Optimal Design of Measurement Network for Neutronic Activity Field Reconstruction by Data Assimilation

Abstract

Several studies show that data assimilation methods, by merging information both from model and measurement can be used to elaborate an optimal determination of neutronic activity field within a nuclear PWR reactor core. Here the problem addressed and solved is to determine an optimal repartition of the instruments, used to make measurements within the core, to get the best possible reconstructed field using a data assimilation procedure. The position optimization is based on simulated annealing, with a Metropolis-Hasting algorithm. To perform the method in the framework of data assimilation, algebraic optimisation related to computing time of data assimilation has been developed and is used here.