Automated BIM-FEA Pipeline Utilizing Large Language Models

In Ho Lee; Yeong Jae Lee; Dabeom Kim; Moon Soo Kang; Jun Won Kang

doi:10.7734/COSEIK.2026.39.2.81

All Issue

2026 Vol.39, Issue 2 Next Page

Research Paper

Automated BIM-FEA Pipeline Utilizing Large Language Models 거대언어모델을 활용한 BIM-유한요소해석 자동화 파이프라인

30 April 2026. pp. 81-88

PDF XML

Abstract

Artificial intelligence (AI) technologies require large, high-quality structural response data to estimate structural behavior or system characteristics under various scenarios. Although experimental results can serve as such data, large-scale numerical analysis outputs are more commonly used to reduce data-acquisition costs and improve the reliability of AI training. Currently, building information modeling data based on the Industry Foundation Classes (IFC) standard typically requires manual preprocessing before it can be utilized for structural analysis. Hence, this study proposes a unidirectional automated pipeline that encompasses the entire process from extracting geometric information from IFC-based building information models to conducting structural analysis as well as automatically converting both input data and analysis results into data formats suitable for AI training. In particular, an algorithm was developed that assigns member properties, boundary conditions, and load conditions to geometries extracted automatically from building information models using Python scripts based on the ABAQUS application programming interface, thereby enabling the automatic generation of the ABAQUS output database ODB file. Subsequently, the generated ODB file is automatically converted into data formats such as CSV, JSON, and prompt using Python scripts, thus allowing large language models to process them effectively. Sensitivity analysis and information-indexing tests show that the combination of GPT-4 LLM and the CSV data format affords the highest information-indexing performance. The outcomes of this study can contribute to the development of an advanced automated design system capable of verifying and improving the structural validity of building information models.

Keywords

Artificial intelligence (AI)

building information modeling (BIM)

structural analysis

large language model

Python

AI 기술을 활용해 다양한 시나리오에서 구조물의 응답 AI 또는 구조시스템의 특성을 추정하기 위해서는 대규모 고품질 구조 응답 데이터가 필요하다. 이러한 데이터로서 구조 실험 결과가 활용되기도 하지만, 데이터 구축 비용 절감과 AI 학습의 신뢰도 향상을 위하여 대규모 구조해석 결과가 사용되는 것이 일반적이다. 현재는 Industry Foundation Classes(IFC) 표준을 따르는 Building Information Modeling(BIM) 데이터를 구조해석에 활용하기 위하여 수작업으로 데이터를 전처리하는 경우가 많다. 이를 개선하기 위하여 이 연구에서는 IFC 기반 BIM 모델에서 기하 정보를 추출하는 것부터 구조해석까지 전 과정을 자동화하고 입력 및 해석 결과를 AI가 학습 가능한 데이터 형식으로 자동 변환하는 단방향 파이프라인을 제시한다. 특히 ABAQUS API 기반 Python 스크립트를 활용해 자동 추출한 BIM 기하 정보에 부재 특성과 경계 및 하중 조건을 부여하고 ABAQUS 해석 결과 파일인 ODB 파일을 자동으로 생성하는 알고리즘을 구현하였다. 생성된 ODB 파일은 Python 스크립트를 통해 대규모 언어 모델이 학습 가능한 CSV, JSON, prompt 등의 데이터 형식으로 자동 변환되도록 하였으며, 성능 민감도 및 정보 인덱싱 능력을 실험한 결과 GPT-4 LLM과 CSV 데이터 형식의 조합이 가장 높은 정보 인덱싱 성능을 보였다. 이 연구의 결과는 향후 BIM 모델의 구조적 타당성을 자동으로 검증하고 보완하는 혁신적 설계 자동화 시스템의 개발에 활용될 수 있다.

키워드

BIM

구조해석

거대 언어 모델

Python

References

Ahn, P.Y., Lee, C.H. (2025) Research on Efficient Preprocessing Method of Unstructured Data for AI Transformation, J. Inst. Internet, Broadcast. & Commun., 25(3), pp.183~192.

Antunes, M.L.R., Júnior, K.M.L.C., Ribeiro, J.C.L., de Oliveira, D.S., de Carvalho, J.M.F. (2024) Analysis of IFC Interoperability Data Schema for Project Representation, Autom. Constr., 166, 105650.

10.1016/j.autcon.2024.105650

Chen, H., Jiao, F., Ravaut, M., Farruque, N., Nguyen, X.P., Qin, C., Dey, M., Ding, B., Xiong, C., Joty, S., Zhou, Y. (2025) StructTest: Benchmarking LLMs' Reasoning through Compositional Structured Outputs, arXiv preprint arXiv:2412.18011.

Hojda, M. (2025) Using Large Languge Models for Processing Sensor Data, Sens., 25(14), p.4380.

10.3390/s2514438040732510PMC12299101

Migilinskas, D., Popov, V., Juocevicius, V., Ustinovichius, L. (2013) The Benefits, Obstacles and Problems of Practical BIM Implementation, Procedia Eng., 57, pp.767~774.

10.1016/j.proeng.2013.04.097

Mior, M.J. (2024) Large Language Models for JSON Schema Discovery, arXiv, p.2407.03286.

Ramaji, I.J., Memari, A.M. (2018) Interpretation of Structural Analytical Models from the Coordination View in Building Information Models, Autom. Constr., 90, pp.117~133.

10.1016/j.autcon.2018.02.025

Tapeh, A.T.G., Naser, M.Z. (2023) Artificial Intelligence, Machine Learning, and Deep Learning in Structural Engineering: A Scientometrics Review of Trends and Best Practices, Arch. Comput. Methods Eng., 30(1), pp.115~159.

10.1007/s11831-022-09793-w

Tian, C., Zhang, Y. (2024) Optimizing Collaboration of Llm based Agents for Finite Element Analysis, arXiv preprint arXiv:2408.13406.

10.2139/ssrn.4972059

Wedholm, W. (2024) Exploring the Influence of Data Formats on the Consistency of Large Language Models Outputs, Bachelor's thesis, Mälardalen University, Västerås, Sweden.

Wei, B. (2024) Requirements are All You Need: From Requirements to Code with Llms, In 2024 IEEE 32nd International Requirements Engineering Conference(RE) pp.416~422.

10.1109/RE59067.2024.00049

Yang, J., Jiang, D., He, L., Siu, S., Zhang, Y., Liao, D., Li, Z., Zeng, H., Jia, Y., Wang, H., Schneider, B., Ruan, C., Ma, W., Lyu, Z., Wang, Y., Lu, Y., Do, Q.D., Jiang, Z., Nie, P., Chen, W. (2025) StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs, arXiv preprint arXiv:2505.20139.

Yusuf, A.O. (2020) Building Information Modeling (BIM): A Necessity for the 21st Century Civil Engineer in Nigeria. In Being a Paper Presented at the 2nd International Civil Engineering Conference (ICEC2020). Department of Civil Engineering, Federal University of Technology, Minna, Nigeria.

10.26634/jmt.6.1.15702

Information

Publisher :Computational Structural Engineering Institute of Korea
Publisher(Ko) :한국전산구조공학회
Journal Title :Journal of the Computational Structural Engineering Institute of Korea
Journal Title(Ko) :한국전산구조공학회 논문집
Volume : 39
No :2
Pages :81-88
Received Date : 2025-12-22
Revised Date : 2026-01-16
Accepted Date : 2026-01-19
DOI :https://doi.org/10.7734/COSEIK.2026.39.2.81

Journal of the Computational Structural Engineering Institute of Korea ISSN:1229-3059(Print) 2287-2302(Online) 한국전산구조공학회 논문집

All Issue