基于UIE框架的电网故障处置预案实体和事件识别方法

doi:10.11930/j.issn.1004-9649.202307076

中国电力 ›› 2023, Vol. 56 ›› Issue (12): 138-146.DOI: 10.11930/j.issn.1004-9649.202307076

• 电力系统弹性提升关键技术 • 上一篇下一篇

基于UIE框架的电网故障处置预案实体和事件识别方法

皮俊波¹(), 齐世雄¹(), 孙文多², 楼贤嗣², 沃建栋³, 张越⁴^,⁵(), 姜涛⁴^,⁵, 单连飞⁴^,⁵

1. 国家电网有限公司国家电力调度控制中心，北京　100031
2. 国网浙江电力调度控制中心，浙江杭州　310007
3. 国网浙江省电力有限公司金华供电公司，浙江金华　321016
4. 南瑞集团有限公司（国网电力科学研究院有限公司），江苏南京　211106
5. 北京科东电力控制系统有限责任公司，北京　100192

收稿日期:2023-07-19 接受日期:2023-11-07 出版日期:2023-12-28 发布日期:2023-12-28
作者简介:皮俊波（1984—），男，硕士，高级工程师，从事大电网调控运行控制及管理工作，E-mail： pi-junbo@sgcc.com.cn
齐世雄（1994—），男，硕士，工程师，从事电力系统优化调度、大电网运行控制研究，E-mail： qi-shixiong@sgcc.com.cn
张越（1989—），男，通信作者，硕士，工程师，从事人工智能技术在电力系统的应用研究，E-mail： zhangyue3655@163.com
基金资助:
国家电网有限公司科技项目（5108-202319067A-1-1-ZN）。

Entity and Event Recognition Method for Power Grid Fault Handling Plan Based on UIE Framework

Junbo PI¹(), Shixiong QI¹(), Wenduo SUN², Xiansi LOU², Jiandong WO³, Yue ZHANG⁴^,⁵(), Tao JIANG⁴^,⁵, Lianfei SHAN⁴^,⁵

1. National Electric Power Dispatching and Control Center of State Grid Corporation of China, Beijing 100031, China
2. State Grid Zhejiang Electric Power Dispatching and Control Center, Hangzhou 310007, China
3. Jinhua Power Supply Company of State Grid Zhejiang Electric Power Co., Ltd., Jinhua 321016, China
4. NARI Group Corporation Co., Ltd. (State Grid Electric Power Research Institute Co., Ltd.), Nanjing 211106, China
5. Beijing KeDong Electric Power Control System Co., Ltd., Beijing 100192, China

Received:2023-07-19 Accepted:2023-11-07 Online:2023-12-28 Published:2023-12-28
Supported by:
This work is supported by Science and Technology Project of SGCC (No.5108-202319067A-1-1-ZN).

摘要/Abstract

摘要：

电网故障处置预案区域差异化明显且实体嵌套复杂，仅凭实体类型识别难以准确地将其结构化。提出一种基于通用信息抽取（universal information extraction，UIE）框架的故障处置预案实体和事件识别方法。首先，提出基于句法分析的故障处置预案实体标记方法，生成故障处置预案标记实体；然后，接入ERNIE 3.0编码及双指针解码模块替代UIE框架中生成式模型，通过调节超参数训练故障处置预案实体与实体标签在高维空间的映射关系及实体嵌套关系；最后，基于句法结构组合嵌套复杂的预案实体得到预案事件。通过不同区域电网调控中心预案验证，与其他算法相比，所提方法对故障处置预案具有较高的实体和事件识别精度，能够准确识别预案中故障处置策略和恢复策略，为故障情况下区域电网弹性提升提供支撑。

关键词: 故障处置预案, 通用信息抽取框架, 实体识别, 事件抽取

Abstract:

The power grid fault handling plan is obvious in regional differentiation and the nested entities are complex, so it is difficult to accurately structure the fault handling plan only by entity type identification. An entity and event recognition method for power grid fault handling plan based on an universal information extraction (UIE) framework is proposed. Firstly, an entity marking method for fault handling plan based on syntactic analysis is proposed to generate fault handling plan marking entity. Then, the ERNIE 3.0 encoding and double-pointer decoding module are used to replace the generative model in the UIE framework, and the mapping relationship between entity and entity labels of fault handling plan and the nested entity relationship are trained through adjusting the hyper-parameters of the UIE framework. Finally, the complex nested plan entities are combined to obtain the plan event based on the syntactic structure. Through verification by the plans of different regional power grid dispatch and control centers, the proposed method has higher entity and event recognition accuracy for fault handling plans than other algorithms. It can accurately identify the fault handling strategies and recovery strategies in the plan, and provide support for the improvement of the regional power grid elasticity in the case of fault.

Key words: fault handling plan, universal information extraction framework, entity recognition, event extraction

皮俊波, 齐世雄, 孙文多, 楼贤嗣, 沃建栋, 张越, 姜涛, 单连飞. 基于UIE框架的电网故障处置预案实体和事件识别方法[J]. 中国电力, 2023, 56(12): 138-146.

Junbo PI, Shixiong QI, Wenduo SUN, Xiansi LOU, Jiandong WO, Yue ZHANG, Tao JIANG, Lianfei SHAN. Entity and Event Recognition Method for Power Grid Fault Handling Plan Based on UIE Framework[J]. Electric Power, 2023, 56(12): 138-146.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202307076

https://www.electricpower.com.cn/CN/Y2023/V56/I12/138

图/表 9

参考文献 26

1	陈国平, 李明节, 许涛. 特高压交直流电网系统保护及其关键技术[J]. 电力系统自动化, 2018, 42 (22): 2- 10.
	CHEN Guoping, LI Mingjie, XU Tao. System protection and its key technologies of UHV AC and DC power grid[J]. Automation of Electric Power Systems, 2018, 42 (22): 2- 10.
2	李明节, 陶洪铸, 许洪强, 等. 电网调控领域人工智能技术框架与应用展望[J]. 电网技术, 2020, 44 (2): 393- 400.
	LI Mingjie, TAO Hongzhu, XU Hongqiang, et al. The technical framework and application prospect of artificial intelligence application in the field of power grid dispatching and control[J]. Power System Technology, 2020, 44 (2): 393- 400.
3	许洪强, 姚建国, 南贵林, 等. 未来电网调度控制系统应用功能的新特征[J]. 电力系统自动化, 2018, 42 (1): 1- 7.
	XU Hongqiang, YAO Jianguo, NAN Guilin, et al. New features of application function for future dispatching and control systems[J]. Automation of Electric Power Systems, 2018, 42 (1): 1- 7.
4	张晓华, 冯长有, 王永明, 等. 电网调控机器人设计思路[J]. 电力系统自动化, 2019, 43 (13): 1- 8.
	ZHANG Xiaohua, FENG Changyou, WANG Yongming, et al. Design ideas of robotic dispatcher for power grid[J]. Automation of Electric Power Systems, 2019, 43 (13): 1- 8.
5	陶洪铸, 翟明玉, 许洪强, 等. 适应调控领域应用场景的人工智能平台体系架构及关键技术[J]. 电网技术, 2020, 44 (2): 412- 419.
	TAO Hongzhu, ZHAI Mingyu, XU Hongqiang, et al. Architecture and key technologies of artificial intelligence platform oriented for power grid dispatching and control application scenarios[J]. Power System Technology, 2020, 44 (2): 412- 419.
6	王福贺, 海威, 张越, 等. 电网线路故障处置智能调度机器人研究及应用[J]. 电气自动化, 2021, 43 (3): 1- 3, 23.
	WANG Fuhe, HAI Wei, ZHANG Yue, et al. Research and application of intelligent dispatching robots handling grid line faults[J]. Electrical Automation, 2021, 43 (3): 1- 3, 23.
7	余建明, 王小海, 张越, 等. 面向智能调控领域的知识图谱构建与应用[J]. 电力系统保护与控制, 2020, 48 (3): 29- 35.
	YU Jianming, WANG Xiaohai, ZHANG Yue, et al. Construction and application of knowledge graph for intelligent dispatching and control[J]. Power System Protection and Control, 2020, 48 (3): 29- 35.
8	胡怀伟, 富英, 张越, 等. 基于自然语言理解的故障处置预案语义建模研究及应用[J]. 电力信息与通信技术, 2022, 20 (5): 68- 73.
	HU Huaiwei, FU Ying, ZHANG Yue, et al. Research and application of semantic modeling of fault handling plan based on natural language understanding[J]. Electric Power Information and Communication Technology, 2022, 20 (5): 68- 73.
9	戴宇欣, 张俊, 季知祥, 等. 基于功能缺陷文本的电力系统二次设备智能诊断与辅助决策[J]. 电力自动化设备, 2014, 40 (8): 1537- 1562.
	DAI Yuxin, ZHANG Jun, JI Zhixiang, et al. Intelligent diagnosis and auxiliary decision of power system secondary equipment based on functional defect text[J]. Electric Power Automation Equipment, 2014, 40 (8): 1537- 1562.
10	. [J]. 2014, 40 (8): 1537- 1562.
	YANG Jinfeng, YU Qiubin, GUAN Yi, et al. An overview of research on electronic medical record oriented named entity recognition and entity relation extraction[J]. Acta Automatica Sinica, 2014, 40 (8): 1537- 1562.
11	鲁华永, 袁越, 郭泓佐, 等. 基于正则表达式的变电站集中监控信息解析方法[J]. 电力系统自动化, 2017, 41 (5): 78- 83.
	LU Huayong, YUAN Yue, GUO Hongzuo, et al. Regular expressions based information analytic method for substation centralized monitoring[J]. Automation of Electric Power Systems, 2017, 41 (5): 78- 83.
12	李慧林, 柴玉梅, 孙穆祯. 面向文本命名实体识别的深层网络模型[J]. 小型微型计算机系统, 2019, 40 (1): 50- 57.
	LI Huilin, CHAI Yumei, SUN Muzhen. Deep network model for text named entity recognition[J]. Journal of Chinese Computer Systems, 2019, 40 (1): 50- 57.
13	陈斌, 周勇, 刘兵. 基于卷积双向长短期记忆网络的事件触发词抽取[J]. 计算机工程, 2019, 45 (1): 153- 158.
	CHEN Bin, ZHOU Yong, LIU Bing. Event trigger word extraction based on convolutional bidirectional long short term memory network[J]. Computer Engineering, 2019, 45 (1): 153- 158.
14	吴文涛, 李培峰, 朱巧明. 基于混合神经网络的实体和事件联合抽取方法[J]. 中文信息学报, 2019, 33 (8): 77- 83.
	WU Wentao, LI Peifeng, ZHU Qiaoming. Joint extraction of entities and events by a hybrid neural network[J]. Journal of Chinese Information Processing, 2019, 33 (8): 77- 83.
15	江叶峰, 孙少华, 仇晨光, 等. 电网故障处置预案文本中的命名实体识别研究[J]. 电力工程技术, 2021, 40 (5): 177- 183.
	JIANG Yefeng, SUN Shaohua, QIU Chenguang, et al. Named entity recognition in power fault disposal preplan text[J]. Electric Power Engineering Technology, 2021, 40 (5): 177- 183.
16	单连飞, 张越. 电网调度专业语料库构建方法研究及应用[J]. 机械与电子, 2022, 40 (4): 73- 76, 80.
	SHAN Lianfei, ZHANG Yue. Research and application for construction method of power grid dispatching professional corpus[J]. Machinery & Electronics, 2022, 40 (4): 73- 76, 80.
17	丁禹, 尚学伟, 米为民. 基于深度学习的电网调控文本知识抽取方法[J]. 电力系统自动化, 2020, 44 (24): 161- 168.
	DING Yu, SHANG Xuewei, MI Weimin. Deep learning based knowledge extraction method for text of power grid dispatch and control[J]. Automation of Electric Power Systems, 2020, 44 (24): 161- 168.
18	佟佳弘, 武志刚, 管霖, 等. 电力调度文本的自然语言理解与解析技术及应用[J]. 电网技术, 2020, 44 (11): 4148- 4156.
	TONGJIA Hong, WU Zhigang, GUAN Lin, et al. Power dispatching text analysis and application based on natural language understanding[J]. Power System Technology, 2020, 44 (11): 4148- 4156.
19	蒋晨, 王渊, 胡俊华, 等. 基于深度学习的电力实体信息识别方法[J]. 电网技术, 2021, 45 (6): 2141- 2149.
	JIANG Chen, WANG Yuan, HU Junhua, et al. Power entity information recognition based on deep learning[J]. Power System Technology, 2021, 45 (6): 2141- 2149.
20	LU Y J, LIU Q, DAI D, et al. Unified structure generation for universal information extraction[EB/OL]. 2022: arXiv: 2203.12277.https://arxiv.org/abs/2203.12277.pdf.
21	SUN Y, WANG S H, FENG S K, et al. ERNIE 3.0: large-scale knowledge enhanced pre-training for language understanding and generation[EB/OL]. 2021: arXiv: 2107.02137.https://arxiv.org/abs/2107.02137.pdf.
22	BONETTA G, CANCELLIERE R, LIU D, et al. Retrieval-augmented transformer-XL for close-domain dialog generation[J]. The International FLAIRS Conference Proceedings, 2021, 34 (1): 2021. DOI
23	纪鹏志, 李光肖, 王琳, 等. 基于Transformer深度学习网络的主动配电网多元源荷灾损辨识方法[J]. 电力建设, 2023, 44(3): 56–65.
	JI Pengzhi, LI Guangxiao, WANG Lin, et al. Transformer-based evaluation method of power outage in active distribution networks with multiple sources and loads[J]. Electric Power Construction, 2023, 44(3): 56–65.
24	DAI Z H, YANG Z L, YANG Y M, et al. Transformer-XL: attentive language models beyond a fixed-length context[EB/OL]. 2019: arXiv: 1901.02860.https://arxiv.org/abs/1901.02860.pdf
25	LI X, FENG J, MENG Y, et al. A unified MRC framework for named entity recognition[C]//Meeting of the Association for Computational Linguistics. Association for Computational Linguistics, 2020. DOI:10.18653/V1/2020.ACL-MAIN.519.
26	YANG R Q, ZHANG J H, GAO X, et al. Simple and effective text matching with richer alignment features[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Florence, Italy. Stroudsburg, PA, USA: Association for Computational Linguistics, 2019: 4699–4709.

编辑推荐 0

Metrics

阅读次数

全文

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	18	0	0	54

来源	本网站	其他网站

次数	52	20
比例	72%	28%

摘要

最新录用	在线预览	正式出版

0	0	47

来源	本网站	其他网站

次数	16	32
比例	33%	67%

参数	设置大小	说明
per_eval_batch_size	32	批处理大小
learning_rate	1×10^–5	最大学习率
num_train_epochs	250	训练周期
attention_dropout	0.1	attention网络失活比例
hidden_act	"gelu"	隐藏层激活函数
hidden_dropout_prob	0.1	隐藏层dropout比例
hidden_size	768	隐藏层神经元个数
initializer_range	0.02	标准偏差
intermediate_size	3072	中间层神经元个数
max_position_emb	2048	序列最大长度
num_attention_heads	12	attention头数
num_hidden_layers	12	transformer层数
pool_act	"tanh"	池化层激活函数
vocab_size	40000	词典大小

参数	设置大小	说明
per_eval_batch_size	32	批处理大小
learning_rate	1×10^–5	最大学习率
num_train_epochs	250	训练周期
attention_dropout	0.1	attention网络失活比例
hidden_act	"gelu"	隐藏层激活函数
hidden_dropout_prob	0.1	隐藏层dropout比例
hidden_size	768	隐藏层神经元个数
initializer_range	0.02	标准偏差
intermediate_size	3072	中间层神经元个数
max_position_emb	2048	序列最大长度
num_attention_heads	12	attention头数
num_hidden_layers	12	transformer层数
pool_act	"tanh"	池化层激活函数
vocab_size	40000	词典大小

序号		事件生成模式
1		act-obj-com-endadv
2		sub-act-obj-com-endadv
3		preadv-sub-act-obj-com-endadv
4		act-preind-ind-ind-endind-com
5		preadv-act-preind-ind-ind-endind-com-endadv
6		sub-act-preind-ind-ind-endind-com-endadv

序号		事件生成模式
1		act-obj-com-endadv
2		sub-act-obj-com-endadv
3		preadv-sub-act-obj-com-endadv
4		act-preind-ind-ind-endind-com
5		preadv-act-preind-ind-ind-endind-com-endadv
6		sub-act-preind-ind-ind-endind-com-endadv

实体类别	精确率/%	召回率/%	F₁/%
sub	98.44	98.44	98.44
act	100.00	98.73	99.36
obj	98.25	98.82	98.53
com	100.00	100.00	100.00
preadv	100.00	100.00	100.00
endadv	94.12	100.00	96.97
preind	92.60	92.60	92.60
endind	97.78	100.00	98.88
ind	97.30	95.58	96.43

基于UIE框架的电网故障处置预案实体和事件识别方法

Entity and Event Recognition Method for Power Grid Fault Handling Plan Based on UIE Framework

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 26

相关文章 1

编辑推荐 0

Metrics

模型	精确率/%	召回率/%	F₁值/%
UIE框架	97.61	98.24	97.91
BERT+BiLSTM-CRF	94.69	95.22	94.93
word2vec+BiLSTM-CRF	93.34	92.78	92.98

实体识别		内容展示
输入模板		模板提示 = ['sub'，'act'，'obj'，'com'，'preadv'，'endadv'，'preind'，'endind'，'ind']
输入文本		"xx省调增加xx机组出力，控制xx双回线潮流不超过10万千瓦。"
实体识别结果		'sub':[{'text': 'xx省调', 'start': 0, 'end': 4}] 'act':[{'text': '增加', 'start':4, 'end': 6}, 　　 {'text': '控制', 'start': 13, 'end': 15,}] 'obj':[{'text': 'xx机组', 'start':6, 'end': 10,}, 　　 {'text': 'xx双回线', 'start':15, 'end':20}] 'com':[{'text': '出力', 'start':10, 'end':12 }, 　　 {'text': '潮流', 'start':20, 'end':22}] 'preadv':[] 'endadv':[{'text': '不超过10万千瓦', 'start':22, 'end':30}] 'preind':[] 'endind':[] 'ind':[]

模态框（Modal）标题

基于UIE框架的电网故障处置预案实体和事件识别方法

Entity and Event Recognition Method for Power Grid Fault Handling Plan Based on UIE Framework

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 26

相关文章 1

编辑推荐 0

Metrics