Sei sulla pagina 1di 308

The Pennsylvania State University The Graduate School Department of Mechanical and Nuclear Engineering

MULTI LEVEL OPTIMIZATION OF BURNABLE POISON UTILIZATION FOR ADVANCED PWR FUEL MANAGEMENT

A Thesis in Nuclear Engineering

by Serkan Yilmaz

2005 Serkan Yilmaz

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy August 2005

The thesis of Serkan Yilmaz was reviewed and approved* by the following:

Kostadin Ivanov Professor of Nuclear Engineering Thesis Advisor Co-Chair of Committee Yousry Y., Azmy Professor of Nuclear Engineering Co-Chair of Committee Arthur T., Motta Professor of Nuclear Engineering Samuel Levine Professor Emeritus of Nuclear Engineering Soundar Kumara Distinguished Professor of Industrial Engineering Moussa Mahgerefteh Senior Nuclear Engineer Exelon Nuclear Company Special Member

Jack Brenizer, Jr. Professor of Nuclear Engineering Chair of Nuclear Engineering Program

*Signatures are on file in the Graduate School

ABSTRACT
The objective of this study was to develop a unique methodology and a practical tool for designing burnable poison (BP) pattern for a given PWR core. Two techniques were studied in developing this tool. First, the deterministic technique called Modified Power Shape Forced Diffusion (MPSFD) method followed by a fine tuning algorithm, based on some heuristic rules, was developed to achieve this goal. Second, an efficient and a practical genetic algorithm (GA) tool was developed and applied successfully to Burnable Poisons (BPs) placement optimization problem for a reference Three Mile Island-1 (TMI-1) core. This thesis presents the step by step progress in developing such a tool.

The developed deterministic method appeared to perform as expected. The GA technique produced excellent BP designs. It was discovered that the Beginning of Cycle (BOC) Kinf of a BP fuel assembly (FA) design is a good filter to eliminate invalid BP designs created during the optimization process. By eliminating all BP designs having BOC Kinf above a set limit, the computational time was greatly reduced since the evaluation process with reactor physics calculations for an invalid solution is canceled.

Moreover, the GA was applied to develop the BP loading pattern to minimize the total Gadolinium (Gd) amount in the core together with the residual binding at End-ofCycle (EOC) and to keep the maximum peak pin power during core depletion and Soluble boron concentration at BOC both less than their limit values. The number of

iv UO2/Gd2O3 pins and Gd2O3 concentrations for each fresh fuel location in the core are the decision variables and the total amount of the Gd in the core and maximum peak pin power during core depletion are in the fitness functions. The use of different fitness function definition and forcing the solution movement towards to desired region in the solution space accelerated the GA runs. Special emphasize is given to minimizing the residual binding to increase core lifetime as well as minimizing the total Gd amount in the core.

The GA code developed many good solutions that satisfy all of the design constraints. For these solutions, the EOC soluble boron concentration changes from 68.9 to 97.2 ppm. It is important to note that the difference of 28.3 ppm between the best and the worst solution in the good solutions region represent the potential of 12.5 EffectiveFull-Power-Day (EPFD) savings in cycle length. As a comparison, the best BP loading design has 97.2 ppm soluble boron concentration at EOC while the BP loading with available vendors U/Gd FA designs has 94.4 ppm SOB at EOC. It was estimated that the difference of 2.8 ppm reflected the potential savings of 1.25 EFPD in cycle length. Moreover, the total Gd amount was reduced by 6.89 % in mass that provided extra savings in fuel cost compared to the BP loading pattern with available vendors U/Gd FA designs.

In addition, enriched Gd-155 and Gd-157 optimized BP designs were developed to further reduce residual binding of natural high (7-8 w/o) Gd2O3 concentrated U/Gd FAs. The optimized enriched Gd-155 designs successfully replaced the natural high (7-8

v w/o) Gd2O3 concentrated U/Gd FAs in the reference BP loading. The residual Gd reactivity effect was analyzed for these optimized designs. Cost analysis comparison was performed to evaluate the potential economical benefits of the optimized BP designs as compared to the conventional BP designs and to determine break-even point for unit Gd enrichment cost.

TABLE OF CONTENTS
LIST OF FIGURES .......................................................................................................... xii LIST OF TABLES........................................................................................................... xxi ACKNOWLEDGMENTS ............................................................................................. xxiv CHAPTER 1 INTRODUCTION ....................................................................................... 1 1.1 Problem Overview .................................................................................................... 1 1.2 Background ............................................................................................................... 3 1.2.1 Burnable Absorbers and Gadolinium Utilization .............................................. 3 1.2.2 Deterministic and Stochastic Optimization Techniques for Loading Pattern (LP) and BP design problems ..................................................................................... 4 1.2.3 Optimized Enriched Gadolinia Concept ............................................................ 9 1.2.4 Burnable Poison Fabrication Cost ................................................................... 12 1.3 Statement of Objective and Thesis Organization.................................................... 14 CHAPTER 2 DETERMINISTIC METHODOLOGY...................................................... 16 2.1 Introduction............................................................................................................. 16 2.2 The MPSFD Computer Program ............................................................................ 21 2.3 Fine-tuning Heuristic Selection Technique ............................................................ 25 2.4 Final Optimal BP Loading Design.......................................................................... 28 2.5 Verification of SIMULATE Results for Benchmark Core ..................................... 29 2.6 Fuel Rod Axial Zoning Model................................................................................ 35 2.7 Reference U/Gd FA designs ................................................................................... 37 2.8 Reference TMI-1 Core Loading.............................................................................. 39

vii 2.9 Superior BP Loading Design .................................................................................. 44 2.10 Final TMI-1 BP Loading Design Using Fine-Tuning Technique......................... 47 2.10 Conclusions........................................................................................................... 50 CHAPTER 3 OPTIMIZATION OF UO2/Gd2O3 FUEL PIN CONFIGURATIONS ...... 52 3.1 Introduction............................................................................................................. 52 3.2 Problem Definition.................................................................................................. 55 3.3 Genetic Algorithm Model ....................................................................................... 57 3.4 Input Parameters and Decision Variables ............................................................... 58 3.4.1 Design Constraints ........................................................................................... 62 3.5 Solution Representation, Encoding and Decoding Process, and Eliminating Bias 63 3.6 Genetic Operators ................................................................................................... 67 3.7 Optimization Calculations on a Single Fuel Assembly .......................................... 71 3.7.1 Minimizing Peak Pin Power within a Single Fuel Assembly during Depletion in an Infinite Array.................................................................................................... 71 3.7.2 GA Flow Diagram............................................................................................ 72 3.7.3 Results.............................................................................................................. 74 3.7.4 Core Depletion Comparison with Available U/Gd BP Designs ...................... 75 3.8 The Relation between BOC Kinf and RPD of a Fuel Assembly.............................. 79 3.8.1 Objective Function of Minimizing BOC Kinf for U/Gd FA Design ................ 79 3.8.2 Results.............................................................................................................. 80 3.9 Kinf Filter Use in Coupled Optimization Calculations ............................................ 82 3.9.1 The Relation between Kinf, Maximum Radial Power Distribution and Maximum Peak Pin Power........................................................................................ 82

viii 3.9.2 GA Algorithm with Kinf Filter Use .................................................................. 86 3.9.3 Objective Functions Definitions and Constraints Handling ............................ 89 3.9.4 Results.............................................................................................................. 91 3.10 Comparison of Methods........................................................................................ 98 3.11 The U/Gd Fuel Assembly Design with 20 U/Gd Fuel Pins ................................ 101 3.12 U/Gd Fuel Assembly Designs............................................................................. 103 3.13 Automation of CASMO Image File Generation................................................. 105 3.14 Conclusions......................................................................................................... 110 CHAPTER 4 OPTIMIZED ENRICHED BURNABLE POISON DESIGN.................. 111 4.1 Introduction........................................................................................................... 111 4.2 The Effect of Gd Absorber on Neutron Spectra and Pu Production..................... 113 4.3 Natural and Enriched Gd BP Design .................................................................... 114 4.4 Modeling Enriched Gadolinium and Genetic Algorithm Model .......................... 119 4.5 Optimization of Enriched Gadolinium Use in the Fresh Fuel Assemblies........... 119 4.5.1 100 w/o Gd-155 enriched with 2 w/o Gd2O3 (20 U/Gd fuel pins)................ 120 4.5.2 100 w/o Gd-155 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)................. 122 4.5.3 70 w/o Gd-157 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)................... 124 4.5.4 100 w/o Gd-157 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)................ 125 4.5.5 Summary of Results....................................................................................... 127 4.6 Enriched Gd-155 BP Design Characteristics........................................................ 128 4.7 Combined Enriched Gd-155 and Gd-157 BP Designs ......................................... 131 4.8 Optimized Pure (100 w/o) Gd-155 and Gd-157 Designs with 2 w/o Gd2O3 Concentration.............................................................................................................. 133

ix 4.9 Equivalent Enriched U/Gd Fuel Assembly Design .............................................. 135 4.10 Enriched Gd-157 and Gd-155 FA Designs in the Core Depletion Calculations 136 4.11 Pure (100 w/o) Gd-155 Enriched Designs with 2 w/o Gd2O3 Concentration..... 140 4.12 Final Heuristics BP Loading Pattern with 4 Optimized U/Gd FA Designs........ 142 4.13 Superior BP Loading Pattern with 16 Optimized U/Gd FA Designs ................. 143 4.14 Conclusions......................................................................................................... 146 CHAPTER 5 FUEL COST ANALYSIS ........................................................................ 147 5.1 Introduction..................................................................................................... 147

5.2 Calculation of Cost Components ...................................................................... 148 5.2.1 Calculation of Each Cost Components .......................................................... 149 5.2.2 Discounting and Levelizing Fuel Cycle Costs............................................... 151 5.3 Fuel Cycle and Cost Data ..................................................................................... 152 5.4 Fuel Cost Analysis ................................................................................................ 154 CHAPTER 6 CORE BURNABLE POISON OPTIMIZATION WITH GENETIC ALGORITHMS .............................................................................................................. 163 6.1 Introduction........................................................................................................... 163 6.2 Problem Definition................................................................................................ 165 6.2.1 Decision Variables ......................................................................................... 168 6.2.1 Design Constraints ......................................................................................... 169 6.3 Genetic Algorithm Model ..................................................................................... 170 6.4 Solution Representation, Phenotype and Genotype Structure .............................. 173 6.5 Mutation Probability Analysis .............................................................................. 182 6.6 GA Input Data....................................................................................................... 185

x 6.7 Test Runs .............................................................................................................. 186 6.7.1 Minimizing Maximum Peak Pin Power during Core Depletion.................... 186 6.7.2 Minimizing Total Gd Amount in the Core without any Constraint............... 190 6.8 Objective Function Definition and Constraints Handling .................................... 194 6.9 Results................................................................................................................... 196 6.9.1 Core BP Optimization with Reference UO2/Gd2O3 FA BP Designs............. 197 6.9.2 Core BP Optimization with Good UO2/Gd2O3 BP Designs .......................... 207 6.10 Comparison of Results........................................................................................ 218 CHAPTER 7 CONCLUSIONS AND FUTURE WORK............................................... 220 7.1 Conclusions........................................................................................................... 220 7.2 Summary of Contributions.................................................................................... 224 7.3 Suggestions for Future Work ................................................................................ 227 7.3.1 Hybrid Genetic Algorithm-Neural Network Application .............................. 227 7.3.2 Sensitivity on Fitness Function Definition .................................................... 232 7.3.3 Multi Objectives Optimization with Genetic Algorithm ............................... 232 7.3.4 Fuel Cost Analysis ......................................................................................... 233 7.3.5 Parallel Genetic Algorithms........................................................................... 234 7.3.6 Different GA Operators ................................................................................. 234 7.3.7 Simultaneous Core Loading and BP Placement Optimization ...................... 235 7.3.8 GA Methodology for Different Integral Fuel Burnable Absorbers ............... 235 7.3.9 Optimization of Burnable Absorber Distribution for Boling Water Reactor Fuel Assembly ........................................................................................................ 236 BIBLIOGRAPHY.......................................................................................................... 237

xi Appendix 1 Fuel Assembly Absorption Cross Section Data Base ................................. 242 Appendix 2 Sample CASMO-4 Branch Input ................................................................ 243 Appendix 3 TABLES Library Generation...................................................................... 246 Appendix 4 Sample SIMULATE Input .......................................................................... 255 Appendix 5 GA Code Input and Output Data for the Optimization of U/Gd Fuel Pin Configurations................................................................................................................. 274 Appendix 6 Core BP Optimization Input and Output Data ............................................ 280 Appendix 7 CASMO-4 Image Generator Input Files ..................................................... 282

LIST OF FIGURES
Figure 2-1 Algorithm for Gd placement ........................................................................... 24 Figure 2-2 Heuristic selection technique flow diagram.................................................... 27 Figure 2-3 Flow diagram of the minimal BP design algorithm ........................................ 28 Figure 2-4 Reference TMI-1 octant core loading ............................................................. 32 Figure 2-5 Reference TMI-1 core depletion results with available fuel assembly design on Exelon computer with RESTART files .................................................................... 33 Figure 2-6 Reference TMI-1 core depletion results with available fuel assembly design on Exelon computer without RESTART files ............................................................... 34 Figure 2-7 Reference TMI-1 core depletion results with available fuel assembly design on PSU computer without RESTART files ................................................................... 34 Figure 2-8 A sample UO2/Gd2O3 fuel rod axial zoning model for a TMI-1 fuel ............. 36 Figure 2-9 The vendor reference configuration for 4 U/Gd fuel pins............................... 37 Figure 2-10 The vendor reference configuration for 8 U/Gd fuel pins............................. 38 Figure 2-11 The vendor reference configuration for 16 U/Gd fuel pins........................... 38 Figure 2-12 Reference TMI-1 core loading with fresh fuel assemblies without BA........ 40 Figure 2-13 Reference TMI-1 Core burnup distribution without BA............................... 41 Figure 2-14 Radial Normalized power distribution for reference core loading without BA ................................................................................................................................... 42 Figure 2-15 Peak pin power distribution for reference TMI-1 initial core loading without BA ............................................................................................................................. 42 Figure 2-16 Soluble boron concentration change with cycle average exposure............... 43

xiii Figure 2-17 Superior TMI-1 BP loading pattern .............................................................. 45 Figure 2-18 Peak pin power distributions for superior BP loading core .......................... 46 Figure 2-19 Radial normalized power distributions for superior BP loading core........... 46 Figure 2-20 Final TMI-1 BP Loading pattern with deterministic method and fine tuning technique ................................................................................................................... 48 Figure 2-21 Peak pin power distributions for reference TMI-1 final BP loading design . 49 Figure 2-22 Radial normalized power distributions for reference TMI-1 final BP loading design ........................................................................................................................ 49 Figure 3-1 A reference TMI-1 octant PWR fuel assembly model.................................... 56 Figure 3-2 Data flow between Generic Code and Reactor Core or Lattice Physics Codes for a given input parameters ..................................................................................... 58 Figure 3-3 Decision space for the selection of UO2/ Gd2O3 fuel pin positions ................ 61 Figure 3-4 Geometric symmetry factors for octant fuel assembly ................................... 62 Figure 3-5 Encoding process for decision variable of UO2/Gd2O3 fuel pin position ....... 65 Figure 3-6 Sample phenotype and genotype structure...................................................... 67 Figure 3-7 Tournaments between the six population members and the resulting mating pool ........................................................................................................................... 69 Figure 3-8 Single Point and Uniform Cross over operators ............................................. 70 Figure 3-9 Bit-Wise mutation operator............................................................................. 71 Figure 3-10 Sample Genetic Algorithm flow diagram ..................................................... 73 Figure 3-11 TMI-1 reference core with available fuel assembly design .......................... 77 Figure 3-12 TMI-1 reference core design with fuel assembly design (at position H-09) with minimum assembly internal peak power during depletion............................... 78

xiv Figure 3-13 Max RPD power vs. Kinf at BOC for U/Gd FA designs ............................... 85 Figure 3-14 Max peak pin power vs. Kinf at BOC for U/Gd FA designs with Kinf filter.. 85 Figure 3-15 Coupled fuel lattice and core depletion calculations in GA optimization algorithm with BOC Kinf filter use............................................................................ 88 Figure 3-16 Solution space demonstration and its objective functions ............................ 89 Figure 3-17 Max. peak pin power vs. BOC Kinf value during GA run............................. 92 Figure 3-18 Max. peak pin power vs. EOC soluble boron concentration during GA run 93 Figure 3-19 Max. peak pin power vs. EOC soluble boron concentration for good designs ................................................................................................................................... 93 Figure 3-20 Best BP design within good solutions........................................................... 94 Figure 3-21 Design No 2 within good solutions............................................................... 95 Figure 3-22 Comparison of U/Gd fuel pin arrangements for different objective functions ................................................................................................................................. 100 Figure 3-23 CASMO-3/TABLES-3/SIMULATE-3 fuel management code package.... 106 Figure 3-24 Flow Diagram of CASMO-4 image file generator ..................................... 109 Figure 4-1 Neutron capture cross section of Gd isotopes versus energy........................ 112 Figure 4-2 Neutron capture cross sections of Pu-239 and U-238 vs. energy [49].......... 114 Figure 4-3 Comparison of Kinf curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel.................................................................................. 116 Figure 4-4 Comparison of residual Kinf curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel...................................................................... 117 Figure 4-5 Comparison of assembly peak pin power curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel....................................... 118

xv Figure 4-6 Max. peak pin power vs. EOC soluble boron concentration during GA run 121 Figure 4-7 Max. peak pin power vs. EOC soluble boron concentration during GA run 123 Figure 4-8 Max. peak pin power vs. EOC soluble boron concentration during GA run 125 Figure 4-9 Max. peak pin power vs. EOC soluble boron concentration during GA run 127 Figure 4-10 Kinf Comparison of Gd-155 and Gd-157 enriched U/Gd fuels ................... 129 Figure 4-11 Residual binding (delta k) profiles of Gd-155 and Gd-157 enriched U/Gd fuels......................................................................................................................... 130 Figure 4-12 Assembly peak pin power profiles of Gd-155 and Gd-157 enriched U/Gd fuels......................................................................................................................... 130 Figure 4-13 Kinf Comparison of different cases.............................................................. 132 Figure 4-14 Residual binding (delta k) comparison of different cases........................... 132 Figure 4-15 Peak pin power vs. exposure for the different designs................................ 133 Figure 4-16 Kinf profiles comparison for 100 w/o Gd-155 and Gd-157 enriched UO2/Gd2O3 fuel and natural design with 7 w/o concentration................................ 134 Figure 4-17 Residual delta k profiles for 100 w/o Gd-155 and Gd-157 enriched UO2/Gd2O3 fuel and natural design with 7 w/o concentration................................ 134 Figure 4-18 Kinf profiles of natural and enriched Gd U/Gd FA designs for different U-235 enrichment............................................................................................................... 135 Figure 4-19 Residual delta k profiles of natural and enriched Gd U/Gd FA designs for different U-235 enrichment (w/o)........................................................................... 136 Table 4-9 Optimized designs for position H-09 in the core loading pattern .................. 137 Figure 4-20 Burnable poison loading pattern developed with deterministic and heuristic methods ................................................................................................................... 138

xvi Figure 4-21 Pin power profiles of position H-09 for different optimized fuel designs .. 139 Figure 4-22 RPD power profiles of position H-09 for different optimized fuel designs 139 Figure 4-23 Pin power profiles of position H-09 for different Gd fuel designs ............. 141 Figure 4-24 RPD power profiles of position H-09 for different Gd fuel designs........... 141 Figure 4-25 Peak pin power distributions of final BP loading with 4 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs ................................. 142 Figure 4-26 Normalized power distributions of optimal BP loading with 4 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs ........................... 143 Figure 4-27 Superior BP loading pattern developed by deterministic method............... 144 Figure 4-28 Pin power profiles for superior BP loading with 16 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs ............................................... 145 Figure 4-29 RPD distributions for superior BP loading with 16 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA design ................................................. 145 Figure 5-1 Escalation and discounting method.............................................................. 152 Figure 5-2 Sensitivity of eq. fuel cycle cost for different scenarios for TMI-1 reference cycle with natural U/Gd FA designs....................................................................... 156 Figure 5-3 Sensitivity of eq. fuel cycle cost for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs ............................................................... 156 Figure 5-4 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs.................................................................................... 157 Figure 5-5 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs.................................................................................... 157

xvii Figure 5-6 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs.................................................................................... 158 Figure 5-7 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs.................................................................................... 158 Figure 5-8 Sensitivity of fuel cycle cost versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs .............................................. 159 Figure 5-9 Sensitivity of levelized fuel cycle cost versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs .................................. 159 Figure 5-10 Sensitivity of gross uranium savings and net savings versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs .. 160 Figure 5-11 Sensitivity of fuel cycle cost versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle........................................................ 161 Figure 5-12 Sensitivity of gross uranium savings and net savings versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle.......... 161 Figure 5-13 Sensitivity of levelized fuel cycle costs versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle..................................... 162 Figure 6-1 Reference initial octant core loading with fresh and used fuels.................... 166 Figure 6-2 Selection of decision variables of number of UO2/Gd2O3 fuel pins and Gd2O3 concentration for each fresh fuel position in the reference initial core loading...... 169 Figure 6-3 Genetic algorithm flow diagram and interaction with the reactor physics code ................................................................................................................................. 172 Figure 6-4 Conversion to integer number and encoding process for decision variable of number of UO2/Gd2O3 fuel pins (* mixed concentrations) ..................................... 177

xviii Figure 6-5 Conversion to integer number and encoding process for decision variable of Gd2O3 concentration ............................................................................................... 178 Figure 6-6 Sample encoding and decoding process for global BP optimization problem ................................................................................................................................. 179 Figure 6-7 Sample BP map encoding process and genotype structure for number of UO2/Gd2O3 fuel pins of fresh fuel positions (* mixed concentrations)................... 180 Figure 6-8 Sample BP map encoding process and genotype structure for Gd2O3 concentrations of fresh fuel positions ..................................................................... 181 Figure 6-9 Comparison of fitness value for high and low mutation probability ............ 183 Figure 6-10 Comparison of average and best fitness value for low and high mutation probability ............................................................................................................... 184 Figure 6-11 Ratio of average fitness to best fitness value for low and high mutation probability ............................................................................................................... 184 Figure 6-12 Effect of mutation probability change during GA evaluation..................... 185 Figure 6-13 Fitness value during GA evaluation............................................................ 188 Figure 6-14 Maximum peak pin power profiles during evaluation ................................ 188 Figure 6-15 Maximum peak pin power versus Gd amount during evaluation ............... 189 Figure 6-16 Best fitness value change with generation number ..................................... 189 Figure 6-17 Solution space for core BP optimization problem ...................................... 192 Figure 6-18 Maximum peak pin power versus Gd amount in the solution space........... 193 Figure 6-19 Fitness value profile during evaluation ....................................................... 193 Figure 6-20 Solution space demonstration and its objective functions .......................... 195 Figure 6-21 Maximum peak pin power versus Gd amount during evaluation ............... 198

xix Figure 6-22 Fitness value change during evaluation ...................................................... 199 Figure 6-23 BOC soluble boron versus total Gd amount (SUM) during evaluation ...... 199 Figure 6-24 BOC soluble boron change during evaluation ............................................ 200 Figure 6-25 Max. peak pin power vs. BOC soluble boron concentration .................... 200 Figure 6-26 Good solutions within maximum peak pin power and BOC soluble boron constraints ............................................................................................................... 201 Figure 6-27 Best fitness value vs. generation number.................................................... 201 Figure 6-28 Total Gd amount (SUM) vs EOC soluble boron concentration for first 10 good designs............................................................................................................ 206 Figure 6-29 Maximum peak pin power vs BOC soluble boron concentration for first 10 good designs............................................................................................................ 206 Figure 6-30 Gd amount vs EOC soluble boron concentration for good designs ............ 207 Figure 6-31 Maximum peak pin power versus Gd amount during evaluation ............... 209 Figure 6-33 BOC soluble boron versus total Gd amount (SUM) during evaluation ...... 210 Figure 6-34 BOC soluble boron change during evaluation ............................................ 210 Figure 6-35 Max. peak pin power vs. BOC soluble boron concentration during evaluation ................................................................................................................................. 211 Figure 6-36 Good solutions within max. peak pin power and SOB constraints............. 211 Figure 6-37 Best fitness value change during evaluation ............................................... 212 Figure 6-38 Total Gd amount (SUM) vs. EOC soluble boron concentration................ 217 Figure 6-39 Maximum peak pin power vs BOC soluble boron concentration ............... 217 Figure 6-40 Gd amount vs. EOC soluble boron concentration for good designs ........... 218 Figure 7-1 Max pin power vs. BOC soluble boron concentration during evaluation..... 229

xx Figure 7-2 Hybrid GA-NN methodology ....................................................................... 231

LIST OF TABLES
Table 1-1 General Characteristics of burnable absorber given for a reference 18 M cycle PWR reactor.............................................................................................................. 13 Table 2-1 TMI-1 reference core data ................................................................................ 31 Table 2-2 Reference TMI-1 core with available FA designs (HFP depletion)................. 35 Table 2-3 Comparison of data for superior and final BP loading design ......................... 50 Table 3-1 Important GA input data for GA code.............................................................. 74 Table 3-2 Optimum UO2/Gd2O3 pin positions in octant assembly model for different assembly uranium enrichment and number of UO2/Gd2O3 pins (4 w/o Gd2O3)....... 75 Table 3-3 U-235=4.5 w/o, Gd2O3=4 w/o CASMO-4 GA optimization results with minimizing BOC Kinf objective ................................................................................ 80 Table 3-4 Reference core with available FA designs and new optimal designs with minimum BOC Kinf ................................................................................................... 80 Table 3-5 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with mixed concentration of Gd2O3) ......................................... 91 Table 3-6 Probability of having U/Gd fuel pin for each position in good designs........... 97 Table 3-7 Comparison of available FA designs and optimal designs............................... 98 Table 3-8 U/Gd fuel pin configurations and core depletion parameters for good solutions (20 U/Gd fuel pins with mixed concentration of Gd2O3) ....................................... 101 Table 3-9 Probability of having U/Gd fuel pin for each position in good designs......... 102 Table 3-10 U/Gd BP designs (No of U/Gd fuel pins, Gd2O3 w/o, Segment Name)....... 104 Table 4-1 Gd isotopes weight fractions for different enrichment of Gd-157 isotope..... 118 Table 4-2 Gd isotopes weight fractions for different enrichment of Gd-155 isotope..... 119

xxii Table 4-3 U/Gd fuel pin configurations and core depletion parameters for good solutions (20 U/Gd fuel pins with 100 w/o enriched Gd-155 and 2 w/o concentration of Gd2O3)..................................................................................................................... 121 Table 4-4 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 100 w/o enriched Gd-155 and 2 w/o concentration of Gd2O3)..................................................................................................................... 123 Table 4-5 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 70 w/o enriched Gd-157 and 2 w/o concentration of Gd2O3) ................................................................................................................................. 124 Table 4-6 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 100 w/o enriched Gd-157 and 2 w/o concentration of Gd2O3)..................................................................................................................... 126 Table 4-7 Comparison of EOC soluble boron data and residual binding savings in ppm ................................................................................................................................. 128 Table 4-8 Gd2O3 concentrations and enrichment ratio of Gd-155 and Gd-157 isotopes for different cases ......................................................................................................... 131 Table 4-9 Optimized designs for position H-09 in the core loading pattern .................. 137 Table 5-1 Parameter Notation for Fuel Cycle Cost Calculations ................................... 149 Table 5-2 Fuel Cycle Data [4, 5] .................................................................................... 153 Table 5-3 Unit Prices for Each Component [4, 5] .......................................................... 153 Table 5-4 Economic Parameters [4, 5] ........................................................................... 153 Table 5-5 Batch regions and feed fuel assembly numbers of the TMI-1 reference core loading..................................................................................................................... 154

xxiii Table 6-1 Initial core loading and input data for the reference TMI-1 core ................... 167 Table 6-2 Sample BP map encoding process.................................................................. 179 Table 6-3 Genetic algorithm input data and operators type............................................ 185 Table 6-4 Best solution and core global BP map for the reference TMI-1 core with available vendor reference fuel assembly design.................................................... 203 Table 6-5 Best individual in GA solutions ..................................................................... 204 Table 6-6 Good designs data from archived solutions.................................................... 204 Table 6-7 Good designs and no of U/Gd pins in the fresh fuel positions....................... 205 Table 6-8 Good designs and Gd2O3 concentrations (w/o) in fresh fuel positions from the archived solution..................................................................................................... 205 Table 6-9 Best solution and core BP map for TMI-1 Reference Core with optimal BP fuel assembly designs..................................................................................................... 214 Table 6-10 Best individual in GA solutions ................................................................... 214 Table 6-11 History data during evaluations.................................................................... 215 Table 6-12 Good designs data from the archived solutions............................................ 215 Table 6-13 Good designs and number of U/Gd pins in the fresh fuel positions from the archived solutions ................................................................................................... 216 Table 6-14 Good designs and Gd2O3 concentrations (w/o) in fresh fuel positions from the archived solution..................................................................................................... 216 Table 6-15 Comparison of core BP optimization results for different BP designs ........ 219

ACKNOWLEDGMENTS
The author would like to thank his advisor, Prof. Kostadin Ivanov for his guidance, constructive suggestions, support, and unlimited patience throughout this study. He always believed in me, encouraged me and stated his support during hopeless time of the research. Special thanks to Prof. Samuel Levine for his close interest, feedback, knowledge, and corrections on my English grammar without limited tolerance. I would also like to thank Prof. Arthur Motta, Yousry Azmy, and Prof. Soundar Kumara for their valuable comments as committee members.

I wish to specially thank Moussa Mahgerefteh from Exelon Nuclear Company for their valuable support, time, comment, and feedback during this study. This research would not be successful without his real industry experience and expertise. The author would also like to thank for his unlimited tolerance and permission to work with company server whenever it was necessary. Special thanks also to Robert Wolfgang from Exelon Nuclear Company for his constructive cooperation during headquarter visits.

Special thanks are due to all of my family members for their encouragement, support and the patience throughout this work. I wish to express my deep appreciation and affection to all of my friends who shared my feelings and worries to overcome encountering difficulties during the long period of this study.

My special thanks to my beloved wife, Mine Ozdemir Yilmaz, for being with me and sharing all of the happiness and difficulties during this period. I would like to express

xxv my feelings with the dedication of following lyrics. She, May be the reason I survive, The why and wherefore I'm alive, The one I'll care for through the rough in ready years, Me, I'll take her laughter and her tears, And make them all my souvenirs, For where she goes I've got to be, The meaning of my life is, she, She, oh she.

CHAPTER 1 INTRODUCTION
1.1 Problem Overview
Advanced measures to minimize fuel cycle costs are receiving increasing attention within the deregulated power generation market [1, 2, 3, 4, 5, 6, and 7]. The economic benefits of minimizing the amount of enriched Uranium (U) and the Burnable Poison (BP) are significant. The role of the in-core fuel management in supporting this goal consists in the development of advanced low-leakage longer cycle loading patterns. A prerequisite of this type of loading pattern is the use of optimized BP designs and their placement into the core [8]. Modern Pressurized Water Reactor (PWR) cores employ BPs to reduce the soluble boron (SOB) concentration at the beginning-of-cycle (BOC) and to maintain the peak pin power below its constraint during the core depletion. The maximum peak pin power is a constraint (1.55 for TMI-1 core) during the depletion calculation and the BOC soluble boron must be below the equivalent concentration (1700 ppm for TMI-1 core) to prevent a positive moderator temperature coefficient (MTC).

Advancements in BPs designs have provided the nuclear fuel manager with a flexible means to improve the performance of the reactor core in terms of reducing the cores maximum peak power so that low leakage cores are useable. Designing a BP pattern in the most effective manner is difficult because of the complex nonlinear

2 changes in power distribution that occur with change in BP pattern. Inserting BPs in the fuel assembly lowers its Kinf and thus, lowers its power but increases the power in other fuel assemblies to conserve the total core power. This process is highly non-linear making it difficult to achieve optimum results.

In practice, fuel loading pattern design starts with used and fresh fuels in the inventory and some heuristic rules are applied to place them in the core. Then, the BPs are inserted into those fresh fuel assemblies in which power limit constraint is exceeded [9]. The available burnable absorber designs are inserted into fuel inventory and then fuel loading optimization calculations are carried out by using a limited number of burnable absorber types in the fuel inventory. The result of such a calculation does not reflect the real optimal solution, and it just presents the solution in a practical way. This study developed a unique technique and practical tool for designing BP placement of a given core loading pattern.

Gadolinium (Gd) is used as integral BP in PWRs. However, the use of Gd as BP has some inherent penalties: reduced average fuel enrichment in the UO2/Gd2O3 fuel pins and increased residual reactivity binding. The research presented in this thesis focused on reducing these drawbacks by first optimizing the UO2/Gd2O3 fuel pin configurations for all feasible U/Gd fuel assembly designs, and then it developed a BP loading pattern by using these feasible designs in the core to minimize the total Gd amount and its residual binding during reactor operation.

3 The loading procedure for optimizing a two-year cycle necessitates the use of BPs in the core design. The End-of-Cycle (EOC) low leakage core design mandates an optimum BP design, which must prevent the maximum peak pin power from exceeding its calculated constraint value during the cycle depletion. Thus, the use of an automatic BP design algorithm becomes much more important.

1.2 Background
A detailed literature search was performed to present what has been done before related with this research. All of the reference studies summarized in the following sections contributed to the direction of the research presented in this thesis. 1.2.1 Burnable Absorbers and Gadolinium Utilization Kapil, S., K., Secker, J., R, Keller, H.,W., [10] emphasized that extended cycle lengths, increased fuel burnups and reduced neutron leakage fuel management strategies contributes to use of BAs. The study presented the total potential fuel cost improvements due to use of different BA designs by reducing residual reactivity penalty further for 24 month cycle PWR core.

Skogen, F.,B., Kempf, F., J., [11], studied on Gadolinia Integral Burnable Absorber. The study emphasized that gadolinia presents broad flexibility to fine-tune core power distributions as function of cycle burnup yielding, and maximum low leakage fuel management strategy can benefit from it without exceeding licensed limits. As a result, the study emphasized that Gd bearing core presented more margin to licensed thermal margin limits.

4 Yokote, M., et al [12] emphasized that the number and optimum location of gadolina rods in an assembly can be determined by considering reactivity control capability and peaking factors. Only 12 BP rods were selected for 1414 FA, and 16 BP rods for 1515 FA for the calculations. The study determined the optimal locations of gadolinia rods by considering bundle parameters such as assembly power peaking factors and negative residual binding that provided us starting points for determining the optimal U/Gd fuel rod positions this thesis.

Wagner, J., and Sanders, C., [13] examined the effect of fixed absorbers including integral burnable absorbers (IBAs) and burnable poison rods (BPRs) on the reactivity of PWR spent nuclear fuel in support of burnup-credit criticality safety analyses. The study presented detailed discussions for understanding of Kinf and residual delta binding profiles of IFBA designs including UO2 pellet with a thin ZrB2 coating on the outer surface, UO2-Gd2O3 rods, UO2-Er2O3 rods and Al2O3-B4C rods. Analysis showed that UO2-Gd2O3 has negative residual delta binding value and it is always less reactive than the fuel without gadolinia. This study contributed to the understanding the effect of both use different concentration and number of gadolinia rods on assembly Kinf and residual reactivity behavior during depletion. 1.2.2 Deterministic and Stochastic Optimization Techniques for Loading Pattern (LP) and BP design problems Deterministic and Stochastic Methods are widely used optimization techniques in nuclear fuel management. Li, Z., [14] developed an automatic PWR reload design expert system computer code. This study used two important deterministic techniques to

5 develop an optimum PWR reload pattern. The first one was to develop a priority scheme, which represents the optimum placement of the fuel in the core for maximization of the cycle length or using minimum fresh fuel for a given cycle length. The second technique called Power Shape Driven Progressive Iteration (PSDPI) was to determine the burnable poison loading in the fresh fuel assemblies. This technique was significantly modified for use in developing the deterministic optimization method presented subsequently in this thesis.

Haibach, B., [15] performed a deterministic fuel management study to optimize IFBA designs for PWRs. He studied on certain number of IFBA pins within a fuel assembly (i.e. 64, 80 and 160) by using CASMO-3 assembly depletion code. The placements of the IFBA pins within an assembly were chosen to maximize the effectiveness of the IFBA pins. Once a loading was found to produce power distributions lower than the vendors design, the iterations continued until a loading was found that compares to the vendors assembly lifetime. Their design reduced the maximum peak pin power by as much as 5 % over the vendors IFBA configuration. This research determined the optimal IFBA configurations by considering assembly lifetime and lower BOC Kinf value of the design. In this thesis, BOC Kinf value of the U/Gd fuel assembly design was discovered and used as a filter in the coupled bundle and core depletion calculations.

De Chaine [16], performed stochastic analysis by using GA optimization tool, CIGARO by using SIMULATE-3 reactor physics code. The study developed optimal

6 loading pattern by using different number of IFBA fuel assemblies in the core. The study emphasized that Genetic Algorithm core loading problem is highly dependent on the fuel inventory selection. The study did some BP placement by inserting fuel types, which all have the same number and enrichment of IFBA designs for all fresh fuels. The GA code structure was very complex and developed for experienced fuel engineer use. This study contributed to the development of designing an algorithm that developed a BP loading pattern with variable number of U/Gd fuel pins and Gd2O3 concentrations for each fresh fuel positions in the reference core.

De Chaine [16] also emphasized that additional fuel types with BPs could be added into fuel inventory, and the optional bits could also be included into genotype structure. The author also suggested that a third option could be separating the fuel placement problem from the BP placement problem. The fuel placement could be solved using a Haling power distribution to estimate the peak normalized power (NP). Then, leaving the fuel placement fixed, the BP loading could be solved again using Genetic Algorithms. This could be done by using the different fuel types to represent different amounts of BPs in the fresh fuel assemblies. Separating the entire problem into two sequential parts may prove to be the most efficient option, and was utilized in this research.

Hida, K., [17] studied on the optimization of axial enrichment and burnable absorber distributions for Boiling Water Reactor Fuel Assembly with successive linear programming method. He defined rod power peaking factor as a decision filter whether

7 continue iteration with core depletion calculations or not. This study contributed to the development of coupled bundle and core depletion calculations by using BOC Kinf filter in GA calculations.

Guler, C. [18], and Hongchun, W. [19], also performed core optimization studies by using genetic algorithms. They all determined the optimum core configuration for a given cycle. Standard bit-based genetic operators were used to optimize the arrangement of assemblies, burnable absorber, and used assembly orientations. Burnable absorber types and numbers were included into available fuel inventory of a reference core. The genetic algorithms code determined the optimal core loading pattern by using available fuel inventory without solving for the burnable poison placement.

Keller, P., [20] reintroduced genetic algorithm methodology into a modern version of the FORMOSA-P code [21], which was developed to determine the family of near-optimum LPs for PWRs by utilizing Simulated Annealing (SA) optimization methodology [221]. The reintroduction was motivated by the inconsistency of the existing SA algorithms in determining near-optimal fresh feed fuel patterns. The GA has the capability to perform multi-objective optimization across multiple feed fuel patterns. This research showed the advantage of use of GA methodology in the LP optimization calculations. The advanced GA reduced the average minimum fresh feed fuel enrichment by 0.019 w/o for a test problem, and reduced the standard deviation of the determined minimum feed fuel enrichment for multiple optimizations from 8.510-2 to 2.210-2.

8 Maldonado, G., T., et al [23] coupled the simulated annealing algorithm to the collision probability-based lattice physics code CPM-2 to produce a within-bundle loading optimization code. The primary optimization control variables include the pinby-pin placement and/or the loading of nuclear fuel, burnable absorbers, or other materials and their characteristics within a fuel assembly. The objective functions were selected as minimization of 235U and minimization of the relative pin power peaking. The study uses a penalty-based option to handle constraints. The results show decrease in assembly average enrichment from an initial value of 4.920 % down to 4.873 %. This study contributed to the development of solution methodology, problem definition and constraint handling technique that will be presented in this thesis for determining UO2/Gd2O3 fuel pin positions for a reference PWR fuel assembly.

Another study done by Maldonado, G., T., et al [24], presents the development of an optimization tool which has been coupled to the lattice-physics code CPM-2. The study also uses Simulated Annealing (SA) algorithm to optimize the pin-by-pin placement and loading of nuclear fuel and burnable absorbers. The tool has user input upper and lower bounds on Kinf. The tool reduces the BOC pin power peak from about 1.674 to 1.093 and average enrichment of the assembly from 4.920 to 4.855.

Lam, H., Q., et al [25] developed a new loading pattern search tool (LP-Fun) that guides the user in exploration of the LP domain defined by the chosen fuel inventory and the available burnable absorbers (IFBA ZrB2 rods). The objective was to minimize pin peaking factor without decreasing cycle length. The tool uses the BOC screening step that

9 rejects LP pattern with very high FdH on the basis that batch refinement and burnable absorbers will never be able to lower the peaking factors to acceptable levels. These patterns were subjected for peak power minimization, and then the best pattern was used to assign burnable absorber for fresh fuel batches with the objective of end-of-cycle boron and the constraint of FdH. This research contributed to the idea of using filter value on fuel assembly parameters before performing core depletion calculations for accelerating the solution process.

Also, Parks G., T. [26] has investigated on the multi-objective GA objective function PWR reload core design problem by using a non-dominated GA search. He worked on the trade-offs between various objectives such as minimizing feed enrichment, maximizing discharge burnup, and minimizing radial power peaking. In this thesis, special emphasize was given to minimizing the residual binding to increase core lifetime as well as minimizing the total Gd amount in the core. 1.2.3 Optimized Enriched Gadolinia Concept Schlieck, M., Berger, H.,D., and Neufert, A., [27] focused on the reduction of inherent penalties of Gd-FA designs. The study showed that the most effective way to overcome these drawbacks is the reduction of the Gd2O3 concentration to values ~ 2 w/o according to recent measurements of the heat conductivity of modern Gd fuels [28], the reduction of the fissile content in the Gd rods is no longer necessary. Therefore, in this thesis, enriched Gd designs was only considered for fuels with low (~2-3) Gd2O3 concentrations.

10 Spetz, S., and Hove, C., [29] performed a reference study, which discussed the effect of enriched gadolinium on the residual absorption and on the absorber depletion rate. The study also estimated the reactivity benefits and front-end fuel cost savings. Because of its high absorption cross section, the study focused on Gd-157 isotope and its enrichment ratio in all other Gd isotopes. The EOC residual binding was reduced by a factor of three to six by enriching this isotope above its natural abundance of 15.7 % to the range of 80 % to 100 %. The study presented the results of the work, which was performed by plasma separation at the TRW Company in Los Angeles [30], for enriched Gd product. The study presented the enrichment cost less than $ 10 /gram-Gd-157. As a result, the savings were about $ 450,000 for core basis at $ 20 /lb for yellowcake and $ 110 per uranium separative work unit. The Gd-157 enrichment cost was subtracted from the gross front-end savings. The net fuel cycle cost savings was determined to be $ 250,000. This study focused on the economical benefits of using enriched Gd-157 for a given reference core. The important contribution was to provide a direction for fuel cycle cost methodology and calculation of fuel cost savings due to use of enriched Gd-155 fuel, which was developed in this study for a given reference TMI-1 core.

Santala, M., et al [31] demonstrated experimentally that significant enrichment can be achieved by laser ionization. The results showed that Gd-157 fraction can be raised from natural 30 % to nearly 70 %. In this thesis, 70 w/o Gd-157 and 100 w/o Gd155 enrichment ratios for optimized designs were considered to be used in the reference TMI-1 core.

11 Renier. J.P.A, and Grossbeck, M.L., from Oak Ridge National Laboratory

published a Phase 1 report on the development of improved burnable poisons for commercial nuclear power reactors in October 2001 [32]. The Gd isotopes were assumed to be separated completely in this study. The study focused on the neutronic performances of enriched 100 % Gd-157 designs. It was estimated that in a single pass in the plasma separation process (PSP) process to obtain Gd-157 would results in 46 % to 73 % enrichment depending upon the magnetic field strength used. The use of single-pass burnable poison enrichments was left to second phase of the project.

Grossbeck, M.L., Renier, J.P.A, and Bigelow, T., from The University of Tennessee published a final report on the development of improved burnable poisons for commercial nuclear power reactors in September 2003 [33]. The study discussed plasma separation process for candidate enriched burnable absorbers in detail, and it concluded that the cost of producing separated Gd-157 isotope is on the order of $ 600 to $1 600 per gram using the present PSP facility (1.8 Tesla Magnet) used for this study. A device with a 6 Tesla magnet was predicted to reduce the price on the order of $ 60 to $ 160 per gram. The study stressed that the improved techniques of operation could make the process more cost effective in the future.

Bejmer, K. and Severborn, O. [34] performed a study on enriched gadolinium with 70 w/o Gd-157. The core depletion results were compared for fuel with 5 w/o gadolinium with natural isotopic composition and fuel with 2 w/o gadolinium with 70 w/o Gd-157. The results showed that the gains in average for the five cycles studied were

12 about 70 EFPH per cycle. This study showed the potential EFPH savings resulting from the use of enriched Gd designs in the reference core. 1.2.4 Burnable Poison Fabrication Cost All of the literature studies presented in this section contributed to the development of fuel cycle cost methodology and helped to estimate the potential savings amount in dollar due to use of less gadolinium in the core after performing optimization calculations.

Lee, J., et al [35] performed fuel cycle cost sensitivity study on burnable absorber and feed assemblies. The study focused on both the variation in BA market price and the economic evaluation of BA by maximizing cycle length. The BA fabrication costs used in this study are as follows: IFBA, WABA, gadolinia, and erbium are $ 360, $ 1250, $ 1 100, and $ 110 per pin, respectively.

Stephenson, L., et al [36] performed a comparative study between three types of integral burnable absorbers: gadolinia, zirconium diboride and erbia. They developed advanced gadolinia fuel design to minimize the use of higher concentration, and to maximize the U-235 enrichment in gadolinia-bearing fuel. It resulted in using low w/o gadolinia concentrations to control soluble boron concentration. They performed equilibrium fuel cycle comparison and the conclusion was that the enrichment requirements for the gadolinia and the zirconium diboride equilibrium fuel cycles are approximately the same, while those for the erbia design are significantly higher. The 0.02 w/o U-235 batch average enrichment difference between the gadolinia and the

13 zirconium diboride fuel cycles is equivalent to approximately 7 $/kgU in fuel cycle cost. While the study presented some economical analysis for three types, it did not mention about the detailed fabrication cost of Gd.

Another study performed by Asou, M., and Porta, J., [37], compared rare earth nuclides gadolinium, erbium, europium, and dysprosium for the control of potential core reactivity in PWRs to increase cycle length. The author presented the detailed comparison of different types including unit prices, thermal cross section, and residual reactivity at EOC. The study takes Gd2O3 prices as 2200 FF/kg (~$ 358 per kg). Table 11 summarizes the general characteristics of burnable absorbers given for a reference 18 month cycle PWR reactor. Table 1-1 General Characteristics of burnable absorber given for a reference 18 M cycle PWR reactor Property Conductivity Gd2O3 Depend on BA w/o Sm2O3 Small change Eu2O3 temperature Dy2O3 Temperature 5500 930 930 164, 165 6840 Er2O3 Small change 2000 162 162 167 5196

Price 2200 6500 10 400 (FF/kg) Thermal 49000 5800 4600 x-sec(barn) Resonance 390 5800 4600 x-sec (barn) Absorber 155, 157* 149, 151 151, to 155 isotopes Residual (pcm) at 420 2604 3636 EOC *Equal initial worth, heterogonous poison mode

Cacciapouti, R., Weader, R., and Malone J., [38] evaluated the relative neutronic efficiency and fuel cycle cost benefits of PWR burnable absorbers for a 3411 MWt

14 typical Westinghouse, 193 assemblies PWR core. The studies performed in this paper are that designing low leakage equilibrium in-core fuel management plans for 12, 18 and 24 month cycles, reviewing of the fuel management impact of IFBA, erbium and gadolinium, calculation of U3O8, UF6, SWU, fuel fabrication and BA absorber requirements, and estimation of fuel cycle costs and comparative savings of different BA in dollar equivalent per kg-U of fabricated fuel. The study presented the methodology of the fuel cycle cost evaluation, but unit cost of each process at the front-end part of the cycle was unknown. The fuel cost methodology was consistent with the Penn State Equilibrium Fuel Cost Methodology [39].

1.3 Statement of Objective and Thesis Organization


The objective of this study was to develop a unique technique and practical tool for designing BP placement for a given core loading pattern. The first task was to develop a deterministic technique, which calculates equivalent thermal absorption cross section necessary to restrict assembly RPD value under the calculated power limit value. The second task was to develop Genetic Algorithms Model for determining UO2/Gd2O3 fuel pin configurations for all feasible BP designs. The decision variables of the problem were UO2/Gd2O3 fuel pin positions within an octant fuel assembly symmetry. The objective was to design good UO2/Gd2O3 fuel assembly types, which do not violate maximum peak pin power constraint and reduce its residual binding when they are placed in the hot fresh fuel assembly positions in the core loading pattern. The third task was to develop optimized low concentration and enriched gadolinia designs, which would replace with high concentrated natural Gd enriched fuel assembly in the core.

15 Fourth task of the study was to develop a separate genetic algorithm methodology to determine the optimal BP pattern, which minimizes Gd amount in the core together with residual binding. The constraints were the maximum peak pin power during depletion and SOB concentration at BOC. The number of UO2/Gd2O3 pins and Gd2O3 concentrations for each fresh fuel location were the decision variables and the total amount of Gd in the core and maximum peak pin power were in the fitness functions.

The organization of thesis is given here. Chapter 2 describes the Modified Power Shape Forced Diffusion (MPSFD) methodology and deterministic optimization program algorithms. The Genetic algorithm techniques and its application for determining the good UO2/Gd2O3 fuel pin configurations are presented in Chapter 3. Chapter 4 introduces the optimized enriched Gd concept and applies the same GA technique to develop enriched Gd-155 and Gd-157 designs. Core depletion results using equivalent optimized design instead of natural FA design are also presented in Chapter 4. Economical analysis of optimized designs concept and sensitivity studies are summarized in Chapter 5. The developed GA methodology and the solution techniques for Core BP optimization problem is given in detail in Chapter 6. Finally, Chapter 7 presents a summary of this study with conclusions, summary of contributions and future work for further study.

CHAPTER 2 DETERMINISTIC METHODOLOGY


2.1 Introduction
Li Z. [14] developed a deterministic technique called the Power Shape Driven Progressive Iteration (PSDPI) method to help design the BP loading pattern in the core. The PSDPI method assigns the burnable poison into the fresh fuel assemblies by first searching for the equivalent total thermal absorption cross section that attempts to match the core power distribution with the Haling power distribution. The technique developed an empirical formula for determining the total thermal absorption cross section of the fuel assembly by using some crude assumptions (e.g. one group theory). The burnable poison equivalent thermal absorption cross sections are first isolated from the total thermal absorption cross section and then converted to the number of burnable poison rods for the given burnable poison type using a burnable poison database. The methodology is based on the one group expression for the Kinf of a fuel assembly as expressed in the formula below:
k = f =

a + a + a
F M

F P

(Equation 2.1)

Where is the number of fission neutrons produced per thermal neutron absorption by the fuel, f is fuel utilization factor, and

,
F a

M a

and a are the thermal


P

absorption cross sections of the fuel, the moderator, and the fission products. The

17 methodology used constant power distribution in the core during the depletion. Some burnable poison is removed during the depletion to keep Kinf constant. The Kinf of the poisoned assembly is written as:

BP

a + a + a
F M

F BP

(Equation 2.2)

The power distribution in the core is a function of the Kinf of the fuel assemblies. The methodology tries to match the actual power distribution with the target power distribution (haling power distribution) by making the difference of Kinf values in Equation (2.3) equal to zero.

BP

F ,H a

+ a + a
M

F ,H BP , H

a + a + a
F M

F BP

=0

(Equation 2.3)

Mathematical derivations ended up with the empirical formula below.

a =
BP

NP NP H NP H

* a

(Equation 2.4)

BP

= the change in BP thermal absorption cross section during each iteration, which

attempts to change the power in the FA, NP, to where it equals the NPH.

18 NP = the calculated normalized power of an assembly, which is equal to the radial relative power density (RPD) of this assembly NPH= target Haling power limit

* a

= the assembly averaged thermal absorption x-section without burnable

poison; is a convergence acceleration parameter to be determined. This methodology did not work at first attempt.

In a PWR reactor, burnable poisons are always loaded in the fresh fuel assemblies, since the new fuel assemblies have the highest Kinf, they normally produce peak power beyond the imposed limit. In this research, burnable poisons are presently placed only in the fresh fuel assemblies. In real life, core power distribution never matches the haling power distribution. Therefore, the Haling power distribution was used as guidance for the beginning of iteration.

Li studied the problem of non convergence and, based on his study, added a second parameter to improve the convergence characteristics during the iterations. The new equation, Equation (2.5), is

a =
BP

NP NP H * a H NP

(Equation 2.5)

19 Li determined that the Equation (2.5) could be made to converge if the appropriate value of is determined. And in fact, it turns out that this new empirical formula will converge without the need for parameter given in Equation (2.6).

a =
BP

NP NP H * a H NP

(Equation 2.6)

A close examination of the Equation (2.6) reveals that this equation may be used to force the NP to equal NPs others than the NPH. By changing the a during iteration, the NP can be made to converge to any reasonable value of NPX. This was shown to be the case. In this research the NPH was replaced with maximum allowed value of NP, NPlimit. This new method was directed to finding a true solution for the BP designs by only requiring that NPs are below their constraint. It then became a practical method to place BPs in the core in an excellent manner to satisfy the maximum peak pin power or normalized power constraints when the calculations are performed with the reactor physics code such as SIMULATE-3. This code must be regimented through a shell because the source code is not available. The new method places Gd only in the fresh fuel. However, it is possible to change the code in the future to add the boron BPs, if it is necessary.

The new method, which is called Modified Power Shape Forced Diffusion (MPSFD), incorporates the RPDlimit (NPlimit) in place of the RPDHPD as expressed in Equation 2.6. The Equation 2.7 is employed only when an assembly radial power exceeds the RPDlimit and the solution is only a first step. The MPSFD method calculates the

20 burnable poison equivalent thermal absorption cross section for all of the fresh fuel assemblies, which have RPDs greater than the limit, with the following equation:

a =
BP , i

NP NP lim it NP lim it

base , i

(Equation 2.7)

BP , i

= the adjustment of the burnable poison equivalent thermal absorption x-

sec for assembly i at an iteration step, this parameter approaches zero as the iteration converges; NPi = the calculated normalized power of assembly i, NPlimit= normalized power limit,
base ,i

= the assembly averaged thermal absorption x-section without burnable

poison, The total thermal absorption cross section of a fuel assembly is written in the following equation.
total ,i

a2 =

base ,i

+ a

BP ,i

(Equation 2.8)

If calculated total thermal absorption cross section given in Equation 2.8 is greater than the total thermal absorption cross section of the maximum allowed number of U/Gd fuel pins and Gd2O3 concentration design in the data base, the program automatically assigns maximum values of 16 U/Gd fuel pins and 8 w/o Gd2O3 concentration, and it continues to iteration. A sample data from the U/Gd fuel assembly cross section data base is presented in Appendix 1.

21 Although, Equation 2.7 was derived using simplified assumptions, it becomes an excellent empirical formula for establishing a workable initial BP loading for the core.

By using the RPDlimit in Equation 2.7., the solution probably does not converge to the optimum BP design wherein the core design uses the minimum Gd. For this reason a second algorithm is introduced to further reduce the Gd in the core while meeting all safety constraints. This new algorithm is a heuristic method. This heuristic method is followed by a fine-tuning method to find the optimum BP loading.

The core was depleted during these calculations using the codes, CASMO3/SIMULATE-3 to satisfy the maximum peak pin power and BOC soluble boron constraints. The MPSFD was designed to first load the Gd as the BP. The codes were operated through a shell.

2.2 The MPSFD Computer Program


A computer program was developed to design a BP placement for a given loading pattern using fuel assembly designs with vendors reference UO2/Gd2O3 pin arrangements within the designs. The initial step was to use only Gd to affect the optimum power shape while remaining within the Gd guideline restrictions. If the Gd does not succeed in meeting the peak pin power constraints, boron BPs can be added to further reduce the peak pin power.

22 Figure 2-1 shows the automated program algorithm. The program first depletes the given core loading to obtain a solution without any burnable absorber. Then, it determines the burnup step where the maximum normalized powers occur for each fresh fuel location. It next determines where the RPDs exceed their limit and calculates the absorption cross section using MPSFD method to determine the amount of Gd to be added to these fuel assemblies to reduce their power (See Equation (2.7)). Otherwise, it skips those locations where the RPD was below their limit.

The program uses the backward depletion subroutine to determine the BOC Gd2O3 concentrations and the number of U/Gd pins in the assembly, which match with absorption cross sections at BOC. The backward depletion subroutine determines the BOC BP content from the burnable absorber data base prepared for Gd fueled assemblies. This data base contains all of the cross section data for the available fuel assemblies, which Gd2O3 concentration varies from 2 to 8 w/o, and numbers of U/Gd pins are 4, 8 and 16. After completing the backward depletion calculation for each fresh fuel location, which has higher RPD than the limit value, the program generates a new SIMULATE input deck with the new BP loading automatically. Then, SIMULATE-3 depletes the cycle using the cross section library generated previously.

Next, the guideline constraints are checked again after cycle depletion until all constraints are satisfied. If they are not satisfied and the iterations do not improve the power distribution, the program uses a new method called Gd adjustment to try to further

23 reduce the Gd and continue to satisfy the RPD constraints. This method is effective in developing an excellent solution in which the maximum RPD is less than the constraint for all fresh fuel assemblies. It increases the RPD in the low RPD fuel assemblies and decreases the RPD in the other fuel assemblies to bring all RPDs within their constraint. After completing the Gd adjustment routine, the program depletes the cycle producing a superior BP design for the given core configuration.

24

Given Core Loading


iteration=1

Cycle Depletion
i=1, Total No of fresh fuel i=i+1

Select Burnup step where maximum NP occurs


Skip this location i=i+1

If NPi > NPlimit

Yes

Calculate Absorption x-sec to reduce the NP

No

yes

If NPi < NPlimit

Determine Gd2O3 SIMULATE Depletion yes concentration and assembly type Check i=Total No of fresh fuel (yes or no) Gd adjustment algorithm to satisfy
No

No

Gd BP design

NPlimit constraints

iteration=iteration+1

BP loading for next iteration

Figure 2-1 Algorithm for Gd placement

25

2.3 Fine-tuning Heuristic Selection Technique


A fine-tuning heuristic BP selective process was designed to reduce amount of BPs in the core that meets all of the guideline constraints. The superior BP design, which was developed by using the MPSFD methodology, is an initial BP loading pattern. Hence, a heuristic selection method was developed to further minimize the amount of BPs in the core if such is possible. It calculates differences between the RPD limit and the assembly RPD (RPD) for each fresh fuel location in the core with the superior BP design. It then arranges the delta RPD values in sequential order where the highest one is first. It removes 1 w/o Gd2O3 concentration, which is the smallest possible increment value, from the fuel assembly with the highest delta RPD. A standard core depletion calculation takes place with the new slightly different BP design.

Next, the program determines the delta RPDs for this depletion calculation and arranges them in sequential order again. It calculates the maximum delta RPD for the new BP design. It is observed that the effect is a very small change on all RPDs as well as the soluble boron concentration at BOC. The process continues checking the guideline constraints after each BP change. When Gd2O3 concentration reaches 2 w/o in a fresh fuel assembly, the process skips to next fresh fuel position, which has next highest

RPD. RPD is newly selected to continue the process by removing 1 w/o Gd2O3 from
it. The guideline constraints are examined after each core depletion calculation. If either the RPD is within 0.02 of its constraint or the soluble boron concentration is within 5 ppm boron of its constraint, the process is assumed to have converged. Otherwise, the selection process continues following the first selection rule. The calculation converged

26 before the Gd2O3 w/o went below 2 w/o for the reference core. The technique decreases Gd loading by reducing concentration and keeping number of UO2/Gd2O3 fuel rods fixed in the assembly.

Moreover, the Linear Programming by Primal Simplex Method (LPPRIM) [40] was studied to perform the fine-tuning calculation, because the variables should be linear for small changes. However, the errors in the gradient calculations were too large for this method to be applicable because the gradients were very small. When the Gd2O3 concentration changes 1 w/o, the RPD power change is too small and about 0.001 in iteration. The heuristic selection fine-tuning program worked successfully for the reference TMI-1 core. Figure 2-2 shows the flow diagram of the heuristic selection technique.

27 Superior BP Design Calculate RPDn= (RPDlimit RPDn)

Arrange RPDn in sequential order Highest first n=1,totalfreshfuel Remove 1 w/o Gd2O3 From the fuel assembly n and deplete the core Calculate RPDi = (RPDlimit RPDi) i=1, totalfreshfuel Arrange RPDi in sequential order No Check new maximum rpd (is that converge?) =abs (newrpd-NPlimit)<0.02 or Gd w/o = 2 w/o, Check Design Constraints

n=n+1 Skip to next position

Yes

Continue with Same fresh FA position

BP Pattern with minimum Gd amount

Figure 2-2 Heuristic selection technique flow diagram

28

2.4 Final Optimal BP Loading Design

All of the above individual BP design algorithms were coupled into one final BP design program to minimize the BP content in the core. The flow diagram of this program is given in Figure 2-3. The final program starts with the MPSFD and determines a superior BP design with all Gd loaded into the fresh fuel assemblies. The program starts with a given core configuration and first determines if the Gd can be decreased or if it must be increased. If the Gd must be increased, the Gd increase is made until the Gd reaches its guideline limit. The final program calls the heuristic selection program to further reduce the Gd if possible. When the constraints for the maximum pin power and BOC soluble boron concentration are very close to the calculated values the program ceases the calculations and prints out the final design.

BP design with Gd only

Fine Tuning Heuristic Selection Routine

Final Minimal BP Design

Figure 2-3 Flow diagram of the minimal BP design algorithm

29

2.5 Verification of SIMULATE Results for Benchmark Core


The TMI-1 reference core data is summarized in Table 2-1. The TMI-1 reference core loading pattern was modeled both in Exelon and Penn State University (PSU) computers for the PSU SIMULATE-3 code version verification purposes. Figure 2-4 shows reference TMI-1 Octant Core Loading pattern used to verify the models. The PSU SIMULATE-3 code version results should be verified and compared with the Exelon results for the same reference core structure before further use in the optimization calculations. The purpose was to make sure that PSU model is working within enough accuracy, and it is a reliable and valid model to perform optimization calculations.

The main difference between two models is that PSU SIMULATE-3 model did not use the restart file options for used fuel trajectories. Instead, PSU model defined all of the used fuels with their initial exposure data by using FUE.NEW card. However, the use of restart file is very common tool, it requires previous cycle information and depletion calculations. Hot Full Power (HFP) operation of the reference cycle was performed on the Exelon server with restart file option, and the results were presented in Figure 2-5.

Due to lack of opportunity to use the same binary restart file at PSU, the reference Exelon SIMULATE-3 run was modified to be able perform depletion calculation without restart file option and to be able to make a comparison with PSU run. Next, core depletion calculations were repeated both at Exelon and PSU computers for the same SIMULATE models. The detailed results are presented in Figure 2-6 and 2-7. SIMULATE runs both in the Exelon and PSU computer were consistent and acceptable

30 for the reference TMI-1 core loading. There was a slight (0.712 vs. 0.715 g/cc) density difference in PSU and Exelon runs for HFP operation. PSU SIMULATE water density is slightly less than the Exelon SIMULATE water density. Table 2-2 presented a comparison of the results for three different cases. It can be seen from Table 2-2 that the core depletion results such as BOC and EOC Soluble Boron (SOB), maximum RPD, and maximum 3-D peak pin power are consistent with each other. These results validated the PSU SIMULATE model.

31 Table 2-1 TMI-1 reference core data Core Operating Data Design Heat Output MWt Vessel Coolant Inlet Temperature, F Vessel Coolant Outlet Temperature, F Core Outlet Temperature, F: Core Operating Pressure, psig: Core and Fuel Assemblies Total No. of Fuel Assemblies in Core No. of Fuel Rods per Fuel Assembly No. of Control Rod Guide tubes per Assembly No. of In-Core Instru. Positions per Fuel Assembly Fuel Rod outside Diameter, in Cladding Thickness, in Fuel Rod Pitch, in Fuel assembly Pitch Spacing, in Cladding Material Fuel Material Form Cylindrical Pellets Pellet Diameter, in Active Length, in Density, % of theoretical Power Generated in Fuel and Cladding, % Total UO2 (Metric Tons) Control Data Control Rod Material No. of Full-Length CRAs No. of APSRs Worth of 61 Full-Length CRAs (k/k)% Control Rod Cladding Material Ag-In-Cd 61 8 ~8 SS-304 2568 555 604 606 2200 177 208 16 1 0.430 0.025 0.568 8.587 M-5 UO2 Dished-End, 0.3735 143 96 97.3 86.7

32

8
C-03 0 4.00 00 - 00

9
F 1 4.80 16*-03 1 O-13 0 4.90 08 - 02 180

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 16*-03 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 08 - 04 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 16*-03 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 16*-03 2 L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 16*-03 6 F 1 4.90 16*-03 9 F 1 4.90 08 - 02 10 N-13 0 4.85 08 - 04 180

14
M-11 0 4.30 08 04 90 F 1 4.87 08 - 04 5 F 1 4.87 08 - 04 7 L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

Degree Rotation

Figure 2-4 Reference TMI-1 octant core loading

TMI-1 Cycle 16 Core Depletion with Available Fuel Assembly Design on Exelon Server with RESTART files (HFP operation) CASE STEP 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 STEP EXP 0.0 4.0 10.0 25.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0 550.0 600.0 650.0 680.0 693.0 BOR - - - - - PEAK POWERS - - - - DENS POWER NQ PPM AX / K A-O RAD NODE 3PIN G/CC % 7 7 4 4 6 6 6 6 6 6 2195 1815 1786 1735 1690 1602 1513 1421 1314 1189 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1 1 1 1 1 1 1 1 1 1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.35 1.34 1.34 1.34 1.34 1.34 1.35 1.36 1.37 1.37 1.37 1.37 1.38 1.38 1.38 1.37 1.36 1.35 1.35 1.37 1.36 1.36 1.36 1.36 1.36 1.37 1.38 1.38 1.38 1.38 1.38 1.38 1.39 1.38 1.38 1.36 1.35 1.35 1.52 1.51 1.51 1.51 1.51 1.49 1.47 1.49 1.50 1.49 1.48 1.48 1.48 1.49 1.48 1.48 1.46 1.45 1.44 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 FLOW % CRD PRES INLET CORE POS PSIA T (F) EXP 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 19.438 19.557 19.735 20.181 20.924 22.410 23.897 25.383 26.869 28.356 29.842 31.328 32.815 34.301 35.787 37.274 38.760 39.652 40.038 TITLE EXP 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

K-EFF 1.00000 1.00000 0.99999 0.99995 0.99999 0.99999 0.99998 0.99999 0.99999 0.99999 0.99999 0.99999 1.00000 1.00000 0.99999 0.99999 0.99999 1.00879 1.00586

SET 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

6 1055 1.00/ 1 6 916 1.00/ 1 7 778 1.00/ 1 7 641 1.00/ 1 6 495 1.00/ 1 6 5 8 6 350 205 5 5 1.00/ 1.00/ 1.00/ 1.00/ 1 1 1 1

2 14 2 15 3 0 3 1

Figure 2-5 Reference TMI-1 core depletion results with available fuel assembly design on Exelon computer with RESTART files

33

TMI-1 Reference Core Depletion with Available Fuel Assembly Design on Exelon Server (HFP operation) CASE STEP BOR - - - - - PEAK POWERS - - - - DENS POWER FLOW CRD PRES INLET CORE STEP EXP K-EFF NQ PPM AX / K A-O RAD NODE 3PIN G/CC % % POS PSIA T (F) EXP 1 2 2 2 2 2 2 2 2 2 3 0 0 1 2 3 4 5 6 7 8 0 0.000 0.100 0.200 0.400 0.800 1.000 5.000 10.000 15.000 20.214 20.601 1.00000 1.00000 0.99999 0.99999 0.99999 8 7 4 6 5 2212 1826 1793 1736 1686 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1 1 1 1 1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.36 1.33 1.33 1.32 1.33 1.33 1.36 1.37 1.39 1.36 1.35 1.39 1.37 1.37 1.36 1.36 1.36 1.38 1.38 1.40 1.36 1.36 1.54 1.50 1.49 1.48 1.48 1.48 1.50 1.50 1.52 1.46 1.46 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 0.715 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 19.445 19.545 19.645 19.845 20.245 20.445 24.445 29.445 34.445 39.659

TITLE EXP 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

SET 0 0 0 0 0 0 0 0 0 0 0

0.99999 5 1672 1.00/ 1 0.99999 7 1460 1.00/ 1 1.00000 8 1092 1.00/ 1 1.00000 13 646 1.00/ 1 1.00000 11 138 1.00/ 1 1.00000 5 101 1.00/ 1

100.0 100.0

555.5 40.046

Figure 2-6 Reference TMI-1 core depletion results with available fuel assembly design on Exelon computer without RESTART files
TMI-1 Reference Core Depletion with Available Fuel Assembly Design on PSU Server (HFP operation) CASE STEP 1 2 2 2 2 2 2 2 2 2 3 0 0 1 2 3 4 5 6 7 8 0 STEP EXP .000 .100 .200 .400 .800 1.000 5.000 10.000 15.000 20.214 20.601 K-EFF BOR PPM 2209 1825 1790 1733 1681 1669 1457 1091 647 138 100 - - - - - PEAK POWERS - - - AX / K A-O RAD NODE PIN 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1.00/ 1 1 1 1 1 1 1 1 1 1 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 1.36 1.33 1.33 1.32 1.33 1.33 1.36 1.37 1.39 1.36 1.35 1.38 1.37 1.37 1.36 1.36 1.36 1.38 1.38 1.40 1.36 1.36 1.54 1.50 1.49 1.48 1.48 1.48 1.50 1.50 1.52 1.47 1.46 DENS G/CC .712 .712 .712 .712 .712 .712 .712 .712 .712 .712 .712 POWER % 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 FLOW % 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 CRD POS 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 6296 PRES PSIA 2170 2170 2170 2170 2170 2170 2170 2170 2170 2170 INLET T (F) 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 555.5 CORE EXP 19.478 19.578 19.678 19.879 20.280 20.480 24.492 29.506 34.520 39.749 40.137 TITLE EXP SET .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 0 0 0 0 0 0 0 0 0 0 0

1.00000 1.00001 1.00000 1.00000 .99999 1.00000 .99999 1.00000 1.00000 1.00000 1.00000

1.00/ 1

100.0 100.0

2170 555.5

Figure 2-7 Reference TMI-1 core depletion results with available fuel assembly design on PSU computer without RESTART files

34

35 Table 2-2 Reference TMI-1 core with available FA designs (HFP depletion) Data BOC SOB (ppm) RPD Node 3-D Pin Power EOC SOB (ppm) Exelon Simulate run (With Restart File) 1815 1.38 1.39 1.52 Exelon Simulate Run (without Restart file) 1826 1.39 1.40 1.54 101 PSU Simulate run (without restart file) 1825 1.39 1.40 1.54 100

2.6 Fuel Rod Axial Zoning Model


Figure 2-8 shows the sample fuel rod axial zoning model, fuel enrichment and gadolina distributions. The Gd2O3 was used in the fuel matrix together with UO2 as a mixture in the intermediate zone as shown in the figure. The blanket regions at the upper and lower regions were designed to use low enriched fuel (3.0 w/o U-235). The only difference between UO2 and UO2/Gd2O3 fuel pin designs is to use of the fuel/absorber mixture in the intermediate zone for BP placement.

36

Top reflector Blanket Fuel 3.0 w/o U235 363.22 (cm)

348.08 (cm)

Intermediate Zone BP Placement Region 4.8 w/o U-235 UO2/Gd2O3

14.88 (cm) Blanket Fuel 3.0 w/o U235 Bottom reflector

0 (cm)

Figure 2-8 A sample UO2/Gd2O3 fuel rod axial zoning model for a TMI-1 fuel

37

2.7 Reference U/Gd FA designs


The reference U/Gd fuel assembly designs are summarized in this section. These fuel assembly designs were used in deterministic calculations. Figure 2-9, 2-10, and 2-11 show the vendor reference configurations for 4, 8, and 16 U/Gd fuel pins within a fuel assembly.

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 NoGd NoGd U/Gd Pin

NoGd NoGd NoGd NoGd NoGd NoGd

Figure 2-9 The vendor reference configuration for 4 U/Gd fuel pins

38

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 U/Gd Pin

NoGd NoGd NoGd NoGd NoGd NoGd NoGd NoGd


Figure 2-10 The vendor reference configuration for 8 U/Gd fuel pins

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 U/Gd Pin

NoGd NoGd NoGd NoGd NoGd NoGd NoGd NoGd


Figure 2-11 The vendor reference configuration for 16 U/Gd fuel pins

39

2.8 Reference TMI-1 Core Loading


Figure 2-12 shows the reference TMI-1 initial core loading pattern with fresh fuel locations in green color and used fuel locations in yellow. There are a total 10 fresh fuel locations given in the reference core loading. A fuel assembly is described in the Figure 2-12 by the following parameters: 1) Previous location of the FA except when it is fresh fuel (F). 2) FA type: The 0 represents used fuel without any boron absorber, 1 represents fresh fuel, and 2 represents used fuel with boron absorber., 3) Assembly uranium enrichment, 4) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o). A fuel assembly without Gd2O3 is shown with 00 - 00 notation. For example, Notation 16* - 03 shows 16 U/Gd fuel pins with mixed concentrations 8 and 3 w/o Gd2O3 or 08 - 02 shows 8 U/Gd fuel pins with 2 w/o Gd2O3 concentration. The explanation for the alphanumeric presentation in each fuel assembly in this figure is given in the lower left corner.

The reference TMI-1 core has 10 fresh fuel locations, and maximum fuel enrichment is 4.90 w/o U-235. Core assembly exposure distribution is given in Figure 2-13 for this loading pattern. Figure 2-14 and 2-15 summarize the results of the cycle depletion for normalized (radial) power and peak pin power distribution for fresh fuel assembly positions. The fresh fuel positions H-09, K-10, L-13, H-13 and M-13 violated maximum peak pin power constraint, which is 1.55 during the depletion calculations when no BP was used in these positions. The SOB let down curve is given in Figure 2-16 for this reference TMI-1 core loading pattern.

40

8
C-03 0 4.00 00 - 00

9
F 1 4.80 00-00 1 O-13 0 4.90 08 - 02 180

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 00-00 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 00-00 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 00-00 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 00-00 2 L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 00-00 6 F 1 4.90 00-00 9 F 1 4.90 00-00 10 N-13 0 4.85 08 - 04 180

14
M-11 0 4.30 08 04 90 F 1 4.87 00-00 5 F 1 4.87 00-00 7 L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

Degree Rotation

Figure 2-12 Reference TMI-1 core loading with fresh fuel assemblies without BA

41

8
C-03 12 13N 34.24 58.11 23.87

9
F 0.00 29.02 29.02 O-13 16C2 35.15 60.03 24.88

10
H-09 17F 26.52 51.54 25.02 F 0.00 27.68 27.68 L-09 17G 27.14 48.22 21.08

11
H-13 17B 26.15 48.65 22.5 K-12 17B2 25.54 48.20 22.60 M-14 16C2 38.85 58.93 20.08 F 0.00 23.77 23.77

12
K-10 17G 27.18 48.68 21.50 F 0.00 26.05 26.05 N-14 16C2 34.78 55.56 20.78 M-13 17D 26.32 46.26 19.40 O-12 17E 23.44 42.74 19.30

13
F 0.00 26.45 26.45 L-13 17F 26.73 50.78 24.05 F 0.00 26.58 26.58 F 0.00 25.29 25.29 F 0.00 20.85 20.85 N-13 17E 23.41 34.43 11.02

14
M-11 17A 24.99 42.16 17.17 F 0.00 22.83 22.83 F 0.00 23.28 22.28 L-14 17C 23.41 37.77 14.36 K-14 17C 24.17 33.58 9.41

15
N-12 16C2 47.69 56.23 8.54 K-11 16A2 47.00 55.63 8.63 M-12 16B2 47.93 55.43 7.50

Previous Location (F=Fresh) Previous Cycle (Reinserts) Batch Number BOC Burnup (GWD/MTU) EOC Burnup (GWD/MTU) Burnup Diff. (GWD/MTU)

Figure 2-13 Reference TMI-1 Core burnup distribution without BA

42
Radial Power distribution in the reference TMI-1 core for fresh fuel assembly locations without any BA core 1.8

1.7

1.6 Radial (Normalized) Power

1.5

1.4

1.3

1.2

1.1

1 0 5 10 15 Assembly Average Burnup (GWd/MTU) H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 M-13 N-13 20 25 30

Figure 2-14 Radial Normalized power distribution for reference core loading without BA

Peak Pin Power distribution in the reference TMI-1 core for fresh fuel assembly locations without any BA core 1.9

1.8

1.7

Peak Pin Power

1.6

1.5

1.4

1.3

1.2

1.1 0 5 10 15 Assembly Average Burnup (GWd/MTU) H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 M-13 N-13 20 25 30

Figure 2-15 Peak pin power distribution for reference TMI-1 initial core loading without BA

43

SOB Concentration Change versus Cycle Burnup for reference TMI-1 Core Loading without BA 3500 3000 2500 SOB (ppm) 2000 1500 1000 500 0 0 5 10 15 20 25 Cycle Average Burnup (GWd/MTU)

Figure 2-16 Soluble boron concentration change with cycle average exposure

44

2.9 Superior BP Loading Design


The superior BP loading was developed after 5th iteration, and it is shown in Figure 2-17. The Gd use in the core loading pattern satisfied enough absorption to keep all of the safety parameters below their constraints. The maximum Gd2O3 concentration used in the superior BP loading was 8 w/o.

Figures 2-18 and 2-19 show the results of the cycle depletion for superior BP loading core. Maximum normalized power occurred at the core position H-09, and it was 1.392. The maximum peak pin power was 1.548 and took place at position K-14 shown in figure 2-18. Soluble boron concentration was reduced continuously by iterations. BOC soluble boron concentration at equilibrium xenon (Xenon) was 1664 ppm for this superior loading pattern. The automated and modular program took 1.16 CPU min on IBM RS/6000 workstation to develop the superior BP pattern.

45

8
C-03 0 4.00 00 - 00

9
F 1 4.80 16* - 07 1 O-13 0 4.90 08 - 02 180

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 16* - 07 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 16* - 08 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 16* - 06 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 16* - 06 2 L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 16* - 02 6 F 1 4.90 16* - 06 9 F 1 4.90 00-00 10 N-13 0 4.85 08 - 04 180

14
M-11 0 4.30 08 04 90 F 1 4.87 00-00 5 F 1 4.87 00-00 7 L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

Degree Rotation

Figure 2-17 Superior TMI-1 BP loading pattern

46

Cycle 16 - Superior Core loading Maximum pin power distribution in the core for fresh fuel assembly locations 1.6 1.55 1.5 Max pin power 1.45 1.4 1.35 1.3 1.25 1.2 1.15 1.1 0 H-09 H-13 5 K-10 K-12 10 K-14 15 20 M-13 25 N-13 30 pin power limit Burnup (GWd/MTU) L-13 L-14 M-11
Pin power constraint

Figure 2-18 Peak pin power distributions for superior BP loading core

Cycle 16 - optimal core loading Normalized power distribution in the core for fresh fuel assembly locations 1.45 1.4 1.35 Normalized power 1.3 1.25 1.2 1.15 1.1 1.05 1 0 H-09 H-13 5 K-10 K-12 10 K-14 15 Burnup (GWd/MTU) L-13 L-14 M-11 20 M-13 25 N-13 30 pin power limit
Normalized power constraint

Figure 2-19 Radial normalized power distributions for superior BP loading core

2.10 Final TMI-1 BP Loading Design Using Fine-Tuning Technique


Figure 2-20 presents the final BP loading design for the reference TMI-1 core. The explanation for the alphanumeric presentation in each fuel assembly in this figure is given in the lower left corner. Maximum Gd2O3 concentration used in the optimal core is 7 w/o. The maximum peak pin power was 1.546 and took place at K-14 assembly location shown in Figure 2-21. Maximum RPD power was observed in position H-09, and it was 1.366 during the depletion shown in Figure 2-22. BOC soluble boron concentration at equilibrium Xe is determined as 1694 ppm for this final BP loading pattern.

Table 2-3 presented the comparison of core data for superior and final BP loading. The maximum RPD power reduced from 1.377 to 1.366 with heuristic selection technique. The final BP loading pattern has 73.485 kg heavy Gd in the core, which represents 22 % mass savings compared to superior BP design. However, the maximum pin power value decreased to 1.546, and it occurred at position K-14.

48

8
C-03 0 4.00 00 - 00

9
F 1 4.80 16* - 07 1 O-13 0 4.90 08 - 02 180

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 16* - 02 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 16* - 03 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 16* - 02 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 16* - 02 2 L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 16* - 02 6 F 1 4.90 16* - 02 9 F 1 4.90 00-00 10 N-13 0 4.85 08 - 04 180

14
M-11 0 4.30 08 04 90 F 1 4.87 00-00 5 F 1 4.87 00-00 7 L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

Degree Rotation

Figure 2-20 Final TMI-1 BP Loading pattern with deterministic method and fine tuning technique

49

Cycle 16 - optimal core loading Maximum pin power distribution in the core for fresh fuel assembly locations 1.6 1.55 1.5 Max pin power 1.45 1.4 1.35 1.3 1.25 1.2 1.15 1.1 0 H-09 H-13 5 K-10 K-12 10 K-14 L-13 15 Burnup (GWd/MTU) L-14 M-11 M-13 N-13 pin power limit 20 25 30
Pin power constraint

Figure 2-21 Peak pin power distributions for reference TMI-1 final BP loading design
Cycle 16 - optimal core loading Normalized power distribution in the core for fresh fuel assembly locations 1.45 1.4 1.35 1.3 Normalized power 1.25 1.2 1.15 1.1 1.05 1 0 5 10 15 Burnup (GWd/MTU) H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 M-13 N-13 pin power limit 20 25 30
Normalized power constraint

Figure 2-22 Radial normalized power distributions for reference TMI-1 final BP loading design

50 Table 2-3 Comparison of data for superior and final BP loading design Superior BP Design Final BP Design

Parameter

Maximum peak pin power and Position Maximum RPD power and position Number of U/Gd pins in Octant Core Total heavy Gd Mass in the Core (kg) Soluble Boron (ppm) at Equilibrium Xe CPU time (min)

1.548 (K-14) 1.377 (H-09) 112 94.365 1682 1.16

1.546 (K-14) 1.366 (H-09) 112 73.485 1694 6.43

2.10 Conclusions
The developed deterministic method appears to perform as expected. This method first modifies Li's [14] method to automatically obtain a good initial BP design followed by a further reduction in Gd using a special Gd adjustment algorithm. The method then fine-tunes the design using a heuristic BP selection process. The new deterministic method presented in this study does not use the HPD as a guide, but it uses the calculated radial power constraint as a guide. It employs a recursion formula to affect an acceptable power design during depletion. It also allows greater flexibility in designing the BPs in the core by not trying to force the power distribution to perform close to the HPD, which is impractical to achieve. The method develops a RPD

51 that satisfies power constraints during depletion, which is the objective of the program. It is a practical method that satisfies the maximum peak pin power and SOB constraints during depletion calculations by implementing sufficient margin in the RPD constraint. Only fine-tuning of the maximum peak pin power is required to finalize the optimal BP design.

CHAPTER 3 OPTIMIZATION OF UO2/Gd2O3 FUEL PIN CONFIGURATIONS


3.1 Introduction
Chapter 3 presents the methodologies to determine the optimal UO2/Gd2O3 fuel pin configurations within a fresh fuel assembly. The UO2/Gd2O3 fuel pin configurations within a fuel assembly affects the fuel assembly internal pin power distribution, assembly power itself, neighboring fuel assembly powers in the core, and residual binding at EOC. Inserting UO2/Gd2O3 fuel pins in the fuel assembly lowers its Kinf and thus, lowers its normalized power in the core while it causes the power to increase in the neighboring fuel assemblies to conserve the total core power.

The optimal BP design is one that minimizes the total Gadolinium (Gd) or BP in the core. It should not violate the maximum peak pin power constraint during depletion, and it should also minimize the residual binding of gadolinium absorber. This process is highly non-linear making it difficult to achieve optimum results. The problem definition, genetic algorithm model and its operators, input parameters, decision variables, design constraints, and the solution representation are defined and explained in the following sections. In order to find the best method to optimize the BP design, different objective function and constraint definitions were studied in detail and the comparison of the results are presented in this Chapter.

53 The solution methods were separated into three phases. The first phase performed the optimization calculations only on a single fuel assembly. The objective function was to minimize the peak pin power within a single fuel assembly in an infinite array. It was assumed that minimizing the peak pin power within a single fuel assembly would also minimize the peak pin power during core depletion. This did not happen; however, the technique is presented here to show the initial development of the GA algorithm.

In this first study those UO2/Gd2O3 fuel pin configurations were identified that had the potential to minimize peak pin power when depleted in the core. Having selected these UO2/Gd2O3 fuel pin configurations, the next step was to give preference to those BP designs when evaluating the BP designs in the core that minimizes the total Gd in the core. The design constraints were checked by performing reactor physics calculations for a reference TMI-1 fuel assembly with CASMO-3 lattice physics code. This first phase is called Optimization Calculations on a Single Fuel Assembly, and it is presented in Section 3.7.

The second phase also involved a single fuel assembly optimization problem as performed above, but in this case the Kinf of the BP design was to be minimized. The Kinf of a fuel assembly at BOC calculated by CASMO-3/4 is a meaningful parameter when evaluating the effect of a BP design on the maximum peak pin power during core depletion. It was important to find a direct relation between single fuel assembly parameters and its corresponding physics parameters characteristics during core depletion for the U/Gd Fuel Assembly (FA) BP designs. It is known that Kinf affects the power

54 distribution in the core during depletion, which relates to the maximum peak pin power in the core during depletion, which is the main design constraint. The Kinf of the fuel assembly is a function of BP arrangements within a fresh fuel assembly and thus is related to the maximum peak pin power constraint. However, the Kinf of the FA maintains its same value when placed in the core.

The relation between a fuel assembly Kinf, its normalized power, and its peak pin power during core depletion is presented in Section 3.8. The objective is to determine the best single fuel assembly parameter to correlate with the actual peak pin power in the core during depletion. The purpose was to find a single fuel assembly parameter that could be used to correlate with the core peak pin power in such a way as to define invalid single fuel assembly BP designs. In this way, the single FA parameter could be used to eliminate the use of reactor physics code when invalid or bad BP designs appear in the optimization process. This technique could potentially reduce the computational time during optimization calculations. It was shown that the FA Kinf can be this parameter. As a consequence, in this section the objective function was to minimize the BOC Kinf of a single fuel assembly BP design.

In the third or final phase core calculations were performed in the optimization calculations to meet the objective function while meeting all of the design and safety constraints. This phase took advantage of the effectiveness that BOC Kinf of a fuel assembly could be used as a filter in the optimization calculations, and it was used to filter out bad designs to eliminate the need for evaluation process with reactor physics

55 calculations. The actual calculations were performed in a systematic way to affect the optimal solutions in an efficient process. First, the calculation determined the BP designs that were below the core maximum peak pin power constraint of 1.55. The Kinf filter was defined to eliminate bad BP designs without performing unnecessary reactor physics calculations. The GA code assigned high negative penalty to reduce the probability of selecting bad designs in future generations. The next step was to meet the objective functions defined in GA code. It should be mentioned that on minimizing the total Gd amount in the core, the GA attempts also to minimize residual binding. The objective function was defined to minimize residual binding, which means to maximize EOC soluble boron (ppm) without violating maximum peak pin power constraint.

3.2 Problem Definition


A TMI-1 (1515) PWR sample fuel assembly design was used in the optimization calculations. The fuel assembly has an octant symmetry shown in Figure 3-1, and the model includes 4 guide tube (GT) locations and 32 total fuel pin locations including the positions not allowing any UO2/Gd2O3 fuel pin placement. The UO2/ Gd2O3 fuel pins can be located in any of the 22 different locations numbered in the figure.

The objective was to determine the optimal UO2/ Gd2O3 fuel pin positions within an octant fuel assembly model, which minimizes the total Gd in the core without violating any of the design constraints. The calculations proceed by altering the BP design in the reference TMI-1 core. The GA code creates new BP designs, which are placed in the reference core. And then the core is depleted to validate the constraints. The

56 optimal solution of this problem is highly complex requiring special technique to obtain the solution efficiently.

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 NoGd NoGd Possible U/Gd Pin Positions

NoGd NoGd NoGd NoGd NoGd NoGd

Figure 3-1 A reference TMI-1 octant PWR fuel assembly model

57

3.3 Genetic Algorithm Model


The genetic algorithm (GA), which is a stochastic method, provides an opportunity to work conveniently with discrete functions and without any derivative information. It is based on concepts from biological genetics, where improving the characteristics in a population of organisms is viewed as an optimization problem, with the best individual surviving and producing the most offspring [16]. A general GA code [41] designed to be very flexible to optimize a general class of different systems was modified to improve the mutation operator. Some modifications were made specific to this problem to improve the codes efficiency.

In the GA, the objective function for the optimization can be viewed as a black box in which the details of evaluation of the objective function are not important [16]. For the pin arrangement problem, the objective function is evaluated by reactor physics codes, and the black box approach means any reactor physics code producing acceptable results can be used for the GA optimization. Figure 3-2 shows the data flow methodology between the reactor physics code and the genetic algorithm code.

58

Genetic Algorithm Optimization Code

Interface Reactor Core/Lattice Physics Codes Code-Specific Results (Peak pin power, Kinf)

Interface

Code-Specific Assembly/Core Loading Input Deck

Figure 3-2 Data flow between Generic Code and Reactor Core or Lattice Physics Codes for a given input parameters

3.4 Input Parameters and Decision Variables


The U-235 enrichment of the uranium fueled pins, the Gd2O3 concentration, and the number of UO2/Gd2O3 pins within a fuel assembly, and Gd isotopes (Gd-152, Gd154, Gd-155, Gd-156, Gd-157, Gd-158, and Gd-160) weight fractions are the input parameters for this problem.

The UO2/Gd2O3 fuel pin positions within the octant fuel assembly symmetry are the decision variables to be optimized during the GA search. Maximum three UO2/Gd2O3 fuel pins should be placed in the octant fuel assembly model shown in Figure 3-1 to simulate 20 UO2/Gd2O3 fuel pins within a fuel assembly design. Thus, there are three decision variables used in the GA calculations each representing one of UO2/ Gd2O3 fuel pin location in the model (see Figure 3-3). Input parameters such as number of UO2/Gd2O3 fuel pins and Gd2O3 concentration should be changed within allowed range to

59 develop good BP designs. Good BP designs were determined by performing GA run each time for different input parameters. Then, optimal BP fuel pin configurations were used to create a huge burnable absorber library with cross section generation code (CASMO3), which is described later in this chapter, and they were used in the core BP optimization problem solution process to perform reactor physics calculations for evaluating solutions.

Decision space contains allowable values of the decision variables used in these GA calculations. Figure 3-3 shows the decision space for the selection of UO2/Gd2O3 fuel pin positions in the octant fuel assembly symmetry. The decision space was separated into two parts shown in Figure 3-3, a feasible and an infeasible decision space. The feasible decision space represents all of the possible 22 fuel pin positions, which can be used during GA run, and infeasible decision space shows the rest of the empty slots in 5 bit representation. In order to eliminate the bias in the solutions, GA evaluates the solution individual as far as they are in the feasible decision space while it assigns high negative penalty for infeasible decision space members.

The geometric symmetry factors within an octant fuel assembly model for each position are shown in Figure 3-4. The number of UO2/Gd2O3 fuel pins in the input data should be exactly the same with the calculated number by using geometric symmetry factors defined for each UO2/Gd2O3 fuel pin position to evaluate the solution with reactor physics calculations. Otherwise, GA algorithm model skips this solution without evaluating, and it assigns high negative penalty to reduce the probability of being selected

60 for the next generation. This procedure saves considerable run time in GA runs, and it eliminates the possible source of bias in the solution. The number of U/Gd fuel pins within a model is calculated by using the following formula:

Number of UO2/Gd2O3 fuel pins = 8 (SF1 + SF2 + SF3), where SF: Symmetry Factor of UO2/Gd2O3 fuel pin positions in the model SF=1.0, means that the model includes the location, SF=0.5, means that the model includes the half of the location,

61

Figure 3-3 Decision space for the selection of UO2/ Gd2O3 fuel pin positions

62

Guide Tube NoGd NoGd 0.5 0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.5 1.0 1.0 1.0 1.0 1.0 0.5 1.0 0.5 NoGd NoGd Possible U/Gd Pin Positions

NoGd NoGd NoGd NoGd NoGd NoGd

Figure 3-4 Geometric symmetry factors for octant fuel assembly

3.4.1 Design Constraints Design constraints for the optimization of UO2/ Gd2O3 fuel pin configurations problem are given as follows: 1. Gd2O3 concentration must be 8 %, and number of UO2/Gd2O3 fuel pins per Fuel Assembly 20 pins

63 2. The UO2/Gd2O3 fuel pins must be loaded in a symmetric pattern in the fuel assembly 3. The fuel enrichment of a UO2/ Gd2O3 pin should be less than the nominal enrichment of the fuel assembly for natural enriched Gd2O3 use due to maximum fuel temperature limitation for UO2/Gd2O3 fuel 4. UO2/ Gd2O3 fuel pins can not be located next to each other 5. UO2/ Gd2O3 fuel pins cannot be located in the instrumentation tube (IT) or in the guide tube (GT) location. 6. Maintain octant assembly symmetry in each iteration 7. Maximum peak pin power in core depletion calculations should be less than 1.55

3.5 Solution Representation, Encoding and Decoding Process, and Eliminating Bias
The decision variable of a UO2/Gd2O3 fuel pin position within a fuel assembly is shown in Figure 3-5. The decision space was separated into two parts shown in Figure 35, a feasible and an infeasible decision space. Feasible decision variables represent the possible position numbers within a fuel assembly in the octant symmetry. Feasible decision region members do not violate the problem geometry, and they represent potential solutions for the decision variable. Infeasible decision variable exceeds the total possible number of fuel pin positions in an octant fuel assembly symmetry. In order to eliminate the bias in the solutions, GA evaluates the solution individual as far as they are

64 in the feasible decision space while it assigns high negative penalty for infeasible decision space members. For example, the GA operators can create the solution, which may have the decision variable of UO2/Gd2O3 fuel pin position larger than 22. In that case, the GA code assigns high negative penalty (-2000) without performing reactor physics calculations to evaluate this solution.

Deb K. [42] suggested a method so that the objective function value is not computed for any infeasible solution. The objective function is used to evaluate all feasible solutions in the population. All infeasible solutions are evaluated according to their constraint violations by using penalty terms. The objective function value and constraint violation are not combined for any solution in the population. This penalty shows that the solution is in the infeasible solution space, and it violates the problem geometry. The objectives of this constraint handling technique are to eliminate the bias which can be produced by using non-uniform decision variable distribution in binary representation, and to remove individuals from the mating pool by reducing their probability of being selected for the next generation due to their low fitness value. With this constraint handling technique any feasible solution was preferred to any infeasible solution. Among two feasible solutions, the one having a better objective function value was selected for the next generation [43]. It is important to understand the bias concept, which may be produced by using classical binary GA approach.

65

Figure 3-5 Encoding process for decision variable of UO2/Gd2O3 fuel pin position

Classical approach of reducing the bias in the solution was to prefer some decision variables more than the other to fill out the possible empty integer slots in binary representation [16]. However, the genotype is biased towards some fuel pin positions more than others in this approach. This bias cannot be eliminated for this representation,

66 and it can be only reduced by adding additional bits to genes to some certain degree for classical approach. However, this approach causes to enlarge the problem size and the total calculations due to addition of extra bits.

The phenotype represents the actual UO2/Gd2O3 fuel pin position numbers within a fuel assembly as applied to the reactor physics code, and the genotype is a bit-string representation of the phenotype for use in the genetic algorithm operations. A gene is composed of bits, each of which may exhibit 0 or 1 called an allele. The conversion process of phenotype (UO2/Gd2O3 fuel pin positions) to genotype (binary form of UO2/Gd2O3 fuel pin positions) is called encoding, and the reverse process is named as decoding.

Both a sample phenotype and corresponding genotype structure are shown in Figure 3-6 for a sample UO2/Gd2O3 fuel pin configuration. This figure shows how the phenotype of integer numbers is encoded into a genotype structure for the UO2/Gd2O3 fuel pin configuration. The three UO2/Gd2O3 fuel pin positions in the configuration require using a total of three decision variables. The genotype has three genes each represented by a 5 bit binary gene. The total of 15 bits representation is required to encode the decision variables representative of UO2/Gd2O3 fuel pin positions. Each of the genes in genotype structure is then decoded into phenotypes and evaluated by a reactor physics code for evaluation of its fitness value. By choosing high fitness values for the next generation, the GA continues to improve the BP pattern design until the optimum is attained.

67 A phenotype and a genotype structure are summarized as follows: Phenotype = [|UO2/Gd2O3 fuel pin position1| | UO2/Gd2O3 fuel pin position2| | UO2/Gd2O3 fuel pin position3|| And, Genotype = [|5 bit||5 bit||5 bit| ] = 15 bit

Sample Phenotype and Genotype Structure Gene No Phenotype Genotype Gene Length 1 0 1 22 1 5 bit 0 1 0 0 2 2 0 5 bit 0 1 0 0 3 8 1 5 bit 1 1

Figure 3-6 Sample phenotype and genotype structure

3.6 Genetic Operators


Selection, cross-over and mutation operators were used to generate more individuals during evaluation. Selection operator promotes better solutions by selecting good solutions to replace less optimal solutions. The main objective is to increase the number of good solutions and eliminate bad solutions in a population, while keeping the population size constant. The elitism feature is used to inject best solutions into next generation. It selects the individual with the highest fitness and returns it as a parent for next generation. New population individuals are checked to make sure whether best individual of the previous generation is replicated or not. If it is not replicated, the best parent is reproduced for the current generation by randomly selecting its slot in the

68 population members. It ensures that the maximum objective function value within a population can never reduce from one generation to the next [42, 43].

Tournament selection methods are utilized based on holding tournaments between a number of individuals (two or more) and selecting the best member of a tournament group for survival to the next generation and placing in the mating pool. Then, two other solutions are randomly chosen and the better one is placed in the mating pool again. Each solution in the population can participate in two tournaments. The best solution wins in both tournaments and makes two copies of itself in the new population. This process is repeated N times in order to select N members of the population. The bad solutions lose in the tournament, and they are eliminated from the population during the process. Figure 3-7 shows the sample schematic diagram on how tournaments happen between the six population members and the resulting mating pool members.

69

Figure 3-7 Tournaments between the six population members and the resulting mating pool

However, the selection operator makes more copies of good solutions, it can not create any new solutions in the population. New solutions are created by cross over and mutation operators during GA calculations. Two solutions are randomly picked from the mating pool, and their bit string forms (genotypes) exchange bit values by cross-over operator to generate two new off-springs. Single point cross over operator randomly chooses a crossing point along bit string and exchanges all the bits on the right side of this point. Uniform cross over operator does not have any fixed points along the bit strings, and each allele can be swapped with given probability by considering all binary bits. Two parents are selected and two children (offspring) produced. For each bit

70 position on the two children, it is decided randomly which parent contributes its bit value to which child. Uniform cross over probability of 0.5 is used to choose every bit for the production of one child. Figure 3-8 shows the single point and uniform cross over operators used in the binary GA code. The first, fifth, and eleventh bits are exchanged between the parents for uniform cross over operator shown in the figure. Parents Children 111100010100

Solution 1

111101111101

Solution 2

001110010100

001111111101

Single Point Cross Over Parents 1111011111011 Children 0111111111001

Solution 1

Solution 2

001110010100

101100010110

Uniform Cross Over

Figure 3-8 Single Point and Uniform Cross over operators

Mutation operator creates new individuals by making changes to a single solution. Its main purpose is to maintain diversity within the population and reduce fast convergence to local optima solutions [44]. Mutation alone performs a random walk through the search space changing bits. Figure 3-9 shows how a new individual can be

71 obtained after bit-wise mutation operation. A random number is created between 0 and 1, and then it is used to determine the number of bits to be skipped from the current bit for performing the mutation operation. Figure 3-9 shows how the fifth bit in the string is mutated to create a new child.

Parent
1111011111011

Child
1111111111011

Solution

Figure 3-9 Bit-Wise mutation operator

3.7 Optimization Calculations on a Single Fuel Assembly


Optimization calculations performed only on a single fuel assembly used an objective function to minimize the peak pin power within a single fuel assembly in an infinite array during fuel assembly depletion. The GA created to utilize this objective function minimized the peak pin power so as to identify those UO2/Gd2O3 fuel pin configurations that minimizes peak pin power when depleted in an infinite array. Thus, it is informative to describe this first GA. 3.7.1 Minimizing Peak Pin Power within a Single Fuel Assembly during Depletion in an Infinite Array The objective function defined to minimize the assembly internal peak pin power during the depletion is as follows:

F ( P) = w p P P = Peak pin power during depletion in an infinite array wp = -100 (weight Factor, minus sign represent the minimization problem)

72 F(P) = Fitness value of the individual The fitness value of the individual reflects the peak pin power value of the assembly. A lower fitness value means a higher peak pin power.
3.7.2 GA Flow Diagram

The GA starts with an initial population. The GA code checks randomly created population members to determine whether they are in the feasible decision space or not. Population members in the feasible decision space are evaluated by using the fuel assembly lattice physics code, CASMO-3. High fitness penalty is assigned for infeasible population members created during GA run without performing lattice physics calculations. It uses selection, crossover and mutation operators to create more generations after every evaluation. Figure 3-10 outlines the flow diagram of the GA, which was used for a single fuel assembly optimization problem.

73

Fuel data: U enrichment No of Gd pins Gd2O3 concentration

Create Population Randomly selected U/Gd fuel pin positions Mutation Operators Feasible Solution? Cross Over Operators yes Selection Operators No
Assign high penalty without performing reactor physics calculations

Evaluate the solution

Reactor physics/Fuel lattice physics code

Create next generation

No

Last generation?

yes END

Figure 3-10 Sample Genetic Algorithm flow diagram

74
3.7.3 Results

The initial population size of 30 solutions was created to make the problem feasible and to provide efficient run time for GA [45]. The base mutation probability was considered to be the inverse of the population size (1/Npop). The elitism selection procedure was used to ensure that the best solution in each generation was always selected for the next generation. The previously tested and recommended David Carrolls [41] genetic parameters and operators were used in the calculations. Some input data are presented in the Table 3-1 for the GA code.
Table 3-1 Important GA input data for GA code

Parameter Population Size (Npop) Max. Generation Number Pmutation Cross over probability

Value 30 50 (1/Npop) = 0.033 0.5

Table 3-2 summarizes the genetic algorithm results for within a fuel assembly in an infinite array optimization problem. The optimum arrangement of the UO2/Gd2O3 fuel pins within the assembly is not a function of uranium enrichment and Gd2O3 concentration. However, minimum peak power reduces by increasing uranium enrichment. The optimum locations depend on the total number of UO2/Gd2O3 fueled pins within the assembly. The minimum peak power increases with an increase in the number of UO2/Gd2O3 pins. The increased amounts are respectively 4.3 %, 3.7 %, and 3.2 % for 4.0 w/o, 4.5 w/o and 5.0 w/o uranium enrichments.

75
Table 3-2 Optimum UO2/Gd2O3 pin positions in octant assembly model for different assembly uranium enrichment and number of UO2/Gd2O3 pins (4 w/o Gd2O3)

Total Number of Gd pin 12 16 16* 20

U=4.0 w/o Optimum Gd pin positions 6,7,15 3,14,15 3,14,15 5,13,22 Minimum Peak Pin Power 1.102 1.113 1.120 1.150

U=4.5 w/o Optimum Gd pin positions 6,7,15 3,14,15 3,14,15 5,13,22 Minimum Peak Pin Power 1.099 1.111 1.118 1.140

U=5.0 w/o Optimum Gd pin positions 6,7,15 3,14,15 3,14,15 5,13,22 Minimum Peak Pin Power 1.097 1.111 1.117 1.133

* Mixed Concentrations 8 w/o and 4 w/o of Gd2O3

GA applications reduced run- time to determine optimal arrangement as compared to deterministic calculations. GA converged to optimal configurations after 10 generation with 10 minutes CPU time. It saved considerable man-power to solve single fuel assembly problem.
3.7.4 Core Depletion Comparison with Available U/Gd BP Designs

Optimal U/Gd fuel pin arrangement, which minimized assembly internal power peaking by considering a single fuel assembly calculation, replaced the available BP design in the hot fuel assembly position (Position H-09) and the core was depleted. It resulted in higher maximum pin power value for the given reference TMI-1 core compared to using available design in the core. The maximum peak pin power was 1.629 for optimal arrangement and it was 1.535 for reference fuel assembly design shown in Figures 3-11 and 3-12.

76 This result showed that the BP design of minimizing peak power within a single fuel assembly in an infinite array would not necessarily end up with less peak power during core depletion. The fuel assemblies in the core are always represented with their Kinf values. The available BP fuel assembly design had a BOC Kinf value of 1.059, which is less than the BOC Kinf value of the optimal BP design (1.065) shown in Figures 3-11 and 3-12. Both figures show the effect of the core environment on both the available and minimum internal peak power U/Gd FA designs. The Kinf increase of a single fuel assembly causes the RPD and peak power to increase in the core. In the fuel management optimization calculations, a fuel assembly RPD and peak pin power values for any fuel position are reduced by choosing a lower Kinf design.

77

10

Fresh 4.80 TYPE17GD03 Kinf: 1.058 RPD:1.361 Peak Pin Power:1.535

Used 4.90 TYPE08GD02 Kinf: 0.942 RPD:1.155 Peak Pin Power:1.226

Fresh 4.80 TYPE17GD03 Kinf: 1.058 RPD:1.358 Peak Pin Power:1.536

Cycle 12 4.00 TYPE00GD00 Kinf: 0.881 RPD:.994 Peak Pin Power:1.053

Fresh (H-09) 4.80 TYPE17GD03 Kinf: 1.059 RPD: 1.364 Peak Pin Power:1.539

Used 4.90 TYPE08BP02 Kinf: 0.998 RPD:1.271 Peak Pin Power:1.333

Fresh 4.80 TYPE17GD03 Kinf: 1.058 RPD:1.361 Peak Pin Power:1.535

Used 4.90 TYPE08GD02 Kinf: 0.942 RPD:1.155 Peak Pin Power:1.226

Fresh 4.80 TYPE17GD03 Kinf: 1.058 RPD:1.358 Peak Pin Power:1.536

Figure 3-11 TMI-1 reference core with available fuel assembly designs

78

10

Fresh 4.80 TYPE17GD03 Kinf: 1.065 RPD:1.426 Peak Pin Power:1.629

Used 4.90 TYPE08GD02 Kinf: 0.938 RPD: 1.187 Peak Pin Power:1.254

Fresh 4.80 TYPE17GD03 Kinf: 1.065 RPD:1.408 Peak Pin Power:1.615

Cycle 12 4.00 TYPE00GD00 Kinf: .878 RPD:1.028 Peak Pin Power:1.084

Fresh (H-09) 4.80 TYPE17GD03 Kinf: 1.065 RPD: 1.426 Peak Pin Power:1.629

Used 4.90 TYPE08BP02 Kinf:0.994 RPD:1.295 Peak Pin Power:1.363

Fresh 4.80 TYPE17GD03 Kinf: 1.065 RPD:1.426 Peak Pin Power: 1.629

Used 4.90 TYPE08GD02 Kinf: 0.938 RPD:1.187 Peak Pin Power: 1.254

Fresh 4.80 TYPE17GD03 Kinf: 1.065 RPD:1.408 Peak Pin Power:1.615

Figure 3-12 TMI-1 reference core design with fuel assembly design (at position H-09) with minimum assembly internal peak power during depletion

79

3.8 The Relation between BOC Kinf and RPD of a Fuel Assembly
The GA code used in Section 3.7 is now modified to use a different objective function. The objective function in this section develops BP configuration that minimized the BOC Kinf of the BP designs. Application of this objective function is presented in this section.

It was expected that if the BOC Kinf value of U/Gd fuel pin designs is reduced, this would lead to a decrease in both the RPD and the peak pin power during core depletion for a given fresh fuel assembly position. The detailed sensitivity studies on BOC Kinf value of a BP design, U/Gd fuel pin configurations, and resulting parameters in the core environment were performed to clarify this relation. Therefore, the objective function was changed to minimize BOC Kinf values instead of minimizing peak pin power for a single fuel assembly. The purpose was to avoid long reactor physics calculations by eliminating the need to perform reactor physics calculations for solutions evaluations. The genetic algorithm calculation created new U/Gd fuel pin configurations within a single fuel assembly, which minimized Kinf at BOC, and then this design was used in the core with reactor physics code to determine the acceptability of the results.
3.8.1 Objective Function of Minimizing BOC Kinf for U/Gd FA Design

The objective function was defined to minimize BOC Kinf value of the U/Gd FA design. GA population members were evaluated by using CASMO-3 lattice physics code by using BOC Kinf parameter in the objective or fitness function as following: Fitness=f1= w1BOCKinf (minimizing BOC Kinf value) where w1 = - 100 (- sign shows maximization problem)

80
3.8.2 Results

GA runs were performed for different number of U/Gd fuel pins, and the results are summarized in the following Table 3-3. Because, TMI-1 reference core uses 16 U/Gd fuel pins with mixed Gd2O3 concentrations (4 and 8 w/o together in the BP design) at the hot fuel assembly position in the core loading, the optimal U/Gd fuel BP design for 16 U/Gd fuel pins was selected to test its design performance in the core. Table 3-4 presents reactor physics calculation results for both the reference core with available vendors U/Gd FA designs and new optimal designs with minimum BOC Kinf.

Table 3-3 U-235=4.5 w/o, Gd2O3=4 w/o CASMO-4 GA optimization results with minimizing BOC Kinf objective

Optimum U/Gd Fuel Pin Total Number Positions Within an of Gd pin Octant FA Model 12 6,15,11 16 15,6,12 16 mixed 15,6*,18* 20 15,18,9 20 mixed 15,18*,9* * Fixed Gd2O3 at 8 w/o

Peak Pin Power

Minimum Kinf at BOC 1.179 1.136 1.1157 1.096 1.0716

Fitness

1.102 1.133 1.165 1.178 1.195

-117.86 -113.65 -111.57 -109.57 -107.16

Table 3-4 Reference core with available FA designs and new optimal designs with minimum BOC Kinf

Data RPD Node 3-D Max Peak Pin Power Kinf at position H-09

Reference FA design 1.39 1.40 1.54 1.059

Minimum assembly peak power 1.44 1.46 1.65 1.065

Minimum BOC Kinf 1.43 1.43 1.56 1.051

81 Table 3.4 shows that Kinf of the fuel assembly at position H-09 and K-10 was reduced from 1.059 to 1.051 with using in optimization process the objective function of minimizing BOC Kinf instead of minimizing assembly peak pin power during the depletion. Kinf of the fuel assembly represent the fuel assembly power. However, the fuel assembly power is also function of neighboring fuel assemblies. The flux spectrum is affected by fuel assembly arrangements in the core. Therefore, U/Gd fuel pin arrangements within a fresh fuel assembly also affects the Kinf and RPD distribution within this region of the core. There is a slight increase in Kinf values of H-08, H-09, K08 and K-10 positions. RPD power of H-09 drops from 1.361 to 1.330 with new arrangements due to Kinf reduction.

On the other hand, peak pin power increases from 1.539 to 1.546 for new optimal design of H-09 fuel position. There is a strong relation between RPD of an assembly and its Kinf value. They follow the similar behavior in the core. But, peak pin power has different tendency as could be observed from the above data. Peak pin power is a function of U/Gd pin arrangements, plus core environment, and other factors. However, 3-D maximum peak pin power value resulting from the use of a new optimal design was slightly larger than the one resulting from use of available design. It was 1.56, which is larger than the design constraint of 1.55. The Kinf of the optimal design drops gradually from 1.059 to 1.051. Thus, maximum RPD power in the core shifts to another location to conserve the total core power. The sensitivity analysis showed that Kinf value of U/Gd fuel assembly design is a one of important factor, but it did not decide whether this design was acceptable or not. Therefore, coupled lattice and core physics calculations

82 should be performed to evaluate the performance of the design in the core during depletion calculations.

3.9 Kinf Filter Use in Coupled Optimization Calculations


Section 3.9 summarizes the coupled lattice and core optimization calculations with Kinf filter use. The main objective is to reduce the calculation time and to evaluate the solutions in the core by performing reactor physics core depletion calculations. Maximum peak pin power and EOC soluble boron concentration are two parameters to be used for the evaluation of a candidate solution during a GA run.
3.9.1 The Relation between Kinf, Maximum Radial Power Distribution and Maximum Peak Pin Power

The Kinf represents the criticality of an infinite array of the fuel assembly, and it is a function of the combination of U-235 enrichment, number of U/Gd fuel pins, and their Gd2O3 concentrations. The Kinfs distribution determines the power distribution in the core. One should study the relation between a fuel assembly Kinfs and its peak pin power and radial power distribution value to understand the phenomena.

Peak pin power is a parameter of a fuel assembly, and it represents the maximum power value of a fuel pin within a fuel assembly. This parameter was affected by the fuel pin configurations within a fuel assembly, and the placement of the U/Gd fuel pins. The Kinf profiles show some differences for different arrangements of U/Gd fuel pins [46]. 3D maximum peak pin power value in the core calculation has very complex relationship,

83 and it is a function of Kinf distribution, inter-assembly leakages, and depletion characteristics of each fuel assembly.

On the other hand, RPD value of a fuel assembly means the normalized total power value of this fuel assembly in radial plane. The maximum radial power of a FA is represented as maximum RPD power in the depletion calculations. Reducing the fuel assembly Kinf reduces the FA power (RPD), but provides no information about the ratio of peak pin power to the FA power. It happens that a lower Kinf lowers the RPD but the peak pin power sometimes increases during depletion.

Figure 3-13 shows max. RPD power versus Kinf at BOC for U/Gd FA designs during GA runs. It can be seen from the figure that maximum RPD in the core depletion calculations increase with BOC Kinf value of the fresh fuel assembly designs. However, some designs may have different max. RPD power while they have the same BOC Kinf value, and the maximum RPD may not increase with less than 1 % increase in Kinf. The complexity of the relation between maximum peak pin power in core depletion calculations with the BOC Kinf values are shown in Figure 3-14. For the same BOC Kinf designs, maximum peak pin power could show big differences due to neighboring fuel assembly, inter assembly leakage effects, and depletion characteristics of the core. However, the maximum peak pin power is more sensitive to the increase in Kinf than the RPD.

84 Detailed reactor physics calculations were performed to evaluate the fuel assembly designs. The reactor physics calculation was a three step process, and it required a long CPU time as first to generate cross sections, then to create library, and third to perform core depletion calculations. The simple CASMO-4 calculations required more than 10 minutes CPU time due to Gd2O3 in the fuel. Cross section generation code (CASMO-4) performed the calculations in small burnup steps to reduce the self-shielding effect of Gd. This made the optimization calculations inefficient.

Therefore, Kinf filter value was defined to reduce the calculation time and to prevent performing unnecessary calculations. The Kinf filter value is shown in Figure 314, and it was chosen to guarantee that fuel assembly design having BOC Kinf value higher than the filter value would present a design, which results with the maximum peak pin power value violating the design constraint of 1.55 with some margin. The Kinf filter of 1.065 was defined as the cut off value and shown in Figure 3-14.

85

Max RPD Power vs Kinf at BOC for U/Gd Fuel Assembly designs (16 Gd pin with mixed concentration of 3 and 8 w/o)

1.8 1.75 1.7 Max RPD Power 1.65 1.6 1.55 1.5 1.45 1.4 1.35 1.055

1.06

1.065

1.07

1.075

1.08 Kinf

1.085

1.09

1.095

1.1

1.105

Figure 3-13 Max RPD power vs. Kinf at BOC for U/Gd FA designs
Max Peak Pin Power vs Kinf at BOC for U/Gd Fuel Assembly designs (16 Gd pin with mixed concentration of 3 and 8 w/o)

2 1.95 1.9 1.85 Max Peak Pin Power 1.8 1.75 1.7 1.65 1.6 1.55 1.5 1.45 1.055
kinf filter 1.065

1.06

1.065

1.07

1.075

1.08 Kinf

1.085

1.09

1.095

1.1

1.105

Figure 3-14 Max peak pin power vs. Kinf at BOC for U/Gd FA designs with Kinf filter

86
3.9.2 GA Algorithm with Kinf Filter Use

Figure 3-15 demonstrates the genetic algorithm flow diagram and its interaction with the lattice and reactor physics codes. The genetic algorithm code starts with a randomly generated initial population. Then, GA checks the individuals whether they are in the feasible solution space or not. If they are not in the feasible solution space, GA skips reactor physics calculations by assigning high negative fitness without using interface code and then continue with the next generation. If the individuals are in the feasible solution space as explained before, CASMO-4 [47] base depletion calculations are performed only at BOC to compare the individual BOC Kinf value with the Kinf filter value. The evaluation of individual, which its Kinf value less than the filter value, is possible with the reactor physics calculations starting from CASMO-4 image file generation for base depletion, core depletion library generation with TABLES code, and then the individual is placed into the TMI-1 reference core loading for core depletion calculations with SIMULATE-3 code. If the BOC Kinf value of the fuel assembly is larger than the Kinf filter value, GA codes assigns high negative penalty fitness for this individual without performing reactor physics calculations and then continue with the next population individual. The individuals, which are in the feasible solution space and their BOC Kinf values less than the filter value, are evaluated with the complete fuel management package.

The reactor physics calculations were performed with the help of interface code developed for this purpose. It helps to evaluate the solutions by providing data transfer between GA code and the reactor physics code. The main objective of this interface code

87 was to convert decoded GA population members into the form of fuel assembly loading (U enrichment, positions of UO2/Gd2O3 rods, and Gd2O3 concentration), and then, to perform reactor physics calculations for determining core depletion parameters such as maximum peak pin power, maximum RPD power, BOC, and EOC soluble boron

concentrations (ppm). Finally, core depletion parameters were used to evaluate the population members by calculating their fitness values in the GA code. Selection, crossover and mutation operators were used to generate next generation population members. The convergence criterion of the GA code was established by assigning a maximum generation number. The solutions were attained when the maximum generation number was reached. Data storage was performed during calculations and the good solutions were archived into a file to study the results later.

88

Fuel data: U enrichment No of Gd pins Gd2O3 concentration

i=1

Create Initial Population Randomly selected U/Gd pin positions

Feasible solution?
yes

No

CASMO-4 Base depletion at BOC CASMO-4 Base depletion Interface code


Create SIMULATE library with TABLES-3 code
yes

BOC Kinf less than Kinf filter


No

High Penalty and Fitness Definition to remove it from pool


Max peak pin power Max. RPD power SOB at BOC SOB at EOC

Core Depletion Calculations with SIMULATE-3

Genetic Operators to create next generation individuals

Fitness Calculation

i=i+1

no

Last Generati on?

yes

Determine Optimal U/Gd pin positions within a lattice

Figure 3-15 Coupled fuel lattice and core depletion calculations in GA optimization algorithm with BOC Kinf filter use

89
3.9.3 Objective Functions Definitions and Constraints Handling

The solution space was divided into two regions based on maximum peak pin power constraint (1.55) shown in Figure 3-16. Each region uses different objective functions. The good solutions (Region 1) do not violate the design constraint of maximum peak pin power, and the fitness of a solution is defined by its EOC soluble boron concentration (ppm). The objective is to maximize EOC soluble boron concentration while reducing the EOC residual binding to attain good designs. The region 2 represents the population members having maximum peak pin power value greater than the limit value of 1.55, and the fitness of an individual is a function of its maximum peak pin power value with the objective of minimizing maximum peak pin power value during the core depletion.

Solution moves from top to bottom

Region 2: Objective: Minimize peak pin power

Max. Peak Pin Power

Plimit= 1.55
Solution moves from left to right Region 1: Good Solutions Objective: Maximize EOC SOB

EOC SOB (ppm)

Figure 3-16 Solution space demonstration and its objective functions

90 The main objective is to move GA solutions from Region 2 to Region 1 (Good Solutions Region). The objective function of Region 2 is to minimize maximum peak pin power value during evaluation. The best solution in GA run improves in each generation, and later it reaches to maximum peak pin power limit value by moving from Region 2 to Region 1. Then, the objective function is changed to maximize EOC SOB concentration in the core depletion calculations by keeping the individuals below the maximum peak pin power constraint. With this approach the genetic algorithm fitness outputs are forced to move from top to bottom first by minimizing maximum peak pin power value in the Region 2, and then it follows the route from right to left in the solution space to maximize EOC soluble boron concentration shown in Figure 3-16.

Negative high fitness penalty (-2000) was assigned for the individuals which represent infeasible decision variables and violate the problem geometry. The second negative high fitness penalty (-1000) was assigned for the individuals having BOC Kinf value greater than the filter value. The main objective of these constraint handling techniques is to remove these individuals from the pool by reducing their probability of being selected for the next generation due to their low fitness value. The GA code skips reactor physics calculations without performing core depletion calculations, and then it continues with the creation and evaluation of next population members.

The fitness definitions given below were used for the solution space, where SOB means soluble boron concentration.
Region 1: Fitness=f1= w1EOCSOB (maximizing EOC SOB (ppm))

91
Region 2: Fitness=f2= w2maxpeakpinpower (minimizing maximum peak pin power)

where the weight coefficients are w1 = 1, w2=-100.


3.9.4 Results

Table 3-5 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, fitness value of the BP design, and EOC soluble boron concentration. Calculations were performed to determine the good designs for 16 U/Gd pins with mixed (8 and 3 w/o) Gd2O3 concentrations. The TMI-1 reference core loading pattern was used to evaluate the fuel assembly design in the core. Of importance is to use developed U/Gd fuel assembly design in fresh fuel assembly positions for the reference core loading pattern presented in Chapter 2 (See Figure 2-5). The newly developed good U/Gd fuel assembly design was placed into 6 fresh fuel assembly positions (H-09, H-13, K-10, K-12, L-13 and M-13) in the reference core loading pattern since all of them use the same number of U/Gd pins and Gd2O3 concentration. The best BP design was determined by looking at the less residual binding at the EOC.
Table 3-5 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with mixed concentration of Gd2O3)
Design No 1 2 3 4 5 6 7 8 9 10 U/Gd fuel pin positions 1 6 3 18 1 13 3 13 7 13 1 9 11 9 11 13 11 15 1 9 BOCKinf 18 6 6 6 6 11 6 6 5 15 1.056 1.057 1.058 1.058 1.058 1.06 1.062 1.057 1.059 1.057 MaxRPD 1.41 1.41 1.41 1.41 1.41 1.42 1.42 1.41 1.41 1.41 MaxPinPower 1.53 1.53 1.54 1.54 1.54 1.55 1.55 1.55 1.55 1.53 Fitness 144 144 143 143 143 143 143 143 143 142 EOCSOB 144 144 143 143 143 143 143 143 143 142

92 Figure 3-17 shows maximum peak pin power versus BOC Kinf value of the good designs. Designs 1 and 2 have the highest EOC soluble boron concentration. The solution space representation is shown in Figure 3-18, which has the graph of maximum peak pin power versus the EOC soluble boron concentration for the good designs.

Figure 3-19 shows good BP designs developed during GA run, and the Designs 1 and 2 lie on the right corner of the solution space represented in Region 1. Design 1 has the U/Gd fuel pin positions at locations 1, 6, and 18, and Design 2 has the U/Gd fuel pin positions at locations 3, 6, 18. The best designs and resulting reactor physics parameters are shown in Figures 3-20 and 3-21.
Max. Peak Pin Power vs BOC kinf during GA evaluation 16 U/Gd FA with mixed concentrations (3 and 8 w/o) in the core) 1.8

1.75 Max. Peak Pin Power

1.7

1.65

1.6

1.55

1.5 1.055

1.056

1.057

1.058

1.059

1.06 BOC kinf

1.061

1.062

1.063

1.064

1.065

Figure 3-17 Max. peak pin power vs. BOC Kinf value during GA run

93

Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration during GA evaluation (16 U/Gd FA with mixed concentrations (3 and 8 w/o) in the core)

1.8

1.75 Maximum Peak Pin Power

1.7

1.65
Region 2

Max. Peak Pin Power Constraint

1.6

1.55
Region 1

1.5 136

138

140

142

144

146

148

150

EOC SOB (ppm)

Figure 3-18 Max. peak pin power vs. EOC soluble boron concentration during GA run
Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration for 10 good design (16 U/Gd FA with mixed concentrations (3 and 8 w/o) in the core) 1.55

6 ,7 ,8 ,9

1.545 Max. Peak Pin Power

1 2 3 ,4 ,5 3 4 5 6 7 8 10 1 2 9 10

1.54

1.535

1.53

1.525 142

142.5

143

143.5

144

144.5

EOC SOB (ppm)

Figure 3-19 Max. peak pin power vs. EOC soluble boron concentration for good designs

94

Design 1

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 NoGd NoGd U/Gd Pin Kinf=1.056 MaxRPD=1.41 Max 3-D pin Power=1.53 EOC SOB=144 ppm

NoGd NoGd NoGd NoGd NoGd NoGd

Figure 3-20 Best BP design within good solutions

95

Design 2

Guide Tube NoGd NoGd 1 3 7 11 16 2 4 8 12 17 18 5 9 6 10 13 19 14 20 15 21 22 NoGd NoGd U/Gd Pin Kinf=1.057 MaxRPD=1.41 Max 3-D pin Power=1.53 EOC SOB=144 ppm

NoGd NoGd NoGd NoGd NoGd NoGd

Figure 3-21 Design No 2 within good solutions

Table 3-6 presents the history data during GA evaluation for the probability of having U/Gd fuel pin for total 22 fuel positions in the model. Table 3-6 summarizes the probability of occurrence in good designs. Some heuristics rules were summarized as follows:

96 Position 6 and 13 had the highest probability of occurrence (~ % 25) in the good U/Gd BP designs. This result showed that position 6 and 13 were the most effective ones in U/Gd BP designs. Positions 3 and 11 had the second highest probability of occurrence (~ 15 %) Good BP designs did not use lower right edge positions 20, 21 and 22. It was not allowed to have U/Gd fuel pins next to each other. Because most of the good designs included position 6, the neighboring fuel pin positions 5, 9 and 10 had very low probability to have U/Gd fuel pins. The probability of having such a BP design could be reduced by assigning low priority for this location in the future calculations. Priority table was prepared by ranking the probability of occurrence for each position given in Table 3-6. This may help to reduce the total calculation time significantly. The total evaluation time can be reduced by eliminating the possible U/Gd fuel pin configurations with low probability of occurrence for future calculations. These results present heuristic rules for BP design problem for future applications.

97
Table 3-6 Probability of having U/Gd fuel pin for each position in good designs Position No Number of Occurrence in good designs 12 0 39 0 1 76 3 0 16 2 42 0 69 1 4 0 0 2 0 0 0 0 267 Percentage Probability (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Total

4.5 0 14.6 0 0.4 28.5 1.1 0 6 0.7 15.7 0 25.8 0.4 1.5 0 0 0.7 0 0 0 0 100

98

3.10 Comparison of Methods


Three different techniques were studied on determining the good U/Gd FA design in the core. Table 3-7 presents the comparison of the available design data and the different optimal design data. Table 3-7 clearly shows the relation between Kinf, maximum peak pin power and EOC soluble boron. The best design, which uses Kinf filter value in the calculations, has the highest EOC soluble boron value and it does not violate the max. peak pin power design constraint of 1.55. The optimal positions of U/Gd fuel pins were determined to be at position 1, 6* and 18* in the octant PWR fuel assembly model (* Fixed at 8 w/o Gd2O3 concentration) for 16 U/Gd fuel pins with mixed Gd2O3 concentration of 3 and 8 w/o. The reference core loading with fresh fuel positions should be known to evaluate the performance of the design in the core. The important conclusion is that the optimal designs are different for different core loading pattern to reduce residual binding.
Table 3-7 Comparison of available FA designs and optimal designs

Data

Reference FA design

Minimizing assembly internal peak power 118 1.44 1.65 1.065 14, 15*,3*

Minimizing BOC Kinf 100 1.43 1.56 1.051 15, 6*, 18*

EOC SOB (ppm) 101 Max RPD 1.39 Max 1.54 3-D Pin Power Kinf 1.059 Optimal U/Gd pin 3*,6*,19 positions * (represent 8 w/o Gd2O3)

Use Kinf filter and coupled core depletion calculations 144 1.41 1.53 1.056 1, 6*, 18*

Figure 3-22 shows the comparison of U/Gd pin arrangements for different objective functions and methods described in this study. The important conclusion of

99 Figure 3-22 is that U/Gd pin arrangements change with the objective function selected. The internal power peaking can be minimized by arranging U/Gd fuel pins, but it behaves different in the core calculations. The peak pin power is a function of arrangements within an assembly, Kinf distributions, and core depletion characteristics. One should calculate the maximum peak pin power during core depletion and residual binding at EOC to determine the optimal arrangement of the U/Gd pins. In that case, core reload pattern and fresh fuel assembly positions in the core should be known beforehand to determine maximum peak pin power values during core depletion. Hot fresh fuel assemblies should be replaced with the optimal U/Gd FA designs to evaluate the core performance.

100

Figure 3-22 Comparison of U/Gd fuel pin arrangements for different objective functions

101

3.11 The U/Gd Fuel Assembly Design with 20 U/Gd Fuel Pins
Table 3-8 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, fitness value of the BP design, and EOC soluble boron concentration. Calculations were performed to determine the good designs for 20 U/Gd pins with mixed Gd2O3 concentrations. The TMI-1 reference core loading pattern was used to evaluate the fuel assembly design in the core.
Table 3-8 U/Gd fuel pin configurations and core depletion parameters for good solutions (20 U/Gd fuel pins with mixed concentration of Gd2O3)

Design No 1 2 3 4 5 6 7 8 9 10

U/Gd fuel pin positions 4 12 20 11 4 15 8 11 2 7 19 15 8 16 5 11 14 3 19 15

BOCKinf MaxRPD MaxPinPower Fitness EOCSOB 6 2 5 2 5 2 4 8 5 2 1.042 1.041 1.048 1.052 1.054 1.040 1.052 1.050 1.042 1.04 1.43 1.43 1.43 1.43 1.44 1.43 1.43 1.43 1.43 1.43 1.54 1.54 1.54 1.55 1.54 1.54 1.54 1.54 1.54 1.54 139 139 139 139 138 138 138 138 138 138 139 139 139 139 138 138 138 138 138 138

Table 3-9 presents the history data during GA evaluation for the probability of having U/Gd fuel pins for total 22 fuel positions in the model. Table 3-9 summarizes the probability of occurrence in good designs. Positions 4, 11 and 20 had the highest probability of occurrence. This result showed that these positions in the model were the most effective ones for this BP design.

102
Table 3-9 Probability of having U/Gd fuel pin for each position in good designs Position No Number of Occurrence in good designs 15 42 40 91 35 7 20 36 6 11 68 20 3 7 20 6 8 18 15 58 6 2 267 Percentage Probability (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Total

2.8 7.9 7.5 17 6.6 1.3 3.7 6.7 1.1 2.1 12.7 3.7 0.6 1.3 3.7 1.1 1.5 3.4 2.8 10.9 1.1 0.4 100

103

3.12 U/Gd Fuel Assembly Designs


Table 3-10 shows the all of the feasible U/Gd BP designs including number of U/Gd fuel pins, Gd2O3 concentration, and their segment name data used during cross section generation. The number of U/Gd fuel pins represented with 17 and 21 shows the mixed Gd2O3 concentrations use (8 pins fixed at 8/wo, others varied) for 16 and 20 U/Gd fuel pins respectively within octant symmetry. These representations were used to separate different fuel types during calculations. Segment name of a BP design represents the number of U/Gd fuel pins and Gd2O3 concentration. For example, 16 U/Gd fuel pins with 3 w/o Gd2O3 was represented by TYPE16GD03 in cross section generation and reactor physics calculations and 16 U/Gd fuel pins with mixed concentrations of 3 w/o and 8 w/o Gd2O3 was represented by TYPE17GD03.

104
Table 3-10 U/Gd BP designs (No of U/Gd fuel pins, Gd2O3 w/o, Segment Name)
Case No No of U/Gd pin Gd2O3 (w/o) Segment Name 0 0 0 TYPE00GD00 1 4 2 TYPE04GD02 2 4 3 TYPE04GD03 3 4 4 TYPE04GD04 4 4 5 TYPE04GD05 5 4 6 TYPE04GD06 6 4 7 TYPE04GD07 7 4 8 TYPE04GD08 8 8 2 TYPE08GD02 9 8 3 TYPE08GD03 10 8 4 TYPE08GD04 11 8 5 TYPE08GD05 12 8 6 TYPE08GD06 13 8 7 TYPE08GD07 14 8 8 TYPE08GD08 15 12 2 TYPE12GD02 16 12 3 TYPE12GD03 17 12 4 TYPE12GD04 18 12 5 TYPE12GD05 19 12 6 TYPE12GD06 20 12 7 TYPE12GD07 21 12 8 TYPE12GD08 22 16 2 TYPE16GD02 23 16 3 TYPE16GD03 24 16 4 TYPE16GD04 25 16 5 TYPE16GD05 26 16 6 TYPE16GD06 27 16 7 TYPE16GD07 28 16 8 TYPE16GD08 29 17* 2 TYPE17GD02 30 17 3 TYPE17GD03 31 17 4 TYPE17GD04 32 17 5 TYPE17GD05 33 17 6 TYPE17GD06 34 17 7 TYPE17GD07 35 17 8 TYPE17GD08 36 20 2 TYPE20GD02 37 20 3 TYPE20GD03 38 20 4 TYPE20GD04 39 20 5 TYPE20GD05 40 20 6 TYPE20GD06 41 20 7 TYPE20GD07 42 20 8 TYPE20GD08 43 21* 2 TYPE21GD02 44 21 3 TYPE21GD03 45 21 4 TYPE21GD04 46 21 5 TYPE21GD05 47 21 6 TYPE21GD06 48 21 7 TYPE21GD07 49 21 8 TYPE21GD08 * : 17 and 21 shows the mixed Gd2O3 concentrations use (8 pins fixed at 8/wo, others variable)

105

3.13 Automation of CASMO Image File Generation


TABLES-3 [48] is a data processing code that links CASMO to SIMULATE. The code reads the CASMO card image files and produces a master binary library in the format required by SIMULATE. The code was used to add or delete data from the library. TABLES-3 produces one, two or three dimensional tables of partials for each of the types of data with the exception of pin power reconstruction data, kinetics data, and isotopic data. Functional capability in TABLES is generalized to allow the user to define the partials based upon his modeling needs. Appendix 3 presents the sample TABLES-3 code input deck for BP library generation. Figure 3-23 shows the procedures to generate the assembly cross sections, formatted library file, and cycle depletion with CASMO3/TABLES-3/SIMULATE-3 package.

106
Unformatted BA, gamma, neutron lib

MICBURN

Formatted BA library

CASLIB

CASMO -3

Formatted Gamma, neutron library

LIG- gamma LIN- neutron PBA-absorber

Unformatted X-sec history And instantaneous dependencies

Core Depletion

SIMULATE

Formatted X-sec history to be used in simulate

TABLES - 3

Figure 3-23 CASMO-3/TABLES-3/SIMULATE-3 fuel management code package

CASMO-4 does not need to use MICBURN gamma and neutron libraries for gadolinium burnable absorbers, and performs gadolinium depletion in a small burnup steps. In this study, CASMO-4 code was used to generate the cross section library of the burnable absorbers.

107 A user-friendly program was developed in this study, which automatically generates all necessary CASMO-4 input decks, and performs the calculations by using interfaces. The optimization library contains the image files of the all feasible U/Gd and Boron fueled fuel assembly designs. It is a huge job, and it requires very long CPU time.

Optimization process gained flexibility to vary the Gd2O3 concentration from 2 to 8 w/o by developing this program. Segment names used in CASMO-3 and SIMULATE-3 input deck for the fuel assemblies represent the Gd2O3 concentration and the type of the fuel assemblies. CASMO-3 runs contain the possibilities of 3 U enrichments (4.0, 4.5 and 5.0), 7 Gd concentrations (2, 3, 4, 5, 6, 7, and 8), and 7 FA types (0, 4, 8, 16, 16 mixed, 20 and 20 mixed U/Gd pins in the assembly). There are also base, low moderator temperature, high boron concentration, and branch CASMO-3 cases for each run. One should also consider creating the image files for blanket, reflector, and used fuel with boron cases. Each segment name given in Table 3-10 needs to complete at least 12 CASMO-runs for creating cross section tables including 3 fuel uranium enrichment (4.0, 4.5, and 5.0), and base, low moderator temperature, high boron concentration, and branch cases.

The total CASMO-4 calculation requires completing 4912=588 runs to create huge burnable absorber library. This process is the basis of the optimization calculations, and it requires huge man-power. Therefore, the automation of the image file generation process is very important, and it presents great benefits to the user by savings the

108 enormous amount of time and complexity. It is very simple to run and to create all of the image files by using the input parameters of U enrichment, number of U/Gd fuel pins, Gd2O3 concentrations, and U/Gd fuel pin positions within an octant fuel assembly.

The flow chart of CASMO-4 image file generator program is given in Figure 324. As the Gd2O3 is used in the fuel matrix mixed together with uranium, the different number of UO2/Gd2O3 pins and concentration of Gd2O3 fuel represent different fuel types in the lattice and reactor physics code structures such as CASMO-3 and SIMULATE-3, and they have to be named with different segment names each representing fuel type characteristics such as uranium enrichment, number of UO2/Gd2O3 pins and Gd2O3 concentrations. The program assigns a segment name according to the fuel type given in Table 3-10, and it performs the necessary CASMO calculations and locates the image files for the library use later. The program continues until covering all of the possible image files.

109

START i=1

Read Fuel Data Fuel Enrichment No Of U/Gd fuel pins and positions (Type) Gd2O3 concentration Gd isotopes weight fractions

Assign Segment Name

Increment the fuel data For next run

Create CASMO-4 files (base, lowtmo, highboron and branch cases)

i=i+1

Perform CASMO-4 Calculations

No

Cover all Image files?

Locate image files to directory destination for the library use

yes

END

Figure 3-24 Flow Diagram of CASMO-4 image file generator

110

3.14 Conclusions
The important conclusion is that optimal U/Gd fuel pin arrangements show differences with the objective function selected as shown in section 3.10. The assembly internal power peaking can be minimized by arranging U/Gd fuel pins, but it behaves different in the core calculations and causes the peak pin power increase at some other locations in the core.

Hence, coupled lattice and core depletion calculations should be performed to find good U/Gd fuel pin configurations in a fuel assembly design process. A GA methodology was developed for the solution of this problem. It was shown that defined BOC Kinf limit value can be used as a filter to reduce the total run time during optimization calculations. The studies showed that BOC Kinf of 1.065 was a good filter to eliminate the evaluation of bad solutions in GA calculations.

CHAPTER 4 OPTIMIZED ENRICHED BURNABLE POISON DESIGN


4.1 Introduction
Seven naturally occurring Gd-152, Gd-154, Gd-155, Gd-156, Gd-157, Gd-158, and Gd-160 isotopes were identified as gadolinium constituents. The two Gd isotopes, Gd-155 and Gd-157, have 14.51 and 15.68 % natural abundances and 58,000 x 10-24 and 240,000 x 10-24 cm2 cross-sections, respectively. Gd-152, Gd-159 and Gd-160 isotopes have very low absorption cross-sections in thermal region compared to the others. Gd152 and Gd-159 have also very low natural abundance ratio and their contribution can be negligible. Absorption cross section of each isotope over energy spectrum is given in the Figure 4-1 [49].

Sensitivity analysis was performed on the effect of using the enriched Gd2O3 fuel pins for neutronic parameters of a FA such as Kinf, peak pin power, RPD power and the residual binding at EOC. This part of the research concentrated on the reduction of the inherent penalties of the conventional gadolinia fuel assembly designs, i.e. reduced average fuel assembly enrichment and heavy metal content, and the residual reactivity binding. The negative residual reactivity associates with the presence of the remaining minor gadolinium isotopes (Gd-152, Gd-154, Gd-156, Gd-158 and Gd-160) that are not significant neutron absorbers, and a corresponding decrease in the mass of uranium by

112 simple displacement. Because, Gadolinia displaces uranium in the fuel matrix, it causes a reduced heavy metal mass. The maximum fuel temperature design criterion is satisfied by reducing the enrichment in UO2/Gd2O3 fuel rods due to lower heat conductivity of the UO2/Gd2O3 FA in the conventional designs [27, 28]. The residual negative reactivity remains after the Gd-155 and Gd-157 depletion at EOC due to presence of gadolinium isotopes Gd-152, Gd-154, Gd-156, Gd-158 and Gd-160.

Figure 4-1 Neutron capture cross section of Gd isotopes versus energy

Uranium enrichment penalty was defined for UO2/Gd2O3 fuel rods due to thermal conductivity limitation for highly concentrated natural Gd enriched UO2/Gd2O3 fuel. This penalty was addressed by using 0.05 w/o less uranium enrichment per Gd2O3 concentration in conventional natural Gd enriched design. Optimized enriched Gd

113 designs removed this penalty and used the same uranium enrichment for all fuel rods with low concentrated (2~3 w/o) UO2/Gd2O3 fuels. The reason is due to recent measurement results of thermal conductivity values of both types of fuel [29]. For example, the average FA enrichment was 4.5 w/o without U-235 enrichment penalty, and it was 4.473 w/o with the penalty due to use of high Gd2O3 concentration.

4.2 The Effect of Gd Absorber on Neutron Spectra and Pu Production


Since Gd absorbers have 1/v cross sections as shown in Figure 4-1, it is expected to deplete low energy end of spectrum due to absorption. Therefore, it causes the shift of Maxwellian spectrum to higher energies. Pu production rate depends on two parameters such as Pu production rate due to U-238 resonance absorption ratio and the absorption of neutrons by Pu. Figure 4-2 shows the neutron capture cross sections of Pu-239 and U-238 over energy spectrum. It can be seen from the figure that U-238 capture cross section for neutrons is constant in the thermal energy region. The relation between a fuel assembly Kinf parameter and Pu production rate is explained as followings for fuel with Gd and fuel with no Gd: Neutron absorption reaction rate for Pu = Ra =a Gd < NoGd , RGd < RNoGd, and it results with an increase in the production ratio of Pu. The reason is due to the reduction in Pu absorption (loss) rate for fuel with Gd compared to fuel without Gd. Kinf of a FA depends on two reaction rates, which are U consumption rate minus Pu production rate. Kinf= Kinf (U consumption rate, Pu production rate)

114 Pu production rate dominates the U consumption rate until Gd burnout point where its Kinf value reaches the peak value shown in the following sections for UO2/Gd2O3 fuels.

Figure 4-2 Neutron capture cross sections of Pu-239 and U-238 vs. energy [49]

4.3 Natural and Enriched Gd BP Design


A sample 1515 PWR fuel assembly was modeled by defining all of the Gd isotopes in the fuel material card with HELIOS 2-D [50] transport theory code based on the method of collision probabilities. The code is capable of explicitly modeling heterogeneous assembly lattices with integral BPs [13]. Sample U-235 enrichment was 4.5 w/o for UO2 pins, and it was 4.38 for enriched UO2/Gd2O3 pin by defining 0.12 w/o penalty.

115 The deterministic method results presented in Chapter 2 developed a BP loading pattern containing 16 U/Gd fuel pins with high (7-8 w/o) Gd2O3 concentration in some of fresh fuel locations for the reference TMI-1 core loading. Therefore, the use of 16 UO2/Gd2O3 pins in the assembly was selected as reference case for this analysis.

Figure 4-3 shows the Kinf values as a function of burnup for different Gd2O3 concentrations in the reference (1515) TMI-1 PWR fuel assembly with 4.5 w/o U-235 enriched and 16 U/Gd pins. Kinf value decreases by increasing Gd2O3 concentration at BOC and different Gd2O3 concentrations presents different burnout rates. The presence of UO2-Gd2O3 rods during the depletion hardens the neutron spectrum, resulting in greater production of fissile plutonium isotopes and reduced
235

U depletion. As a result, the

reactivity of an assembly depleted with UO2-Gd2O3 rods may be higher than that of an assembly depleted without BAs rods. Gd2O3 concentration change affect Gd burnout rate and depletion kinetics. However, Kinf values at BOC shows slight differences, and more important one is that as Gd2O3 concentration increases the time for Gd burnout also increases as shown in Figure 4-3.

116

1.4

Kinf profiles for different Gd2O3 concentration of natural enriched 15x15 U/Gd FA (U-235=4.5 w/o,16 Gd pin) No Gd 2 w/o Conc 3 w/o conc 4 w/o conc 5 w/o conc 6 w/o conc 7 w/o conc 8 w/o conc

1.3

1.2 Kinf

1.1

0.9

0.8 0 10000 20000 30000 Burnup (MWD/MTU) 40000 50000 60000

Figure 4-3 Comparison of Kinf curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel

A calculation was also performed in which each of the UO2-Gd2O3 fuel rods (with U-235 penalty and lower fuel enrichment) was replaced by an equivalent enriched (4.5 w/o U-235 enriched) UO2 fuel rods. The differences in the residual binding (delta k values) between different Gd2O3 amounts are shown in Figure 4-4. As expected, the negative delta k values in the figure indicate that the gadolinia-bearing fuel is less reactive than the fuel without gadolinia, which has zero residual binding. The fuel assembly becomes less reactive when the weight percent of Gd2O3 is increased. Increased gadolinia loading leads to increased concentrations of the Gd isotopes, including the isotopes that are not significant neutron absorbers, and a corresponding decrease in the mass of uranium by simple displacement. The negative residual reactivity increases with gadolinia loading even more than compensates any positive reactivity effect due to harder

117 neutron spectrum linked to the presence of the gadolinia. Gd2O3 concentration affects the residual binding at the EOC, and high concentration results with high residual binding due to higher fraction of Gd low absorbing isotopes in the fuel and the late Gd burnout during the depletion.
Residual Delta k profiles for different Gd2O3 concentration for natural enriched Gd FA (U=4.5 w/o, 16 Gd pin) 0 -0.02 Delta k = (kinf(U/Gd FA)-kinf(U FA) -0.04 -0.06 -0.08 -0.1 -0.12 -0.14 -0.16 -0.18 -0.2 0 10000 20000 30000 Burnup (MWd/MTU) 40000 50000 60000 2 w/o Conc 3 w/o conc 4 w/o conc 5 w/o conc 6 w/o conc 7 w/o conc 8 w/o conc

Figure 4-4 Comparison of residual Kinf curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel

Figure 4-5 shows the assembly peak pin power curves as a function of burnup for a given fuel assembly for 16 U/Gd fuel pins in the assembly, which has uranium enrichment of 4.5 w/o and different Gd2O3 w/o concentration. This analysis shows that the assembly peak pin power increases when the Gd2O3 w/o is increased. There is a 14 % increase in assembly peak pin power at BOC for 8 w/o Gd2O3 while the increase is 12 % for 2 w/o Gd2O3 concentration. It is important to note that the fuel assembly Gd2O3 concentration should be kept as low as possible to reduce assembly internal power

118 peaking during the depletion and to reduce the residual binding at EOC. Assembly peak pin power curves arrive to the region, which is less than 1 % from without Gd fuel after Gd burnout shown in Figure 4-5.
Peak Pin Power profiles for different concentration of natural enriched U/Gd FA (U-235=4.5 w/o, 16 Gd pin)
1.16

1.14

1.12

Peak Pin Power

1.1

1.08

No Gd 2 w/o Conc 3 w/o conc 4 w/o conc 5 w/o conc 6 w/o conc 7 w/o conc 8 w/o conc

1.06 1.04

1.02 1 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000 60000 Burnup (MWd/MTU)

Figure 4-5 Comparison of assembly peak pin power curves for different Gd2O3 w/o concentration with natural Gd enriched UO2-Gd2O3 fuel

Table 4-1 and 4.2 summarize the distribution of isotopic weight fractions of Gd isotopes considered in this study for enriched Gd-157 and Gd-155 BP design.
Table 4-1 Gd isotopes weight fractions for different enrichment of Gd-157 isotope
Isotope Gd-152 Gd-154 Gd-155 Gd-156 Gd-157 Gd-158 Gd-160 Natural Gd-157 0.2 2.1 14.51 20.3 15.68 24.98 21.86 50 w/o Gd-157 0 2 14 14.9 50 19.1 0 Isotope Weight Fractions (w/o) 60 w/o 70 w/o 80 w/o GdGd-157 Gd-157 157 0 0 0 1 1 1 10 5 7 11.9 6.9 9.9 60 80 70 17.1 7.1 12.1 0 0 0 100 w/o Gd-157 0 0 0 0 100 0 0

119
Table 4-2 Gd isotopes weight fractions for different enrichment of Gd-155 isotope

Isotope Gd-152 Gd-154 Gd-155 Gd-156 Gd-157 Gd-158 Gd-160

50 w/o Gd-155 0 0 50 0 50 0 0

Isotope Weight Fractions (w/o) 60 w/o 70 w/o Gd-155 Gd-155 0 0 0 0 60 70 0 0 40 30 0 0 0 0

100 w/o Gd-155 0 0 100 0 0 0 0

4.4 Modeling Enriched Gadolinium and Genetic Algorithm Model


CASMO-3 was not able to deplete enriched Gd2O3 fuel. Both Gd-155 and Gd-157 isotopes were defined as fission products, and they were only produced by depletion of the other isotopes in the decay chain. MICBURN-3 [51] studies were also performed to generate the burnable absorber library of UO2/Gd2O3 fuel. However, MICBURN-3 and CASMO-3 have natural compositions of Gd isotopes defined as hard-wired option in the code structure. Therefore, all of the optimization studies were performed with CASMO-4 Code [47], which was able to deplete enriched Gd isotopes. The same genetic algorithm code and fuel assembly model as explained in Chapter 3 were used to design the enriched Gd BPs.

4.5 Optimization of Enriched Gadolinium Use in the Fresh Fuel Assemblies


Genetic algorithm runs for enriched Gd designs are summarized in this section. The sensitivity studies focused on low concentrated enriched Gd-155 and Gd-157 with 16 and 20 U/Gd fuel pins designs due to their frequent use in the final core loading pattern.

120 The TMI-1 reference core loading pattern defined in Chapter 2 (see Figure 2-5) was used to evaluate the fuel assembly design in the core.
4.5.1 100 w/o Gd-155 enriched with 2 w/o Gd2O3 (20 U/Gd fuel pins)

This section presents the results of genetic algorithm run for 4.5 w/o U-235, 100 w/o Gd-155 enriched with 2 w/o Gd2O3 concentrated fuel design. Figure 4-6 shows maximum peak pin power versus EOC soluble boron concentration (ppm) during GA evaluation. Residual binding was reduced by increasing EOC soluble boron concentration from 139 ppm to 170 ppm for the enriched Gd-155 and low concentrated U/Gd BP FA designs compared to the natural Gd enriched design presented in Chapter 3. The difference of 31 ppm is due to residual binding savings.

Table 4-3 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, GA code fitness value of the BP design, and EOC soluble boron concentrations. Calculations were performed to determine the good designs for 20 U/Gd pins within the fuel assembly design having 100 w/o Gd-155 enrichment with 2 w/o Gd2O3 concentration.

121

Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration during GA evaluation (20 U/Gd FA with 100 w/o Gd-155 and 2 w/o Gd2O3)

1.85 1.8 Maximum Peak Pin Power 1.75 1.7 1.65 1.6 1.55
Region 1 Region 2 Max. Peak Pin Power Constraint

1.5 160

162

164

166 EOC SOB (ppm)

168

170

172

Figure 4-6 Max. peak pin power vs. EOC soluble boron concentration during GA run Table 4-3 U/Gd fuel pin configurations and core depletion parameters for good solutions (20 U/Gd fuel pins with 100 w/o enriched Gd-155 and 2 w/o concentration of Gd2O3)

Design No 1 2 3 4 5 6 7 8 9 10

U/Gd fuel pin positions 19 19 19 19 18 17 18 19 18 7 15 2 15 4 12 1 8 1 15 2 15 4 15 8 11 4 15 4 14 17

BOC Kinf 1.048 1.046 1.046 1.048 1.046 1.048 1.046 1.048 1.044 1.049

Max RPD 1.48 1.48 1.48 1.48 1.48 1.48 1.48 1.48 1.48 1.48

Max. Peak Pin Power 1.55 1.55 1.55 1.55 1.55 1.55 1.55 1.55 1.55 1.55

Fitness 170 170 170 170 170 170 170 170 170 169

EOCSOB (ppm) 170 170 170 170 170 170 170 170 170 169

122
4.5.2 100 w/o Gd-155 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)

This section summarizes the results of GA run for fuel assembly design containing 16 U/Gd fuel pins. The BP design has 100 w/o Gd-155 enrichment with 2 w/o Gd2O3 concentration. Figure 4-7 shows maximum peak pin power vs. EOC soluble boron (ppm) during GA evaluation. EOC soluble boron was increased from 144 ppm to 172 ppm for the 100 w/o enriched Gd-155 and low concentrated U/Gd BP FA designs compared to the natural Gd enriched design presented in Chapter 3. The difference of 28 ppm is due to residual binding savings of the enriched Gd-155 design.

Table 4-4 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, GA code fitness value of the BP design, and EOC soluble boron concentration. The GA calculations were performed to determine the optimal designs.

123

Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration during GA evaluation (16 U/Gd FA with 100 w/o Gd-155 and 2 w/o Gd2O3)

1.80

1.75 Maximum Peak Pin Power

1.70
Region 2

1.65

1.60

Max. Peak Pin Power Constraint

1.55
Region 1

1.50 165

166

167

168

169

170

171

172

173

174

175

176

177

EOC SOB (ppm)

Figure 4-7 Max. peak pin power vs. EOC soluble boron concentration during GA run Table 4-4 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 100 w/o enriched Gd-155 and 2 w/o concentration of Gd2O3)

Design No 1 2 3 4 5

U/Gd fuel pin positions 1 13 6 6 13 1 3 13 6 6 13 3 7 13 1

BOC Kinf 1.068 1.068 1.068 1.068 1.068

Max RPD 1.43 1.43 1.43 1.43 1.43

Max. Peak Pin Power 1.55 1.55 1.55 1.55 1.55

Fitness 172 172 171 171 171

EOCSOB (ppm) 172 172 171 171 171

124
4.5.3 70 w/o Gd-157 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)

This section summarizes the results of GA run for fuel assembly design containing 16 U/Gd fuel pins. The BP design has 70 w/o Gd-157 enrichment with 2 w/o Gd2O3 concentration. Figure 4-8 shows maximum peak pin power vs. EOC soluble boron (ppm) during GA evaluation. EOC soluble boron was increased from 144 ppm to 176 ppm for the 70 w/o enriched Gd-157 and low concentrated U/Gd BP FA designs compared to the natural Gd enriched design presented in Chapter 3. The difference of 32 ppm is due to residual binding savings.

Table 4-5 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, GA code fitness value of the BP design, and EOC soluble boron concentrations. The GA calculations were performed to determine the good designs for this case. The TMI-1 reference core loading pattern was used to evaluate the fuel assembly design in the core.
Table 4-5 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 70 w/o enriched Gd-157 and 2 w/o concentration of Gd2O3)

Design No 1 2 3 4

U/Gd fuel pin positions 10 11 3 10 11 1 13 3 7 13 3 7

BOC Kinf 1.068 1.066 1.068 1.068

Max RPD 1.48 1.49 1.47 1.47

Max. Peak Pin Power 1.55 1.55 1.55 1.55

Fitness 176 176 175 175

EOCSOB (ppm) 176 176 175 175

125

Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration during GA evaluation (16 U/Gd FA with 70 w/o Gd-157 and 2 w/o Gd2O3)

1.8

1.75 Maximum Peak Pin Power

1.7
Region 2

1.65

1.6

Max. Peak Pin Power Constraint

1.55
Region 1

1.5 173

174

175 EOC SOB (ppm)

176

177

Figure 4-8 Max. peak pin power vs. EOC soluble boron concentration during GA run 4.5.4 100 w/o Gd-157 enriched with 2 w/o Gd2O3 (16 U/Gd fuel pins)

This section summarizes the results of GA run for fuel assembly design containing 16 U/Gd fuel pins. The BP design has 100 w/o Gd-157 enrichment with 2 w/o Gd2O3 concentration. Figure 4-9 shows maximum peak pin power vs. EOC soluble boron (ppm) during GA evaluation. EOC soluble boron was increased from 144 ppm to 188 ppm for the 100 w/o enriched Gd-157 and low concentrated U/Gd BP FA designs compared to the natural Gd enriched design presented in Chapter 3. The difference of 44 ppm is due to residual binding savings and it is the greatest savings amount between enriched designs. There is a small fraction of residual isotopes due to Gd-157 absorption at EOC than what it is in enriched Gd-155 design.

126 Table 4-6 presents the good BP designs with U/Gd fuel pin positions within an octant fuel assembly symmetry, and the core depletion parameters such as BOC Kinf value of the design, maximum RPD power, and maximum peak pin power, GA code fitness value of the BP design, and EOC soluble boron concentrations. Calculations were performed to determine the good designs for 16 U/Gd pins with 100 w/o Gd-157 and 2 w/o Gd2O3 concentration. The TMI-1 reference core loading pattern was used to evaluate the fuel assembly design in the core.
Table 4-6 U/Gd fuel pin configurations and core depletion parameters for good solutions (16 U/Gd fuel pins with 100 w/o enriched Gd-157 and 2 w/o concentration of Gd2O3)

Design No 1 2 3 4 5

U/Gd fuel pin positions 18 9 1 18 18 1 6 1.06 1.06 1.06 1.06 1.06 1.47 1.47 1.47 1.47 1.47 15 16 16 13 7 7 6 6 BOC Kinf Max RPD

Max. Peak Pin Power 1.55 1.55 1.55 1.55 1.55 188 187 187 187 187 Fitness

EOCSOB (ppm) 188 187 187 187 187

127

Max Peak Pin Power in core depletion vs EOC SOB (ppm) concentration during GA evaluation (16 U/Gd FA with 100 w/o Gd-157 and 2 w/o Gd2O3)

1.8

1.75 Maximum Peak Pin Power

1.7
Region 2

1.65
Max. Peak Pin Power Constraint

1.6

1.55
Region 1

1.5 165

170

175

180

185

190

EOC SOB (ppm)

Figure 4-9 Max. peak pin power vs. EOC soluble boron concentration during GA run

4.5.5 Summary of Results

This section summarizes the results of GA calculations and compares the resulting residual binding savings achieved by the enriched Gd designs as compared with the conventional BP design. The comparison was performed by considering the natural Gd enriched design (16 U/Gd fuel pins) presented in Chapter 3 as a basis. Table 4.7 presents the comparison of EOC soluble boron data and residual binding savings in ppm for enriched Gd designs for 16 U/Gd fuel pins with 2 w/o Gd2O3 concentration. The biggest residual binding saving were achieved in 100 w/o Gd-157 designs compared to the others. Gd-157 absorption produces Gd-158 isotope. Due to its small absorption cross section, Gd-158 has minimal contribution to the residual negative reactivity. Therefore, the 100 w/o Gd-157 designs provided 44 ppm savings as compared to the natural U/Gd

128 designs, which is larger than the other designs. On the other hand, 100 w/o Gd-155 designs presents extra 28 ppm residual binding savings as compared to the natural U/Gd fuel designs.
Table 4-7 Comparison of EOC soluble boron data and residual binding savings in ppm

Design Natural Gd and 16 U/Gd fuel pins 100 w/o Gd-155 and 16 U/Gd fuel pins 70 w/o Gd-157 and 16 U/Gd fuel pins 100 w/o Gd-157 and 16 U/Gd fuel pins

EOC SOB (ppm) 144 172 176 188

Residual Binding Savings (Delta ppm) 28 32 44

4.6 Enriched Gd-155 BP Design Characteristics


Sensitivity study was performed on enriched Gd-155 BP designs with enrichment ratios of 50, 60 and 70 w/o Gd-155 given in Table 4.2. The BP designs contain 16 U/Gd fuel rods with 2 w/o Gd2O3 concentration. Figures 4-10, 4-11 and 4-12 show the comparison of Kinf, residual delta k and peak pin power during the depletion for natural Gd enriched, 70 w/o enriched Gd-157, 50 w/o, 60 w/o and 70 w/o enriched Gd-155 designs. There is an assembly internal peak power increase, which starts earlier for 70 w/o enriched Gd-157 fuels due to rapid Gd burnout. Assembly internal peak power increase starts to occur late in the mid cycle for enriched Gd-155. Gd-155 has absorption cross section, which is approximately 1/6 of the absorption cross section of Gd-157 isotope. For natural design assembly internal peak power continues to drop for higher burnups. Figure 4-10 shows the effect of Gd-155 and Gd-157 enrichment on Kinf. Enriched Gd-155 has slower and smoother Gd depletion than the enriched Gd-157. From

129 Figure 4-11, it is clear that the slope of the delta k graph, which represents Gd depletion rate is much bigger for enriched Gd-157 than for the others. All of the enriched Gd-155 designs have similar depletion ratio and residual binding.

Kinf Comparison of Gd-157 and Gd-155 enrichments (U-235=4.5 w/o, 16 Gd pin)


1.4 No Gd Natural 7 w/o Gd2O3 1.3 Natural 8 w/o Gd2O3 Gd-157=70, 2 w/o Gd2O3 Gd-155=50,2 w/o Gd2O3 1.2 Gd-155=60,2 w/o Gd2O3 Gd-155=70,2 w/o Gd2O3 Kinf

1.1

0.9

0.8 0 10000 20000 30000 Burnup (MWd/MTU) 40000 50000 60000

Figure 4-10 Kinf Comparison of Gd-155 and Gd-157 enriched U/Gd fuels

130

Residual Delta k profiles for different weight fractions of Gd-157 and Gd-155 (U-235=4.5 w/o, 16 Gd pin)
0.02 0 -0.02 -0.04 -0.06 Delta k -0.08 -0.1
Natural 7 w/o Gd2O3

-0.12
Natural 8 w/o Gd2O3

-0.14 -0.16 -0.18 -0.2 0 10000 20000 30000 Burnup (MWd/MTU) 40000

Gd-157=70, 2 w/o Gd2O3 Gd-155=50,2 w/o Gd2O3 Gd-155=60,2 w/o Gd2O3 Gd-155=70,2 w/o Gd2O3

50000

60000

Figure 4-11 Residual binding (delta k) profiles of Gd-155 and Gd-157 enriched U/Gd fuels
Peak Pin Power profiles for different weight percent of Gd-155 and Gd-157 (U-235=4.5 w/o, 16 Gd pin) 1.16
No Gd

1.14 1.12 Peak Pin Power 1.1 1.08 1.06 1.04 1.02 1 0 10000 20000 30000 Burnup (GWd/MTU) 40000

Natural 7 w/o Gd2O3 Natural 8 w/o Gd2O3 Gd-157=70, 2 w/o Gd2O3 Gd-155=50,2 w/o Gd2O3 Gd-155=60,2 w/o Gd2O3 Gd-155=70,2 w/o Gd2O3

50000

60000

Figure 4-12 Assembly peak pin power profiles of Gd-155 and Gd-157 enriched U/Gd fuels

131

4.7 Combined Enriched Gd-155 and Gd-157 BP Designs


The sensitivity study was performed on enriched Gd-155 and Gd-157 combined in the fuel assembly. The enriched Gd-155 and Gd-157 are used in different fuel rods. In this analysis, 8 U/Gd fuel rods used enriched Gd-155, and the others 8 used enriched Gd157 in Gd2O3 mixture of the fuel. Enriched BP designs with their isotopes weight fractions are given in Table 4.2. The BP designs contain 16 U/Gd fuel rods with 2 w/o Gd2O3 concentration.

Table 4.8 summarizes the enriched Gd-155 and Gd-157 BP designs for different Gd2O3 concentrations. Figure 4-13 shows the Kinf profiles of each case compared to the natural design. However, Case 1 has very similar Gd depletion profile with the natural high concentrated design shown in Figure 4-14. It has the highest peak power value during the depletion in Figure 4-15. The compatibility criterion with natural and optimized design is to match the rate of decrease with burnup of the natural Gd2O3 absorption.
Table 4-8 Gd2O3 concentrations and enrichment ratio of Gd-155 and Gd-157 isotopes for different cases

Case Number 1 2 3 4 5

Gd2O3 Concentration(w/o) 3 3 3 1 2

Gd-155 w/o 50 50 50 50 50

Gd2O3 Concentration(w/o) 3 1 2 3 3

Gd-157 w/o 70 70 70 70 70

Figure 4-15 shows the peak power profiles of natural and enriched Gd-155 and Gd-157 designs. Maximum peak pin power reduces with optimized designs compared to

132 the natural designs. However, we observed an increase in peak pin power after 20,000 MWd/MTU.
Kinf Comparison of for different cases (U-235=4.5 w/o, 16 Gd pin, 3 w/o Gd2O3 Concentration) 1.4
No Gd Natural 8 w/o conc Gd-157=70 Gd-155=50 Case 1 Case 2 Case 3 Case 4 Case 5

1.3

1.2

Kinf

1.1

0.9

0.8 0 10000 20000 30000 Burnup (MWd/MTU) 40000 50000 60000

Figure 4-13 Kinf Comparison of different cases


Residual Delta k profiles for different weight fractions of Gd-157 and Gd-155 (U-235=4.5 w/o, 16 Gd pin) 0.05

Residual Delta k

-0.05

Natural

-0.1
Natural 8 w/o conc Gd-157=70 Gd-155=50 Case 1 Case 2 Case 3 Case 4 Case 5

-0.15

-0.2 0 10000 20000 30000 Burnup (MWd/MTU) 40000 50000 60000

Figure 4-14 Residual binding (delta k) comparison of different cases

133

Peak Pin Power profiles for different weight percent of Gd-155 and Gd-157 (U-235=4.5 w/o, 16 Gd pin) 1.16 1.14 1.12 Peak Pin Power 1.1 1.08 1.06 1.04 1.02 1 0 10000 20000 30000 Burnup (GWd/MTU) 40000 50000 60000
No Gd 8 w/o Natural Gd2O3 Gd-157=70 Gd-155=50 Case 1 Case 2 Case 3 Case 4 Case 5

Figure 4-15 Peak pin power vs. exposure for the different designs

4.8 Optimized Pure (100 w/o) Gd-155 and Gd-157 Designs with 2 w/o Gd2O3 Concentration
This section summarizes the sensitivity study on 100 w/o enriched Gd-155 and Gd-157 BP designs. The fuel assembly design contains 16 U/Gd fuel rods with 2 w/o Gd2O3 concentrated fuels. Figure 4-16 shows comparison of the Kinf profiles of 100 w/o Gd-155 and Gd-157 enriched fuel with natural and 70 w/o Gd-157 enriched fuel. Because Gd-157 has very large absorption cross-section (240,000 barn), enriched Gd-157 designs has very large Gd depletion rate as can be seen in Figure 4-17. Gd-155 has absorption cross section of 48,000 barn, which is much less than the Gd-157 isotope cross section. It provides less Gd depletion rate for the same concentration of optimized design. 100 w/o Gd-155 enriched design has very slow Gd depletion rate as shown in Figure 4-17.

134

Kinf Comparison of Gd-157 and Gd-155 enrichments (U-235=4.5 w/o, 16 Gd pin) 1.2
Natural 7 w/o Gd2O3

1.15 1.1 1.05 Kinf 1 0.95 0.9 0.85 0.8 0 10000 20000 30000 Burnup (MWd/MTU) 40000

Gd-157=70, 2 w/o Gd2O3 Gd-155=50,2 w/o Gd2O3 Gd-155=100,2 w/o Gd2O3 Gd-157=100, 2 w/o Gd2O3

50000

60000

Figure 4-16 Kinf profiles comparison for 100 w/o Gd-155 and Gd-157 enriched UO2/Gd2O3 fuel and natural design with 7 w/o concentration
Residual Delta k profiles for different weight fractions of Gd-157 and Gd-155 (U-235=4.5 w/o, 16 Gd pin)
0.02 0 -0.02 -0.04 -0.06 Delta k -0.08 -0.1 -0.12 -0.14 -0.16 -0.18 -0.2 0 10000 20000 30000 Burnup (MWd/MTU) 40000 50000 60000
Natural 7 w/o Gd2O3 Gd-157=70, 2 w/o Gd2O3 Gd-155=50,2 w/o Gd2O3 Gd-155=100,2 w/o Gd2O3 Gd-155=100,2 w/o Gd2O3 Gd-157=100, 2 w/o Gd2O3

Figure 4-17 Residual delta k profiles for 100 w/o Gd-155 and Gd-157 enriched UO2/Gd2O3 fuel and natural design with 7 w/o concentration

135

4.9 Equivalent Enriched U/Gd Fuel Assembly Design


Optimized fuel assembly design (enriched Gd isotopes with low concentration) presents lower residual absorption at EOC, which means a longer fuel cycle for a given uranium enrichment or lower uranium enrichment for a given fuel cycle length. Figure 418 shows Kinf profiles of natural and enriched Gd optimized designs for different U-235 enrichment w/o. Equivalent optimized fuel assembly design was decided by comparing EOC Kinf and residual reactivity values in Figure 4-18 and 4-19. For a fixed fuel cycle length, optimized design decreases the uranium enrichment requirement from 4.80 w/o to 4.64 w/o for assembly position H-09 in the core to provide the same profile of Kinf and residual binding. The EOC soluble boron concentrations are the same for natural and equivalent optimized enriched Gd designs in core depletion.
Kinf profiles of natural and optimized design for different U-235 enrichment (w/o) 1.2 1.15 1.1 1.05 Kinf 1 0.95 0.9 0.85 0.8 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000 60000 Burnup (MWd/MTU) Natural, 7 w/o Gd2O3, 4.80 w/o U-235 Optimized, 4.80 w/o U-235 Optimized, 4.64 w/o U-235 Optimized, 4.50 w/o U-235 Optimized, 3.50 w/o U-235

Figure 4-18 Kinf profiles of natural and enriched Gd U/Gd FA designs for different U235 enrichment

136

Residual delta k profiles of natural and optimized design for different U-235 enrichment (w/o)
0 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 -0.12 -0.14 -0.16 -0.18 Natural, 7 w/o Gd2O3, 4.80 w/o U-235 Optimized, 4.80 w/o U-235 Optimized, 4.64 w/o U-235 Optimized, 4.50 w/o U-235 Optimized, 3.50 w/o U-235 Residual reactivity saving 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000 60000

Figure 4-19 Residual delta k profiles of natural and enriched Gd U/Gd FA designs for different U-235 enrichment (w/o)

4.10 Enriched Gd-157 and Gd-155 FA Designs in the Core Depletion Calculations
Due to their similar Kinf profiles, seven different cases shown in Table 4-9 were compared with each other in core depletion calculation by looking at their RPD and pin power values. The fuel assembly with 7 w/o Gd2O3 concentration (position H-09) in the BP pattern was matched with the designs given below and then guideline constraints were checked out during the depletion. The previously developed deterministic final BP loading pattern is shown in Figure 4-20. Due to its ability to model and to deplete enriched gadolina fuel, the CASMO-4 and SIMULATE-3 package was used to perform core calculations. CASMO-4 image files were generated for each optimized design defined below. The entire image files were included to the optimization library developed

137 for natural gadolinia fuels. All optimized designs use the equivalent uranium enrichment to provide the same EOC residual binding and soluble boron concentration.
Table 4-9 Optimized designs for position H-09 in the core loading pattern Case No 1 Definition Natural 7 w/o Gd2O3 Uniform concentrations, 2 w/o Gd2O3 with 70 w/o Gd-157 Uniform concentrations, 3 w/o Gd2O3 with 70 w/o Gd-157 Mixed concentrations, 1 w/o and 3 w/o Gd2O3 with 70 w/o Gd-157 Mixed concentrations, 2 w/o and 3 w/o Gd2O3 with 70 w/o Gd-157 Uniform concentrations, 3 w/o Gd2O3 with 70 w/o Gd-157 and 3 w/o Gd2O3 with 50 w/o Gd-155 Mixed concentrations, 1 w/o Gd2O3 with 50 w/o Gd-155 and 3 w/o Gd2O3 with 70 w/o Gd-157

2 3 4 5 6 7

All of the designs result with peak pin power increase around mid cycle in the core depletion calculations. Figure 4-21 shows that natural enriched design has pin power value less than the limit value of 1.55. The main reason for this increase is the fast depletion of enriched Gd-157. Same behaviors can also be observed for RPD power in Figure 4-22. Case 4 design has the closest RPD power profiles to the natural design shown in Figure 4-22, however, its peak pin power profile exceeds the limit value given in Figure 4-21.

138

8
C-03 0 4.00 00 - 00

9
F 1 4.80 16* - 07 1

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 16* - 02 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 16* - 03 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 16* - 02 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 16* - 02 2

14
M-11 0 4.30 08 04 90 F 1 4.87
5 F 1 4.87 7

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

O-13 0 4.90 08 - 02 180

L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 16* - 02 6


F 1 4.90 16* - 02 9 F 1 4.90 10

L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron ) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

N-13 0 4.85 08 - 04 180

Degree Rotation

Figure 4-20 Burnable poison loading pattern developed with deterministic and heuristic methods

139

Comparison of peak pin power value of H-09 for different optimized designs during cycle depletion 1.8 1.7 1.6 Peak Pin Power 1.5 1.4 1.3 1.2 1.1 Case 5 1 0 5 10 15 Cycle Burnup (GWd/MTU) 20 25 30 Case 6 Case 7 limit Case 1 Case 2 Case 3 Case 4

Figure 4-21 Pin power profiles of position H-09 for different optimized fuel designs
Comparison of RPD of H-09 for different optimized design during cycle depletion 1.6

1.5

1.4 RPD Power

1.3

1.2

1.1

Case 1 - Natural Case 2 Case 3 Case 4 Case 5 Case 6 Case 7

1 0 5 10 15 Cycle Burnup (GWd/MTU) 20 25 30

Figure 4-22 RPD power profiles of position H-09 for different optimized fuel designs

140

4.11 Pure (100 w/o) Gd-155 Enriched Designs with 2 w/o Gd2O3 Concentration
In this part, pure (100 w/o) Gd-155 enriched BP design with 2 w/o Gd2O3 concentration replaced the high (7 w/o) Gd2O3 concentrated natural Gd enriched fuel assembly design at position H-09 in the given core loading pattern. Due to residual binding savings, the fuel assembly average enrichment was reduced to match with the similar peak pin power and RPD power profile of conventional design. Case 8 has same uranium enrichment of 4.80 w/o with conventional U/Gd FA design. Case 9 has equivalent uranium enrichment of 4.64 w/o, which is less than 4.80 w/o or the enrichment of the natural Gd fuel design. Figures 4-23 and 4-24 show the pin power profiles and RPD power during the core depletion. Both designs provide good results in the core depletion calculations for the reference core. Equivalent enriched fuel assembly design provided uranium savings due to elimination of U-235 w/o penalty for enriched Gd fuel rods. It has very similar pin power and RPD power profiles with natural design both under the constraint values.

141

Comparison of pin power value of H-09 different optimized during cycle depletion 1.6 1.55 1.5 1.45 Pin Power 1.4 1.35
Case 1: U-235=4.80 w/o, Natural Gd enriched, 7 w/o Gd2O3

1.3 1.25 1.2 0 5 10

Case 8: U-235=4.80 w/o, Gd-155=100 w/o, 2 w/o Gd2O3 Case 9: U-235=4.64 w/o, Gd-155=100 w/o, 2 w/o Gd2O3 limit

15 Cycle Burnup (GWd/MTU)

20

25

30

Figure 4-23 Pin power profiles of position H-09 for different Gd fuel designs
Comparison of RPD of H-09 different optimized during cycle depletion 1.4

1.35

1.3 RPD Power

1.25 Case 1: U-235=4.80 w/o, Natural Gd enriched, 7 w/o Gd2O3 1.2 Case 8: U-235=4.80 w/o, Natural Gd enriched, 2 w/o Gd2O3 1.15 Case 9: U-235=4.64 w/o, Natural Gd enriched, 2 w/o Gd2O3 1.1 0 5 10 15 Cycle Burnup (GWd/MTU) 20 25 30

Figure 4-24 RPD power profiles of position H-09 for different Gd fuel designs

142

4.12 Final Heuristics BP Loading Pattern with 4 Optimized U/Gd FA Designs


The pure (100 w/o) Gd-155 enriched BP design with 2 w/o Gd2O3 concentration was replaced to highly (7 w/o) Gd2O3 concentrated conventional BP fuel assembly design at position H-09 in the final heuristics BP loading pattern. The total 4 fuel assemblies were replaced with the new design. The new design used uranium enrichment 4.64 instead of 4.80. Figure 4-25 shows the pin power profiles of fresh fuel locations for this BP loading pattern replaced. Maximum peak pin power occurs at position K-14 in the loading, and it is 1.546 less than the guideline constraint of 1.55. Figure 4-26 shows the RPD profiles for each fresh fuel locations. Position H-09 in the core loading has maximum RPD of 1.366 less than the RPD limit.

Cycle 16 - BP loading with optimized design Peak pin power distribution in the core for fresh fuel assembly locations (4 U/Gd FA replaces with optimized design) 1.6 1.55 1.5 Peak pin power 1.45 1.4 1.35 1.3 1.25 1.2 1.15 1.1 0 H-09 H-13 5 K-10 K-12 10 15 20 Assembly Average Burnup (GWd/MTU) K-14 L-13 L-14 M-11 M-13

Peak Pin power constraint

25 N-13

30 pin power limit

Figure 4-25 Peak pin power distributions of final BP loading with 4 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs

143
Cycle 16 - final heuristic BP loading Normalized radial power distribution in the core for fresh fuel assembly locations (4 Natural U/Gd replaces with the optimized designs) 1.4 1.35 Radial power 1.3 1.25 1.2 1.15 1.1 1.05 1 0 H-09 H-13 5 K-10 10 15 20 Fuel Assembly Average Burnup (GWd/MTU) K-12 K-14 L-13 L-14 M-11 25 M-13 30 N-13

Figure 4-26 Normalized power distributions of optimal BP loading with 4 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs

4.13 Superior BP Loading Pattern with 16 Optimized U/Gd FA Designs


The pure (100 w/o) Gd-155 enriched BP design with 2 w/o Gd2O3 concentration was replaced the high (7-8 w/o) Gd2O3 concentrated conventional BP fuel assembly designs in the superior BP loading pattern. Figure 4-27 shows previously developed superior BP loading design by using the deterministic methodology. This BP loading pattern had fresh fuel locations with high concentrated natural design. For optimized core, the positions H-09, K-10, and M-11 was loaded with the optimized FA designs of 100 w/o Gd-155 enriched U/Gd FAs. Total of 16 U/Gd FA optimized designs were placed in the loading. The results of the cycle depletions given in Figure 4-29 and 4-30 show that pin power and RPD powers were satisfactory enough and they did not violate power constraints during the depletion. Maximum peak pin power occurs at position K14, and it is 1.549 as shown in Figure 4-28. Maximum RPD power is 1.385, and it occurs at position H-09 as shown in Figure 4-29.

144

8
C-03 0 4.00 00 - 00

9
F 1 4.80 16* - 07 1

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80 16* - 07 3 L-09 0 4.90 16* - 03 270

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55 16* - 08 8

12
K-10 0 4.90 16* - 03 180 F 1 4.55 16* - 06 4 N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

13
F 1 4.70 16* - 06 2

14
M-11 0 4.30 08 04 90 F 1 4.87
5 F 1 4.87 7

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

O-13 0 4.90 08 - 02 180

L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55 16* - 06 6


F 1 4.90 16* - 06 9 F 1 4.90 10

L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron ) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

N-13 0 4.85 08 - 04 180

Degree Rotation

Figure 4-27 Superior BP loading pattern developed by deterministic method

145
Cycle 16 - Superior optimized core loading Peak pin power distribution in the core for fresh fuel assembly locations (16 Natural U/Gd replaces with the optimized designs)
1.6 1.55 1.5 Peak pin power 1.45 1.4 1.35 1.3 1.25 1.2 1.15 1.1 0 5 10 15 20 Assembly Average Burnup (GWd/MTU) 25 30

Peak Pin power limitconstraint

H-09

H-13

K-10

K-12

K-14

L-13

L-14

M-11

M-13

N-13

pin power limit

Figure 4-28 Pin power profiles for superior BP loading with 16 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA designs

1.45 1.4 1.35 R adial pow er 1.3 1.25 1.2 1.15 1.1 1.05 1 0 H-09

Cycle 16 - final heuristic BP loading Normalized radial power distribution in the core for fresh fuel assembly locations (16 Natural U/Gd replaces with the optimized designs)

5 H-13 K-10

10 15 20 Assembly Average Burnup (GWd/MTU) K-12 K-14 L-13 L-14 M-11

25 M-13

30 N-13

Figure 4-29 RPD distributions for superior BP loading with 16 natural U/Gd FA replaces with optimized enriched Gd-155 U/Gd FA design

146

4.14 Conclusions
The natural high (7-8 w/o) Gd2O3 concentrated U/Gd FA designs were successfully replaced with the optimized enriched Gd-155 designs in the reference core loading. However, 100 w/o Gd-157 design provided 44 ppm more residual binding savings, it causes maximum peak pin power violation when it is placed to the core. On the other hand, 100 w/o Gd-155 design presented only 28 ppm more residual binding savings compared to natural U/Gd fuel designs. Sensitivity analysis on optimized designs for core depletion showed that 100 % enriched Gd-155 provides the best peak power and RPD profiles during the depletion.

147

CHAPTER 5 FUEL COST ANALYSIS


5.1 Introduction
Penn State equilibrium fuel cycle cost methodology [39] was updated to include Gd enrichment cost into nuclear fuel cycle cost components. Levelized fuel cycle cost and net fuel cost savings amount due to use of enriched Gd designs in the core were calculated by using this methodology for different cost scenarios.

The nuclear fuel cycle cost consists of the front-end and the back-end cost components. However, the cost of the irradiation process in the reactor is traditionally covered under the capital or operating and maintenance costs of the nuclear power plant. Moreover, the front-end includes the cost of components such as uranium ore, conversion, and enrichment (if applicable), fuel assembly fabrication including transport, and the costs for fuel management activities. The back-end contains the cost for reprocessing, if applicable, and the final storage of high level radioactive wastes. If required, the above-mentioned costs have to be assessed separately for the first core, and for the reload fuel assemblies [2].

The general procedure for arriving at different cost components of the nuclear fuel cycle is divided into the following three steps:

148 Assessment of costs incurred during the front-end activities (mining, conversion, enrichment, fuel assembly fabrication) Assessment of cost incurred during intermediate storage of spent fuel in the plant Assessment of cost incurred during the back-end activities (storage of spent fuel, transport, reprocessing, if applicable, final disposal of high level radioactive waste)

5.2 Calculation of Cost Components


In order to calculate the overall fuel cycle cost, one must first calculate the magnitude of each cost component, and identify the appropriate point in time at which the quantities of fuel are obtained from the reactor neutronic calculations. These quantities of material are adjusted to allow for process losses in the various stages of the nuclear fuel cycle, and then multiplied by the unit cost to obtain component cost. Table 51 summarizes the parameter notation for the fuel cycle cost calculations below.

149
Table 5-1 Parameter Notation for Fuel Cycle Cost Calculations

Parameter Discount rate Time Base date of monetary unit Date of fuel loading Fuel residence time Mass of uranium feed (kg) Mass of uranium charged in reactor (kg) Mass of uranium in the tails (kg) Fraction of U235 in the uranium feed Fraction of U235 charged in reactor Fraction of U235 in the tails Fraction of U235 discharged Conversion factor from kg U to lb U3O8 (a lb U3O8 per kg U) Total component cost Unit cost Escalation ratio Material losses Total loss factor Lead or lag time

Symbol R T tb (01/01/2004) tc Tr Mf Mp Mt Xf (0.711 %) Xp Xt xd a (2.6) Fi Pi si li fi ti

Where i denotes fuel cycle process as follows: i=1 symbolizes uranium purchase, i=2 symbolizes conversion process , i=3 symbolizes enrichment process, i=4 symbolizes fabrication process, i=5 symbolizes spent fuel storage process Monetary units for each fuel cycle process are given as: P1 = monetary units per lb U3O8, P2=monetary units per kg U, P3=monetary units per SWU, P4=monetary units per kg U fabricated, P5=monetary units per assembly ($/Assembly), P6=monetary units per g of enriched Gd product
5.2.1 Calculation of Each Cost Components

For each component, process cost for a given equilibrium fuel cycle can be written in the following equation [39]:

150
Cost of Uranium
F1 = M f a f1 P1 (1 + s1 ) t tb

(Equation 5.1)

Where:
f 1 = (1 + l 2 )(1 + l 3 )(1 + l 4 ) , M =[(x
f p-xt)/(xf-xt)]

Mp ,

Date of front-end components: t=tc-ti


Cost of Conversion
F2 = M f f 2 P2 (1 + s2 )t t b

(Equation 5.2)

Where:
f 2 = (1 + l 2 )(1 + l 3 )(1 + l 4 )

Cost of Enrichment

F3 = S f 3 P3 (1 + s3 )t t b Where: S=Separative Work Units = MpVp + MtVt MfVf Mt=Mf-Mp,

(Equation 5.3)

Va= (2xa-1) ln[xa/(1-xa)] and a is subscript for f, p or t.

f 3 = (1 + l 3 )(1 + l 4 )

Enrichment cost is calculated by considering 0.3 w/o tails assay ratio.


Cost of Fabrication
F4 = M p f 4 P4 (1 + s4 )t t b

(Equation 5.4)

Where:
f 4 = (1 + l 4 ) =

1.

Cost of Spent Fuel Storage


F5 = Batchsize P5 (1 + s5 ) t tb

(Equation 5.5)

Date of back-end components, t= tc+Tr+t5 Where: t5=0 and zero loss factor during irradiation are assumed for spent fuel storage.
Cost of Enriched Gadolinium

F 6 = M Gd P6 M Gd= Total mass of enriched Gd product used in the core

(Equation 5.6)

151
5.2.2 Discounting and Levelizing Fuel Cycle Costs

The total fuel cost is the summation of the all component costs. Each component cost uses the selected base date for discounting back. After calculating discounted component costs, the present value of the total fuel cost will be determined by adding component costs.

The total discounted cost of the nuclear fuel cycle can be written as:

Fi(t ) /(1 + r )
i t =t o T1

t =to +Tr +T2

( t to )

(Equation 5.7)

where: Fi(t)= cost for the i th component at time t , to= reference date, Tr = fuel residence time, T1= maximum value of lead time (in front-end), T2= maximum value of lag time (in back-end) i= cost component number

If C is the constant levelized fuel cost per unit of electricity sent out by a reactor, the total cost of fuel cycle can also be determined in the following equation:
t o + TR t =to

C E (t ) /(1 + r )

t to

(Equation 5.8)

where: E(t) = net electrical output at time t. From the balance of (5.9) and (5.10), the levelized fuel cycle cost can be calculated by the following equation [39]:

C=

t = t c + TR + T2 t = t c T1

Fi(t ) /(1 + r )

( t t mid )

t c + TR t =tc

E (t ) /(1 + r )

t t mid

(Equation 5.9)

152 The cycle energy production is assumed to be at the midpoint of the cycle for an equilibrium cycle. We define tmid as the midpoint time of the cycle, and tc as the date of fuels reload. front-end fuel cycle back-end fuel cycle

Uranium, conversion, enrichment, fabrication

Irradiation (reference time)

Fuel Storage

Time Discounting of front-end fuel cycle cost components Discounting of back-end fuel cycle cost components

Figure 5-1 Escalation and discounting method

5.3 Fuel Cycle and Cost Data


Tables 5-2, 5-3, 5-4 and 5-5 summarize the reference fuel cycle data, unit prices for each cost component [4, 5], economic parameters and batch data used in equilibrium fuel cost analysis for reference natural and optimized core design. Penn State equilibrium fuel cost methodology was used to determine fuel cost for different unit cost and economic scenarios given below. The utilized data and performed calculations are given just as one example to demonstrate the functionality of the Penn State methodology for sensitivity analysis.

153
Table 5-2 Fuel Cycle Data [4, 5]

Monetary base time year Lead time of uranium purchase Lead time for conversion Lead time for enrichment Lead time for fabrication Lag time for spent fuel storage Loss factor for conversion Loss factor for fabrication Tail Assays ratio

2004 10 weeks 8 weeks 4 weeks 2 weeks 0 days 0.5 w/o 0.0 w/o 0.3 w/o

Table 5-3 Unit Prices for Each Component [4, 5] Component Low Nominal High

Uranium ($/kg U) Conversion ($/kg U) Enrichment ($/SWU) Fabrication ($/kg U) PWR Spent fuel storage ($/Assembly) PWR

18.30 3.50 80 185 0

31.35 5.00 105 210 50,000

47.00 7.00 130 275 100,000

Table 5-4 Economic Parameters [4, 5] Parameter Low Nominal High

Escalation rate for spent fuel storage (% per year) Discount Rate (% per year)

2 7

3 9.5

5 12

154
Table 5-5 Batch regions and feed fuel assembly numbers of the TMI-1 reference core loading

Batch Enrichment (w/o U-235) Batch No 1 2 3 4 5 6 Total Reference Core 4.55 4.70 4.80 4.80 4.87 4.90 Optimized Core 4.55 4.70 4.64 4.80 4.87 4.90 -

Total No of Feed Assemblies In the core 20 4 4 8 16 16 68

5.4 Fuel Cost Analysis


Economic analyses of the TMI-1 reference core loaded with natural and optimized UO2/Gd2O3 are presented by the equilibrium fuel cost graphs, which reflects sensitivity studies on Gd enrichment cost, uranium (gross fuel) savings and net fuel savings. Figure 5-2 shows the sensitivity analysis of equilibrium fuel cycle cost for different scenarios for TMI-1 reference cycle with natural UO2/Gd2O3 FA designs. The nominal fuel cycle cost was $ 43,906,432, and it varied from $ 49,373,064 to $ 38,728,160 depending on the upper and lower enrichment unit cost scenario. The high Gd2O3 concentration fuel assembly in core position H-09 was replaced with the low concentrated optimized fuel assembly design. This position required the use of total 4 FAs with the optimized UO2/Gd2O3 design in the core, and the optimized design provided 0.16 w/o enrichment savings because of lower residual binding at EOC. The total amount of required enriched Gd-155 was 2.70 kg for 4 U/Gd FAs replaced with the optimized enriched designs. The fuel cost resulted with a gross fuel savings of $ 80,780 and, the fuel cost decreased to $ 43,906,432. Figure 5-3 and 5-4 show sensitivity of equilibrium fuel cycle cost and uranium and net savings for different scenarios for TMI-1 reference

155 cycle with optimized four UO2/Gd2O3 FA designs. The net fuel cycle cost savings was calculated from the difference of uranium savings and Gd enrichment cost component, and it was $ 53, 992 for nominal unit costs and $ 10 /g-enriched Gd product.

Figure 5-5, 5-6 and 5-7 show the sensitivity of uranium and net savings amount on Gd enrichment costs. The net savings amount increased to $ 67, 384 by reducing Gd enrichment cost to half of the nominal value shown in Figure 5-5. The net savings amount drops to $ 27, 204 and almost to zero for optimized four UO2/Gd2O3 FA designs used in the core by increasing unit cost to $ 20 and $ 30 /g-Gd-155 respectively as shown in Figure 5-6 and 5-7.

Figure 5-8 shows sensitivity of fuel cycle cost on unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four UO2/Gd2O3 FA designs. The fuel cost of optimized four UO2/Gd2O3 FA designs increases with Gd enrichment cost, and it covers the uranium fuel savings where Gd enrichment cost is around $ 30 /g-Gd-155. As we expected, there is a small increase in levelized cost for optimized core design in Figure 59 because of smaller Gd amount in the core. Figure 5-10 presents the sensitivity results of gross uranium savings and net savings on unit Gd-155 enrichment cost for the TMI-1 current cycle with 4 optimized U/Gd FA designs. Net savings amount drops linearly with increasing the Gd enrichment unit cost.

156

Sensitivity of Equilbirum Fuel Cycle Cost for different scnearios TMI-1 Reference Cycle 16 with natural U/Gd FA designs

55,000,000 50,000,000 45,000,000 Fuel Cost ($) 40,000,000 35,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000

Figure 5-2 Sensitivity of eq. fuel cycle cost for different scenarios for TMI-1 reference cycle with natural U/Gd FA designs
Sensitivity of Equilbirum Fuel Cycle Cost for different scnearios TMI-1 Reference Cycle 16 with 4 optimized U/Gd FA designs (Gd-155 unit cost= 10 $/g)

55,000,000 50,000,000 45,000,000


Fuel Cost ($)

40,000,000 35,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000


St or ag e Lo w Lo U w H SW ig h U D is co un t Lo Lo w w E Fa sc br Lo i c w at i C on on ve rs io n N H ig om h in C al on ve rs io n H H i gh ig h Fa Es c br ic at io H n ig h SW Lo w U D is co un t H ig H h ig U h St or ag e

Lo w

Figure 5-3 Sensitivity of eq. fuel cycle cost for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs

St or ag e Lo w Lo U w H S ig W h U D is co un t Lo Lo w w E Fa sc br Lo ic w at C on ion ve rs io n H N ig om h i C on nal ve rs io n H H ig ig h h Fa Es c br ic at io H n ig h SW Lo w U D is co un t H ig H h ig U h St or ag e
Scenario

Lo w

Scenario

157

Sensitivity of Net Fuel Cost savings of optimized equilbirum fuel cycle for different scnearios, TMI-1 Reference Cycle 16 with 4 optimized U/Gd FA designs (Gd-155 unit cost= 10 $/g) 100,000 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0

Savings ($)

Figure 5-4 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs
Sensitivity of Uranium and Net Fuel Cost savings for different scnearios, TMI-1 Reference Cycle 16 with 4 optimized U/Gd FA designs (Gd-155 unit cost= 5 $/g) 100,000 90,000 Savings ($) 80,000 70,000 60,000 50,000 40,000 30,000

Figure 5-5 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs

St or ag e Lo w U Lo w H S W ig h U D is co un t Lo Lo w w E Fa sc br Lo i c w at i C o n on ve rs io n N H o ig m h in C al on ve rs io n H H i g ig h h Fa E s c br ic at io H n ig h SW Lo w U D is co un t H ig H h ig U h St or ag e
Scenario Net Savings U savings

Lo w

or ag e Lo w U Lo w H S W ig h U D is co un t Lo Lo w w E Fa sc br Lo i ca w ti C on on ve rs io n N H om ig h in C al on ve rs io n H H ig ig h h Fa Esc br ic at io H n ig h S Lo W w U D is co un t H ig H h ig U h St or ag e
Scenario Net Savings U savings

Lo w

St

158

Sensitivity of Net Fuel Cost savings of optimized equilbirum fuel cycle for different scnearios, TMI-1 Reference Cycle 16 with 4 optimized U/Gd FA designs (Gd-155 unit cost= 20 $/g)

100,000 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0
Es c b Lo ric w at i C on on ve rs io n N H om ig h in C al on ve rs io n H H ig ig h h Fa Esc br ic at io H n ig h SW Lo w U D is co un t H ig H h ig U h St or ag e Fa St or ag e U U SW ig h D is Lo w co Lo w un t

Savings ($)

Lo w

Lo w

Lo w

Scenario Net Savings U savings

Figure 5-6 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs
Sensitivity of Net Fuel Cost savings of optimized equilbirum fuel cycle for different scnearios, TMI-1 Reference Cycle 16 with 4 optimized U/Gd FA designs (Gd-155 unit cost= 30 $/g)

100,000 80,000
Savings ($)

60,000 40,000 20,000 0 -20,000


ou nt L ow Lo w F a Esc br Lo ic w at i C on on ve rs io n N H om ig h in C al on ve rs io n H H i g ig h h Fa Esc br ic at io H n ig h S Lo W w U D is co un t H ig H h ig U h St or ag e ag e U Lo w or SW H ig h U D is c

St

Lo w

Lo w

Scenario Net Savings U savings

Figure 5-7 Sensitivity of savings for different scenarios for TMI-1 reference cycle with 4 optimized U/Gd FA designs

159

Sensitivity of fuel cycle cost of current and optimized equlibirium cycle on enriched Gd-155 unit cost (4 FA replaces with optimized design for position H-09)

$43,920,000 $43,910,000 $43,900,000


Fuel Cycle Cost ($ )

$43,890,000 $43,880,000 $43,870,000 $43,860,000 $43,850,000 $43,840,000 $43,830,000 $43,820,000 0 5 10 15 20 25 30 35 Unit Cost of Gd-155 ($/g) Total Optimized Fuel cost Current Fuel Cost

Figure 5-8 Sensitivity of fuel cycle cost versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs
Sensitivity of Levelized Cost on Gd-155 enrichment cost for natural and optimized fuel cycle designs (4 FA replaces with optimized design for position H-09)
4.017 Levelized Equilbirum Fuel Cost (mill/kWh) 4.016 4.015 4.014 4.013 4.012 4.011 4.010 4.009 4.008 4.007 0 5 10 15 20 25 30 35 Gd-155 enrichment cost ($/g-product)

Optimized Design Natural design

Figure 5-9 Sensitivity of levelized fuel cycle cost versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs

160

Sensitivity of gross uranium savings and net savings for optimized cycle on enriched Gd-155 unit cost (4 FA replaces with optimized design for position H-09)

$90,000 $80,000 $70,000 $60,000


Savings ($ )

$50,000 $40,000 $30,000 $20,000 $10,000 $0 0 5 10 15 20 25 30 35 Unit Cost of Gd-155 ($/g) Gross Uranium Savings Enriched Gd cost Net Savings

Figure 5-10 Sensitivity of gross uranium savings and net savings versus unit Gd-155 enrichment cost for TMI-1 current cycle with optimized four U/Gd FA designs

All of the above results were calculated by replacing the natural high concentrated UO2/Gd2O3 FA design with the optimized design. Economic calculation results were projected to different number of optimized fuel assemblies in the core by using the same equivalent fuel assembly approach and reducing the U-235 enrichment of each batch in the core. Results were summarized in Figure 5-11 and 5-12 for equilibrium fuel cost and savings. There was a potential of ~ $ 1, 377, 000 for gross savings and ~ $ 916,000 net savings, if all of the 68 feed assemblies would be replaced with the optimized design. Figure 5-13 shows the sensitivity of levelized fuel cycle costs as a function of the number of optimized U/Gd FA designs for the TMI-1 reference core. The levelized cost showed an approximately 2 % drop for the 68 FA optimized designs used in the core.

161

Sensitivity of fuel cycle cost of current and optimized equlibirium cycle on # of Enriched Gd-155 FA in the core

$44,000,000 $43,800,000 $43,600,000


Fuel Cycle Cost ($ )

$43,400,000 $43,200,000 $43,000,000 $42,800,000 $42,600,000 $42,400,000 0 10 20 30 40 50 60 70 80 # of Enriched Gd-155 FA in the core Optimized Fuel Cost Current Fuel Cost

Figure 5-11 Sensitivity of fuel cycle cost versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle
Sensitivity of gross uranium savings and net savings for optimized cycle on # of Enriched Gd-155 FA in the core $1,600,000 $1,400,000 $1,200,000 Savings ($ ) $1,000,000 $800,000 $600,000 $400,000 $200,000 $0 0 10 20 30 40 50 60 70 80 # of Enriched Gd-155 FA in the core Gross Uranium Savings Enriched Gd cost Net Savings

Figure 5-12 Sensitivity of gross uranium savings and net savings versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle

162

Sensitivity of Levelized Cost on # of Enriched Gd-155 FA in the core for natural and optimized fuel cycle designs 4.030 4.020
Levelized Equilbirum Fuel Cost (mill/kWh)

4.010 4.000 3.990 3.980 3.970 3.960 3.950 Optimized Design 3.940 3.930 3.920 0 10 20 30 40 50 60 70 80 # of Enriched Gd-155 FA in the core Natural design

Figure 5-13 Sensitivity of levelized fuel cycle costs versus number of optimized U/Gd FA designs for TMI-1 current cycle with optimized cycle

CHAPTER 6 CORE BURNABLE POISON OPTIMIZATION WITH GENETIC ALGORITHMS


6.1 Introduction
Chapter 6 utilizes the results for the U/Gd BP FA designs developed in Chapter 3, which provide the minimum peak pin power and residual binding during code depletion in the TMI-1 reference core. These BP designs are now utilized in a GA BP placement optimization as applied to all fresh fuel assembly locations in the core (see Figure 6-1). The optimization problem is defined in Section 6.2, and the problem geometry, decision variables and design constraints are summarized in detail in the following sections. The genetic algorithm methodology and the development of the code and its application to the TMI-1 core are also presented in detail.

The core BP optimization problem develops a BP loading map for a given core loading configuration that minimizes the total Gd in the core without violating any design constraints. The number of UO2/Gd2O3 pins and Gd2O3 concentrations for each fresh fuel location are the decision variables and the total amount of Gd in the core is in the objective function. The constraints are the maximum peak pin power during depletion and Soluble Boron Concentration at BOC. The core BP optimization problem starts with the given equilibrium reference core loading pattern. The core BP optimization problem is directed toward optimizing BP patterns in the fresh fuel assemblies in the reference

164 core. The number of UO2/Gd2O3 pins and Gd2O3 concentrations are two parameters that can be changed during evaluation to determine the best BP loading pattern; however, the selection of configurations of UO2/Gd2O3 fuel pins is restricted to those designs determined as optimum in Chapter 3.

The different number of UO2/Gd2O3 pins and concentration of Gd2O3 fuel represents different fuel type in the CASMO/SIMULATE-3 reactor core design system. Each different type has to be given different segment name. Therefore, the core BP optimization calculations require a comprehensive library, which includes the generation of all possible segment names and cross section data for different fuel types having different number of UO2/Gd2O3 pins and different Gd2O3 concentrations. This comprehensive library is generated once for all future SIMULATE-3 reactor physics calculations with optimal UO2/Gd2O3 pins configurations determined before in Chapter 3 for each fuel type.

The developed GA code starts with an initial randomly generated BP map for all of the fresh fuel assemblies in the core, which represents an individual in the population. Then GA continues with genetic operators to create next generation individuals. The SIMULATE-3 reactor physics code is used to calculate the fitness of each member of the population and to check the design constraints during core depletion. The new technique is introduced to divide the solution space into three regions based on the BOC SOB concentration and maximum peak pin power constraints. Each region uses different objective functions. The good solutions (region 1) do not violate any of the constraints,

165 and the fitness of a solution is represented with its total Gd amounts. The region 2 represents the population members having maximum peak pin power value greater than the limit value of 1.55, but the BOC soluble boron concentration are less than the constraint value of 1700 ppm. The fitness of an individual is calculated by using its maximum peak pin power value with the objective of minimizing maximum peak pin power value in the core depletion to obtain all individuals that do not violate either of the two design constraints. The GA utilizes these individuals to determine the optimum configurations. When individuals are produced during the GA process that violates the BOC soluble boron limit (region 3), they are assigned a high negative penalty to reduce the probability of being selected for the next generation.

6.2 Problem Definition


Figure 6-1 shows the reference initial core loading model with fresh fuel locations in green color and used fuel locations in yellow. There are a total 10 fresh fuel locations given in the reference core loading.

Table 6-1 presents the detailed initial core loading data for the reference TMI-1 core. This table shows the required input file format in order to initialize the calculations. There are 29 fuel assembly positions in the octant core symmetry, each numbered in Table 6-1 according to their positions in the core. The initial core loading pattern has a total of 3 different fuel types, and each type is represented with different numbers as presented in Table 6-1. The 0 represents used fuel without any boron absorber, 1 represents fresh fuel, and 2 represents used fuel with boron absorber. The number of

166 U/Gd pins within a fuel assembly and their concentrations are also shown in Figure 6-1 and Table 6-1.

8
C-03 0 4.00 00 - 00

9
F 1 4.80 1

10
H-09 2 4.90 08 - 02 2.4P 180 F 1 4.80
3

11
H-13 2 4.75 08 - 02 2.7P 180 K-12 2 4.75 08 - 02 2.7P 270 M-14 0 4.90 08 - 02 90 F 1 4.55
8

12
K-10 0 4.90 16* - 03 180 F 1 4.55
4

13
F 1 4.70 2

14
M-11 0 4.30 08 04 90 F 1 4.87
5 F 1 4.87 7

15
N-12 2 4.90 08 - 02 2.3P 270 K-11 2 4.45 08 - 02 1.4P 90 M-12 2 4.62 08 - 02 3.5P 90

O-13 0 4.90 08 - 02 180

L-09 0 4.90 16* - 03 270

N-14 0 4.90 08 - 02 180 M-13 0 4.75 16* - 03 180 O-12 0 4.85 08 04 180

L-13 2 4.90 08 - 02 3.5P 90 F 1 4.55


6 F 1 4.90 9 F 1 4.90 10

L-14 0 4.75 08 - 04 180 K-14 0 4.75 08 - 04 180

Previous Location (F=Fresh) Type (0:Used fuel, 1:Fresh fuel, 2:Used fuel with boron ) Fuel Enrichment (U-235 w/o) No of U/Gd fuel pins (*: Mixed) Gd2O3 concentration (w/o)
BP Loading (w/o) (P for Pulled) or fresh fuel number

N-13 0 4.85 08 - 04 180

Degree Rotation

Figure 6-1 Reference initial octant core loading with fresh and used fuels

167
Table 6-1 Initial core loading and input data for the reference TMI-1 core

Fuel No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Position in octant core H-08 H-09 H-10 H-11 H-12 H-13 H-14 H-15 K-09 K-10 K-11 K-12 K-13 K-14 K-15 L-10 L-11 L-12 L-13 L-14 L-15 M-11 M-12 M-13 M-14 N-12 N-13 N-14 O-13

U-235 Enrichment 4 4.8 4.9 4.75 4.9 4.7 4.3 4.9 4.9 4.8 4.75 4.55 4.9 4.87 4.45 4.9 4.9 4.9 4.55 4.87 4.62 4.55 4.75 4.9 4.75 4.85 4.9 4.75 4.85

Type 0 1 2 2 0 1 0 2 2 1 2 1 2 1 2 0 0 0 1 1 2 1 0 1 0 0 1 0 0

Number of U/Gd pins 0 0 8 8 16* 0 8 8 8 0 8 0 8 0 8 16* 8 8 0 0 8 0 16* 0 8 8 0 8 8

1 *

Gd2O3 Concentration (w/o) 0 0 2 2 3 0 4 2 2 0 2 0 2 0 2 3 2 2 0 0 2 0 3 0 4 4 0 4 4

: 0=Used Fuel without Boron, 1=Fresh Fuel, 2=Used Fuel with Boron : Mixed concentrations (8 pins fixed with 8 w/o Gd2O3 others with variable concentration)

168
6.2.1 Decision Variables

The two main decision variables and their possible values are explained below: 1. No of U/Gd fueled pins for each fresh fuel position (0, 4, 8, 12, 16, 16 mixed, 20, and 20 mixed). Mixed case represents 8 pins fixed at 8 w/o, and others with variable concentration 2. Gd2O3 w/o concentration for each fresh fuel position (0, 2, 3, 4, 5, 6, 7, and 8 w/o)

The overall problem decision space includes all of the fresh fuel locations with their selected decision variables. Figure 6-2 shows the selection of the decision variables in the decision space. The selection process includes all fresh fuel positions in the core. There are total 88 (64) possible combinations of decision variables available for each fresh fuel position shown in Figure 6-2.

169

Figure 6-2 Selection of decision variables of number of UO2/Gd2O3 fuel pins and Gd2O3 concentration for each fresh fuel position in the reference initial core loading 6.2.1 Design Constraints

The design constraints for BP placement optimization problem are summarized as follows; 1. Maximum peak pin power should be less than 1.55 until 450 EFPD (or 12.5 GWd/MTU cycle exposure), and then it should be less than 1.45 until end of cycle of 680 EFPD (or 20.68 GWd/MTU cycle exposure) 2. U-235 enrichment of the fuel assemblies in the reference core loading should be fixed during the calculations 3. Do not insert any BP into the used fuel

170 4. Use fresh fuel locations in 1/8 core symmetry for the reference core and keep their positions fixed in the core during optimization calculations 5. Use maximum 20 U/Gd pins assembly 6. Use maximum 8 w/o Gd2O3 concentration for U/Gd fuel 7. BOC soluble boron concentration should be less than 1700 ppm at equilibrium Xe and Sm and follow symmetric pattern within a fuel

6.3 Genetic Algorithm Model


The genetic algorithm code explained in Chapter 3 was modified and updated to apply for this problem. Figure 6-3 demonstrates the genetic algorithm flow diagram and its interaction with the reactor physics code. The genetic algorithm code starts with a randomly generated initial population. The evaluation of each population member is possible with the interface code written in FORTRAN-90 programming language and shown in the dashed box. The main objective of this interface code is to convert decoded GA population member into the form of BP loading (U enrichment, number of UO2/Gd2O3 fuel pins, and Gd2O3 concentration), and to perform reactor physics calculations for determining core depletion parameters such as maximum peak pin power, maximum RPD power, BOC and EOC soluble boron concentration (ppm). Finally, core depletion parameters are used to evaluate the population members by calculating their fitness values in the GA code. Selection, cross-over and mutation operators are used to generate next generation population members. These genetic operators and their use and types in the code were defined before in Chapter 3. The convergence criterion of the GA code was established by assigning a maximum generation number. The solutions are

171 attained when the maximum generation number is reached. Data storage was also performed during calculations and the good solutions were archived into a file to study the results.

172

START

Interface code Create Population

i=1
Evaluate Population Decode genotype and create BP map Prepare SIMULATE-3 Input

No
Do More Generation?

Core Parameters: Gd amount Max Peak Pin Power Max. RPD Power BOC SOB ppm EOC SOB ppm

Reactor Physics Calculations (SIMULATE-3)

i=i+1

yes

Selection Operator

END
Cross Over Operator

Mutation Operator

Figure 6-3 Genetic algorithm flow diagram and interaction with the reactor physics code

173

6.4 Solution Representation, Phenotype and Genotype Structure


Binary GA requires the representation of decision variables in binary strings [43]. The decision variables, the number of UO2/Gd2O3 fuel pins and the Gd2O3 concentrations and their possible values are shown in Figure 6-4. These decision variables are assigned integer number based on their values also shown in Figure 6-4. The main reason for assigning integer numbers is due to the manufacturer design preferences for the number of U/Gd pins and Gd2O3 concentrations within a fuel assembly. Non-integer values can also be used by defining upper and lower bound of decision variables, and the decision variables are mapped uniformly to 2n (n=number of bits) discrete variables, which may not match the manufacturer practical designs exactly.

The next step converts the integer numbers into binary bits using 3 bit binary representation. The main objective of this conversion is to reduce the genotype length, which results with less CPU time during the calculations. Three bit binary representation of a decision variable has total 8 or (23) different gene values each referring to different solutions as shown in Figure 6-4. Any possible bias in the calculations is eliminated with this representation, because the number of decision variables is equal to a possible number of gene values.

Classical binary representation [16] requires at least 5 bit binary bits to include the maximum value of decision variable for this problem. It means that the maximum UO2/Gd2O3 fuel pin number, which is 20 in this case, is directly transferred to the GA code without using any conversion to integer number. Five bit representation has total of

174 32 (25) different gene values or solutions and that would increase the problem size by a factor of 40. In that case, the length of genotype (encoded UO2/Gd2O3 fuel pins number of BP loading pattern) would increase from 30 bits (10 gene3 bit/gene) to 50 bits (10 gene5 bit/gene). Thus, the GA code would take much more CPU time to achieve good results due to the problem size. The optimization of UO2/Gd2O3 fuel pins configurations were explained and presented in Chapter 3. The global GA code runs were performed with the available reference designs first, and then with the optimal designs for comparison purposes.

The phenotype represents the original BP loading pattern as applied to the reactor physics code, and the genotype is the bit-string representation and formulation of the phenotype for the genetic algorithm operations. A gene is composed of bits, each of which may exhibit 0 or 1 called an allele. The conversion process of phenotype (BP loading pattern) to genotype (binary form of BP loading pattern) is called encoding, and the reverse process is named decoding.

Figure 6-5 shows the conversion to integer number and the encoding process of the decision variable of Gd2O3 concentrations. Three bit binary representation is performed to cover all the possible decision variables. The total of 8 different decision variables of Gd2O3 concentrations are represented with 8 different 3 bit binary genes as shown in the figure.

175 Both a sample phenotype and corresponding genotype structure are shown in Figure 6-6 for a sample BP loading. This figure shows how the phenotype of integer numbers is encoded into a genotype structure for the sample BP loading. The total of 10 fresh fuel positions in the reference core loading pattern requires using a total of 20 decision variables. The genotype has 20 genes each represented by a 3 bit binary gene. The total of 60 bits representation is required to encode the decision variables representative of the number of UO2/Gd2O3 fuel pins and their corresponding Gd2O3 concentrations. The GA code alters the population of BP patterns into new BP patterns by various selection methods. Each of BP patterns in a population is then decoded into phenotype and evaluated by a reactor physics code to determine its fitness value. By choosing high fitness values for the next generation, the GA continues to improve the BP pattern design until the optimum is attained.

A phenotype and a Genotype structure are summarizes as follows: Phenotype = [|GdNo1| |GdNo2| |GdNo3|..|GdNo10| |Gdwo1| |Gdwo2|

|Gdwo3|..|Gdwo10|]| and, Genotype = [|3 bit||3 bit|.|3 bit| |3 bit| |3 bit| |3 bit| |3 bit| |3 bit| ] Genotype length = [320 bit] = 60 bit

Table 6-2 presents the detailed encoding process, phenotype, and genotype representation of a sample BP loading pattern. The GAs code performs encoding process and creates genes in binary bits as shown in the table. The combination of genes values

176 for each position creates the genotype structure for this BP loading pattern. Sample BP map encoding process and genotype structures for number of UO2/Gd2O3 fuel pins and its Gd2O3 concentrations of fresh fuel positions are shown in Figures 6-7 and 6-8.

177
Decision Variable: No of UO2/Gd2O3

Integer Coding

3 Bit Binary Genes 0 Encoding to binary numbers 0 0

0 Convert to integer number 4

12

16

16*

20

20*

Figure 6-4 Conversion to integer number and encoding process for decision variable of number of UO2/Gd2O3 fuel pins (* mixed concentrations)

178 Integer Coding

Decision Variable: Gd2O3 concentration

3 Bit Binary Genes 0 Encoding to binary numbers 0 0

0 Convert to integer number 2

Figure 6-5 Conversion to integer number and encoding process for decision variable of Gd2O3 concentration

179

Figure 6-6 Sample encoding and decoding process for global BP optimization problem Table 6-2 Sample BP map encoding process

Fresh No of Gd2O3 Fuel (w/o) Position U/Gd pins No (Phenotype) (Phenotype) in the core loading 1 H-09 8 16* 2 H-13 7 16 3 K-10 5 12 4 K-12 5 12 5 K-14 6 16* 6 L-13 6 4 7 L-14 7 16 8 M-11 3 12 9 M-13 6 8 10 N-13 6 16* * Mixed concentrations (8 pins fixed at 8 w/o)

Integer Coding No of U/Gd pins 6 5 4 4 6 2 5 4 3 6

Integer Coding Gd2O3 (w/o) 8 7 5 5 6 6 7 3 6 6

Genotype Structure Binary Binary Form of Form of Genes Genes 101 110 100 110 011 100 011 100 101 101 001 101 100 110 011 010 010 101 101 101

180
Position

Decision Variable: No of UO2/Gd2O3 Phenotype 16*

Integer Code Decision Variable

3 Bit Binary Genes 1 0 1

H-09

6
Convert to Integer number Encoding to binary genotype

H-13

16

K-10

12

K-12

12

K-14

16*

L-13

L-14

16

M-11

12

M-13

N-13

16*

Figure 6-7 Sample BP map encoding process and genotype structure for number of UO2/Gd2O3 fuel pins of fresh fuel positions (* mixed concentrations)

181
Position
Decision Variable: Gd2O3 concentration

Integer Coded Decision Variable

3 Bit Binary Genes 1 1 1

H-09

8 Convert to integer number

8 Encoding to binary genotype 7

H-13

K-10

K-12

K-14

L-13

L-14

M-11

M-13

N-13

Figure 6-8 Sample BP map encoding process and genotype structure for Gd2O3 concentrations of fresh fuel positions

182

6.5 Mutation Probability Analysis


Sensitivity analysis was performed to study the behavior of the best and average fitness values profiles during GA evaluation, and the automated mutation probability routine was developed and implemented into the GA code for the following reasons. It can be seen from Figure 6-9 that high mutation probability provides diversity in the solution space, and it increases the number of individuals, which violate the BOC soluble boron constraint. The high negative penalty (-10000) on its fitness value is defined to remove this individual from the next generation. The main principle of GA code is the survival of fittest. The individuals with higher fitness value have higher probability to be selected for next generation. High mutation probability distributes the individuals in wide spectrum. Due to higher probability of creating individuals, which violate BOC soluble boron constraint value, the absolute average fitness value increases with the mutation probability increases as shown in Figure 6-9. Low mutation probability can not provide diversity and it may result in increased probability of finding local optimal solutions. However, the best fitness value also changes during the GA evaluation as shown in Figure 6-10. In that case, ratio of average fitness to best fitness value becomes important to decide whether we should increase or decrease the mutation probability more to provide diversity in the solution. The solution space expands for high mutation probability which is good for us, but it behaves like a random search and causes the loss of efficiency (enormous solution time) for a GA run. Base mutation probability equals to 1/Npopsize and then the ratio of these two fitness values is calculated as shown in Figure 6-11. The ratio is greater than 5 for high mutation probability (10/Npopsize), and it starts to behave like random search process. However, it is between 2 and 3 for base mutation

183 probability. If it is less than 2, the solution starts to lose diversity, and the probability of finding local optimum results becomes higher. The new mutation probability routine automatically increases or decreases mutation probability 5 times to keep the ratio of average fitness to best fitness value between 2 and 5 in every 20 generations. Figure 6-12 shows the effect of mutation probability increase or decrease during GA evaluation. It is observed that while the mutation probability increases ten times, the fitness starts to diversify in the solution fitness space (high mutation zone), and then the mutation probability is reduced ten times depending on the ratio of average to best fitness values for eliminating random search start in the calculations.

Comparison of Population Individuals fitness value for high (10/Npopsize) and low (1/Npopsize) mutation probability 0 0 20 40 60 80 100 120 140 160

-2000

-4000 Fitness

-6000

-8000

-10000 Low Mutation High Mutation -12000 Individual No

Figure 6-9 Comparison of fitness value for high and low mutation probability

184

Best Fitness and Average Fitness change with generation for low mutation probability (Pmutate=1/Npopsize) and high mutation probability (Pmutate=10/Npopsize) 70 0.00E+00 -1.00E+03 -2.00E+03 Low Mutation Probability -3.00E+03 Fitness Value -4.00E+03 -5.00E+03 -6.00E+03 -7.00E+03 -8.00E+03 -9.00E+03 -1.00E+04 Generation No Avg.Fitness Best Fitness mutation boundary High Mutation Probability 80 90 100 110 120 130 140 150

Figure 6-10 Comparison of average and best fitness value for low and high mutation probability
Average Fitness to Best Fitness ratio change with generation for low mutation probability (Pmutate=1/Npopsize) and high mutation probability (Pmutate=10/Npopsize) 6.0

5.0

4.0 Ratio

3.0

2.0 Low Mutation Probability 1.0 High Mutation Probability

0.0 70 80 90 100 110 Generation No 120 130 140 150

Figure 6-11 Ratio of average fitness to best fitness value for low and high mutation probability

185

GA run with available fuel assembly design ( BOC SOB constraint=1700 ppm ) Pmutationlow=1/Npop, Pmutationhigh=10/Npop 0 0
Low mutation region
High mutation

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

Low mutation

-2000

-4000
Fitness Value

-6000

-8000

-10000

-12000

Evaluation No

Figure 6-12 Effect of mutation probability change during GA evaluation

6.6 GA Input Data


Table 6-3 summarizes the sample GA input data and operators type used for the global BP optimization problem.
Table 6-3 Genetic algorithm input data and operators type

Parameter Population Size Base Mutation Probability Maximum Generation Number Cross Over Probability Selection Operator Cross Over Operator

Data 150 1/Npop=1/150 100 0.5 Tournament selection with elitism technique Uniform Cross Over

186

6.7 Test Runs


The main objective of this section was to test and to verify the GA code performance with different objective function definitions. The general procedure for testing the performance and accuracy of any code is to create sample problems with known solutions and to perform test runs for the verification process.

Two test run cases were developed. The first case used the objective function of minimizing the maximum peak pin power during core depletion, and the second test case minimized the total Gd amount without any constraint on BOC soluble boron concentration. GA should improve the individual fitness in both cases during evaluations. The expected solutions for test problems were minimum peak pin power value and minimum total Gd amount, which was zero without any BOC soluble boron and power constraints. If there is no constraint on BOC soluble boron concentration and maximum peak pin power value, GA should develop a BP pattern, which has zero Gd amount in the core. The detailed explanations of two test cases are presented in the following sections.
6.7.1 Minimizing Maximum Peak Pin Power during Core Depletion

The main objective of this sensitivity study was to demonstrate that the GA code functions properly and achieves good solutions with the objective function of minimizing maximum peak pin power during the depletion. A total of 100 generations were performed with the GA code. The maximum peak pin power was 1.482 for BP loading pattern, which is very low compared to the limit value of 1.55. Figure 6-13 shows how the fitness value changes during evaluations and reaches its maximum at the end. The lower fitness value represents the higher maximum peak pin power value or vice versa.

187 Maximum peak pin power value during the evaluation is presented in Figure 6-14, and it starts around 1.80 and develops far beyond the maximum peak power constraint value of 1.55. The combination of good solutions can be generated with GA code as shown in the Figure 6-14. Fitness of the individuals was defined as following: Fitness function = w1P w1= weight coefficient = -100 (- sign shows maximization problem) P= Maximum peak pin power value in core depletion calculations

Figure 6-15 shows maximum peak pin power versus Gd amount during evaluation. No constraints on BOC soluble boron concentration or maximum peak pin power were used for test purpose during GA run. It can be observed from Figure 6-16, the individual with the minimum maximum peak pin power value does not necessarily have the minimum Gd amount in the core. Figure 6-16 shows typical GA run output for best fitness value change during evaluations. It shows the best fitness profile during evaluation.

188

Single-Objective Fitness Value (w1=-10, w2=0) (Minimize maximum pin power) 0 -145 -150 -155
FitnessValue

750

1500 2250 3000 3750 4500 5250 6000 6750 7500 8250 9000 9750

-160 -165 -170 -175 -180 No of Evaluation

Figure 6-13 Fitness value during GA evaluation

Maximum Pin Power Value during Evolution for Cycle 16 BP Optimization Problem with available FA designs (Npop=150, Pmutation=1/Npop) 1.8 1.775 Maximum Pin Power during depletion 1.75 1.725 1.7 1.675 1.65 1.625 1.6 1.575 1.55 1.525 1.5 1.475 0 750 1500 2250 3000 3750 4500 5250 6000 6750 7500 8250 9000 9750 No of Evaluation

Figure 6-14 Maximum peak pin power profiles during evaluation

189

Pin Power Versus Gd amount (Single Objective of Minimizing Peak Pin power) 1.85 1.8 1.75 Max. Peak Pin Power 1.7 1.65 1.6 1.55 1.5 1.45 1.4 0 200 400 600 800 1000 1200 1400 1600 Gd Amount (SUM) Pin Power Limit Value

Figure 6-15 Maximum peak pin power versus Gd amount during evaluation

Genetic Algorithm Results for Core BP optimization problem with available FA designs (Best Fitness Value vs No of Generation)
0 -1.46E+02 -1.48E+02 -1.50E+02 -1.52E+02 -1.54E+02 -1.56E+02 Best Fitness -1.58E+02 -1.60E+02 -1.62E+02 -1.64E+02 -1.66E+02 -1.68E+02 No of Generation 10 20 30 40 50 60 70 80 90 100

Best Fitness Value

Figure 6-16 Best fitness value change with generation number

190
6.7.2 Minimizing Total Gd Amount in the Core without any Constraint

The main purpose of this sensitivity study was to demonstrate that GA code functions properly and develops good solutions with the objective function of minimizing total Gd amount in the core without BOC soluble boron concentration and maximum peak pin power constraints. The solution space was divided into regions according to our expectations and the reference core data. Each region has different fitness function definition to guide the GA solutions towards region 1 (Best solution region) and region 2 (Acceptable solution region) as shown in Figure 6-17. High negative penalty definition in the fitness function was used by considering region number and reference Gd amount for an individual when its maximum peak pin power value exceeds the limit values 1.55 and 1.80 shown in Figure 6-17. GA solutions was forced to move from right to left in the solution space to minimize Gd amount in each region.

The total amount of Gd (SUM) in the core, which is a function of the number of U/Gd fuel pins and Gd2O3 concentration, is calculated by the following formula in each region.

SUM=

N Gdw
i =1 i

i=no of fresh fuel position in octant core symmetry

The fitness definitions given below were used in the regions. These definitions force the GA code to minimize the Gd amount in each region and moves from right to left in the solution space as shown in Figure 6-18. i=Region number

191 Region 1: Fitness=f1= -iSUM Region 2: Fitness=f2= -(1-i)(sumref+i)-SUM Region 3: Fitness=f3= -(1-i)(sumref+i)-SUM Region 4: Fitness= f4= -(1-i)(sumref+i)-SUM This definition guarantees that f1>f2>f3>f4. sumref is 640 in the calculations.

Fitness value profile during evaluation is shown in Figure 6-19, and Region 3 and Region 4 fitness values can be seen in this figure. In the first section, BOC soluble boron limit values was not taken into consideration and it was observed that solutions always moved from right to left in the solution space. This was not the deserved situation for this problem. The optimization process never achieves to find an individual, which lays down in region 1 or region 2. The process stayed in region 3 on the left due to the objective of minimizing Gd amount. The Gd amount (SUM) value was reduced to zero (analytical solution for the above expression) in 100 generations, which proves that GA code functions properly. However, the maximum pin power value exceeds the limit value and BOC soluble boron concentration exceeds the limit value.

Three conclusions could be made from this test run. First one is to use BOC soluble boron concentration limit for limiting the Gd amount in the solution space. There is a strong relation between soluble boron concentration and total Gd amount in the core because of the moderator temperature coefficient requirement. Soluble boron concentration also represents the Gd amount in the core. More Gd reduces the soluble

192 boron concentration, which is necessary to keep the reactor critical and moderator temperature coefficient negative. The second conclusion is the necessity to provide diversity in the solution space. This can be done by increasing mutation probability in the GA code. This will reduce the possibility of being trapped at local minimum in the solution space. However, there is a trade-off between using higher mutation probability and efficiency of genetic algorithm. The final conclusion is to force GA results moving from top to bottom of the solution space by using maximum peak power value in the fitness definition. There is need for a technique to force GA individuals to move from upper regions to lower regions. This technique would provide the GA process with capability of developing good solutions in Region 1 and Region 2 below the maximum peak in power constraint value.

Region 4: High Penalty Remove individual from the pool Pupper=1.80

Region 5 Violate SOB constraint


Plimit=1.55

Region 3: Use reactor physics code to evaluate individual (violate constraints)

SOB constraint

Region 1: Good Solutions

Region 2= Acceptable Solutions

SUMref=640

Gd amount (SUM)

Figure 6-17 Solution space for core BP optimization problem

193

Fuzzy Logic GA run (Pmutation=1/Npop) 2 1.95 1.9 1.85 Maximum Pin Power 1.8 1.75 1.7 1.65 1.6 1.55 1.5 1.45 1.4 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Gd amount (SUM) Region 2 Region 1 Region 3 Region 4

Figure 6-18 Maximum peak pin power versus Gd amount in the solution space

Fitness Value Change by evaluation No 0 0 -500 Region 3 -1000


Fitness Value

500

1000

1500

2000

2500

3000

3500

4000

4500

-1500 -2000 -2500 Region 4 -3000 -3500 Evaluation No

Figure 6-19 Fitness value profile during evaluation

194

6.8 Objective Function Definition and Constraints Handling


The solution space was divided into three regions based on the BOC soluble boron concentration (1700 ppm) and maximum peak pin power (1.55) constraints as shown in Figure 6-20. Each region used different objective functions. The good solutions (region 1) did not violate any of the constraints, and the fitness of a solution was represented with its total Gd amounts. The region 2 represented the population members having maximum peak pin power value greater than the limit value of 1.55, and BOC soluble boron concentration less than the constraint value of 1700 ppm, and the fitness of an individual was calculated by using its maximum peak pin power value with the objective of minimizing maximum peak pin power value in the core depletion. The solutions violating BOC soluble boron limit (region 3) were assigned with high negative penalty to reduce the probability of being selected for the next generation from the population.

The main objective of this definition was to promote the GA process developing solutions in the region 1 (good solutions region). The objective function of region 2 was to minimize the maximum peak pin power value during evaluation. The best solution improved in each generation, and later it reached the maximum peak pin power limit value. The best fitness value of the generation moved from Region 2 to Region 1. Then, the objective function was changed to minimize the Gd amount in the core by keeping the individuals below the defined constraint of BOC soluble boron. With this approach genetic algorithm fitness outputs was forced to move from top to bottom first by minimizing maximum peak pin power value in the region 2, and then it followed the

195 route from right to left in the solution space to minimize the Gd amount as shown in Figure 6-20.

Negative high fitness penalty (-10000) was defined in region 3 for the individuals which violated the design constraints of BOC soluble boron concentration (1700 ppm) and the maximum peak pin power during the calculations. The main objective of this constraint handling technique was to remove these individuals from the pool by reducing their probability of being selected for the next generation due to their low fitness value. The BOC soluble boron constraint limits the minimum Gd amount in the core during the calculations represented with a dashed line as the boundary of the region 3 in Figure 620. This relation results from the moderator temperature requirement for PWRs. This region is shown more clearly in the fitness space later in the results section.

Max. Peak Pin Power

Region 3 Violate SOB constraint

Region 2: Use reactor physics code to evaluate individual (violate peak pin power constraint) Objective: Minimize peak pin power SOB constraint High Penalty Remove individual from the pool

Solution moves from top to bottom

Plimit=1.55 Region 1: Good Solutions Objective: Minimize Gd amount

Solution moves from right to left

Gd amount (SUM)
Figure 6-20 Solution space demonstration and its objective functions

196

The total amount of Gd (SUM) in the core, which is a function of number of U/Gd pins and Gd2O3 concentration, is calculated by the following formula in region 1. SUM=

N Gdw
i =1 i

i=no of fresh fuel position in octant core symmetry

SUM= dimensionless fitness value in Region 1 representing total Gd amount in the octant core Ni=No of U/Gd pin for ith fresh FA position in the core loading Gdwi=Gd2O3 concentration for ith fresh FA in the core loading The following fitness definitions were used in the regions; i=Region number
Region 1: Fitness=f1= -iSUM (minimizing Gd amount) Region 2: Fitness=f2= -iw2maxpeakpinpower (minimizing maximum peak pin

power)
Region 3: Fitness=f3= -10000 (Negative high penalty for violating BOC SOB

constraint and cycle exposure constraint on where the maximum peak pin power occurs) This definition guarantees that f1>f2>f3 in Figure 6-20. w2=640 (coefficient guarantee that Region 2 has always smaller fitness value than the Region 1)

6.9 Results
This section summarizes the core BP optimization with GA calculations for different UO2/Gd2O3 fuel pins configurations used in all of fresh fuel assemblies. The first section presents the GA results using the vendors reference U/Gd fuel pin

197 configurations while the second part summarizes the GA results for calculations performed using optimal UO2/Gd2O3 fuel pin configurations developed in Chapter 3.
6.9.1 Core BP Optimization with Reference UO2/Gd2O3 FA BP Designs

A final GA run was performed for core BP optimization problem and the results are presented in Figures 6-21, 6-22, 6-23, 6-24, 6-25, 6-26, and 6-27.

Figure 6-21 shows the maximum peak pin power as a function of the total Gd amount during evaluation. The GA is designed to have the individuals in the solution space move from top to bottom in region 2 and from right to left in region 1. Figure 6-22 shows how the fitness values of the individuals change during evaluation. A high penalty (-10000) is used to remove the individuals violating the BOC soluble boron constraint from the next generation. There are three different fitness functions to represent the three regions as shown in Figure 6-22. The region 1 individuals have greater fitness value than the others. Figure 6-23 shows BOC soluble boron as a function of total Gd amount (SUM) during evaluation, and it displays the relation between the total Gd amount and the BOC soluble boron concentration. BOC soluble boron concentration should be less than 1700 ppm to satisfy the moderator temperature coefficient (MTC) requirement for PWRs.

As the total Gd amount increases, the BOC soluble boron concentration decreases as shown in the Figure 6-23. The BOC soluble boron (ppm) varies between around 2500 ppm and 1350 ppm during the GA evaluations. The BOC soluble boron change during evaluation is shown in Figure 6-24. Figure 6-25 displays maximum pin power as a

198 function of BOC soluble boron concentration during GA evaluations and Figure 6-26 shows the good solutions developed during GA evaluation. The good solutions do not violate any of the design constraints. The GA optimization process enters region 1 at the 25th generation as shown in Figure 6-27.
Maximum Peak Pin Power vs Gd amount during evaluation Genetic Algorithm run with available U/Gd FA designs 2 1.95 1.9
Maximum Peak Pin Power

1.85 1.8 1.75 1.7 1.65 1.6 1.55 1.5 1.45 1.4 0 200 400 600 800 1000 1200
Region 1

Region 2

minimize max. pin power

minimize Gd amount

1400

1600

1800

2000

Gd amount (SUM)

Figure 6-21 Maximum peak pin power versus Gd amount during evaluation

199

Fitness Value versus evaluation no Genetic Algorithm run with available U/Gd FA designs 0

Region 1

-2000

-4000
Fitness Value

Region 2 -6000 Region 3 Violates BOC SOB constraint

-8000

-10000

-12000 0 2000 4000 6000 8000 Evaluation No 10000 12000 14000 16000

Figure 6-22 Fitness value change during evaluation


BOC SOB Concentration vs Gd amount during evaluation Genetic Algorithm run with available U/Gd FA designs 2600 2400 2200 BOC SOB (ppm) 2000 1800 1600 1400 1200 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Gd amount (SUM)

BOC SOB Constraint

Figure 6-23 BOC soluble boron versus total Gd amount (SUM) during evaluation

200

BOC SOB change during evaluation Genetic Algorithm run with available U/Gd FA designs 2700 2500 2300 BOC SOB (ppm) 2100 1900 1700 1500 1300 0 2000 4000 6000 8000 Evaluation No 10000 12000 14000 16000

Figure 6-24 BOC soluble boron change during evaluation


Max Peak Pin Power vs BOC SOB concentration Genetic Algorithm run with available U/Gd FA designs

4.5 4 Max Peak Pin Power 3.5 3 2.5 2 1.5 1 0 500 1000 1500 BOC SOB (ppm) 2000 2500 3000 Peak Pin Power Constraint

good solutions

BOC SOB Constraint

Figure 6-25 Max. peak pin power vs. BOC soluble boron concentration

201

Max Peak Pin Power vs BOC SOB concentration Genetic Algorithm run with available U/Gd FA designs

1.55

1.54 Max Peak Pin Power

Peak Pin Power Constraint good solutions

1.53

1.52 BOC SOB Constraint 1.51

1.5 0 500 1000 1500 BOC SOB (ppm) 2000 2500 3000

Figure 6-26 Good solutions within maximum peak pin power and BOC soluble boron constraints
Best Fitness Value vs Generation No
0.00E+00
Region 2 Minimize Max Pin Power Region 1 Minimize Gd amount

-5.00E+02 Best Fitness Value

-1.00E+03

-1.50E+03

-2.00E+03

-2.50E+03 0 10 20 30 40 50 Generation No 60 70 80 90 100

Figure 6-27 Best fitness value vs. generation number

202 Table 6-4 presents the best solution and core BP map for TMI-1 reference core loading pattern. The developed BP map uses the reference UO2/Gd2O3 FA BP designs. Table 6-4 summarizes how many numbers of U/Gd pins and how much Gd2O3 concentrations are needed to achieve BP design having less Gd and do not violate any of the design constraints. The reactor physics calculation results are summarized in Table 65 for the best design, and it has maximum peak pin power value of 1.53 taking place at 0 GWd/MTU cycle burnup, maximum RPD power value of 1.401, BOC soluble boron concentration value of 1632.8 ppm, which is less than 1700 ppm, EOC soluble boron value of 94.4 ppm, and the total Gd (SUM) value of 1020. Table 6-6 summarizes the maximum peak pin power, the maximum RPD power, the BOC and EOC soluble boron values, the total Gd (SUM), and the fitness values for first ten good solutions. The maximum peak pin power is evaluated during the core depletion and the cycle burnup point where this occurs is also shown in Table 6-6. Furthermore, Tables 6-7 and 6-8 summarize the number of U/Gd pins in the core and their corresponding Gd2O3 concentrations for the good designs.

Figure 6-28 shows the total Gd (SUM) as a function of EOC soluble boron concentration for the first 10 good designs. The best design has 94.4 ppm EOC soluble boron concentration, and it has the minimum total Gd (SUM) value of 1020. There is a 4 ppm difference in the EOC SOB concentration values between the number 1 and 10 designs given in Table 6-8. This difference represents 1.8 EFPD savings in cycle length. Figure 6-29 shows maximum peak pin power as a function of BOC soluble boron concentration for the first 10 good designs. Figure 6-30 shows how the EOC soluble

203 boron concentration changes with the total Gd during the GA run for the good solutions. It can be observed that the EOC SOB changes from 66.6 to 95.8 ppm while the total Gd changes from 1020 to 1300. It is important to note that the difference of 29.2 ppm between the best and the worst solution in the archive was estimated to represent the potential of 13.1 EPFD savings in cycle length. The innovation is the capability to search all of the available and feasible vendors U/Gd fuel assembly designs. Since, the U/Gd fuel pin positions are limited to those used by the vendor in the reference fuel assembly designs, the number of U/Gd fuel pins and Gd2O3 concentrations can only be varied for each U/Gd fuel pin configuration.
Table 6-4 Best solution and core global BP map for the reference TMI-1 core with available vendor reference fuel assembly design

Position in the core H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 M-13 N-13

No of U/Gd pins 20* 20* 20 16 20 20 0 16 20 8

Gd2O3 w/o 7 7 6 4 7 6 0 8 6 3

Mixed concentrations (8 pins fixed at 8 w/o, and others with variable concentration)

204
Table 6-5 Best individual in GA solutions

Parameter Max. Peak Pin Power Max. RPD power Cycle Exposure at maximum peak pin power (GWd/MTU) BOC SOB (ppm) EOC SOB (ppm) Gd amount (SUM) Fitness

Value 1.53 1.401 0 1632.8 94.4 1020 -1020

Table 6-6 Good designs data from archived solutions

Design No

Max Peak Pin Power 1.53 1.547 1.535 1.543 1.546 1.545 1.535 1.539 1.544 1.547

Max RPD power 1.401 1.391 1.399 1.371 1.396 1.419 1.376 1.392 1.432 1.386

1 2 3 4 5 6 7 8 9 10

Cycle Exposure at maximum peak pin power (GWd/MTU) 0 10 10 0 10 0 1 0.8 0 0

BOC SOB (ppm) 1632.8 1624.3 1607.7 1579.6 1625.1 1646.1 1608.1 1608.9 1611.1 1574.9

EOC SOB (ppm) 94.4 91.9 94.1 95.8 91.8 93.1 88.9 91.5 89.8 90.4

Gd amount (SUM) 1020 1024 1032 1036 1036 1040 1052 1056 1068 1072

Fitness

-1020 -1024 -1032 -1036 -1036 -1040 -1052 -1056 -1068 -1072

205
Table 6-7 Good designs and no of U/Gd pins in the fresh fuel positions

No Of U/Gd pins in fresh fuel positions Design No H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 1 20* 20* 20 16 20 20 0 16 2 20* 20* 20 16 20 20 0 16 3 20* 20* 20* 20 8 20* 4 17 4 20 20* 20 20 20 20* 4 16 5 20* 20 20 16 20* 20* 0 17 6 20 20* 20* 16 8 20* 4 16 7 20 20 20 16 20 20 0 16 8 20 20 20 16 20 20* 0 16 9 20* 20 20* 16 20 20 0 16 10 20* 12 20 20 20 20* 4 20* * : Mixed concentrations case (8 pins fixed at 8 w/o)

M-13 20 20 20* 20 20* 20* 20 20* 20 17

N-13 8 8 8 8 8 8 8 8 8 12

Table 6-8 Good designs and Gd2O3 concentrations (w/o) in fresh fuel positions from the archived solution

Design No 1 2 3 4 5 6 7 8 9 10

H-09 7 7 4 7 8 8 7 7 7 8

H-13 7 7 7 8 7 7 7 7 8 7

Gd2O3 concentration (w/o) in fresh fuel positions K-10 K-12 K-14 L-13 L-14 M-11 6 4 7 6 0 8 6 4 7 8 0 5 7 5 8 6 7 3 6 5 4 6 7 7 6 4 6 5 0 5 5 5 8 5 6 5 7 6 6 6 0 5 7 6 5 6 0 7 8 4 7 8 0 4 6 8 5 8 3 4

M-13 6 5 6 6 6 6 8 4 6 6

N-13 3 7 4 4 3 8 7 7 6 3

206

10 Good Designs from the archieved solutions 1080 1 1070 1060 1050 1040 1030 1020 1010 88 89 90 91 92 93 94 95 96 97 EOC SOB (ppm) 2 3 Total Gd amount (SUM) 4 5 6 7 8 9 10

Figure 6-28 Total Gd amount (SUM) vs EOC soluble boron concentration for first 10 good designs
10 Good Designs from the archieved solutions 1.55 10 1.545 9 Max. Peak Pin Power 4 1.54 8 2 5 6 1 2 3 4 5 1.535 3 7 6 7 8 1.53 1 9 10

1.525 1570

1580

1590

1600

1610

1620

1630

1640

1650

BOC SOB (ppm)

Figure 6-29 Maximum peak pin power vs. BOC soluble boron concentration for first 10 good designs

207

EOC SOB (ppm) vs Gd amount for good solutions 1350 1300 1250 Gd amount (SUM) 1200 1150 1100 1050 1000 950 900 60 65 70 75 80 EOC SOB (ppm) 85 90 95 100

Figure 6-30 Gd amount vs. EOC soluble boron concentration for good designs 6.9.2 Core BP Optimization with Good UO2/Gd2O3 BP Designs

The GA run was performed for core BP optimization problem with UO2/Gd2O3 optimal BP designs. The UO2/Gd2O3 BP designs used the configurations developed in Chapter 3. The results are presented in Figures 6-31, 6-32, 6-33, 6-34, 6-35, 6-36, and 637.

Figure 6-31 shows the maximum peak pin power as a function of the total Gd during evaluation. The GA is designed to have the individuals in the solution space move from top to bottom in region 2 and from right to left in region 1. Fitness value change during evaluation is shown in Figure 6-32. A high negative penalty (-10000) fitness is used to remove the individuals violating BOC soluble boron constraint from the next

208 generation. There are three different fitness functions to represent three regions shown in Figure 6-32. The region 1 individuals have greater fitness value than the others.

Figure 6-33 shows BOC soluble boron versus total Gd (SUM) during evaluation, and it displays the relation between total Gd (SUM) and BOC soluble boron concentration. As the total Gd increases, the BOC soluble boron concentration decreases as shown in Figure 6-33. The BOC soluble boron (ppm) varies between around 2500 ppm and 1350 ppm during evaluation depending on Gd amount in the core. BOC soluble boron change during evaluation is shown in Figure 6-34. Figure 6-35 displays maximum pin power as a function of BOC soluble boron concentration during GA evaluations and Figure 6-36 shows the good solutions developed during GA evaluations. The good solutions do not violate the maximum peak pin power and BOC soluble boron constraints as shown in Figure 6-36. The GA run enters the region 1 at 29th generation as shown in Figure 6-37.

209

Maximum Peak Pin Power vs Gd amount during evaluation Genetic Algorithm run with with optimal BP FA designs 2 1.95 1.9 Maximum Peak Pin Power 1.85 1.8 1.75 1.7 1.65 1.6 1.55 1.5 1.45 1.4 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Gd amount (SUM) Region 1 minimize Gd amount minimize max. pin power Region 2

Figure 6-31 Maximum peak pin power versus Gd amount during evaluation

Fitness Value versus evaluation no Genetic Algorithm run with optimal BP FA designs 0

Region 1

-2000

-4000 Fitness Value Region 2 -6000 Region 3 Violates BOC SOB constraint

-8000

-10000

-12000 0 2000 4000 6000 8000 Evaluation No 10000 12000 14000 16000

Figure 6-32 Fitness value change during evaluation

210

BOC SOB Concentration vs Gd amount during evaluation Genetic Algorithm run with optimal BP FA designs 2600 2400 2200 BOC SOB (ppm) 2000 1800 1600 1400 1200 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Gd amount (SUM)

BOC SOB Constraint

Figure 6-33 BOC soluble boron versus total Gd amount (SUM) during evaluation

BOC SOB change during evaluation Genetic Algorithm run with optimal BP FA designs 2700 2500 2300 BOC SOB (ppm) 2100 1900 1700 1500 1300 0 2000 4000 6000 8000 Evaluation No 10000 12000 14000 16000

Figure 6-34 BOC soluble boron change during evaluation

211

Max Peak Pin Power vs BOC SOB concentration Genetic Algorithm run with optimal BP FA designs

4.5 4 Max Peak Pin Power 3.5 3 2.5 2 1.5 1 0 500 1000 1500 BOC SOB (ppm) 2000 2500 3000 BOC SOB Constraint Peak Pin Power Constraint

good solutions

Figure 6-35 Max. peak pin power vs. BOC soluble boron concentration during evaluation
Max Peak Pin Power vs BOC SOB concentration Genetic Algorithm run with optimal BP FA designs

1.55

1.54 Max Peak Pin Power

Peak Pin Power Constraint good solutions

1.53

1.52 BOC SOB Constraint 1.51

1.5 0 500 1000 1500 BOC SOB (ppm) 2000 2500 3000

Figure 6-36 Good solutions within max. peak pin power and SOB constraints

212

Best Fitness value change during GA evaluation 0.0

-500.0 Best Fitness Value

-1000.0
Region 2 Region 1

-1500.0

-2000.0

-2500.0 0 10 20 30 40 50 Generation No 60 70 80 90 100

Figure 6-37 Best fitness value change during evaluation

Table 6-9 presents the best solution and core BP map for TMI-1 reference core loading. The developed BP map used the optimal BP design configurations developed in Chapter 3. It summarizes how many numbers of U/Gd pins and which Gd2O3 concentrations are needed to achieve a BP design having less Gd amount and do not violate any design constraints. The reactor physics calculations results are summarized in Table 6-10 for the best design, and it has the maximum peak pin power value of 1.528 taking place at 0 GWd/MTU cycle burnup, the maximum RPD power value of 1.384, the BOC soluble boron concentration value of 1664.7 ppm (less than 1700 ppm), the EOC SOB value of 97.2 ppm, and the total Gd (SUM) value of 896.

213 History data counter is used to count the number of solutions, which lie in the solution space according to defined solution regions. Table 6-11 presents the history data during GA evaluations. An approximately 10 % of the all evaluations is in good solution region (region 1) as given in Table 6-11. Region 2 has a 75 % fraction of total evaluations, and region 3 represents a 13 % fraction of total evaluations. The fraction of solutions violating cycle exposure constraints is less than 1 %, and it is almost 0.6 %.

Table 6-12 summarizes the maximum peak pin power, the maximum RPD power, cycle exposure data where the maximum peak pin power occurs, the BOC and EOC soluble boron values, the total Gd (SUM), and the fitness values for first ten good solutions developed during GA evaluations. Table 6-13 and 6-14 summarize the good designs for number of U/Gd pins and their corresponding Gd2O3 concentrations for good designs.

Figure 6-38 shows the total Gd (SUM) versus EOC SOB concentration for the first 10 good designs. The best design has 97.2 ppm concentration, and it has the minimum total Gd (SUM) value of 896 in the results. There is a 8.1 ppm difference in EOC soluble boron concentration between the number 1 and 10 designs given in Table 612. This difference was estimated to refer 3.493 EFPD savings in cycle length between the designs no 1 and 10 in the calculations.

Figure 6-39 shows maximum peak pin power vs. BOC soluble boron concentration for first 10 good designs. Figure 6-40 presents how to EOC soluble boron

214 concentration and Gd amount changed and developed during GA run for the archived good solutions. It can bee seen from Figure 6-40 that the EOC SOB changes from 68.9 to 97.2 ppm when the total Gd changes from 896 to 1264. It is important to note that the difference of 28.3 ppm between the best and the worst solutions in the archive was estimated to represent the potential of 12.5 EPFD savings in cycle length.
Table 6-9 Best solution and core BP map for TMI-1 Reference Core with optimal BP fuel assembly designs

Position in the core H-09 H-13 K-10 K-12 K-14 L-13 L-14 M-11 M-13 N-13

No of U/Gd pins 16* 16* 20 16 20 20 0 12 16 8

Gd2O3 w/o 7 7 6 7 5 6 0 5 6 6

Mixed concentrations (8 pins fixed at 8 w/o, and others with variable concentration)

Table 6-10 Best individual in GA solutions

Parameter Max. Peak Pin Power Max. RPD power Cycle Exposure at maximum peak pin power (GWd/MTU) BOC SOB (ppm) EOC SOB (ppm) Total Gd (SUM) Fitness

Value 1.528 1.384 0 1664.7 97.2 896 -896

215
Table 6-11 History data during evaluations

Definition of individuals Region 1 Good Solutions Region 2 Violates only maximum peak pin power Region 3 Violates BOC SOB constraint Violation of Cycle Exposure Constraint Total

Number of Solutions 1561 11377 1972 90 15000

Percentage fraction (%) 10.41 75.85 13.15 0.60 100.00

Table 6-12 Good designs data from the archived solutions

Design No

Max Peak Pin Power 1.528 1.548 1.543 1.549 1.545 1.543 1.539 1.535 1.548 1.548

Max RPD power 1.384 1.378 1.378 1.387 1.394 1.388 1.392 1.37 1.399 1.38

1 2 3 4 5 6 7 8 9 10

Cycle Exposure at maximum peak pin power (GWd/MTU) 0 0.4 0 0 10 0.4 10 0 0 0

BOC SOB (ppm) 1664.7 1676.6 1659 1650.5 1647.1 1663.9 1644.7 1638.4 1640.5 1641.6

EOC SOB (ppm) 97.2 94.8 95.9 90.3 85.7 93.5 90.1 88.3 90.7 89.1

Gd amount (SUM) 896 908 916 960 964 968 968 972 976 980

Fitness

-896 -908 -916 -960 -964 -968 -968 -972 -976 -980

216
Table 6-13 Good designs and number of U/Gd pins in the fresh fuel positions from the archived solutions
Design No 1 2 3 4 5 6 7 8 9 10
*

H-09 16* 16* 16* 16* 16* 16* 16* 16* 16* 16*

H-13 16* 16* 16* 16 16* 16* 16* 16 16* 16

K-10 20 20 20 20 21 20 20 21 20 20

No Of U/Gd pins in fresh fuel positions K-12 K-14 L-13 L-14 M-11 16 20 20 0 12 16 16* 20 0 12 16 20 20 0 12 16* 21 20 0 12 16 21 20 0 12 16 16 20 0 12 16 20 20 0 12 16* 20 21 0 12 16* 20 20 0 12 16 20 20 0 12

M-13 16 16 16 16 16* 16 16* 16 16 16*

N-13 8 8 8 8 8 8 8 8 8 8

: Mixed concentrations case (8 pins fixed at 8 w/o)

Table 6-14 Good designs and Gd2O3 concentrations (w/o) in fresh fuel positions from the archived solution
Design No 1 2 3 4 5 6 7 8 9 10 H-09 7 8 7 8 8 8 8 8 6 7 H-13 7 7 7 7 8 7 7 7 7 7 Gd2O3 concentration (w/o) in fresh fuel positions K-10 K-12 K-14 L-13 L-14 M-11 6 7 5 6 0 5 6 7 5 6 0 5 6 7 5 7 0 5 6 8 5 6 0 6 5 5 6 6 0 5 7 7 7 8 0 5 7 7 6 6 0 5 4 5 5 8 0 6 7 7 6 8 0 5 7 7 6 7 0 5 M-13 6 6 6 6 6 6 6 8 6 8 N-13 6 6 6 6 7 5 7 5 6 6

217

10 Good Designs from the archieved solutions 990 980 970 Total Gd amount (SUM) 5 960 950 940 930 920 3 910 900 1 890 84 86 88 90 92 94 96 98 EOC SOB (ppm) 2 4 8 7 6 10 9 1 2 3 4 5 6 7 8 9 10

Figure 6-38 Total Gd amount (SUM) vs. EOC soluble boron concentration
10 Good Designs from the archieved solutions 1.55 4 9 10 1.545 Max. Peak Pin Power 5 3 1.54 7 6 2

1 2 3 4 5 6 7 8 9 10 1

1.535

1.53

1.525 1635

1640

1645

1650

1655

1660

1665

1670

1675

1680

BOC SOB (ppm)

Figure 6-39 Maximum peak pin power vs. BOC soluble boron concentration

218

EOC SOB (ppm) vs Gd amount for good solutions 1300 1250 1200 Total Gd amount (SUM) 1150 1100 1050 1000 950 900 850 60 65 70 75 80 EOC SOB (ppm) 85 90 95 100

Figure 6-40 Gd amount vs. EOC soluble boron concentration for good designs

6.10 Comparison of Results


Table 6.15 summarizes the core BP optimization results for different BP fuel assembly designs. The optimal BP loading design has 97.2 ppm soluble boron at EOC while the BP loading with reference FA designs has 94.4 ppm soluble boron at EOC. The difference of 2.8 ppm reflected the potential savings of 1.25 EFPD in cycle length. Moreover, the total Gd amount was reduced by 6.89 % in mass that provided extra savings in fuel cost. One should also calculate the total income due to electricity sale resulting from the potential savings in cycle length.

219
Table 6-15 Comparison of core BP optimization results for different BP designs

Parameter

BP loading with reference BP FA designs 1.53 1.401 0 1632.8 94.4 1020 -1020 125.225 -

BP loading with optimal BP fuel assembly design 1.528 1.384 0 1664.7 97.2 896 -896 116.597 6.89

Max. Peak Pin Power Max. RPD power Cycle Exposure at maximum peak pin power (GWd/MTU) BOC SOB (ppm) EOC SOB (ppm) Total Gd (SUM) Fitness Gd amount (kg) Mass Savings (%)

CHAPTER 7 CONCLUSIONS AND FUTURE WORK


7.1 Conclusions
The main outcome of this study is a unique methodology and a practical tool for designing a BP placement for a given PWR core loading using the TMI-1 as an example. First, the deterministic technique called MPSFD method followed by fine-tuning algorithm was successfully developed. Second, an efficient and a practical GA to optimize the BP placement in PWRs was developed and applied successfully to a referenceTMI-1 core.

The deterministic method presented in this study was shown to operate successfully. However, it used CASMO-3 cross section data without MICBURN libraries for gadolinium fuel that shows some differences for UO2/Gd2O3 fuel depletion calculations as compared to CASMO-4 cases. The method did work with CASMO-3 cross section data, but it needs to be applied with CASMO-4 code in the future for verification.

Two genetic algorithm techniques were developed in this study. The first GA developed the best U/Gd fuel pin configuration for only the few hot fuel assemblies in the core by keeping the number of U/Gd fuel pins and Gd2O3 concentration fixed for each run. The best U/Gd fuel pin configurations were determined by this GA for all designs

221 allowed by the manufacturer that fall within the maximum allowed number of U/Gd fuel pins and Gd2O3 concentration range. This genetic algorithm performs coupled fuel assembly and core depletion calculations to evaluate the BP fuel pin configurations for constraint violations. It successfully incorporates a Kinf filter. This is very important because of the very long computational time required for each population member evaluation in the GA process. Each coupled lattice (CASMO-4) and reactor physics (SIMULATE-3) calculation of a new GA solution takes 10 minutes of computational time for an evaluation process. Long CPU time is needed to prepare the cross section library and to perform core depletion calculations. For example, the total of 1,500 GA evaluations consumes 15, 000 CPU minutes or 10.4 CPU days. It was shown that the BOC Kinf, used as a filter can eliminate most of the new genotypes created during the GA process by assigning a high negative penalty to all genotypes that have a BOC Kinf greater than some limit. This limit is 1.065 for the TMI-1 core. In doing so, this filter allows total evaluation of all new genotypes within one day. Different Kinf filter values may be developed for different types of BPs.

The second GA placed BPs in all fresh fuel assembly locations in the core. It utilized the results of the first technique by keeping the U/Gd fuel pin configurations fixed as determined before for each of the different allowed number of U/Gd fuel pins and Gd2O3 concentrations. It then randomly varied the number of U/Gd fuel pins and Gd2O3 concentrations, using the above various fixed U/Gd fuel pin positions, in all fresh fuel positions in the core to minimize both the total Gd amount and the residual binding in the core. It produced excellent results.

222 Optimizing the BP placement in a fuel assembly in an infinite array does not give correct results for the same fuel assembly when placed in the core. In an infinite array increased BP concentrations and number of BP fuel pins increases the peak pin power however in the core increasing BP content in the fuel assembly lowers its Kinf , which also lowers peak pin power when depleted in the actual core. Hence, lower Kinf takes precedence over the result obtained in an infinite array because of the lowering of the fuel assembly power in the core. Results of this study provided a more profound understanding of how the peak pin power is affected by the different BP designs during core depletion.

Enriched Gd-155 and Gd-157 optimized BP designs were studied to further reduce residual binding of natural high (7-8 w/o) Gd2O3 concentrated U/Gd FA designs. The natural high (7-8 w/o) Gd2O3 concentrated U/Gd FA designs were successfully replaced with the optimized enriched Gd-155 designs in the reference core loading. The 100 w/o Gd-157 design provided 44 ppm more residual binding savings while the 100 w/o Gd-155 design presented only 28 ppm more residual binding savings compared to natural U/Gd fuel designs. However, the 100 w/o Gd-157 design did not keep the peak pin power below its constraint in core depletion calculations.

Sensitivity analysis on optimized designs for core depletion showed that the 100 w/o enriched Gd-155 provides best peak pin power and RPD profiles during the depletion. The deterministic BP loading pattern was used to evaluate the neutronic performances of the optimized designs in the core depletion calculations. The net savings

223 amount was calculated to be $ 53,992 for BP loading pattern developed with deterministic and heuristics techniques with current unit cost scenario. The total potential savings was projected to be ~ $ 1, 377, 000 for gross savings and ~ $ 916,000 for net savings, if all of the 68 feed assemblies would be replaced with the optimized designs. The break-even point for unit Gd-155 enrichment cost was determined to be around ~ $ 30 /gram-Gd-155.

The optimal BP fuel designs were develop by using the Kinf filter in the first technique. Then, these designs were utilized in the BP library. Next step was to further minimize the total Gd amount and residual binding by respectively utilizing vendors reference and newly developed optimal BP designs in the calculations. The GA code was modified and applied for this purpose. The GA process searched all of the feasible U/Gd fuel pin designs with variable number of U/Gd pins and concentration of Gd2O3 in core BP optimization calculations. This study provides flexibility to the core designer for evaluating any feasible and practical BP pattern for a given PWR core loading.

The GA code developed many good solutions that satisfy all of the design constraints. The total CPU time is approximately 41.9 hours on IBM RS/6000 UNIX Workstation for a sample GA run. For these solutions, the EOC soluble boron changes from 68.9 to 97.2 ppm. It was estimated that the difference of 28.3 ppm between the best and the worst solution in the good solutions region represents the potential of 12.5 EPFD savings in cycle length. As a comparison, the best BP loading design has 97.2 ppm soluble boron at EOC while the BP loading with the available vendors U/Gd FA designs

224 has 94.4 ppm soluble boron at EOC. The difference of 2.8 ppm was estimated to reflect the potential savings of 1.25 EFPD in cycle length. Moreover, the total Gd amount was reduced by 6.89 % in mass that provided extra savings in fuel cost as compared to the BP loading pattern with the available vendors U/Gd FA designs.

7.2 Summary of Contributions


The unique contributions of this PhD study are summarized as following:

A new deterministic BP loading technique, which restricts the maximum normalized power under the limit value, was successfully developed and applied to a given reference core loading pattern. The program further reduces the amount of gadolinium in the core with fine tuning algorithm. Then, it develops a BP loading pattern, which meets all of the design constraints.

An efficient and a practical genetic algorithm was developed and applied successfully to the BPs placement optimization problem for the reference TMI-1 core.

First, a new genetic algorithm methodology and technique were developed and applied to UO2/Gd2O3 fuel pin configuration optimization. The main contribution is that the BOC Kinf of a fuel assembly was determined to be used as a filter in the optimization calculations. It was used to filter out bad designs and to eliminate the need for evaluation process with reactor physics calculations of these designs. Then, the results were used to

225 develop an extended burnable absorber library, which included all of the feasible U/Gd FA designs with optimal UO2/Gd2O3 fuel pin arrangements. A user-friendly program, CASMO-4 image file generator, was developed in this study, which automatically generates all the necessary CASMO-4 input decks, and runs CASMO-4 cases by using interfaces. The total CASMO-4 calculations consist of 4912=588 runs to create extended burnable absorber library as explained in Chapter 3. This process is the basis of the optimization calculations, and it requires significant manpower. Therefore, the automation of the image file generation process is very important, and it presents great benefits to the user by savings the enormous amount of time and complexity. It is very simple to run and to create all of the image files by using the input parameters of U enrichment, number of U/Gd fuel pins, Gd2O3 concentrations, U/Gd fuel pin positions within an octant fuel assembly model.

Another contribution to the fuel management area was to prepare all of the CASMO-4 image files (more than 2 GB total capacity) and cross section data for Gd2O3 burnable absorber designs. The future studies and applications can greatly benefit from this data base and save huge man-power.

Enriched Gd-155 and Gd-157 optimized BP designs were studied to reduce residual binding of natural high Gd2O3 concentrated U/Gd FAs. Optimal UO2/Gd2O3 fuel pin arrangements were determined by using developed Genetic Algorithm code, which works with the CASMO-4/TABLES/SIMULATE-3 package. The natural high (7-8 w/o)

226 Gd2O3 concentrated U/Gd FA designs were successfully replaced with the optimized enriched Gd-155 designs in the reference core loading.

The Gd enrichment cost was included into equilibrium fuel cycle cost components and detailed fuel cost sensitivity analysis was performed for a given core loading. The potential for fuel cost savings was demonstrated utilizing optimized enriched Gd designs in the loading. The break-even point for Gd enrichment cost was determined for the reference industry unit costs.

The GA methodology was modified and applied successfully to the BP placement optimization problem in the TMI-1 core. The number of UO2/Gd2O3 fuel pins and Gd2O3 concentrations for each fresh fuel location in the core were the decision variables and minimizing the total amount of the Gd in the core is the objective function for good solution region. The innovation was to search all of the possible U/Gd fuel pin designs with variable number of U/Gd pins and concentration of Gd2O3 in core BP optimization calculations to further reduce the total Gd amount in the core.

This study presents a unique methodology and a practical tool for utilities and vendors use for PWRs fuel management. First, the study develops good BP loading patterns with optimal UO2/Gd2O3 fuel pins configurations within a fuel assembly for a given core. Then, it utilizes optimized enriched Gd designs in the core to replace the natural high Gd2O3 concentrated U/Gd FAs. Finally, it automates the fuel cost analysis for the equilibrium core with optimal BP pattern.

227

7.3 Suggestions for Future Work


This section presents the suggestions and recommendations for future work. Each future work is explained in detail in the following sections.
7.3.1 Hybrid Genetic Algorithm-Neural Network Application

The hybrid use of GAs and Neural Networks (NNs) is being studied to improve and to speed up the optimization process. The initial studies were performed and the detailed application was left for future work. Neural networks sometimes referred to as Artificial Neural Networks (ANNs), have been applied as a tool for simulating several reactor physics parameters such as RPD, the peak pin power distribution, the soluble boron concentration at the BOC and the EOC elsewhere [52- 56].

In GA calculations presented in this thesis, all solutions were forwarded to a reactor physics code for determining the fitness value of each solution regardless of how unacceptable that solution might be. The reactor physics code involves long running times for each computation (~10 sec for one SIMULATE evaluation due to 9 depletion steps used in the real core depletion calculations). The NNs could make it possible to determine soluble boron concentration at the BOC and EOC, and RPD in the core with sufficient accuracy to identify GA solutions that are truly not valid without performing the reactor physics calculations reducing the computational time by an order of magnitude or more. The total CPU time was approximately 41.9 hours on IBM RS/6000 Unix workstation for a sample GA optimization run. The NNs have potential to reduce the total CPU time by filtering out many invalid solutions that violate the constraints.

228

The NN allowed obtaining good predictions of the performance of the different BP loading patterns. Initial studies showed that NNs can predict the soluble boron concentration at BOC very accurately with the developed architecture of the network. The training of the network needs a computer time around minutes. After training and establishing the network, the NN can be used to predict the reactor core physics parameters for determining the fitness and checking the constraint violation of the individuals during evaluation process in milliseconds. The PSU experience proved that NNs are good tools to predict the RPD distribution in the core by using the Kinfs distribution as input to the network. Therefore, the NN Design can predict the RPD power distribution. However, the peak pin power value has a very complex relation with the core input data. It is proposed to use NN to predict the BOC soluble boron concentration and use it as a filter during calculations. The NN can be used to eliminate BP patterns that produce BOC soluble boron somewhat above the 1700 ppm constraint. As an example, initial results showed that by using BOC soluble boron value as a filter approximately 12 % of the total populations can be eliminated as shown in Figure 7-1.

229

Max Peak Pin Power vs BOC SOB concentration Genetic Algorithm run with available U/Gd FA designs

4.5 4 Max Peak Pin Power 3.5 3 2.5 2 1.5 1 0 500 1000 1500 BOC SOB (ppm) 2000 2500 3000 Peak Pin Power Constraint

Detailed Reactor Physics Calculations

Neural Network Estimation

good solutions

BOC SOB Constraint

Figure 7-1 Max pin power vs. BOC soluble boron concentration during evaluation

Figure 7-2 shows proposed hybrid GA-NN algorithm for a future work. Neural network routine has two basic modules for training and simulation. The training module will divide the previously generated data set including BOC Kinf distribution in the core, radial power distribution, peak pin power distribution, BOC and EOC soluble boron concentrations into three sets, which are training, validation, and test data sets. It will perform learning process by reducing the network training performance of medium square error, and then it will validate training results by using the validation data set. Validation vectors can be used to stop training early if the network performance on the validation vectors fails to improve or remains the same for maximum validation failures. Finally, test data will be used to determine the performance of the network for different input data set. The test vectors can be used as a further check that the network is generalizing well, but do not have any effect on training.

230

The simulation module will be used to predict the trained network for a given input Kinf distribution vectors. The basic reactor physics parameters of peak pin power distribution, radial power distribution, BOC and EOC soluble boron concentration will be estimated by the simulation module. Neural Network filter value (~1700 ppm) can be defined for BOC soluble boron concentration parameter to filter out the solutions violating the BOC soluble boron constraint value shown in Figure 7-1. More sensitivity studies should also be performed on the neural network filter definitions for fitness values and maximum RPD value during calculations. Furthermore, the solutions moves to good solution region during GA run, and the similarities between solutions increase. The performance of the neural network can change and it may need to be re-trained. Therefore, the training process should be performed continuously during GA run to provide good estimation within the desired accuracy.

231

START
Interface code

Create Population

Decode chromosome and create BP map

Convert BP map to kinfs map

Evaluate Population

Neural Network Routine

yes

Training Necessary?

No

No
Do More Generation?

Fitness Parameters: Gd amount Max Peak Pin Power Max. RPD Power BOC SOB ppm EOC SOB ppm

Training Module

Simulation Module

yes

Predicted Parameters Max Peak Pin Power Max. RPD Power BOC SOB ppm EOC SOB ppm

Selection Operator
Use NN filters?

yes Cross Over Operator No Reactor Physics Calculations (SIMULATE-3) Mutation Operator

END

Figure 7-2 Hybrid GA-NN methodology

232
7.3.2 Sensitivity on Fitness Function Definition

In this study, the objective function was defined to minimize the total Gd amount in the core. The total Gd amount in the core was represented as a function of the multiplication of number of U/Gd fuel pins and Gd2O3 concentration for each fresh fuel assembly position, which is proportional to the mass of Gd. The main assumption was to use same priority and weights for both parameters when calculating the total Gd amount. However, the studies on U/Gd FA designs show that the main source of residual binding at the EOC is due to non absorber Gd isotopes and their residual remains after the depletion of main neutron absorber isotopes of Gd-155 and Gd-157. The residual binding of U/Gd FA designs increase with the increase in the concentration of Gd2O3. Therefore, it may be worthwhile to use different weight coefficient for U/Gd fuel pin number and Gd2O3 concentration while calculating the fitness value. In that case, higher Gd2O3 concentration will have higher weight coefficient in the fitness function that will penalize the use of high concentration. The GA will improve the results by increasing the probability of selecting low concentrated Gd2O3 due to their high fitness. This technique may allow reducing further the residual binding that means further increase in cycle length.
7.3.3 Multi Objectives Optimization with Genetic Algorithm

More sensitivity studies on multi-objective optimization on EOC soluble boron concentration and the total Gd amount in the core should be performed as a future work. Pareto optimality concept and the definition of non-dominated set should be investigated for the two objectives of maximization of EOC soluble boron and minimization of total Gd amount in the core. Optimal set of solutions should be determined by identifying the

233 non-dominated set with using domination approach [43, 57]. The GA code used in this thesis can be modified and updated by adding the procedures for finding a non-dominated set.
7.3.4 Fuel Cost Analysis

The fuel cycle cost analysis should also include the sensitivity studies on the savings of total Gd cost and EFPD in cycle length due to better use of BP during the depletion. Equilibrium fuel cost analysis presented in this study should include UO2/Gd2O3 fabrication cost component as a separate item than the total fuel assembly fabrication cost. In this thesis, it was assumed that the PWR fuel assembly fabrication cost is the same for any type of fuel assembly design regardless of use of Gd or not. The savings on fuel cost will be presented much clear due to reduction of total Gd amount that means less fabrication cost for U/Gd fuels. Longer cycle length can be achieved by reducing the residual binding of Gd for a given core that means more energy production or cost benefits from the reference cycle.

However, this study presented a break-even point for unit cost of enriched Gd155. The main objective was to show the technical feasibility of using Gd-155 in the core depletion calculations instead of using natural enriched Gd design. More sensitivity analysis should be done on Gd isotope enrichment cost due to lack of reliable unit cost data in the literature. Gd isotopes were separated by using the plasma separation process or other methods, which were designed for small scale experimental studies. Yet, it is still not at the commercial production stage.

234
7.3.5 Parallel Genetic Algorithms

GAs using parallel computing may present an effective way to speed up the calculations. It is a promising and challenging future research area in the optimization calculations. GA source should be organized in a way to be applicable for parallel computing purposes. The population is divided into a number of equally sized subpopulations. The calculations, named as parallel genetic algorithms (PGA), can be performed by the parallel execution of a number of classic GA each operating on one of the subpopulations. This process is called coarse-grained PGA. On the other hand, in fine-grained PGAs, the population can also be mapped onto a grid and a neighborhood structure can be defined for each individual on this grid. The selection and cross over step should be restricted to the individuals in a neighborhood [58]. A parallel process can be assigned to each individual. The research on PGAs is promising and will increase the speed of calculations, and it will permit to work with large populations.
7.3.6 Different GA Operators

More sensitivity studies should be performed for using different techniques on GA operators to increase the efficiency of the algorithm. The developed heuristic rules and the results of this thesis can be used in the GA code to improve the efficiency and reduce the CPU time for developing optimum design. Integer and real coded GA application can be considered for BP pattern optimization in order to make a comparison of the efficiency of the different representations.

235
7.3.7 Simultaneous Core Loading and BP Placement Optimization

The complete fuel management optimization tool should include both fuel loading pattern and burnable poison placement calculations simultaneously in some way. Further research should be performed to combine these two problems. First one can combine fuel loading optimization package, which uses Haling depletion calculations for a core design with no BPs loaded, and the burnable poison placement optimization program given in this thesis. The second step can be carried out by inserting all of the feasible U/Gd FA design into fuel inventory. The GA code can perform the core loading pattern optimization with this inventory and can insert the U/Gd fuel pins into a fuel assembly during calculation if its power exceeds the defined limit value. The third option can be the addition of extra bits into genotype structure representing different absorber amounts for each fuel position in the core. In that case, the problem dimension can be very big, and the optimization process should be speeded up with PGAs and NNs applications.
7.3.8 GA Methodology for Different Integral Fuel Burnable Absorbers

The developed GA methodology can be applied to any other integral burnable absorber design as for example the Westinghouse IFBAs and their use in the core. The methodology presented in this thesis can be easily extended for the developing optimal uranium-erbium fuel rod configurations within fresh PWR fuel assembly and their use in the core depletion calculations. One should further investigate the use of erbium as a burnable poison in PWRs by using the objective function of minimizing its residual binding and content in a sample 24 month cycle.

236
7.3.9 Optimization of Burnable Absorber Distribution for Boling Water Reactor Fuel Assembly

The GA methodology can be further extended to determine optimal axial BP distribution for a Boiling Water Reactor (BWR) fuel assembly under control rod programming. This process includes 3-D lattice optimization calculations coupled with core depletion and the use of more than one Gd2O3 concentrations in axial direction to control the power shape and reduce the residual binding. Decision variables of the process should be the number of U/Gd fuel pins and their concentrations for each axial node. The core loading pattern and control rod movement schedule should be assumed known for a given BWR. The main design constraint for a BWR core is minimum critical power ratio. Coupled lattice, core depletion and thermal-hydraulic calculations require high computational time, which makes the solution process very demanding. The BOC Kinf filter definition should be further investigated for BWR fuel assembly design to reduce the computational time by eliminating the need for core depletion and thermal hydraulic calculations in the optimization process. The complete optimization process may require the use of parallel computing techniques to make the process more efficient.

BIBLIOGRAPHY
1. LWR Fuel-Cycle Costs as a Function of Burnup, EPRI Report NP-3788, November, (1984). 2. The Economics of the Nuclear Fuel Cycle, Annex A-9 Nuclear Fuel Costs, OECD/Nuclear Energy Agency, (1994). 3. Optimum Discharge Burnup for Nuclear Fuel: A Comprehensive Study of Duke Powers Reactors. EPRI Report TR-112571, (1999). 4. Optimum Cycle length and Discharge Burnup for Nuclear Fuel A Comprehensive Study for BWRs and PWRs, Phase 1: Results Achievable Within the 5 % Enrichment Limit, EPRI Report 1003133, (2001). 5. Optimum Cycle length and Discharge Burnup for Nuclear Fuel A Comprehensive Study for BWRs and PWRs, Phase 2: Results Achievable With Enrichments Greater than 5 w/o, EPRI Report 1003217, (2002). 6. Economic Aspects of High Burnup Fuel Options in PWRs. Xu Z., Driscoll, M., and Kazimi S. M., Advances in Nuclear Fuel Management III, Hilton Head, SC, October (2003). 7. Gregg, R., and Worral, A. , Effect of highly enriched Robust Fuel on Nuclear Design Parameters, Advances in Nuclear Fuel Management III, Hilton Head, SC, October (2003). 8. Schlieck, M., et al, Optimized gadolinia concepts for advanced in-core fuel management in PWRs, Nucl. Eng. Design, 205, 191-198, (2001). 9. Westinghouse Proprietary Class 2, Core Design Training Course, Presented to Penn State University, February, (2004). 10. Kapil, S. K., Secker, J.R , Keller, H.W., Achievements in Westinghouse Discrete and Integral Fuel Burnable Absorber Usage, ANS Trans. Vol:57 (1988). 11. Skogen, F.B, Kempf, F.J., Advanced Nuclear Fuels Corporations Achievements with Gadolinia Integral Burnable Absorber, ANS Trans. Vol:57 (1988). 12. Yokote, M. , Azumi, S., Mori, K., Kobayashi, S., Nakura, S., Mukai, M., A Japanese Gadolinia Fuel Usage and Demonstration Program, ANS Trans. Vol:57 (1988). 13. Wagner, J., and Sanders, C. E., Investigation of the effect of fixed absorbers on the reactivity of PWR spent fuel nuclear fuel for burnup credit, Nuclear Technology, Vol.139, p 91-125, August, (2002).

238

14. Li Z., An Automatic Optimal Pressurized Water Reactor Reload Design Expert System, Ph.D., A Thesis in Nuclear Engineering, The Penn State University, (1993). 15. Haibach, B., A Study of the Optimization of IFBA Using Deterministic Methods and the Genetic Algorithm Based CIGARO Fuel Management System, M.S. Thesis, The Pennsylvania State University, (1996). 16. DeChaine, M., Stochastic Fuel Management Optimization Using Genetic Algorithm and Heuristic Rules, Ph. D., Thesis in Nuclear Engineering, The Penn State University, (1995). 17. Hida, K., Optimization of Axial Enrichment and Burnable Absorber Distributions for Boiling Water Reactor Fuel Assembly, Doctoral Thesis, Tokyo Institute of Technology, November 30, (1998). 18. Guler, C., Development of VVER Core Loading Optimization System, M.S Thesis, The Penn State University, August (2002). 19. Hongchun W., Pressurized Water Reactor Reloading Optimization Using Genetic Algorithms, Annals of Nuclear Energy, 28, 1329-1341, (2001). 20. Keller P., M., FORMOSA-P Advanced GA Methodology, Advances in Nuclear Fuel Management III, Hilton Head, SC, October 5-8, (2003). 21. Keller P., M., and Turinsky P., J., FORMASA-P Multiple Adaptive Penalty Function Methodology, Trans. Am. Nucl. Soc., 75, pp. 341-342, Washington, DC (1996). 22. S. KIRKI PATRICK, C.D. GELATT et all Optimization by Simulated Annealing, Science, 220, 671 (1983). 23. G.I. Maldonado and T. Guo, Penalty-Based Constraints Applied to WithinBundle Loading Optimization, Trans. Am. Nucl. Soc.,78,235 (1998). 24. E. Klee, T., Guo, Maldonado A Coupling of Within-Bundle Loading Optimization to Lattice Physics Calculations, Proc. Topl. Advances in Nuclear Fuel Management II, Vol 2, South Carolina, (1997). 25. Lam, H., Q., Sato, D., Little, D., C., and Popa, F., D., LP-Fun Applications to Reload Design, Proc. Topl. Advances in Nuclear Fuel Management III, Hilton Head Island, South Carolina, USA, October 5-8 (2003). 26. G.T. Parks Multiobjective PWR Reload Core Design by Nondominated Genetic Algorithm Search, Nucl. Sci. Eng. 124, 178-187 (1996).

239 27. M. Schlieck et all, Optimized gadolinia concepts for advanced in-core fuel management in PWRs, Nucl. Eng. Design, 205, 191-198, (2001). 28. Sontheimer, F., Landskron, H. Billaux, M.R., A Fuel Thermal Conductivity Correlation Based on the Latest Experimental Results, Seminar Proceedings, Thermal Performance of High Burn-up LWR Fuel, , Cadarache, March (1998). 29. Spetz S.,W., Hove C.,M., Trends in Fuel Cycle and Burnable Poison Applications, Fuel Conference, Lynchburg, Virginia, May 5-7, (1986). 30. PSP: The Plasma Separation Process for Isotope Separation, S.L. Korn, et.al., Technology at TRW Electronic and Defense Sector, 6, 1 Winter, (1982/83). 31. Santala, M.I.K, Daavittila, A.S, Lauranto H.M., Salomaa, R.R.E, Odd-isotope enrichment studies of Gd by double resonance laser-ionization for the production of burnable nuclear reactor poison, Applied Physics B 64.339 347 (1997). 32. Renier, J.P.A, and Grossbeck, M.L., Development of Improved Burnable Poisons For Commercial Nuclear Power Reactors, Oak Ridge National Laboratory, Report on Phase 1 of NERI project number 99-0074, ORNL/TM2001/238, (2001). 33. Grossbeck, M.L., Renier, J.P.A, and Bigelow, T., Development of Improved Burnable Poisons For Commercial Nuclear Power Reactors, Department of Nuclear Engineering The University of Tennessee, Final Report on NERI project number 99-0074, September, (2003). 34. Bejmer, K.H, Severborn, O., Enriched Gadolinium as Burnable Absorber for PWR, PHYSOR 2004, Illinois, April 25-29, 2004, ANS, Lagrange Park, IL (2004). 35. Lee D., J., et all., Fuel Cycle Cost Sensitivity to Burnable Absorber and Feed Assemblies , Proc. Of the ANS Topical Mtg. Advances in Nuclear Fuel Management II, (1997). 36. L. Stephenson and A. Neufert, Gadolinia in Economically Competitive PWR Fuel Cycles, Proc. Of the ANS Topical Mtg. Advances in Nuclear Fuel Management II, Volume 1, p.7-17, (1997). 37. Asou, M. and Porta, J., Prospects for Poisoning Reactor Cores of the Future, Nucl. Eng. Design, 168, 261-270, (1997). 38. R. R. Cacciapouti et al, PWR Burnable Absorber Evaluation, Nucl. Energy, 34, No. 5, 303-305, (1995). 39. YILMAZ S., Optimum Cycle Length and Discharge Burnup For BWRs and PWRs, A Master of Science Thesis in Nuclear Engineering, The Pennsylvania State University, December (2002).

240 40. DOiT (Division of Information Technology) of University of Wisconsin, LPPRIM (Linear Programming by Primal Simplex Method), (1998). 41. David L. Carroll, FORTRAN Source of Genetic Algorithm Code. 42. Deb, K., Multi-Objective Optimization Using Evolutionary Algorithms, John Wiley & Sons, LTD. (2003). 43. Back, T., Fogel, D., B., and Michalewicz, Z., Evolutionary Computation 1 Advanced Algorithms and Operators, IOP Publishing Ltd., (2000). 44. Goldberg, D.E., Deb, K. and Clark, J.H. Genetic Algorithms, noise, and the sizing of populations. Complex Systems 6 (4), 333-362, (1992). 45. Reed, P., B. Minsker, and D. E.Goldberg, Designing a Competent Simple Genetic Algorithm For Search and Optimization, Water Resources Research, 36(12), 3757-3761, (2000). 46. Yilmaz, S., K. Ivanov, S. Levine, Optimizing the Burnable poisons Placement in PWRs - Phase 2 Fuel Enrichment Higher than 5 w/o, Final Report to FERMI, The Pennsylvania State University, May (2003). 47. CASMO-4, A STUDSVIK/SOA Fuel Assembly Burnup Program Users Manual,

48. TABLES-3 Users Manual STUDSVIK/SOA-95/16-REV 0. 49. KAERI Online Plotter for MCNP and http://atom.kaeri.re.kr/cgi-bin/endfplot.pl . ENDF cross section data

50. Helios Methods, Studsvik Scandpower, 01 April, (2000). 51. MICBURN-3 Users Manual, Studsvik/NFA-89/12. 52. Juan Jose Ortiz, Ignacio Requena, Using a multi-state recurrent neural network to optimize loading patterns in BWRs, Annals of Nuclear Energy, 31, 789803, (2004) 53. Erdogan, A., Geckinli, M., A PWR reload optimisation code (XCore) using artificial neural networks and genetic algorithms, Annals of Nuclear Energy, 30, 3553, (2003) 54. Ziver, A.K., Pain, C.C., Carter, J.N., Oliveira, C.R.E., Goddard, A.J.H., Overton, R.S., Genetic algorithms and artificial neural networks for loading pattern optimisation of advanced gas-cooled reactors, Annals of Nuclear Energy, 31, 431457, (2004)

241 55. Mehrdad Boroushaki, Mohammad B. Ghofrani, Caro Lucas, Identification of a nuclear reactor core (VVER) using recurrent neural networks, Annals of Nuclear Energy, 29, 12251240, (2002). 56. Eduardo Fernandes Faris, C. Pereira, Nuclear Fuel Loading Pattern Optimisation using a Neural Network, Annals of Nuclear Energy, 30, 603-613, (2003). 57. Deb, K. and Horn, J. Introduction to the special issue: Multicriterion optimization. Evolutionary Computation Journal 8(2), iii-iv, (2000). 58. Back, T., Fogel, D., B., and Michalewicz, Z., Evolutionary Computation 2 Advanced Algorithms and Operators, IOP Publishing Ltd., (2000).

Appendix 1 Fuel Assembly Absorption Cross Section Data Base


File Name: BPdatabase.dat Definition: This file has thermal absorption cross section data base of limited reference U/Gd fuel assembly designs.

Following is an example of fuel assembly thermal absorption cross section data base of BP designs. The deterministic method uses this data base to convert calculated absorption cross section to practical BP designs. Burnup (GWd/MTU) - Case No (4 represents 16 U/Gd fuel pins) Gd2O3 concentration (w/o) Assembly thermal cross section . . ..(Short version of the data is shown here, it continues from previous sections in reality) START CASE =28 0.0 4 4.00 2.00 0.11587000 0.5 4 4.00 2.00 0.11696000 1.0 4 4.00 2.00 0.11741000 1.5 4 4.00 2.00 0.11762000 2.0 4 4.00 2.00 0.11772000 2.5 4 4.00 2.00 0.11769000 3.0 4 4.00 2.00 0.11756000 3.5 4 4.00 2.00 0.11737000 4.0 4 4.00 2.00 0.11711000 4.5 4 4.00 2.00 0.11683000 5.0 4 4.00 2.00 0.11655000 5.5 4 4.00 2.00 0.11632000 6.0 4 4.00 2.00 0.11612000 6.5 4 4.00 2.00 0.11596000 7.0 4 4.00 2.00 0.11582000 7.5 4 4.00 2.00 0.11570000 8.0 4 4.00 2.00 0.11557000 8.5 4 4.00 2.00 0.11545000 9.0 4 4.00 2.00 0.11530000 9.5 4 4.00 2.00 0.11513000 10.0 4 4.00 2.00 0.11493000 10.5 4 4.00 2.00 0.11470000 11.0 4 4.00 2.00 0.11445000 11.5 4 4.00 2.00 0.11417000 12.0 4 4.00 2.00 0.11386000 12.5 4 4.00 2.00 0.11354000

Appendix 2 Sample CASMO-4 Branch Input


*This is CASMO-4 input for the optimal U/Gd fuel placement library generation *This is CASMO input which was prepared by using FORTRAN-90 program *casmo4caxgeneratorbranch.f90 has been used to create this file *Copyright@ RDFMG, Prepared by Serkan YILMAZ, 09/01/2004 * No of Gd pin=16 Gd2O3 concentration=4 w/o natural Gd2O3 PWR 15 1.4427 21.8110 *15X15 PWR ASSEMBLY PRE 151.6800 *2200PSIA IN BAR * *FOLLOWING CARDS DEFINE MATERIAL COMPOSITION CRA 10.1586/47000=80.000 49000=14.500 48000= 4.500 *Ag-In-Cd CR ABSORBER CRS 8.0272 1.782e-5/28000=10.000 26000=68.000 *SS-304 CR TUBE 25000= 2.000 24000=19.000 14000= 1.000 SPA 13.828 7.0E-6, , 6.5878/302=100.0 *ZR SPACER GRIDS MI1 12.4399/45001=100.0 *RH DET WIRE MI2 2.776/13000=52.94 8000=47.06 *AL2O3 INSULATION MI3 8.43/28000=74.35 24000=15.5 26000=8.0 *INCONEL-600 FOR DETECTOR 6000=0.15 29063=0.5 14000=0.5 25000=1.0 MI4 5.0373/28000=70.3725 24000=14.6708 26000=7.572 *INCONEL-600 & AL2O3 25000=0.9465 29063=0.4733 14000=0.4733 *HOMOGENIZED FOR REGIONS 6000=0.1420 8000=2.5175 13000=2.8322 *INSIDE DETECTOR ASS. MI5 3.7 /13000=51.8824 8000=46.1176 *B4C FOR BP ROD 6000=0.4347 5010= 0.2869 5011=1.2784 *2.0% WT% B4C MI6 6.5878 7.000E-6/ 302=100.000 *BP CLAD MI7 8.4424 1.40e-5/28000=61.000 26000= 2.500 *INCONEL-625 CR TUBE 25000= 0.250 24000=21.500 14000= 0.250 13000= 0.200 6000= 0.050 0=14.25 *ID=0 FORCES SUM=100 * *FOLLOWING CARDS DEFINE PIN GEOMETRY PIN 1 0.4743 0.4826 0.5461 *PELLET DIA 0.3735" PIN 2 0.56007 0.62611 0.63754 0.70358 /'COO' 'BOX' 'COO' 'BOX' *INST SLEEVE & GUIDE TUBE PIN 3 0.6325 0.6731/'COO' 'BOX' *CR GUIDE TUBE

244 DPI 4 0.02286 0.05334 0.07874 0.11811 0.37084 *DETECTOR PIN 0.56007 0.62611 0.63754 0.70358 /'MI1' 'MI2' 'MI3' 'AIR' 'MI4' 'COO' 'BOX' 'COO' 'BOX' / /1 'CE' 1 1 45001 PIN 3 0.4902 0.5029 0.5601 0.6325 0.6731 *EXTENDED LIFE CR /'CRA' 'AIR' 'MI7' 'COO' 'BOX' / /1 'CR1' 'ROD' PIN 3 0.4763 0.4902 0.5588 0.6325 0.6731 *GRAY APSR /'MI3' 'AIR' 'CRS' 'COO' 'BOX' / /1 'CR3' 'ROD' *INCONEL 600 ABSORBING * FUEL PIN LAYOUT FOR ONE-EIGHTH OF AN ASSEMBLY LPI 2 1 1 1 1 3 1 1 1 1 1 1 1 1 3 *3: CR Guide Tube 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 TFU 921.88 TMO 579.11 BOR 800.00 *AVG CORE CONDITIONS SIM 'TYPE16GD04' 4.00 0.0 0 0 *SIM-3 SEG NAME * *FUEL COMPOSITIONS FUE 1 10.4038/4.00 * FUE 2 10.2492/2.708 92234=0.022 92238=81.894 8000=12.380 64152=0.0069 64154=0.0729 64155=0.5036 64156=0.7045 64157=0.5442 64158=0.8669 64160=0.7586 * 4 w/o Gd2O3 Concentration * LFU 0 11 110 1112 11110 120112 1111111 11111111 * PDE 29.609 DEP 0 0.5 1. 2. 3. 4. 5. 6 7 8 9 10 12.5 15 17.5 20 25 30 35 40 45 50 60 65 70

245 STA *FOLLOWING ARE CHANGE CARDS FOR MARK B10K *TMOD, BORON, DOPPLER, * TIT *LOW AND HIGH MODERATOR TEMP BRACH RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 TMO 550.770 605.220 STA * TIT *LOW AND HIGH BORON CONCENTRATION BRANCH RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 BOR 0.000 1600.000 STA * TIT *FUEL TEMPERATURE BRANCH RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 TFU 550.770 STA * TIT *Full lenght CR BRANCH CR1 RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 * FUEL PIN LAYOUT FOR ONE-EIGHTH OF AN ASSEMBLY ROD 'CR1' * STA TIT *APSR BRANCH CR3 RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 * FUEL PIN LAYOUT FOR ONE-EIGHTH OF AN ASSEMBLY ROD 'CR3' * STA TIT *SDC BRANCH RES ' ' 0.0 1.0 5.0 10.0 15.0 20.0 25.0 30.0 40.0 50.0 60.0 70.0 SDC 200 1991.5 2191.5 4383 13149 21915 21915 21915 /'DT' *YR: 0.023 0.25 0.5 1 2.5 5 7.5 10 * STA END

Appendix 3 TABLES Library Generation


'COM' TABLES PWR INPUT FOR SAMPLE CORE 'COM' GENERATED BY SERKAN YILMAZ 06/28/04 'COM' TABLES FOR OPTIMIZATION CODE 'COM' REFLECTOR ASSEMBLY AROUND CORE PERIPHERY 'PWR' 0, 151.68/ 'COM' There must be less than 20 files in one case!!!!!!!/ 'COM' For each fuel type is defined as in case !!!!!!!!!/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' SegmentName:TYPE-(#ofGdpin)-GD-(Gdw/o) 'COM' FRESH FUEL, NO GD/ 'TIT' 'TYPE00, FUEL 4.0- 5.0 w/o, NO GD' 'COM' ***************STARTING NEW CASE **********************TYPE00GD00*********** 'LIB' 'ADD' 'TYPE00GD00'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type00gd0/base3w.cax' 'TYPE00GD00'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type00gd0/branch.cax' 'TYPE00GD00' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type00gd0/lowtmo.cax' 'TYPE00GD00' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type00gd0/highbo.cax' 'TYPE00GD00' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' '/home/varna/serkan/cas4lib/natural/U450type00gd0/base3w.cax' 'TYPE00GD00'/ '/home/varna/serkan/cas4lib/natural/U450type00gd0/branch.cax' 'TYPE00GD00' / '/home/varna/serkan/cas4lib/natural/U450type00gd0/lowtmo.cax' 'TYPE00GD00' / '/home/varna/serkan/cas4lib/natural/U450type00gd0/highbo.cax' 'TYPE00GD00' / '/home/varna/serkan/cas4lib/natural/U500type00gd0/base3w.cax' 'TYPE00GD00'/ '/home/varna/serkan/cas4lib/natural/U500type00gd0/branch.cax' 'TYPE00GD00' / '/home/varna/serkan/cas4lib/natural/U500type00gd0/lowtmo.cax' 'TYPE00GD00' / '/home/varna/serkan/cas4lib/natural/U500type00gd0/highbo.cax' 'TYPE00GD00' /

'EXP' 25,1,0, 0.5, 1, 2, 3,4,5,6,7,8,9,10,12.5,15,17.5,20,25,30,35,40,45,50,60,65,70/ 'TMO' 3 2 550.78 579.11 605.22/ 'HTMO' 2 2 550.78 579.11/ 'BOR' 3 2 0 800 1600/ 'HBOR' 2 1 800 1600/ 'TFU' 2 2 550.77 921.89/ 'ENR' 3 2 4.0 4.5 5.0/ 'CRD' 3 1 ' ' 'CR1' 'CR3'/

247 'SDC' 7 1 0 0.200 1.9915 2.1915 4.383 13.149 21.915/ 'BAS.MAC',3,'EXP','HTMO','ENR',,'TMO'/ 'SUB.MAC','HTMO','TMO','HTMO','TMO'/ 'DEL.MAC',3,'EXP','HBOR','ENR',,'BOR'/ 'SUB.MAC','HBOR','BOR','HBOR','BOR'/ 'DEL.MAC',3,'EXP','TMO','ENR'/ 'DEL.MAC',3,'EXP','BOR','ENR'/ 'DEL.MAC',3,'EXP','TFU','ENR'/ 'DEL.MAC',3,'EXP','CRD','ENR'/ 'DEL.MAC',2,'EXP','SDC'/ 'PRI.MAC',18*0/ 'BAS.FPD',3,'EXP','HTMO','ENR',,'TMO'/ 'SUB.FPD','HTMO','TMO','HTMO','TMO'/ 'DEL.FPD',3,'EXP','HBOR','ENR',,'BOR'/ 'SUB.FPD','HBOR','BOR','HBOR','BOR'/ 'DEL.FPD',3,'EXP','TFU','ENR'/ 'DEL.FPD',3,'EXP','TMO','ENR'/ 'DEL.FPD',3,'EXP','BOR','ENR'/ 'DEL.FPD',3,'EXP','CRD','ENR'/ 'DEL.FPD',2,'EXP','SDC'/ 'PRI.FPD',18*0/ 'BAS.DFS',3,'EXP','TMO','ENR'/ 'DEL.DFS',3,'EXP','BOR','ENR'/ 'DEL.DFS',3,'EXP','TFU','ENR'/ 'DEL.DFS',3,'EXP','CRD','ENR'/ 'DEL.DFS',2,'EXP','SDC'/ 'PRI.DFS',18*0/ 'PIN.PIN'/ 'STA'/ 'END'/ 'TIT' 'TYPE04, FUEL 4.0- 5.0 w/o, NO GD' 'COM' ***************STARTING NEW CASE *****TYPE04GD02*********** 'LIB' 'ADD' 'TYPE04GD02'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=2 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd2/base3w.cax' 'TYPE04GD02'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd2/branch.cax' 'TYPE04GD02' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd2/lowtmo.cax' 'TYPE04GD02' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd2/highbo.cax' 'TYPE04GD02' / 'CAS' '/home/varna/serkan/cas4lib/natural/U450type04gd2/base3w.cax' 'TYPE04GD02'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U450type04gd2/branch.cax' 'TYPE04GD02' / 'CAS' '/home/varna/serkan/cas4lib/natural/U450type04gd2/lowtmo.cax' 'TYPE04GD02' /

248 'CAS' '/home/varna/serkan/cas4lib/natural/U450type04gd2/highbo.cax' 'TYPE04GD02' / 'CAS' 'CAS' 'CAS' 'CAS' '/home/varna/serkan/cas4lib/natural/U500type04gd2/base3w.cax' 'TYPE04GD02'/ '/home/varna/serkan/cas4lib/natural/U500type04gd2/branch.cax' 'TYPE04GD02' / '/home/varna/serkan/cas4lib/natural/U500type04gd2/lowtmo.cax' 'TYPE04GD02' / '/home/varna/serkan/cas4lib/natural/U500type04gd2/highbo.cax' 'TYPE04GD02' /

'EXP' 25,1,0, 0.5, 1, 2, 3,4,5,6,7,8,9,10,12.5,15,17.5,20,25,30,35,40,45,50,60,65,70/ 'TMO' 3 2 550.78 579.11 605.22/ 'HTMO' 2 2 550.78 579.11/ 'BOR' 3 2 0 800 1600/ 'HBOR' 2 1 800 1600/ 'TFU' 2 2 550.77 921.89/ 'ENR' 3 2 4.0 4.5 5.0/ 'CRD' 3 1 ' ' 'CR1' 'CR3'/ 'SDC' 7 1 0 0.200 1.9915 2.1915 4.383 13.149 21.915/ 'BAS.MAC',3,'EXP','HTMO','ENR',,'TMO'/ 'SUB.MAC','HTMO','TMO','HTMO','TMO'/ 'DEL.MAC',3,'EXP','HBOR','ENR',,'BOR'/ 'SUB.MAC','HBOR','BOR','HBOR','BOR'/ 'DEL.MAC',3,'EXP','TMO','ENR'/ 'DEL.MAC',3,'EXP','BOR','ENR'/ 'DEL.MAC',3,'EXP','TFU','ENR'/ 'DEL.MAC',3,'EXP','CRD','ENR'/ 'DEL.MAC',2,'EXP','SDC'/ 'PRI.MAC',18*0/ 'BAS.FPD',3,'EXP','HTMO','ENR',,'TMO'/ 'SUB.FPD','HTMO','TMO','HTMO','TMO'/ 'DEL.FPD',3,'EXP','HBOR','ENR',,'BOR'/ 'SUB.FPD','HBOR','BOR','HBOR','BOR'/ 'DEL.FPD',3,'EXP','TFU','ENR'/ 'DEL.FPD',3,'EXP','TMO','ENR'/ 'DEL.FPD',3,'EXP','BOR','ENR'/ 'DEL.FPD',3,'EXP','CRD','ENR'/ 'DEL.FPD',2,'EXP','SDC'/ 'PRI.FPD',18*0/ 'BAS.DFS',3,'EXP','TMO','ENR'/ 'DEL.DFS',3,'EXP','BOR','ENR'/ 'DEL.DFS',3,'EXP','TFU','ENR'/ 'DEL.DFS',3,'EXP','CRD','ENR'/ 'DEL.DFS',2,'EXP','SDC'/ 'PRI.DFS',18*0/ 'PIN.PIN'/ 'STA'/

249 'COM' ***************STARTING NEW CASE********TYPE04Gd03*********** 'LIB' 'ADD' 'TYPE04GD03'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=3 w/o 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'STA'/ 'COM' ***************STARTING NEW CASE ******TYPE04GD04*********** 'LIB' 'ADD' 'TYPE04GD04'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=4 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd4/base3w.cax' 'TYPE04GD04'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd4/branch.cax' 'TYPE04GD04' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd4/lowtmo.cax' 'TYPE04GD04' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd4/highbo.cax' 'TYPE04GD04' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'STA'/ '/home/varna/serkan/cas4lib/natural/U450type04gd4/base3w.cax' 'TYPE04GD04'/ '/home/varna/serkan/cas4lib/natural/U450type04gd4/branch.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U450type04gd4/lowtmo.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U450type04gd4/highbo.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U500type04gd4/base3w.cax' 'TYPE04GD04'/ '/home/varna/serkan/cas4lib/natural/U500type04gd4/branch.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U500type04gd4/lowtmo.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U500type04gd4/highbo.cax' 'TYPE04GD04' / '/home/varna/serkan/cas4lib/natural/U400type04gd3/base3w.cax' 'TYPE04GD03'/ '/home/varna/serkan/cas4lib/natural/U400type04gd3/branch.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U400type04gd3/lowtmo.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U400type04gd3/highbo.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U450type04gd3/base3w.cax' 'TYPE04GD03'/ '/home/varna/serkan/cas4lib/natural/U450type04gd3/branch.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U450type04gd3/lowtmo.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U450type04gd3/highbo.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U500type04gd3/base3w.cax' 'TYPE04GD03'/ '/home/varna/serkan/cas4lib/natural/U500type04gd3/branch.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U500type04gd3/lowtmo.cax' 'TYPE04GD03' / '/home/varna/serkan/cas4lib/natural/U500type04gd3/highbo.cax' 'TYPE04GD03' /

250

'COM' ***************STARTING NEW CASE ******TYPE04GD05*********** 'LIB' 'ADD' 'TYPE04GD05'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=5 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd5/base3w.cax' 'TYPE04GD05'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd5/branch.cax' 'TYPE04GD05' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd5/lowtmo.cax' 'TYPE04GD05' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd5/highbo.cax' 'TYPE04GD05' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'STA'/ 'COM' ***************STARTING NEW CASE *****TYPE04GD06*********** 'LIB' 'ADD' 'TYPE04GD06'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=6 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd6/base3w.cax' 'TYPE04GD06'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd6/branch.cax' 'TYPE04GD06' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd6/lowtmo.cax' 'TYPE04GD06' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd6/highbo.cax' 'TYPE04GD06' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'STA'/ '/home/varna/serkan/cas4lib/natural/U450type04gd6/base3w.cax' 'TYPE04GD06'/ '/home/varna/serkan/cas4lib/natural/U450type04gd6/branch.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U450type04gd6/lowtmo.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U450type04gd6/highbo.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U500type04gd6/base3w.cax' 'TYPE04GD06'/ '/home/varna/serkan/cas4lib/natural/U500type04gd6/branch.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U500type04gd6/lowtmo.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U500type04gd6/highbo.cax' 'TYPE04GD06' / '/home/varna/serkan/cas4lib/natural/U450type04gd5/base3w.cax' 'TYPE04GD05'/ '/home/varna/serkan/cas4lib/natural/U450type04gd5/branch.cax' 'TYPE04GD05' / '/home/varna/serkan/cas4lib/natural/U450type04gd5/lowtmo.cax' 'TYPE04GD05' / '/home/varna/serkan/cas4lib/natural/U450type04gd5/highbo.cax' 'TYPE04GD05' / '/home/varna/serkan/cas4lib/natural/U500type04gd5/base3w.cax' 'TYPE04GD05'/ '/home/varna/serkan/cas4lib/natural/U500type04gd5/branch.cax' 'TYPE04GD05' / '/home/varna/serkan/cas4lib/natural/U500type04gd5/lowtmo.cax' 'TYPE04GD05' / '/home/varna/serkan/cas4lib/natural/U500type04gd5/highbo.cax' 'TYPE04GD05' /

251

'COM' ***************STARTING NEW CASE *****TYPE04GD07*********** 'LIB' 'ADD' 'TYPE04GD07'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=7 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd7/base3w.cax' 'TYPE04GD07'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd7/branch.cax' 'TYPE04GD07' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd7/lowtmo.cax' 'TYPE04GD07' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd7/highbo.cax' 'TYPE04GD07' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' '/home/varna/serkan/cas4lib/natural/U450type04gd7/base3w.cax' 'TYPE04GD07'/ '/home/varna/serkan/cas4lib/natural/U450type04gd7/branch.cax' 'TYPE04GD07' / '/home/varna/serkan/cas4lib/natural/U450type04gd7/lowtmo.cax' 'TYPE04GD07' / '/home/varna/serkan/cas4lib/natural/U450type04gd7/highbo.cax' 'TYPE04GD07' / '/home/varna/serkan/cas4lib/natural/U500type04gd7/base3w.cax' 'TYPE04GD07'/ '/home/varna/serkan/cas4lib/natural/U500type04gd7/branch.cax' 'TYPE04GD07' / '/home/varna/serkan/cas4lib/natural/U500type04gd7/lowtmo.cax' 'TYPE04GD07' / '/home/varna/serkan/cas4lib/natural/U500type04gd7/highbo.cax' 'TYPE04GD07' /

'STA'/ 'COM' ***************STARTING NEW CASE ***TYPE04GD08*********** 'LIB' 'ADD' 'TYPE04GD08'/ 'OPT' 4,1/ 'COM' CASMO FILES USED FOR THIS SEGMENT 'COM' FRESH FUEL, NO GD, and NO BORON/ 'COM' No of Gd pins=4, Gd2O3 concentration=8 w/o 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd8/base3w.cax' 'TYPE04GD08'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd8/branch.cax' 'TYPE04GD08' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd8/lowtmo.cax' 'TYPE04GD08' / 'CAS' '/home/varna/serkan/cas4lib/natural/U400type04gd8/highbo.cax' 'TYPE04GD08' / 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'CAS' 'STA'/ 'END'/ '/home/varna/serkan/cas4lib/natural/U450type04gd8/base3w.cax' 'TYPE04GD08'/ '/home/varna/serkan/cas4lib/natural/U450type04gd8/branch.cax' 'TYPE04GD08' / '/home/varna/serkan/cas4lib/natural/U450type04gd8/lowtmo.cax' 'TYPE04GD08' / '/home/varna/serkan/cas4lib/natural/U450type04gd8/highbo.cax' 'TYPE04GD08' / '/home/varna/serkan/cas4lib/natural/U500type04gd8/base3w.cax' 'TYPE04GD08'/ '/home/varna/serkan/cas4lib/natural/U500type04gd8/branch.cax' 'TYPE04GD08' / '/home/varna/serkan/cas4lib/natural/U500type04gd8/lowtmo.cax' 'TYPE04GD08' / '/home/varna/serkan/cas4lib/natural/U500type04gd8/highbo.cax' 'TYPE04GD08' /

252
* IMPORTANT NOTE: *Calculations should continue with all of different number of U/Gd fuel pins (8, 12, *16 mixed, 20, 20 mixed) with different concentration of Gd2O3 * I skip all of these calculations here, but someone should consider to define image *file paths similar to above case (4 U/Gd fuel pins) for the other cases

'COM' Reflector cases!!!!!!!!!!/ 'TIT', 'BOTTOM AXIAL REFLECTOR, 1D TABLE VS. BORON'/ 'REF', 'AXIAL',0,151.68/ 'LIB', 'ADD','BREF',0/ 'OPT', 0,0/ 'CAS', '/home/varna/serkan/cas4lib/reflector/bottomref.cax','BREF'/BOTTOM REF 'TMO', 3,2,550.78,574.67,594/ 'HTMO', 3,2,550.78,574.67,594/ 'BOR', 4,2,0,600,1200,2000/ 'HBOR', 4,2,0,600,1200,2000/ 'BAS.MAC',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.MAC',10*0.0/ 'PRI.MAC',18*0/ 'BAS.DFS',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.DFS',10*0.0/ 'PRI.DFS',18*0/ 'STA'/ 'END'/ 'TIT', 'TOP AXIAL REFLECTOR, 1D TABLE VS. BORON'/ 'REF', 'AXIAL',0,151.68/ 'LIB', 'ADD','TREF',0/ 'OPT', 0,0/ 'CAS', '/home/varna/serkan/cas4lib/reflector/toprefbranch.cax','TREF'/TOPR 'TMO', 3,2,550.78,574.67,594/ 'HTMO', 3,2,550.78,574.67,594/ 'BOR', 4,2,0,600,1200,2000/ 'HBOR', 4,2,0,600,1200,2000/ 'BAS.MAC',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.MAC',10*0.0/ 'PRI.MAC',18*0/ 'BAS.DFS',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.DFS',10*0.0/ 'PRI.DFS',18*0/ 'STA'/ 'END'/ 'TIT', 'RADIAL REFLECTOR 3 REGIONS: MOD, SHROUD, MOD, 1D TABLE VS. BORON'/ 'REF', 'RADIAL',0,151.68/ 'LIB', 'ADD','RREF',0/

253 'OPT', 0,0/ 'CAS', '/home/varna/serkan/cas4lib/reflector/rref.cax' 'RREF'/ 'TMO', 3,2,550.78,574.67,594/ 'HTMO', 3,2,550.78,574.67,594/ 'BOR', 4,2,0,600,1200,2000/ 'HBOR', 4,2,0,600,1200,2000/ 'BAS.MAC',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.MAC',10*0.0/ 'PRI.MAC',18*0/ 'BAS.DFS',2,'BOR','TMO',,'HBOR','HTMO'/ 'EPS.DFS',10*0.0/ 'PRI.DFS',18*0/ 'STA'/ 'END'/ 'TIT' 'BLANKET250 2.5 W/0 Blanket, NO BP, blanket' 'PWR' 0, 151.68/ 'LIB' 'ADD' 'BLANKET250'/ 'OPT' 4,1/ 'CAS' '/home/varna/serkan/cas4lib/natural/U250type00gd0/base3w.cax' 'BLANKET250'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U250type00gd0/branch.cax' 'BLANKET250' / 'CAS' '/home/varna/serkan/cas4lib/natural/U250type00gd0/lowtmo.cax' 'BLANKET250' / 'CAS' '/home/varna/serkan/cas4lib/natural/U250type00gd0/highbo.cax' 'BLANKET250 ' / 'EXP' 25,1,0, 0.5, 1, 2, 3,4,5,6,7,8,9,10,12.5,15,17.5,20,25,30,35,40,45,50,60,65,70/ 'TMO' 3 2 550.78 579.11 605.22/ 'HTMO' 2 2 550.78 579.11/ 'BOR' 3 2 0 800 1600/ 'HBOR' 2 1 800 1600/ 'TFU' 2 2 550.77 921.89/ 'CRD' 3 1 ' ' 'CR1' 'CR3'/ 'SDC' 7 1 0 0.200 1.9915 2.1915 4.383 13.149 21.915/ 'BAS.MAC',2,'EXP','HTMO',,,'TMO'/ 'SUB.MAC','HTMO','TMO','HTMO','TMO'/ 'DEL.MAC',2,'EXP','HBOR',,,'BOR'/ 'SUB.MAC','HBOR','BOR','HBOR','BOR'/ 'DEL.MAC',2,'EXP','TMO'/ 'DEL.MAC',2,'EXP','BOR'/ 'DEL.MAC',2,'EXP','TFU'/ 'DEL.MAC',2,'EXP','CRD'/ 'DEL.MAC',2,'EXP','SDC'/ 'PRI.MAC',18*0/

254 'BAS.FPD',2,'EXP','HTMO',,,'TMO'/ 'SUB.FPD','HTMO','TMO','HTMO','TMO'/ 'DEL.FPD',2,'EXP','HBOR',,,'BOR'/ 'SUB.FPD','HBOR','BOR','HBOR','BOR'/ 'DEL.FPD',2,'EXP','TFU'/ 'DEL.FPD',2,'EXP','TMO'/ 'DEL.FPD',2,'EXP','BOR'/ 'DEL.FPD',2,'EXP','CRD'/ 'DEL.FPD',2,'EXP','SDC'/ 'PRI.FPD',18*0/ 'BAS.DFS',2,'EXP','TMO'/ 'DEL.DFS',2,'EXP','BOR'/ 'DEL.DFS',2,'EXP','TFU'/ 'DEL.DFS',2,'EXP','CRD'/ 'PRI.DFS',18*0/ 'PIN.PIN'/ 'STA'/ 'TIT' 'BLANKET300 3.0 W/0 Blanket, NO BP, blanket'/ 'LIB' 'ADD' 'BLANKET300'/ 'OPT' 4, 1/ 'CAS' '/home/varna/serkan/cas4lib/natural/U300type00gd0/base3w.cax' 'BLANKET300'/ 'CAS' '/home/varna/serkan/cas4lib/natural/U300type00gd0/branch.cax' 'BLANKET300' / 'CAS' '/home/varna/serkan/cas4lib/natural/U300type00gd0/lowtmo.cax' 'BLANKET300' / 'CAS' '/home/varna/serkan/cas4lib/natural/U300type00gd0/highbo.cax' 'BLANKET300 ' / 'STA'/ 'END'/

Appendix 4 Sample SIMULATE Input


'COM' @RDFMG,written at 07/16/2004 by Serkan Yilmaz 'COM' Run Date(yyyy/mm/dd):2005/ 1/ 9 Time(hh/mm/sec):22:47:42 'COM' CASMO-4/SIM-3 Global Genetic Code fitness evaluation base input 'DIM.PWR',15/ 'DIM.CAL',1,2,2,1/ 'DIM.DEP','EXP','SAM','HTMO','HBOR','XEN','HCRD' 'PIN' 'EBP'/ 'TIT.PRO','CYCLE 16 DESIGN' 'TIT.RUN','FFCD - 68 Feed, Axial Blanket of 3.0 w/o' 'TIT.CAS','cycle 16 depletion with PSU library & Used fuel' 'FUE.LAB',6/ 1 1 15M-04 15K-05 15D-04 15K-11 15M-12 2 1 15P-09 15P-10 TYPE07 TYPE05 15M-11 TYPE05 TYPE07 15P-06 15P-07 3 1 15N-13 TYPE10 TYPE09 TYPE06 15L-03 TYPE02 15L-13 TYPE06 TYPE09 TYPE10 15O-04 4 1 15K-14 TYPE10 15O-12 15O-11 15P-12 TYPE04 15L-07 TYPE04 15P-04 15O-05 15N-03 TYPE10 15K-02 5 1 15L-14 TYPE09 15M-13 TYPE08 15M-02 15G-12 15O-08 15G-04 15M-14 TYPE08 15M-03 TYPE09 15L-02 6 1 15D-11 TYPE07 TYPE06 15N-14 15B-11 15K-06 TYPE03 15K-08 TYPE03 15F07 15B-05 15N-02 TYPE06 TYPE07 15D-05 7 1 15E-09 TYPE05 15C-10 TYPE04 15N-07 TYPE03 15O-13 TYPE01 15O-03 TYPE03 15N-09 TYPE04 15C-06 TYPE05 15E-07 8 1 15N-04 15E-11 TYPE02 15K-10 15H-13 15H-09 TYPE01 12C-03 TYPE01 15H-07 15H-03 15G-06 TYPE02 15M-05 15D-12 9 1 15M-09 TYPE05 15O-10 TYPE04 15D-07 TYPE03 15C-13 TYPE01 15C-03 TYPE03 15D-09 TYPE04 15O-06 TYPE05 15M-07 10 1 15N-11 TYPE07 TYPE06 15D-14 15P-11 15L-09 TYPE03 15G-08 TYPE03 15G10 15P-05 15D-02 TYPE06 TYPE07 15N-05 11 1 15F-14 TYPE09 15E-13 TYPE08 15E-02 15K-12 15C-08 15K-04 15E-14 TYPE08 15E-03 TYPE09 15F-02 12 1 15G-14 TYPE10 15D-13 15C-11 15B-12 TYPE04 15F-09 TYPE04 15B-04 15C-05 15C-04 TYPE10 15G-02 13 1 15C-12 TYPE10 TYPE09 TYPE06 15F-03 TYPE02 15F-13 TYPE06 TYPE09 TYPE10 15D-03 14 1 15B-09 15B-10 TYPE07 TYPE05 15E-05 TYPE05 TYPE07 15B-06 15B-07 15 1 15E-04 15G-05 15N-12 15G-11 15E-12

256

'NOECHO'/ 'FUE.NEW','TYPE01','AA010',4, 2,,, 1/ 'FUE.NEW','TYPE02','AA020',4, 6,,, 2/ 'FUE.NEW','TYPE03','AA030',8,10,,, 3/ 'FUE.NEW','TYPE04','AA040',8,12,,, 4/ 'FUE.NEW','TYPE05','AA050',8,14,,, 5/ 'FUE.NEW','TYPE06','AA060',8,19,,, 6/ 'FUE.NEW','TYPE07','AA070',8,20,,, 7/ 'FUE.NEW','TYPE08','AA080',4,22,,, 8/ 'FUE.NEW','TYPE09','AA090',8,24,,, 9/ 'FUE.NEW','TYPE10','AA100',8,27,,,10/ 'FUE.NEW' '12C-03' '12C-03' 1,1,0,,,,,1, 1 34.24 7 0.7 9 400/ 'FUE.NEW','15H-07','15H-07' 1,3,2,,,,,1, 1 26.52 7 0.7 9 400 25 25.00/ 'FUE.NEW','15H-03','15H-03' 1,4,2,,,,,1, 1 26.15 7 0.7 9 400 25 25.00/ 'FUE.NEW','15G-06','15G-06' 1,5,2,,,,,1, 1 27.18 7 0.7 9 400/ 'FUE.NEW','15M-05','15M-05' 1,7,1,,,,,1, 1 24.99 7 0.7 9 400/ 'FUE.NEW','15D-12','15D-12' 1,8,3,,,,,1, 1 47.69 7 0.7 9 400 25 25.00/ 'FUE.NEW','15C-03','15C-03' 1,9,2,,,,,1, 1 35.15

257 7 0.7 9 400/ 'FUE.NEW','15D-09','15D-09' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400/ 'FUE.NEW','15O-06','15O-06' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15M-07','15M-07' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15G-10','15G-10' 1,16,3,,,,,1, 1 27.14 7 0.7 9 400/ 'FUE.NEW','15P-05','15P-05' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15D-02','15D-02' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15N-05','15C-03' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400 25 25.00/ 'FUE.NEW','15E-03','15E-03' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15F-02','15F-02' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15C-04','15C-04' 1,26,2,,,,,1, 1 23.44 7 0.7 9 400/ 'FUE.NEW','15G-02','15G-02' 1,28,2,,,,,1,

258 1 24.17 7 0.7 9 400/ 'FUE.NEW','15D-03','15D-07' 1,29,2,,,,,1, 1 23.41 7 0.7 9 400/ 'COM' * SYMMETRY REGION 2 'FUE.NEW','15G-08','15G-08' 1,3,2,,,,,1, 1 26.52 7 0.7 9 400 25 25.00/ 'FUE.NEW','15C-08','15C-08' 1,4,2,,,,,1, 1 26.15 7 0.7 9 400 25 25.00/ 'FUE.NEW','15F-09','15F-09' 1,5,2,,,,,1, 1 27.18 7 0.7 9 400/ 'FUE.NEW','15E-05','15E-05' 1,7,1,,,,,1, 1 24.99 7 0.7 9 400/ 'FUE.NEW','15N-12','15N-12' 1,8,3,,,,,1, 1 47.69 7 0.7 9 400 25 25.00/ 'FUE.NEW','15K-04','15K-04' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400 25 25.00/ 'FUE.NEW','15F-13','15F-13' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15G-11','15G-11' 1,15,1,,,,,1, 1 47.00 7 0.7

259 9 400 25 25.00/ 'FUE.NEW','15E-14','15E-14' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15B-04','15B-04' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15E-12','15C-12' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400 25 25.00/ 'FUE.NEW','15C-05','15C-05' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15B-06','15B-06' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15B-07','15B-07' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'COM' SYMMETRY REGION 3 'FUE.NEW','15H-09','15H-09' 1,3,2,,,,,1, 1 26.52 7 0.7 9 400 25 25.00/ 'FUE.NEW','15H-13','15H-13' 1,4,2,,,,,1, 1 26.15 7 0.7 9 400 25 25.00/ 'FUE.NEW','15K-10','15K-10' 1,5,2,,,,,1, 1 27.18 7 0.7 9 400/ 'FUE.NEW','15E-11','15E-11' 1,7,1,,,,,1, 1 24.99

260 7 0.7 9 400/ 'FUE.NEW','15N-04','15N-04' 1,8,3,,,,,1, 1 47.69 7 0.7 9 400 25 25.00/ 'FUE.NEW','15C-13','15C-13' 1,9,2,,,,,1, 1 35.15 7 0.7 9 400/ 'FUE.NEW','15D-07','15D-07' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400 25 25.00/ 'FUE.NEW','15O-10','15O-10' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400/ 'FUE.NEW','15M-09','15M-09' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15L-09','15L-09' 1,16,3,,,,,1, 1 27.14 7 0.7 9 400/ 'FUE.NEW','15P-11','15P-11' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15D-14','15D-14' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15N-11','15N-11' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400 25 25.00/ 'FUE.NEW','15E-13','15E-13' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/

261 'FUE.NEW','15F-14','15F-14' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15D-13','15D-13' 1,26,2,,,,,1, 1 23.44 7 0.7 9 400/ 'FUE.NEW','15G-14','15G-14' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'FUE.NEW','15C-12','15C-12' 1,29,2,,,,,1, 1 23.41 7 0.7 9 400/ 'COM' SYMMETRY REGION 4 'FUE.NEW','15K-12','15K-12' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400 25 25.00/ 'FUE.NEW','15F-03','15F-03' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15G-05','15G-05' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15E-02','15E-02' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15B-12','15B-12' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15E-04','15E-04' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400

262 25 25.00/ 'FUE.NEW','15C-11','15C-11' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15B-10','15B-10' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15B-09','15B-09' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'COM' SYMMETRY REGION 5 'FUE.NEW','15O-03','15O-03' 1,9,2,,,,,1, 1 35.15 7 0.7 9 400/ 'FUE.NEW','15N-09','15N-09' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400/ 'FUE.NEW','15C-06','15C-06' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400/ 'FUE.NEW','15E-07','15E-07' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15F-07','15F-07' 1,16,3,,,,,1, 1 27.14 7 0.7 9 400/ 'FUE.NEW','15B-05','15B-05' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15N-02','15N-02' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15D-05','15D-03' 1,21,1,,,,,1, 1 47.93

263 7 0.7 9 400 25 25.00/ 'FUE.NEW','15M-03','15M-03' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15L-02','15L-02' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15N-03','15N-03' 1,26,2,,,,,1, 1 23.44 7 0.7 9 400/ 'FUE.NEW','15K-02','15K-02' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'FUE.NEW','15O-04','15O-04' 1,29,2,,,,,1, 1 23.41 7 0.7 9 400/ 'COM' SYMETRY REGION 6 'FUE.NEW','15K-08','15K-08' 1,3,2,,,,,1, 1 26.52 7 0.7 9 400 25 25.00/ 'FUE.NEW','15O-08','15O-03' 1,4,2,,,,,1, 1 26.15 7 0.7 9 400 25 25.00/ 'FUE.NEW','15L-07','15L-07' 1,5,2,,,,,1, 1 27.18 7 0.7 9 400/ 'FUE.NEW','15M-11','15M-11' 1,7,1,,,,,1, 1 24.99 7 0.7 9 400/ 'FUE.NEW','15D-04','15D-04' 1,8,3,,,,,1,

264 1 47.69 7 0.7 9 400 25 25.00/ 'FUE.NEW','15G-04','15G-04' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400 25 25.00/ 'FUE.NEW','15L-13','15L-13' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15K-11','15K-11' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15M-14','15M-14' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15P-04','15P-04' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15M-12','15M-12' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400 25 25.00/ 'FUE.NEW','15O-05','15O-05' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15P-06','15P-06' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15P-07','15P-07' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/

265

'COM' SYMMETRY REGION 7 'FUE.NEW','15O-13','15O-13' 1,9,2,,,,,1, 1 35.15 7 0.7 9 400/ 'FUE.NEW','15N-07','15N-07' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400/ 'FUE.NEW','15C-10','15C-10' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15E-09','15E-09' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15K-06','15K-06' 1,16,3,,,,,1, 1 27.14 7 0.7 9 400/ 'FUE.NEW','15B-11','15B-11' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15N-14','15N-14' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15D-11','15D-11' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400/ 'FUE.NEW','15M-13','15M-13' 1,23,2,,,,,1, 1 26.32 7 0.7 9 400/ 'FUE.NEW','15L-14','15L-14' 1,25,2,,,,,1 1 23.41 7 0.7 9 400/ 'FUE.NEW','15O-12','15O-12' 1,26,2,,,,,1,

266 1 23.44 7 0.7 9 400/ 'FUE.NEW','15K-14','15K-14' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'FUE.NEW','15N-13','15N-13' 1,29,2,,,,,1, 1 23.41 7 0.7 9 400/ 'COM' SYMMETRY REGION 8 'FUE.NEW','15G-12','15G-12' 1,11,3,,,,,1, 1 25.54 7 0.7 9 400 25 25.00/ 'FUE.NEW','15L-03','15L-03' 1,13,1,,,,,1, 1 26.73 7 0.7 9 400 25 25.00/ 'FUE.NEW','15K-05','15K-05' 1,15,1,,,,,1, 1 47.00 7 0.7 9 400 25 25.00/ 'FUE.NEW','15M-02','15M-02' 1,17,1,,,,,1, 1 38.85 7 0.7 9 400/ 'FUE.NEW','15P-12','15P-12' 1,18,2,,,,,1, 1 34.78 7 0.7 9 400/ 'FUE.NEW','15M-04','15M-04' 1,21,1,,,,,1, 1 47.93 7 0.7 9 400 25 25.00/ 'FUE.NEW','15O-11','15O-11' 1,23,2,,,,,1, 1 26.32 7 0.7

267 9 400/ 'FUE.NEW','15P-10','15P-10' 1,25,2,,,,,1, 1 23.41 7 0.7 9 400/ 'FUE.NEW','15P-09','15P-09' 1,28,2,,,,,1, 1 24.17 7 0.7 9 400/ 'ECHO'/ 'RES', 'NEWFUEL'/ *LOADS FUE.NEW DATA INTO CORE 'ERR.CHK','PERMIT'/ 'ERR.SUP', 'SEGTFU3'/ 'LIB','/home/varna/serkan/BP/geneticmodel/tables/cas4libexelonpsu.lib'/ 'SEG.LIB', 1,'TYPE00GD00'/ , 2,'TYPE20GD08'/ , 3,'TYPE08BP02'/ , 4,'TYPE08BP02'/ , 5,'TYPE17GD03'/ , 6,'TYPE21GD05'/ , 7,'TYPE08GD04'/ , 8,'TYPE08BP02'/ , 9,'TYPE08BP02'/ ,10,'TYPE20GD08'/ ,11,'TYPE08BP02'/ ,12,'TYPE16GD05'/ ,13,'TYPE08BP02'/ ,14,'TYPE16GD07'/ ,15,'TYPE08BP02'/ ,16,'TYPE17GD03'/ ,17,'TYPE08GD02'/ ,18,'TYPE08GD02'/ ,19,'TYPE21GD06'/ ,20,'TYPE04GD06'/ ,21,'TYPE08BP02'/ ,22,'TYPE16GD05'/ ,23,'TYPE17GD03'/ ,24,'TYPE20GD08'/ ,25,'TYPE08GD04'/ ,26,'TYPE08GD04'/ ,27,'TYPE08GD06'/ ,28,'TYPE08GD04'/ ,29,'TYPE08GD04'/ ,30,'TYPE00GD00'/

268 ,31,'TYPE00GD00'/ ,32,'TYPE00GD00'/ ,33,'TYPE00GD00'/ ,34,'TYPE00GD00'/ ,35,'TYPE00GD00'/ ,36,'TYPE00GD00'/ ,37,'TYPE00GD00'/ ,38,'TYPE00GD00'/ ,39,'TYPE00GD00'/ ,40,'TYPE00GD00'/ ,41,'TYPE00GD00'/ ,42,'TYPE00GD00'/ ,43,'TYPE00GD00'/ ,44,'TYPE00GD00'/ ,45,'TYPE00GD00'/ ,46,'TYPE00GD00'/ ,47,'TYPE00GD00'/ ,48,'TYPE00GD00'/ ,49,'TYPE00GD00'/ ,50,'TYPE00GD00'/ ,51,'TYPE00GD00'/ ,52,'TYPE00GD00'/ ,53,'TYPE00GD00'/ ,54,'TYPE00GD00'/ ,55,'TYPE00GD00'/ ,56,'TYPE00GD00'/ ,57,'TYPE00GD00'/ ,58,'TYPE00GD00'/ ,59,'BLANKET250'/ ,60,'BLANKET300'/ 'REF.LIB',61,'BREF'/ ,62,'RREF'/ ,63,'TREF'/ 'SEG.DAT', 1,4.00,0.0, 0, 0/ , 2,4.80,0.0, 0, 0/ , 3,4.90,2.4, 0,16/ , 4,4.75,2.7, 0,16/ , 5,4.90,0.0, 0, 0/ , 6,4.70,0.0, 0, 0/ , 7,4.30,0.0, 0, 0/ , 8,4.90,2.3, 0,16/ , 9,4.90,0.0, 0, 0/ ,10,4.80,0.0, 0, 0/ ,11,4.75,2.7, 0,16/

269 ,12,4.55,0.0, 0, 0/ ,13,4.90,3.5, 0,16/ ,14,4.87,0.0, 0, 0/ ,15,4.45,1.4, 0,16/ ,16,4.90,0.0, 0, 0/ ,17,4.90,0.0, 0, 0/ ,18,4.90,0.0, 0, 0/ ,19,4.55,0.0, 0, 0/ ,20,4.87,0.0, 0, 0/ ,21,4.62,3.5, 0,16/ ,22,4.55,0.0, 0, 0/ ,23,4.75,0.0, 0, 0/ ,24,4.90,0.0, 0, 0/ ,25,4.75,0.0, 0, 0/ ,26,4.85,0.0, 0, 0/ ,27,4.90,0.0, 0, 0/ ,28,4.75,0.0, 0, 0/ ,29,4.85,0.0, 0, 0/ ,30,4.00,0.0, 0, 0/ ,31,4.80,0.0, 0, 0/ ,32,4.90,0.0, 0, 0/ ,33,4.75,0.0, 0, 0/ ,34,4.90,0.0, 0, 0/ ,35,4.70,0.0, 0, 0/ ,36,4.30,0.0, 0, 0/ ,37,4.90,0.0, 0, 0/ ,38,4.90,0.0, 0, 0/ ,39,4.80,0.0, 0, 0/ ,40,4.75,0.0, 0, 0/ ,41,4.55,0.0, 0, 0/ ,42,4.90,0.0, 0, 0/ ,43,4.87,0.0, 0, 0/ ,44,4.45,0.0, 0, 0/ ,45,4.90,0.0, 0, 0/ ,46,4.90,0.0, 0, 0/ ,47,4.90,0.0, 0, 0/ ,48,4.55,0.0, 0, 0/ ,49,4.87,0.0, 0, 0/ ,50,4.62,0.0, 0, 0/ ,51,4.55,0.0, 0, 0/ ,52,4.75,0.0, 0, 0/ ,53,4.90,0.0, 0, 0/ ,54,4.75,0.0, 0, 0/ ,55,4.85,0.0, 0, 0/ ,56,4.90,0.0, 0, 0/ ,57,4.75,0.0, 0, 0/

270 ,58,4.85,0.0, 0, 0/ ,59,2.50,0.0, 0, 0/ ,60,3.00,0.0, 0, 0/ ,61,0.00,0.0, 0, 0/ ,62,0.00,0.0, 0, 0/ ,63,0.00,0.0, 0, 0/ 'SEG.ADF', 0 4*2 4*6/ 'SEG.LOA', 0 -100/ 'FUE.ZON', 1,1,'LOCAT-01', 61, 0.0, 1, 14.88, 1, 24.89, 1, 332.230, 1, 342.244, 1, 357.124, 63/ , 2,1,'LOCAT-02', 61, 0.0, 60, 14.88, 31, 25.15, 2, 338.070, 31, 348.086, 60, 363.220, 63/ , 3,1,'LOCAT-03', 61, 0.0, 60, 14.88, 32, 25.15, 3, 338.070, 32, 348.086, 60, 363.220, 63/ , 4,1,'LOCAT-04', 61, 0.0, 60, 14.88, 33, 25.15, 4, 338.070, 33, 348.086, 60, 363.220, 63/ , 5,1,'LOCAT-05', 61, 0.0, 60, 14.88, 34, 25.15, 5, 338.070, 34, 348.086, 60, 363.220, 63/ , 6,1,'LOCAT-06', 61, 0.0, 60, 14.88, 35, 25.15, 6, 338.070, 35, 348.086, 60, 363.220, 63/ , 7,1,'LOCAT-07', 61, 0.0, 60, 14.88, 36, 25.15, 7, 338.070, 36, 348.086, 60, 363.220, 63/ , 8,1,'LOCAT-08', 61, 0.0, 60, 14.88, 37, 25.15, 8, 338.070, 37, 348.086, 60, 363.220, 63/ , 9,1,'LOCAT-09', 61, 0.0, 60, 14.88, 38, 25.15, 9, 338.070, 38, 348.086, 60, 363.220, 63/ ,10,1,'LOCAT-10', 61, 0.0, 60, 14.88, 39, 25.15, 10, 338.070, 39, 348.086, 60, 363.220, 63/ ,11,1,'LOCAT-11', 61, 0.0, 60, 14.88, 40, 25.15, 11, 338.070, 40, 348.086, 60, 363.220, 63/ ,12,1,'LOCAT-12', 61, 0.0, 60, 14.88, 41, 25.15, 12, 338.070, 41, 348.086, 60, 363.220, 63/ ,13,1,'LOCAT-13', 61, 0.0, 60, 14.88, 42, 25.15, 13, 338.070, 42, 348.086, 60, 363.220, 63/ ,14,1,'LOCAT-14', 61, 0.0, 60, 14.88, 43, 25.15, 14, 338.070, 43, 348.086, 60, 363.220, 63/ ,15,1,'LOCAT-15', 61, 0.0, 60, 14.88, 44, 25.15, 15, 338.070, 44, 348.086, 60, 363.220, 63/ ,16,1,'LOCAT-16', 61, 0.0, 60, 14.88, 45, 25.15, 16, 338.070, 45, 348.086, 60, 363.220, 63/ ,17,1,'LOCAT-17', 61, 0.0, 60, 14.88, 46, 25.15, 17, 338.070, 46, 348.086, 60, 363.220, 63/ ,18,1,'LOCAT-18', 61, 0.0, 60, 14.88, 47, 25.15, 18, 338.070, 47, 348.086, 60, 363.220, 63/

271 ,19,1,'LOCAT-19', 61, 0.0, 60, 14.88, 48, 25.15, 19, 338.070, 48, 348.086, 60, 363.220, 63/ ,20,1,'LOCAT-20', 61, 0.0, 60, 14.88, 49, 25.15, 20, 338.070, 49, 348.086, 60, 363.220, 63/ ,21,1,'LOCAT-21', 61, 0.0, 60, 14.88, 50, 25.15, 21, 338.070, 50, 348.086, 60, 363.220, 63/ ,22,1,'LOCAT-22', 61, 0.0, 60, 14.88, 51, 25.15, 22, 338.070, 51, 348.086, 60, 363.220, 63/ ,23,1,'LOCAT-23', 61, 0.0, 60, 14.88, 52, 25.15, 23, 338.070, 52, 348.086, 60, 363.220, 63/ ,24,1,'LOCAT-24', 61, 0.0, 60, 14.88, 53, 25.15, 24, 338.070, 53, 348.086, 60, 363.220, 63/ ,25,1,'LOCAT-25', 61, 0.0, 60, 14.88, 54, 25.15, 25, 338.070, 54, 348.086, 60, 363.220, 63/ ,26,1,'LOCAT-26', 61, 0.0, 60, 14.88, 55, 25.15, 26, 338.070, 55, 348.086, 60, 363.220, 63/ ,27,1,'LOCAT-27', 61, 0.0, 60, 14.88, 56, 25.15, 27, 338.070, 56, 348.086, 60, 363.220, 63/ ,28,1,'LOCAT-28', 61, 0.0, 60, 14.88, 57, 25.15, 28, 338.070, 57, 348.086, 60, 363.220, 63/ ,29,1,'LOCAT-29', 61, 0.0, 60, 14.88, 58, 25.15, 29, 338.070, 58, 348.086, 60, 363.220, 63/ ,30,1,'RREF', 61,0.0, 62, 363.220, 63/ BAFFLE & REFLECTOR 'FUE.TYP',1, 1, 2, 3, 4, 5, 6, 7, 8, 30, 2, 9, 10, 11, 12, 13, 14, 15, 30, 3, 10, 16, 17, 18, 19, 20, 21, 30, 4, 11, 17, 22, 23, 24, 25, 30, 30, 5, 12, 18, 23, 26, 27, 28, 30, 0, 6, 13, 19, 24, 27, 29, 30, 30, 0, 7, 14, 20, 25, 28, 30, 30, 0, 0, 8, 15, 21, 30, 30, 30, 0, 0, 0, 30, 30, 30, 30, 30, 30, 0, 0, 0/ 'SEG.TFU',0,0, 414.8, -32.4340/ 'COM' 'TABLE 2:BASED ON ESCORE CALCULATED MARK B8 FUEL TEMPERATURE ' 'TAB.TFU',1, 0,'EXP',9,'POW',5, 0 3.0 5.0 10.0 20.0 30.0 40.0 50.0 60.0 0.339 40.6 44.0 36.6 8.8 -34.9 -72.1 -104.4 -110.4 -104.1 1.000 17.4 20.2 13.4 -12.2 -53.0 -88.8 -118.1 -119.7 -112.7 1.694 22.8 25.4 18.9 -5.7 -45.1 -75.3 -84.9 -81.5 -73.5 2.372 25.8 28.3 22.0 -1.8 -39.7 -40.2 -39.4 -35.8 -27.1 3.049 22.2 24.7 18.9 -3.7 -13.5 2.8 3.6 4.2 11.1/2568MW

272

'COR.DAT',21.811,363.220,83.974,719.79,555.55/ 'COR.SYM','ROT'/ 'PWR.OPT', 'ON'/ 'PWR.CTP', 0, 15, 100/ core power coordinates 'PWR.TIN', 532, 575, 555.55/ core inlet temp in F 'ITE.BOR',1700/ 'ITE.LIM' 0.999, 1/ 'ITE.EXT' 0.03 .99/ 'HYD.ITE',/ 'CRD.GRP',1 5*0 00000 5*0 3*0 0 0 1 0 6 0 1 0 0 3*0 2*0 0 0 3 0 5 0 5 0 3 0 0 2*0 0 0070807080700 0 0 0305040405030 0 010806020608010 005040202040500 060702070207060 005040202040500 010806020608010 0 0305040405030 0 0 0070807080700 0 2*0 0 0 3 0 5 0 5 0 3 0 0 2*0 3*0 0 0 1 0 6 0 1 0 0 3*0 5*0 00000 5*0/ 'CRD.DAT', 988, 0.3531/ 'CRD.BNK',6*988 900 300/ 'CRD.ZON',1,1,'ELCRA ',0,14.4145,10,363.220/ 'CRD.ZON',2,1,'G-APSR',0,16.9545,20,176.9745, 0,363.220/ 'CRD.TYP',1 000001111100000 000111111111000 001111111111100 011112111211110 011111111111110 111211111112111 111111111111111 111111111111111 111111111111111 111211111112111 011111111111110

273 011112111211110 001111111111100 000111111111000 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0/ 'PIN.EDT',50*' '/ 'PIN.SUM',50*' '/ 'PIN.FIL', 'ON','C16.PIN','2-D'/ 'FUE.JAX','01','02','03','04','05','06','07','08','09','10','11','12','13','14','15'/ 'FUE.IAX','A-','B-','C-','D-','E-','F-','G-','H-','K-','L-','M-','N-','O-','P-','R-'/ 'FUE.INI','CIJLAB'/ 'PRI.INP','ON'/ 'PRI.MAC', 'FULL'/ 'PRI.STA','2RPF' '2KINF' '2FLX' '2EXP' '2EBP'/ 'PRI.FTD','OFF'/ 'PIN.EDT','ON','SUMM','2PIN','2P/B','2XPO'/ 'COM' DEPLETION CARDS 'DEP.CYC','CYC16',0.0,16/ 'COR.OPE' 100.0 100.0,2170/ Hot Full Power - HFP 'DEP.FPD', 8/ No Xe, Peak Sm 'STA'/ 'TIT.CAS','BASE DEPLETION to 693 EFPD' 'DEP.STA','AVE',0.1,0.2,0.4,0.8,1.0,5.0,10.0,20.601/ 'DEP.FPD',2/ Eq Xe, Sm by Depletion 'WRE','C16BASE.RES',20.601/ 'STA'/ 'END'/

Appendix 5 GA Code Input and Output Data for the Optimization of U/Gd Fuel Pin Configurations
File Name: fueldata.inp Definition: This file should contain the information about fuel assembly and the Gd isotopes weight fractions. The user should use this file together with ga.inp file.

4.80 * Fuel Enrichment U-235 (w/o) 16 * Number of U/Gd fuel pins 17 * Mixed concentration (17 and 21), if not use the number of U/Gd fuel pins here 3 * Gd2O3 concentration (w/o) 0.2 * Gd-152 (w/o) 2.1 * Gd-154 (w/o) 14.51 * Gd-155 (w/o) 20.3 * Gd-156 (w/o) 15.68 * Gd-157 (w/o) 24.98 * Gd-158 (w/o) 21.86 * Gd-160 (w/o)
File Name: ga.inp Definition: This file has the necessary data for genetic algorithm code. It should have data such as probability of each operators, operator types, problem size, population size, number of decision variables in genotype, and generation number. The user should change this data, if he/she wants to perform different probability in GA code.

&ga irestrt=0, microga=0, npopsiz=30, nparam=3, pmutate=0.0333d0, maxgen=30, idum=-1000, pcross=0.5d0, itourny=1, ielite=1, icreep=1, pcreep=0.04d0, iunifrm=1,

275 iniche=0, nchild=1, iskip= 0, iend= 0, nowrite=1, kountmx=5, parmin= 1.0d0 1.0d0 1.0d0, parmax= 32.0d0 32.0d0 32.0d0, nposibl= 32 32 32, nichflg=3*1, / Input variable definitions in modified and updated David Carrolls Code [49] c c icreep = 0 for no creep mutations c = 1 for creep mutations; creep mutations are recommended. c idum The initial random number seed for the GA run. Must equal c a negative integer, e.g. idum=-1000. c ielite = 0 for no elitism (best individual not necessarily c replicated from one generation to the next). c = 1 for elitism to be invoked (best individual replicated c into next generation); elitism is recommended. c iend = 0 for normal GA run (this is standard). c = number of last population member to be looked at in a set c of individuals. Setting iend-0 is only used for debugging c purposes and is commonly used in conjunction with iskip. c iniche = 0 for no niching c = 1 for niching; niching is recommended. c irestrt = 0 for a new GA run, or for a single function evaluation c = 1 for a restart continuation of a GA run. c iskip = 0 for normal GA run (this is standard). c = number in population to look at a specific individual or c set of individuals. Setting iskip-0 is only used for c debugging purposes. c itourny No longer used. The GA is presently set up for only c tournament selection. c iunifrm = 0 for single-point crossover c = 1 for uniform crossover; uniform crossover is recommended. c kountmx = the maximum value of kount before a new restart file is c written; presently set to write every fifth generation. c Increasing this value will reduce I/O time requirements c and reduce wear and tear on your storage device c maxgen The maximum number of generations to run by the GA. c For a single function evaluation, set equal to 1. c microga = 0 for normal conventional GA operation c = 1 for micro-GA operation (this will automatically reset c some of the other input flags). I recommend using

276 c c c c c c c c c c c c c c c c c c c c c c c c c c c npopsiz=5 when microga=1. nchild = 1 for one child per pair of parents (this is what I typically use). = 2 for two children per pair of parents (2 is more common in GA work). nichflg = array of 1/0 flags for whether or not niching occurs on a particular parameter. Set to 0 for no niching on a parameter, set to 1 for niching to operate on parameter. The default value is 1, but the implementation of niching is still controlled by the flag iniche. nowrite = 0 to write detailed mutation and parameter adjustments = 1 to not write detailed mutation and parameter adjustments nparam Number of parameters (groups of bits) of each individual. Make sure that nparam matches the number of values in the parmin, parmax and nposibl input arrays. npopsiz The population size of a GA run (typically 100 works well). For a single calculation, set equal to 1. nposibl = array of integer number of possibilities per parameter. For optimal code efficiency set nposibl=2**n, i.e. 2, 4, 8, 16, 32, 64, etc. parmax = array of the maximum allowed values of the parameters parmin = array of the minimum allowed values of the parameters pcreep The creep mutation probability. Typically set this = (nchrome/nparam)/npopsiz. pcross The crossover probability. For single-point crossover, a value of 0.6 or 0.7 is recommended. For uniform crossover, a value of 0.5 is suggested.

File Name: ga.out Definition: The output file of a GA run. This file summarizes the each population individual with its fitness and binary representation form. The parameters of average Fitness, best Fitness, average fitness to best fitness Ratio to reduce or increase the mutation probability, and the history of good solutions for each fuel pin locations are summarized for user analysis. The format of the output file is given as follows:
################# Generation # Binary Code 1 100011111101010 18.00 2 001001000001101 5.00 3 000110010010001 4.00 4 101001110011100 21.00 5 110010101011001 26.00 6 110110101011010 28.00 7 010100111000101 11.00 8 110111111010001 28.00 9 111011111100000 30.00 10 100100111110101 19.00 11 010101110010010 11.00 1 ################# P1 P2 P3 Fitness 32.00 11.00 -2000.00 17.00 14.00 -2000.00 5.00 18.00 -2000.00 29.00 29.00 -2000.00 11.00 26.00 -2000.00 11.00 27.00 -2000.00 15.00 6.00 -2000.00 31.00 18.00 -2000.00 32.00 1.00 -2000.00 16.00 22.00 -2000.00 29.00 19.00 -2000.00

277
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 111001101010100 101011011110001 100100011101111 101101010100000 010111101010000 001000110010110 101110110011101 101110111111111 011111010110101 110111111111001 010100111010010 101010101111100 000011010000011 001000101110001 111111000101111 000010011000100 001011100111011 100110101010110 001110011101110 29.00 22.00 19.00 23.00 12.00 5.00 24.00 24.00 16.00 28.00 11.00 22.00 2.00 5.00 32.00 2.00 6.00 20.00 8.00 27.00 24.00 8.00 22.00 27.00 13.00 13.00 16.00 22.00 32.00 15.00 12.00 21.00 12.00 18.00 7.00 26.00 11.00 8.00 21.00 18.00 16.00 1.00 17.00 23.00 30.00 32.00 22.00 26.00 19.00 29.00 4.00 18.00 16.00 5.00 28.00 23.00 15.00 -2000.00 -2000.00 137.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 -2000.00 138.00 -2000.00 -2000.00 -2000.00 18.47 -

Average Values: 1857.50

17.03

18.73

Average Function Value of Generation= -1857.50 Maximum Funcsubtion Value = 138.00

Number of Crossovers

214

Number of Jump Mutations = 16 Number of Creep Mutations = 2 Elitist Reproduction on Individual

27

*Continue until the end of total number of generation and then summarize the results at *the end of file
Summary of Output Generation Evaluations 0.1000E+01 0.3000E+02 0.2000E+01 0.6000E+02 0.3000E+01 0.9000E+02 0.4000E+01 0.1200E+03 0.5000E+01 0.1500E+03 0.6000E+01 0.1800E+03 0.7000E+01 0.2100E+03 0.8000E+01 0.2400E+03 0.9000E+01 0.2700E+03 0.1000E+02 0.3000E+03 0.1100E+02 0.3300E+03 0.1200E+02 0.3600E+03 0.1300E+02 0.3900E+03 0.1400E+02 0.4200E+03 0.1500E+02 0.4500E+03 0.1600E+02 0.4800E+03 0.1700E+02 0.5100E+03 0.1800E+02 0.5400E+03 Avg.Fitness -.1858E+04 -.1673E+04 -.1602E+04 -.1602E+04 -.1602E+04 -.1479E+04 -.1470E+04 -.1040E+04 -.1612E+04 -.1133E+04 -.1305E+04 -.1356E+04 -.1479E+04 -.1234E+04 -.9977E+03 -.8168E+03 -.1204E+04 -.9706E+03 Best Fitness 0.13800E+03 0.13800E+03 0.13800E+03 0.13800E+03 0.13800E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03

278
0.1900E+02 0.2000E+02 0.2100E+02 0.2200E+02 0.2300E+02 0.2400E+02 0.2500E+02 0.2600E+02 0.2700E+02 0.2800E+02 0.2900E+02 0.3000E+02 0.5700E+03 0.6000E+03 0.6300E+03 0.6600E+03 0.6900E+03 0.7200E+03 0.7500E+03 0.7800E+03 0.8100E+03 0.8400E+03 0.8700E+03 0.9000E+03 -.1165E+04 -.9194E+03 -.1246E+04 -.1133E+04 -.1020E+04 -.1317E+04 -.9707E+03 -.1165E+04 -.9189E+03 -.1165E+04 -.1246E+04 -.9803E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03 0.13900E+03

*******HISTORY DATA FOR EACH POSITION******* PositionNo--#ofHistories--Percentage 1 15 2.8 2 42 7.9 3 40 7.5 4 91 17.0 5 35 6.6 6 7 1.3 7 20 3.7 8 36 6.7 9 6 1.1 10 11 2.1 11 68 12.7 12 20 3.7 13 3 0.6 14 7 1.3 15 20 3.7 16 6 1.1 17 8 1.5 18 18 3.4 19 15 2.8 20 58 10.9 21 6 1.1 22 2 0.4 Total Histories: 534

File Name: ga.restart Definition: This file is the restart file of a last GA run. The restart option in ga.inp should be turn on by using irestrt=1 after first run. This gives flexibility to re-run GA code from the point where the code stops without performing previous calculations again. File Name: kinf+peakpower+RPD+SOB.dat Definition: This file summarizes the positions numbers, BOC Kinf value of a design, its maximum peak pin power in the core and maximum radial power value, fitness value, and the resulting BOC and ECO SOB concentration with the region number defined in solution space.

279

Example of the data is given below: U/Gd pin positions BOCkinf MaxRPD MaxPinPower Fitness BOCSOB EOCSOB -100.0000 0. 19 16 22 1.065 1.460 1.740 -174.00 1920 146 2 11 15 19 1.059 1.400 1.650 -165.00 1872 141 2 13 11 22 1.062 1.410 1.660 -166.00 1891 146 2
File Name: interfacecoupling.f90 Definition: This file is an interface code to perform CASMO-4 cross section generation, TABLES-3 calculations to create core depletion library, and SIMULATE-3 reactor physics calculations. File Name: checkkinf.f90 Definition: This file is to compare individuals BOC Kinf with the defined BOC Kinf filter value. It performs CASMO-4 calculations at only BOC, and then it decides for further calculations depending on Kinf filter definition.

Appendix 6 Core BP Optimization Input and Output Data


File Name: ga.inp Definition: GA code input data &ga irestrt=0, microga=0, npopsiz= 150, nparam= 20, pmutate=0.00666d0, maxgen=100, idum=-1000, pcross=0.5d0, itourny=1, ielite=1, icreep=1, pcreep=0.02d0, iunifrm=1, iniche=0, nchild=1, iskip= 0, iend= 0, nowrite=1, kountmx=5, parmin= 20*1.0d0, parmax= 20*8.0d0, nposibl=20*8, nichflg=3*1, / File Name: archive. out Definition: This file contains the good solutions BP loading with their generation No, population number, Number of U/Gd fuel pins (17 and 21 for mixed concentrations), Gd2O3 concentrations, maximum peak pin power, max. RPD power, cycle exposure where max. peak pin power occurs, BOC and EOC SOB, Gd amount (SUM), Fitness, Region Number in the solution space. One should look at this file to investigate good solutions. An example of the data is given below:

29 148 20 21 21 16 20 21 0 12 17 12 1552.0 68.9 1224.00 -1224.00 30 77 20 21 21 16 20 21 0 12 17 12 1552.0 68.9 1224.00 -1224.00

8 8 8 8 8 7 0 8 6 8 1.542 1.372 .80 1 8 8 8 8 8 7 0 8 6 8 1.542 1.372 .80 1

281 31 27 20 20 21 16 20 21 0 12 16 12 8 8 8 8 8 7 0 7 7 7 1.541 1.389 .40 1559.0 73.1 1200.00 -1200.00 1


File Name: multiobjectives.dat Definition: This file contains the every solutions created during GA evaluation. The data contains, maximum peak pin power, max. RPD power, cycle exposure where max. peak pin power occurs, BOC and EOC SOB, Gd amount (SUM), Fitness, Region Number in the solution space. An example of the data is given below:

MaximumPinPower MaxRPDpower PeakPowerExposure BOCppm EOCppm Summation Fitness RegionNo 2.004 1.807 .00 2080.0 114.5 460.00 -10000.00 3 2.109 1.848 .00 1914.4 101.0 732.00 -10000.00 3
File Name: GAglobalpinpower2region.f Definition: This file contains the source of modified GA code. This code includes the new solution space definition, encoding and decoding process, and new mutation probability subroutine. File Name: CPUtime Definition: The total CPU time of the process is stored here. File Name: coreloading.inp Definition: This file should have the core loading information defined in Chapter 6 (see Table 6.1 for data format)

Appendix 7 CASMO-4 Image Generator Input Files


File Name: fueldata.inp Definition: This file has Gd isotope fractions and the positions of UO2/Gd2O3 fuel pin within an octant fuel assembly model explained in Chapter 3. The CASMO-4 image generator (casmo4caxgenerator.f90) produces all of the base, low moderator temperature, high boron concentration, and branch CASMO-4 runs and image (.cax) files for cross section library generation for 4.0, 4.5 and 5.0 w/o U-235 enriched fuels and 2, 3, 4, 5, 6, 7, and 8 w/o Gd2O3 concentrations. User should define the main location of library in the program. (ex. /home/varna/serkan/cas4lib/). The program creates directory structure under defined main folder structure and locates all of the image files ready to be used for library generation. Example of input data is given below:

0.2 *wGd152 ! Gd isotope fractions 2.1 *wGd154 14.51 *wGd155 ! Gd-155 Absorber isotope fraction SA2=58,000 barn 20.3 *wGd156 15.68 *wGd157 ! Gd-57 SA2=240,000 barn 24.98 *wGd158 21.86 *wGd160 6 *position1 ! Read U/Gd fuel pin positions from fueldata.inp 15 *position2 22 *position3
File Name: Segmentname.txt Definition: Contains segment names

17 17 17 17 17 17 17 END

2 3 4 5 6 7 8

TYPE17GD02 TYPE17GD03 TYPE17GD04 TYPE17GD05 TYPE17GD06 TYPE17GD07 TYPE17GD08

VITA
Serkan Yilmaz was born on August 31, 1975 in Inegol, Bursa, Turkey. He started his college experience at Hacettepe University, Nuclear Engineering Department, Ankara, Turkey in 1992. He obtained his B.S. degree from the department in 1997. He started his career as a research and teaching assistant after his graduation. Then, he transferred to Turkish Electricity Utility in 1998. He had contributed to nuclear power plants department activities as a nuclear fuel cycle engineer during Akkuyu nuclear power plant project until January 2001. During this period, he also achieved M.S. degree from Hacettepe University, Nuclear Engineering Department and graduated in December 2000. Afterward, he decided to join graduate school and to continue his career with Ph.D. studies in the United States of America.

He joined the Mechanical and Nuclear Engineering Department at Penn State University in January 2001. He received his second M.S. degree in December 2002 from the department. He performed fuel cycle cost analysis and evaluations in the development of Electric Power Research Institue (EPRI) optimization studies on fuel discharge burnup and cycle length for light water reactors (PWRs and BWRs) in colaboration with Westinghouse and Exelon Nuclear Company during his M.S. study.

He got married on January 2, 2004 with his beloved wife Mine Ozdemir Yilmaz. He passed Ph.D. final defense exam in April 2005 and graduated in August 2005. His research areas are reactor physics, fuel management, fuel cycle cost analysis, core reload and burnable absorber design optimization for LWRs.