Many efforts are being made to enhance user productivity and promote collaboration while ensuring the economic efficiency of office buildings. Analyzing space utilization, indicating how users utilize spaces, has been a crucial factor in these efforts. Appropriate space utilization enhances building maintenance and space layout design, reducing unnecessary energy waste and under-occupied spaces. Recognizing the importance of space utilization, there have been several studies to predict space utilization using information about users, activities, and spaces. These studies suggested an ontology of the information and implemented automated activity-space mapping as part of space utilization prediction. Despite the existing studies, there remains a gap in integrating space utilization prediction with automated space layout design. As a foundational study to bridge this gap, our study proposes a novel methodology that automatically generates office rooms based on space utilization optimization. This methodology consists of three modules: Activity-space mapping, Space utilization calculation, and Room generation. The first two modules use data on space types and user activity types as input to calculate and optimize space utilization through requirement-based activity-space mapping. After optimizing the space utilization value within an appropriate range, the number and area of each space type are determined. The Room generation module then automatically generates rooms with optimized areas and numbers. The practical application of the developed methodology is demonstrated, highlighting its effectiveness in fabricated case scenario. By automatically generating rooms with optimal space utilization, our methodology shows potential for expanding to automated generation of optimized space layout design based on space utilization.

This study aims to simplify the rebar quantity estimation process during the detailed design stage of buildings by addressing identified issues in the existing rebar quantity estimation process and establishing a BIM-based simplification method. To validate the applicability of the proposed method, it was applied to a case study of 'K' High School, with the results compared to rebar quantities from shop drawings to assess usability in real projects. It showed an error rate of 1.4% for SHD22S and 2.63% for HD10 among the types of rebar used in the case model. SHD22S, used as the main rebar in the columns, exhibited this error due to differences in the calculation method for splice and anchorage lengths. HD10, used as stirrup rebar in the columns, showed errors due to the omission of hook length considerations in the simplified estimation formula. Although these discrepancies resulted in cumulative errors of 10 to 20 meters, the error rates of 1.4% and 2.63% respectively fall within the generally accepted rebar overage rate of 5%, suggesting that this method is sufficiently accurate for quantity estimation during the detailed design stage. This study presents a method for efficiently estimating rebar quantities based on BIM during the detailed design stage. While traditional 2D quantity estimation and BIM-based methods often suffer from inefficiencies due to repetitive tasks and high modeling complexity, the proposed method assumes BIM models at LOD200, which are less complex and automate most of the structural information required for rebar quantity estimation, thereby reducing repetitive tasks. However, to further minimize errors, especially those related to hook and anchorage length calculations, additional research is needed to refine the estimation process and expand its applicability to the entire building's rebar quantity estimation.

Recently, LLM (Large Language Model), a rapidly developing generative AI technology, is receiving much attention in the smart construction field. This study proposes a methodology for implementing an knowledge expert system by linking BIM (Building Information Modeling), which supports data hub functions in the smart construction domain with LLM. In order to effectively utilize LLM in a BIM expert system, excessive model learning costs, BIM big data processing, and hallucination problems must be solved. This study proposes an LLM-based BIM expert system architecture that considers these problems. This study focuses on the RAG (Retrieval-Augmented Generation) document generation method and search algorithm for effective BIM data retrieval, with the goal of implementing an LLM-based BIM expert system within a small GPU resource. For performance comparison and analysis, a prototype of the designed system is developed, and implications to be considered when developing an LLM-based BIM expert system are derived.

This study proposes a BIM-based space model generation framework for managing maintenance history in apartments. While BIM adoption in design and construction phases is increasing, its use in operation and maintenance remains limited, particularly in the continuous utilization of as-built BIM models. To address this, the study introduces a methodology for automated space model generation using DYNAMO, analyzing 1,875 maintenance records and long-term repair plans from two apartment projects over two years. The framework was applied to case projects, achieving a 100% space model creation rate for dedicated spaces and 108.91% for public spaces, proving its applicability. This suggests that the automated space model generation process proposed in this study is applicable. The significance of this study is that it presents a model transition framework for the operation and maintenance phase of the as-built BIM model. In addition, it is of research value that a simplified BIM space model is proposed to overcome the limitation of not being able to manage the history of each part in the current maintenance phase. This study can be utilized as a basis for future BIM guidelines applicable to operation and maintenance phases and delivery guidelines for as-built BIM models.