Study of Total Transfer Capability Based on Deep Learning Theory

Total Transfer Capability (TTC) is of great significance in monitoring the stability of the power system, guiding the optimal grid operation and power market construction. In this paper, the commonly used Repeated Power Flow algorithm (RPF) considering transient stability to calculate TTC is explained and analyzed in detail. Secondly, the basic architectures of Fully Connected Neural Network (FCNN), Stacked Autoencoder (SAE), Convolutional Neural Network (CNN), and Graph Neural Network (GNN) are introduced, and the RPF method is utilized to generate the TTC sample data to train the FCNN, SAE, CNN, and GNN models, and the models are compared in terms of their speeds and accuracies on the test set. It is found that, in this paper's regression prediction case study the FCNN model has better accuracy and speed, and the complex model is still to be optimized.