Machine Learning Approach for Predictive Maintenance in Hydroelectric Power Plants

Abstract

The future of hydropower industry has as key elements, optimization in operation and maintenance, costs reduction and increase of reliability. This means greater challenges in the operation of hydroelectric power plants, therefore, greater demands in maintenance. With technology advances and its role in the industrial sector through the revolution 4.0 or Industry 4.0, artificial intelligence and machine learning applications enables the development and modernization of current maintenance techniques in hydropower plants, through condition monitoring, fault diagnosis and predictive maintenance, thus, an early detection can save a lot of time and money.In this study, two techniques are proposed to enable predictive maintenance in the Peña Blanca hydroelectric power plant, using two deep learning models for anomaly detection. The first one consists of a Deep Neural Network with Logistic Regression to classify various types of failures, for the second one a Recurrent Long Short-Term Memory neural network (LSTM) with Autoencoder is used to classify various flaws. With the first model it was found that it is possible to generalize several types of failures, while the LSTM model adjust better on detecting high temperatures on generator bearings since was a failure that occurred frequently during the study.