清洁能源电网大直径电缆破损故障无人机勘探研究

A Study on the UAV Exploration of Large Diameter Cable Breakage Faults in Clean Energy Grids

中文关键词:清洁能源电网  点云数据  电力线建模  故障检测  子空间特征差异算法  高程密度分割算法  无人机

keywords:clean energy grid  point cloud data  power line modeling  fault detection  subspace feature difference algorithm  elevation density segmentation algorithm  UAV

中文摘要:电网电缆线路普遍较长，破损后局部放电信号在电缆内衰减较快，使故障位置的精确定位更加困难，导致故障检测难度较大，从而降低了检测的准确性。提出清洁能源电网大直径电缆破损故障无人机勘探方法。应用无人机采集电力线点云数据，利用子空间特征差异算法优化点云数据的高程阈值，采用高程密度分割算法完成塔杆与高程剩余物的区分识别；构建水平投影来拟合电缆外部绝缘层直线模型和抛物线模型，识别电力线形态。将电缆形态数据输入至基于决策规则集所建立的粗径向基融合模型中，完成大直径电缆破损故障检测。实验结果表明：该方法能够精准提取电网的点云数据，电网大直径电缆破损故障检测准确度高于90%，检测时间低于10 min，证明了所提方法的故障检测精度高、检测效率高，整体检测效果好。

Abstract:Power grid cable lines are typically extensive， and the partial discharge signal diminishes rapidly after the cable damage， complicating the accurate location of the fault. This difficulty in fault detection consequently reduces the accuracy of detection. To tackle this challenge， a drone-based exploration method for detecting damage faults in large-diameter cables within clean energy power grids is proposed. Unmanned Aerial Vehicles （UAVs） are employed to gather power line point cloud data. A subspace feature difference algorithm is utilized to optimize the elevation threshold of the point cloud data， and an elevation density segmentation algorithm is used to differentiate tower poles from elevation residues. The external insulation layer of the cable is modeled using horizontal projection fitting to construct linear and parabolic models， identifying the power line shape. The cable shape data is then input into a rough radial basis fusion model， established on the basis of a decision rule set， to complete the detection of large-diameter cable damage. Experimental results demonstrate that this method can accurately extract point cloud data from power grids. The detection accuracy for large-diameter cable breakage faults exceeds 90%， and the detection time is under 10 minutes， indicating that the proposed method offers high fault detection accuracy， efficiency， and an overall superior detection effect.