Sensitivity study of nuclear heating in PWRs to nuclear data

The work presented in this report aims to study the sensitivity of nuclear heating calculations in Pressurized Water Reactors (PWRs) to the nuclear data library used as calculation input. Accurate understanding and prediction (by means of calculations) of spatial distributions of nuclear heating in the different PWR materials is one of the majors stakes for PWR design and safety studies. Nuclear heating directly affects fuel and material temperatures and is a main parameter for the design of the different components of a reactor. Nuclear heating also affects the location of the power distribution “hot spot” of a reactor which is crucial in safety studies for proper design and sizing of the reactor. Neutron reactions in PWR generate gamma photons and both neutrons and photons interact with reactor materials, causing local energy deposition and increase in the temperature of the materials. The simulation of both neutron and photon transport is thus required to compute nuclear heating. The Monte Carlo neutron- and photon-transport code TRIPOLI-4 developed at the French Atomic Energy Commission (CEA) is used in this report in coupled neutron- photon transport mode to determine values of nuclear heating in three typical PWR assembly configurations taken from the international VERA benchmark: a standard assembly, an assembly with inserted AIC (silver – indium – cadmium) neutron-absorbing control rods and an assembly with fixed burnable neutron-absorbing gadolinium poisons. The nuclear heating is evaluated with three different nuclear data libraries: two different releases from the European JEFF library, namely the JEFF-3.1.1 and JEFF-3.3 releases, and a release from the American ENDF/B library, namely the ENDF/B-VIII.0 release. Shortcomings in the JEFF-3.1.1 and ENDF/B-VIII.0 libraries are identified as those shortcomings lead to wrong values of calculated nuclear heating when in the presence of AIC control rods. Conversely, the JEFF-3.3 library does not suffer from those shortcomings and predicts more reasonable nuclear heating values when in the presence of AIC control rods. For the standard assembly and the assembly with burnable gadolinium poison, the performances from the three libraries are similar.