UConn, power company join to predict storms’ impact

University created model that accurately forecast Sandy damage, outages

Storrs — Two days before Superstorm Sandy struck Connecticut in 2012, UConn scientists fed measurements from the storm and historical data from others into a computer model they developed and forecast where the most damage and power outages would occur.

Their forecast was very accurate in predicting both the scope and the location of the outages: The model predicted 13,500 damage locations; the storm created about 15,000.

Now, Eversource, the state’s largest electric utility, and the university have teamed up to create a center devoted to studying ways to better predict and prepare the state’s power infrastructure for natural disasters.

The Eversource Energy Center at the University of Connecticut officially opened last month after getting approval from the school’s board of trustees. It includes faculty and graduate students in environmental, civil, electrical and structural engineering, computer science and forest management.

In addition to storm forecasting, the center’s scientists will study the resiliency of the state’s electric infrastructure and the trees that grow near power lines.

Eversource has committed $9 million over five years to fund the center, an investment the utility believes will pay off.

“If we can shave hours off small events and days off large events by prestaging the right amount of re-

sources, by making our vegetation management and infrastructure hardening programs better and more cost effective, I think it could pay back several times over,” said Ken Bowes, Eversource’s vice president of engineering.

Center director Emmanouil Anagnostou, an environmental engineering professor, says the center also will eventually include the university’s business school, which will help Eversource with risk analysis, deciding where to spend money on reinforcing the system and creating an optimal storm response plan.

There are other power companies working with scientists at schools across the nation on vegetation or storm forecasting issues, Anagnostou said. But the UConn center is the first to combine multiple disciplines to attack the problems in one place.

He said they hope the center will help attract students and grant money, and give current graduate students some real-world experience that will help them land jobs.

The center will refine the school’s storm forecasting model with data from each subsequent event, making it easier to predict where winds, rain, tides, snow and ice are likely to present the most problems, Anagnostou said.

The school put the power-outage forecast model to work last month for a relatively small coastal storm.

“It predicted a midrange of about 350 trouble spots and we ended up with about 400 trouble spots,” Bowes said. “So, it’s pretty accurate.”

For now, the storm prediction model is being shared only between the school and Eversource. But as it becomes refined the center may decide to share its predictions with the public to alert home and business owners in areas projected to be hard hit, officials said.

Scientists at the center also are using laser technology to create 3D maps of the state’s electric infrastructure and surrounding vegetation that can show the size of trees and where they are relative to power lines. It also will show whether power poles are leaning and may be vulnerable.

And natural resources scientists are putting computer sensors on trees that measure out how the different species react to wind, rain and ice — whether standing alone or in a forest.

“About 90 percent of our electrical outages are related to down trees or tree limbs breaking,” said Bowes. “And we’re hopeful that with this research we’ll be able to prune and trim the right trees, not all the trees.”