Probabilistic method for reliability assessment of metal oxide varistors under lightning stress

Surge protective devices are used to limit the maximum voltage on protected circuits. In the last decade, metal oxide varistors (MOVs) have been the most used protective devices. Since many of the outages experienced on sensitive circuits are due to overvoltages originated by the effects of direct and nearby lightning flashes, it is important to evaluate the reliability of such protective devices under lightning stress. Actually, MOVs are affected by ageing, essentially due to the number and amplitude of stresses, and also other factors such as overheating, pollution and humidity. The aim of this paper is to present a method, based on probabilistic arguments, to evaluate the ageing process of MOVs The expected fife, so obtained can be used to decide when the MOV must be changed before its failure occurs (reliability assessment), since the main standards do not give definitive indications about such features. The estimated expected fife can be used in the design of logistic/maintenance procedures.