Hacker News Reader

Comment by WalterBright | Hacker News Reader

5 hours of grid-scale storage seems like a lot.

On its own, it does seem like a lot.

Thinking about the nature of renewables where I live, it seems so small as to be not worth mentioning; January will often have a week to several weeks of heavy cloud cover and little wind. A mere 5 hours of capacity is a pittance.

Australia, on the other hand, could easily make more sense, as I imagine the outback to be sunnier year round and maybe windier.

Cloud doesn't mean solar stops working - the decrease is 30-60% depending on thickness, but there's still very meaningful generation.

Especially depending on the season - clouds in the summer also typically correspond with a big drop in AC demand, reducing the amount of electricity required. That's why OPs modeling is so interesting, "real" daily data for an entire year is great for this type of planning.

the solar panel's I've used in off-grid applications put out 10% on cloudy days.

We get about nine and a half hours of sun in January, assuming no clouds. That leaves around 14-15 hours of no production, assuming there isn't sufficient wind (also often the case in January).

Panels operating at 30% capacity for 9 hours cannot be offset by 5 hours worth of storage. It simply isn't enough storage, even if the panels were operating at 100% capacity. Adding more storage really isn't a great solution either; I'm guessing that you would be hard pressed to get 15 hours worth of storage out of 9 hours of sunlight when the hours of darkness start at 4:30 PM, especially if everyone moves to electrical cooking, water heating and clothes drying, which some states are starting to compel.

So, the two obvious solutions remain- import stored energy from out of state, or using non-renewables as a base load a few months out of the year. We'll need to pay for burying tonnes of HVDC lines (stringing them up in tornado prone areas isn't a great call, imho) or the expense of more nuclear plants (unpopular and, currently at least, uneconomical) or pulling CO2 out of the air to offset natural gas (also expensive).

HVDC lines seem the cheapest, though sourcing the energy will take years of production and inter-state agreements to get to the point where the south is producing (and storing!) enough to send the excess north.

You can also over-provision solar and wind to factors well above 100%, and still be more economical than burning fossil fuels.

The US can't even provision 100% capacity because of BANANA, let alone over-provision.

it also creates another problem - solar generation drop off very quickly when the sun is obscured. going from 100% to 10% or 20% in many cases. Where doe the power come from during these swings? What generation source can ramp up and down at that speed?

Synchronous condensers or smart batteries. Australia has deployed a lot of the former to deal with this problem at solar farms.

Australia uses synchronous condensers and batteries for frequency regulation, not energy delivery.

The grid is always managed such that any single plant or transmission line can trip at any time without causing voltage or frequency collapse. Clouds going over your solar farm is like the plant tripping but not as bad.

What percentage of the population lives in climates like yours? If I had to guess, maybe something like 10-20% of the worlds population?

My gut tells me the vast majority of the worlds population lives in locations that are more similar to Australia's climate than in climates like yours. So we start there, and overtime things get cheaper, we find better solutions to long duration storage, and then we start tackling the harder climates/geographies.

Many of the countries with climate similar to Australia do not have the purchasing power of most northern countries, so they're currently building new cold power at a rapid pace.

Is it really though? How many gigawatts of gas peaker plants and all the associated infrastructure sit around waiting for the tens or hundred hours they're needed?

Turns out, it's a lot - and it costs the ratepayers:

https://www.greentechmedia.com/articles/read/just-how-much-b...

> Wood Mackenzie calculates the average peaker plant capacity factor will sit between 5 and 6 percent over the next two decades. That amounts to a low number of hours for storage to replicate.

In New York City, ratepayers are paying billions of dollars for polluting gas peakers that sit idle.

https://www.cleanegroup.org/ceg-projects/phase-out-peakers/

* An estimated $4.5 billion in ratepayers funds have gone to support the continued operation of the city’s peaker plants – most of which operate no more than a few hundred hours each year.

* Even though the power plants don’t run often, they can be significant contributors to local air pollution – accounting for more than 10 percent of nitrous oxide (NOx) emissions on high ozone days.

Luckily we have the technology to both start replacing gas peakers with storage, and roll out demand response and load shedding to reduce peak demand.

For example, my employer is a big power user. On a couple days in the summer during peak heat, the utility requests curtailment. The plant managers can respond by switching HVAC operation plans, shift some loads to the overnight - the utility saves money on peak generation, and the company saves a ton of demand charges. Win-win.

This is going to come up a lot in future, so we should be careful to seperate two things:

1. Money spent on having generators on standby when they are used only 5% of the time

2. Money spent on fossil fuels for those plants when they actually run.

1 is money well spent, 2 is a problem to be solved but just carbon taxing it is probably enough to consider it solved.

Luckily, the high cost of 1, makes it easier to justify adding batteries or a thousand other things that help to reduce 2. But you know people are going to point to the same peaker plants we've had for years and say "look how expensive renewables are!".

Also, we've already built the gas plants. Which is probably a big chunk of the cost (they'll just be spreading that cost across the hours that they intend to run during their working life).

Developed countries have a couple of weeks worth of natural gas in strategic reserves. I don't see why we couldn't do that with gas made from CO2, water, and electricity. Or ammonia. Or maybe hydrogen.

Based on some back-of-the envelope calculations, a single Powerwall would provide about 5-10 hours of "house run" for the average house, and residential use is about 1/3 of all electrical use.

So a Powerwall on every house gets us a third of the way there. That seems entirely within possibility.

The "Victorian Big Battery" cost about 100 million to build and provides 450MWh, so at those prices the 120GWh they calculate for 5 hours would cost about 25 billion.

Not outrageous in the scale of nation building projects.

They won't use batteries for grid scale storage, that's a bit too expensive. They'll probably use hydro, some sort of heated mass or compressed air. Or gas, it's much better than coal at least.

I think in this model it is actually batteries. They pair pretty economically with solar to smooth output on the daily cycle.

They also include existing Hydro in their model, plus 'other' which is what mostly covers the storage on longer scales.

But there's easy answers for the rest too. It's just not modelled in this very simple model because most of the heavy lifting is always going to be solar and wind and that's probably the point they're trying to make.

I think Australia will use mostly batteries, but we are building more hydro. I think hydro is miscast as being green though, when in reality it causes enormous ecological destruction.

Hydro is also intermittent - it requires the right amount of rainfall, currently hydro in Australia is not able to reach full generation because it will flood downstream as it's been a wet year.

Hydro is not that cheap (depending on the metric you need) - see an example[1] comparing Australian hydro to the aforementioned Victorian Big Battery.

[1] https://www.solarquotes.com.au/blog/snowy-2-vs-battery-stora...

For 25 billion, you could have two Olkiluoto-like nuclear blocks that actually produce energy and not just store it. For several decades.

(Olkiluoto is considered notoriously over budget.)

That wouldn't solve the problem though, because the NEM is a very long thin grid.

Prices have roughly halved since then, and will likely halve again when a winner of the not-lithium battery competition emerges (which is looking very much like sodium ion as it's entering mass production). Everyone is expecting another halving after that but before 2035 even with no single clear winning outcome. After that prices become dominated by steps that don't typically get cheaper, so prices much below $30 (or a bit over 10% of the Big Battery) are unclear.

It also doesn't sound like actually enough. They never have outages longer than 5 hours?

And even with fully renewable sources they would never have periods longer than 5 hours without much generation?

Hence the overgeneration.

Looks like they used Gas peaker plants (labeled "Other") to make up the difference during the times that was necessary.

Not if we figure out vehicle-to-grid.

Reply on news.ycombinator.com