Smooth Switching Control Method for Parallel and Off Grid of Distributed Photovoltaic Power Grid Based on Deep Reinforcement Learning

The parallel and off grid switching of distributed photovoltaic power grid will cause sudden changes in voltage and current, which is a key factor affecting its stable operation. Therefore, a research on the parallel and off grid smooth switching control method of distributed photovoltaic power grid based on deep reinforcement learning is proposed. The in-depth reinforcement learning method DQN algorithm is used to build the energy management model of the distributed photovoltaic power grid, explore the characteristics and laws of the distributed photovoltaic power grid, and on this basis, in-depth analysis of the transient phenomenon of the parallel off grid switching of the distributed photovoltaic power grid is carried out. Based on the PQ and VF control principles, the grid connected controller and the off grid controller are designed, and the smooth parallel off grid switching control strategy is formulated, The smooth switching control of the parallel and off grid of the distributed photovoltaic power grid can be achieved by implementing the strategy. The experimental data show that the minimum value of the sudden change coefficient of voltage and current obtained by the proposed method is 0.1 and 0.2, which fully proves that the proposed method has better control effect of parallel and off grid switching.