There's a pumped-water electricity storage facility near where I live, called Northfield Mountain. It entered service in 1972 and has been in use since then.
When electricity demand is low, the grid operator uses the excess to pump water uphill. When electrity use is high, the water flows back down into the Connecticut River. Horthfield Mountain can deliver just over a gigawatt of energy for eight hours with the generators running at full bore.
It was originally built as a complement to the (now-decomissioned) Vermont Yankee nuclear power plant. That plant generated 0.6 gigawatts continuously, and functioned most efficienty if it wasn't ramped up and down to meet instantaneous demand. So, the storage facility provided a way to vary the power output from the grid.
The Times piece contained these words:
...utility-scale lithium-ion batteries cost 26 cents a kilowatt-hour, compared with 15 cents for a pumped-storage hydroelectric project. The typical household pays about 12.5 cents a kilowatt-hour for electricity.
It's a little misleading. The 26 cents and 15 cents figures are for capital expenditure to build storage capacity, which can then be used over and over (for 45 years in Northfield Mountain's case). The 12.5 cents is for a kilowatt hour consumed.
Also, I fear the 12.5 number represents lazy journalism. The current residential rate in Los Angeles is 17.8 cents/kWh, and in New York City (home of the Times) it's 21.0 cents/kWh. Rates vary a LOT geographically.
I hope journalists can learn to understand energy economics better, because they are a way for the rest of us to learn about them. And learn about them we must when fossil fuel use diminishes.
Northfield Mountain is a great plant. The company I work for used to own it before we foolishly sold it off. A pension plan bought it, which is a good indication of how stable it's economic outlook is. It's also an important black-start resource in the northeast.
The difference between Northfield and Lake Mead / Hoover Dam is that dedicated pumped-storage plants have two reservoirs: upper and lower. They are generally able to move water between them as needed with few restrictions because there is no downstream water user who's going to suffer damage because water was released from the upper reservoir (or taken from the lower reservoir).
The Colorado River is pretty much the ultimate opposite case. The existing Hoover Dam operations are dictated entirely by downstream flow requirements. If water has to flow downstream (because it's needed by agricultural users), it flows, and power generation follows. If flow needs to be restricted for flood control, power generation is reduced accordingly. Other than emergency spillways, there's no way for water to flow downstream except through the turbines, so there's no flexibility based on the needs or conditions of the power grid. In power industry terminology this is called a 'must-run' resource.
This article doesn't address that at all, although I imagine that if there is a serious engineering paper proposing this that it would look at the consequences of this in detail. It's extremely questionable to invest this scale of money in an energy storage resource that will likely never attain its full operational potential because the operator would not be allowed to ramp up or down arbitrarily. Batteries are more expensive - right now - but far less risky and can easily be deployed to any location, in precisely the size and scale required, with minimal financial risk or exposure to changing climate patterns.
Similarly for the batteries. If lithium-ion could be had for $0.26/kWh then a Tesla Model S battery would be $26.
As it stands, Tesla batteries are estimated to be around the $100/kWh level for individual cells and projected to fall below $100/kWh at the pack level in a couple of years [1].
The 26 cent figure for a kWh of storage must be the cost per cycle, and assuming it gets thousands of cycles. 1 kWh of Li-ion battery capacity is orders of magnitude more expensive than 26 cents.
Also they fail to mention cyclic efficiency and life cycle cost. Hydrolic storage are most efficient with their short comingss being too large and location dependaance.
The elephant in the room that I'm surprised wasn't mentioned is that there's no spare water in the Colorado. The Colorado River Compact allocated more than 100% of the current flow of the Colorado River to the different states (and a separate agreement adds up Mexico as well too), and it's at the point where a "good" winter snowpack means that the water level doesn't fall instead of providing meaningful recharge. The snow in 2016-2017 meant a return to 2014's low, but 2014 itself saw a miserable 20-foot fall in lake levels that there's no prospect for reversing. So where will the water to pump into the lake come from?
First, the water is lost - because evaporation over the large area of the reservoir only increases with its area, which increases with its level
Second, the scheme decreases flow, and that means level of the downstream. This means negative effects on ecosystems downstream.
Third, large hydroelectric dams require subordinate downstream reservoirs to tackle exactly the second problem - keep the downstream flow of the system (of the two+ dams) at least at some acceptable level to hide the variations in the main dam's
outflow which are due to daily/weekly/monthly/yearly fluctuations in inflow and power requirements.
Fourth - hydroelectric turbines require some amount of backpressure to operate. If the backpressure (that is the level of water downstream of the dam) gets too low, they can't operate unless risking damage. (that's actually the original reason behind the subordinate downstream dams)
So the project just does not make any sense. Where are they going to get the water?
Can you spell this out in more detail? I would have thought that in the near future, it would just time shift flow from daytime (when solar production is high) to nighttime, but perhaps my mental model is over simplified?
The most obvious explanation is that for an unit of power (not energy, mind) one has to spend a unit of flow (which is unit of volume per unit of time) falling over a unit of height (which the dam is for).
Now if you pump that unit of flow back up, total downstream loses that same unit of flow.
Second order effect is that when you pump water back up into the reservoir, it is evaporating at an increased rate.
Determining the evoparation increase is not trivial, since the reservoir surface area increases much faster than linear compared to its level at the dam. If they didn't, the dam would've been that much higher, because evaporative losses of the future reservoir are a major factor in the design of the dam.
I'm only asking about the first order effect (your second point, not your first). My question was: if you're letting that water back out every night, don't you get the unit of flow back then? It's delayed by up to 12 hours, but is that a really big effect?
You have to keep enough water somewhere downstream of the dam for the night so that you have something to pump up during the day. If you don't, there's either not enough water to pump up, or severe level fluctuations. If you do, it requires another dam, evapo losses, etc, etc.
Nit: Lake Mead is the battery. Hoover's generators are extracting power from that "battery" constantly.
The plan from TFA: Take water that's already passed through the generators at Hoover and, using solar and wind power, pump it back upstream into Mead to be processed again. This gives us the ability to supplement hydro with solar and wind. I guess if this is more efficient than other batteries, it's not a completely crazy idea. It does, however, have potential negative impact on downstream ecosystems.
Also: I assume that we don't extract 100% of the available potential from the dammed water: how much water flows past generators (instead of through them) and how does that change when upstream water supply changes? How much of a decrease in this "loss" can we get from this solar/wind plan?
At Hoover Dam, 100% of the river flows through the turbines with no bypass. The dam was originally built for flood control. Power generation is driven by downstream water demands. [1]
But it isn’t 100% efficient in converting potential energy into electricity. Even if the generators were 100% efficient, resistance in the pipes leading to the turbines slow down the water, not all the water hits the turbine blades, the water that does doesn’t transfer all its energy (the water would have to have zero speed after the turbine to do that, and the turbines would have to be at the bottom water level)
I took the OP's "how much water flows past generators" to address whether all the dam's outflow goes through the turbines or whether some is diverted around them.
The former is the case. The latter only happens when spillways are in use, which is rare.
> The amount of electricity generated by Hoover Dam has been decreasing along with the falling water level in Lake Mead due to the prolonged drought in the 2010s and high demand for the Colorado River's water. Lake Mead fell to a new record low elevation of 1,071.61 feet (326.63 m) on July 1, 2016 before beginning to rebound slowly.[95] Under its original design, the dam will no longer be able to generate power once the water level falls below 1,050 feet (320 m), which could occur as early as 2017. To lower the minimum power pool elevation from 1,050 to 950 feet (320 to 290 m), five wide-head turbines, designed to work efficiently with less flow, were installed.[96] Due to the low water levels, by 2014 it was providing power only during periods of peak demand.[97]
> The Los Angeles Department of Water and Power, the nation’s largest municipal utility, says its proposal would increase the productivity of the dam, which operates at just 20 percent of its potential, to avoid releasing too much water at once and flooding towns downstream.
Fixing that (digging the river deeper in downsteam towns) would probably be a cheaper way to allow Hoover to scale up and down to more variable demand from wind and solar.
This might sound insane, but the great lakes might work, with a joint Canada-US plan. Raising one of them by a foot would be all you'd need because their area is so huge.
If you're willing to look outside the US, a wild but interesting mega project might be to try to do the same thing with the South Aral Sea. Pump water all the way from the Caspian. The North Aral is cut off by a dam/dyke and the south Aral barely exists anymore, causing major economic harm. Fill it with salt water and salt water fish to revive the fishing economy while using it as a giant battery.
Niagara falls (Erie -> Ontario) is already a huge hydro eletric power generator with variable flow. Instead of storing water though they decide how much water to allow over the falls for tourists. More during the day, less at night.
Thinking about the terrain, it would probably be relatively easy to setup a dam somewhere that acted as a battery, downstream is basically already a giant bowl... I suspect the tourist industry would complain though.
During the Post-Great Big Blackout period, the Falls power generation ran with night-time levels of diversion during the day-time. The tourist boats that go near the falls got furloughed.
Not if the lake is going to be there in either case.
However, storing months of power is silly, you are much better off with extra ~6c/kWh wind/solar that you use 1/2 the year than trying to store ~6kWh energy for six months.
Renewable are the new base load power becase they are Cheap and you don’t gain from not using that energy. They still need peaking power or storage to follow the demand curve.
PS: Say it three times fast ‘base load power is a downside.’
"Water released for power in excess of local and downstream requirements is conserved by pumpback operation during off-peak hours through both powerplants into Lake Oroville to be subsequently released for power generation during periods of peak power demand. The Plant has 4 units: 1 generating unit and 3 pumping–generating units, with an installed capacity of 120 MW, maximum flow rate of 17,400 cu ft/s (490 m3/s).[5]"
Summer days are always dry and sunny, so a large solar array might work well here.
Crater lake doesn't currently have Rivers out or in. The water level is maintained by precipitation and evaporation - so you are right. Opposition from basically everyone.
Death valley varies from outflow to inflow every 400 to 500 years currently no outflow iirc so if we want to wait a few hundred years...
Interesting ideas though
I'm visiting near the Salton Sea right now and have been learning more about its fascinating history. I don't know about the technical advantages or disadvantages (evaporation?) but there would be minimal local opposition to such a project, that's for sure.
When there is a dam already producing electricity, you just reduce the flow through the dam rather than pumping the water back uphill. That is by far the most efficient way to handle excess generation from other sources in this case. This is a politically driven proposal whose implementation will be a boon to private producers at the expense of taxpayers.
This could certainly be a boondoggle, but reducing flow through a dam is limited to storing excess power at the natural flow rate of the river, whereas there is no limit in principle to how fast you could pump water from a big reservoir below a dam to a big reservoir above the dam, or how fast you could draw energy from greatly increased flow through your generators. So pumped hydro storage isn't an inherently silly idea.
A hydro turbine on a dam with no penstocks can probably go from 0-100% in 5 to 10 seconds.
The limit is pressure waves in the water conveyance system due to changes in flow through the turbine. If the water conveyance system is only the width of the dam and encased in concrete it is going to have minimal transients compared to a 7km long steel or HDPE penstock on which the time from 0-100% is more like 2 minutes
After moving to the Columbia Gorge I became fascinated with
the dams and energy production on the Columbia. Over the last 75ish years the whole river has been turned into a battery through a series of dams and reservoirs. The scale is astonishing - Grand Coulee itself produces ~6800 megawatts (about 3x of Hoover Dam).
For those interested [A River Lost: The Life and Death of the Columbia](https://www.amazon.com/River-Lost-Life-Death-Columbia/dp/039...) is a great book documenting the creation and long term implications (from cheap electricity for the Manhattan Project, to agriculture, to the collapse of the Salmon population) of the project.
A similar project has been on construction in Europe
https://de.m.wikipedia.org/wiki/Nant_de_Drance
Cost of $2B. I had the chance to visit the caverne where the turbines will be, really amazing engineering!
Pump storages are already all over Europe for decades. The most famous one being Kaprun/Austria, built by the Nazis in the 30ies. They are all over the Alps, as the green party lets them.
Those pump storages provide the very expensive peek electricity all over Europe, needed for the spikes in the morning and noon.
Nevada probably didn't need it back then, Hoover Dam was built for flood protection, and since then the levels are too low to work efficiently as pump storage. But since peek energy is in high demand they think it over.
If you've not been to Hoover Dam, it's spectacular. Not just the engineering, but the gorgeous art deco everywhere which has been preserved rather than 'modernized'.
If you look at the topography of the area in google maps you will see that the sides of the canyon are somewhat steep and rugged for quite a ways downstream of the dam. Go a bit further and you find Willow Beach, which has a nice flat run north and then west back to the reservoir. I am guessing that being able to run the pipe using trenching equipment rather than tunneling through canyon walls is a big advantage.
No, any extra height to pump water up compared to the height of the dam is just wasted power. They wouldn't do that without a good reason.
My understanding is you can't just stick an inlet on the river, otherwise you could only pump water up when the dam was running (defeating the point). You need to build some kind of reservoir.
Additionally, with hydropower it's not enough to build a dam above your generators. You also need to get rid of the water exiting your generators fast enough. If there is any restiction on the outflow, the back pressure will decrease the generation capacity.
My guess is the distance is needed to meet these requirements. Enough space to let the water slow down, then enough space to build a reservoir.
Why would it be wasted power? Height difference = stored potential power. I was talking about that maybe 20 miles downstream the pump-size/-energy consumption to stored power is more efficent.
Because the height difference that matters is the difference between the water level in the lake and the turbines.
If your pumping station inlet is 20m below the turbines, then any energy used pumping water upto the height of the turbines is wasted, because it can't be recovered when the water comes back down.
This seems incredibly stupid. If you use excess electricity to pump water backward at the dam, you may as well just shut off the flow through the dam instead. That would be more efficient and cost nothing to implement.
If we're talking about so much excess capacity that the dam would be closed AND water pumped back up, that would be interesting. But wouldn't the river go dry? It's not like there's a lake below the dam for that.
This whole thing seems more political than practical. I'm guessing the private producers will somehow get a financial benefit from this at the expense of the taxpayer.
The proposal isn’t to use the dam’s excess power to pump the water back uphill. They want to use excess power from other solar and wind plants to effectively store any excess energy that they produce. Since these alternative sources cannot consistently produce energy 24 hours per day, and the times that they are able to produce it do not necessarily match up with consumer demand, they need to store it somehow. This is being proposed as a more efficient and cost effective form of storage than batteries (though I am curious how much energy will be lost due to evaporation after the water has been pumped in Hoover dam’s hot climate and other inefficiencies).
You might have two simultaneous problems: The water level of the lake behind the dam might be lower than desired, and the rate of water flow into that lake from the upstream river might also be lower than desired. In that case, it'd be advantageous to use the excess power to recapture some of the "downstream" water and pump it back into the lake, thereby restoring the lake level more quickly than would be possible with just the (inadequate) incoming upstream river water.
I don't think a facile analysis of the economics is interesting.
For instance, if you shut the dam off, then it is generating $0 revenue for that period. Is that a better use of capital than adding seemingly redundant pumps to recapture some of the (time shifted) flow?
I'm not sure why you are being downvoted, you are correct. If the normal state of the dam is to generate electricity then the obvious response to excess supply is to stop the flow. No need to pump anything back uphill unless the excess exceeds the normal capacity of the dam.
> If the normal state of the dam is to generate electricity
It isn't, at least not in the sense that you likely mean.
That seems to be a common misconception about the Hoover Dam. Its ability to generate electricity is completely secondary to its primary purpose: controlling the flow of water (e.g. flood control, irrigation).
Unless/until there is a lower reservoir (as proposed by the project in the article) that takes over that primary function, changing the flow through the dam for any electricity-related reasons is simply not an option.
When electricity demand is low, the grid operator uses the excess to pump water uphill. When electrity use is high, the water flows back down into the Connecticut River. Horthfield Mountain can deliver just over a gigawatt of energy for eight hours with the generators running at full bore.
It was originally built as a complement to the (now-decomissioned) Vermont Yankee nuclear power plant. That plant generated 0.6 gigawatts continuously, and functioned most efficienty if it wasn't ramped up and down to meet instantaneous demand. So, the storage facility provided a way to vary the power output from the grid.
The Times piece contained these words:
...utility-scale lithium-ion batteries cost 26 cents a kilowatt-hour, compared with 15 cents for a pumped-storage hydroelectric project. The typical household pays about 12.5 cents a kilowatt-hour for electricity.
It's a little misleading. The 26 cents and 15 cents figures are for capital expenditure to build storage capacity, which can then be used over and over (for 45 years in Northfield Mountain's case). The 12.5 cents is for a kilowatt hour consumed.
Also, I fear the 12.5 number represents lazy journalism. The current residential rate in Los Angeles is 17.8 cents/kWh, and in New York City (home of the Times) it's 21.0 cents/kWh. Rates vary a LOT geographically.
I hope journalists can learn to understand energy economics better, because they are a way for the rest of us to learn about them. And learn about them we must when fossil fuel use diminishes.