電梯為高層與多樓層建築中不可或缺的垂直運輸系統，其設計品質與法規合規性攸關使用安全與建築效能。依據《CNS 15827 電梯設施設計規範》，設計階段需針對車廂尺寸、井道淨寬、開門寬度與機坑深度等條件進行規劃與檢核。然而，臺灣電梯設計流程在實務中仍面臨多項挑戰，包括專業人力短缺、作業流程緊縮與法規判讀困難，已影響設計品質與產業效率。
具備法規理解與空間邏輯能力之設計人才長期不足，導致作業高度依賴資深技術者，知識難以有效傳承。設計流程常受限於工期與市場壓力，錯誤率上升、修正頻繁，品質控管困難。此外，設計成果缺乏統一格式，審查與協作過程中各方理解落差大，溝通成本與責任歸屬爭議明顯增加。
現行法規檢核多仰賴人工查閱與經驗判斷，缺乏即時回饋機制與結構化工具。條文內容常具有條件邏輯、跨條號引用與例外規定，使新手設計人員難以準確轉譯條文至空間參數，導致錯誤風險提升與學習曲線過於陡峭。
隨著自然語言處理（NLP）與生成式人工智慧（Generative AI）發展日益成熟，特別是大型語言模型（LLMs）於語意解析與條文結構理解方面的應用，提供了解決上述問題的新契機。透過語意標註與結構化運算，可將條文轉換為機器可讀之規則格式，支援參數推導與法規檢核任務之自動化。
本研究據此提出一套整合 GPT 模型之電梯設計輔助平台。系統可根據輸入條件，自動產出建議參數與法規比對結果，並生成對應之設計報表與 2D 圖說，協助使用者快速完成判斷與配置。此外，平台亦具教學與訓練潛能，有助於新進人員理解條文邏輯、預防錯誤並提升學習效率。整體而言，本研究展示生成式 AI 技術應用於電梯設計流程之可行性，並提出一套具實務價值的自動化設計與合規檢核解決方案。

Elevators are essential vertical transportation in high-rise and multi-story buildings; design quality and regulatory compliance directly affect user safety and building performance. Under Taiwan’s CNS 15827 Elevator Design Standard, designers must verify car dimensions, shaft clearances, door widths, pit depths, and related constraints.
In practice, however, Taiwan’s workflow faces talent shortages, compressed schedules, and difficult code interpretation, which undermine quality and efficiency. Overreliance on senior experts hinders knowledge transfer; tight timelines increase errors, rework, and inconsistent deliverables; and nonstandard outputs raise review friction and responsibility disputes. Compliance checking remains largely manual, lacking real-time feedback and structured tools. Conditional logic, cross-references, and exceptions make clauses hard to translate into spatial parameters, steepening the learning curve for junior staff.
This study proposes an AI-assisted elevator design platform integrating GPT-based large language models. We apply semantic annotation and structured computation to convert regulatory text into machine-readable rules for automated parameter derivation and constraint checking. Given basic project inputs, the system recommends compliant configurations, generates 2D schematic drawings, and produces traceable compliance reports that cite governing clauses. Beyond automation, it serves as a training companion that explains code logic, highlights edge cases, and standardizes deliverables for collaboration. Overall, the results demonstrate the feasibility and practical value of generative-AI-enabled automated design and compliance checking for elevator projects.

[1]Al-Homoud, M. S. (2001). Computer-aided building energy analysis techniques. Building and Environment, 36(4), 421–433.
[2]Al-Sharif, L. (2017). The design of elevator systems in high rise buildings, Part 1. Lift Report, 43(5), 46–62.
[3]Chen, N., Lin, X., Jiang, H., & An, Y. (2024). Automated Building Information Modeling compliance check through a large language model combined with deep learning and ontology. Buildings, 14(7), 1983.
[4]Cicala, G., Demarchi, S., Menapace, M., & Tacchella, A. (2022). A comparison of declarative AI techniques for computer automated design of elevator systems. Intelligenza Artificiale.
[5]Dahl, M., Magesh, V., Suzgun, M., & Ho, D. E. (2024). Profiling legal hallucinations in large language models. Journal of Legal Analysis, 16(1), 64–116. (also on arXiv).
[6]Eappen, R. (2017). Elevator maintenance considerations for supertall buildings. CTBUH Journal, 2017(Issue II), 28–33.
[7]El-Diraby, T. (2019). Beyond e-permitting: Framing the business case for automated rule checking in AEC in the era of big data (Technical Report No. TR-1012, v1.1). buildingSMART International Regulatory Room.
[8]Emaminejad, N., & Akhavian, R. (2022). Trustworthy AI and robotics: Implications for the AEC industry. Advanced Engineering Informatics, 54, 101819.
[9]Ghimire, P., Kim, K., & Acharya, M. (2024). Opportunities and challenges of generative AI in construction industry: Focusing on adoption of text-based models. Buildings, 14(1), 220.
[10]Guha, N., Nyarko, J., Ho, D. E., Henderson, P., & others. (2023). LegalBench: A collaboratively built benchmark for measuring legal reasoning in large language models. arXiv.
[11]Liang, C.-J., Le, T.-H., Ham, Y., Mantha, B. R. K., Cheng, M. H., & Lin, J. J. (2024). Ethics of artificial intelligence and robotics in the architecture, engineering, and construction industry. Advanced Engineering Informatics, 60, 102449.
[12]Madireddy, S., Gao, L., Din, Z. U., Kim, K., Senouci, A., Han, Z., & Zhang, Y. (2025). Large language model-driven code compliance checking in building information modeling. Electronics, 14(11), 2146.
[13]Magesh, V., Surani, F., Dahl, M., Suzgun, M., Manning, C. D., & Ho, D. E. (2024). Hallucination-free? Assessing the reliability of leading AI legal research tools. arXiv preprint arXiv:2405.20362.
[14]Moundas, M., White, J., & Schmidt, D. C. (2024). Prompt patterns for structured data extraction from unstructured text. In Proceedings of the IEEE International Conference on Big Data (pp. 4426–4435). IEEE.
[15]Onatayo, D., Onososen, A., Oyediran, A. O., Oyediran, H., Arowoiya, V., & Onatayo, E. (2024). Generative AI applications in architecture, engineering, and construction: Trends, implications for practice, education & imperatives for upskilling—A review. Architecture, 4(4), 877–902.
[16]Peng, J., & Liu, X. (2023). Automated code compliance checking research based on BIM and knowledge graph. Scientific Reports, 13, 7065.
[17]Pipitone, N., & Houir Alami, G. (2024). LegalBench-RAG: A benchmark for retrieval-augmented generation in the legal domain. arXiv.
[18]Preidel, C., & Borrmann, A. (2016). Towards code compliance checking on the basis of a visual programming language. Journal of Information Technology in Construction, 21, 402–421.
[19]Prieto, S. A., Mengiste, E. T., & García de Soto, B. (2023). Investigating the use of ChatGPT for the scheduling of construction projects. Buildings, 13(4), 857.
[20]PUBs, F.（1993）. Integration definition for function modelling（IDEF0）. Federal information processing standards publication, 183.
[21]Stępień, M., Vonthron, A., & König, M. (2021). OpenBIMRL—An open format for code compliance checking of complex functional requirements from a regulation to the BIM model. In EG-ICE 2021 Workshop.
[22]White, J., Fu, Q., Hays, S., Sandborn, M., Olea, C., Gilbert, H., Elnashar, A., Spencer-Smith, J., & Schmidt, D. C. (2023). A prompt pattern catalog to enhance prompt engineering with ChatGPT (arXiv:2302.11382). arXiv.
[23]Zhang, Z., Ma, L., & Nisbet, N. (2023). Unpacking ambiguity in building requirements to support automated compliance checking. Journal of Management in Engineering, 39(5), 04023033.
[1]中華民國電梯協會，http://www.elevator.org.tw/web/index/index.jsp
[2]台灣升降設備檢查促進安全協會，https://www.tesa-society.org.tw/decree
[3]胡亦涵（2019）。法規參數化智慧元件於建築設計之應用—以樓梯建置為例，碩士論文，國立高雄大學創意設計與建築學系，臺灣高雄。