基于NSST域增强的直流XLPE电缆局放识别方法

发布时间：2018-11-01 19:19

【摘要】：针对直流交联聚乙烯(XLPE)电缆局放信号图的对比度差和边缘清晰度低,影响电缆缺陷识别率的问题,提出一种基于非下采样剪切波变换(NSST)域增强的直流电缆局放识别方法。首先采用NSST对不同缺陷类型的电缆局放信号图进行分解,得到低频子带图和高频方向子带图,其中,低频子带图进行布谷鸟优化多尺度Retinex算法(CS-MSR)增强处理,高频方向子带图进行亮度保持动态模糊直方图均衡算法(BPDFHE)增强处理。然后对NSST各子带图进行加权重构,得到增强的直流电缆局放信号图。最后提取直流电缆局放信号图中每个子带图的最大放电次数、平均放电时间间隔、平均放电量等共计72个特征参数,并代入到线性核支持向量机(L-SVM),高斯核支持向量机(G-SVM),多项式核支持向量机(P-SVM)和多核支持向量机(M-SVM)识别模型中进行分析。实验结果证明:基于NSST域增强的直流电缆局放信号图与原始信号图相比,在M-SVM情况下,整体缺陷识别准确率提高9.46%以上。运用新提出的方法进行局放识别时,信号图细节更加丰富,且提高了电缆局放信号图的缺陷识别率,为直流电缆局放缺陷识别提供了新的思路。
[Abstract]:Aiming at the problem of low contrast and low edge definition of DC crosslinked polyethylene (XLPE) cable partial discharge signal, the defect recognition rate of cable is affected. A method of DC cable partial discharge recognition based on non-downsampling shear wave transform (NSST) domain enhancement is proposed. Firstly, the low-frequency sub-band map and the high-frequency directional sub-band map are obtained by using NSST to decompose the partial discharge signals of different defect types of cables. Among them, the low-frequency sub-band map is enhanced by the Cuckoo optimization multi-scale Retinex algorithm (CS-MSR). The high frequency directional subband map is enhanced by (BPDFHE), a dynamic fuzzy histogram equalization algorithm. Then, the weighted reconstruction of each sub-band diagram of NSST is carried out, and the enhanced DC cable partial discharge signal diagram is obtained. Finally, the maximum discharge times, the average discharge interval and the average discharge amount in each sub-band diagram of DC cable partial discharge signal diagram are extracted, and a total of 72 characteristic parameters are obtained, which are replaced by linear kernel support vector machine (L-SVM). Gao Si kernel support vector machine (G-SVM), polynomial kernel support vector machine (P-SVM) and multi-kernel support vector machine (M-SVM) are analyzed. The experimental results show that, compared with the original signal diagram, the accuracy of the whole defect recognition is increased by more than 9.46% in the case of M-SVM, compared with the original signal diagram. The enhancement of DC cable partial discharge signal based on NSST domain is more than 9.46%. When using the new method for partial discharge recognition, the detail of signal diagram is more abundant, and the defect recognition rate of partial discharge signal diagram of cable is improved, which provides a new way of thinking for partial discharge defect recognition of DC cable.
【作者单位】：上海交通大学电气工程系;国网浙江省电力公司电力科学研究院;
【基金】：国家重点研发计划(2016YFB0900705)~~
【分类号】：TM855

【相似文献】