1.Southern Power Grid Company;2.Changsha University of Science &3.Technology;4.hunan university
特高频（UHF）法局部放电检测已经成为高压GIS及变压器等设备常用的运维检测手段，但由于该方法是通过测量电磁波间接检测局部放电强度，使得基于该方法的局部放电模式识别技术准确率不高，而且一直没有得到很好的解决。本文引入TEm1模截止频率作为PD UHF信号时频分布图的分割依据，对单一绝缘缺陷激发的PD UHF信号进行预处理、时频分布图分割、特征参数提取和选择、PSO-ELM识别等操作，最终准确地判断出PD信号类型，提高了GIS绝缘缺陷类型的识别准确率。这种分割方法增强了图谱特征的空间分布信息，对改善特高频法识别局部放电的准确率具有实际应用价值。
Detection of partial discharge by UHF signal has become a common operation and maintenance detection means for High Voltage GIS, transformers and other equipment. However, due to the indirect detection of PD intensity by measuring electromagnetic wave, the PD pattern recognition based on this method has no high accuracy and not been well developed. In this paper, the cutoff frequency of TEm1 mode is introduced as the division basis of time-frequency distribution of PD UHF signal. The PD UHF signal excited by a single insulation defect is used signal preprocessing, time-frequency distribution segmentation, feature parameter extraction and selection, PSO-ELM recognition and other operations. Finally, the PD signal type was determined, which improves the accuracy of GIS insulation defect. This segmentation method enhances the spatial distribution information of image features, and it is of great practical value to improve the accuracy of UHF method in PD identification.