AI could save future firefighters from deadly explosions • The Register

According to new research, AI could help save firefighters’ lives by predicting flashovers before they happen released in this week.

Flashovers occur when combustible material in a room suddenly begins to ignite all at once, resulting in a tremendous heat blast and combustible gases that can rupture walls and windows. Over a 10-year period, from 2008 to 2018, around 800 firefighters were killed and more than 320,000 injured in the United States, and it is estimated that 13 percent of these accidents are due to flashover events.

Firefighters must rely on their experience to predict when a flashover is imminent, such as from smoke and heat levels, but it’s not easy given how quickly they can sneak up. Computer scientists have spent the last two decades trying to develop methods to detect flashovers in real time, but modeling something so unpredictable is a difficult task.

Researchers at the US government’s National Institute of Standards and Technology (NIST), Google, Hong Kong Polytechnic University and China University of Petroleum have developed a system that uses graph neural networks (GNN) to determine relationships between different Learn data sources, represented as nodes and edges, from simulated fires.

“GNNs are commonly used for estimated time of arrival or ETA in traffic, where you can analyze 10 to 50 different roads.” Eugene Yujun Fu, co-first author of the study and research assistant professor at Hong Kong Polytechnic University, said in an opinion.

“It is very complicated to use this kind of information correctly at the same time, so we came up with the idea of ​​using GNNs. Aside from our application, we consider spaces instead of streets and predict flashover events instead of ETA in traffic.”

The team simulated everything from building footprints, surface materials, fire conditions, ventilation configurations, smoke detector locations and room temperature profiles to modeling 41,000 mock fires in 17 different building types. A total of 25,000 fire cases were used to train the model and the remaining 16,000 were used to refine and test it.

The performance of the GNN was judged on its ability to predict if a flashover event would occur within the next 30 seconds. Initial results showed that the model had an accuracy of 92.1 percent at best.

Called FlashNet, the system is more advanced than the team’s previous machine learning model P flash.

“Our previous model only had to account for four or five rooms in a layout, but when the layout switches and you have 13 or 14 rooms, it can be a nightmare for the model,” said Wai Cheong Tam, the article’s co-first author and mechanical engineer at NIST. “For real-world application, we believe the key is to move to a generalized model that works for many different buildings.”

FlashNet may appear promising, but has yet to be tested with data from real fire departments. To do this, the model would need to analyze data from thermostats, carbon monoxide and smoke detectors in smart homes, Tam explained The registry. How firefighters could then be made aware of the model’s predictions is unclear.

“The focus of the research was to rely on building data that is provided, or could easily be provided, by available building sensors. One way to turn the research into reality is to integrate the model with a smart fire control panel that collects the temperature data from installed heat detectors and includes a computer module that can process the data and make real-time predictions.”

“From the fire alarm panel or other appropriate device, the prediction would be sent to the incident commander or individual firefighters when deemed appropriate. The exact mechanism for providing such predictive analytics is not specified and would require input from the fire department to develop consensus,” concluded Tam. ® AI could save future firefighters from deadly explosions • The Register

Laura Coffey

World Time Todays is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – The content will be deleted within 24 hours.

Related Articles

Back to top button