As winter storms and summer heat waves increasingly stress the nation’s power grids, Stevens researchers have developed a new way to identify the homes most vulnerable to blackouts — without even visiting them.
The timing couldn’t be more critical. With more than a quarter of U.S. homes already fully electric, and solar installations set to triple during the next five years, understanding vulnerabilities has become critical for emergency planning and public safety.
“We’re racing toward electrification to combat climate change, but we must also understand the risks involved,” says Stevens professor Philip Odonkor, who led the research project. “So, what happens to these solar and electric homes when the power goes out?”
Summer strength, winter blues
Odonkor, with recent graduates and AI summer fellows Andrew Majowicz M.Eng. ’24 and Chetan Popli M.S. ’24, set out to answer that question.
In a new study published in the Journal of Smart Cities and Society, they explore the future of electrified American homes by leveraging AI and analyzing Department of Energy (DOE) building-stock data.
The team dug deep into the energy patterns of 129,000 single-family homes across eight states. Their goal? Uncover the hidden energy “signatures” that distinguish fully electrified homes — those powered entirely by electricity — from those that use a mix of energy sources.
They didn’t stop there, however. For identified mixed-energy homes, the team also worked to pinpoint exactly which appliances have made the shifts to electric power and which haven’t.
After processing and analyzing the dataset, Odonkor’s team found that homes’ energy signatures were not only distinguishable, but they also granted critical insights into the resilience of individual homes.
Solar-powered homes, for example, demonstrated impressive resilience during summer heat waves. However, they proved remarkably vulnerable during winter storms; in fact, fully electrified homes were nearly three times more vulnerable to winter outages, compared to those drawing power from mixed energy sources.
“Think about Texas in 2021, when millions lost power during a winter storm,” Odonkor explains. “As more homes go fully electric, we need to prepare for these scenarios.” “Solar panels help in summer, but they can’t meet the intense heating demands that occur during winter blackouts.”
New methods to inform planning and response
The study wasn’t only pathbreaking for its findings; it was only notable for the innovative AI-powered methods that were used to conduct the analyses.
Odonkor’s team developed novel machine-learning models capable of identifying an individual home’s energy systems and vulnerabilities with over 95% accuracy, using only its energy-consumption patterns. The new approach enables utilities and emergency responders to pinpoint at-risk households across entire neighborhoods, without the need for invasive surveys or inspections.
“Until now, we actually had to go door-to-door to determine if a home was fully electric,” notes Odonkor. “Now, we can automatically identify the most vulnerable homes while still safeguarding people’s privacy.” “This will shift the way we prepare for and respond to extreme weather, enabling faster, and more targeted action when it’s needed most.”
The study’s potential benefits extend beyond empowering individual homeowners. As cities work to build climate resilience, these new tools could help community emergency-service units prioritize responses during outages. It could also assist urban planners in the long-term development of more resilient housing stock and neighborhoods.
That’s key, because communities nationwide are grappling with a one-two punch of aging power grids subjected to more frequent episodes of severe weather.
As we increasingly transition to electric homes to cope with climate change, the team’s findings serve as a warning that we will need implement strategies that protect vulnerable solar and electric households during winter emergencies.
“The path to sustainable cities isn’t just about going green; it’s about staying resilient,” he emphasizes. “As we shape the future of urban housing, understanding vulnerabilities isn’t just a luxury — it’s essential to keeping communities safe.”
