PHYSOR 2006Vancouver, Canada (2006)
This paper presents a new multi-objective function which is made up of
the main objective term as well as penalty terms related to the constraints.
All the terms are represented in the same functional form and the coefficient
of each term is normalized so that each term has equal weighting in the
subsequent simulated annealing optimization calculations. The screening
technique introduced in the previous work is also adopted in order to save
computer time in 3-D neutronics evaluation of trial loading patterns. For
numerical test of the new multi-objective function in the loading pattern
optimization, the optimum loading patterns for the initial and the cycle 7
reload PWR core of Yonggwang Unit 4 are calculated by the simulated
annealing algorithm with screening technique. A total of 10 optimum
loading patterns are obtained for the initial core through 10 independent
simulated annealing optimization runs. For the cycle 7 reload core one
optimum loading pattern has been obtained from a single simulated
annealing optimization run. More SA optimization runs will be conducted to
optimum loading patterns for the cycle 7 reload core and results will be
presented in the further work.
