电缆接头缺陷在线监测技术研究

Study of Online Monitoring of Cable Joint Defects

随着电力能源需求增长,电缆化率不断提高,电缆接头缺陷易导致绝缘击穿,最终演变成永久性故障,因此迫切需要一种能对电缆接头缺陷实现有效在线监测的技术手段。针对XLPE电缆结构,分析了4类典型电缆缺陷,推导了局部放电信号的传输特性。利用光纤光栅超声传感器实现局部放电信号的在线采集,并通过小波分析法实现降噪处理。对处理后的波形进行特征提取,并输入到BP神经网络样本模型中进行多轮训练,输出4种典型电缆接头缺陷的辨识结果。

With the increasing demand for electricity and energy,the rate of cable utilization is constantly increasing.Defects in cable joints can easily lead to insulation breakdown and eventually result in permanent faults.Effective technological methods to online monitor cable joint defects are therefore of urgent necessity in grid operation.Four types of typical cable defects are analyzed with respect to XLPE cable structure,and transmission characteristics of partial discharge signals are derived.Utilizing fiber Bragg grating ultrasonic sensors,online collection of partial discharge signals is achieved and denoising is performed through wavelet analysis.By eigenvalueing the processed waveform and inputing it into BP neural network sample model for multiple rounds of training,fault distinguishing of different types of faults is realized.