The sun is trying to shine on Rutland, Vermont. It’s a gray morning in April, but a few beams cut the clouds as Nathan Adams ’96 and I turn off Route 4 and head up City Dump Road. We pass two trucks unloading garbage at a transfer station. Then we get out of the car and start walking through the mud toward the top of this now-closed landfill. Amidst piles of melting snow, 7,722 silicon solar panels cover ten acres like so many rows of purple tabletops tipped toward the south.
I’ve come here looking for the smart grid revolution. Whatever that is.
The solar panels were erected on this hill by Vermont’s largest electricity company, Green Mountain Power, where Adams works as vice-president of strategy and innovation. When they’re turned on this summer, the panels will generate 2.5 megawatt hours of electricity, enough to power about 2,000 homes during full sun.
Next to the panels, eight metal storage buildings hold racks of both lead-acid and lithium-ion batteries. The batteries will store four megawatts of energy. The hill slopes gently down to the Rutland High School. In the event of a blackout—say the next Hurricane Sandy—the high school can be “islanded,” Adams tells me, separated from the regional electrical grid. With only solar power and batteries, GMP will be able to keep the lights and heat on at the school “as long as they’re needed,” Adams says.
The U.S. Department of Energy has identified this $10 million project, the Stafford Hill Solar Farm, as the first and only all-solar “microgrid” in the United States—and it goes far beyond providing power to an emergency shelter. In two hundred participating houses in the surrounding neighborhood, the hot-water tanks will soon have sensors that can wirelessly transmit their temperature back to Green Mountain Power. “Our customers will always have hot water,” Adams says, but new, sophisticated algorithms in GMP’s control center will be able to automatically shut down already-hot tanks for short periods, reducing demand when this East Rutland circuit nears peak load.
Ditto for the batteries—but in the other direction: as demand spikes, say on a hot summer afternoon when people flip on air conditioners, GMP will be able to quickly turn on the batteries, pumping power into the grid from Stafford Hill instead of having to buy it from expensive “peaker” plants in the New England wholesale electricity market. It’s solar power being collected all day, then “dispatched” as Adams says, when it’s most needed. “We’re developing very local energy generation and storage to optimize the efficiency of the grid,” Adams says. Sounds pretty smart, maybe even revolutionary.
----
Thomas Edison received a patent for the incandescent light bulb in 1880 and by the end of that decade the first electric grids were providing a one-direction flow of electricity from generators to transmission lines to transformers to light bulbs. Over the next century, the U.S. grid spread in size and complexity such that, in 2000, the National Academy of Engineering named the electric power grid the greatest engineering accomplishment of the twentieth century. But the twentieth century is getting to be a while ago—and the grid is showing its age. “A lot of the technology being used today in the electric grid would not have looked unusual to Edison,” says UVM engineering professor Jeff Marshall, “It’s really old stuff.”
A patchwork product, the grid is increasingly congested as more and more of our economy and culture requires plugging in. Large-scale power failures—like the Northeast blackout of 2003 and the one in India in 2012 that left ten percent of humanity in the dark—highlight ongoing concerns about the reliability of the grid, whether stressed by extreme weather or threatened by terrorist attack. And the burning of fossil fuels to create electricity is a major contributor to climate change; more than a third of U.S. greenhouse gases come from the production of electricity.
Marshall leads a group of twenty-one UVM faculty who participate in the University of Vermont’s Smart Grid IGERT program. Funded by the National Science Foundation, and in partnership with Sandia National Laboratories, the five-year, $3 million program will train nearly two dozen doctoral students in an interdisciplinary cross-section of fields from engineering to psychology—all with an eye toward making the grid, well, smarter.
“My definition of smart grid is the use of information technology to make the grid work better. And by working better I mean it’s more reliable, greener, and more cost-efficient,” says UVM professor of engineering Paul Hines, an expert on power systems. Roughly speaking, in the smart grid, electricity meets the internet, and power and information flow in two directions.
But underneath that tidy definition, Hines is quick to say, lies a set of hugely complex technological and cultural challenges—and a brewing set of fights about whose vision of a smarter grid will get funded. For example, to improve reliability, is it smarter to invest in hardening the grid against storm damage—Hines and colleagues ask in a recent opinion article in the Proceedings of the National Academy of Sciences—or is it wiser to build “distributed generation” that spreads power production out to, say, hundreds of wind farms and thousands of solar rooftops?
“Smart grid has become a catch-all phrase to represent the potential benefits of a revamped and more sophisticated electricity system,” writes Hines’ colleague, Jennie Stephens, in Smart Grid (R)Evolution: Electric Power Struggles, a new book that she co-authored. It’s “a vague, politically attractive, seemingly benign, and somewhat ambiguous phrase,” she writes. “After all, who would argue for a ‘dumb grid?’”
“I don’t even say smart grid anymore,” Nathan Adams tells me. “It’s one of those ubiquitous terms everyone is using and nobody knows what it means.” But he and everyone I talked to in my hunt for the smart grid agree that big changes are coming to how we produce, move, and manage electricity.
“We inevitably have to shift from fossil fuels to renewables,” says Jennie Stephens, who has faculty appointments in both the Rubenstein School of Environment and Natural Resources and the College of Engineering and Mathematical Sciences, “the question is how fast? And how is it going happen?”
Last year, four percent of the U.S. electricity supply came from wind, according to the U.S. Energy Information Administration—and just 0.4 percent from solar. One of the central challenges in moving those percentages up is how to overcome what engineers call “the intermittency problem.” Wind is wonderful, until it stops blowing. Solar is swell, until a cloud passes over. “This causes the energy supply to go from full power to zero very quickly,” Jeff Marshall explains. “There is no storage in a traditional power grid. At any instant, the supply has to be equal to demand.” So, until recently, most utilities have seen these fluctuating power sources as no more than a boutique part of their supply.
To help, Marshall, Hines, and several of their students in the UVM smart grid program are designing better control algorithms for managing wind farms that optimize their power—using on-site battery storage to smooth output. And energy storage may become the most revolutionary part of a smart grid revolution, says Mads Almassalkhi, a new professor in the College of Engineering and Mathematical Sciences who specializes in power systems.
For example, a new generation of electric cars could fundamentally change how we think about where our power is coming from: instead of just transportation, your e-car could become part of the grid—“a mobile energy-storage device,” Almassalkhi says—drawing power when you need it to drive, and pumping juice back into the grid when your house or your utility company need it. Charge it with solar panels in the backyard and you have a micro powerplant sitting in the driveway.
This is where the smart grid becomes necessary. Going from managing the output of a few coal or natural gas power plants to organizing perhaps millions of scattered solar panels and car-sized power nodes becomes an astronomically more complex task. Which is why Almassalkhi, Hines, and others in the UVM smart grid team are exploring the best ways to coordinate vehicle charging. At the same time, Christopher Clement, a doctoral student in natural resources and trainee in UVM’s Smart Grid IGERT, is modeling what could happen to Vermont’s landscape if we shift most of our electrical demand to distributed solar and wind. “It’s a major impact on tens of thousands of acres,” Clement says, “that may have real conflict with current land uses, including dairy and forestry.”
Various versions of this solar-meets-battery scenario have some electricity companies running scared. Today, grid-scale batteries are “tremendously expensive,” Jeff Marshall says, but that’s changing fast. Rising electricity prices and declining costs for both solar panels and batteries mean that grid-connected solar-plus-battery energy systems will be “economic within the next ten to fifteen years,” the Rocky Mountain Institute forecasts, “and could soon supply a majority of customers’ needs.” To prevent large-scale “grid defection,” where homeowners and businesses decide that their smartest grid is no grid at all, may call for new rate structures and regulations—and may require utilities to reimagine their business as something other than selling kilowatts pumped in from far-away power stations.
----
I’ve come to UVM’s Gund Institute for Ecological Economics to continue my search for the smart grid. It’s a fine, sunny day in May, but Dan Fredman ’03 picks up his smart phone and texts his wife back at their apartment—and asks her to turn on all the lights and crank the air-conditioning. Next to him on the table is a device that looks like a digital picture frame. It’s an “in-home display” and it may be the way the smart grid, someday, gets into your home.
For many people, smart grid—if it means anything at all—means a smart meter. In the wake of the economic crisis, $4.5 billion in federal stimulus money was directed in 2009 toward building the U.S. smart grid. In Vermont, $69 million came in, matched by an equal amount of in-state funds, for installation of digital smart meters and now about ninety-two percent of Vermont households have one.
It used to be that once a month a meter reader would come walking by. “Now, in Burlington, smart meters collect how many kilowatt-hours are being consumed every fifteen minutes, and automatically transmit this information back to the Burlington Electric Department every eight hours.” If nothing else, smart meters—combined with other grid sensors—have helped utilities to more quickly respond to power outages.
But one of the central promises of this massive investment of government resources in smart meters was that it would allow homeowners to see how much electricity they’re consuming, giving them a tool to conserve energy and save money. So far, that hasn’t happened much.
There are many reasons why the smart meter promise hasn’t yet been met. One is that, to most people, the readout on a smart meter is gobbledygook: who balances their checkbook in kilowatt-hours? That’s where an in-home display can help: it collects information from the smart meter in real-time and translates it into dollars and cents. As part of their doctoral research in the UVM smart grid effort, Dan Fredman and Elizabeth Palchak are working with Burlington Electric to deploy two hundred in-home displays to volunteers in rental apartments around town.
Fredman points to his phone, which is receiving a signal from an in-home display back at his apartment. Before his wife turned on all the lights, the screen showed that he was spending three cents an hour on electricity. Now it’s showing thirty-three cents, a ten-fold increase. “If we kept this going 24/7 that’s like $300 a month,’” Fredman says—and texts his wife again asking her to turn it all off again.
“The fundamental point of this study is to see what happens to people’s behavior when they get real-time information about their energy use and costs,” he explains. One cohort of volunteers in the study will be offered an incentive—their monthly bill paid off—it they’re able reduce their consumption the most in that group. Does clear information, or competition, or a combination of both, compel renters—who don’t have a financial incentive to take on traditional efficiency investments like insulation or new appliances—to reduce their energy use?
Fredman, like many other scientists studying the social nature of our energy systems, wants to know “what moves people and what behavioral strategies work best to help them to conserve?” Digging under the layers of new technology, there’s a growing body of evidence that the grid will only be as smart as we are.
Feature story by Joshua Brown. From the Summer 2015 edition of Vermont Quarterly, UVM's alumni magazine.