Reliability Assessment of Distribution Networks Considering Wind Power and Source Load Uncertainty

With the large-scale grid-connection of wind power, its output and source-load uncertainties pose challenges for reliability assessment. To address the limitations of traditional methods in terms of dynamic probabilistic modelling and computational efficiency, this paper proposes a framework that combines Markov processes, sequential Monte Carlo(SMC) and BP neural networks(BPNN). A Markov chain-based wind speed state transfer model captures the variability of wind turbine output, while a transmission line fault model takes into account load, aging, and weather effects. A BPNN optimises the sampling efficiency of the SMC approach. Simulation results on the IEEE RTS BUS 6 system highlight the accuracy and efficiency of the method for planning and real-time analysis of highly integrated distribution networks for renewable energy sources.