Hacker News Reader

Comment by danuker | Hacker News Reader

It worries me that if industrial-purpose batteries cost the same and have similar degradation, grid services will also decide not to charge/discharge if the profit is less than 8 cents.

This tells me that Li-Ion batteries are far from the ideal medium for grid storage.

View on HN

Most cars use NCA or NMC chemistry which has 500ish cycles. Grid storage will soon use LFP which is cheaper to begin with and has 5000+ cycles. It has a lower density than other chemistries and very poor cold weather performance, making it less ideal for vehicles (though Tesla is using it in the standard range model 3 now). Those downsides don't matter for grid storage and really change the economics of charge/discharge.

> has 500ish cycles

That's on the low end, like if you're fully charging/discharging the battery frequently.

For grid scale storage Li-Ion is a transitional technology at best. It sees interest right now because it's quick to deploy (unlike the much cheaper pumped hydro), and because so much money had been pumped into Li-Ion research and production that other battery technologies have a hard time competing.

If it turns out people don't need all the cycles out their car battery then there's value in capturing that, but new large scale installations will move to other technologies over the next decade or two.

Batteries are extremely inefficient. Capacitors are much more resilient in terms of charge cycles (by orders of magnitude).

If one were to build a "facility" to hold short term load, that's likely what they'd use. They just can't be used in cars because they're incredibly volatile and don't hold a charge for a long period of time.

The idea of super-capacitors in cars is positively horrifying, knowing how explosive just the small ones can be. At grid level though, is there anything that speaks against them?

Edit: to answer my own question, they look pretty good [1]. Significantly lower storage density than batteries (roughly 50Wh/L versus 420Wh/L for Li-ion), but still dense enough for this to be workable at grid level. The linked presentation proposes converting decommissioned power plants into grid-level capacitor storage facilities, since the transmission switchyards are often intact. Furthermore, all the technology is available today, and when built, capacitors require almost zero maintenance.

[1]: https://www.energy.gov/sites/prod/files/piprod/documents/Ses...

Todays grids don't have much need for energy storage on the sub-10 second timeframe. Ie. there is rarely any money to be made by taking power from the grid now and returning it in 10 seconds.

For technical reasons, thats because the 'spinning reserve' - which is momentum of every synchronous motor and generator across the nation - already adequately handles this.

Also, large generation stations are required to have a 'load line' which damps high frequency oscillations. The load line can best be described as 'whenever the grid frequency starts slowing down, generators must put more energy in automatically'. Things like wind and solar typically don't have the ability to do that.

For human reasons, it's because electricity markets tend to be minute by minute at most.

For all those reasons, I don't think you'll make any money with capacitor energy storage banks (not to be confused with capacitors for grid scale power factor correction, which can be profitable).

Isn't it expected that 'spinning reserve' will be decreasing because we shutting down dirty power plants? Possibly it is needed in the future.

> Batteries are extremely inefficient. Capacitors are much more resilient in terms of charge cycles (by orders of magnitude). If one were to build a "facility" to hold short term load,that's likely what they'd use

The problem is that capacitors (assuming you mean super and ultra capacitors) don't have much capacity compared to batteries, so their cost per MWH of storage is high even if their cost per MW is low.

capacitors are very low energy density and very high cost per joule of storage so they are never used this way

There are various niches in grid storage.

Li-ion are good for frequency stabilisation: initial response times of the order of ten milliseconds, run times up to several tens of minutes maybe.

For longer durations, flow batteries and other chemistries are probably better. They win because of very good cycle life and calendar life (20_000 cycles, 50 years) but tend to take longer to start up.

Flow batteries include vanadium redox, zinc-bromine, iron-saltwater (being piloted). Other chemistries: sodium-sulfur (NaS, developed by NGK and sold by BASF in Europe/NA), carbon polymer based (PolyJoule), and a great number of experimental types.

There are also thermal batteries and compressed air energy storage, and pumped hydro.

Pumped hydro is by far the biggest form of grid storage today and is not as limited by geography or cost as one might think. It can be used in the "hours to months" range of energy delivery durations.

These up and coming battery types look fantastic, but everyone seems to forget that lead-acid batteries are already way better than Li-ion or LiFePo for non-mobile energy storage, in terms of price per kWh stored.

Reply on news.ycombinator.com