| Title |
Visual Risk Analysis of Formwork Operations Using Large Language Models and Four-Dimensional Simulation |
| Authors |
박희지(Park, Hee-Ji) ; 김상용(Kim, Sang-Yong) |
| DOI |
https://doi.org/10.5659/JAIK.2026.42.10.417 |
| Keywords |
Construction Safety Management; Large Language Model (LLM); 4D BIM; Data-Driven Risk Analysis |
| Abstract |
This study proposes a methodology for analyzing and visualizing process-specific formwork risk factors at construction sites using large
language model-based data analysis and four-dimensional building information modeling simulation. Accident case data collected through
public application programming interfaces were analyzed using large language models, morphological analysis, term frequency?inverse
document frequency vectorization, and k-means clustering to identify major risk factors and similar risk groups. The identified risk factors
were visualized as process-specific risk levels in a Revit model using Dynamo Script, while four-dimensional simulation linked to the
construction schedule was implemented through Navisworks Timeliner. Process-specific risk factors and accident causes were also displayed
dynamically in the Revit property window to support intuitive risk assessment. The proposed system integrates large language model-based
analysis with four-dimensional building information modeling simulation to enable data-driven prediction and visualization of process-specific
risk factors, with the potential for application across a wide range of construction processes and projects. |