Research on Load State Identification Method Based on CNN-Transformer

Load monitoring is a critical link in the development of power grid technology, and the traditional load feature information is feature extraction and feature mining by simple convolutional neural network due to the inability to effectively combine multiple input timing component information when using CNN for load state identification. For this reason, a load state identification method based on the combination of CNN and Transformer is proposed. CNN mines the features of steady-state load and abnormal operating condition load data. In contrast, Transformer’s multi-attention mechanism is utilized to obtain the effects of abnormal load changes on steady-state load, and simulation and experimental results verify the correctness of the proposed method.