Multivariable modeling of J-R data

Current methods for evaluating safety and remaining life issues for reactor pressure vessels (RPV's) and piping are based on the J-R curve. Unfortunately, field-aged J-specimens are not always available for the specific weld or heat in service or at the radiation conditions which match the case to be analyzed. In order to best evaluate plant safety, therefore, it is necessary to reliably estimate a J-R curve from available data, such as material chemistry, radiation exposure, tensile properties, and Charpy data. A multivariable model for predicting J-R curves from available data, such as material chemistry, radiation exposure, and Charpy data, is being developed. This is an interim report on a project to collect public test data, apply advanced pattern recognition tools, and fit an improved model. A four-parameter model for the dependence of Jd on crack extension, Charpy energy, temperature, and specimen thickness is presented