Preventive Control Method for Transient Stability of Independent Microgrid Based on Deep Learning

Abstract

In view of the difficulty in solving the transient stability constraint part in the microgrid transient stability preventive control, this paper proposes a microgrid transient stability preventive control method based on the deep feedforward neural network to deal with the transient stability constraint. The transient stability prediction model of microgrid built by using deep feedforward neural network can effectively fit the mapping relationship between the active output of generators and transient stability index in the microgrid system. This method takes the active power output of generators in microgrid as input, realizes the fast transient stability prediction of microgrid driven by data, and replaces the solution of differential algebraic equations in the optimal power flow model with traditional transient stability constraints. In this paper, the off-line trained deep feedforward neural network transient stability prediction model is embedded into the preventive control model as a transient stability constraint. Combined with the intelligent algorithm, the active power rescheduling preventive control method under the expected fault is realized to minimize the total cost of active power output adjustment of the generator in the microgrid. In this paper, the feasibility and effectiveness of the proposed method are verified in the three-machine nine-node microgrid system with virtual synchronous machines, with the current limitation of virtual synchronous machines considered and without considering the current limitation of virtual synchronous machines as the research objects.