隨著智慧製造與數位轉型發展，企業累積了大量設備維修紀錄。然而，實務維修資料多由現場人員以非結構化文字記錄，存在語意不一致、缺乏關聯結構與不易再利用等共同問題，導致維修決策仍高度依賴人員經驗，且缺乏系統化推論之決策支援機制。
本研究自製造業維修管理之共同挑戰出發，提出一套結合生成式人工智慧（Generative Artificial Intelligence, GAI）與知識圖譜（Knowledge Graph, KG）之通用維修決策支援架構。本架構透過生成式人工智慧語意解析，自非結構化文字中提取設備類型、零組件、異常狀況、影響結果與維修方式等結構化語意資料，並建構多層式設備維修知識圖譜（Multi-layer Equipment Maintenance Knowledge Graph, MEM-KG），將分散的經驗紀錄轉化為具關聯性與可追溯性之知識結構。
在決策支援方面，本研究發展一套混合式推論機制，整合關係式推論、相似案例分析與生成式人工智慧語意輔助推論，即時提供維修方案建議、相似歷史案例與推論依據。相較於單純文字生成，多層式設備維修知識圖譜可呈現異常事件與維修方式間的關聯路徑，使維修建議具備可追溯性，作為現場人員判斷異常之決策輔助工具。
本研究以生技製造業實務資料進行概念驗證（Proof of Concept, POC），結果顯示候選案例覆蓋率達 86.67%、Top-1 建議一致率為 63.33%、Top-3 命中率為 73.33%，且系統推論回應時間穩定於 15 秒以內，初步驗證此通用架構之技術可行性與實務潛在效益。未來研究可朝向多模態資料擴展（如設備影像、異常聲音）以及結合圖檢索增強生成（Graph Retrieval-Augmented Generation, Graph-RAG）技術架構發展，以持續優化系統推論之穩定性與實務應用價值。

Enterprises have accumulated vast amounts of equipment maintenance records; however, most remain as unstructured text, resulting in semantic inconsistency, fragmented maintenance knowledge, and a heavy reliance on technicians' personal experience. Such limitations make historical maintenance cases difficult to retrieve and reuse systematically, thereby reducing maintenance efficiency and decision consistency. To address these challenges, this study proposes a general maintenance decision-support framework integrating Generative Artificial Intelligence (GAI) and Knowledge Graphs (KGs). Through GAI-based semantic parsing, the framework automatically extracts structured semantic information from historical maintenance records and transforms it into a Multilayer Equipment Maintenance Knowledge Graph (MEM-KG), converting fragmented maintenance experience into a structured, traceable, and reusable knowledge base.
Furthermore, a hybrid reasoning mechanism integrates relational reasoning, similarcase retrieval, and GAI-assisted semantic explanations to generate maintenance recommendations together with interpretable reasoning paths. Historical maintenance records collected from a biopharmaceutical manufacturing company were adopted as the proof-of-concept dataset for evaluation. Experimental results demonstrate a candidate case coverage of 86.67%, a Top-1 recommendation consistency of 63.33%, and a Top-3 hit rate of 73.33%, while maintaining an average inference response time below 15 seconds. Overall, the proposed framework demonstrates the feasibility of integrating GAI and KGs iv for explainable maintenance decision support and provides a practical approach for organizing dispersed maintenance knowledge into an intelligent and explainable decisionsupport system.

Alavi, M., Leidner, D., & Mousavi, R. (2024). Knowledge Management Perspective of Generative Artificial Intelligence. J. Assoc. Inf. Syst., 25, 15. https://doi.org/10.17705/1jais.00859
Bekar, E., Nyqvist, P., & Skoogh, A. (2020). An intelligent approach for data pre-processing and analysis in predictive maintenance with an industrial case study. Advances in Mechanical Engineering, 12. https://doi.org/10.1177/1687814020919207
Bouzidi, A., Rajaoarisoa, L., & Claeys, L. (2024). Explainable Artificial Intelligence Approach for Improving Head-Mounted Fault Display Systems. Future Internet, 16, 282. https://doi.org/10.3390/fi16080282
Dai, D., Zhao, B., Yu, Z., Franciosa, P., & Ceglarek, D. (2025). Generative and Predictive AI for digital twin systems in manufacturing. Frontiers in Artificial Intelligence, 8. https://doi.org/10.3389/frai.2025.1655470
Ding, Y., Li, H., Zhu, F., Wang, Z., Peng, W., & Xie, M. (2024). A Semi-Supervised Failure Knowledge Graph Construction Method for Decision Support in Operations and Maintenance. IEEE Transactions on Industrial Informatics, 20(3), 3104–3114. https://doi.org/10.1109/tii.2023.3299078
Dongbao. (2025). Predictive maintenance technology for industrial production equipment using cloud platform. Diagnostyka. https://doi.org/10.29354/diag/205035
Jasthi, V. N. (2024). Clinical Decision Support Systems Powered by Hybrid Knowledge Graph and ML Models. International Journal of Multidisciplinary Research and Growth Evaluation. https://doi.org/10.54660/.ijmrge.2024.5.3.1109-1115
Khalil, A., & Rostam, S. (2024). Machine Learning-based Predictive Maintenance for Fault Detection in Rotating Machinery: A Case Study. Engineering, Technology & Applied Science Research. https://doi.org/10.48084/etasr.6813
Le, T.-T.-H., Adiputra, A. A., Yun, J., & Kim, H. (2025). Anomaly Detection in Industrial Machine Sounds Using High-Frequency Features and Gate Recurrent Unit Networks. IEEE Access, 13, 77165–77186. https://doi.org/10.1109/access.2025.3565812
Lim, K. Y. H., Yosal, T. S., Chen, C.-H., Zheng, P., Wang, L., & Xu, X. (2023). Graph-enabled cognitive digital twins for causal inference in maintenance processes. International Journal of Production Research, 62, 4717–4734. https://doi.org/10.1080/00207543.2023.2274335
Lou, P., Yu, D., Jiang, X., Hu, J., Zeng, Y., & Fan, C. (2023). Knowledge Graph Construction Based on a Joint Model for Equipment Maintenance. Mathematics, 11(17). https://doi.org/10.3390/math11173748
Mikołajewska, E., Mikołajewski, D., Mikołajczyk, T., & Pączkowski, T. (2025). Generative AI in AI-Based Digital Twins for Fault Diagnosis for Predictive Maintenance in Industry 4.0/5.0. Applied Sciences. https://doi.org/10.3390/app15063166
Mohapatra, A. (2024). Generative AI for Predictive Maintenance: Predicting Equipment Failures and Optimizing Maintenance Schedules Using AI. International Journal of Scientific Research and Management (IJSRM), 12(11), 1648–1672. https://doi.org/10.18535/ijsrm/v12i11.ec03
Musa, S., O, A., & N, O. (2025). A Machine Learning Framework for Predictive Maintenance in Smart Facilities Using IoT Sensor Data. Journal of Engineering Research and Reports. https://doi.org/10.9734/jerr/2025/v27i81601
Nkrumah, M. A. (2024). Data Mining with Explainable Deep Representation Models for Predicting Equipment Failures in Smart Manufacturing Environments. Magna Scientia Advanced Research and Reviews. https://doi.org/10.30574/msarr.2024.12.1.0179
Paroha, A. D. (2024). Generative AI-based Health Monitoring and Prediction of Electrical Submersible Pumps 2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT), https://ieeexplore.ieee.org/stampPDF/getPDF.jsp?tp=&arnumber=10725256&ref=
Rousopoulou, V., Nizamis, A., Vafeiadis, T., Ioannidis, D., & Tzovaras, D. (2020). Predictive Maintenance for Injection Molding Machines Enabled by Cognitive Analytics for Industry 4.0. Frontiers in Artificial Intelligence, 3. https://doi.org/10.3389/frai.2020.578152
Saiyam, A. (2025). Transforming AI Decision Support System with Knowledge Graphs & CAG. International Journal on Engineering Artificial Intelligence Management, Decision Support, and Policies, 2(2), 15–23. https://doi.org/10.63503/j.ijaimd.2025.110
Stewart, M., Hodkiewicz, M., Liu, W., & French, T. (2022). MWO2KG and Echidna: Constructing and exploring knowledge graphs from maintenance data. Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 238, 920–932. https://doi.org/10.1177/1748006x221131128
Su, C., Jiang, Q., Han, Y., Wang, T., & He, Q. (2025). Knowledge graph-driven decision support for manufacturing process: A graph neural network-based knowledge reasoning approach. Advanced Engineering Informatics, 64. https://doi.org/10.1016/j.aei.2024.103098
Sundaram, S., & Zeid, A. (2024). Technical language processing for Prognostics and Health Management: applying text similarity and topic modeling to maintenance work orders. Journal of Intelligent Manufacturing, 36(3), 1637–1657. https://doi.org/10.1007/s10845-024-02323-4
Tamakloe, E., Kommey, B., Kponyo, J., Tchao, E., Agbemenu, A., & Klogo, G. (2025). Predictive AI Maintenance of Distribution Oil‐Immersed Transformer via Multimodal Data Fusion: A New Dynamic Multiscale Attention CNN‐LSTM Anomaly Detection Model for Industrial Energy Management. IET Electric Power Applications. https://doi.org/10.1049/elp2.70011
Taoufyq, H., Guemmat, K. E., Mansouri, K., & Akef, F. (2025). Predictive maintenance approaches: A systematic literature review. Journal of Industrial Engineering and Management. https://doi.org/10.3926/jiem.8537
Tsallis, C., Papageorgas, P., Piromalis, D., & Munteanu, R. (2025). Application-Wise Review of Machine Learning-Based Predictive Maintenance: Trends, Challenges, and Future Directions. Applied Sciences. https://doi.org/10.3390/app15094898
Verwimp, E., Ben-David, S., Bethge, M., Cossu, A., Gepperth, A., Hayes, T., Hullermeier, E., Kanan, C., Kudithipudi, D., Lampert, C., Mundt, M., Pascanu, R., Popescu, A., Tolias, A., Weijer, J., Liu, B., Lomonaco, V., Tuytelaars, T., & Van De Ven, G. (2023). Continual Learning: Applications and the Road Forward. Trans. Mach. Learn. Res., 2024. https://doi.org/10.48550/arxiv.2311.11908
Walker, C., Agarwal, V., Lin, L., Hall, A., Hill, R., Boring, R., Mortenson, T., & Lybeck, N. (2023). Explainable Artificial Intelligence Technology for Predictive Maintenance. https://doi.org/10.2172/1998555
Wan, Y., Liu, Y., Chen, Z., Chen, C., Li, X., Hu, F., & Packianather, M. (2024). Making knowledge graphs work for smart manufacturing: Research topics, applications and prospects. Journal of Manufacturing Systems, 76, 103–132. https://doi.org/10.1016/j.jmsy.2024.07.009
Wang, R., Sun, Y., Peng, T., Hua, Y., Wang, G., & Yan, Y. (2025). A knowledge graph-aided decision guidance method for product conceptual design. Journal of Engineering Design, 36(2), 216–255. https://doi.org/10.1080/09544828.2024.2368405
Woods, C., Hodkiewicz, M., & French, T. (2024). Semantic Quality Assurance of Industrial Maintenance Procedures. IEEE Access, 12, 122029–122046. https://doi.org/10.1109/access.2024.3441757
Xia, L., Zheng, P., Li, S., Lyu, P., Lee, C. K. M., Zhou, J., & Wang, K. (2022). A Knowledge Graph-based Link Prediction for Interpretable Maintenance Planning in Complex Equipment 2022 13th International Conference on Reliability, Maintainability, and Safety (ICRMS), https://ieeexplore.ieee.org/stampPDF/getPDF.jsp?tp=&arnumber=9944561&ref=
Yang, L., Luo, Z., Zhang, S.-S., Teng, F., & Li, T.-J. (2024). Continual Learning for Smart City: A Survey. IEEE Transactions on Knowledge and Data Engineering, 36, 7805–7824. https://doi.org/10.1109/tkde.2024.3447123