Power distribution system reliability evaluation using weight-optimised ANN approach

Power systems are identified as the most complex infrastructure over the globe. That Power distribution system is a critical network section with the greatest concentration of failure occurrences. An evaluation of the reliability of the power distribution network is exhaustive trouble. However, analysation of effectiveness and failure rate interpretation is the most significant barrier to limiting the real-time application of distribution system reliability evaluation. The methodology proposed in this paper is to integrate deep learning-based techniques implemented to execute network reliability for the preferred distribution test system—execution of reliability evaluations obtained by combining MCS models for improving the accuracy of computing PDN reliability indices. Subsequently, for system state observation, an ANN-based state-space categorisation-dependent (SSC) system with Enriched Binary Particle Swarm Optimization (EBPSO) is introduced to seek failure states throughout uncategorised subspaces. Moreover, the reliability evaluation framework is being implemented in Roy Billinton Test System (RBTS) bus 2 systems, and MATLAB Software analyses its reliability indices. The reliability indices for the third component fault case scenario provide a better outcome than other scenario cases. The reliability indices values such as SAIFI, SAIDI, AENS, and ASAI at scenario 3 condition are 0.232,3.497,17.48,0.99905. The proposed structure's outcome promotes an excellent reliability evaluation pattern, confirming lesser loss and providing robust reliability indices computation.