高压电缆阻水缓冲层局部放电行为与放电烧蚀过程识别  

Partial Discharge Behavior and Discharge Ablation Process Identification of High Voltage Cable Water Barrier Buffer

作　　者：梁帅 白明华 LIANG Shuai;BAI Minghua(Tangshan Zhaoxin Solar Power Co.,Ltd.,Tangshan 063000,China)

机构地区：[1]唐山招新太阳能发电有限公司,河北唐山063000

出　　处：《机械与电子》2025年第1期16-20,28,共6页Machinery & Electronics

摘　　要：为研究故障下的局部放电(PD)行为,根据电缆缓冲层的实际结构进行了交流电压下的局部放电试验。根据PD试验的统计特征,整个放电过程分为初始阶段、发展阶段、停滞阶段和缓冲层烧蚀4个阶段,各阶段PD活动具有不同的特征。基于此试验结果设定了偏斜度、陡峭度、放电因素和相位不对称度多参数放电图谱统计特征参量;并提出基于证据理论多源数据融合(图谱统计特征与图谱图像特征)的放电烧蚀过程识别方法。最后对比了4种识别方法,所提出的通过D-S证据理论进行放电图谱图像(结合CNN)与放电图谱特征(结合SVM)数据融合识别的方法,识别准确率平均水平最高,对放电烧蚀过程的平均识别率为94.13%,最高识别率为96.41%。In order to study the partial discharge(PD)behavior under fault,the partial discharge test under AC voltage is carried out according to the actual structure of cable buffer.According to the statistical characteristics of PD test,the whole discharge process is divided into four stages:initial stage,development stage,stagnation stage and buffer layer ablation.PD activities in each stage have different characteristics.Based on the experimental results,the statistical characteristic parameters of multi parameter discharge patterns,such as skewness,steepness,discharge factor and phase asymmetry,are determined;the recognition method of discharge ablation process based on evidence theory multi-source data fusion(spectrum statistical features and spectrum image features)is proposed.Finally,four identification methods are compared.The method of data fusion identification of discharge pattern image(combined with CNN)and discharge pattern feature(combined with SVM)based on D-S evidence theory proposed in this paper has the highest average recognition accuracy.The average recognition rate of discharge ablation process is 94.13%,and the highest recognition rate is 96.41%.

关 键 词：电缆缓冲层 局部放电 烧蚀故障 数据融合 

分 类 号：TM216[一般工业技术—材料科学与工程]