Nukes, schmukes – fuel cells could power future datacenters Proton-exchange membrane (PEM) fuel cells are more efficient than other alternative generator technologies as a way to provide backup power to datacenters in certain climates, according to a recent study. "Hydrogen fuel cells have emerged as a potentially dependable solution to address the intermittent nature of weather- …
Fuel Cells and Elecrolysers both work, if you have enough Platinum to build them.
Efficiency of the whole process is ~40%, from electricity in to electricity out, so it has to be on the pretense that excess, otherwise-curtailed wind power is 'free'.
But at the moment, wind generators get lots of money in subsidies when they are told to shut down. So is the excess energy really 'free'?
And if everyone tried to build electrolysers and fuel cells all at once, the price of both platinum and copper would rocket.
They work at 1.5 Volts (unlike Lithium batteries which have 4 volts per cell) so you need an awful lot more current and an awful lot more copper per cell, and you have to plate the electrodes in Platinum ...
So it is certainly not cheap. It may yet be cheaper than Nuclear, but that isn't saying much.
"Efficiency of the whole process is ~40%, from electricity in to electricity out, so it has to be on the pretense that excess, otherwise-curtailed wind power is 'free'.
But at the moment, wind generators get lots of money in subsidies when they are told to shut down. So is the excess energy really 'free'?"
It's better than free... rather than paying people a premium to not generate electricity you pay them less than that to generate the electricity... you also then get to *not* pay someone else to burn crap later on.
> It's better than free... rather than paying people a premium to not generate electricity you pay them less than that to generate the electricity... you also then get to *not* pay someone else to burn crap later on.
Of course, I agree with you. But in our weird hodge podge of the worst elements of capitalism and communism where we hand out bungs and subsidies to private companies, the wind farm operators may well prefer to be paid for generating nothing (saving on maintenance costs, perhaps, of certain components), than being paid nothing for generating something.
But even if the energy were free (it isn't, even if we could scrap the subsidies) where are we going to get the raw materials for all these electrolysers and fuel cells? And where and how are we going so store the hydrogen?
It leaks if stored under pressure, and there all kinds of weird and wonderful issues if you store it as a cryofluid. The latter obviously taking a bit more energy, too.
I strongly suspect most providers would rather be generating electricity to be used later than not generating at all...
The point is that the energy generated *which would otherwise be curtailed* is free (since although you're paying for it you're also saving money on both not paying for shutdown and for not having to buy energy later - since it's being returned from storage)
There is a cost of storage, and there are technical challenges with hydrogen, but those are at least technically resolvable in a large scale static setup, in a way that they aren't for millions of small scale installations, many of which would be mobile. The cost of paying for the energy and the storage has to be less than the cost of paying for curtailment *and* the cost of purchasing energy at times of high demand... but that's where the accountants come in.
"I strongly suspect most providers would rather be generating electricity to be used later than not generating at all..."
At some point renewables operators are going to have to assume the "backing store" and "interrmittency" costs. The South Australian Battery has proven that it works but it needs to happen at a farm-by-farm level
It takes about three times as much energy to liquify hydrogen or compress it for transportation purposes as it does to produce the hydrogen in the first place
That _really_ badly knocks the end-to-end efficiency
The embrittlement issues are another matter and I really don't want to be within 1/4 mile of a serious Road Crash involving a COPV hydrogen container
I need the Paris icon again
Have I read that right? Electrolysis of water is no more than 50% efficient, so the "energy used to produce it in the first place" is about twice as much as the energy that it stores. Three times that would be six times the energy of the stored hydrogen.
I know that liquefying Hydrogen (by cooling it to 1/10 of the absolute temperature of LNG) is expensive, but is it really this expensive? Shurely shome mishtake, or else nobody would be wasting their breath debating the usefulness of liquid Hydrogen as an energy store..
..right?
I've worked for a global-scale energy supplier, and that company investigated the options for energy storage and swapping the energy vectors, so the whole wind to hydrogen caboodle, and lots more around that (eg flywheels, compressed air storage, hydrogen to methane) and built a whole series of test plant at reasonable scale. Overall efficiency was dismal, as I recall nowhere near 40% end to end.
If you're dissociating water, you're throwing away the oxygen (uneconomic to collect and use) so that's wasted energy, you've got diabatic and adiabatic losses on compression/decomp, then fuel cell efficiencies that aren't much to write home about. If you're looking at grid scale solutions then you'll usually be looking to burn the hydrogen, and in practical terms that means methanation (more losses) and then feeding into a CCGT (or worse OCGT for peaking). If handling hydrogen it's a pain because it'll leak far more readily than say methane, whilst it's less likely to go bang, if it does it'll go bang more spectacularly, and it doesn't lend itself to high densities of energy storage, menaing your unit costs for storage are proprtionately greater. And the plant costs of any hydrogen plant will be prohibitive if the capacity factor is poor, meaning that "off peak energy" is not going to support any credible business case.
There is no "free" energy, it always has a capital and operating cost, and all of the posturing on hydrogen and fuel cells remains a waste of resources until we develop a source of large scale, ultra low cost electricity. There are no candidates of which I'm aware.
Hydrogen NOW is hideously expensive and it's extracted from hydrocarbons.
Green hydrogen would need oil to be well past $250/bbl to be viable - and only if the input energy can't be used in other more efficient ways (which is highly unlikely)
In short: Hydrogen is a classic example of "Just because you CAN doesn't mean you SHOULD"
Green H2 costs are bound to drop significantly.
Simple proof: price of electrolysers is dropping below the 200$/kW as volumes increase.
Green H2 will benefit from the rush to renewable.
Without producing Green H2, it's much harder to find a business case beyond 50% of market share for renewables.
Said otherwise, new RE investments revenue decreases as the market share of RE increases because in periods of abundance wholesale prices are not paying back.
OTOH, during periods of scarcity, that Green H2 will be so useful, whether it's fed to fuel cell, powering a gas turbine or else.
Claiming today that Green H2 should be kept for the industry is like claiming in the 1900s that oil should be kept to burn in oil lamps.
Said otherwise, physical yields have never determined commercial successes.
"until we develop a source of large scale, ultra low cost electricity. There are no candidates of which I'm aware."
TMSR-LF1 - look it up (It's a rebuild of ORNL MSRE but started on thorium from the outset)
If it work as expected there's a good chance that renewables farms might end up being largely abandoned as far too expensive to operate (in addition to not being remotely capable of meeting the generation requirements for full decarbonisation)
This post has been deleted by its author
When I said 40%, i was only considering the electrolyser and fuel cell components, which are about 50% and 80% respectively.
0.5*0.8=0.4
But as others have pointed out, the cost of compressing the hydrogen for storage is extremely non-trivial, possibly dwarfing the input cost of electrolysis.
Saying we can recover the heat and use it for district heating is the same as saying we can recover low grade heat from gas, coal and nuclear plants. We can, and that's called CHP. But we don't usually bother, because it's very complicated and expensive, and the plant isn't always running at the same time as the heat is needed. It's also a completely separate argument to the Hydrogen debate, because it applies equally to other fuels.
Where is the hydrogen going to come from? And how do we safely store large amounts of it?
A recent plan to use hydrogen for steel making came to nothing as a genuinely sustainable source of the hydrogen could not be found. A scheme to convert a fleet of Council vehicles was aborted for the same reason.
A recent lightning triggered explosion at a methane store in a sewage processing plant highlights the problems of storing inflammable gases.
When I was at school (a long time ago, thank goodness. It seemed like a prison for children who'd committed the crime of just existing) both flammable and inflammable are allowed descriptions for burnable stuff, but using the latter has always confused me.
Speaking of which, why not make an hotel out of the fuel cells on the datacentre site? Because that worked really well on Quantum of Solace.
Source of hydrogen: I recently learned of 'white hydrogen' produced by natural inorganic processes, which gathers in subterranean traps like natural gas does. Only discovered recently.
Is it really more difficult to light? Hydrogen will burn at lower concentrations in air than methane, but the big difference is in explosive potential, where you need higher concentrations of hydrogen to get an explosive mix, against which hydrogen diffuses in air much faster than methane. On the other hand, the flame velocity in hydrogen air mixes is about ten times that of methane, so whilst an explosive mix is generally less likely than with methane, you really don't want to be anywhere near a hydrogen explosion.
"hydrogen is more difficult to light than methane"
wtf are you smoking?
The explosive range in air of methane is 5-17% (LEL-UEL)
The explosive range in air of hydrogen is 4-75% - and it requires a MUCH lower ignition energy source (you can set it off with a red-hot object. Methane needs a spark)
(The minimum ignition energy (MIE) of a hydrogen–air mixture is only 0.019 mJ, whereas that of other flammable gases such as petrol, methane, ethane, propane, butane, and benzene is usually on the order of 0.1 mJ according to Lewis and von Elbe)
When I was a kid we had manometers in most towns storing coal gas - a mixture of flammable and and inflammable gasses which set to feed the local town cooking and heating and suicide facilities - it had levels of CO high enough to make it easy to bump yourself off. I know of no lightning induced associate with manometers.
As a storage method its extreemely cheap and even though 50% H2 coal gas never seemed to have this imaginary leakage problem that's been going around of late.
Hydrogen manometers have negligable pressure in them and they were being constantly topped up
Modern reticulated gas systems are operating at substantial pressures in the distribution network and even in street mains, with pressure reduction to a 2-3psi (or less) just before the residential meter - which is why you don't see pressure dropoff when everyone fires up their central heating, (such dropoffs were a regular feature of the old days of town gas)
How about we just don't tolerate AI ramping up data centre energy usage by several factors?
As domestic users of energy we're being expected to invest heavily in energy saving measures, better insulation, different heating technology, alternative fueled vehicles.
Why is Big Tech not being held to the same standards?
We should put a levy on the usage of the energy according to societal benefit. If its medical research on the cure for cancel then fire away, take what you need, its free. But if its just for banal entertainment reasons (cat videos, zoom backgrounds, life hacks) then you should pay through the nose. Of course no one will honestly report their usage. And "mental health". I *need* cat videos to feel ok about my existence...
"We should put a levy on the usage of the energy according to societal benefit. If its medical research on the cure for cancel then fire away, take what you need, its free."
Devils advocate: Why is curing cancer a better than cat videos at a societal level? In Europe we have a problem with increasingly long lived populations who want unfunded pensions to pay out from unduly early ages. Most people who die of cancer are over 75 (source Cancer Research UK). Ergo a cure for cancer means more mouldery oldies who'll live for longer, and just die a bit later of some other cause.
If you want a cause that is benefit to society, use AI to develop a gene for the human population so that they think lettuce is as good to eat as chips. That will improve the health and life quality of a good number of working age people, and reduce their chance of cancer.
'..use AI to develop a gene for the human population so that they think lettuce is as good to eat as chips.'
Surely this could be seen as eugenics though.. possibly 'new eugenics' if you give the parents the right to decide if their offspring have the gene, but it would still create 2 populations one superior to the other?
"In Europe we have a problem with increasingly long lived populations who want unfunded pensions to pay out from unduly early ages"
The problem with the implicit assumption that these people should be working is that there are enough jobs to go around.
"Full employment" went away a very long time ago - as evidenced by the disappearance of child labor and sweatshops (Social protests have never managed to shut down ANYTHING that was profitable. It's only when the economics are marginal that they can suceed)
Your right. Currently the cost of energy is disproportionately more expensive for low energy users. That is surely totally upside down to the need to limit use.
Take the UK's way of dealing with the hyperinflation of energy. They capped household bills. The level it was set at meant nearly half the population who have always kept a very tight rein on their energy use, never did get close to the cap, and therefore had no support through that mechanism. It was the families with outdoor heated swimming pools that took the largess of that tax funded cashback.
"As domestic users of energy we're being expected to invest heavily in energy saving measures, better insulation, different heating technology, alternative fueled vehicles."
For the moment.
A nuclear fission (molten salt) system has the potential to reduce build/operating costs by around 80% (in addition to all the other advantages) and that kind of cost reduction would make it cheaper than EVERYTHING else (Current water-moderated uranium nuclear started off more expensive than coal and has only become more expensive - it really only existed as a way of justifying the uranium separation process and was pushed for military ends (Whilst we refer to uranium enrichment, _depleted_ uranium is the precursor material to weapons plutonium production. the concentration on enrichment and "weaponisation of civil power" is a classic distraction technique from the REAL weaponisation paths)
"Flow batteries typically have a higher energy efficiency than fuel cells"
"cycle energy efficiency (50–80%)."
https://en.wikipedia.org/wiki/Flow_battery
Flow batteries allow "fuel" to be brought-in by truck or pipeline, but they don't NEED to unless there's an unexpectedly long outage. Instead they can have enough storage tanks on site and just "recharge" them while electricity is available which is far cheaper/more efficient than hauling anything in.
Plus, if you were a truck driver, would you rather haul electrolyte, diesel, or hydrogen?
A battery can never have a higher overall efficiency than a fuel cell because they're not comparable. One uses energy, the other stores energy.
In the context of a system implementation you can compare the system efficiency, but that's a whole different kettle of fish.
There is already a solution for smoothing out renewable peak and troughs and that is rotating stabilizers. They are already running in Scotland and Australia.
They are basically large motors/generators that can quickly be run up and then switched to generate mode.
The fuel cell idea may work in small scale, but it is untested technology and there maybe scaling issues. On the other hand they may provide an option for longer term storage so I could see a combination approach. Rotating stabilsers providing short term smoothing and hydrogen cells storing medium term excess
You seem to be conflating "synchronous condensers" aka rotating stabilisers, with flywheel-based energy storage.
Synchronous condensers are useful for adding AC inertia to a grid, stabilising the frequency, and they are quite well established for that. But they do not store a significant amount of energy. They will spin up and down in minutes or even seconds. Certainly not hours, never mind days
Flywheel energy storage on the other hand, intends to store a much greater amount of energy, and is not well established. It's also incredibly dangerous, far more so than the equivalent sized battery. When flywheels fail, they fail suddenly and spectacularly.
Flywheels absolutely cannot "smooth out renewable peaks and troughs" as you claim. The best they can do is prevent "grid islanding" by keeping the frequency stable for the two seconds that it takes for the grid to re-route following a fault.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
