| 引用本文: | 戴志辉,张富泽,杨 鑫,韩 笑.基于FDTW算法的故障录波数据智能比对方法[J].电力系统保护与控制,2023,51(23):82-91.[点击复制] |
| DAI Zhihui,ZHANG Fuze,YANG Xin,HAN Xiao.Intelligent comparison method for fault record waveform data based on the FDTW algorithm[J].Power System Protection and Control,2023,51(23):82-91[点击复制] |
|
| 摘要: |
| 针对智能变电站保护装置故障录波文件无通道标识信息、难以与同源通道故障录波器录波文件相互对应的问题，提出一种基于快速动态时间规整(fast dynamic time warping, FDTW)算法的故障录波数据智能比对方法。首先，对各厂站提取的故障录波数据进行异常检测，确保故障录波文件中数据的质量。其次，利用拉格朗日插值法进行采样频率转换，解决保护装置与故障录波器采样频率不一致的问题。然后，利用欧氏距离对同源录波文件数据进行波形一致性对齐，以实现时钟同步。最后，使用FDTW算法对无通道标识信息的故障录波数据进行智能比对。算例分析表明，该方法能够对故障录波数据进行一致性处理，可以快速准确地计算待匹配录波波形的相似度，实现同源通道的录波数据匹配，具备较强的稳定性和实时性。 |
| 关键词: 故障录波 FDTW 同源比对 欧氏距离 智能变电站 |
| DOI：10.19783/j.cnki.pspc.230497 |
| 投稿时间：2023-04-28修订日期：2023-06-25 |
| 基金项目:国家自然科学基金项目资助(51877084) |
|
| Intelligent comparison method for fault record waveform data based on the FDTW algorithm |
| DAI Zhihui1,ZHANG Fuze1,YANG Xin2,HAN Xiao1 |
| (1. Department of Electrical Engineering, North China Electric Power University, Baoding 071003, China;
2. China Yangtze Power Co., Ltd., Yichang 443002, China) |
| Abstract: |
| An intelligent comparison method for fault waveform data is proposed based on the fast dynamic time warping (FDTW) algorithm to address the problem of the lack of channel identification information in fault recorder wave files of intelligent substation protection devices. This lack makes it difficult to correspond with the fault recorder wave files of the same source channel. First, anomaly detection is performed on the fault waveform data extracted from each substation to ensure the quality of the data in the fault waveform files. Second, the Lagrange interpolation method is used for sampling frequency conversion to deal with inconsistent sampling frequencies between the protection device and the fault waveform recorder. Then, the Euclidean distance is used to align the waveform consistency of the same source waveform files to achieve clock synchronization. Finally, the FDTW algorithm is used to intelligently compare the fault waveform data without channel identification information. Case studies show that the proposed method can process the fault waveform data into a consistent type, quickly and accurately calculate the similarity of waveforms to be matched, achieve the matching of the fault waveform data from the same source channel, and has strong stability and real-time performance. |
| Key words: fault record FDTW homology comparison Euclidean distance (ED) intelligent substation |
