Reliability Worth Analysis of Distribution Systems Using Cascade Correlation Neural Networks

Reliability worth analysis is of great importance in the area of distribution network planning and operation. The reliability worth's precision can be affected greatly by the customer interruption cost model used. The choice of the cost models can change system and load point reliability indices. In this study, a cascade correlation neural network is adopted to further develop two cost models comprising a probabilistic distribution model and an average or aggregate model. A contingency-based analytical technique is adopted to conduct the reliability worth analysis. Furthermore, the possible effects of adding distributed generation units into the network are evaluated. The proposed approach has been tested on a radial distribution test network evaluating the reliability worth. The results show that the probabilistic distribution model provides a more realistic model for the reliability analysis.