Real-time estimation of the fatigue usage factor of nuclear reactor components are helpful tools for on-demand assessment of component structural integrity. Real-time measurements of field variables, such as stress and strain, along with the use of historic/available stress/strain versus fatigue life data and a Bayesian-probabilistic framework can not only help to estimate the fatigue usage factor of reactor components in real time but also can help to estimate the associated confidence bounds. In this paper, a Gaussian Process based Bayesian-probabilistic framework is proposed for real-time estimation of mean fatigue usages factor and associated probabilistic bounds. The preliminary proof of concept was demonstrated through fatigue experiments with 316 stainless steel specimens under different conditions: a) under 300 °C and pressurized water reactor (PWR) water chemistry, and b) room temperature and in-air condition. At present the proposed probabilistic monitoring approach is not part of any code requirements, however the proposed framework is output of a futuristic basic research work, which requires more advancement. Also the results discussed in the paper are part of very preliminary basic research and requires advance validation such as under realistic nuclear reactor conditions.

This content is only available via PDF.
You do not currently have access to this content.