作者：王宁1, 王雁冰2, 杨健3, 欧阳跃4, 王恩路1, 焦嵩鸣5
作者单位：1. 中国广核新能源控股有限公司运维事业部, 北京 100070;
2. 中广核(北京)新能源科技有限公司, 北京 100071;
3. 中国广核新能源新疆分公司, 新疆 乌鲁木齐 830000;
4. 中国广核新能源贵州分公司, 贵州 贵阳 550081;
5. 华北电力大学自动化系, 河北 保定 071003
关键词：风力发电机;无人机;Yolov5;正对风机;精准巡视
摘要：
应用旋翼无人机对风力发电机进行自主巡视时，需正对风机采集全景图，计算桨叶空间位置，进而对巡视航线和航点做出规划。为此提出一种风机自主巡视的无人机控制策略。首先，采用Yolov5神经网络模型对风机各部位进行提取，同时利用风机各部位像素坐标计算无人机位姿修正量并设计控制器，使无人机到达正对风机的位置对其全景图进行采集；其次，根据全景图计算桨叶叶尖的空间坐标，规划出巡视航线及初始航点，达到对桨叶的全面图像采集；最后，为克服定位误差等因素的干扰，设计控制器使无人机可依据图像信息进行位姿调整，使桨叶保持在图像中心位置，达到对风机精准巡视的目的。仿真试验表明，所提出的策略可达到对风力发电机自主巡视的目的，其中的设计思路也可为其他电力设备的巡检方案提供参考。
Research on control strategy of UAV for autonomous inspection of wind turbine
WANG Ning, WANG Yanbing, YANG Jian, OUYANG Yue, WANG Enlu, JIAO Songming
1. Operation and Maintenance Department, CGN New Energy Holdings Co., Ltd., Beijing 100070, China;
2. CGN (Beijing) New Energy Technology Co., Ltd., Beijing 100071, China;
3. CGN New Energy Holdings Co., Ltd(Xinjiang)., Urumqi 830000, China;
4. CGN New Energy Holdings Co., Ltd(Guizhou)., Guiyang 550081, China;
5. Department of automation, North China Electric Power University, Baoding 071003, China
Abstract: When the rotor UAV is used to inspect the wind turbine independently, it is necessary to collect the panorama of the wind turbine, calculate the spatial position of the blades, and then plan the inspection route and inspection point.Therefore, a UAV control strategy for autonomous inspection of wind turbines is proposed in this paper. Firstly, the Yolov5 neural network model is used to extract each part of the wind turbine. At the same time, the position and attitude correction of the UAV is calculated by using the pixel coordinates of each part of the wind turbine, and the controller is designed to make the UAV reach the position of directly facing the wind turbine to collect its panorama. Secondly, the spatial coordinates of the blade tips are calculated according to the panorama, and the patrol route and initial inspection points are planned to achieve a comprehensive image acquisition of the blades. Finally, in order to overcome the interference of positioning errors and other factors, a controller is designed so that the UAV can adjust the position and attitude according to the image information, so that the blade can be kept in the center of the image, so as to achieve the purpose of precision patrol of the wind turbine. The simulation results show that the strategy proposed in this paper can achieve the purpose of independent inspection of wind turbines, and the design ideas can also provide reference for the inspection schemes of other power equipments.
Keywords: wind turbines;UAV;Yolov5;facing wind turbines;precision inspection
2023, 49(9):156-162,171 收稿日期: 2022-4-7;收到修改稿日期: 2022-6-2
基金项目:
作者简介: 王宁（1983-），男，河北保定市人，高级工程师，硕士，主要从事新能源大数据分析建模。
参考文献
[1] 赵坤. 迎接新能源发展更大动能[N]. 中国电力报, 2022-01-24.
[2] 国家能源局新闻发言人. 国家能源局2021一季度网上新闻发布会文字实录[J]. 中国电业, 2021(2): 26-29
[3] PLIEGO M A, GARCIA M P, LEV B. Optimal decision-making via binary decision diagrams for investments under a risky environment[J]. International Journal of Production Research, 2017, 55(17-18): 5271-5286
[4] SADEGHIAN R, SAREH S. Multifunctional arm for telerobotic wind turbine blade repair [C]. //Proceedings of the 2021 IEEE International Conference on Robotics and Automation (ICRA), 2021: 6883-6889.
[5] 赵元星, 汪建文, 张立茹, 等. 气动力和离心力对风力机叶片应力影响研究[J]. 太阳能学报, 2021, 42(2): 225-232
[6] 张英, 汪勇, 廖如超, 等. 三维激光雷达无人机树障实时测量系统研究[J]. 中国测试, 2021, 47(11): 147-152
[7] MOOLAN-FEROZE O, KARACHALIOS K, NIKOLAIDIS D N, et al. Simultaneous drone localisation and wind turbine model fitting during autonomous surface inspection [C]//Proceedings of the 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019: 2014-2021.
[8] KANELLAKIS C, FRESK E, MANSOURI S S, et al. Towards visual inspection of wind turbines: a case of visual data acquisition using autonomous aerial robots[J]. IEEE Access, 2020, 8: 181650-181661
[9] GUO H, CUI Q, WANG J, et al. Detecting and positioning of wind turbine blade tips for UAV-based automatic inspection [C]//Proceedings of the IGARSS 2019- 2019 IEEE International Geoscience and Remote Sensing Symposium, 2019: 1374-1377.
[10] 吕禾丰, 陆华才. 基于YOLOv5算法的交通标志识别技术研究[J]. 电子测量与仪器学报, 2021, 35(10): 137-144
[11] 沈跃杰, 行鸿彦, 王水璋. 基于粒子群优化的串级模糊PID无人机飞行控制系统[J]. 电子测量技术, 2022, 45(1): 96-103
[12] 郑学汉, 魏振忠, 张广军. 运动目标视觉跟踪测量系统与场地坐标系的快速统一方法[J]. 红外与激光工程, 2015, 44(7): 2175-2181
[13] 方春华, 张峰, 党卫军, 等. 基于Canny算法的10 kV电缆主绝缘表面缺陷识别方法[J]. 光电子·激光, 2021, 32(11): 1188-1195
[14] 佘毅, 李冲, 李昊霖. 应用霍夫变换的航带自动构建技术[J]. 测绘科学, 2020, 45(11): 101-106
[15] 李楠, 邓威, 王晨, 等. 基于K-means聚类与概率神经网络的模拟电路故障诊断方法[J]. 中国测试, 2021, 47(3): 98-103+109
