【作者】 瞿敏；

【导师】 张炳达；

【作者基本信息】 天津大学 ， 电力系统及其自动化， 2004， 硕士

【摘要】 电力电缆故障是电力系统中的常见故障，电缆测距是排除电缆故障的前提条件，准确的电缆测距可以缩短发现故障点的时间，有利于快速排除故障，减小由电缆故障带来的损失。本文对不同长度的电缆进行了低压脉冲实验，通过对数据进行时频分析进一步了解了低压脉冲在电缆中的传播规律，针对电缆测距中出现的问题提出了新的解决方法。由于电缆本身存在损耗，使行波在电缆传播过程中发生畸变现象，反射脉冲上升沿变缓，反射脉冲宽度增大，这不利于反射脉冲起始点的正确判别。本文提出了曲线拟合法确定反射波起始点。所谓曲线拟合法是：利用最小二乘法对反射波上升沿的一段数据进行二次曲线拟合，认为拟合曲线的过零点是反射脉冲起始点。为了进一步提高测距精度，利用二次函数对测距结果进行误差修正。结果表明采用曲线拟合法能使测距绝对误差在3个采样点以内，与传统的反射波起始点确定方法相比有较大优势。电缆测距过程中往往会受到各种噪声的影响，降低了采集信号的信噪比，对测距结果造成了不良影响。本文根据真实信号和噪声信号在小波分解各个尺度上不同的传播规律，提出了改进的软阈值消噪法，有效地促进了信噪分离，提高了采集信号的信噪比。更多还原

【Abstract】 Power cable fault occurs regularly in power system. Fault location is the precondition to eliminate the fault. Accurate fault location can reduce the time to find out the fault point and the cost of power cable cut. This paper has carried on the low voltage pulse experiment to the cable of different lengths. The law in the cable of low voltage pulse is known well through time-frequently analysis on the data, and the new solution to the problem that appeared in the experiment is put forward. Because of the loss of cable, the traveling wave is distorted and the reflect pulse rise becomes slowly, which don’t benefit to the accurate judge of start point of the reflect wave. The curve fit method is proposed to compute the start point of reflection wave, and the least squares method is used to fit the data of reflect pulse rise. The wave cross zero point is the considered as the start point of reflection wave. The conic section method is adopted to revise the result. The results show that the start points of reflect wave can be distinguished accurately. The absolute error is no more than 3 sampling points, so it has more advantages compared with the traditional methods to judge the start point of reflect wave.The real signal often influenced by various kinds of noises, which reduce the SNR of gathering signal and put blight on fault location. The improved wavelet soft-threshold method is presented based on the different propagation property of real signals and white noises under the wavelet transform. It can accelerate the separation of the signal from the noise and get higher SNR for the reconstructed signal.更多还原