Development of continuous-energy sensitivity analysis capability in OpenMC

Abstract

© 2017 Elsevier Ltd The iterated fission probability (IFP) method and the Contributon-Linked eigenvalue sensitivity/Uncertainty estimation via Track length importance CHaracterization (CLUTCH) method have been implemented in several Monte Carlo codes to perform sensitivity calculations. In this work the unified adjoint and generalized adjoint calculation framework combining the IFP method and the CLUTCH method is proposed, and corresponding capabilities for calculating nuclear data sensitivity to the effective multipilication factor and reaction rate ratios are implemented in OpenMC. Results are presented on Godiva, Jezebel and a PWR fuel pin system which demonstrate that OpenMC results agree well with reference direct perturbation sensitivity coefficients. In addition, two main problems of the IFP method related to the number of latent generations, namely convergence and statistical fluctuations are investigated theoretically and numerically. This paper further demonstrates that the fission matrix approach is better suited to generate the adjoint and generalized adjoint source distribution in the CLUTCH method to overcome the drawbacks of the IFP method.