Integrated fault location and isolation strategy in distribution networks using Markov decision process

The data gathered from various intelligent sensors installed throughout the network could be utilized for the fault localization, helping the system restoration, reducing the outage time and improving system reliability. However, due to the distribution network characteristics and particularities, the precise fault location is very hard to determine, especially in isolated neutral networks. Consequently, the restoration process is affected by the fault location error and the probability of fault in the exact location. In this paper, Markov Decision Process is used as a tool for the determination of the faulted feeder section and its isolation from the grid. The algorithm is based on the optimization of several criteria, while the transition probabilities among states are obtained from fault passage indicators status. The posterior probability that the particular section is faulted is obtained using the Bayes probability theory. The methodology is tested on the IEEE 123 distribution test network.