The Levelized Cost of Electric Generation

Guest Post by Willis Eschenbach

In early 2013, the US Energy Information Agency (EIA) released their new figures for the “levelized cost” of new power plants. I just came across them, so I thought I’d pass them on. These are two years more recent than the same EIA cost estimates I discussed in 2011 here. Levelized cost is the average cost of power from a new generating plant over its entire lifetime of service. The use of levelized cost allows us to compare various energy sources on an even basis. Here are the levelized costs of power by fuel source, for plants with construction started now that would enter service in 2018:

us average levelized costs 2018Figure 1. The levelized cost of new power plants that would come on line in 2018. They are divided into dispatchable (blue bars, marked “D:”) and non-dispatchable power sources (gray bars, marked “N:”).

Now, there are two kinds of electric power sources. Power sources that you can call on at any time, day or night, are called “dispatchable”. These are shown in blue above, and include nuclear, geothermal, fossil fuel, and the like. They form the backbone of the generation mix.

On the other hand, intermittent power sources are called “non-dispatchable”. They include wind and solar. Hydro is an odd case, because typically, for part of the year it’s dispatchable, but in the dry season it may not be. Since it’s only seasonally dispatchable, I’ve put it with the non-dispatchable sources.

OK, first rule of the grid. You need to have as much dispatchable generation as is required by your most extreme load, and right then. The power grid is a jealous bitch, there’s not an iota of storage. When the demand rises, you have to meet it immediately, not in a half hour, or the system goes down. You need power sources that you can call on at any time.

You can’t depend on solar or wind for that, because it might not be there when you need it, and you get grid brownout or blackout. Non-dispatchable power doesn’t cut it for that purpose.

This means that if your demand goes up,  even if you’ve added non-dispatchable power sources like wind or solar to your generation mix, you still need to also add dispatchable power equal to the increased demand.

So there are two options. If the demand goes up, either you have to add more dispatchable power, or you can choose to add both more dispatchable power and more non-dispatchable power. Guess which one is more expensive …

And that, in turn means that the numbers above are deceptive—when demand goes up, as it always does, if you add a hundred megawatts of wind at $0.09 per kWh to the system, you also need to add a hundred megawatts of natural gas or geothermal or nuclear to the system.

As a result, for all of the non-dispatchable power sources, those gray bars in Figure 1, you need to add at least seven cents per kilowatt-hour to the prices shown there, so you’ll have dispatchable power when you need it. Otherwise, the electric power will go out, and you’ll have villagers with torches … and pitchforks …

Finally, I’m not sure I believe the maintenance figures in their report about wind. For solar, they put the price of overhead and maintenance at about one cent per kilowatt-hour. OK, that seems fair enough, there are no moving parts at all, just routine cleaning the dust off the panels.

But then, they say that the overhead and maintenance costs for wind are only one point three cents per kilowatt-hour, just 30% more than solar … sorry, that won’t wash. With wind, you have a multi-tonne complex piece of rapidly rotating machinery, sitting on a monstrous bearing way up on top of a huge pipe, with giant propellors attached to it, hanging out where the strongest winds blow. I’m not believing that the maintenance on that monstrosity will cost only 30% more than dusting photovoltaic panels …

Best to all,


Usual Request: If you disagree with what I or someone else says, please QUOTE THE EXACT WORDS you disagree with. That allows everyone to understand exactly what you are objecting to.


newest oldest most voted
Notify of
wayne Job

Thank you Willis,
The cost of free energy seems to come in at around three times the price of unfree energy.


I am surprised that geothermal is so cheap. The plant operating in California near the Salton sea has horrendous maintenance problems with superheated brine that they have to use as a heat source. You also place coal with CCS slightly above coal. CCS has yet to be shown to work so how can theu know this figure?

Dodgy Geezer

Of particular interest is a VERY well-written PhD thesis by Eleanor Denny, which can be found here:
In it she models the Irish Grid, and shows quite conclusively that the dispatchability problems with wind power (and by implication all of the renewable non-dispatchable sources result in an interesting phenomenon.
A small amount of non-dispatchable power causes few problems to a grid, and is therefore of benefit. But when the amount of non-dispatchable power added to a grid is in excess of a certain percentage, the disadvantages start to outweigh the advantages, and eventually adding more non-dispatchable power produces a ‘negative benefit’ – in other words, you are worse off having the non-dispatchable power generation facility than you would be if it were shut down.
The reason for this is that using non-dispatchable power forces dispatchable systems to either remain on standby, or to vary their output inefficiently to counteract the variability of the non-dispatchable power inputs. This creates inefficiency costs which eventually overwhelm the renewable benefits.
The interesting point is where this ‘negative’ point occurs. Things start to slide when there is about 15% renewable on a grid, and by 25-30% the benefits are negative.
This robust finding completely negates the plans to provide large percentages of renewable power to grids. Which is why you don’t hear much about this paper anywhere….


Willis are the non-dispatchables calculated at nominal or real capacity? I think the UK has numbers for actual availability for wind at around 17-25% which would have considerable impact on writing off the capital investment during the alotted 25 year timeframe for a windmill installation. Equally solar has to be discounted at a <40% ability to deliver on sticker output due to weather/daylight availabilities.
Secondly the minimum price of 7 ct/kWh for backup capacity would only apply to a plant that's already running. Power-ups and power-downs to meet spot demand is much more pricey due to non-optimal generation efficiency and increased maintenance costs and wear & tear.


Wind turbine reliability statistics are not widely available, particularly for newer megawatt turbines. Most studies do not include newer larger turbine designs because the newer designs have fewer operational years and the failure rates are higher on the front end. What company wants to publicize that their turbine frequently breaks down after installation.
“However, the group of megawatt WTs show a significantly higher failure rate, which also
declines by increasing age. But, including now more and more mega-watt WT models of the newest generation, the failure rate in the first year of operation is being reduced.”
“Reported data may overestimate real availability and should be treated cautiously”
The older the turbine the higher the maintenance costs- no kidding.
And who gets to pay for those higher maintenance costs? We the people!


Re: Ninh
The figures for the UK have wind at 27% of nameplate and solar at 10%.


They use a capacity factor of 34% for onshore wind and 25% for solar PV both of which seem high.
They also use 30% for natural gas fired conventional combustion turbine and 30% for natural gas fired advanced combustion. In other words they assume these plants will only be used 30% of the time. Presumably, this is because they assume that they will be used to fill in the gaps from the non-dispatchable sources. If they dumped the solar/wind and used these gas sources instead then the costs should more than halve.

Joe Public

Is this the only realm where adding a ‘freebie’ reduces system efficiency?

Coach Springer

Not sure – either – how wind and hydro are equivalent in cost and as cheap as anything but combined gas. Seems like hydro is concentrated into just a few turbines and a condensed energy source with economies of scale. I’ve seen several reports of wind projects and existing fields abandoned due to cost , but hydros seem to only get put out for salmon.
For all the other reasons noted, this EIA comparison has enough NOAA-style “homogenization” to be entirely misleading. Looks like more political battle prep.


The levelised costs factor in a $15 per metric ton cost for coal and gas that don’t have carbon capture.


“But then, they say that the overhead and maintenance costs for wind are only one point three cents per kilowatt-hour, just 30% more than solar … sorry, that won’t wash. With wind, you have a multi-tonne complex piece of rapidly rotating machinery, sitting on a monstrous bearing way up on top of a huge pipe, ”
A 1500-2500 + ton crane would be needed for work on the rotors, generate/ gearbox etc. In some cases a smaller crane (100 ton) is needed to start the lift and then the larger one finishes it. These cranes are NOT kept on site and may take days to weeks to get there, be assembled and do the job. Crane work is charged port to port, by the hour, and isn’t cheap. There are very few out there meaning days travel time and when it arrives, days to assemble, do the job, be disassembles and hauled back.
I see rates at about 4,500 british pounds sterling/hr. ($7515 USD) and those are rental rates in the UK. A fully crewed union unit would be at least double to triple that IMHO. Here in the US, one would need to call for a quote.


Thanks Willis …
These prices actually look fairly close. Wind is cheap if you leverage a single line of technology and rationalize your maintenance and installs. As mentioned you have to cover non-dispatchable power with dispatchable power. There is another issue in that the renewables need massive upgrades to the transmission system. If I take my bill in a deregulated market … the transmission is about 35 to 40 percent of the bill. To accommodate renewables, the system operator is installing a 2 billion dollar upgrade in Alberta, Canada. From everything I’ve read the electrical transmission grid in the US is ill equipped to handle renewables. So if you look at my bill which is locked in at 0.08$/kwh with 0.06$/kwh for transmission and ancillaries, then looking out with the gas/wind mix coming and transmission upgrades, it is pretty conservative to say normalized power delivered to your door in North America will be around 0.2$/kwh in 2018. I have lived in an area where you needed air conditioning and with these prices in those areas, your power bill would get close to the average mortgage payment. I guess this is what regulators wanted, to push everyone into on site generation as there is in Honolulu where most roofs have solar panels on them. At what point do we all start installing gas reciprocating engines like some industrial customers ( Maybe that’s not what the regulators wanted, shopping malls with on site thermal generation because of the degraded infrastructure and ‘all in’ cost of this ‘fuel mix’.

Bill Illis

Looks like Natural Gas Combined Cycle (gas-turbine and steam-generator making use of the excess left-over heat) is the way to go now.
– cheapest cost;
– natural gas reserves are large enough to last for many decades, and could extend for 100s of years given the new technological advances;
– easy to transport natural gas through pipelines over long distances;
– smaller plants can be built which loses less electricity through shorter transmission lines;
– smaller plants and lower GHG emissions less likely to disturb the enviros;
– more reliable supply than a grid system with renewables (might need more back-up systems if supplying smaller areas than large coal or hydro facilities which have excess capacity built-in);
– long-term reliability and life-time of the plants is the issue outstanding.

Mark Jacobson, who is arguably Stephen Schneider’s heir at Stanford, insists that adding 13,000 turbines in the waters off Long Island (which is part of Stanford’s master plan for turning New York State 100 percent wind, water, solar) will produce no added costs to consumers. He considers anyone who links wind, water, and solar (Stanford’s Greek chorus these days) to higher electricity costs to be grievously misinformed.
So in love with wind is Jacobson that he is about to publish a peer-reviewed paper based on computer modeling indicating such offshore wind farms will meaningfully diminish the power of hurricanes as they approach the coast.
I recently asked if he knows anyone on a fixed income who struggles to pay already high electricity bills. I received no answer. From a lofty moral elevation, however, he did deign to call a critique of mine of Stephen Schneider, based on his sliding honesty scale, “sad.”
Thus, a self-appointed spokesman for “science,” one close to totally lacking any kind of practicality, uses new math to justify all manner of nonsense.
As I argue in “Don’t Sell Your Coat,” such nonsense is liable to have grievous consequences if left unchecked.

Tom H

“There are two options”
A third oprion is regular black-outs like 3rd world countries. (Or the Government can curtail your power for you when, in their wisdom, they don’t believe you need all of it.)


Willis, there’s a few things in the EIA’s numbers that are simply not credible. First the capacity factors for wind. The EIA is projecting 34 and 37 per cent for onshore and offshore. This is simply not achievable. For example, Hydro-Quebec’s two wind generating complexes in Gaspe, an ideal location, never achieved better than 18 per cent during their first three years of operation, and have never exceeded 25 per cent.
Second, the solar PV figure is not possible even in theory. Capacity factor is reduced by 50 per cent simply for diurnal variation. Winter insolation loss incurs at least another 15 per cent loss from lower sun angle. But the big one is cloud cover. Even light cloud cover reduces the amount of insolation by at least 80 per cent at the time of coverage. If you assume that 100 days per year will have cloud cover, that’s about another 25 per cent reduction. Leaving a reasonable annual capacity factor of about 10-12 per cent. Not the claimed 25 per cent.
And we haven’t even included the loss of generation from snow cover, or air blown dust. The latter is particularly serious. Even the normal dust particles in the air produce significant deterioration of the PV cells from light surface scratching resulting in loss of generation from reflection and diffraction effects. After as little as five years, operational experience has shown a loss of as much as half of potential generation from this.
John Marshall is entirely right, the CCS numbers are fantasy. It’s expected from the designs for CCS seen thus far that it will require at least 1/3 of total plant output to run the system.
Also not included are higher O&M costs for all other forms of generation that must vary their output because of renewable load on the system. Varying thermal generation output rather than producing at constant value has significant adverse effect over time on steam and regulating systems in all thermal plants. One utility I know has to do at least 100 reactor “maneuvers” weekly just to compensate for sudden surges and drops in wind generation even where that wind generation is less than 10 per cent of total installed capacity on the grid.
Finally, their variable O&M signficantly overstates the cost of nuclear fuel and signficantly understates the cost of coal or gas.


Many states have a renewable portfolio standard. In order to meet that set percentage, you need 4-5 times as much wind capacity as the amount of power needed, since the capacity factors are so low. Even in Kansas, one of the best states for wind, the state only gives credit at about 10 percent of capacity toward meeting the renewable goals. This clearly ups the price of installed wind per (semi) reliable kilowatt hour.

I had looked at the EIA estimates a number of times and had glossed over “dispatchable” and “non-dispatchable”. For reasons known to EIA, it looks like almost everything but variable generation (wind, solar, etc) is dispatchable. I suppose nuclear plants, coal plants and the big combined cycle plants are dispatched and have some load following capability, but when one comes off dispatch it’s a big deal. When the nukes come off dispatch, you get a press report. These units are also in a category known as base load.
In Regional Transmission Organizations (grid control) the terms are base load for continuously running and dispatchable for those that operate to meet changing demand. Wind, solar and some other renewables, depending on contract, end up as base load.
The true dispatchables are the peaking plants, that really can be turned off and on as required for demand. (You aren’t going to get today’s power needs by cold starting a coal plant). These run under several regimes. Some do not run unless there is high (expensive electricity) demands. An oil fired peaker might run only when the price of electricity gets above $275/MWh, which is the cost of the oil to power most of them. Many of these get capacity payments–that is they are paid for not running based on arcane formulas. Some of the others run based on demand and/or bid in on the next day’s market predictions. They may also get capacity payments for being inactive.
As Willis said, the grid is “jealous”. Demand for electricity is reflected in price. Price controls generation. It’s semi-capitalistic. We have regional control organizations that govern generation, set prices and make payments. The past few weeks have been an example of price and demand. Prices in much of the PJM (Pennsylvania-Jersey-Maryland) area which includes all or parts of NC, VA, PA, MD, DE, OH, WV, MI, NJ, KY, TN and DC, ran in the range of $750-$2500/MWh. Since my income is based on peaking plants, using renewable energy, operation, this period was pretty good. It beats the nice mild weather when the rates are i the $60/MWh range.
I wonder how much of EIA’s crystal ball gazing is self-serving. That is to make politically correct forms of generation more palatable.


wayne Job says:
The cost of free energy seems to come in at around three times the price of unfree energy.
That’s cheap at that price! In Britain we have calculated offshore wind turbine-generated electricity to be 24 times the price of coal-generated.


If Wind and Solar are so “cheap”, why do tax payers have to subsidize them??

CCS is expensive. Having done engineering on such facilities, the capture cost is somewhere north of $100 / tonne. Is this scale of number included in the CCS cases? It seems you have included $15 / tonne in the non-CCS, which would make me think the CCS ones are too low.


Most of the logic here is fine, but only if cost is the only thing that matters.


Tom H says:
February 16, 2014 at 6:03 am
“There are two options”
A third oprion is regular black-outs like 3rd world countries. (Or the Government can curtail your power for you when, in their wisdom, they don’t believe you need all of it.)

Already planned, new meters being unstalled, at your expense of course, which can be remotely shut off by radio by the central agency. One could also have various means to decide who gets power and who gets shut off, like who you voted for, whether you have posted anything critical about the government on, say, this site, etc. And if that sounds paranoid, remember the military adage that says it is not about whether you think they will do it, but if they can do it, and they can right now. If they can, sooner or later they will be tempted… Example, the NSA and their ability to spy on everyone is tempted to use that spying for private purposes, and stories now circulate that inside sources say this has already occured.
And, of course, it will be done by area, areas that politicians live in, the elite areas, will keep the lights on, the countryside, where the pheasants live, will not. Feudalism is returning.

Problem #1: The U.S. Dept. of Energy put its big fat federal thumb on the scale.
Down in the middle of the page it says, “In the AEO2013 reference case a 3-percentage point increase in the cost of capital is added when evaluating investments in greenhouse gas (GHG) intensive technologies like coal-fired power and coal-to-liquids (CTL) plants without carbon control and sequestration (CCS). While the 3-percentage point adjustment is somewhat arbitrary, in levelized cost terms its impact is similar to that of an emissions fee of $15 per metric ton of carbon dioxide (CO2) when investing in a new coal plant without CCS, similar to the costs used by utilities and regulators in their resource planning. The adjustment should not be seen as an increase in the actual cost of financing, but rather as representing the implicit hurdle being added to GHG-intensive projects to account for the possibility they may eventually have to purchase allowances or invest in other GHG emission-reducing projects that offset their emissions. As a result, the levelized capital costs of coal-fired plants without CCS are higher than would otherwise be expected.”
In other words, they’ve added an arbitrary fudge factor to penalize coal for enhancing agricultural productivity.
Problem #2: They used unrealistic cost recovery periods.
The EIA page says, “The levelized cost shown for each utility-scale generation technology in the tables in this discussion are calculated based on a 30-year cost recovery period, using a real after tax weighted average cost of capital (WACC) of 6.6 percent. In reality, the cost recovery period and cost of capital can vary by technology and project type.”
Does anyone really think wind turbines will last 30 years?
Fortunately for Big Wind, they don’t have to. Among the many incentives the federal government offers for investment in Renewable Energy is a Modified (6-year) Accelerated Cost Recovery System (MACRS) Depreciation.


For all the CCS doubters, the penalty is not as high as you’d think relative to ‘Clean Coal’ and you can go to the SaskPower symposium in Regina in October after they’ve run it a couple months. They are promising to have initial numbers on the economics. They’ve been tinkering with this idea for years and given they rebuilt the plant from the ground up (it’s not a pilot plant add on), I’d say this one won’t be cancelled last minute.
It may yet prove not to be viable relative to ‘post fracking gas’ which didn’t exist when we went down this road (kind of frightening if you think about that) …. but with spot prices for ‘nat gas’ hitting 35$/GJ in producing areas recently … who knows. In the very least we are going to have to decide if we want to cover the landscape with wind mills and power lines or install pipe lines 😀 …
There was also an interview with the CEO on BNN.


AllenC says:
February 16, 2014 at 6:34 am
If Wind and Solar are so “cheap”, why do tax payers have to subsidize them??

Exactly. And why the others are so heavily taxed?
Also, Nuclear looks expensive. I thought It was cheaper.


The 25% capacity value for PV they used demands 2190 sun hours per year. Or 6 sunhours a day.
Parts of Africa and parts of Australia will achieve that. And a few spots in the Andes:


Taliesyn says:
February 16, 2014 at 6:36 am
CCS is expensive.

And stupid.


“Advanced” seems an odd tag for technology. The next generation of power engineers will probably get a good laugh.

Rhys Jaggar

I”m not an expert on this, but I have read before that you can use solar farms to create gas, which is a despatch able source. So, in theory, you can build up a reservoir of gas before you ‘switch on’ the solar source, always maintaining a sufficient back up store of gas to make the source, in effect, despatchable.
I’ve no knowledge about the economics of that but has that option been investigated in your data??


Bob Green says “The true dispatchables are the peaking plants”
You are right in that the true dispatchable units are peaking plants, especially those that provide ancillary capacitance and then ramp up from zero almost instantaneously (typically very small plants).
However, all plants have a turn down and one thing not mentioned here is the initial install of wind generation in the MidWestern US destroyed a lot of coal plants which although equipped to move up and down, were never meant to move around in the way needed to off-set renewables. So in a regulated environment, there was a cost to wind generation that was just ignored … kind of like ignoring the cost a two footed driver has on your car.
Having had to suffer an old coal plant through this new dispatch regime, I am assured the new 800-1000 MW natural gas stations will be highly dispatchable with a sizeable turn down.
What I’m curious about as the power price normalizes to the price of burning Natural Gas plus efficiency losses, plus transmission, plus utility margins … how this plays out with many low efficiency plants being built as a hedge. My LED lighting might not be enough, I might have to properly install on site generation to mitigate the risk.

cgh says, “a few things in the EIA’s numbers… are simply not credible. First the capacity factors for wind. The EIA is projecting 34 and 37 per cent for onshore and offshore. This is simply not achievable.”
You’re right, cgh. The average capacity factor for all German wind turbines was only 17.5% in 2012.
Additionally, 25% is an unrealistically optimistic number for solar PV capacity factor. Germany has the most solar generating capacity of any country in the world, but its 2012 average capacity factor was just 10.5%. An Arizona site might double that, but there’s no way they’ll get to 25%.
Based on this 2012 report, the German nation-wide average wind capacity factor for 2012 was apparently just under 17.5% [45867 GWh/yr / (29.9 GW * 24 * 366) = 0.1746]. I.e., actual generated power was 17.5% of nameplate capacity, despite the fact that a lot of their windmills are pretty new, and electricity prices there are so extraordinarily high there (thanks to “green” politics) that there’s a strong incentive to keep the turbines well-maintained and running. Where electricity prices are lower, that incentive fades. I expect wind to blow, but that really sucks. Believe it or not, PV fared even worse — it’s average capacity factor in Germany in 2012 was only about 10.5%.
Caveat: There’s a new 2013 version of that German report out, but I’ve not yet redone the arithmetic.

Joe Chang

I am suspicious of the hydro cost. It should be heavily dependent on the location, so a single average cost has no real meaning. Carter tried to kill the water projects when he came in out office, of course that was when oil was $10.


drumphil says:
February 16, 2014 at 6:37 am
Most of the logic here is fine, but only if cost is the only thing that matters.
What if traditional electricity costs 5c/kWhr and renewable costs 20c/kwHr. DId you ever think that maybe in that extra 15c/Kwhr that 6-8 cents of it is to produce energy to create the renewable energy? It is VERY likely that we are using more fossil fuels to produce renewable than we are traditional, it is just hidden in manufacturing in some unknown faraway land.

Truth Disciple

Sorry Willis I don’t believe statistics generated by gob’mt apparatchiks. Over time I have found the newsletter “Access to Energy” most reliable. Maintenance on Wind Turbines must include the raptor and bat skeletons and blood of the 500K endangered species slaughtered by GE turbines.
I’m no environchondriac but these devices do not engender economic sense. And it is cheap energy that makes an economy great.
Chemical Engineering Chair, Lehigh University Professor John Chen, did an Energy Assessment for the globe some years ago. I suggest you read it because it stresses the importance of nuclear power. It implies we are wasting ENERGY & TIME on the alternative intellectual masturbations. It was written before the Balken and Marcellus NG finds. Available on his website or Lehigh’s.
In my humble opinion we will always have plentiful natural carbon energy.


I worked in power cost analysis for 25 years. Some may benefit from understanding that there are two different major components to power cost: fixed cost, and variable cost.
Fixed cost relates to the plant, boilers, turbines, the personnel, the lights in the plant, etc. Fixed cost remains whether any electricity is produced at all.
Variable cost is directly linked to energy production. It goes up and down with production. For example, the cost of coal burned is variable cost.
PV solar and wind have a very high fixed cost, but zero variable cost.
Beyond that, as len said, there is the additional consideration of transmission cost.


“I’m not believing that the maintenance on that monstrosity will cost only 30% more than dusting photovoltaic panels …”
You may be underestimating the cost of maintenance of solar farms Willis. Ahem-
Solar panels only output DC and the trend is for a micro-inverter on each panel so if one inverter goes down you don’t lose the output of many panels. Also inverters don’t last as long as the panels and so monitoring and replacement costs accrue.
I don’t see any security with wind turbines but I wouldn’t want to leave my solar panels within easy reach of sticky fingers in remote locations would you?
Not just a simple matter of dusting off the panels by all accounts-


Rhys Jaggar says:
February 16, 2014 at 6:59 am
“I”m not an expert on this, but I have read before that you can use solar farms to create gas, which is a despatch able source. So, in theory, you can build up a reservoir of gas before you ‘switch on’ the solar source, always maintaining a sufficient back up store of gas to make the source, in effect, despatchable.
I’ve no knowledge about the economics of that but has that option been investigated in your data??”
Such devices get tested in Germany. They electrolyse water, produce SynGas and turn that into Methane. The losses are about 50%. Critics say, as long as NatGas is available, this is just a method to lose energy.
In the graphic above, we should end up with 0.28 codt for energy produced via this detour.
Taking into account the solar capacity factor in Germany (800 sunhours a year, a third of what the EIA used) we would arrive at 0.84 .
It is probably still better than human slave labor.


glenncz says, “What if traditional electricity costs 5c/kWhr and renewable costs 20c/kwHr.”
Sorry, as I mentioned above. If we keep this fuel mix we are building out down to 20c/kwh in 2018, that would be fortunate. I would say 20% higher is likely and much higher possible.

Rud Istvan

Directionally correct. Intermittents are not dispatchable by definition. But neither is base load that is run flat out all the time for cost and efficiency. All nuclear and newer supercritical coal are base load. Dispatchables include peaking gas turbines, spinning reserves (older less efficient coal) and theoretically CCGT (but much more efficient and less expensive if run as base load.
The numbers are fudged with respect to capacity factor, especially solar. They are fudged with respect to lifetime. Supercritical coal is over 40, wind is under 20 due to gearbox failure ( and maintenance is understated). The results have been politicized.
Willis’ point about having to add dispatchables to renewables in like proportion remains the biggest fudge of all.


To really understand the build out of the grid, the Dispatchable sources need to be further subdivided into baseline and peaker units.
Baseline sources such as nuclear and coal take days to weeks to bring on line from a cold plant and need to run as near to full capacity as possible to run efficiently.
Peaker units can be taken on or off line in minutes.
To keep the grid balanced, baseline capacity should not exceed minimum load.

I cannot find the references now (so take this with huge grain of salt) but I thought I had read that the reason the wind mtx costs “look” so odd is due to how the mtx is contracted out.
Also a lot of the end of life mtx costs are either born by the taxpayers during decommissioning or the units are just left standing and inoperative.
so that MAY explain why those costs are so odd, because they are not true.

Steven Kopits

Some of these numbers look a bit suspect to me.
Nat gas power plans are typically intermediate and peaker capacity. Hence capacity factors around 25-50%, not 85%, are more common. I think this understates the cost of natural gas compared to coal. By the way, the last year has been characterized by gas to coal switching, ie, coal is gaining share.
Nuclear cost seems understated. The overnight (if you could build the plant over night) cost is quite competitive for nuclear, but environment and legal challenges tend to make the construction period protracted, and financing costs tend to pile up. These can add 50% to capex–I don’t know if such costs are properly captured here, but they don’t seem to be.
Hydro is a fantastic dispatchable resource. If can be turned on and off very quickly, although it does depend on adequate water in the respective reservoir. The EIA does note this in the footnotes, but hydro would ordinarily be considered a dispatchable resource. Also, hydro pairs very nicely with renewables, which are intermittent. Thus, when considering storage solutions for renewables, hydro is often presented as the preferred solution.
Offshore wind costs look about right at 22 cents per kWh. However, the total capacity which could feasibly be installed on the east coast (really, the DC to Boston corridor) is about 12 GW, representing 2,000-3,000 turbines. I think, in terms of physical spacing and proximity to load centers, I think we’d be challenged to put even this many in the water, but it’s theoretically possible. You can find more on this in our report prepared for DOE, with specific modeling assumptions starting after page 164.

Steven Kopits

And while we’re at it, Willis, here’s my presentation on oil and economy at the Center on Global Energy Policy at Columbia University last Tuesday. You should watch the whole thing, all 70 minutes. At a minimum, however, you should watch min 11-17.

Mike M

Willis: “I’m not believing that the maintenance on that monstrosity will cost only 30% more than dusting photovoltaic panels”
I agree and not only that cost, I’ve got one more fer ya that maybe hasn’t been figured in yet? Disaster/Mayhem/Doomsday insurance cost – (IF IT EXISTS??) Almost two years ago and ignored by MSM concerning a wind turbine erected near the north bank of the Mystic River (1000’s years of silt…) at the border of Everett and Charlestown, Mass:
1)”As reported last week, the new wind turbine in Charlestown has apparently sunk about 2 inches causing significant worry that the structural integrity of the 426,000 pound turbine is at risk.”
2)”The blades for the turbine were produced in China and sent by boat to the US and off-loaded in Chelsea and then trucked to Charlestown.”
As I see it, given that the MRWA was too stupid to pick an intelligent engineering firm to do the work correctly the first time – why should anyone believe they found an intelligent one to fix it?
In the disaster/mayhem/doomsday department there is, just across the street, a (?) MW natural gas fired electric plant by National Grid, (used to be called the “Edison” plant). Having this behemoth fall over (1) onto the plant in a hurricane or having a Chinese blade fly off (2) and hit critical components of the power plant certainly meets my description.
On top of that they erected it only a few hundred yards off the approach to Logan runway 15. With a maximum blade tip height (364 feet), and glide slope altitude at 2.5 miles away from “the numbers” of runway 15 about 700 feet. Not a lot of clearance either way IMO.
Are wind turbines required to be underwritten for calamity and, if so, I wonder how the coverage amount is calculated?
Also –
Supporting the EIA chart’s numbers concerning offshore wind power, back in 2010 National Grid agreed to pay 20.7 cents per KWH for power from the Cape Cod Cape Wind (which is still in limbo but will hopefully DIE a miserable death).

Turbines degrade rapidly with age due to operation in a harsh and unforgiving environment and due to difficulty of maintenance. See full details at REF: The Government chief scientific advisor tried unsuccessfully to refute the results of Prof Gordon Huighes.


You all realize that these “levelized cost” comparisons of competing types of electrical generation are misleading or meaningless due to government environmental mandates and regulations artificially driving up the operating costs of the dispatchable technologies while reducing the operating costs including financing of the non-dispatchable technologies. Remove the Obama EPA onerous and outrageous regulations on dispatchable technologies which drive up the costs, and wind solar etc become irrelevant.


This bloke knows how to keep an eagle eye on our electricity market-


I work at a utility that just bought another 100MW of wind energy. It was $.024 per kWh fixed price for 20 years. That is a firm fixed price where the plant owner takes all of the risks, includes all maintenance costs, etc. The subsidies help get them down to this level from .07, and the site is a very good wind resource, very cheap land, no wheeling, but the other part of it is that the big wind suppliers (NextEra and Duke) are very competitive. They have sweetheart deal on GE turbines, have their own gearbox repair factories, etc.
We’ve been working on permitting a new coal plant for about a decade now. It doesn’t look likely to ever get built. People laugh when you bring up nuclear.
Your only real choices for new generation are wind and gas. You build gas to keep the lights on and you buy wind because it is state mandated, is a hedge against inflation, and may help contain the cost of production if gas prices go up. And you do everything in your power to keep your existing coal plants running forever, because they are your only source of fuel diversity and you’ll never be able to rebuild them.


glenncz said:
“It is VERY likely that we are using more fossil fuels to produce renewable than we are traditional, it is just hidden in manufacturing in some unknown faraway land.”
Can you actually support that with solid figures? It takes a lot of work to figure such things out accurately. Can you point me at some decent papers that have dealt with this issue?