Frequency prediction after disturbance of grid-connected wind power systems based on WOA and Attention-LSTM

Abstract

Under the guidance of the goal of the “carbon peaking and carbon neutrality”, due to the high proportion of renewable energy and the high proportion of power electronic equipment, the power system will bring strong randomness, low inertia and other characteristics, causing a large number of frequency stability problems. In order to solve the problems of traditional power system frequency prediction methods, such as difficulty in modeling and poor prediction accuracy, and to determine whether the frequency stability problem will occur after the wind power grid-connected system is disturbed, the Lasso algorithm is first used to reduce the dimension of the input data, and then the attention mechanism based long and short memory (attention LSTM) neural network is used to predict the output frequency curve, The network parameters are optimized by the global search algorithm Whale Optimization Algorithm (WOA). Finally, the accuracy of the algorithm is verified by taking the improved wind power grid-connected 39 bus as an example. The results show that this method has a good guiding significance for evaluating the frequency stability of the wind power grid-connected system after interference, and can effectively predict the frequency curve of the system after interference, and evaluate the frequency stability of the system after topology change.