Synchronous Condenser Reactive Power Output Model Based on DAG-CNN

The unique output characteristic of synchronous condenser reactive power can compensate reactive power of power grid in real time, but due to strong coupling and nonlinearity, general analysis method is difficult to accurately establish reactive power output model. A synchronous condenser reactive power output model based on deep learning is proposed, which takes field current and field voltage as its inputs, and reactive power and system voltage as its outputs. The simulation results of synchronous condenser reactive power regulation in PSCAD/EMTDC simulation software were used as training samples and test samples, and synchronous condenser reactive power output model based on Directed Acyclic Graph Convolutional Neural Network (DAG-CNN) was established. Simulation results show that the DAG-CNN model for synchronous condenser reactive power output can improve accuracy and generalization ability compared with traditional deep learning model, which can provide some references for the operation state prediction of synchronous condenser.