A computer program for loading pattern optimization is developed.
Optimization is performed through one to one fuel assembly shuffling. Dynamic Programming is used to determine an optimized set of assembly shuffles from the initially-guessed loading pattern. And the peak power of the fuel assembly node is only used for objective function.
The present contains two major functions. The one decides which assembly shuffles are to be examined, judges them by their resulting power distribution, finally determines which are to be included in the sequence leading to an optimal loading pattern. The other calculates 1/8 core power distribution from given information, such as power and reactivity concerning the positions of the fuel assembly, and some core constants.
Much of effort is focused on the improvement of accuracy of the power calculation results. Using the proper model, in which the exact physical meaning of the coupling coefficient is considered, the accuracy of the power calculation is increased.
A sensitivity test of this power reactivity relation is performed for the cycle I of the Young Kwang Unit I. A loading pattern optimization test was performed for four hypothetical loading pattern. It was proved that the power reactivity relation is capable of determining optimum loading pattern in case of well guessed initial loading pattern, but more additional functions have to be adopted to search optimum loading pattern for ill guessed loading pattern.
