Ontology-Grounded Large Language Models for Reliable Querying of Wind Turbine Inspection Knowledge
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Inspection reports of industrial assets contain valuable diagnostic knowledge, but their unstructured nature makes automated reasoning difficult. This paper presents an ontology-grounded question answering framework for querying wind turbine gearbox inspection reports using natural language. Inspection data are automatically parsed into structured representations consisting of a domain ontology and a knowledge graph. On top of this representation, a large language model translates user questions into SPARQL queries. To improve robustness, we employ example-based query generation combined with an Ontology-Based Query Checker (OBQC) that validates generated queries against ontology constraints and iteratively repairs violations before execution. The approach is evaluated on real-world inspection reports using 50 diagnostic prompts of varying complexity, achieving a 96\% successful execution rate. Results demonstrate that combining ontology grounding with constrained LLM-based query generation enables reliable and flexible diagnostic reasoning over inspection documentation.
How to Cite
##plugins.themes.bootstrap3.article.details##
Large Language Models (LLMs), Knowledge Graphs, SPARQL Generation, Ontology, Predictive Maintenance, Wind Energy
Allemang, D., & Sequeda, J. (2024, May). Increasing the LLM accuracy for question answering: Ontologies to the rescue! arXiv. doi: 10.48550/arXiv.2405.11706
Arazzi, M., Ligari, D., Nicolazzo, S., & Nocera, A. (2025, February). Augmented knowledge graph querying leveraging LLMs. arXiv. doi: 10.48550/arXiv.2502.01298
Chatterjee, J., & Dethlefs, N. (2021, February). XAI4Wind: A multimodal knowledge graph database for explainable decision support in operations & maintenance of wind turbines. arXiv. doi: 10.48550/arXiv.2012.10489
Chen, R. (2025, March). Retrieval-augmented generation with knowledge graphs: A survey. In Computer Science Undergraduate Conference 2025 @ XJTU.
D’Abramo, J., Zugarini, A., & Torroni, P. (2025, May). Investigating large language models for text-to-SPARQL. In Proceedings of the 4th International Workshop on Knowledge-Augmented Methods for Natural Language Processing (pp. 66–80). Albuquerque, New Mexico, USA: Association for Computational Linguistics. doi: 10.18653/v1/2025.knowledgenlp-1.5
Emonet, V., Bolleman, J., Duvaud, S., de Farias, T. M., & Sima, A. C. (2025, February). LLM-based SPARQL query generation from natural language over federated knowledge graphs. arXiv. doi: 10.48550/arXiv.2410.06062
Liao, X., Chen, C., Wang, Z., Liu, Y., Wang, T., & Cheng, L. (2025, May). Large language model-assisted fine-grained knowledge graph construction for robotic fault diagnosis. Advanced Engineering Informatics, 65, 103134. doi: 10.1016/j.aei.2025.103134
Löwenmark, K., Taal, C., Schnabel, S., Liwicki, M., & Sandin, F. (2022, October). Technical language supervision for intelligent fault diagnosis in process industry. International Journal of Prognostics and Health Management, 13(2). doi: 10.36001/ijphm.2022.v13i2.3137
Mecharnia, T., & d’Aquin, M. (2025, January). Performance and limitations of fine-tuned LLMs in SPARQL query generation. In G. A. Gesese, H. Sack, H. Paulheim, A. Merono-Penuela, & L. Chen (Eds.), Proceedings of the Workshop on Generative AI and Knowledge Graphs (GenAIK) (pp. 69–77). Abu Dhabi, UAE: International Committee on Computational Linguistics.
Pan, J. Z., Razniewski, S., Kalo, J.-C., Singhania, S., Chen, J., Dietze, S., ... Graux, D. (2023, August). Large language models and knowledge graphs: Opportunities and challenges. arXiv. doi: 10.48550/arXiv.2308.06374
Pan, X., de Boer, V., & van Ossenbruggen, J. (2025, August). FIRESPARQL: A LLM-based framework for SPARQL query generation over scholarly knowledge graphs. arXiv. doi: 10.48550/arXiv.2508.10467
Peng, B., Zhu, Y., Liu, Y., Bo, X., Shi, H., Hong, C., ... Tang, S. (2024, September). Graph retrieval-augmented generation: A survey. arXiv. doi: 10.48550/arXiv.2408.08921
Team, D. S. (2024, August). Docling technical report. arXiv. doi: 10.48550/arXiv.2408.09869
van Cauter, Z., & Yakovets, N. (2024, August). Ontology-guided knowledge graph construction from maintenance short texts. In R. Biswas, L.-A. Kaffee, O. Agarwal, P. Minervini, S. Singh, & G. de Melo (Eds.), Proceedings of the 1st Workshop on Knowledge Graphs and Large Language Models (KaLLM 2024) (pp. 75–84). Bangkok, Thailand: Association for Computational Linguistics. doi: 10.18653/v1/2024.kallm-1.8
Xia, L., Liang, Y., Leng, J., & Zheng, P. (2023, April). Maintenance planning recommendation of complex industrial equipment based on knowledge graph and graph neural network. Reliability Engineering & System Safety, 232, 109068. doi: 10.1016/j.ress.2022.109068
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.