Predictive based monitoring of nuclear plant component degradation using a support vector regression approach

Nuclear power plants are large installations comprised of many active and passive assets. Degradation monitoring of all of these assets is an expensive (labor cost) and highly demanding task. This paper proposes a framework based on Support Vector Regression (SVR) for the online surveillance of critical parameter degradation of plant components. In this case, the on-time replacement or maintenance of components will prevent potential plant malfunctions, and thereby reduce the overall operational cost. In the current work, the researchers apply the SVR model equipped with a Gaussian kernel function in order to monitor components. Monitoring includes the one-step-ahead prediction of the component's respective operational quantity using the SVR model, while the SVR model is trained using a set of previously recorded degradation histories of similar components. The predictive capability of the model is evaluated upon the arrival of a sensor measurement, which is compared to the component failure threshold. A maintenance decision is based on a fuzzy inference system that utilizes three parameters: i) prediction evaluation in the previous steps, ii) predicted value of the current step, and iii) the difference of current predicted value with components failure thresholds. The proposed framework will be tested on turbine blade degradation data.