Detection Method of Power Quality Disturbances Based on Double Resolution S Transform and Improved Multi-scale ResNet
Accurate power quality disturbance detection is significant for improving power quality problems in smart grids and ensuring safe and reliable operation.Therefore,this paper proposes a detection method of power quality disturbance signals based on double-resolution S transform(DRST)and improved multi-scale ResNet.Firstly,the power quality disturbance signal feature vectors are extracted accurately and quickly using the DRST.Secondly,it proposes to use Mish function to improve ResNet instead of the traditional ReLU activation function,and then the improved ResNet with different convolution kernel sizes is utilized for feature learning and classification of complex power quality disturbance signals.Without increasing the network parameters,the paper proposes to use the lightweight efficient channel attention(efficient channel attention,ECA)to assign larger weight values to the important features that have a greater impact on the classification results of power quality disturbance detection.The experimental results show that the proposed method has higher accuracy and better noise immunity compared with other power quality disturbance detection methods.
double resolution S transformpower quality disturbanceresidual network(ResNet)attention mechanismactive function
万方数据
维普
