| 引用本文: | 戴志辉,张富泽,韩 笑.基于同源录波数据比对的继电保护采样回路异常检测方法[J].电力系统保护与控制,2025,53(1):147-159.[点击复制] |
| DAI Zhihui,ZHANG Fuze,HAN Xiao.Anomaly detection method for a relay protection sampling circuit based on homologous waveform data comparison[J].Power System Protection and Control,2025,53(1):147-159[点击复制] |
|
| 摘要: |
| 处于改建阶段的智能变电站采样模式复杂，继电保护装置难以发现采样回路轻微异常，导致回路隐患暴露时间严重滞后。针对上述问题，分析改建时期智能变电站的采样模式和二次设备配置情况，提出基于同源录波数据比对的继电保护采样回路异常检测方法。首先，利用双向编码器表征(bidirectional encoder representations from transformers, BERT)语言模型与余弦相似度算法，实现同源录波数据的通道匹配。然后，利用重采样技术和曼哈顿距离完成波形的采样频率统一与时域对齐。最后，基于动态时间规整(dynamic time warping, DTW)算法提出改进算法，并结合采样点偏移量共同设置采样回路的异常判据。算例分析表明，该方法可以完成录波数据的同源通道匹配，实现波形的一致性对齐，并且相比于传统DTW算法，改进DTW算法对异常状态识别的灵敏性和准确性更高。根据异常判据能够有效检测继电保护采样回路的异常状态，确保了智能变电站的安全可靠运行。 |
| 关键词: 继电保护装置 采样回路 异常检测 改进DTW算法 录波数据 |
| DOI：10.19783/j.cnki.pspc.240542 |
| 投稿时间：2024-05-06修订日期：2024-09-30 |
| 基金项目:国家自然科学基金项目资助(51877084) |
|
| Anomaly detection method for a relay protection sampling circuit based on homologous waveform data comparison |
| DAI Zhihui,ZHANG Fuze,HAN Xiao |
| (Hebei Key Laboratory of Distributed Energy Storage and Microgrid (North China Electric
Power University), Baoding 071003, China) |
| Abstract: |
| The sampling mode of an intelligent substation in a reconstruction phase is complex, and it is difficult for the relay protection device to detect slight abnormalities in the sampling circuit, resulting in a serious lag in the exposure time of hidden dangers in the circuit. To address these issues, the sampling mode and secondary equipment configuration of an intelligent substation during the reconstruction period are analyzed, and a relay protection sampling circuit anomaly detection method based on homologous waveform data comparison is proposed. First, a language model and cosine similarity algorithm of bidirectional encoder representations from transformers (BERT) are used for channel matching of homologous waveform data. Then a resampling technique and Manhattan distance are used to achieve waveform sampling frequency unification and time domain alignment. Lastly, an improved algorithm based on dynamic time warping (DTW) is proposed, and an anomaly criterion of the sampling circuit is set in combination with the offset of the sampling point. Case studies demonstrate that this method can successfully match waveform data homologous channels, achieve waveform consistency alignment. Compared to the traditional DTW algorithm, the improved DTW algorithm exhibits higher sensitivity and accuracy in identifying abnormal states. The abnormal state of the relay protection sampling circuit can be reliably detected from the abnormal criterion, ensuring the safe and reliable operation of the intelligent substation. |
| Key words: relay protection device sampling circuit anomaly detection improved DTW algorithm recording data |
