发表机构
Tsinghua University; Key Laboratory of Digital Construction and Digital Twin, Ministry of Housing and Urban-Rural Development; PetroChina Planning & Engineering Institute（清华大学; 住房和城乡建设部数字建造与数字孪生重点实验室; 中国石油规划总院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出结合领域专用LLM、BIM转文本方法及幻觉控制策略的集成框架，实现BIM设计缺陷的智能识别与修复，识别准确率达85%，合理修复建议生成率94%，构建了端到端原型。
AI 中文摘要
现有方法缺乏通用方案来高效识别和解决BIM中多样的设计缺陷。因此，本研究提出一种集成框架，通过领域专用大语言模型（LLM）识别和修复BIM中的各类缺陷。首先，引入带组件平衡分块的BIM转文本方法，以衔接BIM数据与大语言模型；随后，提出结合规则注入、少样本提示和检索增强生成（RAG）的提示学习方法，用于识别缺陷并生成修复建议；同时，引入结合关键标识符验证与标记长度阈值的幻觉控制策略，以确保可靠性。实验表明，能力扩展后的识别准确率达85%，而传统规则检查的准确率为70%，合理修复建议的生成率达94%；此外，所提幻觉控制策略进一步将准确率从64%提升至85%，在单次干预回合中消除了92.5%的幻觉。本研究建立了从原始BIM数据输入、缺陷识别到修复建议生成的端到端原型。
英文摘要
Existing methods lack a generalized approach to efficiently identify and resolve the diversity of design defects in BIM. Therefore, this study proposes an integrated framework to identify and repair various defects in BIM via domain-specific LLMs. Firstly, a BIM-to-Text method with component-balanced chunking is introduced to bridge BIM data with LLMs. Then, prompt learning with rule injection, few-shot prompting and RAG is proposed to identify defects and generate repair suggestions. Meanwhile, a hallucination control strategy combining key identifier validation and token-length thresholds is introduced to ensure reliability. Experiments show capability expansion yields 85% identification accuracy versus 70% for traditional rule checking, achieving a 94% rate of reasonable repair suggestions. Moreover, the proposed hallucination control further increased accuracy from 64% to 85%, eliminating 92.5% of hallucinations in a single intervention round. This study establishes an end-to-end prototype from raw BIM data input, through defect identification, to repair suggestion generation.
