Computation time reduction of nuclear fuel burnup calculations with the predictor–corrector method using low-order model

Abstract

Special attentions should be paid on the time discretization in nuclear fuel burnup calculations for systems including gadolinia, and several efficient methods based on the predictor–corrector method have been proposed so far. There are still requirements of the further reduction of the computation time, and we propose to use a low-order model (LOM) whose computation time is much shorter than that of a full-order model (FOM) in the corrector calculations. To mitigate the error caused by the LOM adoption, correction factors for one-group quantities are devised and introduced. The proposed method is tested on the lattice physics problems under the condition that FOM is a deterministic method based on MOC and LOM is also a deterministic method based on MOC with the coarse ray-tracing condition with fewer-group cross sections. The proposed method gives results whose accuracy is almost equivalent with that of FOM, and computation time is reduced to 60–65%.