Prediction Analysis of Forced Shutdowns in Brazilian Type I Hydropower Plants

Risk management involves both objective and subjective factors. In principle, objective factors are expressed by numerical values being classified by limit values or ranges, while subjective factors express values that can hardly be represented numerically. The problem of evaluation tends to worsen when events must be compared, and a selection or ordering must be performed. The current work presents a proposal of solution to help manage the demands related to the process of prediction, evaluation and classification of shutdowns in Brazilian Hydropower Plants for a regulatory agency. The method, by weighing a relevant set of factors, aims to indicate the projects which should be prioritized by the inspection team, based on criteria data mining techniques such as logistic regression. Its application allowed to measure and guarantee the consistency of the judgments along the process, minimizing risks, promoting more rationalization and administrative transparency.