思维链增强的机电装备运维方案智能生成方法

doi:10.3785/j.issn.1008-973X.2026.07.014

浙江大学学报(工学版)

2026, Vol. 60

Issue (7): 1515-1527 DOI: 10.3785/j.issn.1008-973X.2026.07.014

机械工程

思维链增强的机电装备运维方案智能生成方法

高一聪1(

),吴栋1,密尚华2,*(

),郑浩2,谭建荣1

1. 浙江大学流体动力基础件与机电系统全国重点实验室，浙江杭州 310027
2. 北京航空航天大学杭州创新研究院，浙江杭州 310056

Chain-of-Thought enhanced intelligent generation method of electromechanical equipment operation and maintenance schemes

Yicong GAO1(

),Dong WU1,Shanghua MI2,*(

),Hao ZHENG2,Jianrong TAN1

1. State Key Laboratory of Fluid Power and Mechatronic Systems, Zhejiang University, Hangzhou 310027, China
2. Hangzhou Innovation Institute of Beihang University, Hangzhou 310056, China

全文: PDF(3817 KB) HTML

摘要：

针对机电装备结构复杂、故障耦合度高导致基于人工经验的运维方案方法效率低、故障追溯性差的问题，提出思维链增强的机电装备运维方案智能生成方法. 利用大语言模型的多源知识融合和知识推理的能力，设计机电装备多源异构运维领域知识的预处理流程，建立思维链增强的机电装备运维领域知识本体模型，通过思维链增强的故障知识注入和大模型微调，构建具备因果链式推理能力的思维链增强领域模型，实现机电装备的“故障现象-原因排序-方案生成”故障溯源推理. 引入图检索增强生成技术，构建具有社区划分的零部件知识图谱，深度融合推理多部件维修知识，提高运维方案的生成质量，实现故障溯源到运维方案生成的智能运维闭环. 对思维链增强领域模型进行性能评估和应用验证，结果表明，所提方法在故障溯源、运维方案生成的任务中表现出优异性能，显著提升了机电装备故障溯源的准确性和运维方案的合理性.

关键词： 思维链增强; 知识图谱; 大语言模型; 机电装备运维; 运维方案

Abstract:

An intelligent generation method of electromechanical equipment operation and maintenance schemes with enhanced Chain-of-Thought was proposed, aiming at the problems of low efficiency and poor traceability of operation and maintenance schemes based on manual experience caused by the complex structure of electromechanical equipment and the high coupling degree of faults. Utilizing the capabilities of multi-source knowledge fusion and knowledge reasoning of large language models, the preprocessing process of multi-source heterogeneous operation and maintenance domain knowledge of electromechanical equipment was designed, and the knowledge ontology model of the operation and maintenance domain of electromechanical equipment with enhanced Chain-of-Thought was established. Through the injection of fault knowledge with enhanced Chain-of-Thought and the fine-tuning of large models, the Chain-of-Thought enhanced domain model with causal chain reasoning ability was constructed. The fault traceability reasoning of “fault phenomenon - cause ranking - scheme generation” for electromechanical equipment has been realized. The graph retrieval-augmented generation technology was introduced to construct a components knowledge graph with community division. The multi-component maintenance knowledge was deeply integrated and reasoned, which improved the generation quality of operation and maintenance schemes and achieved an intelligent operation and maintenance closed loop from fault tracking to operation and maintenance scheme generation. Finally, the performance evaluation and application verification of the Chain-of-Thought enhanced domain model were carried out. The results show that the proposed method demonstrates excellent performance in tasks such as fault tracking and operation and maintenance scheme generation, significantly improving the accuracy of fault tracking and the rationality of operation and maintenance schemes.

Key words: Chain-of-Thought enhancement knowledge graph large language model electromechanical equipment operation and maintenance operation and maintenance schemes

收稿日期: 2025-05-20 出版日期: 2026-05-23

CLC:

TP 391.1

基金资助: 国家自然科学基金资助项目(52375272，52575263）；浙江省高等教育学会高等教育研究课题（KT2025456）；杭州市农业与社会发展项目（20241203A21）.

通讯作者: 密尚华 E-mail: gaoyicong@zju.edu.cn;393185319@qq.com

作者简介: 高一聪（1982—），男，副教授，杭州市钱江特聘专家，博士，从事产品正向设计理论与方法、智能结构创新设计研究. orcid.org/0000-0002-1987-0431. E-mail：gaoyicong@zju.edu.cn

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	作者相关文章
	高一聪
	吴栋
	密尚华
	郑浩
	谭建荣

引用本文:

高一聪,吴栋,密尚华,郑浩,谭建荣. 思维链增强的机电装备运维方案智能生成方法[J]. 浙江大学学报(工学版), 2026, 60(7): 1515-1527.

Yicong GAO,Dong WU,Shanghua MI,Hao ZHENG,Jianrong TAN. Chain-of-Thought enhanced intelligent generation method of electromechanical equipment operation and maintenance schemes. Journal of ZheJiang University (Engineering Science), 2026, 60(7): 1515-1527.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2026.07.014 或 https://www.zjujournals.com/eng/CN/Y2026/V60/I7/1515

图 1 思维链增强的机电装备运维方案智能生成框架

图 2 多源异构数据预处理

表 1 运维文本数据去口语化

表 2 智能立体车库机电系统实体和关系

图 3 智能立体车库机电系统知识本体模型

图 4 思维链增强的机电装备故障溯源模型

表 3 故障知识提取提示词模板

图 5 Neo4j知识存储

图 6 CoT增强的故障溯源流程

图 7 故障知识图谱查询及故障匹配

图 8 社区报告要求

图 9 思维链增强的机电装备运维方案智能生成流程图

表 4 微调实验设置

表 5 CoT增强领域模型与基线模型的性能对比

模型	回答
模型	智能车库升降机对中运行超时是什么原因，该怎么排查？
GPT-4	1.导轨或链条卡滞，解决方案：检查并清理导轨、链条、钢丝绳，必要时更换损坏部件. 2.传动机构松动或磨损，解决方案：检查电机、减速机、链轮等，拧紧松动部件或更换磨损部件. 3.升降平台平衡性问题，解决方案：调整平衡块，重新校准负载分布. 4.位置传感器故障或偏移，解决方案：检查并清洁传感器，重新标定或更换损坏的传感器. 5.PLC程序异常，解决方案：检查控制逻辑，调整时间参数，必要时更新或重写程序. 6.变频器或伺服控制异常，解决方案：检查变频器故障代码，调整参数，检测伺服电机编码器反馈. 7.电缆连接松动或损坏，解决方案：检查所有接线，修复松动或断裂的电缆. 8.负载过大或分布不均，解决方案：检查是否超载，重新分配载荷，避免偏载运行.
Llama3-8B	智能车库升降机对中运行超时，可能的原因有多种，包括但不限于以下几点： 1.电源问题：电源不稳定或电压不足可能导致升降机运行缓慢，甚至超时. 排查时可以检查电源线路、电压稳定性和电源设备是否正常工作. 2.控制系统故障：升降机的控制系统负责发送和接收指令，监控运行状态. 如果控制系统出现故障，可能导致指令无法正确执行或状态无法及时反馈，从而引起超时. 此时，需要检查控制系统的硬件和软件是否正常，包括传感器、控制器和执行器等部件. 3.机械部件卡滞：升降机的导轨、链条、滑轮等机械部件如果出现卡滞或磨损，也可能影响运行速度，导致超时. 定期对这些部件进行清洗、润滑和维护，以及检查其磨损程度，是预防此类问题的重要措施. 4.传感器故障：传感器负责检测升降机的位置和状态，如果传感器出现故障或误判，可能导致控制系统做出错误的决策，从而引起超时. 排查时需要检查传感器的灵敏度、准确性和连接线路是否正常.
领域模型（CoT enhanced）	根据运维知识图谱中的常见故障原因及其出现频次，以下是可能原因的排序和排查过程： 1. 最有可能的故障原因：车辆停放较斜，导致对中阻力增大，影响对中完成. 排查过程：检查车辆是否停放歪斜，导致升降机在对中过程中受阻；观察存车位置，确保车辆完全驶入存车位并居中停放；测试不同车辆在相同位置存取，排除个别车辆形状影响. 2. 中等可能原因：对中光电传感器检测异常，导致超时报警排查过程：检查光电传感器是否有灰尘、异物遮挡或对准偏差；检测光电开关信号，确认其是否能正确感应到对中完成状态；观察光电传感器安装角度，调整为最佳检测位置. 3. 最小可能原因：对中整流器损坏，导致信号无法正确传输排查过程：检查对中整流器是否有明显的损坏或老化迹象；使用万用表检测整流器输出电压，确认是否正常；检查整流器连接线路，排除接触不良或断线情况. 按照以上排查顺序，可以优先解决高频故障，逐步排除其他可能性，以快速恢复升降机正常运行.

表 6 不同模型回答效果对比

图 10 系统问答实例

1	任磊, 贾子翟, 赖李媛君, 等数据驱动的工业智能: 现状与展望[J]. 计算机集成制造系统, 2022, 28 (7): 1913- 1939 REN Lei, JIA Zizhai, LAI Liyuanjun, et al Data-driven industrial intelligence: current status and Future directions[J]. Computer Integrated Manufacturing Systems, 2022, 28 (7): 1913- 1939 doi: 10.13196/j.cims.2022.07.001
2	彭劼扬. 数控机床故障知识建模及故障推理的关键技术研究 [D]. 上海: 同济大学, 2022. PENG Jieyang. Research on the key technology of fault knowledge modeling and fault reasoning of CNC machine tools [D]. Shanghai: Tongji University, 2022.
3	周彬, 花豹, 陆玉前, 等面向设备点检故障根因分析的因果知识建模方法[J]. 计算机集成制造系统, 2023, 29 (8): 2708- 2721 ZHOU Bin, HUA Bao, LU Yuqian, et al Causal knowledge modeling for root cause analysis of equipment spot-inspection failure[J]. Computer Integrated Manufacturing Systems, 2023, 29 (8): 2708- 2721 doi: 10.13196/j.cims.2023.08.017
4	左戴悦, 蒋文波, 郑杭彬, 等面向缺陷识别的可解释视觉问答方法[J]. 计算机集成制造系统, 2025, 31 (6): 2084- 2097 ZUO Daiyue, JIANG Wenbo, ZHENG Hangbin, et al Interpretable visual question answering method for defect recognition[J]. Computer Integrated Manufacturing Systems, 2025, 31 (6): 2084- 2097 doi: 10.13196/j.cims.2024.0303
5	邢雪琪, 丁雨童, 夏唐斌, 等基于知识图谱的商用飞机维修方案推荐系统集成建模[J]. 浙江大学学报: 工学版, 2023, 57 (3): 512- 521 XING Xueqi, DING Yutong, XIA Tangbin, et al Integrated modeling of commercial aircraft maintenance plan recommendation system based on knowledge graph[J]. Journal of Zhejiang University: Engineering Science, 2023, 57 (3): 512- 521
6	张栋豪, 刘振宇, 郏维强, 等知识图谱在智能制造领域的研究现状及其应用前景综述[J]. 机械工程学报, 2021, 57 (5): 90- 113 ZHANG Donghao, LIU Zhenyu, JIA Weiqiang, et al A review on knowledge graph and its application prospects to intelligent manufacturing[J]. Journal of Mechanical Engineering, 2021, 57 (5): 90- 113 doi: 10.3901/JME.2021.05.090
7	SHI D, ZHANG F, WANG N, et al Intelligent electromagnetic compatibility management of cell phones by using knowledge graphs[J]. IEEE Transactions on Industrial Electronics, 2019, 66 (12): 9808- 9816 doi: 10.1109/TIE.2019.2893839
8	刘华一, 鄢萍, 周强, 等基于语义网的机床故障诊断知识扩展方法[J]. 计算机集成制造系统, 2020, 26 (3): 609- 622 LIU Huayi, YAN Ping, ZHOU Qiang, et al Machine tool fault diagnosis knowledge expansion method based on semantic web[J]. Computer Integrated Manufacturing Systems, 2020, 26 (3): 609- 622 doi: 10.13196/j.cims.2020.03.004
9	贝毅君, 周勇, 高克威面向数控机床设备维护的知识问答技术[J]. 计算机集成制造系统, 2022, 28 (9): 2881- 2893 BEI Yijun, ZHOU Yong, GAO Kewei Question answers technology towards maintenance of CNC machine tools[J]. Computer Integrated Manufacturing Systems, 2022, 28 (9): 2881- 2893 doi: 10.13196/j.cims.2022.09.019
10	中国信息通信研究院. 2024年智能运维(AIOps)行业分析: 大模型驱动下运维效率提升至毫秒级[EB/OL]. (2025−04−14) [2025−06−15]. https://www.vzkoo.com/document/202504117ef7f65d71e49fed466d85fb.html.
11	ZHENG S, PAN K, LIU J, et al Empirical study on fine-tuning pre-trained large language models for fault diagnosis of complex systems[J]. Reliability Engineering and System Safety, 2024, 252: 110382 doi: 10.1016/j.ress.2024.110382
12	PANG Z, LUAN Y, CHEN J, et al ParInfoGPT: an LLM-based two-stage framework for reliability assessment of rotating machine under partial information[J]. Reliability Engineering and System Safety, 2024, 250: 110312 doi: 10.1016/j.ress.2024.110312
13	张华, 张美航, 鄢威, 等. 基于混合微调大模型的变工况机械加工能耗预测方法 [EB/OL]. (2025−01−13). https://link.cnki.net/doi/10.13196/j.cims.2024.0426.
14	LIU P, QIAN L, ZHAO X, et al Joint knowledge graph and large language model for fault diagnosis and its application in aviation assembly[J]. IEEE Transactions on Industrial Informatics, 2024, 20 (6): 8160- 8169 doi: 10.1109/TII.2024.3366977
15	WEN Y, WANG Z, SUN J, et al. MindMap: knowledge graph prompting sparks graph of thoughts in large language models [EB/OL]. [2025−02−27]. https://arxiv.org/abs/2308.09729.
16	GAO Y, XIONG Y, GAO X, et al. Retrieval-augmented generation for large language models: a survey [EB/OL]. [2025−02−27]. http://arxiv.org/abs/2312.10997.
17	刘瑞祥, 柳先辉, 赵卫东, 等基于大语言模型的业务流程自动建模方法[J]. 计算机集成制造系统, 2025, 31 (6): 2001- 2014 LIU Ruixiang, LIU Xianhui, ZHAO Weidong, et al Automatic business process modeling method based on large language models[J]. Computer Integrated Manufacturing Systems, 2025, 31 (6): 2001- 2014 doi: 10.13196/j.cims.2024.0399
18	张鹤译, 王鑫, 韩立帆, 等大语言模型融合知识图谱的问答系统研究[J]. 计算机科学与探索, 2023, 17 (10): 2377- 2388 ZHANG Heyi, WANG Xin, HAN Lifan, et al Research on question answering system on joint of knowledge graph and large language models[J]. Journal of Frontiers of Computer Science and Technology, 2023, 17 (10): 2377- 2388
19	李莉, 时榕良, 郭旭, 等融合大模型与图神经网络的电力设备缺陷诊断[J]. 计算机科学与探索, 2024, 18 (10): 2643- 2655 LI Li, SHI Rongliang, GUO Xu, et al Diagnosis of power system defects by large language models and graph neural networks[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18 (10): 2643- 2655 doi: 10.3778/j.issn.1673-9418.2405085
20	OJIMA Y, SAKAJI H, NAKAMURA T, et al. Knowledge management for automobile failure analysis using graph RAG [EB/OL]. [2025−01−14]. http://arxiv.org/abs/2411.19539.
21	LUO H, E H, TANG Z, et al. ChatKBQA: a generate-then-retrieve framework for knowledge base question answering with fine-tuned large language models [EB/OL]. [2025−02−27]. https://arxiv.org/abs/2310.08975.
22	WEI J, WANG X, SCHUURMANS D, et al. Chain-of-thought prompting elicits reasoning in large language models [EB/OL]. [2025−02−27]. http://arxiv.org/abs/2201.11903.
23	WU Y, HUANG Y, HU N, et al. CoTKR: Chain-of-Thought enhanced knowledge rewriting for complex knowledge graph question answering [EB/OL]. [2025−02−27]. https://arxiv.org/abs/2409.19753.
24	LIU P, YUAN W, FU J, et al Pre-train, prompt, and predict: a systematic survey of prompting methods in natural language processing[J]. ACM Computing Surveys, 2023, 55 (9): 1- 35 doi: 10.1145/3560815
25	HU E J, SHEN Y, WALLIS P, et al. LoRA: low-rank adaptation of large language models [EB/OL]. [2025−02−27]. http://arxiv.org/abs/2106.09685.
26	ZHAO J, WANG T, ABID W, et al. LoRA Land: 310 fine-tuned LLMs that rival GPT-4, a technical report [EB/OL]. [2025−02−27]. http://arxiv.org/abs/2405.00732.
27	EDGE D, TRINH H, CHENG N, et al. From local to global: a graph RAG approach to query-focused summarization [EB/OL]. [2025−02−27]. https://arxiv.org/abs/2404.16130.
28	GRATTAFIORI A, DUBEY A, JAUHRI A, et al. The Llama 3 herd of models [EB/OL]. [2025−02−27]. http://arxiv.org/abs/2407.21783.
29	ZHENG Y, ZHANG R, ZHANG J, et al. Llamafactory: unified efficient fine-tuning of 100+ language models [EB/OL]. [2025−02−27]. https://arxiv.org/abs/2403.13372
30	PAPINENI K, ROUKOS S, WARD T, et al. BLEU: a method for automatic evaluation of machine translation [C]// 40th Annual Meeting on Association for Computational Linguistics. Philadelphia, Morristown: ACL, 2001.

[1]	魏蕴田,金苍宏,费峥东,郑铜亚,王晓亮,宋明黎. 大小模型协同优化的兴趣点轨迹预测框架[J]. 浙江大学学报(工学版), 2026, 60(6): 1251-1260.
[2]	陈思如,舒元超. 多模态大模型边缘部署与推理加速技术综述[J]. 浙江大学学报(工学版), 2026, 60(4): 723-737.
[3]	李延松,陈宁,刘锋光,陈盼,黄晓峰,葛慧丽. 基于分阶段语义感知的事件抽取大语言模型框架[J]. 浙江大学学报(工学版), 2026, 60(3): 527-535.
[4]	冯超文,耿程晨,刘英莉. 基于嵌入特征和稀疏矩阵的实体对齐方法[J]. 浙江大学学报(工学版), 2026, 60(2): 379-387.
[5]	黄孝喜,查正超,陆诗佳. 基于大语言模型的中文隐喻多维度评估[J]. 浙江大学学报(工学版), 2026, 60(2): 388-395.
[6]	谢涛,葛慧丽,陈宁,汪晓锋,李延松,黄晓峰. 知识嵌入增强的对比推荐模型[J]. 浙江大学学报(工学版), 2026, 60(1): 90-98.
[7]	竹志超,李建强,齐宏智,赵青,高齐,李思颖,蔡嘉怡,沈金炎. 大模型知识引导的复合多注意力文档级关系抽取方法[J]. 浙江大学学报(工学版), 2025, 59(9): 1793-1802.
[8]	李忠良,陈麒,石琳,杨朝,邹先明. 时间感知组合的动态知识图谱补全[J]. 浙江大学学报(工学版), 2024, 58(8): 1738-1747.
[9]	李劲业,李永强. 融合知识图谱的时空多图卷积交通流量预测[J]. 浙江大学学报(工学版), 2024, 58(7): 1366-1376.
[10]	何勇禧,韩虎,孔博. 基于多依赖图和知识融合的方面级情感分析模型[J]. 浙江大学学报(工学版), 2024, 58(4): 737-747.
[11]	李松,王哲,张丽平. SL-tgStore：新的时序知识图谱存储模型[J]. 浙江大学学报(工学版), 2024, 58(3): 449-458.
[12]	王慧欣,童向荣. 融合知识图谱的推荐系统研究进展[J]. 浙江大学学报(工学版), 2023, 57(8): 1527-1540.
[13]	李松,舒世泰,郝晓红,郝忠孝. 融合文本描述和层次类型的知识表示学习方法[J]. 浙江大学学报(工学版), 2023, 57(5): 911-920.
[14]	邢雪琪,丁雨童,夏唐斌,潘尔顺,奚立峰. 基于知识图谱的商用飞机维修方案推荐系统集成建模[J]. 浙江大学学报(工学版), 2023, 57(3): 512-521.
[15]	苏丰龙,景宁. 基于关系聚合的时序知识图谱表示学习[J]. 浙江大学学报(工学版), 2023, 57(2): 235-242.

Viewed

Full text

Abstract

Cited

Shared

Discussed