An Ohio State Spinout is Harnessing AI to Predict Utility Damage from Weather Events

When a storm hits and the power goes out, utility workers are sent in as quickly as possible to repair the damage and restore power to the people affected. But what if there was a way to anticipate the location most affected by storm damage and the gravity of the damage itself? Columbus startup StormImpact is using artificial intelligence developed at Ohio State to provide insights on disruptive weather, helping utility companies optimize their crew placement and response time to restore power more efficiently.

“StormImpact is an Ohio State University spinoff, licensing technology that was developed at the university,” Co-founder and Chief Scientist Steven Quiring said. “Our focus is on harnessing data science and machine learning, along with weather forecasts, to develop predictive models for electrical utilities. Whether it’s a winter or summer storm, hurricane, or anything in between, giving utilities advanced notice about where power may be disrupted allows them to rapidly and efficiently respond and get power back on for their customers.”

A variety of data informs that predictive model, such as locations of poles and transformers, the age and materials of those assets, vegetation growth, topography, as well as past weather conditions and forecasts. Those environmental and meteorological variables are stored in a database and funneled into an algorithm that helps the startup make storm damage predictions. But StormImpact doesn’t only use historical data, its database is updated in real time to be as accurate as possible.

“We score the performance of our models everyday for all weather conditions,” Quiring said. “Whether it’s a blue sky day where there’s not much going on, a storm, or moderate weather, we report that to the electrical utilities every month. Then we retain our models, taking that most recent month of information, and adding that to our training data set so our models get better over time. One of the advantages of using AI is that it learns from what’s currently going on.”

So how does this predictive information help utility and electrical companies better service their customers? When a storm occurs, utility organizations bring in restoration crews to do repairs that restore power to affected areas, but have a limited amount of crew personnel on staff. The startup’s predictions help utilities anticipate how many responders the storm will require so they can work with other organizations and contractors to gather more people and get them into position to respond quickly to power outages.

“In the days leading up to a storm, they’re trying to predict how bad it’s going to be, how many people they need to bring, and they’re making those decisions before the damage occurs,” he said. “If they’re over prepared, that means they’re spending more money than they need to, but if they under prepare state regulators and customers get upset and they lose revenue because, if the power isn’t on, they’re not making money. It’s an optimization problem we’re trying to solve.”

StormImpact’s technology originated at Ohio State, but Ohio Third Frontier partner Rev1 Ventures has been instrumental in coaching the startup through the entrepreneurial side of the business, connecting it with resources in the Columbus area.

“Rev1 has been a great champion and mentor for us,” Quiring said. “They were absolutely essential because I know a lot about atmospheric science, technology, and machine learning, but not as much about business. They provided us with guidance and connections, legal and accounting expertise, order of operations, and helped us think about how to build and grow the company. They’ve been supporting us along the way, so we’re continuing to work with them and we’re very grateful for that support.”

Exciting things are on the horizon for the startup. In September 2022, it applied for a Small Business and Innovation Research grant through NASA to use satellite earth observations to develop new technology for quantifying vegetation growth near power grids and electrical utilities. Now, the company has applied for phase two funding to take that research to the next level, something Quiring encourages for any small business or startup looking to take their tech to market.

“Trimming trees in and around power lines consumes millions, sometimes hundreds of millions, of dollars for utilities every year,” he said. “It’s one of their largest ongoing expenses, so they’re looking to do that in a more efficient and optimal way. The support from NASA is helping us do that. I strongly encourage any small business or startup to look at these sources of funding as a potential way to build and develop new products — it’s a great source of support.”