Automated quantification of construction workers’ exposure to falling object hazards

Abstract

The construction industry is among the most hazardous industries, and its continuously changing and complex environment results in a labour-intensive task of planning and preventing hazards. Current manual safety planning procedures cannot keep up with construction progress. This leads to unplanned durations and an increased responsibility for the individual workers to analyse and act accordingly to an emerging situation. This work proposes an automated approach to identifying and measuring the amount of struck-by falling object hazard exposure to construction tasks and their assigned work crews. Additionally, falling objects can originate from activities not foreseen or planned due to planning resolution (e.g. crane lift paths or temporarily impassable access routes). Therefore, it is investigated how to extend the current practices of safety analysis in both the planning and construction stages using building information modelling artificial intelligence and sensor techniques. The proposed strategy is to identify hazard sources and subjects based on their topology and nature in a spatio-temporal analysis. The proposed combinatorial analysis approach is validated in a case study performed on a real construction project in Finland. It yields new insights, which can be necessary for construction sequence decisions and convincing workers to improve their safety behaviour.