Hacker News Reader

Comment by java-man | Hacker News Reader

The sooner we switch to solar and grid battery storage, the better. There should be no excuse whatsoever not to do that now.

What reasoning is holding us up?

View on HN

> What reasoning is holding us up?

People turn their lights on when the sun goes down, and efficient energy storage is the holy grail.

Also fun back of the napkin math with overestimates, if we have a totally efficient Li+ battery with 0.25 kWh/kg energy density, and converted all the world's 40 billion kg of lithium in the crust and oceans to batteries, we'd get roughly 10 TWh of energy storage capacity. Which is enough to power California for 19 days (we generate around 200TWh annually).

In reality we only have access to a tiny fraction of the world's Lithium reserves, batteries aren't that efficient, they're heavy, dangerous, and all that makes for an expensive technology ill suited for grid energy storage (and alternatives like lead-acid are even worse).

Energy storage on the grid is going to look more like dams and reservoirs in the mountains than battery arrays. Even that's not great, ideally we'd invest more into nuclear energy for generating clean power on demand.

According to [1], a standard 2 Ah 18650 Li-ion cell has something like 0.6 grams of lithium. (2.0 Ah * 4.2 V / 0.0006 kg of lithium) works out to 14 kWh/kg or 560 TWh of capacity.

That one battery chemistry alone would be enough to power the entire world for a week. Once you eliminate nonessential industrial power consumption - aka bring your own power storage or shut down machinery with the sun - the amount of lithium we'd need to store to make a transition to solar/wind feasible would be far smaller than global reserves.

Edit: as noted by phillipkglass, this math is moot

[1] https://batteryuniversity.com/learn/archive/is_lithium_ion_t...

You are confusing present estimates of lithium resources (a term of art used in the mining industry) with terrestrial abundance.

Your estimate of 40 billion kilograms roughly corresponds to the USGS estimate of world lithium resources in 2015 (39.9 million tons):

Mineral Commodity Summary, Lithium, 2015 [1]:

https://s3-us-west-2.amazonaws.com/prd-wret/assets/palladium...

Note that this year it has risen to 80 million tons:

Mineral Commodity Summary, Lithium, 2020:

Owing to continuing exploration, identified lithium resources have increased substantially worldwide and total about 80 million tons.

https://pubs.usgs.gov/periodicals/mcs2020/mcs2020-lithium.pd...

And at the turn of the millennium identified resources were about 13 million tons:

Mineral Commodity Summary, Lithium, 2001:

https://s3-us-west-2.amazonaws.com/prd-wret/assets/palladium...

The Earth's mass changes negligibly over a few decades. The reason lithium resource estimates change so much over a short time is that people have put in the effort to identify and quantify specific sources of lithium. That's what converts preexisting geology from not-a-resource into what miners call a resource.

In generic geochemical terms, the Earth's crust is estimated to contain 17-20 parts per million of lithium by mass:

https://en.wikipedia.org/wiki/Abundance_of_elements_in_Earth...

If the mass of Earth's continental crust is 2.1 * 10^22 kg, as given in [2], then lithium in continental crust masses about

(2.1 * 10^22) * (17 / 10^6) = 357,000,000,000,000,000

kilograms, or 357 trillion tons. There's another couple hundred billion tons of lithium dissolved in the oceans.

[1] USGS lithium statistics and annual reports from https://www.usgs.gov/centers/nmic/lithium-statistics-and-inf...

[2] https://ui.adsabs.harvard.edu/abs/2007AGUFM.V33A1161P%2F/abs...

Are you saying 19 full days of energy storage? Why do we need 19 days? Isn't it more like N hours per day, which can be recharged overnight and the following day? This seems like a very disingenuous presentation of this information.

I remember seeing academics analysis of various storage methods that showed how none of them could be scaled up to store 7 days worth of energy. And thus solar and wind were infeasible.

There were two problems.

One he was using 7 days worth of storage. It's true that in the fossil fuel industry 7 days of storage is standard. But that's because coal/oil/gas have supply chain issues to deal with. So he was intentionally increasing the required storage 10 fold.

The second. He studiously avoided any mention of thermal storage methods. Or using natural gas as a back up. Both of which can be scaled up to meet his 7 day storage requirement.

I see I pissed someone off. Whatever, you are who you are.

https://www.pv-magazine.com/2020/05/22/borehole-thermal-ener...

Also, we don't have to limit ourselves to lithium. Li makes sense in vahicular applications, but the grid storage could be perfectly fine with Na or any other technology.

Here is, for example, a sodium ion chemistry based on prussian blue:

[0] https://natron.energy/technology/

We need to quickly get to the next generation chloride ion battery batteries! A whole lot more salt lying around than lithium.

Interestingly, batteries don't have to be lithium-based.

https://qz.com/1355672/stacking-concrete-blocks-is-a-surpris...

The point still stands that storing energy in chemical batteries for CA-levels of usage does not math out.

Anti-fossil fuel mandates imposed by the California legislature is part of the reason there is a shortage of power in various areas. Rushing the transition without enough alternative supply in place is a recipe for further problems.

https://www.wsj.com/articles/californias-blackout-warning-11...

https://www.forbes.com/sites/michaelshellenberger/2020/08/15...

They retired base load power (nuclear) and didn't replace it with other power sources that can provide base load. Instead everyone pushed for renewable only and opposed the natural gas power stations.

Build 5,000 MW of natural gas peakers and this problem immediately goes away.

So California has one-upped Germany: instead of replacing nuclear with coal, they're replacing nuclear with rolling blackouts. Ugh.

And to think, if we just hadn't stopped building nuclear, we would already be rid of CO2-intensive electricity. Instead, we've barely begun, and we've only begin through duck-curve denial. Facepalm.

California never had much nuclear to begin with. Only 2GW of capacity has been decommissioned so far this millennium.

The article states that there was no such shortage...

There was no such shortage if you only look at instantaneous figures and ignore the fact that a substantial proportion of the grid's generation capacity was in the process of rapidly disappearing due to a predicable daily process called "sunset", as the article does. See my other comment. I suspect the person you're replying to might be right.

Grid scale battery storage only became practical in the last few years. Now there are a lot of projects under way but they'll take time between approval and completion. One notable example from last week:

"Vistra approved to build a grid battery bigger than all utility-scale battery storage in the US combined"

The past month has been littered with news of exceptionally large battery storage developments, yet none in the world can compare to the news that Vistra’s permit to expand an energy storage system under construction at its natural gas-fired Moss Landing generation station in Monterey County, California to 1,500 MW/6,000 MWh has been approved.

https://pv-magazine-usa.com/2020/08/13/vistra-approved-to-bu...

Does not have to be one giant storage plant. The installation could be decentralized:

[0] https://phys.org/news/2018-02-tesla-australia-homes-power.ht...

Centralized is cheaper to install and maintain, though. Rather than dispatching an electrician to a home every time its battery has a problem, a farm keeps electricians on-staff and replacement parts in-stock. Installation is cheaper for the same reason.

Up to a point yes. Decentralised is also nice if co-sited with substations because the utility can use it to flatten out spikes/dips and generally make other parts of the infrastructure happier, which reduces maintenance costs on those other parts.

When the article calls this a "landmark plan", they mean it's an experimental new plan. Power technologies need to be well-tested, known to be reliable and efficient, if they're going to be deployed at scale.

The first 300-MW phase is planned for completion by the end of 2020, with the second, a 100-MW expansion expected to come a year later in 2021.

> solar

I can't seem to find it right now, but someone in another post noted that peak power demand in California is several hours later than peak solar production, so supposing you have storage, it doesn't really matter what the source of the energy is.

> grid battery storage

I'm really curious if this is more or less expensive than adding production capacity.

California is famous for it's duck curve [0] where peak load & peak solar generation are not correlated. There are many possible solutions, some are technical and some are not. I am currently building one of them [1].

0. https://en.wikipedia.org/wiki/Duck_curve

1. https://www.greentechmedia.com/articles/read/opus-one-tests-...

That's still a pretty easy situation compared to many places where the storage need is over months and not fractions of days. (Though those places can often do a lot shifting with hydro)

Yes, pumped hydro is a very efficient solution for long-term storage. Plus, it is quickly available when needed (on the order of seconds).

While the solutions seem simple, putting all the pieces together is the real hard part. It's pretty common to have no data available on the distribution grid and the sensors themselves are expensive and then you need all the related infrastructure.

California gets 12% of its power from hydro, but I'm sure they're already shifting that production around to compensate for solar.

Time and money.

I doubt it: Musk installed 129MWh battery plant in Australia that already saved the consumers over $100M.

There must be another reason.

[0] https://www.dailymail.co.uk/sciencetech/article-8082841/Elon...

That plant can only store a few minutes of power. It's not a solution to blackouts, just stabilising the grid while reliable sources come online.

MW is a funny unit in this context. Apparently Daily Mail meant MWh.

> It provides a total of 129 megawatt-hours (460 GJ) of storage capable of discharge at 100 megawatts (130,000 hp) into the power grid, which is contractually divided into two parts:[20][needs update]

> - 70 MW running for 10 minutes (11.7 MWh) is contracted to the government to provide stability to the grid (grid services)[21] and prevent load-shedding blackouts[13][22] while other generators are started in the event of sudden drops in wind or other network issues. This service has reduced the cost of grid services to the Australian Energy Market Operator by 90%.[23]

> - 30 MW for 3 hours (90 MWh) is used by Neoen for load management to store energy when prices are low and sell it when de

https://en.wikipedia.org/wiki/Hornsdale_Power_Reserve

Time, money & NIMBY. :p

Pricing and regulation, we need to get out of pretend flat-rate electricity ASAP. This will probably lead to generously overprosivisioning solar & wind + long distance HVDC transmission lines or energy storage for the overproduction.

I don't see how pricing can really fix this, because the price falls on individuals who aren't in a position to fix the grid. Do you think it would be effective by incentivizing people to buy oversized residential solar arrays and significant dispersed battery storage?

How many solar panels and grid storage batteries would you need to support a population of 39 million people running their air conditioners constantly during a heat wave for a week?

How much would all that equipment cost to buy and to maintain?

That's what's holding us up.

Cost

Reply on news.ycombinator.com