Michael Giberson
Wind power has been subsidized by state and federal governments in the United States because it is seen as clean and renewable, and perhaps even because wind power is seen as glamorous. Consumers pay higher electric rates and taxpayers pay higher taxes to support these subsidies, and it is a quite reasonable public policy question to ask whether the benefits are worth the costs. (Of course wind power is not the only energy technology subsidized by government policy.)
The primary external benefits from expanded wind power production comes from emissions avoided due to the reduced use of fossil-fuel fired electric generation, predominantly natural gas and coal. Which fuel is displaced, however, depends in large part on where the wind power project is located and what time of day the wind power is put onto the grid.
Conventionally, an estimate of reduced emissions might be made through an elaborate production cost modeling exercise, comparing overall use of different input fuels against scenarios featuring different levels of installed wind capacity. It is one useful approach, but it would be good as a reality check to test such estimates against actual data. Two recent estimates of fuel displaced by wind power rely on data analysis to get their results.
A relatively straightforward approach to this estimate was taken by Monitoring Analytics, the external market monitor for the PJM market, in preparing “Estimated Marginal Fuel Displacement By Wind Generation in PJM.” The chart was posted online without accompanying documentation, but folks at Monitoring Analytics tell me their estimate was derived from market data on wind power output by hour combined with data on marginal generation by fuel type by hour. As the chart nearby indicates, about 75 to 80 percent of the wind-produced power in PJM displaced coal-fired power. (Coal is the orange portion of the bars.)
Joseph Cullen took a more data-intensive econometric approach to estimating the fuel displaced and related emission reductions in ERCOT due to wind power. Cullen ran regressions on the output of each non-wind generating unit in the ERCOT market against wind power output to identify the actual responsiveness of each generator to changes in wind power. (I’m over-simplifying his methods. See his paper for details.) In ERCOT, for the time period analyzed, Cullen estimated that about 80 percent of the time wind displaced gas-fired generation and about 20 percent of the time wind displaced coal-fired generation.
One of my policy objections to the production tax credit approach to subsidizing wind power is that it offers the same subsidy per MWh output without respect to the environmental benefits provided (if any). Therefore it tends to be more attractive to the developer to invest where wind power output will be high – i.e. West Texas, among other places – and the external benefits relatively muted – instead of where the external benefits would be high, as in PJM. So much wind power capacity has been added in West Texas, relative to the current grid capability, that wind power capacity in effect just displaces other wind power generation during high output periods.
Why should consumers and taxpayers subsidize that?
From a commercial point of view, it certainly makes sense to build wind power where wind power output will be high. I’m not opposed to smart commercial activity. I don’t see that public policies should subsidize it. Rather, public policy should be oriented at achieving external benefits in a cost-effective manner.
Consumers and taxpayers will end up getting more for their money from policies that put a price on the externality.
I run full-information structural power models for a living. My modeling is more long term, but indcates that long term it is gas, and not coal, that is displaced by new wind. Almost entirely gas.
Another damaging fact, seldom mentioned out loud, is that quite often, wind actually displaces nothing. The wind blows, nothing gets shut off (especially not baseload coal), prices crash for a few minutes and then things stabilize.
Bartman,
I am a little puzzled by your comment that wind power often “displaces nothing”. It may be true that wind power can cause prices to crash for a few periods, but the lower prices also induce lower production all else equal. Even from an engineering standpoint when additional wind power comes on the grid some generator has to reduce production to maintain the appropriate voltage unless the excess power is just grounded out. It is my understanding that power is not grounded out (thrown away) since it puts undue stress on the generating equipment. Is there something I am missing here?
Government subsidies for wind power only make people feel good. Beer would be cheaper and more effective.
Joseph:
Coal generators are loath to cycle their boilers up and down, because this greatly increases maintenance costs and potential for unscheduled shutdowns, so they continue to burn the same amount of fuel and simply run a little less steam through their turbines. More steam gets routed to the cooling towers or ponds. So some of the energy extracted from the environment as wind gets put back into the environment as heat at coal plants. No less CO2 is emitted.
I already knew that ERCOT was Energy [something] Texas. Digging a little, I learned that PJM is Pennsylvania, New Jersey, Maryland (and some or all of eight other states).
I’m still a bit confused, though. It appears that we’ve decided that wind generation is a good thing and we should subsidize its contruction & connection to the grid. Ok. Fine.
Constructing & interconnecting the wind to the grid does not render the wind power dispatchable, though.
I mean, I can make a basketball team but I have no quarantee that I’ll be able to play – let alone start.
So don’t we need a second subsidy to pay the 2nd delta between the wind generation and the nuclear generation that is most likely dispatched?
Or am I missing a synapse?
PS – I completely concede your point re: other fuel technologies that are subsidized by the feds.
Katie:
Large scale wind has three components, two of which are seldom talked about: (1) wind turbines, (2) transmission, and (3) balancing energy. Balancing energy can be any of many things – batteries, compressed air storage, pumped hydro, gas turbines – but it is an absolute necessity, and must be considered as part of the total system cost of going to a “25% RPS” world. But, of course, this multiplies the cost of wind generation by some number greater than 3 or 4, so the wind developers don’t want to talk about it, and certinly don’t want to pay for it.
Nonetheless, wind looks to become a boondoggle on a bigger scale than corn ethanol, the last great, stupid, short-sighted, politically motivated energy mistake.
Bill:
ERCOT = Electricity Reliability Council of Texas.
Where to start?
I take issue first with the idea that fuel input to coal-fired boilers is not cycled down when the electrical output of the generators they drive is cycled down. It is true that many coal-fired boilers are less efficient at lower output levels, but for many coal units there is a direct simple relationship between fuel input and unit output over a rather large operating range. And while even the lowest-efficiency coal-unit capacity factors have been increasing in recent years, many of them do still cycle on a daily basis. But there is no need to theorize — the EPA posts hourly records of both heat input and MW output for almost every coal unit in the country. It’s not the easiest data in the world to interpret, but it demonstrates that fuel input and electricity output for coal units do vary up and down together.
As for displacement by wind generation, it is true that an increase in wind-energy injected into the system directly displaces energy injection by other generators, other things being equal. And as my prior paragraph makes clear, it also displaces some fuel input to those plants, even if it might make some of them operate less efficiently. [This is not necessarily the case — some plants may be forced to shut down completely. The fuel efficiency of the system can be better, worse, or the same after wind injection, depending on a variety of details. At the margin – for the last kW of wind output injected – the displaced unit probably does run less efficiently.]
As for what fuel gets displaced by wind generation, it really depends on the power system into which it is being injected, in multiple ways. Two of the most important characteristics are the mix of capacity and fuel types, and the capability of the transmission system. ERCOT is a nice example because it is nearly a stand-alone system, and its wind-producing area is transmission-constrained.
ERCOT’s capacity stack was about 21% coal and 6% nuclear in 2007. Almost all of the rest is gas-fired, and most of that is in efficient combined-cycle configuration. Over half of ERCOT’s electrical energy comes from gas. Even when gas is relatively cheap, as it is now, it is still more expensive than most coal units per unit of electrical output, all expenses considered. So what this implies for ERCOT is that gas-fired generation is “on the margin” almost all of the time. That is, if someone on the system reduces their consumption by 1 kW, then some gas-fired unit reduces its output by 1 kW(+losses). The same is true of wind-energy injection; in ERCOT it almost certainly displaces gas. In fact, it often displaces very efficient gas-fired generation. A declining quarter of ERCOT’s capacity is inefficient gas-fired boiler units, but these units are not on the margin as often as efficient, combined-cycle gas. Nevertheless, if wind output happens to coincide with the peak load, then it can displace inefficient boiler gas in ERCOT. Other systems have different generation mixes and may have coal on the margin more often than ERCOT, and if wind output occurs off-peak then it has a chance of offsetting coal-fired generation. But most of the country has gas on the margin most of the time, so it is getting less and less likely that wind energy will displace coal as long as it remains cheaper than gas to burn.
Then there are the transmission issues. Wind energy can displace energy generated in the broader system only up to the transmission limit of the wind-energy-producing area. After that limit is reached, then additional wind generation can displace only resources within the constrained area. At the extreme, this local surplus can cause low or negative prices for power injection. Negative prices indicate that some injection must be removed from the system to bring it within safe operating limits. (No one “grounds out” electric energy, at least not on purpose. We don’t throw it away, we just don’t generate it. Blades can be trimmed to reduce output, and/or the unit can be taken offline.) Subsidies based on output really distort prices when local systems go into surplus. That’s a different issue that I’d like to avoid here.
Bartman is correct that there are unseen costs associated with wind generation. Busbar cost per kWh is a poor way to compare wind generation with other types, but it is a difficult nut to crack. On this blog I have suggested numerous times that a better question about wind generation is the net *value* of its output at the time and place it is injected onto the system. Negative prices indicate negative value (subsidies aside). How much sense does it make to build a wind generator if its output has negative value, regardless of what it costs? And absent smart-grid-style demand response and/or storage, it is also true that wind generation requires some means of backup. (The West may have some flexibility to deal with vagaries of wind through the use of its abundant hydro resources, but Eastern markets don’t have quite that much flexibility.) In any case, the key is that prices should reflect the value of energy generated or consumed at the time and place where the energy is injected or extracted. Only after some time under those conditions will we ever really find out what amount of backup or storage is needed, given the costs and benefits of both, as well as the costs and benefits of all other related things, including our lights, cars, houses, TV sets…
Doug: The operations and dispatch people in the company I work for, a large generator, tell me that short-lived negative price signals that correspond with spikes in wind output do not cause them to change coal burn rates, because (1) it takes more than five minutes to get an output response, and (2) the negative price signal can often go away in the next five minutes as the wind calms down, so you have to ramp up again. The plants run at the output at which they were dispatched in the day-ahead market.
Generators, as profit maximizers, would love to ramp down so as to be able to “pay” massively negative LMPs, but owners of big coal plants don’t do this because either they can’t, technically, or it isn’t economic to do so.
Yes, coal units can be modulated in response to hourly day-ahead price signals, but that isn’t what I was talking about. The bottom line is, most don’t move hardly at all. You can verify this by looking at PJM bid data – lots of units have the same min and max dispatch rates, along with min uptimes of at least 24 hours.
Over 5-minute intervals, I agree that they don’t react. Coal units can’t ramp that fast. But let’s be clear that that’s what we’re talking about, i.e., very short-term perturbations.
Yes, many coal units are base loaded, but many cycle to some extent every day, in concert with day-ahead prices signals. That is, not in response to day-ahead prices, but as components of the market system that results in the day-ahead prices. In any case, let’s be clear that when you are saying that coal units don’t move at all, that you’re talking about very short periods of time and in response to short price excursions. Because over a 24 hour period they do move, and they do reduce heat input accordingly. And to the extent that wind resources have some average output level, they become involved in the longer-term price setting and dispatch process, too. It is the very short-term volatility that coal plants can’t follow.