Open Access
Yilmaz, Serkan
Graduate Program:
Nuclear Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
April 20, 2005
Committee Members:
  • Kostadin Nikolov Ivanov, Committee Chair
  • Yousry Azmy, Committee Chair
  • Arthur Thompson Motta, Committee Member
  • Samuel H Levine, Committee Member
  • Soundar Kumara, Committee Member
  • Moussa Mahgerefteh, Committee Member
  • Burnable Poison
  • Genetic Algorithm
  • Gadolinium
  • Nuclear
  • Optimization
  • Fuel Management
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-of-Cycle (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 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 Effective-Full-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 vendor’s 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 vendor’s 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 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.