Prediction method of primary frequency modulation capability of power system based on MEA-BP algorithm

Abstract

In order to accurately predict the primary frequency modulation capability of power system when frequency deviation occurs, a method for predicting primary frequency modulation performance of power system based on improved neural network with thought evolution algorithm (MEA) is presented. Nine-dimensional parameters, such as frequency deviation before and after disturbance, system backup capacity, load level, mechanical power and electrical power, are selected as input eigenvalues, and the output is the system power change curve. This paper takes CEPRI36v8 system as an example of simulation, obtains the primary frequency modulation curve of the system by setting the shear load disturbance with PSASP, uses the disturbance data and the primary frequency modulation response data as training set, test set and validation set data, trains the prediction model of primary frequency modulation capability based on MEA-BP, and finds the prediction error is about 0.5% compared with the actual data, which proves that the proposed method can accurately and quickly predict the primary frequency modulation capability after the frequency deviation of the power network. This method can assist power dispatcher to analyze the primary frequency modulation response of power grid after power disturbance.

Keywords

• Primary frequency
• Prediction analysis
• Power system disturbance
• BP neural network
• Mind evolutionary algorithm