Toward a Decision Support System for a Toolbox Meeting Pertaining to Roofing Activities in Construction

According to global accident statistics, construction sites remain the most dangerous of all sectors, with falls from heights being the leading cause of accidents. Current construction safety plans for toolbox meetings to prevent falls rely on paper-based 2D drawings and schedules to instruct workers, comprehend the need for safety equipment installation on a job site, and then mandating regular physical visits for safety plan implementation. Therefore, this paper presents a framework developed to provide a decision support system for a toolbox meeting pertaining to roofing activities in construction. The framework includes three modules: (1) accident data classification (ADC) module, (2) 4D-BIM-based safety information modeling (SIM) module, and (3) model discrepancies detection (MDD) module. The ADC module is used to compute the probability of accidents in roofing activities from historical accident dataset. The 4D-BIM-based SIM module integrates the extracted data of ADC module with 3D-BIM, construction activities, and schedule. The MDD module compares as-is and as-built model of safety guardrails to detect discrepancies with the planned model. The proposed system aids safety managers by extracting specific 4D-BIM-based activities for a given day and predicting associated hazards using historical accident data from the OSHA database. It also incorporates input from the actual job site through an as-built model generated from a point cloud, enabling the identification of changes and discrepancies in the real site before conducting a toolbox meeting. The findings of this study are expected to give important information to support safety manager in comparing the actual job site with the safety plans and provide significant and actual information for a toolbox meeting which could not be assessed with the 2D drawings.