Power quality disturbance identification based on gated space-time graph neural network

With the increasing scale of power systems, the assurance of power quality is facing unprecedented challenges. To improve the operation quality of the power system, a power quality disturbance identification method is studied and designed. The process uses a graph convolutional neural network as the basis to capture local and global features of the power system network, the fine-grained preference information of the signal is exploited, and the weighted sum is performed to obtain the final perturbation representation result. The experimental results show that in the recall test, the research method has a recall value of 0.328 for the number of disturbances of 1 in the voltage dips scenario, and the recall value increases to 0.491 when the number rises to 11. This indicates that the research method is able to effectively and accurately identify the disturbance information.