|
|
|
| Research on Construction Method and Application of Knowledge Graph for Power Transformer Operation and Maintenance Based on ALBERT |
| Xie Qing1,2, Cai Yang1, Xie Jun1, Wang Chunxin1, Zhang Yutong1, Xu Zhikang1 |
1. State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources North China Electric Power University Beijing 102206 China;
2. Hebei Provincial Key Laboratory of Power Transmission Equipment Security Defense North China Electric Power University Baoding 071003 China |
|
|
|
Abstract Knowledge graph can effectively manage massive information and deal with complex and diverse relationships. At the same time, knowledge search and decision-making based on knowledge graph are more in line with human logic and enhance the interpretability of answers. Since the current power transformer operations mainly depends on traditional experience, and has accumulated the massive operations within the power system documentation and the accident report, it is an urgent need to solve the problem on how to read the vast amounts of data quickly and accurately extract the valuable information, so as to use data-driven approach to transformer intelligent operations. Therefore, this paper proposes an ALBERT based knowledge graph construction method for power transformer operation and maintenance.
First of all, due to the small number of publicly available operation and maintenance documents and accident handling reports in the field of power transformer operation and maintenance, it is necessary to build a training dataset for the field of power transformer. The construction process of the dataset is divided into two parts. The first part is to generate samples of the obtained fault analysis report, anomaly detection report and other texts on basis of the method of regular matching rules. The second part is to clean the obtained literature texts and annotate the processed texts with BIO method. Then, the ALBERT-BiLSTM-CRF model was used to extract entities from the operation and maintenance texts of power transformers, and the ALBERT-BiLSTM-Attention model is used to extract relations from the operation and maintenance texts of power transformers. Finally, the Neo4j graph database is used to store and visualize the extracted triple. In the light of the constructed knowledge graph of power transformer operation and maintenance, the auxiliary decision-making function of power transformer operation and maintenance can be realized.
The ALBERT-BiLSTM-CRF model and ALBERT-BiLSTM-Attention model are used to conduct relation extraction experiments on the operation and maintenance texts of power transformers respectively. The accuracy of entity extraction using the ALBERT-BiLSTM-CRF model can reach 94.4%. The F1 score can reach 94.2%, which is 9.1% and 7.9% higher than BiLSTM-CRF respectively. The accuracy of ALBERT-BiLSTM-Attention model can reach 94.1%, and the F1 score can reach 95.1%, which are improved by 3.2% and 2.5% compared with the BiLSTM-Attention model. From the experimental results, it shows that the ALBERT pre-trained model has a good adaptability to extract entities and relations for power transformer operation and maintenance, and can better complete the task of knowledge extraction.
To sum up, this paper proposes a knowledge graph construction method of power transformer operation and maintenance based on ALBERT model. The main conclusions are as follows: ① A training dataset of power transformer operation and maintenance text is constructed based on Selenium framework and the sample generation method of regular matching, which effectively solves the difficult problem of obtaining data sets in the field of power transformer operation and maintenance. ② The ALBERT-BiLSTM-CRF model and ALBERT-BiLSTM-Attention model are used to extract entities and relations from the operation and maintenance texts of power transformers. Compared with the traditional deep learning model, ALBERT model can effectively overcome the problems in power transformer operation and maintenance texts, such as too many proper nouns, letters and symbols among Chinese characters. ③ Based on the constructed power transformer operation and maintenance knowledge graph, the operation and maintenance auxiliary decision-making function can be realized, which provides a new idea for the intelligent operation and maintenance of power transformers.
|
|
Received: 15 September 2022
|
|
|
|
|
|
[1] 李刚, 李银强, 王洪涛, 等. 电力设备健康管理知识图谱:基本概念、关键技术及研究进展[J]. 电力系统自动化, 2022, 46(3): 1-13.
Li Gang, Li Yinqiang, Wang Hongtao, et al.Knowledge graph of power equipment health management: basic concepts, key technologies and research progress[J]. Automation of Electric Power Systems, 2022, 46(3): 1-13.
[2] 谢庆, 杨天驰, 裴少通, 等. 基于多尺度协作模型的电气设备红外图像超分辨率故障辨识方法[J]. 电工技术学报, 2021, 36(21): 4608-4616.
Xie Qing, Yang Tianchi, Pei Shaotong, et al.Super-resolution identification method of electrical equipment fault based on multi-scale cooperation model[J]. Transactions of China Electrotechnical Society, 2021, 36(21): 4608-4616.
[3] 赵鹏, 蒲天骄, 王新迎, 等. 面向能源互联网数字孪生的电力物联网关键技术及展望[J]. 中国电机工程学报, 2022, 42(2): 447-457.
Zhao Peng, Pu Tianjiao, Wang Xinying, et al.Key technologies and perspectives of power Internet of Things facing with digital twins of the energy Internet[J]. Proceedings of the CSEE, 2022, 42(2): 447-457.
[4] 律方成, 牛雷雷, 王胜辉, 等. 基于优化YOLOv4的主要电气设备智能检测及调参策略[J]. 电工技术学报, 2021, 36(22): 4837-4848.
Lü Fangcheng, Niu Leilei, Wang Shenghui, et al.Intelligent detection and parameter adjustment strategy of major electrical equipment based on optimized YOLOv4[J]. Transactions of China Electrotechnical Society, 2021, 36(22): 4837-4848.
[5] 蒋逸雯, 李黎, 李智威, 等. 基于深度语义学习的电力变压器运维文本信息挖掘方法[J]. 中国电机工程学报, 2019, 39(14): 4162-4171.
Jiang Yiwen, Li Li, Li Zhiwei, et al.An information mining method of power transformer operation and maintenance texts based on deep semantic learning[J]. Proceedings of the CSEE, 2019, 39(14): 4162-4171.
[6] 刘梓权, 王慧芳. 基于知识图谱技术的电力设备缺陷记录检索方法[J]. 电力系统自动化, 2018, 42(14): 158-164.
Liu Ziquan, Wang Huifang.Retrieval method for defect records of power equipment based on knowledge graph technology[J]. Automation of Electric Power Systems, 2018, 42(14): 158-164.
[7] Chen Jiahui, Ge Xingtong, Li Weichao, et al.Construction of spatiotemporal knowledge graph for emergency decision making[C]//2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, Brussels, Belgium, 2021: 3920-3923.
[8] Sarazin A, Bascans J, Sciau J B, et al.Expert system dedicated to condition-based maintenance based on a knowledge graph approach: application to an aeronautic system[J]. Expert Systems With Applications, 2021, 186: 115767.
[9] Qin Hao, Jin Jiong.Intelligent maintenance of shield tunelling machine based on knowledge graph[C]// 2020 IEEE 18th International Conference on Industrial Informatics, Warwick, United Kingdom, 2020: 793-797.
[10] 边晓燕, 张璐瑶, 周波, 等. 基于知识图谱的国内外电力市场研究综述[J]. 电工技术学报, 2022, 37(11): 2777-2788.
Bian Xiaoyan, Zhang Luyao, Zhou Bo, et al.Review on domestic and international electricity market research based on knowledge graph[J]. Transactions of China Electrotechnical Society, 2022, 37(11): 2777-2788.
[11] 蒲天骄, 谈元鹏, 彭国政, 等. 电力领域知识图谱的构建与应用[J]. 电网技术, 2021, 45(6): 2080-2091.
Pu Tianjiao, Tan Yuanpeng, Peng Guozheng, et al.Construction and application of knowledge graph in the electric power field[J]. Power System Technology, 2021, 45(6): 2080-2091.
[12] 冯斌, 张又文, 唐昕, 等. 基于BiLSTM-Attention神经网络的电力设备缺陷文本挖掘[J]. 中国电机工程学报, 2020, 40(增刊1): 1-10.
Feng Bin, Zhang Youwen, Tang Xin, et al.Power equipment defect record text mining based on BiLSTM-attention neural network[J]. Proceedings of the CSEE, 2020, 40(S1): 1-10.
[13] 郭榕, 杨群, 刘绍翰, 等. 电网故障处置知识图谱构建研究与应用[J]. 电网技术, 2021, 45(6): 2092-2100.
Guo Rong, Yang Qun, Liu Shaohan, et al.Construction and application of power grid fault handing knowledge graph[J]. Power System Technology, 2021, 45(6): 2092-2100.
[14] Devlin J, Chang Mingwei, Lee K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[J/OL]. https://doi.org/10. 48550/arXiv.1810.04805, 2019
[15] Lan Z, Chen M, Goodman S, et al.ALBERT: a lite BERT for self-supervised learning of language representations[C]//Proceedings of the 8th International Conference on Learning Representations, La Jolla, CA, 2020: 1-17.
[16] 国家能源局. DL/T 573—2010电力变压器检修导则[S]. 北京: 中国电力出版社, 2010.
[17] Fan Tao, Wang Hao.Research of Chinese intangible cultural heritage knowledge graph construction and attribute value extraction with graph attention network[J]. Information Processing & Management, 2022, 59(1): 102753.
[18] zhang Fan, Wu Jizhou, Nie Yingli, et al. Research of knowledge graph technology and its applications in agricultural information consultation field[C]//2020 IEEE 39th International Performance Computing and Communications Conference, Austin, TX, USA, 2020: 1-4.
[19] Zeng Xiangxiang, Tu Xinqi, Liu Yuansheng, et al.Toward better drug discovery with knowledge graph[J]. Current Opinion in Structural Biology, 2022, 72: 114-126.
[20] Ma Xiaogang.Knowledge graph construction and application in geosciences: a review[J]. Computers & Geosciences, 2022, 161: 105082.
[21] 乔骥, 王新迎, 闵睿, 等. 面向电网调度故障处理的知识图谱框架与关键技术初探[J]. 中国电机工程学报, 2020, 40(18): 5837-5848.
Qiao Ji, Wang Xinying, Min Rui, et al.Framework and key technologies of knowledge-graph-based fault handling system in power grid[J]. Proceedings of the CSEE, 2020, 40(18): 5837-5848.
[22] Ma Lianbo, Wang Jingwei, Cheng Jian, et al.MLRP-KG: mine landslide risk prediction based on knowledge graph[J]. IEEE Transactions on Artificial Intelligence, 2022, 3(1): 78-87.
[23] 叶欣智, 尚磊, 董旭柱, 等. 面向配电网故障处置的知识图谱研究与应用[J]. 电网技术, 2022, 46(10): 3739-3749.
Ye Xinzhi, Shang Lei, Dong Xuzhu, et al.Knowledge graph for distribution network fault handling[J]. Power System Technology, 2022, 46(10): 3739-3749.
[24] 贺瑞芳, 段绍杨. 基于多任务学习的中文事件抽取联合模型[J]. 软件学报, 2019, 30(4): 1015-1030.
He Ruifang, Duan Shaoyang.Joint Chinese event extraction based multi-task learning[J]. Journal of Software, 2019, 30(4): 1015-1030.
[25] 杨穗珠, 刘艳霞, 张凯文, 等. 远程监督关系抽取综述[J]. 计算机学报, 2021, 44(8): 1636-1660.
Yang Suizhu, Liu Yanxia, Zhang Kaiwen, et al.Survey on distantly-supervised relation extraction[J]. Chinese Journal of Computers, 2021, 44(8): 1636-1660.
[26] 蔡莉, 王淑婷, 刘俊晖, 等. 数据标注研究综述[J]. 软件学报, 2020, 31(2): 302-320.
Cai Li, Wang Shuting, Liu Junhui, et al.Survey of data annotation[J]. Journal of Software, 2020, 31(2): 302-320.
[27] Li Chun xin, Lu Qian nan, Huang Ming, et al. Construction of radio transmission equipment type approval knowledge graph and its application[C]// 2021 IEEE 2nd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering, Nanchang, China, 2021: 956-961. |
|
|
|
2023, Vol. 38 
