An Adaptive Fuzzy Model for Failure Rates of Overhead Distribution Feeders

This article presents the development of an adaptive-fuzzy model to predict the failure rate of overhead distribution feeders based on factors such as tree density, tree trimming, lightning intensity and wind index. A gradient descent method was used to train the fuzzy model. To check performance of the model, two error terms, the root mean square error (RMSE) and absolute average error (AAE) were observed. A sensitivity analysis was done to evaluate effectiveness of the trained model. Variations of failure rate to various factors obtained from the sensitivity analysis are discussed.