多超声波传感器局部放电源定位方法的研究

Research on partial discharge source localization method of multi-ultrasonic- sensor

随着工业的快速发展,电力设备能否正常运行直接关系到电网的安全、稳定。绝缘劣化是电力设备最易发生的故障,其中最重要的原因就是局部放电。超声波检测方法是局放检测定位最传统也是最主要的方法。文章采用LevenbergMarquardt算法求解一个多元非线性方程组得到局放源坐标。测量时运用多个传感器对局放源进行定位,这样可以避免由于个别传感器测量不准引起的偶然误差。再利用K-means聚类算法对Levenberg-Marquardt算法得到的结果进行聚类,选取元素最多的类作为最佳解组,然后对该组求平均值,得到的结果即为最接近实际局放源的结果,这样得到的最终结果更加准确。

With the rapid development of industry,electrical equipment is important to the safety of electrical network. Deteriorations of the insulation is the major frequent malfunctions,the most crucial cause is partial discharge（ PD）. The method for ultrasonic location of transformer PD is the main method of the current transformer insulation detection. In this paper,Levenberg-Marquardt algorithm is used to solve a nonlinear system of equations,and then it is put in the source coordinates. The multiple-sensor is used to locate the source of PD,which can avoid the error caused by the inaccurate measurement of individual sensor. K-means clustering algorithm is used to get the results of Levenberg-Marquardt clustering,a class of element is selected as the best solution group,and then the average value of the group is solved. The result obtained is the most close to the actual results of the source,so that the final result is more accurate.