Converting engineering drawings from Chinese GB to German DIN standards demands near-perfect accuracy: even a single hallucinated value or a misplaced annotation can render a drawing unsafe for manufacturing. Large language models (LLMs) are fluent, but they lack spatial reasoning and often generate uncontrolled outputs—two fatal flaws in this domain. We therefore design a neuro-symbolic framework that couples an LLM with a knowledge-based engineering shell. The core components are a Trident Router that assigns each annotation to a risk-appropriate path, an ontology-driven retrieval-augmented generation (RAG) module with entity masking, a hard symbolic validation loop, and a spatial-aware CAD layout adapter. Before generation, structural constraints protect critical entities; after generation, symbolic checks reject any output that violates standard or format rules. Together, these mechanisms prevent technical hallucinations (e.g., altered tolerances or standard codes) and layout collisions when the translated text is reinserted into the CAD drawing. By detailing the logic and theoretical operation of each module, we show how deep neuro-symbolic integration turns an LLM from an unconstrained text generator into a reliable, auditable, and standard-compliant industrial assistant.

Chaoyang Zhang, Shu Wu. LLM Hallucinations Suppression in Safety-Critical Engineering Drawing Conversion. Academic Journal of Computing & Information Science (2026), Vol. 9, Issue 5: 48-56. https://doi.org/10.25236/AJCIS.2026.090506.

[1] J. Freimuth, S. Kaiser, M. Schädler (Eds.), Standardization Strategies in China and India: Industrial Policy and Geopolitics and Implications for Europe, Springer Nature, 2024.

[2] R. Jiao, "From 'data-driven' to 'data-informed' design — grounding AI for design in knowledge, context and decisions," Journal of Engineering Design, vol. 37, pp. 1–22, 2026.

[3] N. Melluso, I. Grangel-González, G. Fantoni, "Enhancing Industry 4.0 standards interoperability via knowledge graphs with natural language processing," Computers in Industry, vol. 140, p. 103676, 2022.

[4] P. Kügler, F. Dworschak, B. Schleich, S. Wartzack, "The evolution of knowledge-based engineering from a design research perspective: Literature review 2012–21," Advanced Engineering Informatics, vol. 55, 2021.

[5] Z. Sun, L. Jia, J. Hao, et al., "KT-MDO: a knowledge-template-driven multidisciplinary design optimization framework," Advances in Engineering Software, vol. 214, 2026.

[6] N. Madhusudanan, B. Gurumoorthy, A. Chakrabarti, "From natural language text to rules: knowledge acquisition from formal documents for aircraft assembly," Journal of Engineering Design, vol. 30, pp. 417–444, 2019.

[7] Z. Shi, S. Ergan, "An ontology-based approach for Façades inspection checklist generation," Advanced Engineering Informatics, vol. 72, 2026.

[8] B. Patzák, S. Šulc, V. Šmilauer, "Towards digital twins: Design of an entity data model in the MuPIF simulation platform," Advanced Engineering Software, vol. 197, 2024.

[9] P. Lewis, E. Perez, A. Piktus, et al., "Retrieval-augmented generation for knowledge-intensive NLP," in Advances in Neural Information Processing Systems (NeurIPS), 2020, pp. 9459–9474.

[10] D. Shi, J. Li, O. Meyer, T. Bauernhansl, "Enhancing RAG for cognitive digital twins," Computers in Industry, vol. 171, 2025.

[11] H. Xiao, J. Zhuang, B. Yang, et al., "Generative knowledge-guided review system," Advanced Engineering Informatics, vol. 68, 2025.

[12] C. Yu, Y. Chen, L. Wang, T. Yan, "Reconstructed generative design for industrial products," Journal of Engineering Design, vol. 37, pp. 1234–1254, 2026.

[13] S. Jose, K.T.P. Nguyen, K. Medjaher, et al., "Multimodal industrial diagnostics," Expert Systems with Applications, vol. 255, 2024.

[14] NLP Research Team, "Faithful to the Document or to the World?," in Proceedings of the 62nd Annual Meeting of the ACL, 2024.

[15] R. Louf, B. Willard, "Efficient guided generation for LLM," in Proceedings of the 12th ICLR, 2024.

[16] M.T. Khan, et al., "Vision-Language Model for engineering drawing extraction," in ICIAI, Springer, 2026.

[17] J. Zhou, J.D. Camba, "Multimodal LLM in parametric CAD," Expert Systems with Applications, vol. 282, 2025.

[18] R. Andukuri, "CNN–GNN machining feature recognition," Expert Systems with Applications, vol. 303, 2026.

[19] F. Wang, et al., "AAGATNet: Attentive graph network," Computer-Aided Design, vol. 193, 2026.

[20] S. Yoo, N. Kang, "Explainable AI for manufacturing cost estimation," Expert Systems with Applications, vol. 183, 2021.

[21] F. Boussuge, et al., "Tensor factorisation for CAE model preparation," Computer-Aided Design, vol. 152, 2022.

[22] J.-S. Lee, et al., "Automation of trimming die design inspection," Engineering Applications of Artificial Intelligence, vol. 127, 2024.

[23] J. Song, et al., "P2Seg: Distance query from point to segments," Computer-Aided Design, vol. 189, 2025.