Extreme weather and changing patterns of electricity use have led to blackouts and unpredictable utility bills across the nation.

To address that unpredictability, Cornell researchers are piloting a new payment plan and an app that would provide consumers with more information about their energy use and incentives to reduce use, while also allowing utility companies to respond more nimbly in times of crisis. Avangrid, the parent company of New York State Electric and Gas (NYSEG), is an initial partner.

“We have all been consuming electricity since we were born, but we don’t know how we consume electricity,” said Madhur Srivastava, assistant research professor in chemistry and chemical biology in the College of Arts and Sciences. “We have no understanding of how much electricity each appliance uses, for example, or each light bulb, or how the costs change for a particular time of use. How can we change our behavior if we have no analysis of it?”

Srivastava was drawn to the issue when he noticed how his electricity bills fluctuated each month, without knowing why or having a clear way to control or predict the monthly bill. In interviews with community members, his team found that this variability was widespread and sometimes caused hardship, especially for small businesses operating on narrow profit margins.

As a result, the pilot run by Srivastava’s group (Signal Science Lab) will offer a tiered, subscription-based pricing plan to participants, beginning in late-January. Around 200 NYSEG customers have enrolled, but customers of any utility company in the U.S. can register.

Instead of a variable bill every month, participants will pay a flat rate, choosing from subscription price tiers generated from prior use and essentially committing to a pattern of consumption or a goal for consumption. The app will then provide information to keep participants on track and alert them to possible savings.

Srivastava and his team will adjust their pricing projections as needed and monitor how customers respond to suggestions.

“If you use a washer and dryer after 7 p.m., you’re going to be charged less than if you use it during the daytime, but people do not know that,” said Shikhar Prakash, M.S. ’19, a doctoral student in systems. “If we can provide that information to the customer in an efficient way, even in real time, as they are using their dryer, it provides an incentive to reduce their electricity bill. Currently, there is no platform for the utility companies to reach out to specific segments of customers in this way.”

Srivastava and his students hope the new payment plan and app will help consumers moderate their electricity use, reduce strain on the electrical grid, and allay the need for more expensive infrastructure, which raises costs for consumers. They also hope the project lays the groundwork to impact energy policy both in New York state and nationwide, especially in the western U.S., where summer blackouts have become common. Srivastava is currently in talks with multiple utility companies from across the country – all facing similar challenges – to formally expand the pilot.

Connecting consumers and providers

To design the pricing plan and allow for targeted messaging to consumers, Srivastava and his team had to figure out how to segment customers with different patterns of electricity use – a challenge requiring advanced data analytics.

“We mapped the census data with the electricity consumption data, with population density data, and were able to devise a machine learning algorithm to segregate customers based on their behavior,” Prakash said. “This allows the utility company to reach out to the specific segment of customers – it creates a two-way communication channel between the utility companies and the consumers where they can build trust.”

The ability to communicate with targeted groups could help in emergency situations, Srivastava said. If there’s too much demand on the grid, utility companies could send targeted messages to high-use consumers, or those with a high degree of variability in their electricity use.

“In the West, blackouts are happening every summer and there’s no way of communicating with the consumers,” Prakash said. “Utility companies have to broadcast messages to the entire state, and they don’t know whom to ask to reduce blackout risk.”

James Rincon, manager and corporate innovation lead at Avangrid, said the project is building the foundation for a more modern relationship between consumers and their energy providers.

“In a lot of areas of our lives, we have apps that both simplify the way we interact with companies and also open up a communication channel,” he said. “In our case, this will allow us to help people reach their goals, whether that’s prioritizing renewable energy, or energy efficiency or cost. That will not only help the consumer but will help us operate the grid in a way that makes it safer and more reliable and resilient for everyone.”

This open communication will become even more important, Rincon said, as patterns of energy use continue to shift – with the purchase of more electric vehicles, for example – and as the sources of electricity become more diverse, with consumers adding solar panels, or additional generators or batteries to their homes.

The app, and the data it processes, would also help utility companies plan infrastructure upgrades and investments.

“Consumption patterns are changing, and deciding where to put a new substation, which substations to upgrade – those choices become difficult and can only be solved through advanced data analytics,” Srivastava said. “All the data should be talking to each other in a synchronized way to make sure the consumer gets their good price, the substation is not overloaded, and the companies know which are the critical infrastructure upgrades and which can be delayed.”

A community-driven approach

Srivastava approached the problem of fluctuating energy bills like he does all his research – with the community in mind.

An Engaged Opportunity Grant from the David M. Einhorn Center for Community Engagement jumpstarted the project. Prakash and Srivastava were then able to fully define the challenges the community faces through the National Science Foundation’s Innovation Corps program, a six-week training program that also emphasizes community engagement and using research, technology and innovation to benefit the public.

 As part of the NSF program, Srivastava and Prakash interviewed more than 250 Avangrid customers and industry employees to better understand their pain points and their needs.

The interviews provided key insights – namely how small businesses, hotels, apartment buildings and single-family homes with varying incomes all had different patterns of use and associated challenges. The information gathered underlined the need for an algorithm that could segment and classify customers, which Prakash and Shrivastava developed with the help of then-undergraduates William Bekerman ’22 and Leah Woldemariam ’22, now a doctoral student in electrical engineering who continues to work on the project.

“We have to ask how people want the solution to be and then use science to develop that solution, rather than develop a solution and ask people to change because science said so – that never happens,” Srivastava said. “We need to understand the exact needs of the community and the stakeholders.”

For Avangrid, Rincon said, the partnership with Cornell is essential to moving innovation forward. The company named Srivastava a Faculty Fellow to support the project’s design.

“There’s a lot of inherent risk in innovation,” Rincon said, “so to have a framework to approach a new idea, to have the talent of one the most prestigious and high-performing universities and a researcher like Madhur engage our customers and convert the insights into actionable innovation is extremely valuable, a win-win.”

The opportunity to solve complex problems that benefit society is what drew Prakash back to Cornell to pursue his Ph.D.; when he was a master’s student at Cornell, his team won the Avangrid Innovation Challenge, and he’d taken a job at Avangrid once he completed his degree. In meetings with Srivastava as an employee, he’d seen that the problems they faced required fundamental, innovative research – research he was eager to pursue.

“This is what I like,” Prakash said, “the kind of study that is built from fundamentals, but we’ll go out in the public and provide a direct benefit.”

Srivastava said he loves data science for this reason: because it can easily impact the community and apply directly to so many areas of life.

“I’ve seen this with my most engaged students – they see a bigger picture and every aspect of their lives is an open question,” he said. “This also gives them ownership of the problems they’re trying to solve – because they see themselves as a part of the community they’re helping. I can’t take credit for that, but it makes me proud.”