In 2005, a semi-truck caught the nation’s attention when it crashed and caught fire, igniting 35,000 pounds of explosives it was carrying through Utah’s Spanish Fork Canyon.
Thanks to a brief delay between the truck’s crash and the subsequent explosion, there were no fatalities. But, as evidenced by a number of injuries and a 30-by-70-foot crater taken out of the highway, the results can be crippling.
To shed light on the mechanism that caused the chain reaction and help prevent future occurrences, Professor Martin Berzins and his research team from the University of Utah turned to the Argonne Leadership Computing Facility’s 10-petaflop IBM Blue Gene/Q system, Mira. Their research was the subject of a feature article by Jim Collins on the ALCF website.
In the case of the Utah highway incident, the 8,400 cylinders of explosives in transport should have burned away more slowly in an accidental fire, through a process called deflagration. Instead the cylinders detonated, combusting at supersonic speeds and generating a shockwave that blew out the windows of nearby cars.
The research project is using INCITE funding to recreate the detonation virtually. Getting the simulation to reach the desired state has proved particularly challenging due to the incorporation of multiple spatial and temporal scales, but the team’s perseverance has paid off.
“We set out to simulate one-eighth of the actual semi-truck with the explosives in their original packing configuration, but it was not an easy feat,” says Jacqueline Beckvermit, a PhD student at the University of Utah. “After two years of work and more than 100 million computing hours, we finally reached detonation this fall.”
Based on current simulations, the team has identified two possible scenarios that could have led to the explosion: one involving a high-pressure environment caused by trapped gases from the cylinders, and a similar high-pressure scenario caused by the impact of exploding cylinders.
Optimizing and scaling their Uintah Computational Framework to harness a large number of Mira’s cores was key to the group’s success and plans are in place to scale even higher in the future.
Berzins says their ultimate goal will be to enable strategies to prevent similar accidents from occurring in the future.