基于惯性测量的输电导线覆冰运动特征提取

Covered ice motion feature extraction for transmission lines based on inertial measurement

为识别输电线路的覆冰运动及其运动特性并进行有效阻尼控制,提出一种基于微惯性测量系统(Micro Inertial Measurement System,MIMS)的多维时间序列特征提取方法。设计一种MIMS测量节点,采集输电导线运动时三轴加速度和角速度的原始信号,对原始信号进行预处理并建立多维时变测量矩阵;对相关系数矩阵进行奇异值分解,计算加权特征矩阵作为导线运动幅值和频率的识别特征,采用近邻分类法对覆冰运动进行识别并通过相关分析估计覆冰导线幅值和频率变异程度;构建输电网络覆冰运动模拟实验系统,采集不同运动形式的MIMS信号按照该算法处理,并将三种运动形式特征进行比较。实验结果表明所提算法能够识别输电线路覆冰运动并进行幅频估计,可为输电线路的运动阻尼控制器的设计提供参数。

In order to identify covered ice motions and their features of transmission lines and perform effective damping control, a multi-dimensional time series feature extraction method based on the micro inertial measurement system(MIMS) was proposed. Firstly, a MIMS measurement node was designed to collect original signals of tri-axial accelerations and angular velocities of transmission lines during transmission lines moving. Original signals were pre-processed to establish a multi-dimensional time-varying measurement matrix. Then, the singular value decomposition was done for the correlation coefficient matrix to calculate the weighted feature matrix taken as recognition features of transmission lines’ motion amplitude and frequency. The covered ice motions were recognized using the near neighbor classification method, and variation levels for amplitude and frequency of iced conductors were estimated with correlation analysis. Finally, an analog test system was constructed for transmission lines’ covered ice motions, and MIMS signals of different motion forms were collected and processed according to the proposed algorithm. Features of covered ice’s 3 motion forms were compared. The test results showed that the proposed algorithm can be used to recognize transmission lines’ covered ice motions and estimate their amplitude and frequency, it can provide parameters for design of motion damping controller of transmission lines.