Grid scale buffering with molten salt batteries that run at 500 Celsius | Hacker News Reader

Grid scale buffering with molten salt batteries that run at 500 Celsius

ambri.com

56 pointscar19 comments

I’m so used to seeing “max cycles”, “cycles to 80% capacity”, etc figures that I was a little confused I couldn’t see such a figure mentioned. According to https://ambri.com/benefits/ they essentially do not degrade. Neat.

It appears their main selling point is cost. Given that, they are suspiciously shy about giving an actual figure. The only thing you can find is:

> Even compared to this low future lithium-ion price of $100/kWh, these energy storage systems will be significantly more expensive than Ambri-based battery systems.

$17/kWh

https://ambri.com/benefits/

Edit: this appears to just be the cost for the price of the cathode and the anode...which makes sense since the pricing on the complete unit would probably vary from customer to customer.

This is an interesting system. Ambri says:

> Ambri-based systems generate their own heat during use, thereby eliminating the need for auxiliary power for temperature control. These systems like to be used – a full charge/discharge cycle at least every two days will keep the system at its operating temperature and higher duty cycles will not increase degradation.

And:

> Capacity: 1000 kWh, 250 kW

(per container), so it charges or discharges in 4 hours.

So it’s not terribly useful for seasonal storage, but it’s fine for solar or wind with output variations by time of day. Suppose you amortize the cost over 10 years and you cycle once per day, that $17/kWh comes out to 0.4 cents/kWh. Figure a factor of 10 for costs missing targets and the balance of system, and this storage costs around the LCoE of solar.

So, in effect, one could use this to turn a solar installation into a solar+storage installation that can spread the output over a full day without even doubling the overall cost.

This is pretty awesome! Too bad it doesn’t do much about winter or a cloudy week.

If you scale it up it can probably deal with a week. People have pretty bad intuitions about heat loss for big objects because most hot things in our lives are small, but since heat loss follows a square cube law, every 8x volume increase gives 2x decrease in lost heat. It's also important to remember that this doesn't need to be anywhere near 100% efficient. Given the swings in energy prices, you can probably make a profit at 50% efficiency.

Except it’s a bunch of shipping containers. They scale horizontally but not vertically (sorry). And you probably need to leave clearance between them, for service if nothing else, so the thermal performance is what it is. You could add insulation.

I totally believe that the risk of this stuff is "less" than an uncontained lithium fire, but 500C is nothing to laugh at. There are significant risks to the people around this, contact with water, water vapour, volatiles released from plastics.

I don't want to "but.." because lots of industrial processes are inherently dangerous, and we still do them. The thing is to design the systems to contain the risk. Thats cost, and may go to the same place as the containment costs for lithium systems. For the right engineering outcome in terms of operational cost and responsiveness, it's worth it. For energy density, it may be really good value.

They say that after initial heating, the process is inherently exothermic and so maintains it's own heat. Thats beneficial because it means external heat sources don't have to be continuously maintained.

Antimony is poisonous. Assuming normal operations don't volatalize this, or create dust and aerosols, the problem would be what happens if there is an uncontained fire, or some problem which causes it to be released in smoke or water. Again, not a problem unique to this technology, proposed energy storage with Ammonia is going to have to confront NH fume risks. (for instance)

These aren't things people will build out in their back yard. Things like flow batteries are much more deployable into rural and remote: maybe this technology is deployed in light industrial, heavy industry or at the electricity substation and plant.

The good news is that there's a really easy solution to 500C being dangerous near people: don't put it near people. You don't want one of these in your car or your basement, but you can just put a ton of them in a big gray building that isn't next to anything and it's pretty much completely safe.

> maybe this technology is deployed in light industrial, heavy industry or at the electricity substation and plant.

With all the wonderful people shooting at electrical substation in populated areas maybe this isn't the best place? They aren't well secured and there is already enough risk there from the high voltage - not to mention the end user impact when they go off line.

While 500C is a lot when compared to the average process in a kitchen, on an industrial scale it's not really even all that impressive. Any steel mill deals with temperatures over twice as high. Never mind that a steel mill does a lot of moving around of steel in crucibles and such which is the main danger, which I don't reckon would be required for this process. The local trash incineration facility gets hot enough to melt copper, which is over twice as hot as this molten salt process.

Heck, the inside of your average wood stove gets up to these temperatures.

Painting 500C as some sort of very high temperature that's hard and dangerous to contain isn't quite fair, seeing as there are myriad industrial processes that are generally considered safe that run at much much higher temperatures.

My grill can hit 700 if I want to make a pizza. Handwringing over this is ridiculous

$17/kWh and no capacity degradation. Daaammmmmn! For comparison the cheapest LiIon batteries are like $150/kWh today, and degrade significantly over a decade.

I guess grid scale storage is a solved problem then? And Tesla battery wall is pointless.

Do you mean the Tesla power wall? This thing has quite a large footprint (10ft container) so I can see the power wall still being used household backup solution.

I think he means the mega/gigapack.

These are for buffering/load-balancing, not storage. Note the <500 milliseconds response time, that's way faster than pump hydro can respond, even with compressed air.

A typical UK household uses close to 50kWh of energy a day, most of it on heating and cars. So, in a zero-carbon future, you'd need a 10-foot shipping container per 20 homes to give them each a single day of storage. You'd need >10x that to ride out a windless cold snap (common in the UK where I am—we even had 2 weeks of no wind power in the height of summer last year so ran 60% gas during the hottest week ever recorded).

Grid-scale storage is tens of TWh, still way beyond any battery technology. These are MWh-scale systems. Useful, but not grid-scale storage. A grid would need tens of millions of these. At $17/kWh we're talking $500 billion.

They would be useful for keeping the air-con running on hot nights in solar-powered countries though.

Scaling up manufacturing, and finding places to put (bury) these large batteries, will be a significant effort, but seems straightforward enough that projected future capacity estimates are probably pretty reliable. The hardest part will be effective load balancing strategies, IMO. (Very excited for distributed optimization systems research in this area, personally.)

In the real world you can't just announce theoretical cost and then say 'we win give us the money'.

Ambri has existed for 10+ years now and they are only gone deploy their first grid battery in 1-2 years.

They might have a great product, but they are far away from impacting the market to such a degree that any other battery company is 'pointless'.

Cost is not the only variable. Efficiency and charging/discharging capacity and time are also important metrics.

Basically storing heat and converting it back to electricity, you are going to lose quite a bit of energy. And that's before you consider cooling effects. Heating up a large mass is not particularly fast. It takes some time. And then getting a lot of energy out rapidly is also not really that easy.

A lithium ion battery has probably well over 90-95% efficiency. It keeps its charge. And you can charge/discharge rapidly. Basically, you cans switch them on fast enough that switching from the grid to a battery when there's an outage, you probably won't notice the switch over.

Both have their strong and weak points and there is probably going to be a use for both on the grid.

This is the company cofounded Prof. Donald Sadoway from MIT. I have been following their progress for a while. He made a lot of media around this technology 10 years ago. See the TedTalk:

https://www.youtube.com/watch?v=Sddb0Khx0yA

So this is one of those case where you hear some news about amazing battery tech, and the reality is it minimum 10 years before it ever on the market.

The company had ups and downs and it seems they are actually pretty close to actually deploying a major system within the next year or so. This if further then most other grit batteries companies get.

This is a bit of newer talk with some more information: https://www.youtube.com/watch?v=p2N3QAMhtPU

He also worked on a really cool idea for Green Steel making, producing iron from iron ore like we do aluminum from Bauxite. The startup company for that is called 'Boston Steel'.

Short video: https://www.youtube.com/watch?v=-pYNzxorJs0

Full presentation for more technical people: https://www.youtube.com/watch?v=ZhQaxFZOptE

Reply on news.ycombinator.com