【摘要】 针对传统电能质量扰动识别方法在处理复合扰动数据时特征提取复杂、识别精度低和运算量大等问题，提出了一种基于Lissajous曲线与YOLOv8网络的复合电能质量扰动识别新方法。该方法首先依据Lissajous曲线参数方程将一维电能质量扰动数据转换为二维椭圆轨迹图像，以增强复合扰动特征的显性表达能力；随后，使用YOLOv8网络作为识别模块对生成的Lissajous曲线图进行分类训练，显著提升对复合扰动的辨识能力；最后，对不同识别方法进行对比分析。实验结果表明，该方法能有效提取电能质量扰动特征，在不同噪声环境下对17类电能质量扰动均实现了高精度识别，显著改善了传统方法在复合扰动识别中精度不足的问题。更多还原

【Abstract】 To address the issues of complex feature extraction, low recognition accuracy, and high computational load in traditional methods for identifying composite power quality disturbances, this paper proposes a novel identification approach based on Lissajous curve and the YOLOv8 network. The method first converts one-dimensional power quality disturbance data into two-dimensional elliptical trajectory images using Lissajous curve parametric equations, thereby enhancing the distinct representation of composite disturbance features. Subsequently, the YOLOv8 network is employed as the recognition module to perform classification training on the generated Lissajous curve figures, significantly improving the identification capability for composite disturbances. Finally, a comparative analysis of different identification methods is conducted. Experimental results demonstrate that the proposed method can effectively extract power quality disturbance features, achieve high-accuracy identification of 17 types of power quality disturbances under different noise environments, and substantially alleviate the insufficient recognition accuracy of traditional methods in handling composite disturbances.更多还原