Analysis of power transients in the CABRI experimental reactor with a multi-physics APOLLO3®/THEDI coupling

•Development of a multi-physics partitioned coupling to simulate power transients in the CABRI research reactor.•Definition and implementation of a step-wise validation process for a multi-physics coupling.•Analysis of the Calculation/Measurement discrepancies to explain their origin.•Implementation of a first methodology to propagate nuclear data covariances in coupled power transient simulations. The CABRI research reactor located at CEA Cadarache is dedicated to the analysis of nuclear fuel behavior during Reactivity-Injection Accident (RIA) in Pressurized Water Reactors (PWRs). The reactor experimentally simulates power pulse transients in the driver zone, which induce a RIA-representative energy deposition in a fuel sample contained in a water-loop that reproduces PWR thermal-hydraulics conditions. Power pulses are initiated by the 3He depressurization of four transient rods, which introduces reactivity into the core. The injection of reactivity causes the reactor power to increase rapidly, resulting in a rise in fuel temperature. This temperature rise mitigates the power excursion by the Doppler effect. Subsequently, heat transfer occurs between the fuel, the cladding and the moderator. The evolution of the core power is measured during power transients, and these measurements can be valorized for the experimental validation of multi-physics simulation tools. Indeed, the experimental validation of these coupling mainly relies on measurement data coming from the current industrial reactor fleet. The measurement data coming from experimental facilities for the validation of multi-physics coupling at the reactor scale are very rare. A multi-physics core modelling based on the APOLLO3® and THEDI tools within the coupling platform C3PO was developed to simulate power transients in the CABRI reactor. A step-wise validation process has been defined and applied to this multi-physics coupling, from the numerical validation against Monte Carlo calculations at the neutron lattice level, to the integral experimental validation at the core level. Two types of transients were simulated: firstly, three neutron-driven supercritical transients (in which reactivity feedbacks are negligible), and secondly, a prompt supercritical transient, which is a high-energy transient with a five-decade power increase. The Calculation/Measurement (C/M) discrepancies on core power during supercritical transients can be explained with the support of the numerical validation results.