A practical method of predicting the failure intensity of hydropower generating units

With an eye toward the reliability and economic efficiency of Hydropower Generating Units, reasonable maintenance strategies play critical role in prolonging their lifetime, and optimal timing of maintenance repairs depends fundamentally on the failure intensity. Modeling the failure intensity of generating units based on both current condition and past operation profile is an important ramification of maintenance scheduling. Considering the complexity of influencing factors of the failure intensity during practical operation of Hydropower Generating Units, several main influencing factors are screened out and discussed, including corrective maintenance, preventive maintenance, and the concomitant information. The concomitant information is extended to environment covariates and response covariates at every inspection time during each interval. In this paper we are investigating the case that the concomitant information varies during each interval. Based on the above idea, an improved method that combines proportional intensities model with historical operation and maintenance records is proposed, which makes a quantitative and comprehensive analysis on the effect of the main influences on the failure intensity of the Hydropower Generating Units as a repairable system. In the case study, our results demonstrate that the GPIM fits these data well, particularly when the covariates at every inspection time during each interval are taken into consideration.