Prediction of radiation induced hardening of reactor pressure vessel steels using artificial neural networks

In this paper, we use an artificial neural network approach to obtain predictions of neutron irradiation induced hardening, more precisely of the change in the yield stress, for reactor pressure vessel steels of pressurized water nuclear reactors. Different training algorithms are proposed and compared, with the goal of identifying the best procedure to follow depending on the needs of the user. The numerical importance of some input variables is also studied. Very accurate numerical regressions are obtained, by taking only four input variables into account: neutron fluence, irradiation temperature, and chemical composition (Cu and Ni content). Accurate extrapolations in term of neutron fluence are obtained.