Reliability assessment based on Bayesian networks and time sequence simulation for distribution systems

As the Bayesian network model suitable for distribution system reliability evaluation is inadequate, the compositions of Bayesian network model are researched, and “unite” relation models and “cause-effect” relation models are set up. As the accurate inference of Bayesian network to large-scale distribution systems reliability indices is difficult to calculate, a time sequence simulation inference algorithm based on Bayesian network and time sequence simulation technique is put forward. This algorithm can generate state of elements and random time period, inference the system states in real-time, aggregate the amount of time, consumer power cut frequency and amount. Using the cumulative amount may not only calculate the reliability indices of the system, but also carry out diagnostic reasoning and causal reasoning, thus can accomplish collectivity evaluation of the system and identify weak taches of strangulation system. By comparing with the result of the simulation inference algorithm and accurate inference algorithm, the simulation inference algorithm is reasonable and effective.