How can AI help protect our drinking water?
The Animas River before three million gallons of toxic waste spilled from the Gold King Mine near Silverton, Colorado. Photo: Getty Images
Artificial intelligence (AI) is now providing more accurate simulations to better prevent water contamination following toxic chemical spills.
Why you should know: When toxic chemicals spill into rivers, they flow downstream and may pollute the water supply for local communities.
For example: In 2014, a chemical facility in West Virginia leaked 10,000 gallons of 4-methylcyclohexylmethanol, a foaming agent used to wash and process coal, into the Elk River one mile upstream from the largest water treatment plant in the state.
The National Science Foundation called it “one of the largest human-made environmental disasters in this century.”
The Elk River feeds into the Ohio River, which is the foremost water source for Cincinnati.
Hours after the spill, Leidos and the Greater Cincinnati Water Works (GCWW) used a software solution called ICWater to stop the chemical from entering the city’s drinking supply.
ICWater also minimized contamination during the Gold King Mine spill near Silverton, Colorado in 2015 and the tsunami-triggered Fukushima Daiichi Nuclear Power Plant incident in Japan in 2011.
How it works: ICWater calculates the time-of-travel and concentration of contaminants based on a river’s real-time flow and velocity from the National Water Model. Leidos created ICWater under contract for the Defense Threat Reduction Agency.
The latest: Leidos experts Rakesh Bahadur, Bill Samuels and Mike Monteith recently upgraded the ICWater software to provide more accurate data for city officials.
- “AI and machine learning have led to new equations with the potential to enhance these simulations,” Bahadur said. “We experimented with a new equation, which accounts for geometric characteristics of a river winding through natural landscapes. We integrated this equation with our latest software, ran simulations and found the new model yielded more accurate results.”
These advanced simulations can now help utilities predict how many hours intakes should be closed to prevent contamination.
Looking ahead: Samuels and his team are expanding the tool’s river network to include regions outside of the U.S. The additional data will allow ICWater to provide better results during international emergencies.
Please contact the Leidos media relations team for more information.