New method of detecting combustible dust uses real-time imaging

West Lafayette, IN — Using newly developed algorithms, researchers from Purdue University have designed an image- and video-based application to detect combustible dust concentrations suspended in the air.

The application, which the researchers say can be used in agricultural, powder-handling or manufacturing settings, involves capturing images of a suspended dust cloud and then analyzing the light extinction coefficient. In testing, the algorithm was able to recognize 95% of sawdust and 93% of cornstarch particulates in the air, a university press release states, adding that the application was able to distinguish suspended dust from “normal background noise.”

Current technology for detecting dust levels can be expensive and difficult to install in workspaces, and separates dust matter into multiple filters that must be weighed and undergo additional analysis, according to the researchers. In contrast, the new application doesn’t require extended training, is location independent and doesn’t need to be permanently installed.

According to data from the Chemical Safety Board, between 2006 and 2017, 111 combustible dust incidents resulted in 66 worker deaths and 337 injuries in the United States.

“Determining suspended dust concentration allows employers to take appropriate safety measures before any location within the industry forms into an explosive atmosphere,” Kingsly Ambrose, an associate professor of agricultural and biological engineering, said in the release.

The research team, which has worked with the Purdue Research Foundation Office of Technology Commercialization to patent the technology, said the application can be used effectively in open and confined spaces.

The study was published online July 21 in the Journal of Loss Prevention in the Process Industries.