Aside from a handful of CSP plants comprising 7.3% of the basic 2050 grid, and another handful comprising the grid’s entire 4.38% overbuild, none of the Roadmap’s farms will have any built-in training wheels.
External backup or energy storage would be the only reliable ways to make the Roadmap’s fuel-free grid work.
Cheap storage is the holy grail of renewable energy, the long sought game-changer that will finally put renewables on par with traditional baseload systems.
With enough cheap storage, the bits and bursts of energy gathered from wind and solar can be collected and later released in a smooth, steady stream that mimics the dependable power we rely upon for modern life.
But the Roadmap has a better idea: Eliminate the need for storage. With enough oversized farms in enough locations backing each other up, who needs storage?
This is an adventurous proposition, because the light and motion that make renewables work isn’t under human control and never will be, unless and until the energy we make with it can be transferred to a storable medium.
Our existing grid has a 2.5X overbuild, consisting of
fossil, nuclear, and hydro plants. All these systems have storage in the form of actual fuel, or in the potential energy of elevated water. And though drought is eroding our hydro capacity, we still have plenty of fossil and nuclear fuel on hand.
Traditional power plants extract energy from their fuel supply (a load of coal, a fuel tank, a fuel rod, or a reservoir of water) and generate power with that energy day and night.
Fuel (or elevated water) enables these plants to adjust their output, sometimes within minutes, to respond to predictable peak loads.
Renewables, not so much.
Overbuilding the grid is all about how much extra generating capacity (fuel
or hydro) can be reliably brought into service with adequate lead time.
The operative concept is “reliability,” but the Roadmap dispenses with this old-school notion. Its 4.38% overbuild of extra CSP farms will rely strictly on sunshine – if and when it’s available.
As you can imagine, this may not suffice to ensure grid stability. For example, California needs to find 60 additional MWs per minute every afternoon, as the sun’s energy fades past its midpoint and solar production declines.
So let’s play gridmaster, and examine the quantity of extra power potentially available in 2050, to see just how wobbly the grid might be without training wheels.
First off, the Roadmap calls for adding 19.2 GWs to our dammed hydro production, from our current 28.7 GWs, for a new peak capacity of 47.9 GWs
(3% of the 2050 grid). This will be done by upgrading the turbines and extending operating hours.
However, the turbines in our large dams weren’t designed to be ramped up and down as fast-response backups, even if they were upgraded. So we’re guessing that the Roadmap’s expansion of our hydro dams would probably include the installation of additional, fast-start hydro turbines.
It’s just our speculation, but our large existing dam turbines would most likely function as they always have, providing both baseload and what Europe calls intermediate load – power that can be slowly ramped up and down to meet predictable daily and seasonal demand. The newly installed turbines would be fast-start models to respond to unexpected peak loads.
We’ll also have our existing 22 GWs (peak capacity) of pumped hydro storage, which can run for about 12 hours – if the reservoirs are full. Which means the Roadmap, through no fault of its own, will actually inherit a tiny amount of storage. And we’ll also have the Roadmap’s 69.7 GWs of overbuild CSP.
Our intrepid gridmasters of 2050 will therefore have the following options:
That’s a theoretical blue-sky maximum of 110.9 GWs (about 7% of the grid.) Which is great, until we have an extended period of unfavorable weather.
And that’s the Roadmap’s entire backup for the 2050 grid. We recommend more.
Staking out a patch of wilderness and waiting for energy
to come along is what we call a Cargo Cult approach to power.
In World War II, remote Pacific Islanders marveled at the bounty that washed ashore from torpedoed cargo ships. As far as they knew, they must have finally thrown the right virgin in the right volcano, because all of a sudden the gods were sending them tons of free goodies.
Mother Nature sends us an inexhaustible bounty of wind, waves, tide and sunlight. WWS “islands” catch this light and motion and make electric power.
But if we can’t use the power as soon as it’s made, we have to store it or dump it. At least cargo-cult Spam had a shelf life.
Lithium-ion batteries are mentioned in the Roadmap’s footnotes, but they aren’t formally factored in, probably because they’re prohibitively expensive and haven’t been tested at scale.
In any case, those that are on the market are only warrantied for 10 years. So even if prices drop like a rock, their lifespans will have to grow by 6X to match
But the battery guys aren’t about to give up.
In July of 2017, Elon Musk promised to build a 100-MW lithium-ion battery in 100 days, to store the excess production of a South Australian wind farm.1
If he makes his self-imposed 100-day deadline, the price is $50 Million. If he misses the deadline, it’s free. (No worries – he can afford it.)
That is one big-ass battery. But it’s still not big enough – like everything else in the field of renewables, it’s a nice idea that doesn’t scale.
One hour of energy storage for the Roadmap’s 1,591-GW grid would require over 122,000 tonnes of lithium (77 tonnes per GW-hr × 1,591.)
That’s more than 3 times the total global production of lithium in 2016, which was 36,000 tonnes.4 Flip the numbers around:
All the lithium mined on planet Earth in 2016 would provide a whopping 18 minutes of all-grid storage for the 2050 Roadmap (36,000 tonnes ÷ 122,000 tonnes = 0.3. And 0.3 hrs = 18 minutes.)
Liquid-flow batteries have been generating considerable interest in renewables circles as a possible solution to mass energy storage. Don’t get too excited.
The battery consists of two tanks of different electrolyte solutions (dissolved minerals.) The liquids are pumped into a divided chamber (the two solutions never touch) to produce a flow of dc (direct current) electricity.
The liquids are returned to their tanks, and the system is recharged for another round, using surplus electricity from a wind or solar farm (when it’s available.)
Vanadium oxide is the electrolyte mineral of choice. Unlike a solid lithium battery, which eventually wears down and must be dismantled to recycle the material, a vanadium electrolyte solution can be used over and over again.
Problem is, a grand total of just 79,400 metric tonnes of vanadium were mined last year, over half of it in China. And they used 90% of it for making steel, not batteries.
Aside from resource availability, the problem with flow batteries is energy density: Vanadium electrolyte solution can only store 0.36 MW-hrs per tonne of vanadium, compared to lithium’s storage capacity of 13 MW-hrs per tonne. That’s a 36:1 ratio.
And, the round-trip efficiency of a flow battery is about 75%, compared to a lithium battery’s 90%. But here’s the real deal breaker:
If all 79,400 tonnes of vanadium that was mined worldwide in 2015 were used to make electrolyte solution for one big-ass flow battery, it would store about one minute of all-grid storage.5
Back to the drawing board.
Underground Thermal Energy Storage (UTES) is mostly done in the form of borehole energy storage (BHES).
Basically, borehole is just a residential heat pump writ large and enhanced with thermal solar panels, those black plastic rooftop panels that heat circulating water.
The hot water is sent through a network of underground pipes and the surrounding soil retains the heat, which can later be retrieved for space heating.
Since different soils retain heat with varying degrees of efficiency, borehole is
a site-dependent technology. But it works well enough that the authors of the Roadmap chose it for the bulk of our heating needs.
Sounds boring, but space heating is a sizable slice of our energy pie, currently consuming about 10% of primary energy.6
(Remember, primary energy is all the energy we use: For air, land, and sea transportation; for construction, industrial process heat, space heating, etc.,
as well as electricity.)
The Roadmap removes space heating from the 2050 primary energy pie because it’ll be produced independently of the grid, using borehole rather than grid electricity.
Fair enough – for our all-nuclear grid we’ve done the same.
With improved efficiency and building insulation, the Roadmap estimates that our national space heating requirements will drop to just 7.2% of our total primary energy, down from the approximately 10% that space heating now consumes (mostly natural gas and heating oil.) That 7.2% amounts to 114.7 GWs, derived from heat, not electricity.
Like a hydroelectric dam, pumped hydro (PHES, or Pumped Hydro Energy Storage) generates energy from the force of falling water.
A superior storage system, pumped hydro comprises nearly 99% of all mass energy storage in the world. The U.S. currently has a peak capacity of 22 GWs, which is about 2% of our grid’s total peak power.7
Essentially, pumped hydro acts like a dam: Pump water uphill when you have the power to spare, and let it run back down through the same reversible turbines when you need the power to satisfy electric demand.
The round-trip efficiency is about 80%: If you use 100 megawatt-hours to pump up the hydro, you’ll get 80-ish MW-hrs back.
Unlike our existing hydroelectric dams, our pumped hydro reservoirs aren’t factored into the 2050 grid. Probably because, strictly speaking, pumped hydro doesn’t generate its own electricity. Rather, it stores and re-generates (most of) the energy that was used to fill its reservoir in the first place.
When it comes to mass energy storage, pumped hydro is the gold standard that has to be matched or beaten by other non-fuel storage technologies. And it’s as low-tech and reliable as gravity.
Like borehole, the cost and efficiency of pumped hydro is site-dependent. If there’s a dammable, high-elevation valley nearby, or an abandoned watertight mine, you’re in luck. Otherwise, you’ll have to build two reservoirs, not one.
This is one reason why pumped hydro prices vary from a few pennies per installed watt-hour, to well over a dollar. To be more than fair, we’ll calculate pumped hydro for the 2050 grid based on the lowest quintile: $0.20 per watt-hour of stored energy. It’ll be interesting to see if P2G can match or beat that $0.20.
So why did we call out a price per “installed watt” for a fueled power plant, if we called out a price per “installed watt-hour” for storage?
Because any fueled generator can continuously produce power at its nameplate rating as long as it has fuel. But storage can only deliver a finite amount of energy before it has to be recharged, refueled, or refilled.
Unlike the heat stored in CSP’s molten salt, pumped hydro
energy doesn’t fade away, unless the water evaporates or
the reservoir leaks. Which is why fresh water is a must: A saltwater leak would be catastrophic to the local flora.
That’s why ancient armies would sow their enemy’s fields with salt.
The drawbacks of large-scale pumped hydro are the gargantuan amounts of water required, and the evaporative loss that’s bound to occur. Not to mention
the permanent inundation of habitable land.
To put the volume of water in perspective, it would take almost 135 days of America’s total fresh water consumption (irrigation, industry, tap water, the works) to store one day of power at 1,591 GWs, or one “grid day”.
So that’s off the table.
We’ll explain how we arrived at 135 days in the section “One ESB” below.
The general consensus of MSR engineers is that Molten Salt Reactors can be built for an average cost of $2.00 an installed watt, which would make them substantially cheaper than a coal plant.8
The reactors of an all-nuclear grid could generate about
550 grid days of power (18 months) before refueling. And that’s if we use traditional solid-fuel reactors.
Liquid-fuel Molten Salt Reactors would be even better, since an MSR can be refueled while it’s running by pouring in more fuel salt.
Like most other Gen IV reactors, a pair of small MSRs would enable switching from one reactor to the other, so the first one can be taken back to the factory for service and refueling.9
ThorCon’s twin 500-MW reactors will feature continuous electrical production 24 / 7, year in and year out. A fresh reactor will come online as a spent reactor is shut down and shipped back to the factory.
This eliminates the downtime that is ordinarily backed up by the rest of the fleet, meaning the entire collection of power plants that generate power for
Since an MSR can run non-stop for 4–10 years before it needs to be serviced, the ability to refuel on the fly, and / or the ability to switch to a fresh on-site reactor, are two major advantages that give Gen IV reactors a capacity factor that will probably exceed 99%.
Of all the Gen IVs, we feel the MSR is the solution for our energy needs, featuring:
We now return to our regularly scheduled program.
We coined the term “ESB” to refer to the volume of one Empire State Building – the amount of falling water required to generate 250 MW-hrs of electric energy.
A 100-meter drop provides a good pressure head, like the standpipe in a skyscraper. Which is what inspired the term.
Imagine if the Empire State Building was made entirely of water, like the water snake in The Abyss.10 Now picture that water draining by gravity through a bank of hydro-turbines in the basement.
That’s one ESB.
Under perfect conditions (i.e., with 100%-efficient machinery), 917,400 cubic meters of water, falling 100 meters, would be all the water you need: That much water has a kinetic energy capable of producing 250 MW-hrs of electric energy.11
But in real-world conditions, hydro turbine-generator efficiency is about 90%. So the actual water volume required to generate 250 MW-hrs is 1,020,000 m3 (917,400 ÷ 0.90 = 1,020,000.)
The ideal water volume of 917,400 m3 is roughly equal
to the visible part of the Empire State Building, which is 900,000 cubic meters. The entire structure, including the basement, comes to 1,100,000 m3. So the ESB acronym works in either case.
Of course, a reservoir would never be shaped like that. We’re just trying to give you a good visual on the volume of water required.
Drop one ESB in one hour and you generate 250 MWs. Drop it in 5 hours and you get 50 MWs for five hours. Drop it in 10 hours and you get 25 MWs for 10 hours. Et cetera. The rule of thumb is: “One ESB = 250 MW-hrs.”
(And no, tonne is not pronounced “tonay” or “tunny.” A tonne is a tonne.)
In the U.S., a tonne is often called a “metric ton” to distinguish it from a 2,000-lb U.S. ton, which is sometimes called a “short ton”. At 2,200 lbs, a metric ton is also called a “long ton”.
Weight is gravity’s pull on mass. So we’re actually talking about a mass of water with a volume of one cubic meter, which just so happens to weigh about 2,200 lbs on this planet. The same mass of water would weigh more, or less, on another world.
All these labels and numbers might seem confusing, but that’s because we in the U.S. adopted the antiquated measurements of inches, feet, yards and pounds from England, while the rest of the world adopted the metric system.
Science uses the metric system because all metric measurements – volume, energy, mass, distance, force, work, power, etc. – fit together like Legos in a simple, elegant, and logical framework. And it all comes back to the weight and volume of water.12
They don’t call this a water planet for nothing.
Pump up the hydro13
At the rate of 1,020,000 m3 for 250 MW-hrs, we would
need about 156 billion cubic meters of water (that’s 156 cubic kilometers) to generate one grid-day of power.
In 2010, the U.S. consumed about 421 cubic kilometers of fresh water. Which means that one grid-day of pumped-hydro energy production requires the fresh water consumption of the entire nation for nearly 135 days.14
Even at the bargain-basement price of $0.20 per installed watt-hour, twenty-four hours of pumped hydro for the
entire 1,591-GW grid would cost $7.6 Trillion.15 Plus you’ll need enough fresh water, and the land to build the reservoirs.
Keep in mind, prices will vary depending on site conditions. But even with the most favorable conditions, one grid day of storage (a prudent insurance policy) would balloon the cost of the Roadmap to nearly $23 Trillion.
That’s a lot of money, and a lot of water – 152,700 ESBs to be exact. If the entire island of Manhattan was carpeted with a solid mass of Empire State Buildings, built window-to-window, you would need 19 Manhattan Islands to accommodate that many ESBs.
All of which is ridiculous. So let’s explore a (slightly) more reasonable scenario:
WWS advocates say that with a smartly managed grid, wind and solar farms will only need a few hours’ storage. Since the number 4 has been bandied about, let’s do the math:
Four hours is 1/6th of a day, which comes to about 23 days of total U.S. fresh water use. Which is still ridiculous.
Granted, the entire electric grid would never have to be backed up all at once, even for 4 hours. But with a major interruption like the Northeast Blackout of August 2003, a substantial portion might need backup for a day or more. These things add up.
In our view, a mere four hours of all-grid backup is pretty darn optimistic for a fuel-free grid. Our existing grid has a safety margin of 150%, and virtually all the power plants we have, including our backup systems, are powered by fuel or hydro, with plenty of extra fuel on hand.
Even then, we still have the occasional wide-area blackout.
Nevertheless, the authors of the Roadmap are confident that we can enjoy a clean and green fuel-free future, utilizing geographic sweet spots in all 50 states.
Distributing, diversifying, and interconnecting our 50,000-plus WWS farms would (hopefully) ensure that the bulk of them would never be idle at the same time, and would always be able to back each other up.
Which looks great on paper, until you look out the window.
A 10.8 volt (V), 8.8 amp-hour (Ah) Li-ion battery contains 7.9 grams (g) lithium.
10.8 V X 8.8 coulombs / sec X 3,600 sec / h = 342e3 joules (J) energy content of battery.
Conversion factor: 1 kWh = 3.6e6 J.
342e3 J X 1 kWh / 3.6e6 J = 0.095 kWh energy content of the battery.
Therefore: 7.9 g Li / 0.095 kWh = 83 g lithium / kWh.
Now click on:
Refer to derivation by Saeed Kazemiabnavi: lithium content = 0.0714 kg /kWh
or 71 g lithium /kWh.
Average the values 71 g and 83 g to obtain 77 g Li /kWh.
100 MW / 129 MW-hrs refers to 129 megawatt-hours of energy storage (energy content, or energy “capacity”), with a maximum power output (discharge rate) of 100 megawatts.
As usual, the word “capacity” is misused here to refer to peak power output.
129e6 W-hrs energy content X 77 g Li /1e3 W-hrs = 9.9e6 g Li, or 9.9 tonnes lithium.
See the 2nd image in the left column. Vanadium flow battery (VFB) storage capacity of 1,600 MW-hrs
uses 8,000 tonnes of V2O5.
Of that 8,000 tonnes, 4,480 tonnes is elemental vanadium.
Vanadium atomic mass = 51, and oxygen atomic mass = 16.
So, V2O5 molecular mass = (2 × 51) + (5 × 16) = (102 + 80) = 182. The vanadium portion = 102 ÷ 182 = 0.56.
8,000 tonnes × 0.56 = 4,480 tonnes elemental vanadium.
VFB specific energy = 1,600e6 W-hrs ÷ 4,480e6 grams = 0.357 W-hr / gram of vanadium.
Alternatively, 2.8 grams / W-hr of stored energy.
See first section titled Global production of vanadium totaled 79,400 tonnes in 2015.
One grid-hour of energy storage is 1,591 GWs × 1 hour = 1,591e9 W-hrs.
1591e9 W-hrs × 2.8 grams / W-hr = 4.45e12 grams of vanadium per grid-hour.
World production of 79,400e6 grams ÷ 4.45e12 grams /grid-hour = 0.018 grid-hour, or about 1 minute.
See: Frame 62, page 61.
See: page 6ff
Acceleration from earth’s gravity (g) = 9.81 meters /second per second (9.81 m /s2).
Force (F) [also called weight] = mass X acceleration = m X g.
F = 1000 kg X 9.81 m / s2 = 9.81e3 newtons (N).
Kinetic energy (NRG) from falling 100 meters onto hydroturbine = F X distance = 9.81e3 N X 100 m =
981e3 joules (J) per cubic meter.
Conversion factor: 1 watt-hour (Wh) = 3.6e3 J.
981e3 J per m3 of water / 3.6e3 J /Wh = 273 Wh of kinetic NRG per m3 of water.
Ideally, 1 ESB = 917,400 m3 (with 100% efficient machinery).
273 W-hrs / m3 X 917,400 m3 = 250e6 W-hrs.
Or 250 megawatt-hours per 1 ESB.
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.
U.S. annual water use = 397 million acre-feet per year, of which 86% was fresh water, so 341 million acre-feet.
Multiply by conversion factor 1.233e-6 km3 / acre-foot.
Obtain 421 km3 / year, or 421e9 m3 / year
1,56e9 m3 per grid-day / 421e9 m3 water usage / year X 365 days per year =
135 days of fresh water usage for one grid-day.