Evaluation of mean life of power equipment based on 3-parameter weibull distribution

Evaluation of the mean life of equipments is a basic issue for effective equipment maintenance. This paper presents a method to estimate the mean life and its standard deviation of power system equipments with limited end-of-life failure data. We propose a model using 3-parameter Weibull distribution based on the results Li (2004) produced. It is assumed that the mean life of power equipments consists of "non end-of-life failure period" and "period of aging end-of-life failure", which is represented by a location parameter of Weibull distribution. Three parameters of Weibull distribution are estimated by applying Levenberg-Marquardt method, which is one of the optimization algorithms. The computational procedures of non-linear regression, which take account both of good convergence and practical use are proposed. Numerical examples show that our model better fits to actual data than 2-parameter Weibull distribution model.