基于相空间重构和卷积神经网络电能质量扰动分类.PDFVIP

第 46 卷 第 14 期 电力系统保护与控制 Vol.46 No.14 2018 年 7 月 16 日 Power System Protection and Control Jul. 16, 2018 OI: 10.7667/PSPC17 1080 基于相空间重构和卷积神经网络的电能质量扰动分类 陈 伟，何家欢，裴喜平 (兰州理工大学电气工程与信息工程学院，甘肃 兰州 730050) 摘要：为了提高电能质量扰动信号分类的准确率，首先利用相空 重构将一维时 序列电能质量扰动信号重构到 多维相空 ，获得扰动信号轨迹并投影到二维相平面，形成二维轨迹图像。然后对该图像进行二值化处理，减少 信号的数据量，凸显轨迹轮廓。最后通过卷积神经网络对处理后的轨迹图像进行特征提取，并对相应的扰动信号 进行分类识别。在卷积神经网络框架 Caffe 下进行仿真实验，仿真结果表明该方法具有很高的识别准确率和良好 的抗噪能力。 关键词：电能质量；扰动分类；相空 重构；深度学习；卷积神经网络 Classification for power quality disturbance based on phase-space reconstruction and convolution neural network CHEN Wei, HE Jiahuan, PEI Xiping (School of Electrical Engineering and Information Engineering, Lanzhou University of Technology, Lanzhou 730050, China) Abstract: In order to improve the accuracy of the power quality disturbance signals classification, firstly, the one-dimensional time series power quality disturbance signals are reconstructed into the multidimensional phase space by phase-space reconstruction theory, thereby obtaining the disturbance signals traj ectory and proj ecting to the two-dimensional plane, and forming a two-dimensional images. Then these images are binarized to reduce the amount of data of the signal and highlight the traj ectory profile. Finally, the convolution neural network is used to extract the features of traj ectory images and achieve the classification and identification of the corresponding disturbance signals. Simulation experiments are carrie