How much wind could the East Coast take if its power grid could take wind?
How much wind power a system can incorporate is a significant problem in power grid management. At a basic level, grid operators can’t control when the wind blows, and the problem is exacerbated by the fact that more traditional generators can’t always be ramped up quickly if the wind suddenly cuts out. If gusts unexpectedly stop gusting on a hot July day and everyone on the East Coast reaches to turn on their air-conditioning at the same time, grid operators can neither generate electricity from the wind, nor can they always immediately generate electricity from more traditional sources, because those need time to ramp up. Thus a generation shortage occurs—and that can lead to a power outage.
A recent pair of studies conducted by the University of Delaware and Princeton University tried to assess how much wind power PJM Interconnection, an East Coast-based grid operator that serves 60 million people in 14 states, could incorporate before the wind’s unpredictability might put grid reliability at risk.
The researchers ran hundreds of simulations to see how well the grid could adjust to five levels of offshore wind farm build-out along the US Atlantic Coast: level 1 added 7 gigawatts of installed capacity and level 5 added 70 gigawatts of installed capacity.
The researchers concluded that PJM could handle quite a lot of offshore wind power on its transmission lines, even using conservative wind forecasting and assuming no advances in turbine or transmission technology. Specifically, PJM could handle 7 GW of installed offshore wind capacity using the most conservative estimates. More generous estimates allowed nearly 36 GW of installed wind capacity. Scenarios permitting level 5 build-out, or 70 GW of installed capacity, required more significant grid and forecasting upgrades.
The answers are blowin’ in the wind
Part one of the researchers’ two papers addressed wind forecasting for the proposed offshore wind farms. The US is far behind Europe in offshore wind installations—our first offshore wind farm only just went live this past December. That meant that, while this paper was being completed, there were no long-term measurements of wind speeds at hub-height for offshore wind turbines along much of the US East Coast. (Although that may have changed since the paper was written, because just this winter New York agreed to lease significant tracts of offshore acreage to wind farm operators.) Still, “the lack of adequate wind speed observations has been identified already as one of the main obstacles to the development of offshore wind farms along the US East Coast,“ the researchers wrote.
In lieu of good offshore wind speed data, the first paper tried to estimate wind potential by using offshore wind forecasts generated by Weather Research and Forecasting (WRF), combined with forecast error probabilities gleaned from 23 Great Plains wind farms operated by PJM. The researchers chose to estimate wind scenarios using data from January, April, July, and October, assuming that one month for each season would reflect seasonal wind conditions adequately.
Put it on the grid
Once the researchers had drawn up estimates of the wind speeds that the proposed offshore turbines would experience, they applied it to their grid simulation. Their second paper outlined the process they used to determine how much wind could be injected into PJM’s system without compromising the operator’s ability to serve all its customers. The researchers built a simulator called Smart-ISO that took into account generator scheduling, transmission capacity, and the wind model from the previous paper.
Smart-ISO balanced predicted injections of wind against a list of 830 real-world generators within the PJM system. The generators were classified as either “must-run” units or as slow, fast, and “other” units depending on how quickly the generator could be ramped up. Must-run units included nuclear-fueled generators and coal-fueled generators “with notification plus warm-up times above 32 hours.” Slow generators needed warm up times of two to 32 hours, fast generators could warm up in under two hours, and “other” indicated hydroelectric generators, pumped storage, and wind.
Smart-ISO was also run assuming either a constrained or an unconstrained grid. An unconstrained grid assumes the power lines are unchanged but that their thermal and electric power carrying capacities can handle as much wind as operators throw at it. A constrained grid assumes power lines are limited to their real thermal capacities.
First, the researchers looked at an unconstrained grid without any additional generation reserves added to PJM’s current system. For those simulations, PJM was able to handle build-out level 1 (7 GW) relatively easily but faced generation shortfalls moving to build-out level 2 (25.3 GW of installed offshore wind capacity) due to a lack of fast-ramping generators.