基于图像处理的GIS局部放电图谱信号恢复（英文）

Signal Recovery of GIS Partial Discharges Graph Based on Image Processing

由于局部放电检测设备厂家的原始信号不公开、厂家间数据格式不统一等原因,现有存储的局部放电数据多以图谱形式存在。在这种情况下,大量原始记录和信息已经丢失,不利于后续的分析和诊断。通过图像处理手段,提出了一种可行的GIS局部放电图谱信号恢复算法。以Hanbit便携式GIS局放设备得到的PRPS图谱为例,首先通过双边梯度提取坐标轴网格线,然后在不同的色彩空间检测信号。结合检测到的网格线和放电信号点,可以恢复放电信号的相位和幅值信息。利用这些重建的信息,可以更加便捷地利用现有成熟的针对信号的局部放电自动诊断算法。将非结构化的图谱数据转化为结构化的信号对于电力大数据的研究也具有重要意义。

Since the raw data of partial discharge is not accessible in most cases and the formats between manufacturers are different,the stored partial discharge data is always in the form of graph.ln this case,much original information has been lost, and it＇s difficult to make further analysis and diagnosis.In this paper,we present a feasible algorithm to reconstruct signals in a GIS partial discharge graph through image processing ways.Take the PRPS graph created by Hanbit portable GIS PD equipment for example, we firstly extract the grid lines of the coordinates by bilateral gradient, and then detect signals in different color spaces.The phase and amplitude could be recovered by combining the grid lines and signal points.With the reconstructed information,it＇s convenient to make use of many state-of-the-art automatic PD fault diagnosis methods aiming signals.To convert the unstructured graph data to structured signals is also important for the study of big data in electric system.