Real-Time Prediction of Electricity Distribution Network Status Using Artificial Neural Network Model: A Case Study in Salihli (Manisa, Turkey)

Abstract

Electricity distribution networks are critical to the delivery of energy and the continuity of the economy. The healthyand efficient operation of these networks depends on the prediction of failures, their early detection and the rapidrecovery of the resulting failures. The causes of failure are internal and external factors. Many studies in differentsectors that use different techniques for failure prediction in the literature. The use of artificial intelligencetechniques, which are becoming increasingly important today, in failure estimates; in terms of estimation successand effectiveness, it brings many privileges compared to other techniques. In this study, a status prediction modelhas been developed by using artificial neural network (ANN) technique for power outages and healthy workingconditions of the electricity distribution network installed in Salihli district of Manisa province. In previous studies,using artificial intelligence techniques in the energy sector generally focused on one component of network, lifetime,energy demand estimation, battery life and goods failures. The effect of meteorological factors has not been studiedon the distribution network situation using artificial intelligence techniques. In this study we use hourly poweroutages and hourly meteorological factors that cause failures or healthy conditions. It is aimed to effective riskmanagement and make anticipation of power outage occurring in electricity transmission network, to makepreventive maintenance for failures, to make suggestions for early intervention and shortening downtime andmaintenance.