基于振荡波的电缆多源局放检测及定位方法

发布时间：2018-09-14 11:29

【摘要】：局部放电是电缆发生故障的前期征兆,在局部放电易发期,通过快速有效的手段,检测电缆绝缘局部放电的程度,根据检测的结果判断电缆的绝缘健康状态,同时,利用电缆局部放电信号,结合行波测距的方法,对放电位置进行精确的测定,从而采取相应的解决措施,对电缆乃至电网的安全运行有着重要的意义。振荡波测试系统(OWTS)是近年来发展起来的用于电缆绝缘检测的新方法,并受到了一致的推广。本文以振荡波测试系统为基础,分析了系统的组成结构、参数设置及具体工作原理,并在仿真平台中搭建了系统模型,模拟局放发生及检测过程;在仿真的基础上,成功研制出了一套小型的振荡波测试系统,通过人为设置电缆尖端缺陷,对系统进行了试验,从试验结果来看,所研制系统成功地模拟了振荡波测试系统的工作过程,并采集到了局放信号,且局放信号的分布特性符合尖端放电的特性,证明了所研制系统的有效性,为后续的研究工作奠定了基础。基于振荡波测试系统所获得的局放信号,其中含有大量的白噪声干扰及窄带干扰,考虑现有的白噪声去噪算法已经较为成熟,本文仅针对局部放电窄带干扰提出一种基于频域估计与时域重构相结合的窄带干扰抑制算法。算法根据局放信号和窄带干扰在时域和频域上的分布差异,对信号采样数据进行分段后,通过分析各段信号频谱的相似性来实现窄带干扰的个数确定及窄带频率的粗略估计,再利用频域插值实现窄带频率的精确估计,进而结合傅里叶级数展开,对窄带干扰进行时域重构,最后与原始信号相减以实现窄带干扰的抑制。通过仿真及实测数据进行算法对比,验证了所提算法在提高窄带干扰参数估计及信号重构精度上的优势,以及在不同工况下良好的窄带抑制效果,同时也验证了所提算法在抗白噪声方面的稳定性,能够满足现场复杂工况下的窄带抑制应用需求。在利用振荡波测试系统所获得的局放信号对局放源进行定位的过程中,为减小噪声干扰及反射干扰的影响,并实现多个局放源的定位,本文提出一种基于局放匹配对识别及利用多组局放匹配对进行多源局放定位的算法,来提高在上述干扰工况中的定位精度。该算法利用局放信号的幅值突变特点以及在电缆中的传播特性来识别局放匹配对,再基于局放匹配对识别结果进行多源局放的定位,进而结合DBSCAN信号聚类方法来修正局放定位结果以实现对多源局放的精确定位。仿真结果表明,该多源局放定位算法能够在不同信噪比工况下,有效识别局放匹配对并精确估计出局放源的个数和位置,同时,由于利用了多组局放匹配对进行定位,其精确性比利用单组局放匹配对定位的精确性更高。理论研究和仿真结果表明,本文提出的基于振荡波的电缆多源局放检测及定位方法,在电缆局部放电信号的检测、去噪以及定位三个工作方面取得了一定效果,对电网的稳定运行提供了重要的安全保障。
[Abstract]:Partial discharge is the precursor of cable failure. In the period of partial discharge, the degree of partial discharge in cable insulation can be detected quickly and effectively, and the insulation health of cable can be judged according to the result of detection. At the same time, the discharge position can be accurately determined by using the partial discharge signal of cable and traveling wave ranging method. The oscillatory wave test system (OWTS) is a new method developed in recent years for cable insulation testing and has been widely used. Based on the oscillatory wave test system, this paper analyzes the structure of the system, the setting of parameters and the application of OWTS. The system model is built in the simulation platform to simulate the occurrence and detection of partial discharge. On the basis of simulation, a set of small-scale oscillatory wave testing system is successfully developed. The system is tested by setting the cable tip defect artificially. From the test results, the system successfully simulates the oscillatory wave. The working process of the system is tested, and the PD signal is collected, and the distribution characteristics of the PD signal conform to the characteristics of the sophisticated discharge, which proves the validity of the developed system and lays a foundation for further research. Considering that the existing white noise denoising algorithms are mature, this paper proposes a narrowband interference suppression algorithm based on the combination of frequency domain estimation and time domain reconstruction for narrowband partial discharge interference. The number of narrow-band interference and the rough estimation of narrow-band frequency can be achieved by the similarity of signal spectrum. Then the accurate estimation of narrow-band frequency can be achieved by frequency-domain interpolation. Then the time-domain reconstruction of narrow-band interference can be carried out by combining Fourier series expansion. Finally, the suppression of narrow-band interference can be achieved by subtracting from the original signal. The results show that the proposed algorithm has advantages in improving the precision of narrowband interference parameter estimation and signal reconstruction, and has good narrowband suppression effect under different operating conditions. At the same time, the stability of the proposed algorithm in white noise suppression is also verified, which can meet the application requirements of narrowband suppression under complex working conditions. In order to reduce the influence of noise interference and reflection interference and realize the localization of multiple PD sources, a localization algorithm based on PD matching pairs and multiple PD matching pairs is proposed in this paper to improve the jammer. The algorithm uses the amplitude mutation characteristics of the PD signal and the propagation characteristics in the cable to identify the PD matching pairs, and then locates the multi-source PD based on the PD matching pairs, and then uses the DBSCAN signal clustering method to modify the localization results of the PD to achieve accurate localization of the multi-source PD. The real results show that the algorithm can effectively identify the matching pairs of PD and accurately estimate the number and location of PD sources under different SNR conditions. At the same time, the accuracy of localization is higher than that of localization using single matching pairs of PD. The results show that the method of detecting and locating multi-source PD signals based on oscillatory waves in this paper has achieved certain results in detecting, denoising and locating the PD signals of cables, which provides an important security guarantee for the stable operation of power grid.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM855

【相似文献】