Preventing high-risk work accidents

With construction sites around the world accounting for 60,000 fatalities each year, an Australian university has come up with a hazard analysis tool to reduce work accidents and improve safety.

A research team at the University of South Australia has built what it calls a ‘knowledge graph’ that predicts hazards on construction sites, traditionally identified manually in a time-intensive process.

The urgency to innovate safety measures has never been greater, according to UniSA Construction Management lecturer and leader of the study, Dr Sonali Pandithawatta.

“ 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 Pandithawatta said.

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

To design the algorithm, the researchers gathered data from incident reports and experts, integrating information such as weather, hazards, job steps, and preventative measures all captured on a graph.

They analysed more than 100 JHA documents and consulted 18 industry experts to build the tool which has demonstrated “exceptional accuracy” of 90% according to the research team.

The tool was capable of analysing both primary and secondary hazards, workplace proximity, weather, and atmospheric hazards in real-time. The next step for the team is to assess other risk factors such as human and managerial influences, and to integrate advanced machine learning techniques.

The study (available in pdf) has been published in the Journal of Engineering, Project and Product Management.

About the author

Desi Corbett

Desi is the Editor of Concrete in Australia and at the helm of our magazine for 8 years. She was behind the Institute's weekly news bulletins from 2016-2021 and is now writing our focused news items. Desi has been an engineering news and features journalist/editor across all disciplines since 2013 - part of a 30-year career writing for a wide range of industries.