Monte Carlo Sensitivity Coefficients and Analytical Benchmarks for Unresolved Resonance Probability Tables

TL;DR

This paper introduces Monte Carlo differential operator sampling for sensitivity analysis of unresolved resonance probability tables, develops analytical benchmarks for verification, and validates the method against benchmark models and MCNP6.2 results.

Contribution

It presents new analytical benchmarks and applies Monte Carlo sensitivity methods to unresolved resonance cross sections, verifying accuracy against established models.

Findings

Method agrees with analytical benchmarks.

Sensitivity coefficients match MCNP6.2 calculations.

Validated on Big Ten and Molten Chloride Reactor models.

Abstract

The Monte Carlo differential operator sampling method is applied to the computation of sensitivity coefficients of unresolved resonance probability table cross sections. Three new analytical benchmarks for verifying unresolved resonance treatments and sensitivity coefficient computations are developed. The method and its research-code implementation are verified against these benchmarks and agreement is observed. Numerical results for unresolved resonance sensitivity coefficients are obtained for the Big Ten benchmark and a simplified Molten Chloride Fast Reactor model. Energy-integrated eigenvalue sensitivity coefficients for the unresolved resonance range agree with MCNP6.2 calculations of these two models.