To enhance the consumption capacity of distributed photovoltaic （PV） and improve the power supply reliability of rural areas, this paper proposes a hierarchical coordinated planning method of PV and energy storage system for rural microgrids. Firstly, the medium and long-term load forecasting model based on the DenseNet-Autoformer network is proposed, which uses the autocorrelation mechanism to extract the deep time-series features, and comprehensively considers factors such as meteorology and economy on long-term monthly electricity consumption; Secondly, the Wasserstein generative adversarial network-gradient penalty method is used to expand the PV power samples. On this basis, the typical source and load scenarios are obtained via Gaussian mixture clustering; Finally, taking into account the electricity self-sufficiency and the off-grid reliability, the PV and storage configuration plan of the distribution transformer area is optimized, and the power supply reliability under extreme scenarios is quantified based on chance constraints. Meanwhile, considering the overall energy self-sustained ability of the 10 kV feeder, the feeder-level PV and storage configuration plan is optimized. The simulation results show that the proposed method can enhance the energy self-sustained ability and PV consumption capacity of rural microgrids and improve the power supply reliability in extreme scenarios.

microgrids  /  artificial intelligence  /  planning  /  electric load forecasting  /  power supply reliability  /  layered consumption