Implementation Of A Bayesian Linear Regression Framework For Nuclear Prognostics

Abstract

Steam turbines are an important asset of nuclear power plants (NPPs), and are required to operate reliably and efficiently. Unplanned outages have a significant impact on the ability of the plant to generate electricity. Therefore, predictive and proactive maintenance which can avoid unplanned outages has the potential to reduce operating costs while increasing the reliability and availability of the plant. A case study from the data of an operational steam turbine of a NPP in the UK was used for the implementation of a Bayesian Linear Regression (BLR) framework. Anappropriate model for the deterioration under study is selected. The BLR framework was applied as a prognostic technique in order to calculate the remaining useful life (RUL). Results show that the accuracy of the technique varies due to the nature of the data that is utilised to estimate the model parameters.