A statistical model for the analysis and prediction of the effect of neutron irradiation on Charpy impact energy curves

The structural integrity safety assessments of nuclear reactor pressure vessels are based, in part, on a prediction of the effect of neutron irradiation on material properties. Databases which monitor this effect are often made up of Charpy absorbed energy measurements. This article presents a generally applicable, new and statistically rigorous method of analysis in which any prior belief as to the form of the material specific, irradiation damage mechanisms can be included. The analytic strategy described can provide predictions of best estimate and confidence limits for use in safety cases. The Charpy absorbed energy measurements obtained over a wide range of temperatures follow a sigmoidal curve which has been modelled by a three parameter curve with the same functional form as a relationship for the Burr distribution function. Charpy transition curves obtained as part of the surveillance of radiation damage of nuclear reactor materials, are found to be displaced to higher temperatures and change their appearance with increasing neutron dose. These changes have been examined by multiple regression analysis of Charpy data using the maximum likelihood estimation method. Relationships which describe the dependence of the upper shelf Charpy energy and the parameters of Burr distribution on irradiation dose and temperature have been derived and linked to the physical processes of irradiation damage. As an illustration of the method, the analysis of a set of Charpy data for nuclear reactor pressure vessel steels is described.