Hacker News Reader

Comment by causi | Hacker News Reader

5000? Have there been significant improvements? Last I heard the expected cycles were something like 1000.

Lithium iron phosphate batteries have drastically improved cycles at the expense of energy density (approximately 14%). I'm seeing multiple sources state that they can achieve over 10,000 cycles in optimal conditions.

I assume you mean that energy density is about 14% less (86%) rather than having 14% of the capacity of lithium ion.

I just wanted to make that clarification in case anyone who isn't familiar with LFP is misled into thinking that LFP-based cars would have a range of like 30 miles.

yup. lfp can cycle every day for over 12 years and still retain 80% capacity. there are a number of home battery vendors doing 5 year 5000 cycle/80% capacity warranties for their LFP batteries already on the market.

Still not there yet. For grid storage we need batteries with 30+ years lifespan in harsh very hot/very cold weather conditions. They also need to be much cheaper per kW/h than current lithium ion tech.

> For grid storage we need batteries with 30+ years lifespan

Why? Lifespan is basically irrelevant even without considering that battery technology continues to rapidly improve: the thing that matters is the cost of upkeep.

> in harsh very hot/very cold weather conditions.

They don't have to be capable of being placed in every single possible location: the whole point of the power grid is that generation doesn't have to be colocated with consumption. But even if they did, EVs have the same constraints, except you can't design an EV to be permanently in the shade.

> They also need to be much cheaper per kW/h than current lithium ion tech.

This depends on what they're competing against. Even current batteries are cost-competitive in some scenarios (i.e. ones where you can't do pumped-storage hydroelectric, either because you don't have water to pump or it'll take too long to build).

>They don't have to be capable of being placed in every single possible location

Yes, but I'm looking at this through the lens of where I live. My country has 40 times less GDP per capita than the US, and electricity has to be (and currently is) very cheap for people to afford it. Also, my country is very hot in summer and can be very cold in winter. It is a dry Central Asian region with up to 45-47C in summer and -10C in winter. Also, we had -27C last January which was a record cold and was a disaster.

Here we have very good conditions for solar. Solar has become very cheap, and we are building several 100-400 MWh plants with more of them coming in the near future. But multi GWh battery storages for solar plants will cost billions. If we need to replace it in 5-10 years, this is not viable for us. But 30+ years or at least 25+ years is definitely worth looking at.

Or, even if the life span is 10 years, it needs to be 3x cheaper per MWh storage than 30+ year option. But in that case, we'll have 3x more stuff to get rid of/recycle/pollute the environment.

Nobody needs either of these technologies for stationary storage, where gravimetric and volumetric energy and power density are all but irrelevant.

Why do you need such long lifespan for grid storage? It’ll be much easier to service and change out modules in a grid storage plant than in a car. Most likely, when recycling tech is more widespread and efficient, you’ll want to recycle the cells quite often to stay at a high % SoH to get the most out of the material, and to exploit the latest chemistry advancements.

Obviously li-ion doesn’t need to get cheaper to be used for grid storage. It’s used profitably for that purpose today. What kind of grid storage are you talking about?

Li-ion is already excellent for frequency regulation and short term energy storage. Li-ion has never and will never be a solution for long term seasonal storage. Then you’ll want pumped hydro or other gravity based storage, flow batteries, thermal storage, liquid air battery, etc.

I think liquid metal batteries might outcompete li-ion in the end. And they do probably have 30 year lifespans with little to no maintenance. But it’ll be the cost and rapid scalability that’s the critical factor.

We want. But we don't need 30 years any more than we "need" 20 or 10. It's economics against utility. We could deploy with 5 years, and replace the short life elements across the long life of the other infrastructure. Batteries are modular.

The perfect is the enemy of the good enough. Many things deploy at scale before true economies of scale are reached.

> Still not there yet. For grid storage we need batteries with 30+ years lifespan in harsh very hot/very cold weather conditions. They also need to be much cheaper per kW/h than current lithium ion tech.

No, not at all. They can be adequately shielded from the elements, they don't need to last so incredibly long, and even currently priced Li/ion batteries make economical sense today in given scenarios.

What's your rationale for these criteria?

Because I live in a third world country which is also very sunny, dry and hot like Arizona in the US.

And while solar has become cheap enough that our utilities are starting implementing it, the storage is still too expensive if you want GWh scale storage. It can literally cost billions, and considering we may need to replace it in like 5-10 years, it is not a viable solution for us. We just can’t afford it.

LTO batteries might be able to do 30 years, unfortunately they cost more than lithium.

I'm actually kind of surprised/disturbed they aren't used in consumer gadgets a lot, since they are affordable in small sizes and the battery is the main thing to wear out on a lot of products.

The volumetric energy density on LTO sucks ass, even before considering the inability to go to dozens of vendors and order a million lipo pouch cells in basically whatever rectangular dimension you want for dirt cheap, or a custom shape for probably still significantly cheaper than an LTO cell would cost you.

Reply on news.ycombinator.com