Prediction Modeling of Frequency Response Characteristic of Power System Based on Historical Data

The disturbance brought by the connection of new energy to the grid poses greater challenges for the automatic generation control system of the power grid. In this paper, a real-time prediction method of frequency response characteristic coefficient of power system based on historical data is proposed, which can be used to improve the frequency deviation of the automatic generation control system coefficient setting. First, the maximal information coefficient is used as the index of the max-relevance and min-redundancy algorithm to extract the features of many factors affecting the frequency response characteristic in the power system. Then, according to the results of feature extraction, a prediction model for the frequency response characteristic of power system was established by using an artificial neural network with dropout. The analysis and comparison of different models using historical operation data of the power grid verified that the model has better prediction accuracy and generalization ability.