.1
Eckert also demonstrated that the gravity eects are negligible11 . Assuming steady state12 and utilizing Bernoullis equation between point (1) on plunger tip and point (3) at the gate area (see Figure 7.1) yields P1 U1 2 P3 U3 2 + = + + h1,3 2 2 where: U h1,3 subscript 1 2 3 velocity of the liquid metal the liquid metal density energy loss between plunger tip and gate exit plunger tip entrance to runner system gate (7.16)
It has been shown that the pressure in the cavity can be assumed to be about atmospheric (for air venting or vacuum venting) providing vents are properly designed BarMeir at el 13 . This assumption is not valid when the vents are poorly designed. When they are poorly designed, the ratio of the vent area to critical vent area determines the build up pressure, P3 , which can be calculated as it is done in BarMeir et al However, this is not a desirable situation since a considerable gas/air porosity is created and should be avoided. It also has been shown that the chemical reactions do not play a signicant role during the lling of the cavity and can be neglected [5]. The resistance in the mold to liquid metal ow depends on the geometry of the part to be produced. If this resistance is signicant, it has to be taken into account calculating the total resistance in the runner. In many geometries, the liquid metal path in the mold is short, then the resistance is insignicant compared to the resistance in the runner and can be ignored. Hence, the pressure at the gate, P3 , can be neglected. Thus, equation (7.16) is reduced to U1 2 U3 2 P1 + = + h1,3 2 2
11 see
(7.17)
for more details chapter 3 in the section 7.4.4 on the transition period of the pQ2 13 Read a more detailed discussion in Chapter 9
12 read
7.4. THE REFORMED PQ2 DIAGRAM The energy loss, h1,3 , can be expressed in terms of the gate velocity as h1,3 = KF
2 U3 2
89
(7.18)
where KF is the resistance coecient, representing a specic runner design and specic gate area. Combining equations (7.7), (7.17) and (7.18) and rearranging yields U3 = CD where CD = f (A3 , A1 ) = 1 1
A3 A1 2
2P1
(7.19)
(7.20) + KF
Converting equation (7.19) into a dimensionless form yields Q= When the Ozer Number is dened as Oz =
CD 2 Pmax 2 Qmax A3
2Oz P
(7.21)
A3 Qmax
CD 2
Pmax
(7.22)
The signicance of the Oz number is that this is the ratio of the eective maximum energy of the hydrostatic pressure to the maximum kinetic energy. Note that the Ozer number is not a parameter that can be calculated a priori since the CD is varying with the operation point. 14 For practical reasons the gate area, A3 cannot be extremely large. On the other hand, the gate area can be relatively small A3 0 in this case Ozer number A3 A3 n where is a number larger then 2 (n > 2). Solving equations (7.21) with (7.15) for P , and taking only the possible physical solution, yields P = 1 1 + 2 Oz (7.23)
should be margin-note and so please ignore this footnote. how Ozer number behaves as a function of the gate area? Oz = A3 Pmax Qmax 2 1 A3 2 + K F A
1
90
7.4.2
This solution provide a powerful tool to examine various parameters and their eects on the design. The important factors that every engineer has to nd from these calculations are: gate area, plunger diameters, the machine size, and machine performance etc15 . These issues are explored further in the following sections. The gate area eects Gate area aects the reduced pressure, P , in two ways: via the Ozer number which include two terms: one, (A3 /Qmax ) and, two, discharge coecient CD . The discharge coecient, CD is also aected by the gate area aects through two dierent terms in the denition (equation 7.20), one, (A3 /A1 )2 and two by KF . Qmax eect is almost invariant with respect to the gate area up to small gate area sizes16 . Hence this part is somewhat clear and no discussion is need. (A3 /A1 )2 eects Lets look at the denition of CD equation (7.20). For illustration purposes let assume that KF is not a function of gate area, KF (A3 ) = constant. A small perturbation of the gate area results in Taylor series, CD = CD (A3 + A3 ) CD (A3 ) 1 A3 A3 = + A3 2 A3 2 1 A 1 2 + KF A1 2 1 A 1 2 + K F
3 A3 2 2 A 2 A1 4 1 A3 2 +KF
1
(7.24)
3 2
1 A 2 A1 2 1 A3 2 +KF
1
A3 2 + O(A3 )
3
A3 2 A1 2
+ KF
In this case equation (7.8) still hold but CD has to be reevaluated. repeat the example ?? with KF = 3.3 First calculate the discharge coecient, CD for various gate area starting from 2.4 106 [m2 ] to 3 104 [m2 ] using equation 7.20. This example demonstrate the very limited importance of the inclusion of the term (A3 /A1 )2 into the calculations. KF eects The change in the gate area increases the resistance to the ow via several contributing factors which include: the change in the ow cross section, change in the direction of the ow, frictional loss due to ow through the gate length, and the loss due to the abrupt expansion after the gate. The loss due to the abrupt expansion is a major contributor and its value changes during the lling process. The liquid metal
15 The machine size also limited by a second parameter known as the clamping forces to be discussed in Chapter 11 16 This is reasonable speculation about this point. More study is well come
91
enters the mold cavity in the initial stage as a free jet and sometime later it turns into an immersed jet which happens in many geometries within 5%-20% of the lling. The change in the ow pattern is believed to be gradual and is a function of the mold geometry. A geometry with many changes in the direction of the ow and/or a narrow mold (relatively thin walls) will have the change to immersed jet earlier. Many sources provide information on KF for various parts of the designs of the runner and gate. Utilizing this information produces the gate velocity as a function of the given geometry. To study further this point consider a case where KF is a simple function of the gate area. When A3 is very large then the eect on KF are relatively small. Conversely, KF when A3 0 the resistance, KF . The simplest function, shown in Figure 7.7, that represent such behavior is KF = C 1 + C2 A3
K0
(7.25)
A3 C1 and C2 are constants and can be calculated (approximated) for a specic geometry. The value of C1 determine the Fig. -7.7. KF as a function of gate area, A3 value of the resistance where A3 eect is minimum and C2 determine the range (point) where A3 plays a signicant eect. In practical, it is found that C2 is in the range where gate area are desired and therefore program such as DiePerfect are important to calculated the actual resistance.
Example 7.1: Under construction create a question with respect to the function 7.25 Solution Under construction The combined eects Consequently, a very small area ratio results in a very large resistance, and when A3 0 therefore the resistance resulting in a zero gate A1 velocity (like a closed valve). Conversely, for a large area ratio, the resistance is insensitive to variations of the gate area and the velocity is reduced with increase the gate area. Therefore a maximum gate velocity must exist, and can be found by dU3 =0 dA3 (7.26)
which can be solved numerically. The solution of equation (7.26) requires full information on the die casting machine. A general complicated runner design conguration can be converted into a straight conduit with trapezoidal crosssection, provided that it was proportionally designed
92
to create equal gate velocity for dierent gate locations17 . The trapezoidal shape is commonly used because of the simplicity, thermal, and for cost reasons. To illustrate only the eects of the gate area change two examples are presented: one, a constant pressure is applied to the runner, two, a constant power is applied to the runner. The resistance to the ow in the shot sleeve is small compared to resistance in the runner, hence, the resistance in the shot sleeve can be neglected. The die casting machine performance characteristics are isolated, and the gate area eects on the the gate velocity can be examined. Typical dimensions of the design are presented in Figure ??. The short conduit of 0.25[m] represents an excellent runner design and the longest conduit of 1.50[m] represents a very poor design. The calculations were carried for aluminum alloy with a density of 2385[kg/m3 ] and a kinematic viscosity of 0.544 106 [m2 /sec] and runner surface roughness of 0.01 [mm]. For the constant pressure case the liquid metal pressure at the runner entrance is assumed to be 1.2[MPa] and for the constant power case the power loss is [1Kw]. lling time that tmax t = V Qmax 2Oz P (7.27)
The gate velocity is exhibited as a function of the ratio of the gate area to the conduit area as shown in Figure ?? for a constant pressure and in Figure ?? for a constant power. General conclusions from example 7.7 For the constant pressure case the common18 assumption yields a constant velocity even for a zero gate area. The solid line in Figure ?? represents the gate velocity calculated based on the common assumption of constant CD while the other lines are based on calculations which take into account the runner geometry and the Re number. The results for constant CD represent averaged of the other results. The calculations of the velocity based on a constant CD value are unrealistic. It overestimates the velocity for large gate area and underestimates for the area ratio below 80% for the short runner and 35% for the long runner. Figure ?? exhibits that there is a clear maximum gate velocity which depends on the runner design (here represented by the conduit length). The maximum indicates that the preferred situation is to be on the right hand side branch because of shorter lling time. The gate velocity is doubled for the excellent design compared with the gate velocity obtained from the poor design. This indicates that the runner design is more important than the specic characteristic of the die casting machine performance. Operating the die casting machine in the right hand side results in smaller requirements on the die casting machine because of a smaller lling time, and therefore will require a smaller die casting machine. For the constant power case, the gate velocity as a function of the area ratio is shown in Figure ??. The common assumption of constant CD yields the gate velocity
17 read 18 As
93
U3 A1 /A3 shown by the solid line. Again, the common assumption produces unrealistic results, with the gate velocity approaching innity as the area ratio approaches zero. Obviously, the results with a constant CD over estimates the gate velocity for large area ratios and underestimates it for small area ratios. The other lines describe the calculated gate velocity based on the runner geometry. As before, a clear maximum can also be observed. For large area ratios, the gate velocity with an excellent design is almost doubled compared to the values obtained with a poor design. However, when the area ratio approaches zero, the gate velocity is insensitive to the runner length and attains a maximum value at almost the same point. In conclusion, this part has been shown that the use of the common pQ2 diagram with the assumption of a constant CD may lead to very serious errors. Using the pQ2 diagram, the engineer has to take into account the eects of the variation of the gate area on the discharge coecient, CD , value. The two examples given inhere do not represent the characteristics of the die casting machine. However, more detailed calculations shows that the constant pressure is in control when the plunger is small compared to the other machine dimensions and when the runner system is very poorly designed. Otherwise, the combination of the pressure and power limitations results in the characteristics of the die casting machine which has to be solved.
U3
pressure
efficiency
common model
realistic velocity
power
A3
Fig a.
power
Fig b.
Fig. -7.8. Velocity, U3 as a function of the gate area, A3 and the general characterstic of a pump
94
The die casting machine characteristic eects There are two type of operation of the die casting machine, one) the die machine is operated directly by hydraulic pump (mostly on the old machines). two) utilizing the non continuous demand for the power, the power is stored in a container and released when need (mostly on the newer machines). The container is normally a large tank contain nitrogen and hydraulic liquid19 . The eects of the tank size and gas/liquid ratio on the pressure and ow rate can easily be derived.
Meta
The power supply from the tank with can consider almost as a constant pressure but the line to actuator is with variable resistance which is a function of the liquid velocity. The resistance can be consider, for a certain range, as a linear function of the velocity square, UB 2 . Hence, the famous a assumption of the common die casting machine p Q2 .
Meta End
The characteristic of the various pumps have been studied extensively in the past [15]. The die casting machine is a pump with some improvements which are patented by dierent manufactures. The new congurations, such as double pushing cylinders, change somewhat the characteristics of the die casting machines. First let discuss some general characteristic of a pump (issues like impeller, speed are out of the scope of this discussion). A pump is mechanical devise that transfers and electrical power (mostly) into hydraulic power. A typical characteristic of a pump are described in Figure ??. Two similar pumps can be connect in two way series and parallel. The serious connection increase mostly the pressure as shown in Figure ??. The series connection if normalized is very close to the original pump. However, the parallel connection when normalized show a better performance. To study the eects of the die casting machine performances, the following Fig. -7.9. Various die casting machine perforfunctions are examined (see Figure 7.9): mances
1 0.8 0.6
0.4
P= 1 Q P= 1 Q P= 1 Q
0.2
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1P 1P 1P
19 This similar to operation of water system in a ship, many of the characteristics are the same. Furthermore, the same dierential equations are governing the situation. The typical questions such as the necessarily container size and the ratio of gas to hydraulic liquid were part of my study in high school (probably the simplied version of the real case). If demand to this material raised, I will insert it here in the future.
95
The functions (??), (??) and (??) represent a die casting machine with a poor performance, the common performance, and a die casting machine with an excellent performance, respectively. Combining equation (7.21) with (7.28) yields 1P = 1P =
4
1P =
rearranging equation (7.31) yields P 2 2(1 + Oz)P + 1 = 1 P (1 + 2Oz) = 4Oz P 2 + P 1 = 0 0 0 (7.34) (7.35) (7.36)
Solving equations (7.34) for P , and taking only the possible physical solution, yields P = P = P = 1 + Oz (2 + Oz) Oz 1 1 + 2 Oz 1 + 16 Oz 2 1 8Oz 2 (7.37) (7.38) (7.39)
The reduced pressure, P , is plotted as a function of the Oz number for the three die casting machine performances as shown in Figure 7.10. Figure 7.10 demonstrates that P monotonically decreases with an increase in the Oz number for all the machine performances. All the three results convert to the same line which is a plateau after Oz = 20. For large Oz numbers the reduced pressure, P , can be considered to be constant P 0.025. The gate velocity, in this case, is U3 0.22CD Pmax (7.40)
The Ozer number strongly depends on the discharge coecient, CD , and Pmax . The value of Qmax is relatively insensitive to the size of the die casting machine. Thus, this equation is applicable to a well designed runner (large CD ) and/or a large die casting machine (large Pmax ).
96
1.00 .. 0.90 0.80 0.70 0.60
P
0.50 0.40 0.30 0.20 0.10 0.00
0.00
3.00
6.00
9.00
12.00
15.00
18.00
21.00
24.00
27.00
30.00
Oz
Fig. -7.10. Reduced pressure, P , for various machine performances as a function of the Oz number.
The reduced pressure for a very small value of the Oz number equals to one, P 1 or Pmax = P1 , due to the large resistance in the runner (when the resistance in the runner approaches inn ity, KF , then P = 1). Hence, the gate velocity is determined by the approximation of
PB 2 D B PB 1
DR
atmospheric pressure P 1 D1
U3
CD
2Pmax
(7.41)
The dierence between the various machine performances is more considerable in the middle range of the Oz numbers. A better machine performance produces a higher reduced pressure, P . The preferred situation is when the Oz number is large and thus indicates that the machine performance is less important than the runner design parameters. This observation is further elucidated in view of Figures ?? and ??.
97
The pressure at the plunger tip can be evaluated from a balance forces acts on the hydraulic piston and plunger as shown in Figure 7.11. The atmospheric pressure that acting on the left side of the plunger is neglected. Assuming a steady state and neglecting the friction, the forces balance on the rod yields DR 2 D1 2 DB 2 (PB 1 PB 2 ) + PB 2 = P1 4 4 4 In particular, in the stationary case the maximum pressure obtains DB 2 DR 2 D1 2 (PB 1 PB 2 )|max + PB 2 |max = P1 |max 4 4 4 (7.43) (7.42)
The equation (7.43) is reduced when the rode area is negligible; plus, notice that P1 |max = Pmax to read DB 2 D1 2 (PB 1 PB 2 )|max = Pmax 4 4 Rearranging equation (7.44) yields DB D1
20 2
(7.44)
DB D1
(7.45)
The gate velocity relates to the liquid metal pressure at plunger tip according to the following equation combining equation (7.5) and (??) yields (PB 1 PB 2 )|max 1+
2 CD A3 Qmax 2 DB D1 2 2
U3 = CD
(7.46)
(PB 1 PB 2 )|max
DB D1
98
1.0 ... .. .. .. .. 0.9 .. .. .. .. 0.8 .. .. .. .. 0.7 .. .. . .. .. 0.6 .. .. .. .. 0.5 .. .. .. .. .. 0.4 .. .. .. .. 0.3 .. .. .. .. 0.2 .. ... .. 0.1 0.0 0.0 1.0 2.0 3.0
............
...
...
....
Fig. -7.12. Reduced liquid metal pressure at the plunger tip and reduced gate velocity as a function of the reduced plunger diameter.
Meta
the solution for the intersection point is given by equation ? To study equation (7.46), lets dene = (PB 1 PB 2 )|max Qmax C D A3 D1 DB (7.47)
= 2 Oz P
(7.50)
99
one can obtain from equation (??) that (make a question about how to do it?) = 2 1 +1 (7.51)
The coecients of P1 in equation (7.50) and D1 in equation (7.47) are assumed constant according to the common pQ2 diagram. Thus, the plot of y and as a function of represent the aect of the plunger diameter on the reduced gate velocity and reduced pressure. The gate velocity and the liquid metal pressure at plunger tip decreases with an increase in the plunger diameters, as shown in Figure 7.12 according to equations (7.49) and (7.51).
Meta End
A control volume as it is shown in Figure 7.13 is constructed to study the eect of the plunger diameter, (which includes the plunger with the rode, hydraulic piston, and shot sleeve, but which does not include the hydraulic liquid or the liquid metal jet). The control volume is stationary around the shot sleeve and is moving with the hydraulic piston. Applying the rst law of thermodynamics, when that the atmospheric pressure is assumed negligible and neglecting the dissipation energy, yields Uin Q + min hin + 2
2
= mout hout +
Uout 2
dm dt
e+
c.v.
Uc.v. 2
(7.52) + Wc.v.
In writing equation (7.52), it should be noticed that the only change in the control volume is in the shot sleeve. The heat transfer can be neglected, since the lling process is very rapid. There is no ow into the control volume (neglecting the air ow into the back side of the plunger and the Fig. -7.13. A general schematic of the control volume change of kinetic energy of the of the hydraulic piston with the plunger and part of the air, why?), and therefore the sec- liquid metal ond term on the right hand side can be omitted. Applying mass conservation on the control volume for the liquid metal yields
the moving part of the control volume UP liquid metal out atmospheric pressure air flow in PB 2 PB 1 friction is neglected
dm dt
= mout
c.v.
(7.53)
The boundary work on the control volume is done by the left hand side of the plunger and can be expressed by Wc.v. = (PB 1 PB 2 )AB U1 (7.54)
100
The mass ow rate out can be related to the gate velocity mout = A3 U3 Mass conservation on the liquid metal in the shot sleeve and the runner yields
2 A1 U1 = A3 = U1 2 = U3
(7.55)
A3 A1
(7.56)
Substituting equations (7.53-7.56) into equation (7.52) yields (PB 1 PB 2 )AB U3 A1 = A3 U3 (hout e) + Rearranging equation (7.57) yields (PB 1 PB 2 ) Solving for U3 yields
A 2 (PB 1 PB 2 ) A1B (hout e) 2 U3 2
A3 A1
(7.57)
AB U2 = (hout e) + 3 A1 2
A3 A1
(7.58)
U3 =
A3 A1
(7.59)
U3 =
(hout e)
2
(7.60)
A3 A1
U3 =
(7.61)
(7.62)
101
The expression (7.62) is a very complicated function of A1 . It can be shown that when the plunger diameter approaches innity, D1 (or when A1 ) then the gate velocity approaches U3 0. Conversely, the gate velocity, U3 0, when the plunger diameter, D1 0. This occurs because mostly K and CD 0. Thus, there is at least one plunger diameter that creates maximum velocity (see gure 7.14). A more detailed study shows that depending on the physics in the situation, more than one local maximum can occur. With a small plunger diameter, the gate velocity approaches zero because CD approaches innity. For a large plunger diameter, the gate velocity approaches zero because the pressure dierence acting on the runner is approaching zero. The mathematical expression for the maximum gate velocity takes several pages, and therefore is not shown here. However, for practical purposes, the maximum velocity can easily (relatively) be calculated by using a computer program such as DiePerfect. Machine size eect
U3 common model
The question how large the die casting machine depends on how ecient it is used. To maximized the utilization of the die casting machine we must understand under what condition it happens. It is important to realize that the injection of the liquid metal into the cavity requires power. The power, we can extract from a ma- Fig. -7.14. The gate velocity, U3 as a function chine, depend on the plunger velocity and of the plunger area, A1 other parameters. We would like to design a process so that power extraction is maximized. Lets dened normalized machine size eect
realistic velocity
po ssib le m ax
A1
pwrm =
QP Pmax Qmax
QP
(7.63)
Every die casting machine has a characteristic curve on the pQ2 diagram as well. Assuming that the die casting machine has the common characteristic, P = 1 Q2 , the normalized power can be expressed pwrm = Q(1 Q2 ) = Q2 Q3 (7.64)
where pwrm is the machine power normalized by Pmax Qmax . The maximum power of this kind of machine is at 2/3 of the normalized ow rate, Q, as shown in Figure ??. It is recommended to design the process so the ow rate occurs at the vicinity of the maximum of the power. For a range of 1/3 of Q that is from 0.5Q to 0.83Q, the average power is 0.1388 Pmax Qmax , as shown in Figure ?? by the shadowed rectangular. One may notice that this value is above the capability of the die casting machine in two ranges of the ow rate. The reason that this number is used is because with some
102
0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 . 0.00 .. .. .. 0.00 ... . ..
..
...
.....
. .... . .
...
0.10
0.20
0.30
0.40
0.50
0.60
0.70
Q
Fig. -7.15. The reduced power of the die casting machine as a function of the normalized ow rate.
improvements of the the runner design the job can be performed on this machine, and there is no need to move the job to a larger machine21 . Precondition eect (wave formation)
Meta
discussion when Q1 = Q3
Meta End
7.4.3 Poor design eects
Meta
discussed the changes when dierent velocities are in dierent gates. Expanded on the sudden change to turbulent ow in one of the branches.
Meta End
7.4.4 Transient eects
that requirements on the clamping force is meet.
Under construction
21 Assuming
7.5. DESIGN PROCESS To put the discussion about the inertia of the system and compressibility. the magnitude analysis before intensication eects
103
insert only general remarks until the paper will submitted for publication insert the notes from the yellow folder
(7.66)
the clamping force, and satisfy the power requirements. For these criteria the designer has to check the runner design to see if gate velocity are around the recommended range. A possible answer has to come from nancial considerations, since we are in the business of die casting to make money. Hence, the optimum diameter is the one which will cost the least (the minimum cost). How, then, does the plunger size determine cost? It has been shown that plunger diameter has a value where maximum gate velocity is created. A very large diameter requires a very large die casting machine (due to physical size and the weight of the plunger). So, one has to chose as rst approximation the largest plunger on a smallest die casting machine. Another factor has to be taken into consideration is the scrap created in the shot sleeve. Obviously, the liquid metal in the sleeve has to be the last place to solidify. This requires the biscuit to be of at least the same thickness as the runner. Trunner = Tbiscuit Therefore, the scrap volume should be D1 2 D1 2 Tbiscuit = Trunner 4 4 (7.68) (7.67)
104
When the scrap in the shot sleeve becomes signicant, compared to scrap of the runner D1 2 Trunner = LTrunner 4 Thus, the plunger diameter has to be in the range of D1 =
To discussed that the plunger diameter should not be use as varying the plunger diameter to determine the gate velocity
(7.69)
4 L
(7.70)
put schematic gure of how it is done from the patent by die casting companies
Meta
why? to put discussion
Analysis of the forces demonstrates that as rst approximation the plunger diameter does not contribute any additional force toward pushing the liquid metal.
Meta End
A very small plunger diameter creates faster solidication, and therefore the actual force is reduced. Conversely, a very large plunger diameter creates a very small pressure for driving the liquid metal.
7.7 Summary
In this chapter it has been shown that the common diagram is not valid and produces unrealistic trends therefore has no value what so ever23 . The reformed pQ2 diagram was introduced. The mathematical theory/presentation based on established scientic principles was introduced. The eects of various important parameters was discussed. The method of designing the die casting process was discussed. The plunger diameter
22 A note for the manufactures, if you would like to have your system described here with its advantages, please drop me a line and I will discuss with you about the material that I need. I will not charge you any money. 23 Beside the historical value
7.8. QUESTIONS
105
has a value for which the gate velocity has a maximum. For D1 0 gate velocity, U3 0 when D1 the same happen U3 0. Thus, this maximum gate velocity determines whether an increase in the plunger diameter will result in an increase in the gate velocity or not. An alternative way has been proposed to determine the plunger diameter.
7.8 Questions
106
Garber concluded that his model was not able to predict an acceptable value for critical velocity for ll percentages lower than 50% . . . Brevick, Ohio
107
108
was in the pre BarMeirs model. First, a description of Garbers model is given later Brevicks two models along with Millers model5 are described briey. Lastly, the EKKs numerical model is described.
8.2.1
Garbers model
1
The description in this section is based on one of the most cited paper in the die castv v ing research [17]. Garbers model deals h v only with a plug ow in a circular cross P h P section. In this section, we improve the model to include any geometry cross section with any velocity prole6 . Consider a duct (any cross section) Fig. -8.1. A schematic of wave formation in with a liquid at level h2 and a plunger mov- stationary coordinates ing from the left to the right, as shown in Figure 8.1. Assuming a quasi steady ow is established after a very short period of time, a unique height, h1 , and a unique wave velocity, Vw , for a given constant plunger velocity, Vp are created. The liquid in the substrate ahead of the wave is still, its height, h2 , is determined by the initial ll. Once the height, h1 , exceeds the height of the shot sleeve, H, there will be splashing. The splashing occurs because no equilibrium can be achieved (see Figure 8.2a). For h1 smaller than H, a reecting wave from the opposite wall appears resulting in an enhanced air entrainment (see Figure 8.2b). Thus, the preferred situation is when h1 = H (in circular shape H = 2R) in which case no splashing or a reecting wave result.
p w
2 H
L(t)
initial height
Fig a.
A schematic of
Fig b.
A schematic of
built wave.
reecting wave.
Fig. -8.2. The left graph depicts the common pQ2 version. The right graph depicts Pmax and Qmax as a function of the plunger diameter according to common model.
It is easier to model the wave with coordinates that move at the wave velocity, as shown in Figure 8.3. With the moving coordinate, the wave is stationary, the plunger moves back at a velocity (Vw Vp ), and the liquid moves from the right to the left. Dashed line shows the stationary control volume.
model was developed at Ohio State University by Millers Group in the early 1990s. addition to the original Garbers paper is derived here. I assumed that in this case, some mathematics will not hurt the presentation.
6 This 5 This
8.2. THE COMMON MODELS Mass conservation of the liquid in the control volume reads: Vw dA =
A2 A1
1
109
v=0 vw
P2 h2
where v1 is the local velocity. Under quasi-steady conditions, the corresponding Fig. -8.3. A schematic of the wave with moving average velocity equals the plunger veloc- coordinates ity: 1 A1 v1 dA = v1 = Vp
A1
(8.2)
What is justication for equation 8.2? Assuming that heat transfer can be neglected because of the short process duration7 . Therefore, the liquid metal density (which is a function of temperature) can be assumed to be constant. Under the above assumptions, equation (8.1) can be simplied to
hi
build a question about what happens if the temperature changes by a few degrees. How much will it aect equation 8.2 and other parameters?
A(hi ) =
0
dA
(8.3)
Where i in this case can take the value of 1 or 2. Thus, Vw = f (h12 ) (Vw Vp ) (8.4)
where f (h12 ) = A(h1 ) is a dimensionless function. Equation (8.4) can be transformed A(h2 ) into a dimensionless form: v ( 1) v f (h12 ) = v = f (h12 ) 1 f (h12 ) = (8.5) (8.6)
where v = Vw . Show that A(h1 ) = 2R2 for h1 = 2R Assuming energy is conserved Vp (the Garbers model assumption), and under conditions of negligible heat transfer, the energy conservation equation for the liquid in the control volume (see Figure 8.3) reads: PB E (Vw Vp )2 + (Vw Vp )dA = 2 P2 Vw 2 + Vw dA 2 (8.7)
A1
A2
where =
7 see
1 A1 (Vw Vp )3
(Vw v1 )3 dA =
A1
1 A1 ( 1)3 v
v
A1
v1 Vp
dA
(8.8)
110
underconstruction The shape factor, E , is introduced to account for possible deviations of the velocity prole at section 1 from a pure plug ow. Note that in die casting, the ow is pushed by the plunger and can be considered as an inlet ow into a duct. The typical Re number is 105 , and for this value the entry length is greater than 50m, which is larger than any shot sleeve by at least two orders of magnitude. The pressure in the gas phase can be assumed to be constant. The hydrostatic y pressure in the liquid can be represent by Rc g [28], where Rc is the center of the y cross section area. For a constant liquid density equation (8.7) can be rewritten as: Rc1 g + E y (Vw Vp )2 Vw 2 (Vw Vp )A(h1 ) = Rc2 g + y Vw A(h2 ) 2 2 (8.9)
Garber (and later Brevick) put this equation plus several geometrical relationships as the solution. Here we continue to obtain an analytical solution. Dening a dimensionless parameter F r as Fr = Rg , Vp 2 (8.10)
Utilizing denition (8.10) and rearranging equation (8.9) yields 2F rE yc1 + E ( 1)2 = 2F rE yc2 + v 2 v Solving equation (8.11) for F rE the latter can be further rearranged to yield: F rE = 2(c1 yc2 ) y (1+E )f (h12 ) E f (h12 )1 (8.12) (8.11)
Given the substrate height, equation (8.12) can be evaluated for the F rE , and the corresponding plunger velocity ,Vp . which is dened by equation (8.10). This solution will be referred herein as the energy solution.
8.2.2
Brevicks Model
The square shot sleeve Since Garbers model never work Brevick and coworkers go on a shing expedition in the uid mechanics literature to nd equations to describe the wave. They found in Lambs book several equations relating the wave velocity to the wave height for a deep liquid (water)8 . Since these equations are for a two dimensional case, Brevick and coworkers built it for a squared shot sleeve. Here are the equations that they used. The instantaneous height dierence (h = h1 h2 ) is given as h = h2
8I
V p + 1 2 gh2
h2
(8.13)
have checked the reference and I still puzzled by the equations they found?
8.2. THE COMMON MODELS This equation (8.13), with little rearranging, obtained a new form Vp = 2 The wave velocity is given by Vw = gh2 3 1+ h 2 h2 gh2 h1 1 h2
111
(8.14)
(8.15)
Brevick introduces the optimal plunger acceleration concept. By plotting the height and position of each incremental wave with time, their model is able to predict the stability of the resulting wave front when the top of the front has traveled the length of the shot sleeve.9 . They then performed experiments on this miracle acceleration 10 .
8.2.3
Probably, because it was clear to the authors that the previous model was only good for a square shot sleeve 11 . They say let reuse Garbers model for every short time steps and with dierent velocity (acceleration).
8.2.4
Miller and his student borrowed a two dimensional model under assumption of turbulent ow. They assumed that the ow is innite turbulence and therefor it is a plug ow12 . Since the solution was for 2D they naturally build model for a square shot sleeve13 . The mass balance for square shot sleeve Vw h2 = (Vw Vp )h1 Momentum balance on the same control volume yield PB (Vw Vp )2 P2 Vw 2 + (Vw Vp )h1 = + Vw h 2 2 2 and the solution of these two equations is F rmiller =
9 What
(8.16)
(8.17)
1 h1 2 h2
h1 +1 h2
(8.18)
an interesting idea?? Any physics? to say this is not good enough a fun idea, they also invented a new acceleration units cm/seccm. 11 It is not clear whether they know that this equations are not applicable even for a square shot sleeve. 12 How they comeout with this conclusion? 13 Why are these two groups from the same university and the same department not familiar with each others work.
10 As
112
A considerable amount of research work has been carried out on this wave, which is known in the scientic literature as the hydraulic jump. The hydraulic jump phenomenon has been studied for the past 200 years. Unfortunately, Garber, ( and later other researchers in die casting such as Brevick and his students from Ohio State University [8], [31])14 , ignored the previous research. This is the real reason that their model never works. Show the relative error created by Garbers model when the substrate height h2 is the varying parameter.
8.3.2
Brevicks models
square model There are two basic mistakes in this model, rst) the basic equations are not applicable to the shot sleeve situation, second) the square geometry is not found in the industry. To illustrate why the equations Brevick chose are not valid, take the case where 1 > h1 /h2 > 4/9. For that case Vw is positive and yet the hydraulic jump opposite to reality (h1 < h2 ). Improved Garbers model Since Garbers model is scientic erroneous any derivative that is based on it no better than its foundation15 .
8.3.3
Millers model
The ow in the shot sleeve in not turbulent16 . The ow is a plug ow because entry length problem17 . Besides all this, the geometry of the shot sleeve is circular. This mistake is discussed in the comparison in the discussion section of this chapter.
with these major mistakes NADCA under the leadership of Gary Pribyl and Steve Udvardy continues to award Mr. Brevick with additional grands to continue this research until now, Why? 15 I wonder how much NADCA paid Brevick for this research? 16 Unless someone can explain and/or prove otherwise. 17 see Chapter 3.
14 Even
113
8.3.4
This model based on numerical simulations based on the following assumptions: 1) the ow is turbulent, 2) turbulence was assume to be isentropic homogeneous every where (k model), 3) unspecied boundary conditions at the free interface (how they solve it with this kind of condition?), and 4) unclear how they dealt with the corner point in which plunger perimeter in which smart way is required to deal with zero velocity of the sleeve and known velocity of plunger. Several other assumptions implicitly are in that work18 such as no heat transfer, a constant pressure in the sleeve etc. According to their calculation a jet exist somewhere in the ow eld. They use the k model for a eld with zero velocity! They claim that they found that the critical velocity to be the same as in Garbers model. The researchers have found same results regardless the model used, turbulent and laminar ow!! One can only wonder if the usage of k model (even for zero velocity eld) was enough to produce these erroneous results or perhaps the problem lays within the code itself19 .
8.4.1
In this section the momentum conservation principle is applied on the control volume in Figure 8.3. For large Re ( 105 ) the wall shear stress can be neglected compared to the inertial terms (the wave is assumed to have a negligible length). The momentum balance reads: P B + M (Vw Vp )2 dA =
A1 A2
P2 + Vw 2 dA
(8.20)
where M =
18 This 19 see
1 A1 (Vw Vp )2
(Vw v1 )2 dA =
A1
1 A1 ( 1)2 v
v
A1
v1 Vp
dA
(8.21)
paper is a good example of poor research related to a poor presentation and text processing. remark on page 44 20 in the last 100 years
114
Given the velocity prole v1 , the shape factor M can be obtained in terms of v . The expressions for M for laminar and turbulent velocity proles at section 1 easily can be calculated. Based on the assumptions used in the previous section, equation (8.20) reads: Rc1 g + M (Vw Vp ) y
2
(8.22)
Rearranging equation (8.22) into a dimensionless form yields: f (h12 ) yc1 F r + M ( 1) v Combining equations (8.5) and (8.23) yields
f (h12 ) f (h12 )1 2 2
= yc2 F r + v 2
(8.23)
F rM =
M f (h12 )
f (h12 ) f (h12 )1
(8.24)
where F rM is the F r number which evolves from the momentum conservation equation. Equation (8.24) is the analogue of equation (8.12) and will be referred herein as the BarMeirs solution.
16.00 14.40 12.80 11.20 9.60 .......... momentum energy Miller Karni Garber .. .. .. .. . .. .. .. . .. .. .. . .. .. .. .. .. .. .. . ... .. ... .. ... ...
Fr
. . . ... ... . . . . . . . . . .... ... ... ... ... ... ... ... ... ... 0.17 0.33 0.49 0.66
.. . .. . . . ... .. . .. . .. . 0.83
.... ..
....
....
0.99
1.16
1.32
1.49
1.65
h2
R
115
ow (Poiseuille paraboloid), are discussed in the Appendix 22 . Note that the Froude number is based on the plunger velocity and not on the upstream velocity commonly used in the twodimensional hydraulic jump. The experimental data obtained by Garber , and Karni and the transition from the free surface ow to pressurized ow represented by equations (8.12) and (8.24) for a circular cross section are presented in Figure 8.4 for a plug ow. The Millers model (two dimensional) of the hydraulic jump is also presented in Figure 8.4. This Figure shows clearly that the BarMeirs solution is in agreement with Karnis experimental results. The agreement between Garbers experimental results and the BarMeirs solution, with the exception of one point (at h2 = R), is good. The experimental results obtained by Karni were taken when the critical velocity was obtained (liquid reached the pipe crown) while the experimental results from Garber are interpretation (kind of average) of subcritical velocities and supercritical velocities with the exception of the one point at h2 /R = 1.3 (which is very closed to the Bar Meirs solution). Hence, it is reasonable to assume that the accuracy of Karnis results is better than Garbers results. However, these data points have to be taken with some caution23 . Non of the experimental data sets were checked if a steady state was achieved and it is not clear how the measurements carried out. It is widely accepted that in the two dimensional hydraulic jump small and large eddies are created which are responsible for the large energy dissipation [19]. Therefore, energy conservation cannot be used to describe the hydraulic jump heights. The same can be said for the hydraulic jump in dierent geometries. Of course, the same has to be said for the circular cross section. Thus, the plunger velocity has to be greater than the one obtained by Garbers model, which can be observed in Figure 8.4. The Froude number for the Garbers model is larger than the Froude number obtained in the experimental results. Froude number inversely proportional to square of the plunger velocity, F r 1/Vp 2 and hence the velocity is smaller. The Garbers model therefore underestimates the plunger velocity.
8.4.2
Design process
To obtain the critical slow plunger velocity, one has to follow this procedure: 1. Calculate/estimate the weight of the liquid metal. 2. Calculate the volume of the liquid metal (make sure that you use the liquid phase property and not the solid phase). 3. Calculate the percentage of lling in the shot sleeve, 4. Find the F r number from Figure 8.4. 5. Use the F r number found to calculate the plunger velocity by using equation (8.10).
22 To 23 Results
height . r
appear in the next addition. of good experiments performed by serious researchers are welcome.
116
8.5 Summary
In this Chapter we analyzed the ow in the shot sleeve and developed a explicit expression to calculated the required plunger velocity. It has been shown that Garbers model is totally wrong and therefore Brevicks model is necessarily erroneous as well. The same can be said to all the other models discussed in this Chapter. The connection between the wave and the hydraulic jump has been explained. The method for calculating the critical slow plunger velocity has been provided.
8.6 Questions
9.1 Introduction
Proper design of the venting system is one of the requirements for reducing air/gas porosity. Porosity due to entrainment of gases constitutes a large portion of the total porosity, especially when the cast walls are very thin (see Figure ??). The main causes of air/gas porosity are insucient vent area, lubricant evaporation (reaction processes), incorrect placement of the vents, and the mixing processes. The present chapter considers the inuence of the vent area (in atmospheric and vacuum venting) on the residual gas (in the die) at the end of the lling process. Atmospheric venting, the most widely used casting method, is one in which the vent is opened to the atmomaximum sphere and is referred herein as air ventshrinkage porosity porosity ing. Only in extreme cases are other solutions required, such as vacuum venting, Pore Free Technique (in zinc and aluminum casting) and squeeze casting. Vacwall thickness uum is applied to extract air/gas from the mold before it has the opportunity to mix Fig. -9.1. The relative shrinkage porosity as a with the liquid metal and it is call vacuum function of the casting thickness. venting. The Pore Free technique is a variation of the vacuum venting in which the oxygen is introduced into the cavity to replace the air and to react with the liquid metal, and therefore creates a vacuum [5]. Squeeze
117
118
casting is a dierent approach in which the surface tension is increased to reduce the possible mixing processes (smaller Re number as well). The gases in the shot sleeve and cavity are made mostly of air and therefore the term air is used hereafter. These three solutions are cumbersome and create a far more expensive process. In this chapter, a qualitative discussion on when these solutions should be used and when they are not needed is presented. Obviously, the best ventilation is achieved when a relatively large vent area is designed. However, to minimize the secondary machining (such as trimming), to ensure freezing within the venting system, and to ensure breakage outside the cast mold, vents have to be very narrow. A typical size of vent thicknesses range from 12[mm]. These conicting requirements on the vent area suggest an optimum area. As usual the common approach is described the errors are presented and the reformed model is described.
9.2.2
Millers model
Miller and his student, in the early 90s, constructed a model to account for the friction in the venting system. They based their model on the following assumptions: 1. No heat transfer
1 Apparently,
119
2. Isothermal ow (constant temperature) in the entrance to vent (according to the authors in the presentation) 3. Fanno ow in the rest vent 4. Air/gas obeys the ideal gas model Miller and his student described the calculation procedures for the two case as choked and unchoked conditions. The calculations for the choked case are standard and can be found in any book about Fanno ow but with an interesting twist. The conditions in the mold and the sleeve are calculated according the ambient condition (see the smart quote of this Chapter)2 . The calculations about unchoked case are very interesting and will be discussed here in a little more details. The calculations procedure for the unchoked as the following:
Assume Min number (entrance Mach number to the vent) lower than Min for choked condition Calculate the corresponding star (choked conditions) the temperature ratio for the assume Min number Calculate the dierence between the calculated
4f L D 4f L D ,
Use the dierence 4f L to calculate the double stars (theoretical exit) conditions D based on the ambient conditions. Calculated the conditions in the die based on the double star conditions.
Now the mass ow out is determined by mass conservation. Of course, these calculations are erroneous. In choked ow, the conditions are determined only and only by upsteam and never by the down steam3 . The calculations for unchoked ow are mathematical wrong. The assumption made in the rst step never was checked. And mathematically speaking, it is equivalent to just guessing solution. These errors are only fraction of the other other in that model which include among other the following: one) assumption of constant temperature in the die is wrong, two) poor assumption of the isothermal ow, three) poor measurements etc. On top of that was is the criterion for required vent area.
120
1) the ow rate out is less than the volume pushed by the piston, 2) the ow rate out is more than the volume pushed by the piston, or 3) the ow rate out is equal to the volume pushed by the piston. The last case is called the critical design, and it is associated with the critical area. Air ows in the venting system can reach very large velocities up to about 350 [m/sec]. The air cannot exceed this velocity without going through a specially congured conduit (converging diverging conduit). This phenomena is known by the name of choked ow. This physical phenomenon is the key to understanding the venting design process. In air venting, the venting system has to be designed so that air velocity does not reach the speed of sound: in other words, the ow is not choked. In vacuum venting, the air velocity reaches the speed of sound almost instantaneously, and the design should be such that it ensures that the air pressure does not exceed the atmospheric pressure. Prior models for predicting the optimum vent area did not consider the resistance in the venting system (pressure ratio of less than 2). The vent design in a commercial system includes at least an exit, several ducts, and several abrupt expansions/contractions in which the resistance coecient, 4f L , is of the order of 37 or D more. Thus, the pressure ratio creating choked ow is at least 3. One of the dierences between vacuum venting and atmospheric venting occurs during the startup time. For vacuum venting, a choking condition is established almost instantaneously (it depends on the air volume in the venting duct), while in the atmospheric case the volume of the air has to be reduced to more than half (depending on the pressure ratio) before the choking condition develops - - and this can happen only when more than 2/3 or more of the piston stroke is elapsed. Moreover, the ow is not necessarily choked in atmospheric venting. Once the ow is choked, there is no dierence in calculating the ow between these two cases. It turns out that the mathematics in both cases are similar, and therefore both cases are presented in the present chapter. The role of the chemical reactions was shown to be insignicant. The dierence in the gas solubility (mostly hydrogen) in liquid and solid can be shown to be insignicant [1]. For example, the maximum hydrogen release during solidication of a kilogram of aluminum is about 7cm3 at atmospheric temperature and pressure. This is less than 3% of the volume needed to be displaced, and can be neglected. Some of the oxygen is depleted during the lling time [5]. The last two eects tend to cancel each other out, and the net eect is minimal. The numerical simulations produce unrealistic results and there is no other quantitative tools for nding the vent locations (the last place(s) to be lled) and this issue is still an open question today. There are, however, qualitative explanations and reasonable guesses that can push the accuracy of the last place (the liquid metal reaches) estimate to be within the last 10%30% of the lling process. This information increases the signicance of the understanding of what is the required vent area. Since most of the air has to be vented during the initial stages of the lling process, in which the vent locations do not play a role. Air venting is the cheapest method of operation, and it should be used unless acceptable results cannot be obtained using it. Acceptable results are dicult to obtain
121
1) when the resistance to the air ow in the mold is more signicant than the resistance in the venting system, and 2) when the mixing processes are augmented by the specic mold geometry. In these cases, the extraction of the air prior to the lling can reduce the air porosity which require the use of other techniques. An additional objective is to provide a tool to combine the actual vent area with the resistance (in the venting system) to the air ow; thus, eliminating the need for calculations of the gas ow in the vent in order to minimize the numerical calculations. Hu et al. and others have shown that the air pressure is practically uniform in the system. Hence, this analysis can also provide the average air pressure that should be used in numerical simulations.
With the above assumptions, the following model as shown in Figure ?? is proposed. A plunger pushes the liquid metal, and both of them (now called as the piston) propel the air through a long, straight conduit.
122
The mass balance of the air in the cylinder yields dm + mout = 0. dt (9.1)
Fig. -9.2. A simplied model for the venting system.
This equation (9.1) is the only equation that needed to be solved. To solve it, the physical properties of the air need to be related to the geometry and the process. According to assumption 4, the air mass can be expressed as m=
PV RT
(9.2)
The volume of the cylinder under assumption 6 can be written as V (t) = V (0) 1 t tmax (9.3)
(9.4)
The lling process occurs within a very short period time [milliseconds], and therefore the heat transfer is insignicant 3. This kind of ow is referred to as Fanno ow4 . The instantaneous ow rate has to be expressed in terms of the resistance to the ow, 4f L , the pressure ratio, and the characteristics of Fanno ow [29]. Knowledge of D Fanno ow is required for expressing the second term in equation (9.1). The mass ow rate can be written as mout = P0 (0)AMmax where k1 2 f [Min (t)] = 1 + (Min (t)) 2
(k+1) 2(k1)
P0 (0) P0 (t)
k+1 2k
(9.5)
(9.6)
The Mach number at the entrance to the conduit, Min (t), is calculated by Fanno ow characteristics for the venting system resistance, 4f L , and the pressure ratio. Mmax is D
4 Fanno ow has been studied extensively, and numerous books describing this ow can be found. Nevertheless, a brief summary on Fanno ow is provided in Appendix A.
123
the maximum value of Min (t). In vacuum venting, the entrance Mach number, Min (t), is constant and equal to Mmax . Substituting equations (9.4) and (9.5) into equation (9.1), and rearranging, yields: k 1 dP = dt
tmax k1 2k tc M f (Min )P
1t
P;
P (0) = 1.
(9.7)
The solution to equation (9.7) can be obtained by numerical integration for P . The residual mass fraction in the cavity as a function of time is then determined using the ideal gas assumption. It is important to point out the signicance of the tmax . This tc parameter represents the ratio between the lling time and the evacuation time. tc is the time which would be required to evacuate the cylinder for a constant mass ow rate at the maximum Mach number when the gas temperature and pressure remain at their initial values, under the condition that the ow is choked, (The pressure dierence between the mold cavity and the outside end of the conduit is large enough to create a choked ow.) and expressed by tc = m(0) AMmax P0 (0)
k RT0 (0)
(9.8)
Critical condition occurs when tc = tmax . In vacuum venting, the volume pushed by the piston is equal to the ow rate, and ensures that the pressure in the cavity does not increase (above the atmospheric pressure). In air venting, the critical condition ensures that the ow is not choked. For this reason, the critical area Ac is dened as the area that makes the time ratio tmax /tc equal to one. This can be done by looking at equation (9.8), in which the value of tc can be varied until it is equal to tmax and so the critical area is Ac = m(0) tmax Mmax P0 (0)
k RT0 (0)
(9.9)
Substituting equation (9.2) into equation (9.9), and using the fact that the sound velocity can be expressed as c = kRT , yields: Ac = V (0) ctmax Mmax (9.10)
where c is the speed of sound at the initial conditions inside the cylinder (ambient conditions). The tmax should be expressed by Eckert/BarMeir equation.
124
A Parameters inuencing the process are the area ratio, Ac , and the friction parameter, 4f L . From other detailed calculations [4] it was found that the inuence of the D parameter 4f L on the pressure development in the cylinder is quite small. The inuence D is small on the residual air mass in the cylinder, but larger on the Mach number, Mexit . A The eects of the area ratio, Ac , are studied here since it is the dominant parameter. Note that tc in air venting is slightly dierent from that in vacuum venting [3] by A a factor of f (Mmax ). This factor has signicance for small 4f L and small Ac when the D Mach number is large, as was shown in other detailed calculations [4]. The denition chosen here is based on the fact that for a small Mach number the factor f (Mmax ) can be ignored. In the majority of the cases Mmax is small. A For values of the area ratio greater than 1.2, Ac > 1.2, the pressure increases the volume ow rate of the air until a quasi steadystate is reached. In air venting, this quasi steadystate is achieved when the volumetric air ow rate out is equal to the volume pushed by the piston. The pressure and the mass ow rate are maintained constant A after this state is reached. The pressure in this quasi steadystate is a function of Ac . A A For small values of Ac there is no steadystate stage. When Ac is greater than one A the pressure is concave upwards, and when Ac is less than one the pressure is concave downwards. These results are in direct contrast to previous molds by Sachs , Draper , Veinik and Lindsey and Wallace , where models assumed that the pressure and mass ow rate remain constant and are attained instantaneously for air venting. To refer to the stroke completion (100% of the stroke) is meaningless since 1) no gas mass is left in the cylinder, thus no pressure can be measured, and 2) the vent can be blocked partially or totally at the end of the stroke. Thus, the completion (end of the process) of the lling process is described when 90% of the stroke is elapsed. Figure ?? presents the nal presA sure ratio as a function Ac for 4f L = 5. Fig. -9.3. The pressure ratios for air and vacD The nal pressure (really the pressure ra- uum venting at 90% of the piston stroke. A tio) depends strongly on Ac as described in Figure ??. The pressure in the die cavity increases by about 85% of its initial value A A when Ac = 1 for air venting. The pressure remains almost constant after Ac reaches A the value of 1.2. This implies that the vent area is suciently large when Ac = 1.2 for A air venting and when Ac = 1 for vacuum venting. Similar results can be observed when the residual mass fraction is plotted. This discussion and these results are perfectly correct in a case where all the assumptions are satised. However, the real world is dierent and the assumptions have to examined and some of them are:
20.0 18.0 16.0 14.0 Vacuum venting Air venting ......... 12.0 P (t =0.9) 10.0 P (t =0) 8.0 6.0 4.0 2.0 0.0 4fL =5.0 D .. .. .. .. .. .. .. .. ... .. . . . .. . .. . ... . ... ... ... ... ... . . . . . ... ... ... . . . ... . .... ... .. . . .. .. ... .. . . . . . . . 1.2 1.5 A Ac 1.8 2.1 2.4 2.7 0.0 0.3 0.6 0.9 3.0
1. Assumption 1 is not a restriction to the model, but rather guide in the design. The engineer has to ensure that the resistance in the mold to air ow (and metal ow) has to be as small as possible. This guide dictates that engineer designs the
9.5. RESULTS AND DISCUSSION path for air (and the liquid metal) as as short as possible.
125
2. Assumptions 3, 4, and 10 are very realistic assumptions. For example, the error in using assumption 4 is less than 0.5%. 3. This model is an indication when assumption 5 is good. In the initial stages (of the lling process) the pressure is very small and in this case the pressure (force) to open the plates is small, and therefore the gap is almost zero. As the lling process progresses, the pressure increases, and therefore the gap is increased. A signicant gap requires very signicant pressure which occurs only at the nal A stages of the lling process and only when the area ratio is small, Ac < 1. Thus, this assumption is very reasonable. 4. Assumption 6 is associated with assumption 9, but is more sensitive. The change in the resistance (a change in assumption in 9 creates consequently a change in the plunger velocity. The plunger reaches the constant velocity very fast, however, this velocity decrease during the duration of the lling process. The change again depends on the resistance in the mold. This can be used as a guide by the engineer and enhances the importance of creating a path with a minimum resistance to the ow. 5. Another guide for the venting system design (in vacuum venting) is assumption 7. The engineer has to reduce the vent volume so that less gas has to be evacuated. This restriction has to be design carefully keeping in mind that the resistance also has to be minimized (some what opposite restriction). In air venting, when this assumption is not valid, a dierent model describes the situation. However, not fullling the assumption can improve the casting because larger portion of the liquid metal which undergoes mixing with the air is exhausted to outside the mold. 6. Assumption 8 is one of the bad assumptions in this model. In many cases there is more than one vent, and the entrance Mach number for dierent vents could be a dierent value. Thus, the suggested method of conversion is not valid, and therefore the value of the critical area is not exact. A better, more complicated model is required. This assumption cannot be used as a guide for the design since as better venting can be achieved (and thus enhancing the quality) without ensuring the same Mach number. 7. Assumption 9 is a partially appropriate assumption. The resistance in venting system is a function of Re and Mach numbers. Yet, here the resistance, 4f L , D is calculated based on the assumption that the Mach number is a constant and equal to Mmax . The error due to this assumption is large in the initial stages where Re and Mach numbers are small. As the lling progress progresses, this error is reduced. In vacuum venting the Mach number reaches the maximum instantly and therefore this assumption is exact. The entrance Mach number is very small (the ow is even not choke ow) in air venting when the area ratio, A Ac >> 1 is very large and therefore the assumption is poor. However, regardless
126
CHAPTER 9. VENTING SYSTEM DESIGN the accuracy of the model, the design achieves its aim and the trends of this model are not aected by this error. Moreover, this model can be improved by taking into consideration the change of the resistance.
8. The change of the vent area does aect the resistance. However, a detailed calculation can show that as long as the vent area is above half of the typical cross section, the error is minimal. If the vent area turns out to be below half of the typical vent cross section a improvement is needed.
9.6 Summary
This analysis (even with the errors) indicates there is a critical vent area below which the ventilation is poor and above which the resistance to air ow is minimal. This critical area depends on the geometry and the lling time. The critical area also provides a mean to combine the actual vent area with the vent resistance for numerical simulations of the cavity lling, taking into account the compressibility of the gas ow. Importance of the design also was shown.
9.7 Questions
Under construction
d dt
h dV =
V A
hvi dA +
A
T dA n
(10.1)
1 A discussion on the mathematical aspects are left out. If explanation on this point will be asked by readers I will added it.
127
128
CHAPTER 10.
vi dA
(10.2)
The equations (10.1) and (10.2) do not have any restrictions of the liquid movement which has to be solved separately. Multiply equation (10.1) by a constant hl results in d hl dV = hl vi dA (10.3) dt V A Subtraction equation (10.3) from equation (10.1) yields d dt (h hl ) dV =
V A
(h hl )vi dA +
A
T dA n
(10.4)
The rst term on the right hand side composed from two contributions: one) from the liquid side and two) from solid side. At the solid side the contribution is vanished because (h hl )vi is zero due to vi is identically zero (no movement of the solid, it is a good assumption). In the liquid phase the term h hl is zero (why? ) thus the whole term is vanished we can write the identity (h hl )vi dA 0
A
(10.5)
where vi is the velocity at the interface. The rst term of equation (10.4 ) can be expressed in the term of the c.v.2 as
solid liquid=0
s A2 (hs hl ) s A1 (hs hl ) + ( (hl hl )) dt V dx = (s (hs hl )) = s (hs hl )vn (10.6) dt liquid side contribution is zero since h hl 0 and the solid contribution appears only in transitional layer due to transformation liquid to solid. The second term on right hand side of equation (10.4) is simply (h hl ) dV = k
A
d dt
T T T dA = ks kl n n n
(10.7)
It is noteworthy that the front propagation is about 10previously was calculated. Equation (10.7) holds as long as the transition into solid is abrupt (sharp transition).
2 please
note some dimensions will canceled each other out and not enter into equationsssss
129
Meta
For the case of where the transition to solid occurs over temperature range we have create three zones. Mathematically, it is convenient to describe the the mushy zone boundaries by two boundary conditions.
The creation of voids is results of density changes which change the heat transfer mechanism from conduction to radiation. The location of the void depends on the crystallization and surface tension eect, etc. The possibility of the liquid channels and the ow of semi-solid and even solid compensate for this void.
Kleins paper
Meta
Yet, one has to take into consideration the pressure eect The liquidation temperature and the latent heat are aected somewhat by the pressure. At pressure between the atmospheric to typical intensication pressure the temperature and latent heat are eected very mildly. However, for pressure near vacuum the latent heat and the temperature are eected more noticeably.3
Meta End
The velocity of the liquid metal due to the phase change can be related to the front propagation utilizing the equation (10.2). The left hand side can be shown to be (s l )vn . The right hand side is reduced into only liquid ow and easily can be shown to be l vl . (s l ) vn + l vl = 0 vl ( 1) = vn where is the density ratio, s /l .
(10.9)
3 I have used Clapyrons equation to estimate the change in temperature to be over 10 degrees (actually about 400 [C]). However, I am not sure of this calculations and I had not enough time to check it in the literature. If you have any knowledge and want to save me a search in the library, please drop me a line.
130
CHAPTER 10.
131
132
It doesnt matter on what machine the product is produce, the price is the same Prof. Al Miller, Ohio
133
134
tion quantity, and 9)secondary machining. After statistical analysis they have done they comeout with the following equation price = 0.485 + 2.20weight 0.505zinc + 0.791mag + 0.292details +0.637tolerance 0.253quantity where mag, zinc, details(<100 dimension), and tolerance are on/o switch. They claim that this formula is good for up to ten pounds (about 4.5[kg]). In summary, if you expect to get equation that does not have much with the actual cost, you got one. (12.1)
scrap cost
machine cost
m3
m1 > m2 > m3
HD
2 The price of a die casting machine increase almost exponentially with the machine size. Thus, nding the smallest die casting machine to run the job is critical importance. 3 The change in the parts numbers per shot will be discussed in section ??.
135
cost components: 1) those which the engineer controls, and 2) those which the engineer does not control. The uncontrolled components include overhead, secondary operations, marketing, space4 , etc. This category should be considered as a constant, since the engineers actions/choices do not aect the cost and therefore do not aect the cost of design decisions. However, the costs of die casting machine capital and operations, personnel cost, melting cost, and scrap cost5 are factors which have to be considered, and are discussed in the succeeding sections. In this analysis it is assumed that the die casting company is here to make a buck, and it is also assumed that competitive price wars for a specic project and/or any other personal reasons inuencing decision making are not relevant6 . This issue is formulated in such a way that the engineer will have the needed tools to make appropriate decisions.
(12.2)
136
In this analysis, it is assumed (or at least hoped) that the other die casting machines have other jobs waiting for them. If a company for a short time is not working to full capacity, the analysis will still be valid with minor modications. However, a longer duration of being below full capacity requires the company to make surgical solutions. The cost of the die casting machine depends on the market and not on the value the accountant has put on the books for that machine. Clearly, if the machine is to be sold/leased, the value obtained will be according to the market as average value. The market value should be used since the machine can be sold and this money can be invested in other possibilities. Amortization is estimated in the same manner. The dierence between the current value and the value at one year older is the depreciation value.9 Having these numbers, the capital cost can be estimated. For example, a one million dollar machine with a 20% money cost and a 5% depreciation cost equals about $250,000.00 a year. To convert this number to an hourly base rate, the number of idle days (on that specic machine) is required, and in many case is about 60 65 days. Thus, hourly capital cost of that specic machine is about $34.70. A change in the capital cost per unit can be via the change in the cycle time. The change in the cycle time is determined mostly by the solidication processes, which are controlled slightly by the runner design. Yet this eect can be diminished by controlling the cooling rate. Hence, the capital price is virtually unaected once the die casting machine has been selected for a specic project. Here, the cost per unit can be expressed as follows: capital = capital cost per hour Nc Np (12.3)
where capital is the capital cost per unit produced, Nc is the number of cycles per hour, and Np the number of parts shot.
137
less than 10[kWhour]. Thus, the total energy cost is in most cases at most $1.00 per hour. The change in the energy is insensitive to the runner and venting system designs and can vary by only 30% (15 cents for a very very large job), which is insignicant compared to all other components. The operation cost can be expressed as operation = operation cost per hour f(machine size, type etc.) Nc Np (12.4)
Vrunner
HD T 2 = LT 4
length
0.10 0.08
(12.5)
..
where HD is the typical size of the hyQ draulic diameter, and LT its length (these values are not the actual values, but they are used to represent the sizes of the run- Fig. -12.2. The reduced power of the die casting ner). From equation (12.5), it is clear that machine as a function of the normalized ow the diameter has one of the greater im- rate. pacts on the scrap cost. The minimum diameter at which a specic machine can produce good quality depends on the required lling time, gate velocity, other runner design characteristics, and the characteristics of the specic machine. Scrap cost is a linear function of the volume10 . The scrap cost per volume/weight consists of three components: 1) the melting cost, 2) the dierence between the buying price and the selling price (assuming that the scrap can be sold), and 3) the handling
0.10 0.20 0.30 0.40 0.50 0.60 0.70 1.00
10 Up
138
cost. The melting cost includes the cost to raise the metal temperature to the melting point, to melt the metal, and to hold the metal temperature above the melting point. The melting cost can be calculated by measuring energy used (crude oil or natural oil in most cases) plus the maintenance cost of the furnace divided by the amount of metal that has been casted (the parts and design scrap). The buying price is the price paid for the raw material; the selling price is the price for selling the scrap. Sometime it is possible to reuse the scrap and to remelt the metal. In some instances, the results of reusing the scrap will be a lower grade of metal in the end product. If reuse is possible, the dierence in cost should be substituted by the lost metal cost, which is the cost of 1) metal that cannot be recycled and 2) metal lost due to the chemical reactions in the furnace. The handling cost is the cost encountered in selling the metal, and it includes changing the mechanical or chemical properties of the scrap, transportation, cost of personnel, storage, etc. Each handling of the metal costs a dierent amount, and the specics can be recorded for the specic metal. Every job/mold has typical ranges for the lling time and gate velocity. Moreover, a rough design for the runner system can be produced for the mold. With these pieces of information in place, one can calculate the gate area (see pQ2 diagram calculations in Chapter 7 for more details, and this part is repeated in that Chapter. I am looking for the readers input to decided what is the best presentation.). Then the ow rate for the mold can be calculated by Q= Vmold Agate Vgate (12.6)
Additionally, the known design of the runner with ow rate yields the pressure dierence in the runner, and this yields the power required for the runner system, Pr = QP or in normalized form, Pr = QP Pmax Qmax QP (12.8) (12.7)
Every die casting machine has a characteristic curve on the pQ2 diagram as well. 2 Assuming that the die casting machine has the common characteristic, P = 1 Q , the normalized power can be expressed Pm = Q(1 Q ) = Q Q
2 2 3
(12.9)
where Pm is the machine power normalized by Pmax Qmax . The maximum power of this kind of machine is at 2/3 of the normalized ow rate, Q, as shown in Figure ??. It is recommended to design the process so the ow rate occurs at the vicinity of the maximum of the power. For a range of 1/3 of Q that is from 0.5Q to 0.83Q, the average power is 0.1388 Pmax Qmax , as shown in Figure ?? by the shadowed rectangular. One may notice that this value is above the capability of the die casting machine in two
139
ranges of the ow rate. The reason that this number is used is because with some improvements of the the runner design the job can be performed on this machine, and there is no need to move the job to a larger machine. If the machine power turns out to be larger than the required power of the runner, Pm > Cs Pr , the job can then be performed on the machine; otherwise, a bigger die casting machine is required. In general, the number of molds castable in a single cycle is given by Np = Pm Cs Pr = Pm Cs Pr (12.10)
The oor symbol being used means that the number is to be rounded down to the nearest integer. Cs denotes the safety factor coecient. In the case that Np is less than one, Np > 1, that specic machine is too small for this specic job. After the number of the parts has been determined (rst approximation) the runner system has to be redesigned so that the required power needed by the runner can be calculated more precisely. Plugging the new numbers into equation (12.10) yields a better estimation of the number of parts. If the number does not change, this is the number of parts that can be produced; otherwise, the procedure must be repeated. In this analysis, the required clamping forces that the die casting machine can produced are not taken into consideration. Analysis of the clamping forces determines the number of possible parts and it is a dierent criterion which required to satised, this will be discussed in more detail in Chapter 11. The actual number of parts that has to be taken into consideration is the smaller of the two criteria. Next, the new volume of the runner system has to be calculated. The cost per cavity is the new volume divided by the number of cavities: scrap = Vrunner (cost per volume) Np (12.11)
140
the shape is, the more likely it is that the number of attempted shots will increase. If it is assumed that the engineer is experienced, the only factor that will aect the number of shots will be the complexity provided that the job can be performed on the same die casting machine. The complexity of the shape should present a general idea of the number of expected attempts, and should be used in calculating the startup cost, startU p = (Cost per attempt) Na Nr (12.12)
where Na is the number of attempts, and Nr is the number of the total parts to be produced.
In toadys market, the operator cost is in the range of $10$20 per hour. When automatization is used, the personnel cost is signicantly reduced to the point that it is insignicant.
141
The ordering occurs once a year with 364 days of stocking cost which must be added.
the weight for the total of the hard drives [in kilograms]. the number is not constant or even seasonal (depend on the season) this model can be expanded by numerical analysis. 13 It is assumed that the year is 365 days which about 75% of the occurrence. There are years with 356 days.
12 If 11 Or
142
750
(12.I.b)
On the rst day has 364 portions (days) to be stocked. On the second day has 363 portions (days) to be stocked. On every sequence day there is one less portions to be stocked. Hence, there is a series of 364 items which starts with 364 and end at zero. This exactly algebraic series which can be calculated as a1 + a364 364 + 0 total = 364 = 364 = 66248 portions 2 2 Thus, the total cost for ordering once is total cost for once = $150 + 66248 0.75 = $49836 ordering (12.I.c)
(12.I.d)
For the purpose of this example, the choice of one time order is better. While the cost of one time ordering looks better in this example, in real life other factors should be considered. In this case case, the erosion of the company credit which was not a considered here. That fact should reversed the decision.
End Solution
Example 12.1 exhibits the eect of the number of orders on the cost of operation. Intuitively, for example, it can be observed that for two (2) orders, as compare to one order, the cost is reduced by half plus $150 (why?). The focus of this discussion is to nd the optimum number of times to order the items per year. Not all years are the same (some years are 366 days) but here it will be assumed that all years are 365 days. In this discussion, months and weeks do not appear and the dissimilarity is not discussed. The number of items ordered per year is assumed to be constant for this discussion and is denoted as N . Hence, the number of order items per day is N/365. The cost per day of store of item is denoted y. The ordering cost is r. The unknown number of periods is denoted as p. The number of days per period is D= 365 p (12.14)
The period cost of storage is the number parts not used which were kept for the following days of the period. The number of items remained for the storage for the rst day of the period are Remainder = (D 1) N 365 (12.15)
12.11. MINIMIZING COST OF SINGLE OPERATION On the second day, the number of items that remain in storage is Remainder = (D 2) N 365
143
(12.16)
There are no items to stored on the last day of the period. The number of days that we had to store is (D 1) (no need to store at the last day). The cost of the storage per period is
rst day second day 3rd last day day before last last day
N 365
+0
N 365 (12.17)
Equation (12.17) can be rearranged (because it is a regular algebraic series) is N number of = 365 items or
averaged to be stored days number of days D1
D1i
i=1
(12.18)
(D 1)
N D (D 1) N = 365 2 365
(12.19)
(12.20)
Or in terms of the number of period, p (see also equation (12.14)) the storage cost is storage cost = per period The yearly storage cost is N y 365
(12.21)
yearly storage = storage cost p = N per period 2 cost The total cost yearly ordering cost, C, is C=
(12.22)
(12.23)
144
In term of the number of periods the total ordering cost is yearly cost = rp+ N 2 365 1 p y (12.24)
The minimum cost expressed by the expression (12.24) a derivative with respect to p number of period as C N y 365 =r+ p 2 The solution of equation (12.25) is p= N y 365 2r (12.26) 1 p2 =0 (12.25)
These calculations were made under the assumption that the number of periods is a real number. However, the number of periods and several other parameters must be an integer. It can be argued that the number of orders can be 6.5 on the account that in two years planning to have 13 orders (13/2 = 6.5) It is more common to have solution with a totally irrational number which leads in practicality solution that cannot be used. The real solution (in a yearly integer planning sense) lay either on one of adjoining sides of the continuous solution. It has to be manually calculated. The calculation can yield a number of periods to be below one. The actual meaning is that the ordering cost is signicant (dominate) so the order must be continuous. On the other extreme, when the ordering cost is so insignicant to order can occur several time a day. Example 12.2: In ABC die casting company has 68000 kg of aluminum a year. The cost of storage of 1 kg a day is $0.04 and cost of ordering is $130. What is the optimum order period? Solution The information provides that p= 68000 0.04 365 = 3.23443016 2 130 (12.II.a)
The solution is either 3 or 4 times when additional consideration has to be taken into account.
End Solution
12.13. SUMMARY
145
(C) and revenue (R) which associated with these items both which are a function of x. Thus, T (x) = R(x) C(x) (12.27)
The revenue is a function of number items, x, the price, P (x) is also a function of x. To add discussion about the discrete v. approximation of the smooth derivation. The typical cost is
cost per item more complex situation
C(x) = f ixed +
x+
f (x)
(12.28)
Hence, the cost per item (if a is a function of x) is c(x) = 1 f (x) C(x) = a + d(x) + x x x (12.29)
Examining the terms in equation (12.29) reveals that there are terms that vanish as the number of items increase to innity x 0(1/x) = 1/ = 0. However, other terms behaves according the supply and demand diagram (pQ2 diagram). In the supply and demand diagram is diagram and was proposed by Marco Fanno (the other Fanno). The interesting part of the supply diagram show that the price decreases as the number of items increases. This part of the analysis shows indicate that the profits are a strong function item parts. These facts shows that situations where the total prots is reduced when the total income is increased. A typical equation repressing this situation is
TA2
(12.30)
12.13 Summary
In this chapter the economy of the design and choices of the casting process have been presented. It is advocated that the averaged approach commonly used in the die casting industry be abandoned. Adopt a more elaborate method, in which more precise calculations are made is also advocated. It is believed that the new method will create the super die casting economy.
12.14 Question
146
This model explains many industrial ow processes which includes emptying of pressured container through a relatively short tube, exhaust system of an internal combustion engine, compressed air systems, etc. As this model raised from need to explain the steam ow in turbines.
A.1 Introduction
Consider a gas owing through a conduit with a friction (see Figure (A.1)). It is advantages to examine the simplest situation and yet without losing the core properties of the process. Later, more general cases will be examined2 .
1 Even 2 Not
the friction does not convert into heat ready yet, discussed on the ideal gas model and the entry length issues.
147
148
The energy conservation (under the assumption that this model is adiabatic ow and the friction is not transformed into thermal energy) reads T0 1 = T0 2 T1 + U1 U2 = T2 + 2cp 2cp
2 2
(A.2)
It is assumed that the ow can be approximated as onedimensional. The force acting on the gas is the friction at the wall and the momentum conservation reads AdP w dAw = mdU It is convenient to dene a hydraulic diameter as DH = Or in other words A= DH 2 4 (A.8) 4 Cross Section Area wetted perimeter (A.7) (A.6)
3 The equation of state is written again here so that all the relevant equations can be found when this chapter is printed separately.
149
It is convenient to substitute D for DH and yet it still will be referred to the same name as the hydraulic diameter. The innitesimal area that shear stress is acting on is dAw = Ddx (A.9)
Introducing the Fanning friction factor as a dimensionless friction factor which is some times referred to as the friction coecient and reads as the following: f= w 1 U 2 2 (A.10)
By utilizing equation (A.2) and substituting equation (A.10) into momentum equation (A.6) yields
A w 2
D dP Ddx f 4
1 2 U 2
m A
= A U dU
(A.11)
Dividing equation (A.11) by the cross section area, A and rearranging yields dP + 4f dx D 1 2 U 2 = U dU (A.12)
The second law is the last equation to be utilized to determine the ow direction. s2 s1 (A.13)
U2 k RT
P
(A.14)
Using the identity in equation (A.14) and substituting it into equation (A.11) and after some rearrangement yields 4f dx DH 1 kP M 2 2 U dU dU = kP M 2 U U
2 U 2
dP +
(A.16)
= kM 2
dU U
(A.17)
It is convenient to relate expressions of (dP/P ) and dU/U in terms of the Mach number and substituting it into equation (A.17). Derivative of mass conservation ((A.2)) results in
dU U
d 1 dU 2 + =0 2 U2
(A.18)
The derivation of the equation of state (A.5) and dividing the results by equation of state (A.5) results dP d dT = + P dT (A.19)
Derivation of the Mach identity equation (A.14) and dividing by equation (A.14) yields d(M 2 ) d(U 2 ) dT = 2 M U2 T (A.20)
Dividing the energy equation (A.4) by Cp and by utilizing the denition Mach number yields dT + T 1 kR (k 1)
Cp
1 U2 d T U2
U2 2
dT (k 1) U 2 + d T kRT U 2
c2
U2 2
dT k 1 2 dU 2 + M =0 T 2 U2
(A.21)
Equations (A.17), (A.18), (A.19), (A.20), and (A.21) need to be solved. These equations are separable so one variable is a function of only single variable (the chosen as the independent variable). Explicit explanation is provided for only two variables, the rest variables can be done in a similar fashion. The dimensionless friction, 4f L , D is chosen as the independent variable since the change in the dimensionless resistance, 4f L D , causes the change in the other variables. Combining equations (A.19) and (A.21) when eliminating dT /T results dP d (k 1)M 2 dU 2 = P 2 U2 (A.22)
151
The term d can be eliminated by utilizing equation (A.18) and substituting it into equation (A.22) and rearrangement yields dP 1 + (k 1)M 2 dU 2 = P 2 U2 The term dU 2 /U 2 can be eliminated by using (A.23) kM 2 1 + (k 1)M 2 4f dx dP = P 2(1 M 2 ) D (A.24) (A.23)
The second equation for Mach number, M variable is obtained by combining equation (A.20) and (A.21) by eliminating dT /T . Then d/ and U are eliminated by utilizing equation (A.18) and equation (A.22). The only variable that is left is P (or dP/P ) which can be eliminated by utilizing equation (A.24) and results in 1 M 2 dM 2 4f dx = D kM 4 (1 + k1 M 2 ) 2 Rearranging equation (A.25) results in kM 2 1 + k1 M 2 4f dx dM 2 2 = M2 1 M2 D (A.26) (A.25)
After similar mathematical manipulation one can get the relationship for the velocity to read dU kM 2 4f dx = U 2 (1 M 2 ) D and the relationship for the temperature is 1 dc k(k 1)M 4 4f dx dT = = T 2 c 2(1 M 2 ) D density is obtained by utilizing equations (A.27) and (A.18) to obtain d kM 2 4f dx = 2 (1 M 2 ) D The stagnation pressure is similarly obtained as kM 2 4f dx dP0 = P0 2 D The second law reads ds = Cp ln dP dT R ln T P (A.31) (A.30) (A.29) (A.28) (A.27)
152
The stagnation temperature expresses as T0 = T (1 + (1 k)/2M 2 ). Taking derivative of this expression when M remains constant yields dT0 = dT (1 + (1 k)/2M 2 ) and thus when these equations are divided they yield dT /T = dT0 /T0 (A.32)
In similar fashion the relationship between the stagnation pressure and the pressure can be substituted into the entropy equation and result in ds = Cp ln dP0 dT0 R ln T0 P0 (A.33)
The rst law requires that the stagnation temperature remains constant, (dT0 = 0). Therefore the entropy change is ds (k 1) dP0 = Cp k P0 Using the equation for stagnation pressure the entropy equation yields ds (k 1)M 2 4f dx = Cp 2 D (A.35) (A.34)
153
The trends or whether the variables are increasing or decreasing can be observed from looking at the equation developed. For example, the pressure can be examined by looking at equation (A.26). It demonstrates that the Mach number increases downstream when the ow is subsonic. On the other hand, when the ow is supersonic, the pressure decreases. The summary of the properties changes on the sides of the branch Subsonic decrease increase increase decrease decrease decrease Supersonic increase decrease decrease increase increase increase
f dx =
L
(A.36)
f dx
(A.37)
In the isothermal ow model it was shown that friction factor is constant through the process if the uid is ideal gas. Here, the Reynolds number dened in equation (??) is not constant because the temperature is not constant. The viscosity even for ideal gas is complex function of the temperature (further reading in Basic of Fluid Mechanics chapter one, Potto Project). However, the temperature variation is very limit. Simple improvement can be done by assuming constant constant viscosity (constant friction factor) and nd the temperature on the two sides of the tube to improve the friction factor for the next iteration. The maximum error can be estimated by looking at the maximum change of the temperature. The temperature can be reduced by less than 20% for most range of the spesic heats ratio. The viscosity change for this change is for many gases about 10%. For these gases the maximum increase of average Reynolds number is only 5%. What this change in Reynolds number does to friction factor? That depend in the range of Reynolds number. For Reynolds number larger than 10,000 the change in friction factor can be considered negligible. For the other extreme, laminar
154
ow it can estimated that change of 5% in Reynolds number change about the same amount in friction factor. With the exception the jump from a laminar ow to a turbulent ow, the change is noticeable but very small. In the light of the about discussion the friction factor is assumed to constant. By utilizing the mean average theorem equation (A.36) yields
k+1 2 4f Lmax 1 1 M2 k+1 2 M = + ln D k M2 2k 1 + k1 M 2 2
(A.38)
It is common to replace the f with f which is adopted in this book. Equations (A.24), (A.27), (A.28), (A.29), (A.29), and (A.30) can be solved. For example, the pressure as written in equation (A.23) is represented by 4f L , and D Mach number. Now equation (A.24) can eliminate term 4f L and describe the pressure D on the Mach number. Dividing equation (A.24) in equation (A.26) yields
dP P dM 2 M2
1 + (k 1M 2 dM 2 2M 2 1 + k1 M 2 2
(A.39)
The symbol * denotes the state when the ow is choked and Mach number is equal to 1. Thus, M = 1 when P = P equation (A.39) can be integrated to yield:
P 1 = P M
1+
k+1 2 k1 2 2 M
(A.40)
k+1 T c2 2 = 2 = T c 1 + k1 M 2 2
(A.41)
1 = M
1+
k1 2 2 M k+1 2
(A.42)
U = U
=M
1+
k+1 2 k1 2 2 M
(A.43)
A.5. THE WORKING EQUATIONS The stagnation pressure decreases and can be expressed by (1+ 1k M 2 ) k1 2 P0 = P0 P0 P P0 P
2 ( k+1 ) k1 k k
155
P (A.44) P
Using the pressure ratio in equation (A.40) and substituting it into equation (A.44) yields P0 = P0 1+
k1 2 2 M k+1 2
k k1
1 M
1+
k1 2 2 M k+1 2
(A.45)
k+1 2(k1)
(A.46)
2M 2
k+1 1 + k1 M 2 2
k+1 k
(A.47)
The results of these equations are plotted in Figure (A.2) The Fanno ow is in many cases shockless and therefore a relationship between two points should be derived. In most times, the star values are imaginary values that represent the value at choking. The real ratio can be obtained by two star ratios as an example T2 = T1
T T M2 T T M1
(A.48)
4f L dx = D
Lmax L1
4f L dx D
Lmax L2
4f L dx D
(A.49)
Hence, 4f Lmax D =
2
4f Lmax D
4f L D
(A.50)
156
* *
Fanno Flow *
1e+01
0.1
0.01
1 Mach number
10
Solution For isentropic, the ow to the pipe inlet, the temperature and the total pressure at the
4 This
157
pipe inlet are the same as those in the reservoir. Thus, nding the total pressure and temperature at the pipe inlet is the solution. With the Mach number and temperature known at the exit, the total temperature at the entrance can be obtained by knowing the 4f L . For given Mach number (M = 0.9) the following is obtained. D M
4fL D P P P0 P0 U U T T
1.0089
1.0934
0.9146
1.0327
T2 =
2
is added as 3.21
4f L D
4f L D
The rest of the parameters can be obtained with the new by interpolations or by utilizing the attached program. M
4fL D P P P0 P0
U U
T T
0.35886 3.2100
3.0140
1.7405
2.5764
0.38814 1.1699
Note that the subsonic branch is chosen. The stagnation ratios has to be added for M = 0.35886 M
T T0 0 A A P P0 AP A P0 F F
0.91484 1.5922
0.78305
The total pressure P01 can be found from the combination of the ratios as follows:
P1 P
P01 = P2
P P
P P
P0 P
158
T1 T
T01 = T2
T T
348K = 75 C
End Solution
Another academic question/example: Example A.2: A system is composed of a convergentdivergent nozzle followed by a tube with length of 2.5 [cm] in diameter and 1.0 [m] long. The system is supplied by a vessel. The vessel conditions are at 29.65 [Bar], 400 K. With these conditions a pipe inlet Mach number is 3.0. A normal shock wave occurs in the tube and the ow discharges to the atmosphere, determine: (a) the mass ow rate through the system; (b) the temperature at the pipe exit; and (c) determine the Mach number when a normal shock wave occurs [Mx ]. Take k = 1.4, R = 287 [J/kgK] and f = 0.005. Solution
atmosphere conditions
(a) Assuming that the pressure vessel is very much larger than the pipe, therefore the velocity in the vessel can be assumed to be small enough so it can be neglected. Thus, the stagnation conditions can be approximated for the condition in the tank. It is further assumed that the ow through the nozzle can be approximated as isentropic. Hence, T01 = 400K and P01 = 29.65[P ar]
159
The mass ow rate through the system is constant and for simplicity point 1 is chosen in which, m = AM c The density and speed of sound are unknowns and need to be computed. With the isentropic relationship the Mach number at point one (1) is known, then the following can be found either from Table (A.1) or the PottoGDC M 3.0000
T T0 0 A A P P0 AP A P0 F F
Using the temperature, the speed of sound can be calculated as c1 = kRT = 1.4 287 142.8 239.54[m/sec] The pressure at point 1 can be calculated as P1 = P1 P01 = 0.027 30 P01 0.81[Bar]
1.97
kg m3
The mass ow rate can be evaluated from equation (A.2) m = 1.97 0.0252 kg 3 239.54 = 0.69 4 sec
(b) First, check whether the ow is shockless by comparing the ow resistance and the maximum possible resistance. From the Table (A.1) or by using the Potto GDC, to obtain the following M 3.0000
4fL D P P P0 P0 U U T T
0.50918 1.9640
0.42857
Since 0.8 > 0.52216 the ow is choked and with a shock wave. The exit pressure determines the location of the shock, if a shock exists, by comparing possible Pexit to PB . Two possibilities are needed to be checked; one, the shock at the entrance of the tube, and two, shock at the exit and comparing the pressure ratios. First, the possibility that the shock wave occurs immediately at the entrance for which the ratio for Mx are (shock wave Table (??))
Mx 3.0000
My 0.47519
Ty Tx
y x
Py Px
P0 y P0x
2.6790
3.8571
10.3333
0.32834
After the shock wave the ow is subsonic with M1 = 0.47519. (Fanno ow Table (A.1))
4fL D
P P
P0 P0
U U
T T
0.47519 1.2919
2.2549
1.3904
1.9640
0.50917 1.1481
0.85676 1.1912
0.65326
P2 = P0 = =
P2 P
161
The actual pressure ratio 1/29.65 = 0.0338 is smaller than the case in which shock occurs at the entrance. Thus, the shock is somewhere downstream. One possible way to nd the exit temperature, T2 is by nding the location of the shock. To nd the location of the shock ratio of the pressure ratio, P2 is needed. P1 With the location of shock, claiming upstream from the exit through shock to the entrance. For example, calculate the parameters for shock location with known 4f L in the y side. Then either by utilizing shock table or the program, D to obtain the upstream Mach number. The procedure for the calculations: Calculate the entrance Mach number assuming the shock occurs at the exit: 1) a) set M2 = 1 assume the ow in the entire tube is supersonic: b) calculated M1 Note this Mach number is the high Value. Calculate the entrance Mach assuming shock at the entrance. a) set M2 = 1 2) b) add 4f L and calculated M1 for subsonic branch D c) calculated Mx for M1 Note this Mach number is the low Value. According your root nding algorithm5 calculate or guess the shock location and then compute as above the new M1 . a) set M2 = 1 3) b) for the new 4f L and compute the new M for the subsonic branch y D c) calculated Mx for the My d) Add the leftover of
4f L D
4) guess new location for the shock according to your nding root procedure and according to the result, repeat previous stage until the solution is obtained. M1 3.0000 M2 1.0000
4fL D up 4fL D down
Mx 1.9899
My 0.57910
0.22019
0.57981
(c) The way of the numerical procedure for solving this problem is by nding
4f L D
that will produce M1 = 3. In the process Mx and My must be calculated (see the chapter on the program with its algorithms.).
End Solution
up
162
and the heat transfer is not sucient to maintain constant temperature. In the Fanno model there is no heat transfer, and, furthermore, because the very limited amount of heat transformed it is closer to an adiabatic ow. The only limitation of the model is its uniform velocity (assuming parabolic ow for laminar and dierent prole for turbulent ow.). The information from the wall to the tube center6 is slower in reality. However, experiments from many starting with 1938 work by Frossel7 has shown that the error is not signicant. Nevertheless, the comparison with reality shows that heat transfer cause changes to the ow and they need/should to be expected. These changes include the choking point at lower Mach number.
4f L D (M
, k = 1.4) = 0.8215
The maximum length of the supersonic ow is limited by the above number. From the above analysis, it can be observed that no matter how high the entrance Mach number will be the tube length is limited and depends only on specic heat ratio, k as shown in Figure (A.5).
163
The discussion has to dierentiate between two ways of feeding the tube: converging nozzle or a converging-diverging nozzle. Three parameters, the dimensionless friction, 4f L , the entrance Mach number, M1 , and the pressure ratio, P2 /P1 are conD trolling the ow. Only a combination of these two parameters is truly independent. However, all the three parameters can be varied and they are discussed separately here.
A.9.1
In the analysis of this eect, it should be assumed that back pressure is constant and/or low as possible as needed to maintain a choked ow. First, the treatment of the two branches are separated. Fanno Flow Subsonic branch For converging nozzle feeding, increasing the tube length results in increasing the exit Mach number (normally denoted herein as M2 ). Once the Mach number reaches maximum (M = 1), no further increase of the exit Mach number can be achieved. In this process, the mass ow rate decreases. It is worth noting that entrance Mach number is reduced (as some might explain it to reduce the ow rate). The entrance temperature increases as can be seen from Figure (A.7). The velocity therefore must decrease because the loss of the enthalpy (stagnation temperature) is used. The density decrease
164
P because = RT and when pressure is remains almost constant the density decreases. Thus, the mass ow rate must decrease. These results are applicable to the converging nozzle.
In the case of the convergingdiverging feeding nozzle, increase of the dimensionless friction, 4f L , results in a similar ow pattern as in the converging nozzle. Once D the ow becomes choked a dierent ow pattern emerges.
Fanno Flow Supersonic Branch There are several transitional points that change the pattern of the ow. Point a is the choking point (for the supersonic branch) in which the exit Mach number reaches to one. Point b is the maximum possible ow for supersonic ow and is not dependent on the nozzle. The next point, referred here as the critical point c, is the point in which no supersonic ow is possible in the tube i.e. the shock reaches to the nozzle. There is another point d, in which no supersonic ow is possible in the entire nozzletube system. Between these transitional points the eect parameters such as mass ow rate, entrance and exit Mach number are discussed. At the starting point the ow is choked in the nozzle, to achieve supersonic ow. The following ranges that has to be discussed includes (see Figure (A.8)):
165
1 1
1 2 2 2
Fanno lines
0
4f L D 4f L D 4f L D choking shockless chokeless
4f L D 4f L D 4f L D 4f L D
4f L D 4f L D 4f L D
0a ab bc c
Fig. -A.8. The Mach numbers at entrance and exit of tube and mass ow rate for Fanno Flow as a function of the 4f L . D
The 0-a range, the mass ow rate is constant because the ow is choked at the nozzle. The entrance Mach number, M1 is constant because it is a function of the nozzle design only. The exit Mach number, M2 decreases (remember this ow is on the supersonic branch) and starts ( 4f L = 0) as M2 = M1 . At the end of the range a, M2 = 1. In the D
166
range of a b the ow is all supersonic. In the next range a b The ow is double choked and make the adjustment for the ow rate at dierent choking points by changing the shock location. The mass ow rate continues to be constant. The entrance Mach continues to be constant and exit Mach number is constant. The total maximum available for supersonic ow b b , 4f L , is only a D max theoretical length in which the supersonic ow can occur if nozzle is provided with a larger Mach number (a change to the nozzle area ratio which also reduces the mass ow rate). In the range b c, it is a more practical point. In semi supersonic ow b c (in which no supersonic is available in the tube but only in the nozzle) the ow is still double choked and the mass ow rate is constant. Notice that exit Mach number, M2 is still one. However, the entrance Mach number, M1 , reduces with the increase of 4f L . D It is worth noticing that in the a c the mass ow rate nozzle entrance velocity and the exit velocity remains constant!10 In the last range c the end is really the pressure limit or the break of the model and the isothermal model is more appropriate to describe the ow. In this range, the ow rate decreases since (m M1 )11 . To summarize the above discussion, Figures (A.8) exhibits the development of M1 , M2 mass ow rate as a function of 4f L . Somewhat dierent then the subsonic D branch the mass ow rate is constant even if the ow in the tube is completely subsonic. This situation is because of the double choked condition in the nozzle. The exit Mach M2 is a continuous monotonic function that decreases with 4f L . The entrance Mach D M1 is a non continuous function with a jump at the point when shock occurs at the entrance moves into the nozzle. Figure (A.9) exhibits the M1 as a function of M2 . The Figure was calculated by utilizing the data from Figure (A.2) by obtaining the 4f L for M2 and subtracting D the given 4f L and nding the corresponding M1 . D The Figure (A.10) exhibits the entrance Mach number as a function of the M2 . Obviously there can be two extreme possibilities for the subsonic exit branch. Subsonic velocity occurs for supersonic entrance velocity, one, when the shock wave occurs at the tube exit and two, at the tube entrance. In Figure (A.10) only for 4f L = 0.1 and D 4f L 4f L D = 0.4 two extremes are shown. For D = 0.2 shown with only shock at the exit only. Obviously, and as can be observed, the larger 4f L creates larger dierences D between exit Mach number for the dierent shock locations. The larger 4f L larger M1 D must occurs even for shock at the entrance. For a given 4f L , below the maximum critical length, the supersonic entrance ow D has three dierent regimes which depends on the back pressure. One, shockless ow,
10 On a personal note, this situation is rather strange to explain. On one hand, the resistance increases and on the other hand, the exit Mach number remains constant and equal to one. Does anyone have an explanation for this strange behavior suitable for nonengineers or engineers without background in uid mechanics? 11 Note that increases with decreases of M but this eect is less signicant. 1 1
max
167
Fanno Flow
M1 as a function of M2 1 0.9 0.8 0.7 Entrace Mach number 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 4fL = 0.1 D = 1.0 = 10.0 = 100.0
tow, shock at the entrance, and three, shock at the exit. Below, the maximum critical length is mathematically 4f L 1 1+k k+1 > + ln D k 2k k1 For cases of 4f L above the maximum critical length no supersonic ow can be over the D whole tube and at some point a shock will occur and the ow becomes subsonic ow12 .
A.9.2
P2 , P1
eects
In this section the studied parameter is the variation of the back pressure and thus, the pressure ratio P2 variations. For very low pressure ratio the ow can be assumed P1 as incompressible with exit Mach number smaller than < 0.3. As the pressure ratio
12 See
168
Fanno Flow
M1 as a function of M2 for the subsonic brench 5 4.5 4 3.5 3 M1 2.5 2 1.5 1 0.5 0 0 0.2 0.4 0.6 0.8 1 M2 1.2 1.4 1.6 1.8 2 4fL = 0.1 D = 0.2 = 0.4 = 0.1 shock = 0.4
increases (smaller back pressure, P2 ), the exit and entrance Mach numbers increase. According to Fanno model the value of 4f L is constant (friction factor, f , is independent D of the parameters such as, Mach number, Reynolds number et cetera) thus the ow remains on the same Fanno line. For cases where the supply come from a reservoir with a constant pressure, the entrance pressure decreases as well because of the increase in the entrance Mach number (velocity). Again a dierentiation of the feeding is important to point out. If the feeding nozzle is converging than the ow will be only subsonic. If the nozzle is converging diverging than in some part supersonic ow is possible. At rst the converging nozzle is presented and later the converging-diverging nozzle is explained. Choking explanation for pressure variation/reduction Decreasing the pressure ratio or in actuality the back pressure, results in increase of the entrance and the exit velocity until a maximum is reached for the exit velocity.
169
critical Point a
criticalPoint b
critical Point c
critical Point d
4f L D
for a short
4f L D
The maximum velocity is when exit Mach number equals one. The Mach number, as it was shown in Chapter (??), can increases only if the area increase. In our model the tube area is postulated as a constant therefore the velocity cannot increase any further. However, for the ow to be continuous the pressure must decrease and for that the velocity must increase. Something must break since there are conicting demands and it result in a jump in the ow. This jump is referred to as a choked ow. Any additional reduction in the back pressure will not change the situation in the tube. The only change will be at tube surroundings which are irrelevant to this discussion. If the feeding nozzle is a convergingdiverging then it has to be dierentiated between two cases; One case is where the 4f L is short or equal to the critical length. The D critical length is the maximum Short
4f L D 4f L D max
Figure (A.12) shows dierent pressure proles for dierent back pressures. Before the ow reaches critical point a (in the Figure) the ow is subsonic. Up to this stage the
170
for a long
4f L D
nozzle feeding the tube increases the mass ow rate (with decreasing back pressure). Between point a and point b the shock is in the nozzle. In this range and further reduction of the pressure the mass ow rate is constant no matter how low the back pressure is reduced. Once the back pressure is less than point b the supersonic reaches to the tube. Note however that exit Mach number, M2 < 1 and is not 1. A back pressure that is at the critical point c results in a shock wave that is at the exit. When the back pressure is below point c, the tube is clean of any shock13 . The back pressure below point c has some adjustment as it occurs with exceptions of point d. Long
4f L D 4f L D
In the case of
>
4f L D
max
<
4f L D
max
171
shock at
75% 50% 5%
Fig. -A.13. The eects of pressure variations on Mach number prole as a function of when the total resistance 4f L = 0.3 for Fanno Flow D
4f L D
exit Mach number is equal to 1 and the ow is double shock. Further reduction of the back pressure at this stage will not move the shock wave downstream the nozzle. At point c or location of the shock wave, is a function entrance Mach number, M1 and the extra 4f L . The is no analytical solution for the location of this point c. The D procedure is (will be) presented in later stage.
A.9.3
In this discussion, the eect of changing the throat area on the nozzle eciency is neglected. In reality these eects have signicance and needs to be accounted for some instances. This dissection deals only with the ow when it reaches the supersonic branch reached otherwise the ow is subsonic with regular eects. It is assumed that in this P2 discussion that the pressure ratio P1 is large enough to create a choked ow and 4f L D is small enough to allow it to happen. The entrance Mach number, M1 is a function of the ratio of the nozzles throat area to the nozzle exit area and its eciency. This eect is the third parameter discussed
172
4f L D
= 0.3
here. Practically, the nozzle area ratio is changed by changing the throat area. As was shown before, there are two dierent maximums for 4f L ; rst is the total D maximum 4f L of the supersonic which depends only on the specic heat, k, and second D the maximum depends on the entrance Mach number, M1 . This analysis deals with the case where 4f L is shorter than total 4f L . D D
max
Obviously, in this situation, the critical point is where 4f L is equal to 4f L D D max as a result in the entrance Mach number. The process of decreasing the convergingdiverging nozzles throat increases the entrance14 Mach number. If the tube contains no supersonic ow then reducing the nozzle throat area wouldnt increase the entrance Mach number. This part is for the case where some part of the tube is under supersonic regime and there is shock as a transition to subsonic branch. Decreasing the nozzle throat area
14 The word entrance referred to the tube and not to the nozzle. The reference to the tube is because it is the focus of the study.
173
shock
moves the shock location downstream. The payment for increase in the supersonic length is by reducing the mass ow. Further, decrease of the throat area results in ushing the shock out of the tube. By doing so, the throat area decreases. The mass ow rate is proportionally linear to the throat area and therefore the mass ow rate reduces. The process of decreasing the throat area also results in increasing the pressure drop of the nozzle (larger resistance in the nozzle15 )16 . In the case of large tube 4f L > 4f L the exit Mach number increases with the D D max decrease of the throat area. Once the exit Mach number reaches one no further increases is possible. However, the location of the shock wave approaches to the theoretical location if entrance Mach, M1 = . The maximum location of the shock The main point in this discussion however, is to nd the furthest shock location downstream. Figure (A.16) shows the possible 4f L as function of retreat of the location of the shock wave from the maximum D location. When the entrance Mach number is innity, M1 = , if the shock location is at the maximum length, then shock at Mx = 1 results in My = 1. The proposed procedure is based on Figure (A.16). i) Calculate the extra 4f L and subtract the actual extra D the left side (at the max length). ii) Calculate the extra 4f L and subtract the actual extra D the right side (at the entrance).
4f L D
4f L D
iii) According to the positive or negative utilizes your root nding procedure.
Frictionless nozzle has a larger resistance when the throat area decreases is one of the strange phenomenon that in one way increasing the resistance (changing the throat area) decreases the ow rate while in a dierent way (increasing the 4f L ) does not aect the ow D rate.
16 It 15 Strange?
174
4f L D
Fig. -A.16. The extra tube length as a function of the shock location,
supersonic branch
From numerical point of view, the Mach number equal innity when left side assumes result in innity length of possible extra (the whole ow in the tube is subsonic). To overcome this numerical problem it is suggested to start the calculation from distance from the right hand side. Let denote 4f L 4f L 4f L = (A.51) D D actual D sup Note that
4f L D sup
is smaller than
4f L D
max
ed is that denote 4f L as dierence between the maximum possible of length D retreat in which the supersonic ow is achieved and the actual length in which the ow is supersonic see Figure (A.15). The retreating length is expressed as subsonic but 4f L D =
retreat
4f L D
max
4f L D
(A.52)
sup
Figure (A.17) shows the entrance Mach number, M1 reduces after the maximum length is exceeded. Example A.3: Calculate the shock location for entrance Mach number M1 = 8 and for assume that k = 1.4 (Mexit = 1).
4f L D
= 0.9
175
4f L D max
4f L D
Fig. -A.17. The maximum entrance Mach number, M1 to the tube as a function of supersonic branch
for k =
exceed the maximum length for this entrance 1.4 is 0.821508116. Hence, 4f L Mach number. The maximum for M1 = 8 is D = 0.76820, thus the extra tube is
4f L D 4f L D
= 0.9 0.76820 = 0.1318. The left side is when the shock occurs at
= 0.76820 (ow is choked and no additional 4f L ). Hence, the value of left side is D 0.1318. The right side is when the shock is at the entrance at which the extra 4f L is D calculated for Mx and My is Mx 8.0000 With (M1 ) M 0.39289 The extra
4fL D P P P0 P0 U U T T
My 0.39289
Ty Tx
y x
Py Px
P0y P0 x
13.3867
5.5652
74.5000
0.00849
2.4417
4f L D
2.7461
1.6136
2.3591
0.42390
1.1641
176
between the negative of left side to the positive of the right side17 . In a summary of the actions is done by the following algorithm: (a) check if the 4f L exceeds the maximum D cordingly continue.
4f L D up 4f L D 4f L D 4f L D max
(b) Guess
max
4f L D up ,
(d) Calculate the associate Mach number, Mx with the Mach number, My calculated previously, (e) Calculate
4f L D
4f L D down
4f L D
4f L D up
(h) Check the new and improved stop or return to stage (b). Shock location are:
4f L D
down
M1 8.0000
M2 1.0000
4fL D up
4fL D down
Mx 1.6706
My 0.64830
0.57068
0.32932
A.10.
177 My
4fL D
Mx 1.3838 1.5286 1.6021 1.6382 1.6554 1.6635 1.6673 1.6691 1.6699 1.6703 1.6705 1.6706 1.6706 1.6706 1.6706 1.6706 1.6706 1.6706
0.67426 0.62170 0.59506 0.58217 0.57605 0.57318 0.57184 0.57122 0.57093 0.57079 0.57073 0.57070 0.57069 0.57068 0.57068 0.57068 0.57068 0.57068
0.22574 0.27830 0.30494 0.31783 0.32395 0.32682 0.32816 0.32878 0.32907 0.32921 0.32927 0.32930 0.32931 0.32932 0.32932 0.32932 0.32932 0.32932
0.74664 0.69119 0.66779 0.65728 0.65246 0.65023 0.64920 0.64872 0.64850 0.64839 0.64834 0.64832 0.64831 0.64831 0.64830 0.64830 0.64830 0.64830
0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000 0.90000
18 This questions were raised from many who didnt nd any book that discuss these practical aspects and send questions to this author.
178
A.10.1
4f L D
This pair of parameters is the most M2 4f L M1 4f L natural to examine because, in most P1 P2 D P = P D cases, this information is the only inM =1 formation that is provided. For a hypothetical section 4f L given pipe , neither the enD trance Mach number nor the exit Fig. -A.18. Unchoked ow calculations showing the Mach number are given (sometimes hypothetical full tube when choked the entrance Mach number is give see the next section). There is no exact analytical solution. There are two possible approaches to solve this problem: one, by building a representative function and nd a root (or roots) of this representative function. Two, the problem can be solved by an iterative procedure. The rst approach require using root nding method and either method of spline method or the half method found to be good. However, this author experience show that these methods in this case were found to be relatively slow. The NewtonRapson method is much faster but not were found to be unstable (at lease in the way that was implemented by this author). The iterative method used to solve constructed on the properties of several physical quantities must be in a certain range. The rst fact is that the pressure ratio P2 /P1 is always between 0 and 1 (see Figure A.18). In the gure, a theoretical extra tube is added in such a length that cause the ow to choke (if it really was there). This length is always positive (at minimum is zero). The procedure for the calculations is as the following: 1) Calculate the entrance Mach number, M1 assuming the (chocked ow);
4f L D
4f L D
max
2) Calculate the minimum pressure ratio (P2 /P1 )min for M1 (look at table (A.1)) 3) Check if the ow is choked: There are two possibilities to check it. a) Check if the given
4f L D
is smaller than
4f L D
b) check if the (P2 /P1 )min is larger than (P2 /P1 ), continue if the criteria is satised. Or if not satised abort this procedure and continue to calculation for choked ow. 4) Calculate the M2 based on the (P /P2 ) = (P1 /P2 ), 5) calculate 4f L based on M2 , D 6) calculate the new (P2 /P1 ), based on the new f (remember that 4f L = D
4f L D 2 4f L D 1
4f L D
),
A.10.
179
7) calculate the corresponding M1 and M2 , 8) calculate the new and improve the 4f L by D 4f L D 4f L D
P2 P1 P2 P1
=
new
old
given old
(A.53)
4f L D
9) Calculate the improved/new M2 based on the improve 4f L D 10) calculate the improved
4f L D
as
4f L D
4f L D
given
+
4f L D .
4f L D
new
12) Compare the abs ((P2 /P1 )new (P2 /P1 )old ) and if not satised returned to stage (6) until the solution is obtained. To demonstrate how this procedure is working consider a typical example of 4f L = D 1.7 and P2 /P1 = 0.5. Using the above algorithm the results are exhibited in the following gure. Figure (A.19) demonstrates that the conversion occur at about 7-8
3.0
M1 M2
4f L D
P2/P1
2.0
4f L D
1.5
1.0
0.5
10
20
30
40
50
60
70
80
90
100
110
120
130
Number of Iterations, i
October 8, 2007
Fig. -A.19. The results of the algorithm showing the conversion rate for unchoked Fanno ow model with a given 4f L and pressure ratio. D
iterations. With better rst guess this conversion procedure will converts much faster (under construction).
180
A.10.2
This situation pose a simple mathematical problem while the physical situation occurs in cases where a specic ow rate is required with a given pressure ratio (range) (this problem was considered by some to be somewhat complicated). The specic ow rate can be converted to entrance Mach number and this simplies the problem. Thus, the problem is reduced to nd for given entrance Mach, M1 , and given pressure ratio calculate the ow parameters, like the exit Mach number, M2 . The procedure is based on the fact that the entrance star pressure ratio can be calculated using M1 . Thus, using the pressure ratio to calculate the star exit pressure ratio provide the exit Mach number, M2 . An example of such issue is the following example that combines also the Naughty professor problems. Example A.4: Calculate the exit Mach number for P2 /P 1 = 0.4 and entrance Mach number M1 = 0.25. Solution The star pressure can be obtained from a table or Potto-GDC as M
4fL D P P P0 P0 U U T T
0.25000 8.4834
4.3546
2.4027
3.6742
0.27217 1.1852
And the star pressure ratio can be calculated at the exit as following P2 P2 P1 = = 0.4 4.3546 = 1.74184 P P1 P And the corresponding exit Mach number for this pressure ratio reads M
4fL D P P P0 P0 U U T T
1.1801
1.5585
0.64165 1.1177
A bit show o the PottoGDC can carry these calculations in one click as M1 0.25000 M2 0.60693
End Solution
4fL D
P2 P1
8.0193
0.40000
While the above example show the most simple from of this question, in reality this question is more complicated. One common problem is situation that the diameter is not given but the ow rate and length and pressure (stagnation or static) with some combination of the temperature. The following example deal with one of such example.
A.10.
181
Example A.5: A tank lled with air at stagnation pressure, 2[Bar] should be connected to a pipe with a friction factor, f = 0.005, and and length of 5[m]. The ow rate is (should be) kg 0.1 sec and the static temperature at the entrance of the pipe was measured to be 27 C. The pressure ratio P2 /P1 should not fall below 0.9 (P2 /P1 > 0.9). Calculate the exit Mach number, M2 , ow rate, and minimum pipe diameter. You can assume that k = 1.4. Solution The direct mathematical solution isnt possible and some kind of iteration procedure or root nding for a representative function. For the rst part the naughty professor procedure cannot be used because m/A is not provided and the other hand 4f L is D not provided (missing Diameter). One possible solution is to guess the entrance Mach and check whether and the mass ow rate with the naughty professor procedure are satised. For Fanno ow at for several Mach numbers the following is obtained M1 0.10000 0.15000 0.20000 M2 0.11109 0.16658 0.22202
4fL D P2 P1
From the last table the diameter can be calculated for example for M1 = 0.2 as D= 4f L
4f L D
The same was done for all the other Mach number. Now the area can be calculated and therefor the m/A can be calculated. With this information the naughty professor is given and the entrance Mach number can be calculated. For example for M1 = 0.2 one can obtain the following: m/A = 0.1/( 0.031362 /4) 129.4666798 The same order as the above table it shown in naughty professor (isentropic table). M
T T0 0 A A P P0 AP A P0 F F
1.5781 0.66752 0.36404 1.2329 0.36221 0.97443 0.93730 1.7268 0.10979 0.99760 0.99400 5.3092
The rst result are not reasonable and this process can continue until the satisfactory solution is achieved. Here an graphical approximation is shown. From this exhibit it can be estimated that M1 = 0.18. For this Mach number the following can be obtained
182
0.4
guessed M1 calculated M1
0.3
0.2
0.1
Solution
0.1 0.15 0.2 0.25 0.3
Fig. -A.20. Diagram for nding solution when the pressure ratio and entrance properties (T and P0 are given
M1 0.18000
M2 0.19985
4fL D
P2 P1
3.9839
0.90000
Thus, the diameter can be obtained as D 0.0251[m] The ow rate is m/A 202.1[kg/sec m2 ] M
T T0 0 A A P P0 AP A P0 F F
0.97978 3.3726
1.4628
A.11
The isothermal ow model has equations that theoreticians nd easier to use and to compare to the Fanno ow model. One must notice that the maximum temperature at the entrance is T0 1 . When the Mach number decreases the temperature approaches the stagnation temperature (T T0 ). Hence, if one allows certain deviation of temperature, say about 1% that ow can be assumed to be isothermal. This tolerance requires that (T0 T )/T0 = 0.99 which requires that enough for M1 < 0.15 even for large k = 1.67. This requirement provides that somewhere (depend) in the vicinity of 4f L = 25 the ow can be assumed D isothermal. Hence the mass ow rate is a function of 4f L because M1 changes. Looking D at the table or Figure (A.2) or the results from PottoGDC attached to this book shows that reduction of the mass ow is very rapid. As it can be seen for the Figure (A.21) the dominating parameter is 4f L . The results are very similar for isothermal ow. The D only dierence is in small dimensionless friction, 4f L . D
A.12.
183
M1 Fanno flow
with comperison to Isothermal Flow
10
20
30
40
50 4fL D
60
70
80
90
100
Fig. -A.21. The entrance Mach number as a function of dimensionless resistance and comparison with Isothermal Flow
A.12
Example A.6: To demonstrate the utility in Figure (A.21) consider the following example. Find the mass ow rate for f = 0.05, L = 4[m], D = 0.02[m] and pressure ratio P2 /P1 = 0.1, 0.3, 0.5, 0.8. The stagnation conditions at the entrance are 300K and 3[bar] air. Solution First calculate the dimensionless resistance,
4f L D .
4f L 4 0.05 4 = = 40 D 0.02 From Figure (A.21) for P2 /P1 = 0.1 M1 0.13 etc. or accurately by utilizing the program as in the following table.
Only for the pressure ratio of 0.1 the ow is choked. M 0.12728 0.12420 0.11392 0.07975
T T0 0 A A P P0 AP A P0
Therefore, T T0 and is the same for the pressure. Hence, the mass rate is a function of the Mach number. The Mach number is indeed a function of the pressure ratio but mass ow rate is a function of pressure ratio only through Mach number. The mass ow rate is m = P AM k 0.022 = 300000 0.127 RT 4 1.4 0.48 287300 kg sec
and for the rest P2 0.1242 = 0.3 0.48 = 0.468 P1 0.1273 P2 0.1139 m = 0.5 0.48 = 0.43 P1 0.1273 P2 0.07975 m = 0.8 0.48 = 0.30 P1 0.1273 m
End Solution
A.13
M 0.03 0.04
4fL D
P P
P0 P0
U U
T T
787.08 440.35
36.5116 27.3817
19.3005 14.4815
30.4318 22.8254
A.13.
185
M 0.05 0.06 0.07 0.08 0.09 0.10 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00
4fL D
P P
P0 P0
U U
T T
280.02 193.03 140.66 106.72 83.4961 66.9216 14.5333 8.4834 5.2993 3.4525 2.3085 1.5664 1.0691 0.72805 0.49082 0.32459 0.20814 0.12728 0.07229 0.03633 0.01451 0.00328 0.0 0.30500 0.52216 0.63306 0.69380 0.72988 0.75280 0.76819 0.77899 0.78683 0.81265 0.81582 0.81755 0.81860 0.81928 0.81975 0.82008
21.9034 11.5914 18.2508 9.6659 15.6416 8.2915 13.6843 7.2616 12.1618 6.4613 10.9435 5.8218 5.4554 2.9635 4.3546 2.4027 3.6191 2.0351 3.0922 1.7780 2.6958 1.5901 2.3865 1.4487 2.1381 1.3398 1.9341 1.2549 1.7634 1.1882 1.6183 1.1356 1.4935 1.0944 1.3848 1.0624 1.2893 1.0382 1.2047 1.0207 1.1291 1.0089 1.061 1.002 1.00000 1.000 0.40825 1.688 0.21822 4.235 0.13363 10.72 0.089443 25.00 0.063758 53.18 0.047619 1.0E+2 0.036860 1.9E+2 0.029348 3.3E+2 0.023905 5.4E+2 0.00609 1.5E+4 0.00390 4.6E+4 0.00271 1.1E+5 0.00200 2.5E+5 0.00153 4.8E+5 0.00121 8.6E+5 0.000979 1.5E+6
18.2620 15.2200 13.0474 11.4182 10.1512 9.1378 4.5826 3.6742 3.0702 2.6400 2.3184 2.0693 1.8708 1.7092 1.5753 1.4626 1.3665 1.2838 1.2119 1.1489 1.0934 1.044 1.000 0.61237 0.50918 0.46771 0.44721 0.43568 0.42857 0.42390 0.42066 0.41833 0.41079 0.40988 0.40938 0.40908 0.40889 0.40875 0.40866
0.05476 0.06570 0.07664 0.08758 0.09851 0.10944 0.21822 0.27217 0.32572 0.37879 0.43133 0.48326 0.53452 0.58506 0.63481 0.68374 0.73179 0.77894 0.82514 0.87037 0.91460 0.95781 1.00 1.633 1.964 2.138 2.236 2.295 2.333 2.359 2.377 2.390 2.434 2.440 2.443 2.445 2.446 2.446 2.447
1.1994 1.1991 1.1988 1.1985 1.1981 1.1976 1.1905 1.1852 1.1788 1.1713 1.1628 1.1533 1.1429 1.1315 1.1194 1.1065 1.0929 1.0787 1.0638 1.0485 1.0327 1.017 1.000 0.66667 0.42857 0.28571 0.20000 0.14634 0.11111 0.086957 0.069767 0.057143 0.014815 0.00952 0.00663 0.00488 0.00374 0.00296 0.00240
186
4fL D
P P
P0 P0
U U
T T
A.14
Constant Zone
Fig. -A.22. Moody diagram on the name Moody who netscaped H. Rouse work to claim as his own. In this section the turbulent area is divided into 3 zones, constant, semiconstant, and linear After S Beck and R. Collins.
The friction factor in equation (A.25) was assumed constant. In Chapter ?? it was shown that the Reynolds number remains constant for ideal gas uid. However, in Fanno ow the temperature does not remain constant hence, as it was discussed before, the Reynolds number is increasing. Thus, the friction decreases with the exception of the switch in the ow pattern (laminar to turbulent ow). For relatively large relative roughness larger /D > 0.004 of 0.4% the friction factor is constant. For smother pipe /D < 0.001 and Reynolds number between 10,000 to a million the friction factor vary between 0.007 to 0.003 with is about factor of two. Thus, the error of 4f L is D limited by a factor of two (2). For this range, the friction factor can be estimated as a linear function of the log10 (Re). The error in this assumption is probably small of the assumption that involve in fanno ow model construction. Hence, f = A log10 (Re) + B (A.54)
A.14.
187
Where the constant A and B are function of the relative roughness. For most practical purposes the slop coecient A can be further assumed constant. The slop coecient A = 0.998125 Thus, to carry this calculation relationship between the viscosity and the temperature. If the viscosity expanded as Taylor or Maclaren series then A1 T = A0 + + 1 T0 Where 1 is the viscosity at the entrance temperature T1 . Thus, Reynolds number is Re = DU 1 A0 + AT0T + DU A0 +
A1 T2 T1
(A.55)
(A.56)
Left hand side of equation (A.25) is a function of the Mach number since it contains the temperature. If the temperature functionality will not vary similarly to the case of constant friction factor then the temperature can be expressed using equation (A.41).
constant
4 A log10 D
A0 + A1
DU 1 + k1 M1 2 2 1+
2 k1 2 M2
+ B
(A.58)
Equation (A.58) is only estimate of the functionally however, this estimate is almost as good as the assumptions of Fanno ow. Equation fanno:eq:d2 can be improved by using equation (A.58)
constant
4 Lmax D
A log10
In the most complicate case where the ow pattern is change from laminar ow to turbulent ow the whole Fanno ow model is questionable and will produce poor results. In summary, in the literature there are three approaches to this issue of non constant friction factor. The friction potential is recommended by a researcher in Germany and it is complicated. The second method substituting this physical approach with numerical iteration. In the numerical iteration method, the expression of the various relationships are inserted into governing dierential equations. The numerical methods does not allow exibility and is very complicated. The methods described here can be expended (if really really needed) and it will be done in very few iteration as it was shown in the Isothermal Chapter.
188
anonymous In this section exhibits the establishment experts reaction the position that the common pQ2 diagram is improper. Their comments are responses to the authors paper: The mathematical theory of the pQ2 diagram (similar to Chapter 7)1 . The paper was submitted to Journal of Manufacturing Science and Engineering. This part is for the Associate Technical Editor Dr. R. E. Smelser. I am sure that you are proud of the referees that you have chosen and that you do not have any objection whatsoever with publishing this information. Please send a copy of this appendix to the referees. I will be glad to hear from them. This concludes comments to the Editor. I believe that you, the reader should judge if the mathematical theory of the pQ2 diagram is correct or whether the experts position is reasonable. For the reader unfamiliar with the journal review process, the associate editor sent the paper to readers (referees) which are anonymous to the authors. They comment on the paper and according to these experts the paper acceptance is determined. I have chosen the unusual step to publish their comments because I believe that other motivations are involved in their responses. Coupled with the response to the publication of a summary of this
1 The exact paper can be obtain free of charge from Minnesota Suppercomputing Institute, http://www2.msi.umn.edu/publications.html report number 99/40 The mathematical theory of the pQ2 diagram or by writing to the Supercomputing Institute, University of Minnesota, 1200 Washington Avenue South, Minneapolis, MN 55415-1227
189
190
paper in the Die Casting Engineer, bring me to think that the best way to remove the information blockage is to open it to the public. Here, the referees can react to this rebuttal and stay anonymous via correspondence with the associated editor. If the referee/s choose to respond to the rebuttal, their comments will appear in the future additions. I will help them as much as I can to show their opinions. I am sure that they are proud of their criticism and are standing behind it 100%. Furthermore, I am absolutely, positively sure that they are so proud of their criticism they glad that it appears in publication such as this book.
191
Let me say what a smart man once said before: I dont need 2000 scientists to tell me that I am wrong. What I need is one scientist to show what is wrong in my theory.
Please read my rebuttal to the points the referees made. The referees version are kept as close as possible to the original. I put some corrections in a square bracket [] to clarify the referees point. Referee comments appear in roman font like this sentence, and rebuttals appear in a courier font as this sentence.
192
APPENDIX B. WHAT THE ESTABLISHMENTS SCIENTISTS SAY any combinations of the two cases. explanation is given in Chapter 7. Nevertheless, an additional
4. The authors approach is useful? Gives perspective to a commonly used process engineering method (pQ2 ) in die casting. Some of the runner lengths chosen (1 meter) might be consider exceptional in die casting yet the author uses this to show how much in error an average value for CD be. The author might also note that the North American Die Casting Association and many practitioners use a A3 /A2 ratio of .65 as a design target for gating. The author analysis reinforces this value as a good target, and that straying far from it may results in poor design part lling problems (Fig. 5) The reviewer refers to several points which are important to address. All the four sizes show large errors (we do not need to take 1[m] to demonstrate that). The one size, the referee referred to as exceptional (1 meter), is not the actual length but the represented length (read the article again). Poor design can be represented by a large length. This situation can be found throughout the die casting industry due to the common model which does not consider runner design. My office is full with runner designs with represent 1 meter length such as one which got NADCAs design award2 . In regards to the area ratio, please compare with the other referee who claim A2 /A2 = 0.8 - 0.95. I am not sure which of you really represent NADCAs position (I didnt find any of NADCAs publication in regards to this point). I do not agree with both referees. This value has to be calculated and cannot be speculated as the referees asserted. Please find an explanation to this point in the paper or in even better in Chapter 7.
B.3 Referee 2
There are several major concerns I have about this paper. The [most] major one [of these] is that [it] is unclear what the paper is attempting to accomplish. Is the paper trying to suggest a new way of designing the rigging for a die casting, or is it trying to add an improvement to the conventional pQ2 solution, or is it trying to somehow suggest a mathematical basis for the pQ2 diagram ? The paper shows that 1) the conventional pQ2 solution is totally wrong, 2) the mathematical analytical solution for the pQ2 provides an excellent tool for studying the effect of various parameters.
2 to
B.3. REFEREE 2
193
The other major concern is the poor organization of the ideas which the authors [are] trying to present. For instance, it is unclear how specic results presented in the results section where obtained ([for instance] how were the results in Figures 5 and 6 calculated?). I do not understand how the organization of the paper relates to the fact that the referee does not understand how Fig 5 and 6 were calculated. The organization of the paper does not have anything do with his understanding the concepts presented. In regard to the understanding of how Figure 5 and 6 were obtained, the referee should referred to an elementary fluid mechanics text book and combined it with the explanation presented in the paper. Several specic comments are written on the manuscript itself; most of these were areas where the reviewer was unclear on what the authors meant or areas where further discussion was necessary. One issue that is particularly irksome is the authors tendency in sections 1 and 2 to wonder [wander] o with editorials and other unsupported comments which have no place in a technical article. Please show me even one unsupported comment!! Other comments/concerns include what does the title have to do with the paper? The paper does not dene what a pQ2 diagram is and the results dont really tie in with the use of such diagrams.
The paper presents the exact analytical solution for the pQ2 diagram. The results tie in very well with the correct pQ2 diagram. Unfortunately, the common model is incorrect and so the results cannot be tied in with it. p.4 The relationship Q P is a result of the application of Bernoullis equation system like that shown in Fig 1. What is the rational or basis behind equation 1; e.g. Q (1 P )n with n =1, 1/2, and 1/4? Here I must thank the referee for his comment! If the referee had serious problem understanding this point, I probably should have considered adding a discussion about this point, such as in Chapter 7.
p.5 The relationship between equation 1(a) to 1(c) and a die casting machine as poor, common, and excellent performance is not clear and needs to be developed, or at least dened.
194
APPENDIX B. WHAT THE ESTABLISHMENTS SCIENTISTS SAY with similar gating layouts in the future. But because most dies have numerous gates branching o of numerous runners, to determine all of the friction factors as a function of Reynolds number would be quite dicult and virtually untractable for design purposes. Generally die casting engineers nd conventional pQ2 approach works quite well for design purposes. This several points comment give me the opportunity to discusses the following points: I would kindly ask the referee, to please provide the names of any companies whom experimentally determine CD . Perhaps they do it down under (Australia) where the regular physics laws do not apply (please, forgive me about being the cynical about this point. I cannot react to this any other way.). Please, show me a company that uses the common pQ2 diagram and it works. Due to the computer revolution, today it is possible to do the calculations of the CD for a specific design with a specific flow rate (die casting machine). In fact, this is exactly what this paper all about. Moreover, today there is a program that already does these kind of calculations, called DiePerfect. Here the referee introduce a new idea of the family -the improved constant CD . In essence, the idea of family is improve constant CD in which one assigned value to a specific group of runners. Since this idea violate the basic physics laws and the produces the opposite to realty trends it must be abandoned. Actually, the idea of family is rather bizarre, because a change in the design can lead to a significant change in the value of CD . Furthermore, the family concept can lead to a poor design (read about this in the section poor design effects of this book). How one can decided which design is part of what family? Even if there were no mistakes, the authors method (calculating the CD ) is of course cheaper and faster than the referees suggestion about family of runner design. In summary, this idea a very bad idea. What is CD =constant? The referee refers to the case where CD is constant for specific runner design but which is not the case in reality. The CD does not depend only on the runner, but on the combination of the runner system with the die casting machine via the Re number. Thus, a specific runner design cannot have CD assigned to it.
B.4. REFEREE 3 The CD has to be calculated for any combination runner system with die casting machine. I would like to find any case where the common pQ2 diagram does work. Please read the proofs in Chapter 7 showing why it cannot work.
Discussion and results A great deal of discussion focuses on the regions where A3 /A2 0.1; yet in typical die casting dies A3 /A2 0.8 to 0.95.
195
Please read the comments to the previous referee In conclusion, its just a plain sloppy piece of work I hope that referee does not mind that I will use it as the chapter quote. (the Authors even have one of the references to their own publications sited incorrectly!). Perhaps, the referee should learn that magazines change names and, that the name appears in the reference is the magazine name at the time of writing the paper.
B.4 Referee 3
The following comments are not arranged in any particular order. General: The text has a number of errors in grammar, usage and spelling that need to be addressed before publication. p 6 1st paragraph - The rsts sentence says that the ow rate is a function of temperature, yet the rest of the paragraph says that it isnt. The rest of the paragraph say the flow rate is a weak function of the temperature and that it explains why. I hope that everyone agrees with me that it is common to state a common assumption and explain why in that particular case it is not important. I wish that more people would do just that. First, it would eliminate many mistakes that are synonymous with research in die casting, because it forces the smart researchers to check the major assumptions they make. Second, it makes clear to the reader why the assumption was made. p 6 - after Eq 2 - Should indicate immediately that the subscript[s] refer to the sections in Figure 1. I will consider this, Yet, I am not sure this is a good idea. p 6 - after equation 2 - There is a major assumption made here that should not pass without some comment[s]3 Assuming steady state - This assumption goes
3 Is
196
to the heart of this approach to the lling calculation and establishes the bounds of its applicability. The authors should discuss this point. Well, I totally disagree with the referee on this point. The major question in die casting is how to ensure the right range of filling time and gate velocity. This papers main concern is how to calculate the CD and determine if the CD be assigned to a specific runner. The unsteady state is only a side effect and has very limited importance due to AESS. Of course the flow is not continuous/steady state and is affected by many parameters such as the piston weight, etc, all of which are related to the transition point and not to the pQ2 diagram per se. The unsteady state exists only in the initial and final stages of the injection. As a general rule, having a well designed pQ2 diagram will produce a significant improvement in the process design. It should be noted that a full paper discussing the unsteady state is being prepared for publication at the present time. In general the organization of the paper is somewhat weak - the introduction especially does not very well set the technical context for the pQ2 method and show how the present work ts into it. The present work does not fit into past work! It shows that the past work is wrong, and builds the mathematical theory behind the pQ2 diagram. The last paragraph of the intro is confused [confusing]. The idea introduced in the last sentence on page 2 is that the CD should vary somehow during the calculation, and subsequently variation with Reynolds number is discussed, but the intervening material about geometry eects is inconsistent with a discussion of things that can vary during the calculation. The last two sentences do not t together well either - the assumption of constant CD is not valid - okay, but is that what you are going to talk about, or are you going to talk about particularly the eects of the gate area? Firstly, CD should not vary during the calculations it is a constant for a specific set of runner system and die casting machine. Secondly, once any parameter is changed, such as gate area, CD has to be recalculated. Now the referees statement CD should vary, isnt right and therefore some of the following discussion is wrong. Now about the fitting question. What do referee means by fit together? Do the paper has to be composed in a rhyming verse? Anyhow, geometrical effects are part of Reynolds number (review fluid mechanics). Hence, the effects of the gate area shows that CD varies as well and has to be recalculated. So what is inconsistent? How do these sentences not fit together? On p 8, after Eq 10 - I think that it would be a good idea to indicate immediately
B.4. REFEREE 3
197
that these equations are plotted in Figure 3, rather than waiting to the next section before referring to Fig 3. Also, making the Oz-axis on this graph logarithmic would help greatly in showing the dierences in the three pump characteristics. Mentioning the figure could be good idea but I dont agree with you about the log scale, I do not see any benefits. On p. 10 after Eq 11 - The solution of Eq 11 requires full information on the die casting machine - According to this model, the machine characterized by Pmax, Qmax and the exponent in Eq 1. The wording of this sentence, however, might be indicating that there is some information to be had on the machine other that these three parameters. I do not think that that is what the authors intend, but this is confusing. This is exactly what the authors intended. The model does not confined to a specific exponent or function, but rather gives limiting cases. Every die casting machine can vary between the two extreme functions, as discussed in the paper. Hence, more information is needed for each individual die casting machine. p 12 - I tend to disagree with the premises of the discussion following Eq 12. I think that Qmax depends more strongly on the machine size than does Pmax. In general, P max is the intensication pressure that one wants to achieve during solidication, and this should not change much with the machine size, whereas the clamping force, the product of this pressure and platten are, goes up. On the other hand, when one has larger area to make larger casting, one wants to increase the volumetric ow rate of metal so that ow rate of metal so that ll times will not go up with the machine size. Commonly, the shot sleeve is larger, while the maximum piston velocity does not change much. Here the referee is confusing so many different concepts that it will take a while to explain it properly. Please find here a attempt to explain it briefly. The intensification pressure has nothing to do with the pQ2 diagram. The pQ2 does not have much to do with the solidification process. It is designed not to have much with the solidification. The intensification pressure is much larger than Pmax . I give up!! It would take a long discussion to teach you the fundamentals of the pQ2 diagram and the die casting process. You confuse so many things that it impossible to unravel it all for you in a short paragraph. Please read Chapter 7 or even better read the whole book. Also, following Eq 13, the authors should indicate what they mean by middle range of the Oz numbers. It is not clear from Fig 3 how close one needs to get to Oz=0 for the three curves to converge again. The mathematical equations are given in the paper. They are very simple that you can use hand calculator to find how much close you
198
need to go to Oz = 0 for your choice of error. A discussion on such issue is below the level from an academic paper. . Besides being illustrative of the results, part of the value of an example calculation comes from it making possible duplication of the results elsewhere. In order to support this, the authors need to include the relationships that used for CD in these calculations. The literature is full of such information. If the referee opens any basic fluid mechanics text book then he can find information about it. The discussion on p 14 of Fig 5 needs a little more consideration. There is a maximum in this curve, but the authors criterion of being on the right hand branch is said to be shorter ll time, which is not a criterion for choosing a location on this curve at all. The ll time is monotone decreasing with increasing A3 at constant A2, since the ow is the product of Vmax and A3. According to this criterion, no calculation is needed - the preferred conguration is no gate whatsoever. Clearly, choosing an operating point requires introduction of other criteria, including those that the authors mention in the intro. And the end of the page 14 discussion that the smaller lling time from using a large gate (or a smaller runner!!??) will lead to a smaller machine just does not follow at all. The machine size is determined by the part size and the required intensication pressure, not by any of this. Once again the referee is confusing many issues; let me interpret again what is the pQ2 diagram is all about. The pQ2 diagram is for having an operational point at the right gate velocity and the right filling time. For any given A2 , there are two possible solutions on the right hand side and one on the left hand side with the same gate velocity. However, the right hand side has smaller filling time. And again, the referee confusing another issue. Like in many engineering situations, we have here a situation in which more than one criterion is needed. The clamping force is one of the criteria that determines what machine should be chosen. The other parameter is the pQ2 diagram. It seems that they authors have obscured some elementary results by doing their calculations.4 For example, the last sentence of the middle paragraph on p 15 illustrates that as CD reaches its limiting value of 1, the discharge velocity reaches its maximum. This is not something we should be publishing in 1998. CD ? There is no mention of the alleged fact of CD reaches its limiting value of 1. There is no discussion in the whole article
4 If it is so elementary how can it be obscured. I have brokenout this paragraph for purposes of illustration.
B.4. REFEREE 3
199
about CD reaching its maximum (CD = 1). Perhaps the referee was mistakenly commenting on different articles (NADCAs book or an other die casting book) which he has confused with this article. Regarding the concluding paragraph on p 15: 1. The use of the word constant is not consistent throughout this paper. Do they mean constant across geometry or constant across Reynolds number, or both. To the readers: The referee means across geometry as different geometry and across Reynolds number as different Re number5 . I really do not understand the difference between the two cases. Arent actually these cases the same? A change in geometry leads to a change in Reynolds number number. Anyhow, the referee did not consider a completely different possibility. Constant CD means that CD is assigned to a specific runner system, or like the common model in which all the runners in the world have the same value. 2. Assuming that they mean constant across geometry, then obviously, using a xed value for all runner/gate systems will sometimes lead to large errors. They did not need to do a lot calculation to determine this. And yet this method is the most used method in the industry(some even will say the exclusive method). 3. Conversely, if they mean constant across Reynolds number, i.e. CD can vary through the run as the velocity varies, then they have not made their case very well. Since they have assumed steady state and the P3 does not enter into the calculation, then the only reason that mention for the velocity to vary during the ll would be because Kf varies as a function of the ll fraction. They have not developed this argument suciently. Let me stress again the main point of the article. CD varies for different runners and/or die casting machines. It is postulated that the velocity does not vary during run. A discussion about P3 is an entirely different issue related to the good venting design for which P3 remains constant. 4. If the examples given in the paper do not represent the characteristics of a typical die casting machine, why to present them at all? Why are the more detailed calculations not presented, instead of the trivial results that are shown?
5 if
the interpretation is not correct I would like to learn what it really mean.
200
APPENDIX B. WHAT THE ESTABLISHMENTS SCIENTISTS SAY These examples demonstrate that the common method is erroneous and that the authors method should be adopted or other methods based on scientific principles. I believe that this is a very good reason.
201
202
203 grant? Working on the same research project with this Mr. Pribyl was Dr. Brevick who also received a grant mentioned above. Is there a connection between the fact that Gary Pribyl cooperated with Dr. B. Brevick on the sensor project and you deciding to renew Dr. Brevicks grant on the critical slow plunger velocity project? I would like to learn what the reasoning for continuing to fund Dr. Brevick after you had learned that his research was problematic. Additionally, I learned that Mr. Steve Udvardy was given a large amount of money to study distance communications. I am sure that Mr. Udvardy can enhance NADCAs ability in distance learning and that this is why he was awarded this grant. I am also sure that Mr. Udvardy has all the credentials needed for such research. One can only wonder why his presentation was not added to the NADCA proceedings. One may also wonder why there is a need to do such research when so much research has already been done in this area by the worlds foremost educational experts. Maybe it is because distance communication works dierently for NADCA. Is there a connection between Mr. Steve Udvardy being awarded this grant and his holding a position as NADCAs research director? I would like to learn the reasons you vouchsafe this money to Mr. Udvardy! I also would like to know if Mr. Udvardys duties as director of education include knowledge and research in this area. If so, why is there a need to pay Mr. Udvardy additional monies to do the work that he was hired for in the rst place? We were informed by Mr. Walkington on the behalf of NADCA in the NovDec 1996 issue that around March or April 1997, we would have the software on the critical slow plunger velocity. Is there a connection between this softwares apparent delayed appearance and the fact that the research in Ohio has produced totally incorrect and obase results? I am sure that there are reasons preventing NADCA from completing and publishing this software; I would just like to know what they are. I am also sure that the date this article came out (Nov/Dec 1996) was only coincidentally immediately after I sent you my paper and proposal on the shot sleeve (September 1996). What do you think? Likewise, I learned that Mr. Walkington, one of the governors of NADCA, also received a grant. Is there a connection between this grant being awarded to Mr. Walkington and his position? What about the connection between his receiving the grant and his former position as the director of NADCA research? I am sure that grant was awarded based on merit only. However, I have serious concerns about his research. I am sure that these concerns are unfounded, but I would like to know what Mr. Walkingtons credentials are in this area of research. The three most important areas in die casting are the critical slow plunger velocity, the pQ2 diagram, and the runner system design. The research sponsored by NADCA on the critical slow plunger velocity is absolutely unfounded because it violates the basic physics laws. The implementation of the pQ2 diagram is also absolutely unsound because again, it violates the basic physics laws. One of the absurdities of the previous model is the idea that plunger diameter has to decrease in order to increase the gate velocity. This conclusion (of the previous model)
204
violates several physics laws. As a direct consequence, the design of the runner system (as published in NADCA literature) is, at best, extremely wasteful. As you also know, NADCA, NSF, the Department of Energy, and others sponsoring research in these areas exceed the tens of millions, and yet produce erroneous results. I am the one who discovered the correct procedure in both areas. It has been my continuous attempt to make NADCA part of these achievements. Yet, you still have not responded to my repeated requests for a 1 grant. Is there a reason that it has taken you 1 2 years to give me a negative answer? Is there a connection between any of the above information and how long it has taken you? Please see the impressive partiallist of the things that I have achieved. I am the one who found Garbers model to be totally and absolutely wrong. I am also the one responsible for nding the pQ2 diagram implementation to be wrong. I am the one who is responsible for nding the correct pQ2 diagram implementation. I am the one who developed the critical area concept. I am the one who developed the economical runner design concept. In my years of research in the area of die casting I have not come across any research that was sponsored by NADCA which was correct and/or which produced useful results!! Is there any correlation between the fact that all the important discoveries (that I am aware of) have been discovered not inbut outside of NADCA? I would like to hear about anything in my area of expertise supported by NADCA which is useful and correct? Is there a connection between the foregoing issues and the fact that so many of the die casting engineers I have met do not believe in science? More recently, I have learned that your secretary/assistant, Tricia Margel, has now been awarded one of your grants and is doing research on pollution. I am sure the grant was given based on qualication and merit only. I would like to know what Ms. Margels credentials in the pollution research area are? Has she done any research on pollution before? If she has done research in that area, where was it published? Why wasnt her research work published? If it was published, where can I obtain a copy of the research? Is this topic part of Ms. Margels duties at her job? If so, isnt this a double payment? Or perhaps, was this an extra separated payment? Where can I obtain the nancial report on how the money was spent? Together we must promote die casting knowledge. I am doing my utmost to increase the competitiveness of the die casting industry with our arch rivals: the plastics industry, the composite material industry, and other industries. I am calling on everyone to join me to advance the knowledge of the die casting process. Thank you for your consideration. Sincerely, Genick BarMeir, Ph.D. cc: NADCA Board of Governors NADCA members Anyone who care about die casting industry
205
206
I do not agree with your statement in your December 6, 1996 letter to me stating that This paper is highly technical-too technical without a less-technical background explanation for our general readers . I do not believe that discounting your readers is helpful. I have met some of your readers and have found them to be very intelligent, and furthermore they really care about the die casting industry. I believe that they can judge for themselves. Nevertheless, I have yielded to your demand and have eliminated many of the mathematical derivations from this paper to satisfy your desire to have a simple presentation. This paper, however, still contains the essentials to be understood clearly. Please note that I will withdraw the paper if I do not receive a reply stating your intentions by January 1, 1998, in writing. I do believe this paper will change the way pQ2 diagram calculations are made. The pQ2 diagram, as you know, is the central part of the calculations and design thus the paper itself is of same importance. I hope that you really do see the importance of advancing knowledge in the die casting industry, and, hope that you will cooperate with those who have made the major progress in this area. Thank you for your consideration. Sincerely, Dr. Genick Bar-Meir cc: Boxter, McClimtic, Scott, Wilson, Holland, Behler, Dupre, and some other NADCA members ps: You probably know by now that Garbers model is totally and absolutely wrong including all the other investigations that where based on it, even if they were sponsored by NADCA. (All the researchers agreed with me in the last congress) Well that letter got him going and he managed to get me a letter in which he claim that he sent me his revisions. Well, read about it in my next letter dated January 7, 1998. Paul M. Bralower, NADCA, Editor 9701 West Higgins Road, Suite 880 Rosemont, IL 60018-4721 Dear Mr. Bralower: Thank you for your fax dated December 29, 1997 in which you alleged that you sent me your revisions to my paper The mathematical theory of the pQ2 diagram. I never receive any such thing!! All the parties that got this information and myself nd this paper to of extreme importance. I did not revise my paper according to your comments on this paper, again, since I did not receive any. I decided to revised the paper since I did not received any reply from you for more than 4 months. I revised according to your comments on my previous paper on the critical slow plunger velocity. As I stated in my letter
207 dated December 6, 1997, I sent you the revised version as I send to all the cc list. I resent you the same version on December 29, 1997. Please note that this is the last time I will send you the same paper since I believe that you will claim again that you do not receive any of my submittal. In case that you claim again that you did not receive the paper you can get a copy from anyone who is on the cc list. Please be aware that I changed the title of the paper (December, 6, 1997 version) to be How to calculate the pQ2 diagram correctly. I would appreciate if you respond to my e-Mail by January 14, 1998. Please consider this paper withdrawn if I will not hear from you by the mentioned date in writing (email is ne) whether the paper is accepted. I hope that you really do see the importance of advancing knowledge in the die casting industry, and, hope that you will cooperate with those who have made the major progress in this area. Sincerely, Dr. Genick Bar-Meir ps: You surely know by now that Garbers model is totally and absolutely wrong including all the other investigations that where based on it He responded to this letter and changed his attitude I thought. January 9, 1998. Dear Mr. Bar-Meir: Thank you for your recent article submission and this follow-up e-mail. I am now in possession of your article How to calculate the pQ2 diagram correctly. It is the version dated Jan. 2, 1998. I have read it and am prepared to recommend it for publication in Die Casting Engineer. I did not receive any earlier submissions of this article, I was confusing it with the earlier article that I returned to you. My apologies. However I am very pleased at the way you have approached this article. It appears to provide valuable information in an objective manner, which is all we have ever asked for. As is my policy for highly technical material, I am requesting technical personnel on the NADCA sta to review the paper as well. I certainly think this paper has a much better chance of approval, and as I said, I will recommend it. I will let you know of our decision in 2-3 weeks. Please do not withdraw itgive us a little more time to review it! I would like to publish it and I think technical reviewer will agree this time. Sincerely, Paul Bralower Well I waited for a while and then I sent Mr. Bralower a letter dated Feb 2, 1998. Dear Mr. Bralower, Apparently, you do not have the time to look over my paper as you promise. Even a negative reply will demonstrate that you have some courtesy. But apparently the paper is not important as your experts told you and I am only a small bothering cockroach. Please see this paper withdrawn!!!!
208
I am sorry that we do not agree that an open discussion on technical issues should be done in your magazine. You or your technical experts do not have to agree with my research. I believe that you have to let your readers to judge. I am sure that there is no other reasons to your decision. I am absolutely sure that you do not take into your consideration the fact that NADCA will have to stop teaching SEVERAL COURSES which are wrong according to this research. Thank you for your precious time!! Dr. Bar-Meir Please note that this letter and the rest of the correspondence with you in this matter will be circulated in the die casting industry. I am sure that you stand by your decision and you would like other to see this correspondence even if they are NADCA members. Here is the letter I received in return a letter from Paul Bralower Feb 5, 1998. Dear Mr. Bar-Meir: Ill have you know that you have inconvenienced me and others on our sta today with your untoward, unnecessary correspondence. If you had a working telephone or fax this e-mail would not be necessary. As it is I must reply to your letter and take it to someone elses oce and have them e-mail it to you right away. I tried to telephone you last week on Thurs. 1/29 with the news that we have agreed to publish your article, How to Calculate the pQ2 diagram correctly. I wanted to ask you to send the entire paper, with graphics and equations, on a disk. Because of the current status of our e-mail system, I would advise you not to e-mail it. Send it on any of the following: Syquest, Omega ZIP or Omega JAZ. Use Microsoft Oce 97, Word 6.0 or Word Perfect 6.0. The problem is I couldnt reach you by phone. I tried sending you a fax several times Thurs. and last Friday. There was no response. We tried a couple of dierent numbers that we had for you. Having no response, I took the fax and mailed it to you as a letter on Monday 2/2. I sent Priority 2-day Mail to your attention at Innovative Filters, 1107 16th Ave. S.E., Minneapolis, Minn, 55414. You should have received it today at latest if this address is correct for you, which it should be since it was on your manuscript. Now, while Im bending over backwards to inform you of your acceptance, you have the nerve to withdraw the paper and threaten to spread negative gossip about me in the industry! I know you couldnt have known I was trying to contact you, but I must inform you that I cant extend any further courtesies to you. As your paper has been accepted, I expect that you will cancel your withdrawl and send me the paper on disk immediately for publication. If not, please do not submit any further articles. My response to Paul M. Bralower. Feb 9, 1998 Dear Mr. Bralower: Thank you for accepting the paper How to calculate the pQ2 diagram correctly. I strongly believe that this paper will enhance the understanding of your readers
209 on this central topic. Therefore, it will help them to make wiser decisions in this area, and thus increase their productivity. I would be happy to see the paper published in Die Casting Engineer. As you know I am zealous for the die casting industry. I am doing my utmost to promote the knowledge and protability of the die casting industry. I do not apologize for doing so. The history of our correspondence makes it look as if you refuse to publish important information about the critical slow plunger velocity. The history shows that you lost this paper when I rst sent it to you in August, and also lost it when I resubmitted it in early December. This, and the fact that I had not heard from you by February 1, 1998, and other information, prompted me to send the email I sent. I am sure that if you were in my shoes you would have done the same. My purpose was not to insults anyone. My only aim is to promote the die casting industry to the best of my ability. I believe that those who do not agree with promoting knowledge in die casting should not be involved in die casting. I strongly believe that the editor of NADCA magazine (Die Casting Engineer) should be interested in articles to promote knowledge. So, if you nd that my article is a contribution to this knowledge, the article should be published. I do not take personal insult and I will be glad to allow you to publish this paper in Die Casting Engineer. I believe that the magazine is an appropriate place for this article. To achieve this publication, I will help you in any way I can. The A paper was written using LTEX, and the graphics are in postscript les. Shortly, I will send you a disc containing all the les. I will also convert the le to Word 6.0. I am afraid that conversion will require retyping of all the equations. As you know, WORD produces low quality setup and requires some time. Would you prefer to have the graphic les to be in TIFF format? or another format? I have enhanced the calculations resolution and please be advised that I have changed slightly the graphics and text. Thank you for your assistance. Sincerely, Dr. Genick Bar-Meir
210
Thank you for your consideration and assistance!! Genick Then I got a surprise: the person dealing with me was changed. Why? (maybe you, the reader, can guess what the reason is). I cannot imagine if the letter was an oer to buy me out. I just wonder why he was concerned about me not submitting proposals (or this matter of submitting for publication). He always returned a prompt response to my proposals, yah sure. Could he possibly have suddenly found my research to be so important. Please read his letter, and you can decide for yourself. Here is Mr. Steve Udvardy response on Fri, 24 Apr 1998 Genick, I have left voice mail for you. I wish to speak with you about what appears to be non-submittal of your proposal I instructed you to forward to CMC for the 1999 call. I can and should also respond to the questions you are posinjg to Paul. I can be reached by phone at 219.288.7552. Thank you, Steve Udvardy Since the deadline for that proposal had passed long before, I wondered if there was any point in submitting any proposal. Or perhaps there were exceptions to be made in my case? No, it couldnt be; I am sure that he was following the exact procedure. So, I then sent Mr. Udvardy the following letter. April 28, 1998 Dear Mr. Udvardy: Thank you very much for your prompt response on the behalf of Paul Bralower. As you know, I am trying to publish the article on the pQ2 diagram. I am sure that you are aware that this issue is central to die casting engineers. A better design and a signicant reduction of cost would result from implementation of the proper pQ2 diagram calculations. As a person who has dedicated the last 12 years of his life to improve the die casting industry, and as one who has tied his life to the success of the die casting industry, I strongly believe that this article should be published. And what better place to publish it than Die Casting Engineer? I have pleaded with everyone to help me publish this article. I hope that you will agree with me that this article should be published. If you would like, I can explain further why I think that this article is important. I am very glad that there are companies who are adopting this technology. I just wish that the whole industry would do the same. Again, thank you for your kind letter. Genick ps: I will be in my oce Tuesday between 9-11 am central time (612) 3782940
211 I am sure that Mr. Udvardy did not receive the comments of/from the referees (see Appendix B). And if he did, I am sure that they did not do have any eect on him whatsoever. Why should it have any eect on him? Anyhow, I just think that he was very busy with other things so he did not have enough time to respond to my letter. So I had to send him another letter. 5/15/98 Dear Mr. Udvardy: I am astonished that you do not nd time to answer my letter dated Sunday, April, 26 1998 (please see below copy of that letter). I am writing you to let you that there is a serious danger in continue to teach the commonly used pQ2 diagram. As you probably know (if you do not know, please check out IFIs web site www.dieperfect.com), the commonly used pQ2 diagram as it appears in NADCAs books violates the rst and the second laws of thermodynamics, besides numerous other common sense things. If NADCA teaches this material, NADCA could be liable for very large sums of money to the students who have taken these courses. As a NADCA member, I strongly recommend that these classes be suspended until the instructors learn the correct procedures. I, as a NADCA member, will not like to see NADCA knowingly teaching the wrong material and moreover being sued for doing so. I feel that it is strange that NADCA did not publish the information about the critical slow plunger velocity and the pQ2 diagram and how to do them correctly. I am sure that NADCA members will benet from such knowledge. I also nd it beyond bizarre that NADCA does not want to cooperate with those who made the most progress in the understanding die casting process. But if NADCA teaching the wrong models might ends up being suicidal and I would like to change that if I can. Thank you for your attention, time, and understanding! Sincerely, Dr. Genick Bar-Meir ps: Here is my previous letter. Now I got a response. What a dierent tone. Note the formality (Dr Bar-Meir as oppose to Genick). May 19, 1998 Dear Dr Bar-Meir, Yes, I am here. I was on vacation and tried to contact you by phone before I left for vacation. During business travel, I was sorry to not be able to call during the time period you indicated. As Paul may have mentioned, we have approved and will be publishing your article on calculating PQ2. The best t for this is an upcoming issue dedicated to process control. Please rest assured that it will show up in this appropriate issue of DCE magazine.
212
Since there has been communications from you to Paul and myself and some of the issues are subsequently presented to our Executive Vice President, Dan Twarog, kindly direct all future communications to him. This will assist in keeping him tied in the loop and assist in getting responses back to you. His e-mail address is [email protected]. Thank you, Steve Udvardy Why does Mr. Udvardy not want to communicate with me and want me to write to Executive Vice President? Why did they change the title of the article and omit the word correctly. I also wonder about the location in the end of the magazine. I have submitted other proposals to NADCA, but really never received a reply. Maybe it isnt expected to be replied to? Or perhaps it just got/was lost?
Is it all coincidental?
I had convinced Larry Winkler in mid 1997 (when he was still working for Hartzell), to ask Mr. Udvardy why NADCA continued support for the wrong models (teaching the erroneous Garbers model and fueling massive grants to Ohio State University). He
213 went to NADCA and talked to Mr. Udvardy about this. After he came back, he explained that they told him that I didnt approach NADCA in the right way. (what is that?) His enthusiasm then evaporated, and he continues to say that, because NADCA likes evolution and not revolution, they cannot support any of my revolutionary ideas. He suggested that I needed to learn to behave before NADCA would ever cooperate with me. I was surprised and shaken. What happened, Larry? I asked him. But I really didnt get any type of real response. Later (end of 1997) I learned he had received NADCAs design award. You, the reader, can conclude what happened; I am just supplying you with the facts. Several manufacturers of die casting machines, Buler, HPM, Prince, and UBE presented their products in Minneapolis in April 1999. When I asked them why they do not adapt the new technologies, with the exception of the Buler, the response was complete silence. And just Buler said that they were interested; however, they never later called. Perhaps, they lost my phone number. A representative from one of the other companies even told me something on the order of Yeah, we know that the Garber and Brevick models are totally wrong, but we do not care; just go awayyou are bothering us!. I have news for you guys: the new knowledge is here to stay and if you want to make the die casting industry prosper, you should adopt the new technologies. You should make the die casting industry prosper so that you will prosper as well; please do not look at the short terms as important. The next issue of the Die Casting Engineer (May/Jun 1999 issue) was dedicated to machine products. Whether this was coincidental, you be the judge. I submitted a proposal to NADCA (November 5, 1996) about Garber/Brevick work (to which I never received a reply). Two things have happened since: I made the proposal(in the proposal I demonstrate that Brevicks work from Ohio is wrong) 1) publishing of the article by Bill Walkington in NADCA magazine about the wonderful research in Ohio State University and the software to come. 2)a scientic article by EKK. During that time EKK also advertised how good their software was for shot sleeve calculations. Have you seen any EKK advertisements on the great success of shot sleeve calculations lately? Here is another interesting coincidence, After 1996, I sent a proposal to NADCA, the cover page of DCE showing the beta version of software for calculating the critical slow plunger velocity. Yet, no software has ever been published. Why? Is it accidental that the author of the article in the same issue was Bill Walkington. And after all this commotion I was surprised to learn in the (May/June 1999) issue of DCE magazine that one of the Brevick group had received a prize (see picture below if I get NADCA permission). I am sure that Brevicks group has made so much progress in the last year that this is why the award was given. I just want to learn what these accomplishments are. For a long time NADCA described the class on the pQ2 diagram as a A close mathematical description. After I sent the paper and told them about how the pQ2 diagram is erroneous, they change the description. Well it is good, yet they have to say that in the past material was wrong and now they are teaching something else. or
214
are they? I have submitted ve (5) papers to the conference (20th in Cleveland) and four (4) have been rejected on the grounds well, you can read the letter yourself: Here is the letter from Mr. Robb. 17 Feb 1999 The International Technical Council (ITC) met on January 20th to review all submitted abstracts. It was at that time that they downselected the abstracts to form the core of each of the 12 sessions. The Call for Papers for the 1999 Congress and Exposition produced 140 possible abstracts from which to choose from, of this number aproximately 90 abstracts were selected to be reviewed as nal papers. I did recieve all 5 abstracts and distribute them to the appropriate Congress Chairmen. The one abstract listed in your acceptance letter is in fact the one for which we would like to review the nal paper. The Congress Chairmen will be reviewing the nal papers and we will be corresponding with all authors as to any changes revisions which are felt to be appropriate. The Congress Chairmen are industry experts and it is there sole discretion as to which papers are solicited based on abstract topic and t to a particular session. It is unfortunate that we cannot accept all abstracts or papers which are submitted. Entering an abstract does not constitue an automatic acceptance of the abstract/or nal paper. Thank you for your inquiry, and we look forward to reviewing your nal paper. Regards, Dennis J. Robb NADCA I must have submitted the worst kind of papers otherwise. How can you explain that only 20% of my papers (1 out of 5) accepted. Note that the other researchers ratio of acceptance on their papers is 65%, which means that other papers are three times better than mine. Please nd here the abstracts and decide if youd like to hear such topics or not. Guess which the topic NADCA chose, in what session and on what day (third day).
215 in other cases the results are not satisfactory. One of the keys to the success is the design of the vacuum system, especially the vacuum tank. The present study deals with what are the design requirements on the vacuum system. Design criteria are presented to achieve an eective vacuum system.
How Cutting Edge technologies can improve your Process Design approach
A proper design of the die casting process can reduce the lead time signicantly. In this paper a discussion on how to achieve a better casting and a shorter lead time utilizing these cutting edge technologies is presented. A particular emphasis is given on the use of the simplied calculations approach.
On the eect of runner design on the reduction of air entrainment: Two Chamber Analysis
Reduction of air entrapment reduces the product rejection rate and always is a major concern by die casting engineers. The eects of runner design on the air entrapment have been disregarded in the past. In present study, eects of the runner design characteristics are studied. Guidelines are presented on how to improve the runner design so that less air/gas are entrapped.
Afterward
At the 1997 NADCA conference I had a long conversation with Mr. Warner Baxter. He told me that I had rued a lot of feathers in NADCA. He suggested that if I wanted to get real results, I should be politically active. He told me how bad the situation had been in the past and how much NADCA had improved. But here is something I cannot understand: isnt there anyone who cares about the die casting industry and who wants it to ourish? If you do care, please join me. I actually have found some individuals who do care and are supporting my eorts to increase scientic knowledge in die casting. Presently, however, they are a minority. I hope that as Linux is liberating the world from Microsoft, so too we can liberate and bring prosperity to the die casting industry.
216
After better than a year since my rst (and unsent) letter to Steve Udvardy, I feel that there are things that I would like to add to the above letter. After my correspondence with Paul Bralower, I had to continue to press them to publish the article about the pQ2 . This process is also described in the preceding section. You, the reader, must be the judge of what is really happening. Additionally, open questions/discussion topics to the whole die casting community are added. What happened to the Brevicks research? Is there still no report? And does this type of research continue to be funded? Can anyone explain to me how NADCA operates? Is NADCA, the organization, more important than the die casting industry?
BIBLIOGRAPHY
[1] ASM. Metals Handbook , volume 13. ASM, Metals Park, Ohio, 1987. [2] G. Backer and Frank Sant. Using Finite Element Simulation for the Development of Shot Sleeve Velocity Proles . In NADCA 19th Congress and exposition, Minneapolis, Minnesota, November 1997. paper T97-014. [3] G. Bar-Meir, E.R.G. Eckert, and R. J. Goldstein. Pressure die casting: A model of vacuum pumping. Journal of Engineering for Industry , 118:001 007, February 1996. [4] Genick Bar-Meir. On gas/air porosity in pressure die casting . PhD thesis, University of Minnesota, 1995. [5] Genick Bar-Meir. Analysis of mass transfer processes in the pore free technique. Journal of Engineering Materials and Technology , 117:215 219, April 1995. [6] A. Bochvar, A., M. Notkin, E., S. I. Spektorova, and N.M. Sadchikova. The Study of Casting Systems by Means of Models. Izvest. Akad. Nauk U.S.S.R. (Bulletin of the Academy of Sciences of U.S.S.R) , pages 875882, 1946. [7] Hydraulics Laboratory Branch. Hydraulic Design of Stilling Basins and Bucket Energy Dissipatiors . U.S. Bureau of Reclamation, Engineering Monograph 25, Denver, Colorado, 1958. [8] Jerald R. Brevick, Dwaine J. Armentrout, and Yeou-Li Chu. Minimization of entrained gas porosity in aluminum horizontal cold chamber die castings . Transactions of NAMRI/SME , 12:4146, November - December 1994. [9] Derek L. Cocks. DCRF Recommended Procedures: Metal Flow Predictor System. American Die Casting Institute, Inc., Des Plaines, Illinois, 1986.
217
218
BIBLIOGRAPHY
[10] Derek L. Cocks and A. J. Wall. Technology transfer in the united kingdom: Progress and prospects. In Transactions 12th International SDCE, Minneapolis, Minnesota, 1983. paper G-T83-074. [11] K Davey and S. Bounds. Modeling the pressure die casting process using boundary and nite element methods. Journal of material Processing Technology, 63:696700, 1997. [12] A. J. Davis. Eects of the relationship between molten metal ow in feed systems and hydraulic uid ow in die casting machines. In Transactions 8th International SDCE, St. Louis, 1975. paper G-T75-124. [13] E.R.G. Eckert. Similarity analysis applied to the Die Casting Process . Journal of Engineering Materials and Technology , 111:393398, 1989. No. 4 Oct. [14] Mohamed El-Mehalawi, Jihua Liu, and R. A. Miller. A cost estimating model for die cast products. In NADCA 19th Congress and exposition, Minneapolis, Minnesota, NOvember 1997. paper T97-044. [15] Fairbanks. Hydraulic HandBook. Mores and Co., Kansas city, Kansas, 1959. [16] H. Fondse, H. Jeijdens, and G Ooms. On the inuence of the exit conditions on the entrainment rate in the development region of a free, round, turbulent jet. Applied Scientic Research, pages 355375, 1983. [17] L. W. Garber. Theoretical analysis and experimental observation of air entrapment during cold chamber lling. Die Casting Engineer , 26 No. 3:33, May - June 1982. [18] G. Hansen, Arthur. Fluid Mechanics. John Wiley and Sons, Inc., New York, New York, 1967. [19] F. M. Henderson. Open Channel Flow. Macmillan Publishing Co., New York, New York, 1966. [20] A. G. Horacio and R. A. Miller. Die casting die deections: computer simulation of causes and eects. In NADCA 19th Congress and exposition, Minneapolis, Minnesota, NOvember 1997. paper T97-023. [21] Henry Hu and S. A. Argyropoulos. Mathematical modeling of solidication and melting: a review . Modeling Simulation Mater. Sci. Eng. , 4:371396, 1996. [22] J. Hu, S. Ramalingam, G. Meyerson, E.R.G. Eckert, and R. J. Goldstein. Experiment and computer modeling of the ows in pressure die casting casings. In ASME/CIE Design, San Francisco, California, 1992. [23] C. M. Kim and Frank J. Sant. An application of 3D solidication analysis to large complex castings. In 2nd Pacic rim international conference on modeling of casting and solidication, Singapore, January 1995.
BIBLIOGRAPHY [24] P.A. Madsen and Svendsen. Turbulent bores and hydraulic jumps. Journal of Fluid Mechanics, 129:125, 1983.
219
[25] Ralph David Maier. Inuence of liquid metal jet character on heat transfer during die casting. PhD thesis, Case Western Reserve University, 1974. Engineering, metallurgy. [26] T Nguyen and J. Carrig. Water Analogue Studies of Gravity Tilt Casting Copper Alloy components. AFS Trans , pages 519528, 1986. [27] Richard H. F. Pao. Fluid Mechanics. John Wiley and Sons, Inc., New York, New York, 1961. [28] N. Rajaratnam. The hydraulic jump as a wall jet. Journal of Hydraulic Div. ASCE, pages 107131, 1965. 91 (HY5). [29] Ascher H. Shapiro. The Dynamics and thermodynamics of Compressible Fluid Flow , volume I. John Wiley and Sons, New York, 1953. [30] W. F. Stuhrke and J. F. Wallace. Gating of die castings. Transactions of American Foundrymens Society , 73:569597, 1966. [31] Marilyn Thome and Jerald R. Brevick. Optimal slow shot velocity proles for cold chamber die casting. In NADCA Congress and exposition, Indianapolis, Indiana, October 1995. paper T95-024. [32] I Wygnanski and F. H. Champan. The origin of pus and slugs and the ow in a turbulent slug. J. Fluid Mechanics , pages 281335, 1973.
220
BIBLIOGRAPHY
Subjects index
A
absolute viscosity, 13 adiabatic nozzle, 25
B
Bernoullis equation, 22
C
convergingdiverging nozzle, 23
F
fanno second law, 145 fanno ow, 143, 4f L 147 D choking, 148 average friction factor, 149 entrance Mach number calculations, 157, 175 entropy, 148 shockless, 155, 156 star condition, 150
M
Mach number, 23
S
shear stress, 12 speed of sound star, 24 speed of sound, what, 21 stagnation state, 23
AUTHORS INDEX
221
Authors index