An artificial neural network model for the lifetime estimation of wood poles supporting the overhead Hellenic electrical distribution network

摘要

Lifetime estimation of power system components is a very important issue for electrical power utilities, since the maintenance scheduling and the replacement decisions are mainly based on such studies. Wood poles may account for about one third of the material costs, and further costs occur due to pole replacements throughout the life time of an overhead distribution line. The Hellenic electrical distribution networks are basically supported with the use of wood poles. Outages due to ageing, degradation or broken wood poles may infer significant costs due to repair and loss of energy to consumers. Therefore the estimation of the life expectancy and the production of survival curve for in-service wood poles are essential. In this paper an artificial neural network (ANN) method is developed and presented for the lifetime estimation of the wood poles supporting the Hellenic electrical overhead distribution network. Actual recorded input and output data collected from the Hellenic distribution network are used for the ANN training. The ANN model is applied on in-service wood poles from several areas of Hellas presented very accurate results concerning their condition. The proposed methodology intends to minimize the life-cycle cost of inspection and refurbishment of wood poles in the distribution network of Hellas, something that is really important in the new competitive and liberating Hellenic electrical energy market.