直流下交联聚乙烯电缆局部放电的试验特性研究

Test Research on DC Partial Discharges of Cross Linked Polyethylene Cable

为研究直流电压下交联聚乙烯(cross linked polyethylene,XLPE)电缆绝缘局部放电特性,搭建了一套XLPE电缆直流局部放电试验及检测系统,使用直流局放仪和高频电流法(high frequency current method,HFCT)检测。使用YJV 120mm2单芯电缆设计并制作了绝缘内部气隙、主绝缘表面划伤、高压端毛刺电晕、半导电层爬电4类绝缘缺陷模型,并使用Comsol Multi-physics软件仿真了不同缺陷模型的电场分布特性。使用恒压法试验,绘制了时间分辨局放图谱(time-resolved partial discharge,TRPD)记录放电发展过程,并基于统计特征绘制了H(Q,Δt)图谱和4类典型指纹图谱。综合分析试验和仿真研究结果表明:1)电晕缺陷模型电场畸变最严重,气隙缺陷次之;2)划伤缺陷和爬电缺陷的放电起始电压明显高于另外两者;3)4类缺陷模型的起始放电阶段和稳定发展阶段均存在差异,划伤缺陷起始放电量很大,并存在两个较明显的放电水平,气隙和电晕缺陷起始阶段放电量和放电重率均随时间递增;4)不同类型放电的H(Q,Δt)图谱差异较明显,指纹图谱随机性较大,其中电晕缺陷的指纹图谱线性规律最明显。

To study the DC partial discharge characteristics of cross linked polyethylene （XLPE） cable, we built a test system of XLPE cable under DC voltage, detecting by DC partial discharge detector and high frequency current method （HFCT）. YJV 120mm2 single core cables were chosen to make 4 kinds of defect models： cavity inside insulation, scratch on the insulation surface, corona of metal burr and creepage on semi-conductive layer. The electric field distributions of different models were simulated by Comsol Multi-physics. Constant DC voltage was applied on every model and the discharge development process was recorded by the time-resolved partial discharge （TRPD） map. The H（Q, At） maps and 4 kinds of fingerprints were drawn based on statistical characteristics. The results of test and simulation show that： 1） the electric field distortion of corona defect is most serious than others and cavity defect takes the second place;2） the initial discharge voltages of scratch and creepage defects are significantly higher than the others;3） There are differences between the initial discharge stage and stable development stage of defect models. The initial discharge capacity of scratch is large and has two apparent discharge levels. By contrast, the discharge capacity and discharge rate of cavity and corona defects increase gradually; 4） there are significant differences among H（Q, At） maps of different defects. However, the fingerprints have great randomness. The fingerprints of corona defect has a better regular.