New York State Electric & Gas says it is developing what it is calling an "outage prediction model," essentially a software program that will use machine learning or artificial intelligence — AI — to predict outages during storm events.
NYSEG and its parent company, Avangrid, along with its sister utility, RG&E, short for Rochester Gas & Electric, are working with researchers at the University at Albany and the University of Connecticut on developing the AI system.
The system will use AI to analyze weather forecasts to predict — or guess — which parts of the electrical grid will be hit hardest by storms. That way the utility can prepare to deploy resources to those areas in advance. Such machine learning gets better over time as the system "learns" from past events and data.
Today, without such a system in place, utilities rely on sensors, customers and time-consuming field surveys to figure out the cause and location of outages — efforts that can take days to complete until linemen are sent in to make repairs.
"We're excited to partner in the creation and piloting of this innovative prediction model that we expect will strengthen our storm response efforts with the potential to shorten durations of outages for the benefit of our customers," Trish Nilsen, Avangrid's director of emergency preparedness said.
Also participating in the project is the New York State Mesonet, a network of 126 weather stations across the state that are run by UAlbany and were developed conjunction with the state Department of Homeland Security and Emergency Services.
UAlbany hosts the National Weather Service's local office and the school also teaches atmospheric sciences and is home to top weather researchers at the university's Atmospheric Sciences Research Center, or ASRC.
"Over the last two decades, New York has experienced increasingly extreme weather that poses a risk for extended power outages and significant economic losses," Chris Thorncroft, director of UAlbany's Atmospheric Sciences Research Center said. "Our scientists at ASRC are focused on creating smart solutions for weather risk management decisions."
The new AI outage prediction system received $600,000 in funding from the New York State Energy Research and Development Authority, a semi-autonomous state agency that funds renewable energy and so-called smart grid technologies that help the state's electrical grid run more efficiently and more resiliently.
And it's not just where the weather is happening that determines outages. There are many variables such as the strength of power lines and the wooden poles that hold them up — as well as the soil the poles are dug into. Trees and other vegetation also tie into the formulas required to predict storm damage, which can cost a utility tens of millions of dollars during the worst storms and lead to millions of dollars in business losses during outages.
NYSEG serves parts of Saratoga, Albany, Washington, Columbia and Greene counties, and combined, NYSEG and RG&E serve 40 percent of the state, with a total of 1.28 million customers.
Avangrid is based in Orange, Conn., but is owned by a Spanish holding company called Iberdola Group. National Grid's ownership is similar, with its parent company based in the United Kingdom.
©2021 Times Union, Distributed by Tribune Content Agency, LLC.
