Integrating Renewables Into the U.S. Electric Grid – a Discussion with Dr. Paul Denholm
Earlier this month, I attended at the American Society of Mechanical Engineers (ASME) Energy Sustainability Conference in Washington, DC. During the conference, I had the opportunity to speak with Dr. Paul Denholm, a senior analyst at the U.S. National Renewable Energy Laboratory (NREL) in Golden, Colorado. Dr. Denholm is a member of the Energy Forecasting and Modeling Group in the Strategic Energy Analysis Center at NREL. He works on figuring out how we can best model electric power systems, including interactions between renewable and conventional energy technologies.
I caught up with Dr. Denholm after he gave a talk on his paper titled “Enabling Technologies for High Penetration of Wind and Solar Energy.”
MCL: Thanks again for taking the time to chat a bit about integrating renewables into the U.S. grid. To begin – would you define the terms “intermittency” and “variable” for us?
PD: Sure. Intermittency is a term that people often apply to wind and solar generation because it is variable by nature – the wind doesn’t always blow and the sun doesn’t always shine. But, actually, more and more people are now using the term variable.
A lot of people, especially in the wind community, find the term intermittent is often used as a negative – almost to attack wind. They believe that the term “variable” is a more technically accurate term. And the reason is, that people use the term intermittent to imply that the wind will just suddenly stop or suddenly start. But, wind will actually slowly slow down and start up. It will very rarely just suddenly change from zero to sixty in a second or two. So, some believe that “variable” is a better term to describe the resource
MCL: Why is this variability a problem as we try to increase the amount of renewable power that we have on our grid?
Variable resources are ones that vary depending on the weather conditions. So, with solar, that’s pretty obvious. If you have a solar generator, it will generate electricity when the sun is shining and it won’t generate at night. Wind will only generate during periods when it’s windy. And, the fact that the wind speed changes means that the output from a wind generator will vary quite a bit over the course of hours and days. That variation of output is very unlike traditional fossil fuel generators where utilities can completely control the output and rely on that generator to be available pretty much all the time.
MCL: Could you also describe what “flexibility” means for our current generation fleet?
PD: Right now, the demand for electricity varies a large amount from hour to hour. In the middle of the night, around 3 to 4 AM – when people are asleep – there is not a lot of demand for electricity. There’s minimum point where utilities kinda turn back all of their power plants and just meet this base load. As people start to turn on lights and go about their day, their demand for electricity increases. And, depending on the season, their demand might peak at about 4pm. In fact, a lot of electricity systems will be using twice as much electricity in the afternoon compared to the middle of the night.
So, the flexibility is the ability of the utility to increase the output of their generation fleet to meet this demand in the middle of the day. The challenge, when you bring in renewables, is that non-renewable generators will have to be able to step on the gas pedal harder to ramp up generation to balance variable renewable power.
MCL: Is our current generation fleet flexible or inflexible?
PD: The flexibility of a power system is dependent on the types of generation technology it includes. So, the best example is – a nuclear power plant is not very flexible. It has a limited ability to rapidly change its output. On the other hand, places where they have a lot of [natural] gas or, even better, a lot of hydro would be very flexible. So, the ability to include lots of wind and solar in the system is largely dependent on what the remaining generators are.
One of the things that we’ve done in the past to understand flexibility is to do wind integrations studies. Here, we take the generators that exist now and put them into a computer model that basically simulates the hour-to-hour operation of the grid for a given year. So, you can take today’s grid and stick a bunch of wind on it to see if the generators that exist now – do they respond to the increased need of flexibility.
MCL: And what did these studies find in terms of the grid’s flexibility?
PD: Both of these studies looked at up to 30% wind penetration – where 30% of annual electricity demand is supplied by wind – and both found that the existing generation fleet is flexible enough to accommodate that level of penetration.
Now, they had to do some changes to how the system is operated – including greater cooperation between utilities and faster communication so they can share resources and new ways of sharing information across large areas. But, the generation fleet is flexible enough to have up to this 30% wind level. Is it flexible to meet more than that? We don’t know. It’s certainly pushing the limits of flexibility.
MCL: Now, this 30% – that’s above the 10-20% limit that people talk about being a realistic limit without storage.
PD: Yeah, it’s really interesting. If you look through the papers that have been published, and columns, and newspaper articles there’s this magical 10-20%. The problem, quite frankly, is that those things that said 10-20%, they were guessing. And the reason I say that they were guessing is that the only way that you can figure out these limits is to do these computer simulations – or well, build all of these generators. But, that’s like how you test bridges by driving bigger and bigger trucks over it. You really want to do something before that.
Before 2 years ago, we couldn’t have done these simulations because we didn’t have simulated wind data over large amounts of the country. We had some anemometers and some wind farms. But you can’t take isolated bits of wind data and extrapolate that to 20 or 30% wind.
Until you do these simulations, you just can’t know where technical and economic limits exist.
MCL: So, how far can we go with wind?
PD: We now know that the percentage is a moving target. We’ve gone up to 30% and said, “Ok, we know we can do that” and I’m sure next it’ll be 35 or 40%. And yes, we’ll see increasing problems with this. But, I don’t think that we’ll find a place where the grid breaks. It’s not like we’re going to see some hard limit where, lets say, at 29% it works and at 29.5% it doesn’t. It’s all just kind of, “OK, we have a new challenge and we need to do ‘this thing’” and this comes with a cost. So there might be an economic limit where society says, “hey, this is now too expensive, we want to do something else.” But, the limit will be economic, not technical.
MCL: What types of things can we do so make the grid more flexible to make renewable integration an easier transition?
PD: The most effective things that we can do to make the grid more flexible are maybe the most boring. And least technical.
To give you an example –
In the eastern U.S., there’s this entity called PJM, an independent system operator that runs the grid through a bunch of states in the mid-Atlantic. It used to be that, in each state- for the most part – the local utility would run the grid and it would be isolated. They didn’t have a while lot of ability to ship power from place to place in real time. So, if they found themselves in a shortfall, it was difficult for them to talk to their neighbors to ask for power. And vice versa, with exporting excess power to help others out. When they formed PJM, they combined all of those areas into a single balancing area. So, now you can ship power from anywhere – Ohio to Pennsylvania to Virginia to Maryland – basically in real time and balance the entire system. That’s going to be very important in the case of wind with its variability and, quite frankly, the limited predictability of the resource. So now, all of the sudden, when we got too much wind we can ship it anywhere as long as transmission’s available.
MCL: So, better communication and connectedness increases flexibility?
PD: Yes. As we’ve consolidated areas and have computers and allowed for more communication, it has allowed us to better share not only wind and solar resources, but also demand and generation resources. And, one of the biggest things that we’ve found is that, allowing for this increased level of cooperation – and it doesn’t have to be new markets – can be the biggest thing that we can do to enable greater flexibility in the grid.
The other thing that we’ve found – the other market piece of this – is allowing demand to respond to either price signals or some other mechanism that incentivizes people to use electricity when it’s cheap, and not when it’s expensive.
In most of the U.S., it’s still the fact that most residential customers pay the same amount for electricity in the middle of the night when it’s really cheap and in the middle of the day when it’s really expensive and, quite frankly, they’re in danger of a system blackout. You still have people running appliances because they don’t know that it’s stressing the system and they have no financial inventive not to do so. If you introduce market mechanisms that we use for any other commodity that’s cost varies seasonally or hourly – airline tickets, strawberries, etc. – that will make the electric power system more efficient and better able to incorporate variable generation resources.
And it doesn’t have to be some kind of “ra! ra! socially good thing.” It can be just a sheer profit, least-cost solution.
MCL: Could you touch on more technology-based solutions that we might also use?
PD: Sure. General Electric, for example, has introduced a new line of gas turbines and combined cycle gas turbines that have greater flexibility. Meaning, their ability to ramp rapidly and over a large range – how fast can they go from “zero to sixty” and what their top and bottom speeds are. This is called ramp rate and range.
Right now, when wind and solar are producing, and you can’t turn thermal power plants [gas, coal, etc.] down any further, you use less renewables. New turbines could allow you to use more renewables but ramping down over a larger range without too big of an efficiency hit. This increases system flexibility.
MCL: What about storage?
PD: A lot of people talk about energy storage as being the holy grail of the grid. And, I think storage is a really cool set of technologies. But, we have to be a little bit careful in assuming that storage will be the savior of the grid and that storage is somehow needed right now.
We do have storage in the grid right now in the U.S. – we have 20 GW or about the equivalent of 20 large power plants of pumped hydro-storage. But, there are a lot of places that don’t have storage and are able to operate the grid fine without storage.
As I’ve said, the range in the demand of electricity varies by a factor of 2 from the middle of the day to the lowest demand period. And, utilities are doing just fine with that level of variability.
The point being is utilities already have to deal with a lot of variability on their grid. So, just because you introduce more variability doesn’t mean all of the suddenly you have to totally change how the grid works and introduce electricity storage.
We’ve found that storage is not needed in the grid, but could be a helpful addition to the grid if it’s low enough cost and widely available. There are a large number of storage technologies available with varying levels of economic competitiveness. And, I think that there’s a lot of hope that energy storage technologies will continue to improve. Because, I think that there’s no doubt that storage technologies would be a great addition to the grid. If we could get low cost storage we could store wind, solar, or other low-cost generation resources when more is being produced than is needed and release that energy during periods of reduce wind and solar output or high energy demand. So, storage could be a great addition to the grid. But, it is just one of the pieces of what we can use to accommodate more variable generation on the grid.
MCL: Thanks, Paul. Appreciate your taking the time.