I operate the electric grid for a living. From a purely professional standpoint, I don’t much care how our electricity is generated, as long as there’s enough of it available to meet demand. And this is where discussions of the annualized output of the various types of generating station can mislead: Unlike more traditional power plants, which can be ordered to a specific output at any time, up to their max output rating, renewables like wind and solar generate only as much as they want to generate at any given moment. We can’t always order them up to their maximum rated capacity when the customers demand it. We can’t even count on ordering them up to their average annual output. The environmental circumstances of the moment simply might not permit it.
With our current mix of energy resources, the renewables are pretty much allowed to generate as much as they care to generate, and the other (mostly fossil) power plants adjust their output accordingly. This arrangement is mostly working acceptably at the current mix of energy resources on the grid. But as we replace more and more of the power plants that we can control with ones that we can’t, we start to put ourselves closer and closer to a scenario where your refrigerator and air conditioner are at the mercy of the wind and sun, with very little that I can do about it.
System load typically peaks in the late afternoon and early evening, as all the “nine-to-fivers” start getting home from work and turning on their televisions, electric ovens, and computers, and adjusting their thermostats to more comfortable temperatures. This load peak, unfortunately, happens after peak solar. So in this early-evening period, we see load rising toward its peak at the same time that solar is trailing off from its peak. Not only do we need to find megawatts to support all that additional customer load, but we also need to find megawatts to replace the diminishing solar output. Where is all that power going to come from? Wind? That might not work as well as you expect. Because wind typically only generates at a small fraction of its capacity rating, and the grid needs to match load with generation now — in real time.
Let’s take a fictitious scenario: Over the course of a year, a certain region gets 60% of its power from fossil, 20% from solar, and 20% from wind. You might be tempted to think that if we tripled our wind and solar capacity, we’d be able to get rid of all our fossil plants with room to spare. You’d be wrong.
Because there was a moment during that year — on a hot night, when the wind was still and the air conditioners were cranking — during which wind was generating less than 5% of the energy being consumed and solar was generating none of it. The fossil plants were generating basically all of the energy being consumed in that moment, even though, at the end of the year, they only ended up generating 60% of the total. Meeting demand means that whatever resources we have available at any given moment must meet the demand in that moment (plus some margin). The annualized numbers don’t even come close to telling the whole story.
There are a few solutions to this, none of them cheap:
Option 1a: Storage with batteries. Batteries degrade over time, and need to be replaced pretty regularly. They also generate a lot of environmental damage of their own. And they tend to have very small storage capacities, relative to the energy demands we’re talking about here. At most, an on-site battery might be able to slow the rate at which output from a wind or solar facility ramps down, so that other resources have time to accommodate the change.
Option 1b: Pumped storage. This is a man-made pond that’s located next to and higher than a natural body of water, into which water is pumped when generation exceeds demand, and from which water is drawn (to spin turbine-generators) when demand exceeds generation. These are quite reliable and long-lived, but they take up a lot of land and they aren’t free. Hydroelectric power plants can serve a similar function by allowing the volume of water being held in their reservoirs to rise and fall as the need for generation falls and rises, respectively.
Option 2: Overbuild the ever-living crap out of our wind infrastructure. As I already mentioned, the annualized load in a region can’t be met by an equivalent amount of annual solar and wind generation capacity unless you have a lot of storage. Without storage, you’d need to instead have a lot of reserve wind capacity, so that even when the wind is still, it’s generating enough to match load. This isn’t going to be cheap, and will consume a lot of real estate.
In reality, our most practical path forward is probably to mix all of our options: Overbuild our wind capacity, install battery storage at the wind and solar sites to limit the rate at which they ramp down when the environmental conditions change, install pumped storage facilities wherever possible, and keep some of our traditional power plants around.
This is going to end up being a lot more expensive than renewable proponents admit — because we’re going to need a lot more than just installing an equivalent capacity of wind and solar to match what we currently have with fossil generation. And the “invisible hand of the free market” isn’t going to make this overbuild happen, because it’s going to require building and maintaining facilities with the understanding that they’ll sit idle for most of their existence. That’s incredibly expensive, and no company wants to do that. We’ll need to create massive incentives for these businesses to build and maintain idle infrastructure that can be called upon when needed. That’s not just up-front money, that’s a recurring expense that will continue forever.
And we’ll need a new way of monitoring and controlling all of this generation infrastructure. The current system simply monitors grid frequency and sends control signals to online generators telling them “more power” or “less power”. The online generators all respond to this common signal, to the extent that they are able.
Wind and solar facilities, by the way, don’t participate in this control scheme. They just generate whatever they’re able, and the controllable generators adjust their output appropriately. That would obviously need to change if most of our generation is coming from renewables. This would involve retrofitting generation control systems to all of our existing wind and solar facilities.
Under this system , individual generators’ available capacities are roughly static (they might call and report a derate, which we then manually plug into the system, but for the most part, max available output is fixed). A future all-renewable (or heavily-renewable) grid control system will need to know how much actual generating capacity and how much storage are available in real-time, as may be limited by environmental conditions, and then make decisions about where and how to draw power to meet load without expending our reserves. That’s an order of magnitude more complicated than what we have right now. As with all the other issues, it isn’t impossible to overcome, but it’ll cost money.
The resistance from people who are invested in fossil power is a very real impediment to converting our energy economy over to 100% renewables. But it isn’t the only impediment — not even the biggest one, in my opinion. There are very real technical issues that I don’t think the supporters of renewable power are adequately discussing when they trot out fossil power companies as simple boogeymen. And I further feel that renewable energy proponents are underestimating the cost and complexity of moving to an all-renewable energy economy.
Marching forward with the installation of wind and solar and with the decommissioning of fossil plants without first carefully considering these technical issues is a recipe for an unreliable electric grid. These issues can be overcome, but we don’t seem to be giving much consideration to them yet. I fear that we won’t start dealing with them until after they’ve already rendered our electric grid unreliable.
