Uncertainty quantification for the Doppler reactivity feedback coefficient of MYRRHA

Safety analyses of MYRRHA require calculating the reactor temperature feedback coefficients of reactivity associated with power operation. Within this framework, a significant emphasis is put on the quantification of the reactivity coefficient uncertainty, which takes into account different uncertainty sources. This work reports the investigation of the fuel Doppler coefficient of reactivity of MYRRHA implemented with the Serpent 2 Monte Carlo particle transport code. Nuclear data and the MYRRHA mixed-oxide (MOX) fuel composition are selected as major sources of uncertainties, and their impact on the Doppler coefficient is assessed with a stochastic sampling approach. Nuclear data covariance matrices are taken from the JEFF-3.3 evaluated library and propagated with the SANDY code. Uncertainties of the fuel composition are derived from declared records for MOX fuel assemblies irradiated in a pressurized water reactor/boiling water reactor. The contribution of the aleatoric Monte Carlo uncertainty could be removed from the stochastic uncertainty estimate of the Doppler coefficient by using a methodology based on conditional estimators. This technique has proved to be tremendously advantageous to quantify the uncertainty of reactivity feedback coefficients using Monte Carlo codes, overcoming some of the limitations associated with sensitivity-based uncertainty propagation methods. The uncertainties of the MYRRHA Doppler constant K_D, considering the variability of nuclear data (only cross sections) and MOX fuel compositions, are, respectively, 3.0% and 1.3%, with a negligible dependence on the considered fuel temperature. The largest nuclear data uncertainty contributions comes from ²³⁸U and ^239,240Pu, while the impact of the Pb and Bi uncertainties is only marginal.