Towards establishment of an efficient approach for validation of PWR full core Monte Carlo simulations at hot zero power conditions

The use of Monte-Carlo (MC) simulations for neutronic reactor core physics has been of high interest since long. Resulting MC models would be relevant, e.g., for high-resolution assessments of the local neutron flux and power gradients, generally beyond the modeling capabilities of engineering deterministic codes for regular full core simulations. Such MC simulations would be of particular importance for the advancement of reactor operation and safety assessments or, for example, to support designing new materials testing experimental programs at operational reactors, designing advanced reactor cores, etc. Nevertheless, MC-based core-follow burnup calculations are still challenging for routine applications. Therefore, the Laboratory for Reactor Physics and Thermal-Hydraulics (LRT) has developed in recent years a “cycle-check-up” (CHUP) concept. It allows to transfer the operating conditions (coolant and fuel temperatures, density of moderator, boron concentration, position of control rods) and burned fuel isotopic compositions from validated reference core-follow models, based on the state-of-the-art deterministic codes CASMO5/SIMULATE5/SNF, to MC codes, such as Serpent 2.2 or MCNP6®. This article presents recent work performed on the optimization of the methodology. This optimization is based on a newly developed adaptive fuel materials clustering to maximize the accuracy of the simulations while keeping the memory consumption of simulations constant. In addition, thanks to the availability of the validated deterministic core-follow models and access to additional reference data for Swiss power plants, such as start-up tests at PWRs, there is a valuable opportunity to extend the validation database for Burnup Credit applications using Swiss reactors data. This paper presents the ongoing work towards the verification and validation (V&V) of MC full core simulations for burned configurations against results of the validated deterministic models. The radial power distribution of the MC full core model using pin-wise composition was verified, yielding relative deviations in the [-9, 6]% range against the validated nodal solver. Additionally, using the developed MC models for hot zero power (HZP) conditions, the analysis of the start-up reactor measurements showed a −100 ± 2 pcm deviation from criticality, which is considered as an excellent agreement.