Calibration of reduced models by multi-objective genetic algorithms

摘要

In practice, a dynamic approach to reliability analysis is likely to require a significant increase in the computational efforts, due to the need of integrating the dynamic evolution into the stochastic failure/repair processes. Thus, it becomes mandatory to resort to simplified models of dynamics. These are typically based on effective parameters whose values need to be estimated so as to best fit to the available plant data. In this paper we propose a multi-objective genetic algorithm approach for estimating the effective parameters of a simplified nuclear reactor dynamics model based on point kinetics to describe the neutron balance in the core and on thermal equilibrium to describe the energy exchanges. The calibration of the effective parameters is achieved by best fitting the model responses to the actual evolution profiles simulated by a design code of literature. The (pseudo)measured quantities of interest are the reactor power and the average fuel temperature.