An online condition monitoring thermal prognostic indicator system for MV cable circuits

Developing a reliable online condition monitoring prognostic indicator tool for MV cables is of great importance as it can predict and prevent upcoming failures of the distribution cable circuits. This paper introduces a thermal prognostic model for MV underground cable terminations based on a support vector regression algorithm. The model is shown to predict the likely temperature along the cable thirty minutes into the future and is able to rapidly identify temperature anomalies which may indicate upcoming failures