【摘要】 针对现有配电网电缆故障定位的准确率和可靠性均偏低的问题，提出一种基于多元变分模态分解（multivariate variational mode decomposition,MVMD）与峭度的故障定位方法。鉴于MVMD模态数K及惩罚因子α对算法与信号的匹配度有较大影响，改进了MVMD的参数；利用改进的MVMD算法分解故障电流暂态分量，获得多个本征模态函数分量后分别计算其峭度，取峭度最大值所对应的模态为特征模态，并标定行波波头；修正电磁暂态下的行波波速，用双端法行波测距实现故障的精准定位。仿真结果表明：故障点距离电缆中间位置越近，其定位精度越高；当故障点靠近电缆两端，则其定位误差相对较大，但均在0.3%内。该方法不受故障电阻的影响，且定位结果具备高精度与高可靠性。更多还原

【Abstract】 To address the low accuracy and reliability of the cable fault location in distribution networks,a new fault method based on multivariate variational mode decomposition(MVMD)and Kurtosis calculation is proposed in this paper. First of all,given the mode number K and penalty factor[ α] of the MVMD have significant impacts on the matching degree between the algorithm and the signal,the MVMD parameters have been improved. The transient component of the fault current is decomposed to obtain multiple intrinsic mode function components,and the kurrosis of each component is calculated separately. The corresponding mode of the maximum kurtosis is taken as the characteristic mode and at the same time the travelling wave head is calibrated. Finally,the velocity of the traveling wave under electromagnetic transient is modified and the accurate fault location is realized with the double-ended travelling wave ranging. The simulation verification results suggest that the closer the fault point is to the middle position of the cable,the higher the precision is,on the contrary,the error is relatively larger,but all within 0.3%,and that the method is not affected by the fault resistance,and the location results are both accurate and reliable.更多还原