The construction of public buildings involves a significant amount of carbon emissions, particularly during the materialization phase of underground construction, which is characterized by a rapid and intense release of carbon emissions. In response to this issue, this study aims to develop a Building Information Modeling (BIM)-based method for assessing the factors influencing carbon emissions during the materialization phase of underground construction in public buildings. To achieve this goal, a process-based quantitative and qualitative method was established to evaluate carbon emissions at various stages, including raw material extraction and processing, building material/equipment production, and on-site processing of components, as well as construction and installation. The proposed method may serve as a valuable tool for selecting design schemes in the early stages of building design. By analyzing 125 actual cases in Xi’an, China and utilizing a carbon emission calculation model, accurate carbon emission data was obtained. This study’s findings indicate a correlation between the embodied carbon emissions and the area of underground space in the cases analyzed. Specifically, larger underground spaces were associated with higher embodied carbon emissions in underground engineering. Furthermore, this study categorized the cases into two groups based on the presence or absence of underground space, resulting in the identification of 16 factors for cases without underground space and 19 factors for cases with underground space. These findings provide valuable insights for understanding and mitigating the impact of carbon emissions during the materialization phase of underground construction in public buildings.

Automated compliance checking for building fire protection design requires accurate extraction of domain-specific entities from technical code provisions. However, building fire protection codes contain highly specialized terminology, hierarchical clause structures, and complex regulatory semantics, which make manual information extraction time-consuming and error-prone. This study develops a domain-specific named entity recognition approach for Chinese building fire protection design codes. An expert-annotated dataset was constructed from five representative codes, containing 2748 annotated provisions and 13,877 entity mentions across nine entity categories. A RoBERTa-BiLSTM-CRF model was then developed to capture contextual semantic representations, bidirectional sequence dependencies, and label-transition constraints. Results suggest that the proposed model achieves 88.34%, 88.31%, and 88.32% for precision, recall, and F1-score, respectively. The extracted entities can be organized into structured and traceable records, providing a foundation for downstream building fire protection knowledge management and automated compliance checking.