基于一维残差收缩网络的电能质量复合扰动识别

Power Quality Compound Disturbance Identification Based on One-dimensional Residual Shrinkage Network

电网中强噪声的干扰会严重影响电能质量复合扰动识别,为提高电能质量复合扰动识别准确率,提出一种基于一维残差收缩网络的电能质量复合扰动识别方法。该方法能够以原始数据作为输入避免有效特征的丢失,采用子网络自动设置阈值对各个特征通道进行软阈值化,并通过加宽卷积层进一步增强网络抗噪性。仿真实验结果表明:所提方法在强噪声干扰下能快速准确识别电能质量复合扰动。

The interference of strong noise in the power grid seriously affects the identification of power quality compound disturbances.In order to improve the identification accuracy of power quality compound disturbances,this paper proposes a power quality compound disturbance identification method based on one-dimensional residual shrinkage network.The method can use the original data as input to avoid the loss of effective features,use the sub-network to automatically set thresholds to soft-threshold each feature channel,and further enhance the noise immunity of the network by widening the convolutional layer.The simulation results show that the method proposed in this paper can quickly and accurately identify the composite disturbance of power quality under strong noise interference.