Optimization strategy for distributed photovoltaic access to distribution networks based on power data

Abstract In rural areas, unreasonable access of distributed photovoltaic (DPV) may pose a significant challenge to the safe operation of the power network. To address this, determining the maximum hosting capacity of network feeders is crucial for rational DPV planning. This study analyzes typical integration modes and voltage deviations. Firstly, the load-source correlation is constructed, and a hybrid model of ARMA and GARCH is proposed to predict photovoltaic (PV) generation, serving as input for dynamic grid evaluation. Leveraging the radial topology, a matrix-based method is employed to streamline power flow calculations. The resulting nonlinear optimization problem is solved to determine the maximum permissible PV capacity at each grid node. A case study demonstrates the calculation of the maximum photovoltaic capacity at each node. The results indicate that the proposed model and method effectively assess the DPV consumption capacity based on feeder distribution, providing planning guidance and technical support for PV integration via overhead lines in rural areas.