Abstract：Aiming at the shortcomings of traditional power quality disturbance classification methods that are difficult to manually select features and low accuracy, on the basis of traditional convolutional networks, drawing on the ideas of inception and residuals, combined with hybrid pooling and efficient channel attention mechanism, a power quality disturbance identification method based on channel selection multi-scale fusion deep residual network (CSSF-ResNet) is proposed. Multi-scale convolution is used to extract features of different scales, and global mixed pooling is combined with efficient channel attention mechanism. Feature screening is carried out in the channel dimension. Effective features are mined, and residual connections are introduced to form a CSSF-ResNet. The simulation results show that the proposed method has the advantages of high classification accuracy and strong noise robustness.
刘伟, 王凯. 基于通道选择多尺度融合深度残差网络的电能质量扰动识别[J]. 电气技术, 2023, 24(5): 11-15.
LIU Wei, WANG Kai. Power quality disturbance identification of multi-scale fusion depth residual network based on channel selection. Electrical Engineering, 2023, 24(5): 11-15.
2023, Vol. 24 
