基于视频监测的输电线路舞动信息提取与频谱分析  

Galloping Information Extraction and Spectral Analysis for Transmission Lines Based on Video Monitoring

作　　者：尹晖[1,2] 张晓鸣 李小祥[1,3] 管保安 吕中宾 谢凯[5] 

机构地区：[1]武汉大学测绘学院,武汉430079 [2]地球空间信息技术协同创新中心,武汉430079 [3]武汉吉威空间信息技术研究院有限公司,武汉430040 [4]武汉大学电气工程学院,武汉430072 [5]国网河南省电力公司电力科学研究院,郑州450002

出　　处：《高电压技术》2017年第9期2889-2895,共7页High Voltage Engineering

基　　金：国家自然科学基金(51077105);国家电网公司科技项目(SGSX0000YJJS(2014)457);国家电网公司科技攻关团队项目(SG11013)~~

摘　　要：输电线路舞动在线监测和特征提取是舞动灾害预警的关键。针对采用分裂导线的(超/特)高压输电线路,提出了一种舞动信息提取与舞动分析的新方法。从视频中捕获每帧图像,选择分裂导线间隔棒为特征物进行边缘检测;通过自动搜索定位各帧图像中间隔棒轮廓位置,计算得到间隔棒中心坐标,以及间隔棒X、Y方向瞬时位移与偏移(距离)的时间序列;最后利用频谱分析得到导线舞动的主频和幅值,探测到主频为0.403 Hz的明显周期信号,舞动幅值相对误差<5%。结果表明,该方法能够有效分析导线舞动的特征和规律,可为输电线舞动灾害预警提供参考依据。Online monitoring and galloping features extraction of transmission lines are keys to galloping disaster prediction.We propose a new solution to galloping extraction and analysis by focusing on the EHV/UHV transmission lines with bundle-conductor. Each frame is captured from video, and the bundle-conductor spacer is selected as a feature object for edge detection. Then spacer outline is positioned by automatic searching, and coordinates of spacer center are calculated, thus time series of instantaneous displacement in X and Y direction and offset（distance） of spacer are generated, respectively. Finally, dominant frequency and amplitude of galloping are obtained though spectral analysis, in which the principal periodic signal of 0.403 Hz is obviously detected and relative error of galloping amplitudes is within 5%. The result shows that the transmission lines galloping features can be analyzed efficiently with this method, which provides reference foundations for galloping disaster prediction.

关 键 词：舞动 视频监测 间隔棒 信息提取 频谱分析 

分 类 号：TM752[电气工程—电力系统及自动化]