Analysis of the TRAPU irradiation in PHENIX with TRIPOLI-4® and DARWIN-3 for the validation of fast reactor fuel depletion calculations

Abstract

The knowledge of the final inventory in the core of a Fast Reactor is required for many purposes, such as safety issues, reprocessing concerns or conception of new reactors. DARWIN-3 is the CEA code that aims to be the new reference for deterministic fuel depletion calculations. Composed of the deterministic neutronic code APOLLO3® and the depletion module MENDEL, it therefore needs to be validated by comparisons to experimental results or other fuel depletion codes.In this paper, the DARWIN-3 package performances are challenged for the validation of final inventory calculations in the core of a sodium-cooled fast reactor (SFR). This is done by reinterpreting the TRAPU experiment, which is a pin irradiation experiment that took place in the center of the PHENIX reactor between 1977 and 1979. During the TRAPU experiment, ten well-characterized fuel pins – of three different enrichment and Pu isotopic vectors – were placed in a subassembly near the center of the PHENIX core.At the end of the irradiation, the final amount of some isotopes results mainly from their initial concentration (e.g. plutonium, 241Am), whereas others – such as curium or 237Np – are produced exclusively during the irradiation; the final amount of these will be much more sensitive to potential modeling or calculation biases.We have computed Calculation/Experiment (C/E) ratios for different isotopic ratios at the end of the irradiation. A flux adjustment was performed in the calculations, in order to match the calculated 148Nd production to the experiment and avoid propagating a bias in the neutron flux calculation. The uncertainties take into account the experimental uncertainty, the statistical dispersion of the TRIPOLI-4® results and the statistical spread of the different sample results. For each calculated isotopic ratio, we have also assessed the associated uncertainty induced by the uncertainties on the nuclear data involved in the calculation.Comparisons with the previous deterministic package, DARWIN-2, composed of the ERANOS2 neutronic code and the PEPIN2 depletion module, as well as with the TRIPOLI-4® stochastic reference code, are performed.It appears that the production of uranium and plutonium are in agreement between DARWIN-3 and TRIPOLI-4®; the discrepancies observed for some americium and curium isotopes are explained by the slightly softer neutron energy spectrum in DARWIN-3 compared to TRIPOLI-4®.The disparities observed between DARWIN-2 and DARWIN-3 are also attributed to a difference in the neutron energy distribution calculated by the associated neutronic codes. This affects the calculation of the production of americium and curium, which are produced by reactions with large cross section resonances at low energy, where a small difference in the energy spectrum can result in a large difference on the reaction rate. The C/E for the production of those isotopes is worse with DARWIN-3 than it was with DARWIN-2; nevertheless, the production of neptunium, uranium and plutonium is compatible with both calculation codes.Due to its flexibility and attractive calculation time, DARWIN-3 allowed us to test some modeling hypothesis concerning the concentrations to be depleted, the type of decay chain, the effect of the nuclear data library and fuel temperature.