Hacker News Reader

Most flywheels in use are on gensets and only last a few seconds, keeping the generator spinning until the genset has ramped up to speed.

There are standalone flywheel systems but they're not nearly as common now that genset flywheels are around, in part because they have atrocious recovery time, they're far more complex and expensive (they tend to spin at significantly higher speeds, do so in a low-viscocity gas or vacuum, etc..and don't utilize a power unit that already exists) and they cannot help start a generator whose starter has failed.

Their main advantage, like you said, is that they don't degrade with age or use, but they're also compact and far less hazardous than large battery rooms.

Their major downside is that their recovery time is atrocious. There was a huge outage at an SF datacenter many years ago that was caused by multiple short outages. The flywheels didn't have time to spool back up between outages and eventually didn't have enough kinetic energy to carry things until the generators came back on.

Pumped hydro operates on a scale orders of magnitude larger than flywheels and batteries and has a response time of seconds, which is plenty sufficient for a grid-scale operator, because the grid has enormous inertia. It's the fastest responding form of large-scale power generation (natural gas takes minutes or more, nuclear takes days.)

One of its major successes has been in the UK, where for decades pumped hydro has helped the grid sustain "TV Pickup" - everyone flipping on their electric kettle between BBC programs.

Yeah, pumped hydro has response times on the order of 10s of seconds, but the capacity is only useful for short term generation. For example, the UK has 24 GWh of installed PHES capacity. Last year the UK used ~915 GWh per day on average, so best case about ~40 minutes worth of consumption.

PHES is for peaking and grid stability, not long term storage.

What is long term storage in the context of renewables? Scaling UKs PHES capacity times 40 so that it lasts a whole day doesnt seem like an impossible task. I might be naive, but it seems extremely possible, given how those 24 GWh capacity were achieved with little to no political support.

Given the ridiculous volatility of energy prices in many places of the world recently (notably Europe), I do wonder why this isn't seen more as a proper investment opportunity?

If you can buy cheap and sell for 4x the price the next day[1], that ought to cover the infrastructure costs very rapidly right?

[1] e.g. https://www.nordpoolgroup.com/en/Market-data1/Dayahead/Area-...

I’ve looked into this. Even sent an email to Amber Kinetics (the flywheel company in the OP) some months back asking for some basic economic parameters, but didn’t get a reply.

My conclusion was that simple electricity arbitrage using refurbished lithium ion batteries would be reasonably profitable right now but that business case depends on volatility staying high. The CAPEX is approximately 500 EUR per kWh storage capacity, and you can maybe make 0.25 EUR per kWh and day doing arbitrage. It’s hard to know what the electricity market will look like in the relevant timeframe.

5.5 years payback time? Yeah, I don’t expect current volatility in European markets last that long. It might, but I don’t expect it.

Yea, that’s the issue. You also need some more time to pay interest and make a profit.

You can do the maths for this yourself to find out :)

Big utility battery installations are around $400/KWh IIRC and can cope with at least 2000 cycles[1]. Looking at the table you provided it looks like you'd be able to do around €50-€100 EUR per MWh sale per week on average if you're good at buying low and selling high but lack a crystal ball. So €0.05 per KWh and 2000 cycles still puts you in the red on your battery installation. :(

[1] https://www.pnnl.gov/sites/default/files/media/file/Final%20...

This is why most BESS's in Europe (and elsewhere) do revenue stacking to be profitable. This means they're not just doing arbitrage of intraday prices, but also providing ancillary services to the grid (so if there's e.g. a fault that causes the grid frequency to drop, a BESS will kick in and start to discharge at a power rate that stabilizes the grid frequency).

There's actually even more volatility than what is revealed in the dayahead table. Check intraday prices. They predictably go down at night and up at day (because while production is unpredictable, consumption is not).

Because anything spinning at high RPMs and containing rotational energy is a massive danger to anything around it in any case of failure.

It's not a much more devastating failure mode than hydro dams, nuclear power plants, even wind power plants.

Besides, a flywheel is perhaps best operated underground where the bedrock can absorb the energy if it should fly apart.

The only installations that I have seen all had significant containment. Also, the rotors were densely wrapped fibers and on impact became a tangled mess that absorbed the kinetic energy. They don’t fragment into sharp, hard chunks.

Massive danger according to whom? By that logic airplanes should not exist.

Pumped hydro needn't bee toooo slow, here's one in wales

"From standstill, a single 450-tonne generator can synchronise and achieve full load in approximately 75 seconds. With all six units synchronised and spinning-in-air [...], 0 MW to 1800 MW load can be achieved in approximately 16 seconds.[13] Once running, at full flow, the station can provide power for up to six hours before running out of water"

https://en.wikipedia.org/wiki/Dinorwig_Power_Station#Specifi...

16 secs ain't too lazy. IIRC they have a spun-up metal flywheel to take the load in those few seconds before handoff to the hydro, though I can't find a reference.

TIL: “The monitoring of popular television channels is an important factor in electricity grid control centres”

I'd say it's critical in the UK: a combination with popular TV programmes and the UK's citizens being likely to own an electric kettle: https://youtu.be/WCAzalhldg8

Millions of people turning on water heaters within a few minutes will have an impact on any electric supply.

Oh, very much so! Ad break during popular soap operas and a lot of kettles go on simultaneously around the country. It's very well known to electricity suppliers.

This makes me think of something...

So, when a TV studio broadcasts, the electricity for everyone to watch it is paid for by the viewers. The studio only pays for the power to either broadcast from the TV antenna, or the power to send the signal to the cable company.

But when you watch a broadcast on the internet, not only are the viewers paying for it (their internet), but also the company is paying for the internet on their end (all clients connecting at once, using all that bandwidth), as well as all the servers to handle all those connections.

If TV broadcasts worked the way the internet did, a broadcasting company would have to be able to handle the incoming power load of every household's power bill combined, simultaneously.

It seems like the internet is poorly architected! The company should be able to send its one broadcast stream out, and it should be distributed to all the client machines, without the broadcast company needing to directly connect to (and duplicate) the signal across every client themselves.

In the early internet, where every device has a public IP and the only firewall is the one you should've set up, multicast penetrated through all networks and a single packet stream could be subscribed to from anywhere, replicated across the internet.

These days, only IPv6 capable networks (so half of the web or so) satisfy the necessary requirements for such a system and internet multicast has wisely been turned off for the enormous DDoS/bandwidth waste it implies.

This mechanism is still used on some TV networks, though, especially digital ones that come over fiber. There is a single stream of packets generated to send to all subscribers that the subscriber devices can then subscribe to with the proper network config. This is often accomplished through IGMP and other such multicast protocols.

As for sending a single broadcast stream out, that's exactly what online streaming services do. A 30mbps Twitch stream with a million viewers doesn't require you to get a data center's worth of internet capacity at home; instead, you upload a single stream to your favourite service and that service replicates the stream for you. You can set up such a system yourself if you want to stream from home, have a cloud server with good internet, but only cable or DSL upload speeds through somerhing as simple as nginx with RTMP enabled.

There are multicast protocols that do that, but I have no idea how widely used they are.

Please consider the implications of your suggestion

Aren’t on demand TV services going to ruin this prediction capability?

Possibly, who knows! On the one hand you lose a clear predictor of sudden power peaks, but on the other hand the ability to pause/resume/pick your own time probably spreads out the usage more across the day so the kettles have less of a direct impact.

I wonder if these days the grid operators also monitor internet statistics in some way.

This might be more applicable in 120V America, but it would be nice to have a kettle that charges a battery for a quicker boil when it needs to be used. Given that, it might not really matter if everyone used their kettles at once, since they would just recharge slowly afterwards.

You're forgetting compressed air storage in caves, which is the one you want for the weeks times scale if you want to go 90% renewables.

I didn’t know how fast it is. I assumed faster than water, but I couldn’t find any reference.

I’ve also read (somewhere) about pumping air into underwater bladders. Sort of a reversed pumped hydro.

How about leaks and thermal losses?

Renewables are so cheap that you need ridiculously large losses before needing to care very much.

My personal preference is a global HVDC power grid, which would be fine from a technical point of view even with current standard cables. (There are non-technical problems, but 60% resistive losses are genuinely ignorable given how cheap optimal PV is).

I have not looked into pressurised air storage, but I can easily believe it’s also something where the losses, whatever they are, are just not a big deal any more.

gravity (notably pumped hydro) is cheap for even more capacity but has limited opportunities, and it’s not nearly as fast to trigger as the other, so it’s better used to store capacity over one week. Evaporation is a problem, but cheaply fixable.

I wonder whatever happened to the earth piston idea I saw floated a bunch of years ago. It involved cutting a cylindrical piston in the earth and drilling beneath it at an angle. You then pump water down into the space beneath the piston, causing it to lift. To extract the stored energy you simply let the pressure provided by the piston force the water back out through a turbine.

The main trick to getting it all to work is with the surfaces. You want the piston to be able to slide up and down smoothly without water infiltrating the earth or going around the side of the bore, so the whole thing needs to be sealed and smooth like an engine cylinder. The other trick is for your pumps, valves, and turbines to be able to operate reliably with huge water pressures.

I haven’t heard about the idea in years so I have no idea how it went, though I would guess it didn’t pan out for some reason.

It sounds like quite an engineering challenge and a lot of ground to excavate to store a significant amount of water.

Geothermic energy and CO2 storage sound a lot more useful if you’re going to dig that much

I think the idea was to avoid excavating all the ground in the middle. You just cut around the circumference of the piston and along the bottom. The thing I never could understand was how they’d seal the walls of the cylinder against the bore so that water doesn’t fill that space and gush right out the top!

That sounds massively expensive. Being very generous and assuming that the piston is the length of the hole and can rise out completely, then the energy storage is still only a few times that required to pump all the water out of the hole. It doesn't start to compare to the volume behind a hydro dam (which may then have a large vertical drop to the generator further down-river), but is still much more complex.

Another proposed idea for places that have deep seas or lakes is pumping air down into a storage at the bottom of the water. This storage can even being a flexible plastic - there isn't any high loading on it because the pressure balances out. The issue with that tech is that it's not that efficient, as compressing the air going down generates a lot of heat - some of that could potentially be recovered, but there's a trade-off with simplicity. Also if the containment fails then a lot of air bubbles to the surface, potentially sinking any ship on the surface at the time.

As ben_w points out elsewhere in the thread, renewables are so cheap that losses matter less than you think.

Put another way: Today in Denmark, electricity is free. Literally 0 cents (øre) before taxes and transport fees (of about 5 cents, 36 øre). Just before Christmas it was about $1 (700 øre) per kWh. There's no inefficiency where it would be bad to store energy with swings like that.

> batteries (notably LFP) ... but expensive to scale with power electronics;

Can you please elaborate on the scaling costs ?

I have been told that packs will output 400vDC as-is. But turning this into a grid AC connection is "complicated" ?

Scaling their Amperage is expensive.

Putting them in parallel increases the amperage. My guess is they mean the inverters are expensive.

Batteries put out DC - the grid is AC.