电力变压器故障诊断多智能体大模型构建与应用

林金山; 李元; 方荠萱; 刘志濠; 李春鹏; 张冠军

doi:10.7652/xjtuxb202510002

您当前的位置：

首页 >

文章列表页 >

电力变压器故障诊断多智能体大模型构建与应用

更新时间：2025-09-28

- 电力变压器故障诊断多智能体大模型构建与应用
  增强出版
- Construction and Application of Multi-Agent Large Language Model for Power Transformer Fault Diagnosis
- 西安交通大学学报 2025年59卷第10期页码：22-31
- 作者机构：
- 作者简介：
  
  作者简介：林金山（1998—），男，博士生；
  李元（通信作者），男，副教授，博士生导师。
- 基金信息：
  
  国家自然科学基金资助项目。(52477159)
- DOI：10.7652/xjtuxb202510002
  中图分类号： TM83
- 收稿：2025-03-06，
  
  纸质出版：2025-10-10
- 稿件说明：
移动端阅览
林金山, 李元, 方荠萱, 等. 电力变压器故障诊断多智能体大模型构建与应用[J]. 西安交通大学学报, 2025,59(10):22-31.

LIN Jinshan, LI Yuan, FANG Qixuan, et al. Construction and Application of Multi-Agent Large Language Model for Power Transformer Fault Diagnosis[J]. Journal of Xi'an Jiaotong University, 2025, 59(10): 22-31.
林金山, 李元, 方荠萱, 等. 电力变压器故障诊断多智能体大模型构建与应用[J]. 西安交通大学学报, 2025,59(10):22-31. DOI： 10.7652/xjtuxb202510002.

LIN Jinshan, LI Yuan, FANG Qixuan, et al. Construction and Application of Multi-Agent Large Language Model for Power Transformer Fault Diagnosis[J]. Journal of Xi'an Jiaotong University, 2025, 59(10): 22-31. DOI： 10.7652/xjtuxb202510002.

摘要

针对电力变压器状态参量多元异构、数据缺失导致的传统小模型适用性受限、诊断效果不佳等问题，提出了一种电力变压器故障诊断多智能体大模型框架。该框架构建了由故障案例知识图谱和3个智能体协同的推理体系。故障案例知识图谱通过约束机制增强大模型对电力专业知识的理解能力，并有效抑制机器幻觉现象。3个智能体模拟专家诊断过程，将复杂的变压器故障诊断任务分解执行：初级诊断智能体通过特征阈值分析实现故障初步识别；专家诊断智能体利用概率图模型对典型故障模式进行不确定性推理；案例分析智能体接入历史故障案例库，完成知识回溯与诊断结果验证。验证结果表明：所提模型对100例故障案例诊断表现优异，诊断准确率可达86%，相比BERT模型提高了33%；故障案例知识图谱的引入提升了大模型输出信息的完整性、语义一致性和专业深度，专家评分平均增幅为50%。相较于单一的大模型，该智能体大模型在降低误判率方面表现突出，可为电力变压器的故障智能诊断与运维提供技术支撑。

Abstract

To address the limited applicability of traditional small language models and the unsatisfactory diagnostic effects caused by the multi-dimensional heterogeneity of state parameters and data loss in power transformers

a multi-agent large language model framework for transformer fault diagnosis is proposed. The framework establishes a collaborative reasoning system with fault cases knowledge graph and three agents. The fault cases knowledge graph constraint mechanism significantly enhances the large language model's understanding of power engineering expertise while effectively mitigating machine hallucination. Three agents simulate expert diagnostic processes by decomposing complex transformer fault diagnosis tasks. Primary diagnostic agents perform preliminary fault identification through feature threshold analysis

expert diagnostic agents conduct uncertainty reasoning on typical fault patterns using probabilistic graphical models

and case analysis agents access a historical fault case database to enable knowledge retrieval and diagnostic result validation. Validation results show that the proposed model achieves excellent performance in diagnosing 100 fault cases

with an accuracy rate of 86%

representing a 33% improvement over the BERT model. The integration of the fault cases knowledge graph enhances the large language model's output in terms of completeness

semantic consistency

and professional depth

with an expert rating average performance increase of 50%.This multi-agent large language model demonstrates superior performance in reducing misjudgment rates compared to monolithic large language models

and can provide technical support for intelligent fault diagnosis and operation and maintenance of power transformers.

关键词

Keywords

references

《新型电力系统发展蓝皮书》编写组 . 新型电力系统发展蓝皮书 [EB/OL ] .[ 2025-02-25 ] . https://www.nea.gov.cn/download/xxdlxtfzlpsgk.pdf .

国家能源局 , 中国电力企业联合会 . 2022年全国电力可靠性年度报告 [EB/OL ] .( 2023-10-18 )[ 2025-02-20 ] . https://prpq.nea.gov.cn/ndbg/9875.html .

刘云鹏 , 许自强 , 李刚 , 等 . 人工智能驱动的数据分析技术在电力变压器状态检修中的应用综述 [J ] . 高电压技术 , 2019 , 45 ( 2 ): 337 - 348 .

LIU Yunpeng , XU Ziqiang , LI Gang , et al . Review on applications of artificial intelligence driven data analysis technology in condition based maintenance of power transformers [J ] . High Voltage Engineering , 2019 , 45 ( 2 ): 337 - 348 .

徐尧宇 , 李元 , 王怡静 , 等 . 信息缺失下变压器状态预警方法 [J ] . 高电压技术 , 2020 , 46 ( 9 ): 3062 - 3069 .

XU Yaoyu , LI Yuan , WANG Yijing , et al . Earlywarning methods of transformer conditions under information deficiency [J ] . High Voltage Engineering , 2020 , 46 ( 9 ): 3062 - 3069 .

张国治 , 陈康 , 方荣行 , 等 . 基于DGA与鲸鱼算法优化LogitBoost-决策树的变压器故障诊断方法 [J ] . 电力系统保护与控制 , 2023 , 51 ( 7 ): 63 - 72 .

ZHANG Guozhi , CHEN Kang , FANG Rongxing , et al . Transformer fault diagnosis based on DGA and a whale algorithm optimizing a LogitBoost-decision tree [J ] . Power System Protection and Control , 2023 , 51 ( 7 ): 63 - 72 .

石宜金 , 谭贵生 , 赵波 , 等 . 基于模糊综合评估模型与信息融合的电力变压器状态评估方法 [J ] . 电力系统保护与控制 , 2022 , 50 ( 21 ): 167 - 176 .

SHI Yijin , TAN Guisheng , ZHAO Bo , et al . Condition assessment method for power transformers based on fuzzy comprehensive evaluation and information fusion [J ] . Power System Protection and Control , 2022 , 50 ( 21 ): 167 - 176 .

曲岳晗 , 赵洪山 , 马利波 , 等 . 多深度神经网络综合的电力变压器故障识别方法 [J ] . 中国电机工程学报 , 2021 , 41 ( 23 ): 8223 - 8230 .

QU Yuehan , ZHAO Hongshan , MA Libo , et al . Multi-depth neural network synthesis method for power transformer fault identification [J ] . Proceedings of the CSEE , 2021 , 41 ( 23 ): 8223 - 8230 .

万吉林 , 吴国强 , 管敏渊 , 等 . 基于RetinaNet的变压器图像小部件智能识别方法 [J ] . 电力系统保护与控制 , 2021 , 49 ( 12 ): 166 - 173 .

WAN Jilin , WU Guoqiang , GUAN Minyuan , et al . Intelligent recognition method for transformer small components based on RetinaNet [J ] . Power System Protection and Control , 2021 , 49 ( 12 ): 166 - 173 .

Stanford University . The 2023 AI index report [EB/OL ] .[ 2025-02-23 ] . https://hai.stanford.edu/ai-index/2023-ai-index-report .

朝乐门 . 数据科学的科学性与科学问题的分析 [J ] . 计算机科学 , 2024 , 51 ( 1 ): 26 - 34 .

ZHAO Lemen . Exploring the scientific nature and scientific questions of data science [J ] . Computer Science , 2024 , 51 ( 1 ): 26 - 34 .

李刚 , 方鸿 , 刘云鹏 , 等 . 新型电力系统中的大模型驱动技术:现状、机遇与挑战 [J ] . 高电压技术 , 2024 , 50 ( 7 ): 2864 - 2878 .

LI Gang , FANG Hong , LIU Yunpeng , et al . Largemodel drive technology in new power system:status, challenges and prospects [J ] . High Voltage Engineer ing , 2024 , 50 ( 7 ): 2864 - 2878 .

ACHIAM J , ADLER S , AGARWAL S , et al . GPT-4 technical report [EB/OL ] .( 2024-03-04 )[ 2025-02-22 ] . https://arxiv.org/abs/2303.08774 .

ANIL R , DAI A M , FIRAT O , et al . PaLM 2 technical report [EB/OL ] .( 2023-09-13 )[ 2025-02-19 ] . https://arxiv.org/abs/2305.10403 .

DAI Damai , DENG Chengqi , ZHAO Chenggang , et al . DeepSeekMoE:towards ultimate expert specialization in mixture-of-experts language models [C ] // Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics . Stroudsburg,PA, USA : ACL , 2024 : 1280 - 1297 .

YAN Ziming , XU Yan . Real-time optimal power flow with linguistic stipulations:integrating GPT-agent and deep reinforcement learning [J ] . IEEE Transactions on Power Systems , 2024 , 39 ( 2 ): 4747 - 4750 .

于硕 , 王司宇 , 王超 , 等 . 类ChatGPT大语言模型在电力系统中的应用前景 [J ] . 电气时代 , 2023 ( 10 ): 50 - 53 .

YU Shuo , WANG Siyu , WANG Chao , et al . Application prospect of ChatGPT-like large language model in power system [J ] . Electric Age , 2023 ( 10 ): 50 - 53 .

赵俊华 , 文福拴 , 黄建伟 , 等 . 基于大语言模型的电力系统通用人工智能展望:理论与应用 [J ] . 电力系统自动化 , 2024 , 48 ( 6 ): 13 - 28 .

ZHAO Junhua , WEN Fushuan , HUANG Jianwei , et al . Prospect of artificial general intelligence for power systems based on large language model:theory and applications [J ] . Automation of Electric Power Systems , 2024 , 48 ( 6 ): 13 - 28 .

黄文栋 , 贾鹏辉 , 朱云佳 . 南方电网首个配网AI大模型实现实体化应用 [J ] . 电力安全技术 , 2024 , 26 ( 5 ): 36 .

HUANG Wendong , JIA Penghui , ZHU Yunjia . The first large-scale AI model of China Southern Power Grid realizes practical application [J ] . Electric Safety Technology , 2024 , 26 ( 5 ): 36 .

李鹏 , 余涛 , 李立浧 , 等 . 电力人工智能的演变与展望:从专业智能走向通用智能 [J ] . 电力系统自动化 , 2024 , 48 ( 16 ): 1 - 17 .

LI Peng , YU Tao , LI Licheng , et al . Retrospect and prospect of artificial intelligence for electric power system:from domain intelligence to general intelligence [J ] . Automation of Electric Power Systems , 2024 , 48 ( 16 ): 1 - 17 .

YU Yangyang , LI Haohang , CHEN Zhi , et al . Fin-Mem: a performance-enhanced LLM trading agent with layered memory and character design [C ] // Proceedings of the 2024 AAAI Spring Symposium Series . Palo Alto,CA,USA : AAAI Press , 2024 : 595 - 597 .

MAO Jiageng , QIAN Yuxi , YE Junjie , et al . GPT-driver:learning to drive withGPT [EB/OL ] .( 2023-12-05 )[ 2025-02-23 ] . https://arxiv.org/abs/2310.01415 .

ZHANG Siyao , FU Daocheng , LIANG Wenzhe , et al . TrafficGPT:viewing,processing and interacting with traffic foundation models [J ] . Transport Policy , 2024 , 150 : 95 - 105 .

余涛 , 王艺澎 , 罗庆全 , 等 . 基于大语言模型的新型电力系统用户特性立体应用模式初探 [J ] . 高电压技术 , 2024 , 50 ( 7 ): 2833 - 2848 .

YU Tao , WANG Yipeng , LUO Qingquan , et al . Preliminary exploration for a large language model-based usercharacteristics stereoscopic application paradigm in new power system [J ] . High Voltage Engineering , 2024 , 50 ( 7 ): 2833 - 2848 .

吴海洋 , 吴子辰 , 吴博科 , 等 . 基于预训练语言模型的电网检修计划异常检测 [J ] . 情报工程 , 2024 , 10 ( 4 ): 14 - 24 .

WU Haiyang , WU Zichen , WU Boke , et al . A pretrained language model-based anomaly detection method for grid maintenance plans [J ] . Technology Intelligence Engineering , 2024 , 10 ( 4 ): 14 - 24 .

谢腾 , 杨俊安 , 刘辉 . 基于BERT-BiLSTM-CRF模型的中文实体识别 [J ] . 计算机系统应用 , 2020 , 29 ( 7 ): 48 - 55 .

XIE Teng , YANG Jun'an , LIU Hui . Chinese entity recognition based on BERT-BiLSTM-CRF model [J ] . Computer Systems & Applications , 2020 , 29 ( 7 ): 48 - 55 .

JEONG S , BAEK J , CHO S , et al . Adaptive-RAG:learning to adapt retrieval-augmented large language models through question complexity [C ] // Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies . Stroudsburg,PA,USA : ACL , 2024 : 7036 - 7050 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

语义区分和子图整合驱动的知识图谱表示学习方法

大型电力变压器非电量保护研究进展与展望

采用双小波字典的齿轮箱复合故障特征提取方法

一维卷积神经网络在机械故障特征提取中的可解释性研究

蜉蝣优化双通道网络在齿轮箱故障诊断中的应用