How 3Tier Wind Forecasts Enable the Northwest Smart Grid
The Northwest is undertaking a grand project that could unlock a future of efficient energy use and smooth integration of renewables—two necessary pieces of a solution to the immense challenge of climate change.
The beating heart of the Pacific Northwest Smart Grid Demonstration Project, one of the largest in the world, is an incentive signal to utilities about the current and future state of the power system, which will allow them to plan for and react to the price of power in ways they never could before.
In a region that already has one of the highest levels of wind capacity in the country—4,711 megawatts of wind, or nearly a quarter of the nameplate capacity of the 31 federal hydro dams in the region—forecasting is fundamental to predicting the future electricity supply and crucial to this project.
One of the Northwest’s quintessential cleantech companies, Seattle’s 3Tier, is providing those forecasts, using a combination of global weather and climate models, computing power, and the expertise gained from tracking wind energy production in the Columbia Gorge for 13 years. (Check out our update of 3Tier’s strategy under new chief executive Craig Husa at the end of this story.)
In mid-October, as fall finally took hold in the Northwest, a front approached pushing strong winds toward the hundreds of wind turbines built in eastern Oregon and Washington in the last decade. 3Tier’s wind energy forecast on Oct. 14 showed the potential for an extraordinary event: wind energy production on the morning of 16th would exceed hydroelectric output for the first time in the history of the Bonneville Power Administration (BPA).
The forecast came to pass. For a time that Tuesday morning, wind energy production on the BPA system was near full capacity and equaled about 85 percent of regional demand. 3Tier touted its prediction because “wind is almost always unfairly characterized as being unpredictable, and we’ve been saying for going on a decade, that’s not true,” says 3Tier chief operations officer Pascal Storck. “Wind is predictable with a high degree of accuracy a day or two in advance. That’s really all the grid needs to successfully integrate it.”
That accurate day-ahead predictability makes wind and solar less different from traditional forms of generation than is commonly believed. Storms, unexpected maintenance problems and natural disasters can knock power plants of all kinds offline at any time.
“Nobody can guarantee that a particular asset is going to be available weeks, months, years into the future,” Storck argues. “Same thing with wind. I can’t tell you that on Aug. 30 of 2013, it’s going to be a windy day in the Pacific Northwest, but the day before that day shows up, we can nail it.”
Combining accurate forecasting with smart-grid capabilities promises to make more efficient use of renewable energy generation and reduce the costs of its integration—a challenge that will grow as wind and solar become an increasingly significant part of the nation’s electricity infrastructure. The $178 million Pacific Northwest Smart Grid Demonstration Project—led by Battelle in collaboration with the BPA and funded with $89 million in federal stimulus from the Department of Energy, which is matched or exceeded by the participants—is testing that promise with data to be collected during the next two years.
The fundamental challenge of the power grid is keeping supply and demand in balance. The supply of electricity used to be easily controllable with generation at dams and coal plants that can be turned up or down to accommodate the somewhat random, yet predictable fluctuations in demand. Today and in the future, as more renewable energy is added to the grid, electricity supply is less controllable. (3Tier didn’t conjure the mid-October storm front that jacked up wind generation; it just predicted it.)
With smart grid, the demand or “load” side of the equation can be throttled up or down, too, adjusting power consumption in response to changing conditions, such as a spike in wind generation. When the smart grid signals utilities that power is relatively cheap, intelligent devices can react to that incentive in a way that benefits the broader system. In response to a low price signal, water heaters turn on, smart buildings rev up the HVAC system, and battery storage units begin charging, or vice versa.
“We have now the means to engage loads, to have loads be an active part of the system rather than a passive part of the system,” Ron Melton, director of the smart grid project, tells me from the Pacific Northwest National Laboratory operated by Battelle in Richland. “This fundamentally changes our challenge in operating the power system.”
The project involves 11 utilities and 60,000 electrical meters in Washington, Oregon, Idaho, Montana, and Wyoming, making it unique in the world in scope and in its ambition of coordinating demand and supply within and across multiple geographically dispersed utilities.
“There’s no other project that I’m aware of in the world that’s trying to provide this mechanism—a smart grid application layer, if you will—that pulls together the ability of coordinating responses across multiple utilities in a region to provide a regional benefit,” says Melton, who presented the project at international smart grid conferences in China and France earlier this year.
Those models take forecast inputs from 3Tier for wind generation and from the BPA and elsewhere on other aspects of the current and future state of the system, computing the cost of power delivered to each utility based on its location in the power system.
3Tier’s Storck says the company did not have to make major adjustments to its forecasts for the smart grid project. He and others argue that the technologies underlying the smart grid are mature. The challenges that remain—and they are substantial—are around technology integration and proving the value of the smart grid.
“If the value is there, then I think this can scale up, not just in the U.S., but there’s other projects all throughout the world,” Storck says, pointing to the potential of new business for the 60-person company.
Indeed, 3Tier’s focus is increasingly global. New U.S. wind project development has all but disappeared in the face of uncertainty about the future of the wind energy production tax credit and the disruption of the electricity market caused by the natural gas boom.
Since becoming chief executive a year and a half ago, Craig Husa has helped refocus 3Tier on the industries, players, and locations that offer the best growth opportunity. Two days after we talked late last month, Husa and Storck jumped on a plane for a two-week trip to India and China.
“We’ve seen a shift in where some of our revenue is coming from, but the good thing is that it’s still growing on an overall basis, and it’s largely because the industry is getting smarter and needs to have better intelligence, which we can provide,” Husa says, noting 3Tier still gets 60-70 percent of its revenue from North America, with growth in solar outpacing the company’s biggest industry segment, wind.
That increasing sophistication reflects the financial motivation that has come to underpin renewable energy development.
Sustainability, climate change, energy security—these are all good reasons to develop renewable energy, Husa tells me. “But there’s also the economics. And for the industry to really become successful and mature, you have to have both.”
3Tier, a “big data” company if ever there was one, has transitioned in the last decade from data generation to data mining—crunching through global weather models produced by the likes of the European Centre for Medium-Range Weather Forecasts, the National Weather Service, and other governments.
Combined with hundreds of other forecasts and bits of information ranging from local topography to wind equipment characteristics and wind farm layout, 3Tier produces customized short- and long-term variability and risk analysis for customers with financial stakes in wind and solar energy production.
“The magic is being able to coax out something that is intelligent and actually actionable,” Husa says.