Reliability assessment in electrical power systems: the Weibull-Markov stochastic model

The field of power system reliability is dominated by the use of homogenous Markov models. The negative exponential distributions in these models, however, are unrealistic in the case of repair or switching times. The use of homogenous Markov models is often justified by arguing that the use of other models makes it impossible to perform analytical or nonsequential calculations. It is shown in this paper that this is not the case. A special kind of semi-Markov model is proposed which uses Weibull distributions for all stochastic durations, which enables analytical calculations. The proposed model contains the homogenous Markov model as a subset, enables bell-shaped duration distributions, and may be used in various analytical and nonsequential reliability assessment methods.