The research team from the University of South Australia has constructed a “knowledge graph” that predicts construction site hazards.

Construction sites account for approximately 20% of occupational deaths worldwide and Dr Sonali Pandithawatta says the urgency to innovate safety measures has never been more significant.

“Traditionally, potential job hazards rely heavily on safety personnel identifying risks and control measures, a process that is prone to inefficiencies and human error,” Dr Sonali says.

“Our research addresses the critical need to automate job hazard analysis (JHA).”

When designing the initial algorithm, researchers accumulated data from incident reports and experts to integrate data. This data covered weather, hazards, preventative measures and job steps which were captured in a graph.

More than 100 JHA documents were examined and included material from 18 industry experts.

Co-author Professor Rameez Rameezdeen says the model had illustrated “exceptional accuracy” with greater than 90% accuracy. It was also competent in analysing atmospheric hazards, workplace proximity and both primary and secondary hazards in real-time.

The next step for the study is to assess other risk factors such as human and managerial influences. They also want to integrate advanced machine learning techniques for a broader use.

The study published in the Journal of Engineering, Project and Production Management, was supported by the South Australian Water Corporation (SA Water) is available here.