ChatGPT를 활용한 BIM 데이터 평가 기준 실시간 변경 대응 시스템 개발 연구 = Development of BIM Data Evaluation Criteria Real-time Change Response System Using ChatGPT

다국어 초록 (Multilingual Abstract)

This research presents a methodology for enhancing automation in design certification evaluation systems by integrating Building Information Modeling (BIM) with OpenAI’s generative AI, ChatGPT. Traditional BIM-based certification systems require frequent manual revisions whenever certification standards change, relying heavily on programming specialists. This leads to inefficiencies and delays in the certification process. To overcome these challenges, this research proposes a real-time interactive system that enables users to modify evaluation criteria directly through a chat-based natural language interface. ChatGPT interprets user instructions and dynamically updates the evaluation logic. The system specifically targets the Barrier-Free (BF) certification, focusing on the clear width criteria applied to BIM data. By connecting ChatGPT with the AIBIM Chat system, users without programming knowledge can interactively modify the evaluation logic. The system extracts object attributes such as IfcDoor parameters from BIM data, applies the modified criteria, and ensures consistency through automated verification. This process streamlines the certification workflow by reducing manual effort and minimizing errors. A case study was conducted to validate the system's accuracy, efficiency, and adaptability. Results confirmed that the system dynamically handles both strengthened and relaxed evaluation criteria, automatically reflects them in BIM data, and provides accurate certification outcomes. This study highlights the potential of AI-driven automation to improve the accessibility, reliability, and efficiency of design certification, offering a practical solution for non-expert users.