A Security Adjustment Framework for Distribution Networks Under the Penetration of Large Scale Renewable Energy

The emergence of large-scale distributed generation (DG) poses challenges to the safe operation of distribution networks. For example, during daytime photovoltaic (PV) power generation may cause reverse currents in the upper power grid lines, resulting in voltage exceeding the limit. The existing distribution network voltage control has the characteristics of local regulation and dependence on external devices. With the depletion of external regulatory resources, it gradually fails to meet the demand for online access to microgrids, thereby limiting the absorption of solar energy. This article constructs a distribution network security regulation framework that takes into account multiple regulation resources, mainly including reactive power optimization and flexible topology optimization. A reinforcement learning-based solution algorithm is proposed for topology optimization. This algorithm designs a multi-agent distributed knowledge storage method to solve the combinatorial explosion problem caused by multiple control variables, and achieves the transfer of prior knowledge by constructing a knowledge base. The feasibility of using network topology as a voltage control resource was verified through case analysis and simulation. Under the adjustment of the topology structure, the distribution network has effectively improved its ability to absorb renewable energy in the microgrids.