An AI-Based Approach for Failure Prediction in Transmission Lines Components

In this paper, a novel AI method for failure prediction in transmission lines components is presented. The method combines machine learning and deep learning capabilities. The approach was tested using degradation simulated data of a composite insulator exposed to different levels of environmental pollution. The failure model was constructed using historical real data of a Mexican utility. Preliminary experimental results shows that the joint use of deterministic forecasting and probabilistic diagnosis methods help determine the future failure of a transmission line component for different time horizons with very acceptable precision rates.