基于超声红外热像的电缆终端局部放电缺陷检测方法

Detection Method of Partial Discharge Defects in Cable Terminals Based on Ultrasonic Infrared Thermography

电缆终端局部放电缺陷特征短暂,缺陷范围与外部环境纠缠,很难准确定位,需要结合温度特征和模式识别特征共同检测,本文利用超声红外热成像的优势,提出基于超声红外热像的电缆终端局部放电缺陷检测方法,方法利用图像梯度化、灰度化处理采集到的电缆终端局部放电缺陷特征超声红外热成像图,并通过智能模式识别处理方法抑制采集图像的复杂背景,删除包含在电缆终端局部放电缺陷特征红外图像中的大面积地物及地面;根据K-means聚类算法,圈定疑似局部放电缺陷特征范围,构建局部放电缺陷范围模板,经匹配参考范围后,得出疑似局部放电缺陷范围的温度特性信息,诊断电缆终端是否存在局部放电缺陷。实验结果表明,该方法可有效获取电缆终端局部放电缺陷部位,检测不同类型的电缆终端局部放电缺陷的平均精准率高达98%,平均漏检率为1%。

The partial discharge defect characteristics of cable terminals are short, and the defect range is entangled with the external environment, making it difficult to accurately locate. It must be detected along with the temperature characteristics and pattern recognition characteristics. In this paper, using the advantages of ultrasonic infrared thermal imaging, a partial discharge defect detection method for cable terminals based on ultrasonic infrared thermal images is proposed. This method uses image gradient grayscale to collect an ultrasonic infrared thermal image of the partial discharge defect characteristics of a cable terminal, suppress the complex background of the collected image via an intelligent pattern recognition processing method, and delete large-area ground objects and surfaces contained in the image. Using the K-means clustering algorithm,the characteristic range of the suspected partial discharge defects is delineated, and the partial discharge defect range template is constructed. After matching the reference range, information on the temperature characteristics of the suspected partial discharge defect range is obtained to diagnose whether there are partial discharge defects in the cable terminal. The experimental results show that this method can effectively obtain the partial discharge defects of cable terminals. The average accuracy of detecting different types of partial discharge defects in cable terminals was as high as 98%, and the average missed detection rate was 1%.