LWR-UOx doppler reactivity coefficient best estimate plus (nuclear and atomic sources of) uncertainties

International audienceThe negative Doppler reactivity coefficient of Light Water Reactors (LWR) is a major safety coefficient. Its calculation according to the full temperature range (from Cold Zero Power -CZP- to Hot Zero Power -HZP- or to Hot Full Power -HFP- or even accidental conditions) has to be precisely done. In the first part, this paper aims at estimating the various approximations of standard calculation routes: the uniform temperature spatial distribution versus the parabolic shape, the free gas model (FGM) broadening for absorption cross section even by using effective theories in place of the accurate cristal lattice model. And finally, the double differential scattering treatment in the neutron epithermal range (crude asymptotic kernel assumption versus more accurate Double Differential Cross Section (DDXS) Doppler broadening accounting for resonant up-scattering within the free gas model) will be discussed. More of that, a tentative to account for cristal lattice effects is proposed. The second part is dedicated to the estimation of the Doppler coefficient uncertainty. The two discussed components are the nuclear data, namely the width channel parameters (neutron and $\gamma$) for the low energy $^{238}$U+n resonances and the atomic data by propagating the phonon Density Of States uncertainty within the frame of the Cristal Lattice Model (CLM). This leads to $\pm$6% for the global Doppler coefficient uncertainty