Future failure rate prediction for cables in power systems

Abstract

In real life, the normal operation of a power system matters the people’s daily life. This dissertation focuses on predicting the failure rate of cables in a power system, to save money and time for maintenance. By using the concept of Root Cause Analysis (RCA), figure out the problem of cables and find out the reasons, as well as how to predict the failure rate with the methods of statistic models and AI. With the cable data in real life, including information on the failed time, use different lifetime statistic models for fitting and predicting data, as well as compare the results of different models. For AI methods, apply decision trees and random forests into predicting the failure rate of the cable data.