CNN based warning algorithm for AC transmission capacity exceeding limit in distribution

Research on a CNN based warning algorithm for AC transmission capacity exceeding limits in distribution networks, in order to adapt to complex and ever-changing power grid environments, and improve the safety and efficiency of power grid operation. Construct a distribution network AC line transmission capacity exceeding limit warning based on CNN-LSTM-PSO. In the offline training stage, the PCA algorithm is used to reduce the dimensionality of the preprocessed historical operation data of the distribution network. After selecting key feature data, training samples are generated. After random sampling, different working condition sample data are obtained. The first fault limit transmission capacity is constructed with the maximum power transmission margin variable as the optimization objective and meets the constraints of stable operation inequality, transient power angle, transient voltage, etc. The continuous power flow algorithm is used to solve the problem, determine the corresponding working condition limit transmission capacity calculation result, and combine it with each input sample to form a complete sample data for training. The warning parameters are optimized using PSO algorithm. In the online warning stage, real-time operation data of the distribution network is used as input. After capturing the input data features using a CNN network, the LSTM network is used to learn the correlation between features. The fully connected layer outputs the transmission capacity prediction result, which is compared with the threshold set according to the maximum transmission capacity to complete the transmission capacity exceeding limit warning. The experimental results show that the algorithm can achieve warning of transmission capacity exceeding the limit under different working conditions, with RMSE, MAPE, and R2 values of 1.993, 0.011, and 0.672, respectively; The predicted loss is around 0.01.