Probabilistic Analysis Enabled Proactive Fault Risk Assessment of Electricity Power Distribution Systems

This paper establishes a probabilistic operational scenario forecasting (POSF)-based method to obtain the actual operational state profile and probabilistic map for power distribution system reliability analysis (PFRA) tasks. The proposed POSF forecasts various probabilistic scenarios with multi-scale lead times based on Wasserstein Generative Adversarial Networks (WGAN), and the developed WGAN model is further incorporated into a stochastic constrained optimization framework to establish the conditional confidence interval and risk assessment index based on past observations. The POSF-based method is extensively evaluated through a range of power outage events in power distribution system transformers’ daily operation realistic measurement records. The simulation outcomes confirm that the developed method can well adaptively forecast operational scenarios under uncertainties along with outage risk assessment compared with the existing benchmarks.