Mike Conley & Tim Maloney
April 17, 2015
(NOTE: This is a work in progress.
It will be a chapter in the forthcoming book
“Power to the Planet” by Mike Conley.)
Four bottom lines up front:
- It would cost over $29 Trillion to generate America’s baseload electric power with a 50 / 50 mix of wind and solar farms, on parcels of land totaling the area of Indiana. Or:
- It would cost over $18 Trillion with Concentrated Solar Power (CSP) farms in the southwest deserts, on parcels of land totaling the area of West Virginia. Or:
- We could do it for less than $3 Trillion with AP-1000 Light Water Reactors, on parcels totaling a few square miles. Or:
- We could do it for $1 Trillion with liquid-fueled Molten Salt Reactors, on the same amount of land, but with no water cooling, no risk of meltdowns, and the ability to use our stockpiles of nuclear “waste” as a secondary fuel.
Whatever we decide, we need to make up our minds, and fast. Carbon fuels are killing us, and killing the planet as well. And good planets are hard to come by.
If you think you can run the country on wind and solar, more power to you.
It’s an attractive idea, but before you become married to it, you should cuddle up with a calculator and figure out exactly what the long-term relationship entails.
This exercise has real-world application. The 620 MW (megawatt) Vermont Yankee nuclear reactor was recently shut down. So were the two SONGS reactors in San Onofre, which generated a combined total of 2.15 GWs (gigawatts). But the public didn’t suddenly go on an energy diet; in the wake of Fukushima, they were just more freaked out than usual about nuclear power.
Regardless, the energy generated by these reactors will have to be replaced, either by building more power plants or by importing the electricity from existing facilities.
To make the numbers easier to think with, we’ll postulate a 555 MW reactor that has an industry-standard 90% online performance (shutting down for refueling and maintenance) and delivers a net of 500 MW, sufficient to provide electricity for 500,000 people living at western standards. The key question is this:
What will it take to replace a reactor that delivers 500 MW of baseload (constant) power with wind or solar?
Once we’ve penciled out our equivalent wind and solar farms, we’ll be able to scale them to see what it would take to power any town, city, state or region—or the entire country—on renewables.
The ground rules.
TheSolutionProject.Org has a detailed proposal to power the entire country with renewables by 2050. It’s an impressive piece of work, presenting a custom blend of renewables tailored for each state, everything from onshore and offshore wind, to wave power, rooftop solar, geothermal, hydroelectric, the list goes on.
Costs are offset by the increased economic activity from building and operating the plants. Other major offsets derive from health care savings, increased productivity, lower mortality rates, reduced air pollution and global warming. But since these offsets also apply to an all-nuclear grid, they cancel themselves out.
Instead of exploring each technology the Solutions Project offers, we’ll simplify things and give them their best advantage by concentrating on their two major technologies—onshore wind and CSP solar (we’ll explain CSP shortly.) Both systems are at the low end of the long-term cost projections for renewables.
In our comparative analysis, we’ll be focusing on seven parameters:
- CO2 (from material production and transport)
- Land area
- Deathprint (casualties from power production)
- Carbon karma (achieving CO2 break-even)
- Construction cost
Most of these are obvious, but “deathprint” and “carbon karma” deserve a bit of explaining. We’ll get into the first one now, and save the other one for later.
No form of energy production is, or ever has been, completely safe. Down through the centuries, countless people have been injured and killed by beasts of burden. More were lost harvesting the wood, peat and whale oil used for cooking, heating, and lamplight. Millions have died from mining coal, and millions more from burning it. America loses 13,000 people a year from health complications attributed to fossil fuel pollution; China loses about 500,000.
Although hydroelectric power is super-green and carbon-free, we too easily forget that in the last century alone, many thousands have died from dam construction and dam failures. Even solar energy has its casualties. In fact, more Americans have died from installing rooftop solar than have ever died from the construction or use of American nuclear power plants. Some people did die in the early days of uranium mining, but the actual cause was inhaling the dust. Proper masks lowered the casualty rates to nearly zero.
Although reactors produce nearly 20% of America’s power, and have been in use for over fifty years, there have been just five deaths from construction and inspection accidents. Only three people have ever died from the actual production of American atomic energy, when an experimental reactor suffered a partial meltdown in 1961. And for all the panic, paranoia, and protests about Three Mile Island, not one person was lost. The worst dose of radiation received by the people closest to the TMI plant was equal to one half of one chest X-ray.
As we contrast and compare the facts and figures for a wind farm, a solar farm, and a reactor, we’ll cite each technology’s “deathprint” as well—the casualties per terawatt-hour (TWh) attributed to that energy source.
[NERD NOTE: A terawatt is a trillion watts. The entire planet’s electrical consumption is right around 5 terawatt-hours. One TWh (terawatt-hour) is a constant flow of a trillion watts of electricity for a period of one hour.]
“Any way the wind blows, doesn’t really matter to me.” — Freddy Mercury
Well, it should. Wind power is all about direction and location. The problem is, climate change may also be changing long-term wind patterns. The polar vortex in the winter of 2013 might be a taste of things to come. Large-scale wind farms could prove to be a very expensive mistake, but we’ll look at them anyway.
At first frostbitten blush, a freight train of Arctic air roaring through the Lower 48 seems to fly in the face of global warming, doesn’t it? But here’s how it works:
Since the Arctic is warming faster than the rest of the world, its air mass is becoming less distinct than Canada’s air mass. This erodes the “thermal wall” of the Jet’s Stream’s arctic corridor, and it’s starting to wander like a drunk, who can usually navigate if he keeps his hand on the wall. But now the wall is starting to disappear, and when it finally goes it’s anyone’s guess where he’ll end up next.
In North America, the median “capacity factor” for wind is 35%.
Some places in America are a lot more windacious than others. But on average, the wind industry claims that a new turbine on U.S. soil will produce around 35% of the power rating on the label, meaning it has a “35% capacity factor.”
One difficulty in exploring renewables is that capacity factor numbers are all over the map. The Energy Information Agency disagrees with the Department of Energy, and the renewables industry disagrees with them both. Manufacturers stay out of the fray, only stating what their device’s “peak capacity” is, meaning the most power it can produce under ideal conditions. Your mileage may vary.
Because wind, like solar, is an “intermittent” source (ebbs and flows, comes and goes) the efficiency of a turbine has to be averaged over the course of a year, depending on where it’s used. But we’ll accept the wind industry’s claim of 35% median capacity factor for new onshore turbines sited in the contiguous states.
And we won’t stop there. Because if we actually do build a national renewables infrastructure, it stands to reason that we’ll concentrate our wind farms where they’ll do the most good, and build branch transmission lines to connect them to the grid. Since the industry claims a maximum U.S. capacity factor of 50% for new turbines and a median of 35%, we’ll split the difference at a generous 43%.
To gather 500 MWavg (megawatts average) of wind energy in a region with a 43% capacity factor (often called “average capacity”), we’ll need enough turbines for a peak capacity of 1,163 MWp (megawatts peak): 500 ÷ 0.43 = 1,163.
Let’s go with General Electric’s enormous model 2.5xl turbines, used at the Shepherd’s Flat wind farm in Oregon, a top-of-the-line machine with a peak capacity of 2.5 MW. That pencils out to 465 “spinners” (1,163 ÷ 2.5 = 465.)
Each assembly is made with 378 tonnes of steel, and the generator has a half-tonne of neodymium magnets, a rare earth element currently available only in China, where it’s mined with an appalling disregard for the environment and worker safety. And, the 300-ft. tower requires a concrete base of 1,080 tonnes.
[NERD NOTE: A “tonne” is a metric ton, which is 1,000 kilograms—2,204.62 lbs to be exact. And no, it’s not pronounced “tonnie” or “tonay.” A tonne is a ton.]
The installed cost of a GE 2.5xl is about $4.7 Million, which includes connecting it to the local grid. That breaks down to $1.9 Million per MWp.
In this exercise, we’re not factoring in the cost of the land, or the cost of a branch transmission line if our renewables farm isn’t next to the grid. But figure about $1 Million a mile for parts and labor to install a branch line, plus the land.
Renewables, like most things, have their own CO2 footprint.
Steel production emits 1.8 tonnes of CO2 per tonne, and concrete production emits 1.2 tonnes of CO2 per tonne. So just the raw material for GE’s 2.5xl turbine alone “costs” 1,976 tonnes of CO2 emissions. [(378 X 1.8) + (1,080 X 1.2) = 1,976.4]
We’ll give them a pass on the CO2 emitted during parts fabrication and assembly, but we really should include the shipping, because these things weigh in at 378 tonnes. And, the motors are made in China and Germany, the blades are made in Brazil, they do some assembly in Florida, and the tower sections are made in Utah. That’s a lot of freight to be slinging around the planet.
But to keep things simple, and to be more than fair, we’ll just figure on shipping everything from China to the west coast, and write off all the CO2 emissions from fabrication and assembly, and the land transportation at both ends. So 378 tonnes at 11 grams of CO2 (equivalent) per ton-mile, shipped 5,586 miles from Shanghai to San Francisco, comes out to 23.2 tonnes per turbine.
Even though we’re not calculating the price of the land, we will be adding up the amount of acreage. Turbines need a lot of elbowroom, because they have to be far enough away from each other to catch an undisturbed breeze. It can be difficult to realize how huge these things are—imagine a 747 with a hub in its belly, hanging off the roof of a 30-story building and spinning like a pinwheel.
Each turbine will need a patch of land 0.23 / km2 (square kilometers), or 550 yards on a side. A rough rule of thumb is to figure on four large turbines per square kilometer, or ten per square mile. But before we put the numbers together, there are two more things to consider.
Wind and solar farms are gas plants.
Don’t take our word for it; listen to this guy instead, one of the most famous voices in the renewable energy movement:
“We need about 3,000 feet of altitude, we need flat land, we need 300 days of sunlight, and we need to be near a gas pipe. Because for all these big solar plants—whether it’s wind or solar—everybody is looking at gas as the supplementary fuel. The plants we’re building, the wind plants and the solar plants, are gas plants.” – Robert F. Kennedy, Jr., board member of BrightSource, builders of the Ivanpah solar farm on the CA / NV border.
Large wind and solar farms are in the embarrassing position of having to use gas-fired generators to smooth out the erratic flow of their intermittent energy. It’s like showing up at an AA meeting with booze on your breath.
Still, it’s considered a halfway decent solution, but only because wind and solar contribute such a small proportion of the energy on the grid. But if renewables ever hope to be more than 15% of our energy picture, they’ll have to lose the training wheels, and there’s only one way to do it. Which brings us to the other thing we need to consider. And this one is a deal-breaker all by itself.
For the wires to sing, you need a choir of generators humming away in perfect harmony. And for intermittent energy farms to join the chorus as full-fledged members, they’ll first have to store all the spurts and torrents of energy they produce, and then release it in a smooth, precisely regulated stream.
Right now, the stuttering contributions that residential solar or the occasional renewables farm feed the grid are no problem. It’s in such small amounts that the “noise” it generates isn’t noticeable. The amount of current on the national grid is massive in comparison, generated by thousands of finely tuned turbines at our carbon-fuel, nuclear, and hydro plants. These gargantuan machines operate 24 / 7 / 365, delivering a rock-solid stream of AC power at a smooth 60Hz.
That’s baseload power, and every piece of gear we have—from Hoover Dam to your doorbell—is designed to produce it, convey it, or run on it. Our entire energy infrastructure has been built around that one idea. Choppy juice simply won’t do.
(For a more detailed explanation of why this so, please see our article “We’re Not Betting the Farm, We’re Betting the Planet.“)
For renewables to be a major player and replace carbon and nuclear fuels, they’ll have to deliver the same high-quality energy, day in and day out. Up to now, computerized controls haven’t been able to smooth out the wrinkles, because the end result of all of their highfalutin calculations comes down to engaging or disengaging mechanical switches. And mechanical switches aren’t nearly as precise as the computers that run them, because they’re made out of metal, which expands and contracts and wears down. Unless this technology is perfected (and it’s a lot harder than it sounds), glitches will resonate through the grid, and with enough glitches we won’t have baseload power, we’ll have chaos.
So while a national renewables infrastructure will have to be built on free federal acreage—the amount of land required is nearly impossible to wrap your mind around, and paying for it is completely out of the question—the cost of energy storage needs to be factored into any grid-worthy plant.
Remember, we’re replacing a reactor. They crank it out day and night, rain or shine, for months at a stretch, with an average online capacity of 90% after shutdowns for refueling and maintenance are factored in. If a renewables farm can’t provide baseload power, it’ll be just another expensive green elephant on the greenwash circuit.
Pumped-Hydro Energy Storage (PHES).
By far, the most cost-effective method of producing baseload power from intermittent energy is with pumped hydro. It’s an idea as simple as gravity: Water is pumped uphill to an enormous basin, and drains back down through precisely regulated turbines to produce a smooth, reliable flow of hydroelectricity.
Thus far, most pumped-hydro systems have used the natural terrain, connecting a high basin with a lower one. Dams that have been shut down by drought or other upstream conditions can also be used. Watertight abandoned mines and quarries, or any large underground chambers at different elevations have potential as well. But if nothing’s readily available, one or both basins can be built. And if we go big on wind and solar, we’ll likely be building a lot of them.
A “closed-loop” PHES has a basin at ground level connected by a series of vertical pipes to another basin deep underground. When energy is needed, water drops through the pipes to a bank of generators below, then collects in the lower basin. Later, when energy production is high and demand is low, the surplus energy is used to pump the water back upstairs.
It sounds great, but the amount of water needed is mind-boggling. To understand why, here’s a rundown of the basic concepts underlying hydroelectric power.
Good old H2O.
The metric system is an amazing, ingenious, brilliant, and stupid-simple method of measurement based on two everyday properties of a common substance that are exactly the same all over the world: the weight and volume of water.
One cubic meter (m3) of pure H2O = one metric ton (~ 2,200 lbs) = 1,000 kilograms = 1,000 liters. And one liter = 1 kilogram (~ 2.2 lbs) = 1,000 grams = 1,000 cm3 (cubic centimeters.) And one cm3 of water = one gram, hence the word “kilogram,” which means 1,000 grams. And a tonne is a million grams.
You may have already deduced that metric linear measurements are related to the same volume of water: A meter is the length of one side of a one-tonne cube of water, and a centimeter is the length of one side of a one-gram cube of water.
Metric energy measurements are based on another thing that’s exactly the same all over the world: the force of falling water. One cubic centimeter (one gram) of water, falling for a distance of 100 meters (about 378 feet) has the energy equivalent of right around one “joule” (James Prescott Joule was a British physicist and brewer in the 1800s who figured a lot of this stuff out.)
One joule per second = one watt. (Energy used or stored over time = power. A joule is energy, a watt is power.) A million grams (one tonne) falling 100 meters per second = a million joules per second = a million watts, or one megawatt (MW). One MW for 3,600 seconds (one hour) = one MWh (megawatt-hour.)
They don’t call this a water planet for nothing.
Which brings us back to Pumped-Hydro Energy Storage.
To store one hour’s worth of energy produced by a 500 MW wind farm, we’ll need to drop 500 metric tonnes (cubic meters) of water each second for an entire hour, down a series of 100-meter-long pipes, to spin a series of turbines at the bottom of the drop. (For right now, we’ll leave out the loss of energy due to friction in the pipes, and the less-than-perfect efficiency of the turbines.)
That’s 1,800,000 tonnes per hour, which is a lot of water. How much, exactly? About twice the volume of the above-ground portion of the Empire State Building, which occupies 1.04 million cubic meters of space (if you throw in the basement.)
Remember, that’s for just one hour of pumped-hydro. To pull it off, our wind farm will need two basins, each one the volume of two Empire State Buildings (!), with a 100-meter drop in elevation between them. And, the basins will have to be enclosed to minimize evaporation.
Two ESBs (Empire State Buildings) is a huge volume of water to devote to one hour of energy storage, particularly when we might be entering a centuries-long drought induced by climate change. Replenishing our water supply because of evaporation won’t be an easy option, and will likely annoy the locals, who will probably be fighting water wars with the folks upstream.
Sorry, no free lunch. Wrong universe.
Converting one form of energy to another always results in a loss, and pumped hydro systems can consume nearly 25% of the energy stored in them. But we’ll be generous and figure on 20%. That still means we have to grow our 465-turbine wind farm to 581 turbines to get the output we need.
And remember, we’re just storing one hour of power. If our wind farm gets two hours of dead calm, we’re out of luck. And two hours of dead calm is nowhere near uncommon. But with a national renewables energy grid, maybe we can import some solar energy from Arizona. Maybe. Unless it’s cloudy in Arizona, or it’s after sundown.
Sigh... When you start thinking it through, it’s becomes pretty clear that you have to figure on at least one full day of storage. Some people will tell you to figure on a week, but as you’ll see, even one day is enough to fry your calculator.
The DoE estimates that closed-loop pumped storage should cost about $2 Billion for one gigawatt-hour, or $2 Million per megawatt-hour. First we’ll add the extra turbines, and then we’ll throw in the PHES. (Are you sitting down?)
A 500 MWavg baseload wind farm with Pumped-Hydro Energy Storage.
To get 500 MWavg in a region with 43% average capacity, we’ll need 465 turbines with a 2.5 MW peak capacity: [(500 ÷ 2.5) = 200. (200 ÷ 0.43) = 465].
On top of that, we’ll need to compensate for the 20% energy loss to pumped-hydro storage, so we’ll need a grand total of 581 turbines (465 ÷ 0.80 = 581.)
- Steel ………………………………………… 219,618 tonnes
- CO2 from steel …………………………… 395,312 t
- Concrete …………………………………… 627,480 t
- CO2 from concrete ……………………… 752,976 t
- CO2 from shipping ……………………… 29,951 t
- CO2 estimate for PSH …………………. 1 Million t
- Total CO2 ………………………………….. 2.17 Million t (see below)
- Land (0.23 km2 / MWp) ……………….. 119 km2 (10.9 km / side) 46 sq. miles (6.78 mi / side)
- Deathprint …………………………………. 0.15 deaths per TWh
- Carbon karma ……………………………. 181 days (see below)
- Turbines (581 X $4.7 M) ……………… $2.7 Billion
- PHES (500MW X 24hrs X $2M) …… $24 Billion
- Total cost ………………………………….. $26.7 Billion
Carbon Karma — achieving the serenity of CO2 break-even.
The entire point of a renewables plant is to make carbon-free energy. But it will “cost” us at least 1.17 Million tonnes of CO2 just to get our turbines built and shipped. And remember, that doesn’t include the CO2 of fabrication, assembly, and the land transport at both ends.
Depending on local conditions, we could get lucky and use an old mine or quarry, or dam up a mountain hollow. But we should figure at least another 1 million tonnes of CO2 in the material and construction of the PHES: Two steel-reinforced concrete basins stacked on top of each other, 350 meters deep and 350 meters on a side, with the floor of the lower one 800 meters underground, plus the 100-meter drop pipes to connect them, with turbines at the bottom of the drop. Plus the diesel fuel needed to excavate and build it.
Burning coal for energy emits about 1 metric ton of CO2 per MWh (megawatt-hour) of energy produced. Since our wind farm will be cranking out 500 clean MWs, it won’t be releasing the 500 tonnes of CO2 / hr normally emitted if we were burning coal. Then again, it took about 2.17 Million tonnes of CO2 emissions to get the place up and running, which is nothing to sneeze at.
To pay off this carbon-karma debt, our wind farm will have to make merit by producing carbon-free energy for at least 4,320 hours, or 181 days. (2.17 Million tonnes of CO2 ÷ 12,000 tonnes per day saved by 500MW of clean energy production = 180.83) Sounds pretty good, until you see how fast a 500 MW reactor redeems itself.
“Direct your feet to the sunny side of the street.” — Louis Armstrong
A good song to live by. Except there’s a good chance that, just like our wind farm, our solar farm will be miles from any street or highway. Like wind, solar needs lots of land, and the cheaper the better. Free is better than cheap, but that means it’ll probably be a bleak patch of federal wilderness 50 miles from nowhere.
In North America, the capacity factor for PV (photo-voltaic) solar panels averages 17% of the peak capacity on the label, due to things like latitude, the seasonal angle of the sun, clouds, and nighttime. Dust on the panels can lower the average to 15%. But we’ll be using a much better technology than PV solar.
Sunshine in a straw.
We’ll model our solar farm after the 150 MWp (megawatts peak) Andasol station in Andalusia, Spain. Its Concentrated Solar Power (CSP) technology is far more efficient and cost-effective than PV panels, and uses just a fraction of the land. Instead of flat panels with photo-electric elements, Andasol has racks of simple parabolic trough mirrors (“sun gutters”) that heat a pipe suspended in the trough, carrying a 60/40 molten salt blend of sodium nitrate and potassium nitrate.
Andasol claims a whopping 41% capacity factor due to their high altitude and semi-arid climate, but it’s actually 37.7%. They say they have a 150 MWp farm that produces a yearly total of 495 GWh, so who do they think they’re fooling?
[NERD NOTE: 150 MWp X 8,760 hrs a year = 1,314 GWh. 495 ÷ 1,314 = 0.3767, or 37.67%. So there.]
But aside from that bit of puffery, they do have a good system, and a big factor is the efficiency of their molten salt heat storage system. Costing just 13% of the entire plant, the storage system can generate peak power for 7.5 hrs at night or on cloudy days. And remember, Andasol’s peak power is 150MW.
This means that in a pinch, they can deliver up to 83% of their daily average capacity from storage alone. (37.7% of 150 MWp = 56.5 MWavg / hr. 56.5 MW X 24 hrs = 1,357 MWavg / day. 150 MWp X 7.5 hrs = 1,125 MW. 1,125 ÷ 1,357 = 0.829, or 83%.) What this also means is that the molten salt storage concept can be exploited to produce baseload power.
The Andasol plant is compact, as far as solar installations go: Using 162.4 t of steel and 520 t of concrete per MWp, the $380 Million (USD) facility produces 56.5 MWavg from 150 MWp on just 2 square kilometers of sunbaked high desert. That’s $2.53 Million per MWp, or about $6.85 Million per MWav.
But since we want to produce true baseload power, we’ll need to re-think the system. Heat storage is all well and good for “load balancing,” which is meant to to smooth out the dips and bumps of production and demand over the course of several hours. But heat dissipates—you either use it or lose it—and baseload is a 24-hour proposition. So there’s a point of diminishing returns for molten salt heat storage, and Andasol figured that 7.5 hrs was about as far as they could push it. We’ll take their advice, and proceed from there.
Producing 500 MW baseload with Concentrated Solar Power.
We’ll have to put all the energy we generate into storage, staggering the feed-in from sunup to sundown. To do this, we’ll have to grow the plant by 3.2 times (24 hrs ÷ 7.5 = 3.2). Like our pumped-storage wind farm, our CSP energy will be distributed from storage at a steady 500 MW of baseload power, with a 24-hr “margin” of continuous operation—meaning if we know we’ll be offline because a big storm is coming in, the masters of the grid will have 24 hours to line up another producer who can fill in. With enough baseload renewables plants in enough regions of the country, 24 hours will (hopefully) be sufficient.
Although solar capacity in the U.S. averages 17%, it’s a dead certainty that if we actually do go with a national renewables infrastructure, we’ll put CSP plants in the southwest deserts where they’ll do the most good. And if some of them end up 50 miles from nowhere, it’ll just be another $50 million a pop (not counting the transmission corridor) to hook them into the grid. Which is chump change, given the overall price tag.
The California deserts have a CSP capacity factor of 33%, so let’s roll with that. Remember, Andasol is high desert, and most of our deserts are at low elevation, with thicker air for the sun to punch through. But the USA is still CSP country.
A 500 MWavg baseload CSP system.
At 33% average capacity, we’ll need 1,515 MWp of CSP (500 ÷ 0.33 = 1,515). Then we grow the plant by 3.2 X to get 24-hour storage, for a total of 4,848 MWp.
- Steel ………………………………………….. 787,315 tonnes
- CO2 (from steel) …………………………… 1.42 Million t
- Concrete …………………………………….. 2.52 Million t
- CO2 (from concrete) ……………………… 3.02 Million t
- Total CO2 ……………………………………. 4.44 Million t
- Land: (0.013 km2 / MWp X 4,848)……. 63 km2 (7.9 km / side)
24.3 sq. miles (4.9 mi / side)
- Deathprint …………………………………… 0.44 deaths per TWh (for solar)
- Carbon karma ……………………………… 370 days
- Cost (4,848 X $2.53 M / MWp) ………. $12.3 Billion
It’s less than one-third the cost of wind, but it’s still enough to make you…
Instead of a budget-busting renewables farm that takes up half the county, we could go with a Gen 3+ reactor instead, such as the advanced, passively safe Westinghouse AP-1000 Light Water Reactor (LWR). Two are under construction in Vogtle, GA for $7 Billion apiece.
Four more are under construction in China. We won’t really know what the Chinese APs will cost until they cut the ribbons, but it’ll certainly be a fraction of our cost, because they’re not paying any interest on the loan, or any insurance premiums, or forking over exorbitant licensing and inspection fees.
They also don’t have to deal with long and pricey delays from lawsuits, protests, and the like. Which don’t just cost a fortune in legal fees; you also get eaten alive paying interest on the loan. So the Chinese are going to find out what it actually costs to just build one. And that will be a very interesting and meaningful number.
With 90% online performance, the 1,117 MWp AP-1000 produces 1,005 MWavg of baseload power. And since the AP has scalable technology, the parts and labor for a mid-size AP should be roughly proportional.
Installing a new 555 MWp / 500 MWavg Gen 3+ Light Water Reactor.
The AP-1000 requires 58,000 tonnes of steel and 93,000 tonnes of concrete. Cutting that roughly in half, our “AP-500” will need:
- Steel …………………………………….. 28,818 tonnes
- CO2 from steel ………………………. 51,872 t
- Concrete ………………………………. 46,208 t
- CO2 from concrete …………………. 55,450 t
- Total CO2 ……………………………… 107,322 t
- Land (same as AP-1000) ………… 0.04 km2 (200 meters / side)
0.015 sq. miles (about 8 football fields)
- Deathprint …………………………….. 0.04 deaths per TWh
- Carbon karma ……………………….. 9 days
- Cost ($7.27 Million X 555) ……… $4.03 Billion
We’ve been cuddled up with a calculator, thinking about whether to go with a 500 MW Light Water Reactor, or a 500 MW wind or solar farm.
So far, wind is weighing in at $26.7 Billion, CSP solar at $12.3 Billion, and a Gen-3+ Light Water Reactor at $4.03 Billion. The land, steel and concrete for the reactor is minuscule, the material for wind or solar is substantially more, and the land for the wind farm is enough to make you faint.
But wait, it gets worse…
A reactor has a 60-year service life. Renewables, not so much.
The industry thinks that wind turbines will last 20-25 years, and that CSP trough mirrors will last 30-40 years. But no one really knows for sure: the earliest large-scale PV arrays, for example, are only 15 years old, and CSP is younger than that. And there’s mounting evidence that wind turbines will only last 15 years.
Of course, when the time comes they’ll probably just replace the generator, not the entire contraption. And to refresh a CSP farm, they’ll probably just swap out the mirrors, and maybe the molten salt pipes, and use the same racks. And we should assume that all the replacement gear will be better, or cheaper, or both.
So out of an abundance of optimism, and an abiding faith in Yankee ingenuity, let’s just tack on another 50% to extend the life of our renewables to 60 years.
Putting it all in perspective.
For a baseload 500 MWavg power plant with a 60-year lifespan, sufficient to provide electricity for 500,000 people living at western standards:
- Wind: 119 km2 ……….. two-thirds of Washington, DC
- CSP: 63 km2 …………… one-third of Washington, DC
- Nuclear: 0.04 km2 ……. one-half of the White House grounds
(0.03% of wind / 0.06% of CSP)
- Wind ……………………… 0.15 deaths / TWh
- CSP ………………………. 0.44 deaths / TWh
- Nuclear ………………….. 0.04 deaths / TWh
(26% of wind / 9% of solar)
- Wind ………………………. 181 days
- CSP ………………………. 370 days
- Nuclear ………………….. 9 days
(7.6% of wind / 3.3% of CSP)
- Wind …………………….. $40 Billion (nearly 10 X nuclear)
- CSP ……………………… $18.5 Billion (over 4.5 X nuclear)
- Nuclear …………………. $ 4.03 Billion
(10% of wind / 22% of CSP)
One step at a time.
Granted, $4.03 Billion is still a hefty buy-in. But power companies will soon be able to buy small factory-built reactors one at a time, and gang them together to match the output of a large reactor. These new reactors will be walk-away safe, with a 30-year fuel load for continuous operation—think “nuclear battery.” Welcome to the world of Small Modular Reactors (SMRs.)
Over the next decade, several Gen-3+ and Gen-4 SMRs are coming to market. The criteria for Gen-4 reactors are a self-contained system with high proliferation resistance, passively cooled, and a very low waste profile. Most Gen-4s won’t need an external cooling system, which requires access to a body of water. They’ll be placed wherever the power is needed, even in the harshest desert.
For a lower buy-in and a much faster start-up time, you’ll be able to install an initial SMR and roll the profits into the next one, building your plant in modular steps and reaching your target capacity as fast, if not faster, than building one big reactor. And you’re producing power for your customers every step of the way.
So instead of securing a loan for $4+ Billion and constructing a single, massive reactor like a hand-built, one-of-a-kind luxury car, you could be up and running with a small mass-produced $1 Billion reactor instead, with perhaps 20% of the output, delivered and installed by the factory. And as soon as you’re in the black, just get another one.
The daunting thing about building a large power plant is more than just the eye-popping buy-in. It’s also the long, slow march through the “Valley of Death”—that stretch of time (it could be years, even decades) when you’re hemorrhaging money and not making a profit, which makes you far more vulnerable to lawsuits, harassments, protests and other delays.
Going big — a carbon-free national energy infrastructure.
A robust power grid would be modeled after the Internet—a network of thousands of right-sized, fully independent nodes. If one node is down, business is simply routed around it. And within these nodes are smaller units that can also stand on their own, interacting with the local area as well as the national system.
Small Modular Reactors can be sited virtually anywhere, changing our grid in fundamental ways—if one reactor needs to be shut down, the entire power plant doesn’t have to go offline. Behemoth power plants, their transmission corridors marching over vast landscapes, will no longer serve as kingpins or fall like dominos. Once a top-down proposition for big players, baseload power will become distributed, networked, local, independent, reliable, safe and cheap.
Aside from the mounting threat of global warming, the productivity and lives lost from rolling blackouts is immense, and will surely get worse with business-as-usual. Ad as our population continues to expand, whatever energy we save will quickly be consumed by even more energy-saving gadgets.
Poverty and energy scarcity strongly correlate, along with poor health and poor nutrition. Unless we start desalinating the water we need, shooting wars will soon be fought over potable water. Energy truly is the lifeblood of civilization.
A word or two about natural gas.
Gas-fired plants are far less expensive than nuclear plants, or even coal plants, which typically go for about $2 an installed watt. Nuclear plants, even in America, could be as cheap as coal plants if the regulatory and construction process were streamlined—assembly-line fabrication alone will be an enormous advance. Still, a gas plant is about a third the price of a coal plant, which sounds great. But the problem with a gas-fired plant is the gas.
CO2 emissions from burning “natural gas” (the polite term for “methane”) are 50% less than coal, which is a substantial improvement, but it’s still contributing to global warming. It’s been said that natural gas is just a slower, cheaper way to kill the planet, and it is. But it’s even worse than most folks realize, because when methane escapes before you can burn it (and any gas infrastructure will leak) it’s a greenhouse gas that’s 105 times more potent than CO2. (If it’s any consolation, that number drops to “only” about 20 times after a few decades.)
Another problem with natural gas is that it’s more expensive overseas. Which at first glance doesn’t seem like much of a problem, since we’ve always wanted a cheap, abundant source of domestic energy. But once we start exporting methane in volume (the specialized ports and tankers are on the drawing board), why would gas farmers sell it here for $3 when they can sell it over there for $12?
A final note on natural gas: Even if all of our shale gas was recoverable (which it’s not), it would only last 80-100 years. But we have enough thorium, an easily mined and cheaply refined nuclear fuel, to last for literally thousands of years.
Natural gas is a cotton candy high. The industry might have 10 years of good times on the horizon, but I wouldn’t convert my car if I were you. Go electric, but when you do, realize that your tailpipe is down at the power plant. So insist on plugging into a carbon-free grid. Otherwise you’ll just be driving a coal burner.
Which brings us back to nuclear vs. renewables, the only two large-scale carbon-free energy sources available to us in the short term. And since all we have is the short term to get this right, we’d better knuckle down and make some decisions.
America has 100 nuclear power plants. We need hundreds more.
Reactors produce nearly 20% of America’s electrical power, virtually all of it carbon-free. And if you’re concerned about the proliferation of nuclear weapons, it may interest you to know that for the last 25 years, half of that power has been generated by the material we recovered from dismantling Soviet nuclear bombs. (And just so you know, power reactors are totally unsuited for producing weapons-grade material, and the traces of plutonium in their spent fuel rods is virtually impossible to use in a weapon. But that’s the subject for another paper.)
Many of our reactors are approaching retirement age, and lately there’s been some clamor about how to replace them. The top candidates—other than a new reactor—are natural gas and renewables. (Nobody’s a big fan of coal, except the coal company fat cats and the folks in the field doing the hard work for them. And of course their lobbyists.)
If the foregoing thicket of numbers hasn’t convinced you thus far, or if you’re still just fundamentally opposed to nuclear energy, let’s apply the numbers to the national grid. Let’s see what it would take to shut down every American reactor, like they shut down Vermont Yankee and San Onofre, and replace them all with wind and solar. And just for fun, we’ll also swap out our fossil fuel power plants, until the entire country is running on clean and green renewables.
A refresher on the ground rules.
TheSolutionsProject.Org has a buffet of renewables that they’ve mixed and matched, depending on the availability of renewable energy in each state. But keep in mind that onshore wind and CSP solar are two of the lowest-cost technologies in their tool kit, and that the actual renewables mix for any one state will probably be more complex—and more expensive—than what we’ll be laying out in the next section.
Thus far, we’ve bent over backwards to give renewables every advantage, from average capacity numbers to CO2 estimates to pumped-hydro efficiency to equipment replacement costs. Projecting how the entire country can run on wind and solar alone is simply an exercise for ballpark comparisons. Your mileage will definitely vary, and probably not in a way you would like.
“Let me live that fantasy.” — Lourde
So after all we’ve been through together, you would still prefer to run the country on wind and solar? Well, okay, then let’s run the numbers and see what it takes.
America’s coal, gas, petroleum and nuclear plants generate a combined baseload power of 405 GWavg, or “gigawatts average.” (Remember, a gigawatt is a thousand megawatts.) Let’s replace all of them with a 50 / 50 mix of onshore wind and CSP, and since our energy needs are constantly growing, let’s round up the total to 500 GWs, which is likely what we’ll need by the time we finish a national project like this. Some folks say that we should level off or reduce our consumption by conserving and using more efficient devices, which is true in principle. But in practice, human nature is such that whatever energy we save, we just gobble up with more gadgets. So we’d better figure on 500 GWs.
To generate this much energy with 1,000 of our 500 MW renewables farms, we’ll put 500 wind farms in the Midwest (and hope the wind patterns don’t change…) and we’ll put 500 CSP farms in the southwest deserts—all of it on free federal land and hooked into the grid. Aside from whatever branch transmission lines we’ll need (which will be chump change), here’s the lowdown:
Powering the U.S. with 500 wind and 500 CSP farms, at 500 MWavg apiece.
- Steel ……………….. 503 Million tonnes (5.6 times annual U.S. production)
- Concrete ………….. 1.57 Billion t (3.2 times annual U.S. production)
- CO2 …………………. 3.3 Billion t (all U.S. passenger cars for 2.5 years)
- Land ………………… 91,000 km2 (302 km / side)
35,135 sq. miles (169 mi / side)
(the size of Indiana)
- 60-year cost ……… $29.25 Trillion
That’s 29 times the 2014 discretionary federal budget.
If we can convince the wind lobby that they’re outclassed by CSP, we could do the entire project for a lot less, and put the whole enchilada in the desert:
Powering the U.S. with 1,000 CSP farms, producing 500 MWavg apiece.
- Steel ………………. 787 Million t (1.6 times annual U.S. production)
- Concrete …………. 2.52 Billion t (5.14 times annual U.S. production)
- CO2 ………………… 3.02 Billion t (all U.S. passenger cars for 2.3 years)
- Land ……………….. 63,000 km2 (251 km / side)
24,234 sq. miles (105.8 mi / side)
(the size of West Virginia)
- 60-year cost ……. $18.45 Trillion
That’s to 18 times the 2014 federal budget.
Or, we could power the U.S. with 500 AP-1000 reactors.
Rated at 1,117 MWp, and with a reactor’s typical uptime of 90%, an AP-1000 will deliver 1,005 MWav. Five hundred APs will produce 502.5 GWav, replacing all existing U.S. electrical power plants, including our aging fleet of reactors.
The AP-1000 uses 5,800 tonnes of steel, 90,000 tonnes of concrete, with a combined carbon karma of 115,000 t of CO2 that can be paid down in less than 5 days. The entire plant requires 0.04km2, a patch of land just 200 meters on a side, next to an ample body of water for cooling. (Remember, it’s a Gen-3+ reactor. Most Gen-4 reactors won’t need external cooling.) Here’s the digits:
- Steel ………. 2.9 Million t (0.5% of W & CSP / 0.36% of CSP)
- Concrete … 46.5 Million t (3.3% of W & CSP / 1.8% of CSP)
- CO2 ……….. 59.8 Million tonnes (2% of W & CSP / 1.5% of CSP)
- Land ………. 20.8 km2 (4.56 km / side) (0.028% W & CSP / 0.07% of CSP)
1.95 sq. miles (1.39 miles / side)
(1.5 times the size of Central Park)
- 60-year cost ……… $2.94 Trillion
That’s 2.9 times the 2014 federal budget.
Small Modular Reactors may cost a quarter or half again as much, but the buy-in is significantly less, the build-out is much faster (picture jetliners rolling off the assembly line), the resources and CO2 are just as minuscule, and they can be more widely distributed, ensuring the resiliency of the grid with multiple nodes.
Or for just $1 Trillion, we could power the entire country with MSRs.
The Molten Salt Reactor was invented by Alvin Weinberg and Eugene Wigner, the same Americans who came up with the Light Water Reactor (LWR). The liquid-fueled MSR showed tremendous promise during more than 20,000 hours of research and development at Oak Ridge National Labs in the late 60s and early 70s, but it was shelved by Richard Nixon to help his cronies in California, who wanted to develop another type of reactor (which didn’t work out so well.)
Today’s MSR proponents are confident that when research and development is resumed and brought up to speed, assembly-line production of MSRs could be initiated within five years. The cost of all this activity would be about $5 Billion—substantially less than the cost of one AP-1000 reactor in Vogtle, Georgia.
Several cost analyses on MSR designs have been done over the years, averaging about $2 an installed watt—cheaper than a coal plant, and far cleaner and safer as well. A true Gen-4 reactor, the MSR has several advantages:
- It can’t melt down
- It doesn’t need an external cooling system
- It’s naturally and automatically self-regulating
- It always operates at atmospheric pressure
- It won’t spread contaminants if damaged or destroyed
- It can be installed literally anywhere
- It can be modified to breed fuel for itself and other reactors
- It is completely impractical for making weapons
- It can be configured to consume nuclear “waste” as fuel
- It can pay for itself through the production of isotopes for medicine, science and industry
- It can be fueled by thorium, four times as abundant as uranium and found all over the world, particularly in America (it’s even in our beach sand.)
Since it never operates under pressure, an MSR doesn’t need a containment dome, one of the most expensive parts of a traditional nuclear plant. And MSRs don’t need exotic high-pressure parts, either. The reactor is simplicity itself.
Overall, an MSR’s steel and concrete requirements will be significantly less than an AP-1000, or any other solid-fuel, high-pressure, water-cooled reactor, including the Small Modular Reactors.
While SMRs are a major advance over the traditional Light Water Reactor, and are far safer machines, the liquid-fueled MSR is in a class all its own. It’s a completely different approach to reactor design, which has always used coolants that are fundamentally—and often violently—incompatible with the fuel.
Like the old saying goes, “Everything’s fine until something goes wrong.” And the few times that LWRs have gone wrong, the entire planet freaked out. In the wake of those three major incidents—only one of which (Chernobyl) has ever killed anyone—the safest form of large-scale carbon-free power production in the history of the world was very nearly shelved for good.
The key differences in MSR design is that the fuel is perfectly compatible with the coolant, because the coolant IS the fuel and the fuel IS the coolant, naturally expanding and contracting to maintain a safe and stable operating temperature.
They used to joke at Oak Ridge that the hardest thing about testing the MSR was finding something to do. The reactor can virtually run itself, and will automatically shut down if there’s a problem—an inherently “walk-away safe” design. And not because of clever engineering, but because of the laws of physics.
Wigner and Weinberg should have gotten the Nobel Prize. The MSR is that different. Liquid fuel changes everything. Liquid fuel is a very big deal.
The bottom line
The only way we’re going to power the nation—let alone the planet—on carbon-free energy is with nuclear power. And the sooner we all realize that, the better.
There’s so much work to do!
SEE another preview chapter We’re not betting the farm. We’re betting the planet.
Copyright © 2015 by Michael Sean Conley. All rights reserved.