基于电弧模型及相关性算法的电缆早期故障检测研究

Research on cable early fault detection based on arc model and correlation algorithm

电力电缆早期故障具有持续时间短、峰值电压低等特点,现有的技术方案较难被检测,且随着时间推移会演变成永久性故障。为此,文中提出一种基于电弧模型及相似性算法的电缆早期故障检测方法。首先采集电缆送端电压信号,并基于互相关相似性函数进行异常变化识别;其次,对正弦波形进行拟合并计算其R^(2)值,根据该值超过阈值的次数做出故障是否存在的判定。所提方法能够实现电缆早期故障与其他类似非故障事件的有效区分。仿真结果表明,该方法对早期故障检测的准确度达到99%,具有较好的识别准确率和先进性。

The early fault of power cable has the characteristics of short duration and low peak voltage.The existing technical solutions are difficult to detect,and will evolve into permanent fault over time.A method of cables early fault detection based on arc model and similarity algorithm is proposed.By collecting the voltage signal at the cable end,abnormal changes are identified based on cross-correlation similarity function.The sine waveform is merged to calculate its R^(2) value and determine whether the fault is existed according to the number of times that the value exceeds the threshold exceeds the threshold.The proposed method can realize the effective distinction between early cable fault and other similar non-fault events.The simulation results show that the accuracy of this method for early fault detection can reach 99%,and it has good recognition accuracy and progressiveness.