Time-dependent reliability assessment of a distributed generation system based on multi-valued decision diagrams and Markov processes

In this paper, we present a new time-dependent reliability method for Distributed Generation (DG) power systems. In the developed method, Markov models are used to describe the time-dependent (failure/repair) behaviors of the power generation components. Multi-valued decision diagrams (MDD) are developed to model the generation capacity of each component. The capacity of the DG system is, then, calculated as a summation of the power generated by all the components. A MDD-based reliability evaluation algorithm is developed to calculate two reliability indexes for the DG system, i.e., Expected Energy Not Supplied (EENS) and Loss of Load Expectation (LOLE). A comparison to existing time-dependent reliability methods shows that the developed method can achieve more accurate results while requiring far less computational costs. Also, a comparison to the static reliability analysis methods shows that the developed method can better estimate the reliability indexes, especially at the early stage of the life cycle.