A novel extension of chord length sampling method for TRISO-type fueled reactor applications

A new chord length distribution model is proposed to characterize the stochastic distribution of TRISO fuel particles in nuclear reactor systems and is used in the chord length sampling (CLS) method to analyze the neutronic behavior of TRISO fuel systems. In this model, the coating layers of fuel particles are homogenized with the background matrix region. The probability density function (PDF) of the chord length between fuel kernels, instead of fuel particles, is developed and is used in the CLS method. We first apply the new CLS model to solving one-group eigenvalue problems in a simplified 3-D stochastic medium system. Good accuracy is obtained in predicting the multiplication factor and fission density distribution. The relative differences are within 1.0% for both the multiplication factor and the total fission density in all the studied scenarios. We then apply the new CLS model to analyzing three realistic reactor designs: two Very High Temperature Gas-cooled Reactor unit cells and one fuel pin unit cell of an innovative light water reactor design with accident tolerant fuel. Infinite multiplication factor and intra-Dancoff factor are evaluated for the three unit cells respectively. Compared with the reference results, predictions from CLS simulations show a relative difference of less than 0.26% for infinite multiplication factors and less than 1.0% for intra-Dancoff factors in all the studied cases. Meanwhile, a significant improvement in the computational efficiency has been observed for the CLS new model compared with the old model, with a speedup of at least 1.70. The CLS method with the new chord length distribution model is verified to be an efficient Monte Carlo method, superior over the old model and without sacrificing the accuracy.