Cable failures are disruptive, costly to repair and have a serious impact on customer confidence. Developing a reliable online condition monitoring prognostic indicator tool is of a great interest as it could predict and prevent upcoming failures in a power system network. Our paper introduces a novel thermal prognostic indicator system for MV underground cable joints based on the machine learning Support Vector Regression (SVR) algorithm that predicts the likely temperature along the cable thirty minutes into the future and is able to detect temperature anomalies, which potentially can indicate upcoming failures.
The above system was developed through an experiment in a distribution substation. Temperature condition monitoring units were installed on four underground cable joints within a 132kV/11kV substation. Real-time current loading data, weather conditions as well as surface temperature by the cable joints were used during the development of the online condition monitoring prognostic indicator tool. Following development, the reliability of the system was subsequently tested over a period of four years in the same substation. This paper presents the findings of the four year testing period of the developed tool installed on one of the investigated cable joints and explores the possibility of further implementation of the system into Industry 4.0 application.
Download the paper below to learn more about the performance results of the developed thermal prognostic system.
