基于深度学习的配电网架空线故障识别技术的研究

Research on Distribution Overhead Line Fault Recognition Technology Based on Deep Learning

针对配电网线路走廊环境多样复杂,传统的人工巡线技术耗费大量人力物力财力的现状,本文提出基于深度学习算法的配电网架空线故障识别技术。通过分析架空线的图像信息,提出了一种基于深度学习的检测方法,将图像的线性特征提取出来以进行分析。并提出了基于K-means聚类算法的架空线故障评估方法,对识别出的架空线故障及故障隐患进行影响程度评估,通过对智能图像采集系统采集到的架空线图像进行故障检测,检测结果表明,该方法可有效检测出有故障隐患的架空线路,对架空线路巡检提供了一种新技术。

Aiming at the situation that the corridor environment of distribution network overhead lines is complicated,and the traditional manual patrol technology consumes a lot of manpower,material and financial resources,a fault recognition technology of distribution network overhead lines based on deep learning algorithm is proposed in this paper.Through the characteristics analysis of the images of overhead lines,the phase-consistent overhead lines image feature detection method based on deep learning is proposed and the linear characteristics are analyzed.What’s more,the overhead lines fault evaluation method based on K-means clustering algorithm is proposed.And the influence degree of the identified overhead lines faults and hidden faults is evaluated.Take the overhead lines images collected by the acquisition system of a city as an example,the test results show that the method can effectively detect the potential faults of overhead lines,which prove to provide a new technology of overhead lines recognition.