Large-Scale Language Model Assisted Construction of Multi-Source Heterogeneous Knowledge Graphs for Marine Renewable Energy

Abstract

Marine renewable energy systems, particularly offshore wind and photovoltaic (PV) installations, generate large volumes of heterogeneous maintenance texts. However, the resulting knowledge remains fragmented due to dispersed sources, diverse formats, and domain-specific terminology. To address these challenges, this study proposes a large-scale language model assisted methodology for constructing a multi-source heterogeneous knowledge graph for intelligent operation and maintenance (O&M). The method integrates unified document preprocessing, domain-oriented prompt engineering, large-scale language model-based entity and relation extraction, and multi-level entity normalization. It systematically transforms unstructured documents (e.g., standards, procedures, manuals, inspection records, and environmental reports) into structured triples, enabling the construction of a dynamically evolving O&M knowledge graph. A rigorous ablation study on real-world offshore wind and PV datasets demonstrates that the proposed workflow exhibits exceptional robustness against OCR noise (e.g., scanned artifacts, stamps, and signatures) and substantially improves extraction volume, accuracy, and coverage compared with traditional methods. In particular, combining high-quality preprocessing and optimized prompts yields the most reliable and semantically coherent results. The study provides a practical technical pathway for automated knowledge management in marine renewable energy and offers a foundation for future applications in intelligent diagnostics, predictive maintenance, and digital-twin systems.

Keywords

Knowledge graph construction / Operation and maintenance / Large-scale language models / Marine renewable energy

Statement of the Use of Generative AI and AI-Assisted Technologies in the Writing Process

During the preparation of this manuscript, the author utilized AI large-scale models to assist with translation. All content generated by these tools has been thoroughly reviewed, revised, and verified by the author, who assumes full responsibility for the accuracy and completeness of the published material.

Acknowledgements

The authors would like to thank the financial support of Nanning city science and Technology Bureau.

Author Contributions

Methodology, M.L.; Software, Z.J.; Validation, Y.Y. and X.M.; Investigation, Z.J. and T.L.; Writing Original Draft Preparation, Z.J.; Review & Editing, J.Z. and Y.W.; Supervision, J.Z. and Q.M.

Ethics Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data will be made available on request.

Funding

This research was funded by the Key R&D Program of the Nanning Science Research and Technology Development Plan grant number (20253057).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

[1] Lian J, Cui L, Fu Q. Offshore Renewable Energy Advance. Mar. Energy Res. 2024, 1, 10006. DOI:10.70322/mer.2024.10006

[2] Gao X, Dong X, Ma Q, Liu M, Li Y, Lian J. Marine Photovoltaic Module Salt Detection via Semantic-Driven Feature Optimization in Mask R-CNN. Mar. Energy Res. 2025, 2, 10015. DOI:10.70322/mer.2025.10015

[3] He Z, Yang J, Tang G, Yang Y, Du Z, Zhang S. Key technologies and development trends of VSC-HVDC transmission for offshore wind power. Renew. Energy Syst. Equip. 2025, 1, 35-49. DOI:10.1016/j.rese.2024.09.001

[4] Gao X, Wang T, Liu M, Lian J, Yao Y, Yu L, et al. A framework to identify guano on photovoltaic modules in offshore floating photovoltaic power plants. Sol. Energy 2024, 274, 112598. DOI:10.1016/j.solener.2024.112598

[5] Ali AAIM, Imran MMH, Jamaludin S, Ayob AFM, Suhrab MIR, Norzeli SM, et al. A review of predictive maintenance approaches for corrosion detection and maintenance of marine structures. J. Sustain. Sci. Manag. 2024, 19, 182-202. DOI:10.46754/jssm.2024.04.014

[6] Fox H, Pillai AC, Friedrich D, Collu M, Dawood T, Johanning L. A review of predictive and prescriptive offshore wind farm operation and maintenance. Energies 2022, 15, 504. DOI:10.3390/en15020504

[7] Yang Y, Wang S, Tan J, Chen L. The Intelligent Operation and Maintenance Management System for Offshore Wind Farms. South. Energy Constr. 2025, 8, 74-79. DOI:10.16516/j.gedi.issn2095-8676.2021.01.011.(InChinese)

[8] Lou P, Yu D, Jiang X, Hu J, Zeng Y, Fan C. Knowledge graph construction based on a joint model for equipment maintenance. Mathematics 2023, 11, 3748. DOI:10.3390/math11173748

[9] Bouwer UI. Maintenance strategies for deep-sea offshore wind turbines. J. Qual. Maint. Eng. 2010, 16, 367-381. DOI:10.1108/13552511011084526

[10] Pi Q, Lu J, Zhu T, Peng Y. A large language model-enhanced zero-shot knowledge extraction method. Comput. Sci. 2025, 52, 22-29. DOI:10.1186/jsjkx.241000049.(InChinese)

[11] Yang Y, Chen S, Zhu Y, Liu X, Ma W, Feng L. Intelligent extraction of reservoir dispatching information integrating large language model and structured prompts. Sci. Rep. 2024, 14, 14140. DOI:10.1038/s41598-024-64954-0

[12] Tang H, Liu Z, Chen D, Chu Q. ChatSOS: A large language model-based question-answering system for safety engineering knowledge. China Saf. Sci. J. 2024, 34, 178-185. DOI:10.16265/j.cnki.issn1003-3033.2024.08.1901.(InChinese)

[13] Sarhan AM, Ali HA, Wagdi M, Ali B, Adel A, Osama R, et al. CV content recognition using YOLOv8 and Tesseract-OCR deep learning. Comput. J. 2025, bxaf124. DOI:10.1093/comjnl/bxaf124

[14] Yang M, Yoo S, Jeong O. DeNERT-KG: Named entity and relation extraction model using DQN, knowledge graph, and BERT. Appl. Sci. 2020, 10, 6429. DOI:10.3390/app10186429

[15] Xu D, Chen W, Peng W, Zhang C, Xu T, Zhao X, et al. Large language models for generative information extraction: A survey. Front. Comput. Sci. 2024, 18, 186357. DOI:10.1007/s11704-024-40555-y

[16] Xing J, Nie X, Li Q, Fan X. Construction of distributed photovoltaic operation and maintenance knowledge base based on time series knowledge map. Discov. Appl. Sci. 2025, 7, 409. DOI:10.1007/s42452-025-06858-w

[17] Zhang H, Zhang H, Yan W, Zhu S, Jiang Z. Precise extraction of remanufacturing process knowledge using chain-of-thought prompting with large language models. Manuf. Technol. Mach. Tool 2025 10, 90-98. DOI:10.19287/j.mtmt.1005-2402.2025.10.007.(InChinese)

[18] Zhang N, Xu X, Tao L, Yu H, Ye H, Qiao S, et al. DeepKE: A deep learning based knowledge extraction toolkit for knowledge base population. arXiv 2025, arXiv:2201.03335. DOI:10.48550/arXiv.2201.03335

[19] Wei Z, Xiao Y, Tong Y, Xu W, Wang Y. Linear building pattern recognition via spatial knowledge graph. Acta Geod. Et Cartogr. Sin. 2023, 52, 1355-1363. DOI:10.11947/j.AGCS.2023.20220121.(InChinese)

[20] Di Paolo G, Rincon-Yanez D, Senatore S. A quick prototype for assessing OpenIE knowledge graph-based question-answering systems. Information 2023, 14, 186. DOI:10.3390/info14030186

[21] Lang TA, Stroup DF. Who knew? The misleading specificity of “double-blind” and what to do about it. Trials 2020, 21, 697. DOI:10.1186/s13063-020-04607-5

[22] Mishra BD, Tandon N, Clark P. Domain-targeted, high precision knowledge extraction. Trans. Assoc. Comput. Linguist. 2017, 5, 233-246. DOI:10.1162/tacl_a_00058

[23] Sabbatini F, Calegari R. On the evaluation of the symbolic knowledge extracted from black boxes. AI Ethics 2024, 4, 65-74. DOI:10.1007/s43681-023-00406-1

[24] Pazhouhan M, Karimi Mazraeshahi A, Jahanbakht M, Rezanejad K, Rohban MH. Wave and tidal energy: A patent landscape study. J. Mar. Sci. Eng. 2024, 12, 1967. DOI:10.3390/jmse12111967

[25] Giannelos S, Konstantelos I, Zhang X, Strabc G. A stochastic optimization model for network expansion planning under exogenous and endogenous uncertainty. Electr. Power Syst. Res. 2025, 248, 111894. DOI:10.1016/j.epsr.2025.111894

[26] Du Z, Yin H, Zhang X, Hu H, Liu T, Hou M, et al. Decarbonisation of Data Centre Networks through Computing Power Migration. In Proceedings of the 2025 IEEE 5th International Conference on Computer Communication and Artificial Intelligence (CCAI), Haikou, China, 23-25 May 2025; pp. 871-876. DOI:10.1109/CCAI65422.2025.11189418