Hacker News Reader

> This outcome is in a particularly distorted badly operated market, one where for no very good reasons regulatory power is not being applied . . .

You described most energy markets on the planet.

It’s particularly bad in Australia and South Australia where this battery is was the worst.

Australia despite having abundant reserves of coal, gas and uranium as well as excellent solar and wind locations has some of their most expensive electricity prices in the world.

This is true and it blows my mind. It also makes me sad. We have some really smart and educated people in this country. What's the problem here?

Mostly the Coalition. Seriously.

It’s clear from the AEMO market makeup etc. that almost all the cost increase was attributable to the cost of gas (due to Santos having contracts to export more gas than they had supply for to Japan etc., while the international gas price has fallen. So they are sucking up gas on the domestic market pushing up domestic prices and their liquefication operation is unprofitable anyway). But the Coalition instead chose to use the cost of electricity as a political tool in a bizarre anti-renewables push.

Economists have suggested that a large enough gas reservation policy (like WA has) would rectify that. The LNP could legislate it in a few weeks at worst. This would also benefit industrial gas users that have been struggling with big increases.

A more minor part (like what this battery is helping to stop) is the privatised generators gaming the market by constraining supply artificially to push up prices at certain high-demand times. Some of this also naturally happens with some of the older coal stations tripping out and falling off for a few hours (I think Liddel is only achieving something like 46% capacity on average or something?).

>almost all the cost increase was attributable to the cost of gas

Unfortunately natgas prices determines price because it's often the price at the margin.

If policy makers care about efficiency, fairness, and reducing pollution, they would better-allow, or outright demand, price signals flow through to the marketplace.

> Mostly the Coalition. Seriously.

Pull the other one, the risk is something like what happened in Germany. The cheap, reliable option in Australia is coal. So an investor has 2 options:

1) Invest in coal. At some point a Labor-Green government gets in and copies the German model - all your capital now makes a loss (ie, the plant you just built).

2) Invest in something else. This isn't as politically risky, but if the government decides that high power prices have to go and pours money into something (eg, new coal plants) then you just got set up as a marginal producer.

I mean, I haven't crunched the numbers, but unless wind/solar are legitimately the most cost effective options (which historically they most certainly have not been) then you would have to be a real risk taker to get involved.

Blaming that situation on the Coalition is too much; the political risks go way beyond them. The issue is that the Coalition wants cheap, the greens want clean, and Labor isn't committed either way as far as I know. And the regulator is a long way from perfect.

Fortunately, people have crunched the numbers:

https://reneweconomy.com.au/plunging-costs-make-solar-wind-a...

The Coalition doesn't want cheap - they've spun coal so hard it's no longer politically tenable to unspin it.

isn't their only major energy investment a bigger hydro dam?

Sorry, that doesn't make any sense. We have no pressing need for additional baseload capacity at the moment. The problem is entirely in high-demand times, and that is entirely sourced from dispatchable generation. This is mostly gas and hydro in Australia, and the price of gas drives the price, given the way the AEMO bidding market works.

Renewable energy only takes pressure off this demand (i.e. solar is fairly well correlated with hot weather, so actually helps a lot in reducing the peak on hot days (presumably from air-conditioning etc.) by hundreds of megawatts).

Australia's number 1 problem is that the grid itself is too expensive. That isn't the Coalition, because it has been happening through the late Howard, Rudd-Gillard-Rudd, Abbott and Turnbull governments. Typical slow government problem, not any particular party. Obviously you aren't talking about that, so I didn't mention it.

Australia's number 2 problem is all our power plants are aging and shutting down. I addressed that; because you seemed to be talking to it (eg, Liddel is only achieving something like 46% capacity). That is a looming issue. The problem here is also not the Coalition, as I mentioned.

Now SA appears to have a tertiary problem which is they have lots of wind energy and not enough storage to balance out the grid. The battery is addressing that and seems to be going well. This is also nothing to do with the Coalition, because the SA state government controls SA and the majority bounces between the two majors. The Weatherill government seems to have had the battery installed extremely quickly despite whatever you think the Coalition did to prevent it.

You aren't assigning blame correctly here. The big picture problem isn't that gas prices are high, and power problems aren't happening because of "mostly the Coalition".

And the grid is too expensive because the regulator allowed an effectively risk-free rate of return on investment that was too good to pass up.

In telecoms wholesale pricing in Australia, there would frequently be adjustments to the claimed capital costs to remove this gold plating incentive, but the energy side seemed to be much more accepting of whatever the distributors claimed.

Coal is not cheap though. Or to be more accurate: coal has a bad ROI, which is what drives investment decisions. At this point in time, solar (using Chinese panels) has an ROI that outstrips coal power by a lot, and has less risk (panels start producing a lot faster and can be expanded online - a coal plant sits there costing money for 5 years till it comes online).

The problem for both the US and Australia is that private investors have no appetite for coal, because the same money can just continuously install solar and get a better return - with no real problems with environmental lobbies, access to fuel, or anything. We're probably a few years away from grid expansions being built on the basis that they'll make money from people installing solar farms in the new areas.

"Ah! But China is subsidizing it's solar industry!" you say - well "who cares?" says I, the investor who China is effectively paying to build solar plants, and who gets to take a bunch of good PR from it.

So why isn't that happening? It seems odd to me that in an ostensibly free market economy, a good investment opportunity would be ignored.

Does the government just say "no" to letting wind and solar producers add capacity?

https://youtu.be/vSz1fItPWIw

I watched this Melbourne University video which discussed this issue in detail. It went into detail on the economics of the Australian electricity market.

I mean, I haven't crunched the numbers, but unless wind/solar are legitimately the most cost effective options (which historically they most certainly have not been) then you would have to be a real risk taker to get involved.

Really?

The 2017 Lazard report (which I think is generally pretty well regarded) has them at:

Coal, min: 60, max: 143

Wind, min: 30, max: 60

https://en.wikipedia.org/wiki/Cost_of_electricity_by_source#...

There are plenty of other sources on that page. You'll quickly find that most show wind power being cheaper than coal since around 2012 (even without a carbon price).

Looking at solar and wind historically doesn’t make any sense. They’re emerging technologies with rapidly decreasing cost. You have to look at their costs now, not from before.

When the resources are internationally tradeable, the size of the local reserves doesn't really matter. The price faced will still be the market price.

Generation is far from the only factor in power prices, anyway, and Australia is not nearly so blessed when it comes to the economics of power transmission.

Not to the point where it costs $14,000 per MW. At that pricepoint hiring humans to hand crank generators is not the laughably ridiculous proposition it should be.

For reference, a megawatt (MW) is 13000 humans cranking generators. You might get by with fewer since the problem being addressed is short bursts of capacity mismatch.

A human might average 100W per hour over a workday at say 10$ per hour to keep the numbers clean (1$ = 10Wh). 14,000$ per 1,000,000Wh is 72wh per dollar so far cheaper. Except this is instantaneous power so they would need 1,000,000 people to sit around doing nothing most of the day and then occasionally have have some or all of them work, because again that battery caps at 100MW.

Really, people in AU don't pay 14$ per kWh making 14$ per MWh meaningless. They are really paying to keep generating capacity online not for actual power. A backup generator that's never used costs infinite dollars per W, that does not make it useless.

PS: Well under 1 million beause people can generate more power in short periods.

> to keep the numbers clean

With Australia's minimum wage, it's likely to be at least double this.

Fair enough, I was thinking people might average more power and you might use someone outside Australia. On the other hand you need to pay for equipment so ~3W/$ may be more accurate but being generous does not change things much.

Anyway ~20$ /h * 24 h a day * 1 million people / -123 MW (Tesla capacity) = ~3,900,000 $ / MWh which is closer than I thought it would be.

> Really, people in AU don't pay 14$ per kWh making 14$ per MWh meaningless.

Agreed that consumers don't pay wholesale pricing - but disagreed it's meaningless. Consumers do (of course) end up paying an amount that includes the amortised cost of wholesale power over the course of their contract.

Additionally, there are a range of incentives that result from these kinds of significant, if ephemeral, variations in wholesale pricing. Some of these incentives are bad for consumer, planet, etc.

> They are really paying to keep generating capacity online not for actual power. A backup generator that's never used costs infinite dollars per W, that does not make it useless.

A backup generator that's never used actually is useless.

> A backup generator that's never used actually is useless.

Having a backup generator for your data-center means you can have less batteries. Even if you never directly use it you are benefiting by reducing overall capital costs while still mitigating the same risks.

Put another way owning a generator can provide a revenue stream because you can sell that risk reduction to someone else. That's not a theoretical benefit that's cold hard cash.

A human might average 100W per hour...

This seems like the wrong unit to use.

[EDIT:] I mean that power is already a "per time period" measure; there's no reason to say "joules per second per hour".

A rule of thumb is a large hamburger with all the fixings is about one kWh.

His point is that a Watt is a unit of power being equal to 1 joule per second. As an aside, this corresponds to an electric circuit with a potential difference of 1 volt and a current of 1 Ampere.

Saying a human produces 100W/hr is wrong. What I assume he meant is this:

A human can produce 100W for 1 hour, on average. This is reasonable, since humans produce about 1/8 Horsepower, on average; and 1 HP = 746W.

He meant a human can produce 100W. He specifically doesn't mean "for one hour"; he says the human works a full workday, producing 800 watt-hours of energy.

Right, but 100W/hr is very different from 800wHr.

Mostly because "100W/hr" doesn't really mean anything sensible? If you wanted to write "100W" like "100W·hr/hr", that would be fine.

I am breaking the day down hour periods, and I think a human might average 100W over each of them.

Being slightly faster in a factory is often not very useful, but being slightly slower will quickly cause problems.

Why not just use 100W, why specify over 1 hour periods? you would likely have a pool of bikers anyways so the averaging is builtin.

3x8 hour shifts adds up to 24 hours. This means our 6/7 of your shift ~everyone has been working for 6/7 hours. You might be able to stagger start times every 4 hours or even every 10 minutes over the day to smooth things out, but doing such an odd schedule would increase costs.

Is this per mWh? That seems like a rediculous amount if true.

It does seem a bit high because that is a peak statistic, not an average. The energy regulator keeps some statistics on the subject at [1].

Regardless, there is a good point to be made that South Australia was a strong competitor in the hypothetical 2017 Worlds Most Expensive Electricity Competition. It is possible that the batteries success is due to some quirk of their particularly poorly run grid. It is also possible that wind-battery combinations are now a competitive option. The situation continues to unfold.

[1] https://www.aer.gov.au/wholesale-markets/wholesale-statistic...

Great and important points, thank you.

Regarding pumped storage, I recently discovered some studies pointing out the potentially massive greenhouse impact of certain reservoirs that we must keep in mind.

http://journals.plos.org/plosone/article?id=10.1371/journal....

Mandating storage or fcas capability for new developments is not the solution, as it would make such developments more expensive and less appetising to developers.

Strengthening the market for those services and working with AEMO to drive the networks or third parties to provide these services is the better option in my opinion.

The recent rule changes for 5 minute settlements in the wholesale spot markets will also alter the financial incentives for very fast start generation and drive investment in battery storage, perhaps even in the residential market.

I don't agree that mandated storage or fcas would reduce investment: FCAS capability is mandated in other economies, which are powering ahead with offering it. There are many other sources of cost and risk exposure which outweigh this one.

Otherwise, I agree with what you say.

Power investment is not 'oh well, I'll put it into fish oil instead' money, these are firms who specialize in this niche and who are looking to invest in this niche. The lack of government commitment to PV and wind is a bigger threat to them than a requirement to spend more capital including storage and frequency supply logic.

Ancillary services like frequency stabilization are the highest ROI application of batteries in the grid and nothing to sneeze at. Without batteries, transmission and especially distribution need to be sized to the highest momentary spikes in demand, and given the long lead time and high costs to upgrade these facilities, batteries are a great solution. They can respond instantly and solve the problem at a very low cost.

As for storing power overnight, batteries are on a similar manufacturing learning curve* as photovoltaic, albeit a few years behind. Most people thought PV could never be cost effective, until suddenly it was the cheapest. Batteries too will surprise the doubters. It's just a matter of time before batteries are more cost-effective than pumped hydro.

*Moore's law-style cost reductions were first observed in airplane manufacturing in 1936. Cost of manufacturing decreases with the volume produced:

http://www.uvm.edu/~pdodds/research/papers/others/1936/wrigh...

The particularly distorted quality of the Australian energy marketplace, distinct from all the other worldwide markets which I absolutely acknowledge are, in their own terms, equally distorted, but the quality here, is the drop in the FCAS service bid by 90%, for what we have to call better frequency stability: The battery responds faster, and there is significantly less visible perturbation to frequency stability, and it does it 9x cheaper than the prior market driven process.

How good is that!

If I understood correctly, they are using batteries in same manner as pumped storage:

https://www.energy.gov/eere/water/pumped-storage-hydropower

We have a few of those for decades, it is actually quite simple, when there is surplus of el. energy it is used to pump water on a hill and acumulate it there. When you need it, the water is than used to generate el. energy. This is the traditional way of flattening the spikes in el. grid.

Now I wonder why the batteries were choosen instead of water, are they cheaper? Does it acumulate more energy on less space than water, it is due to efficiency (there is some loss when you pump water up) or for Austrialia, the water based storage is just not an option (not enough water / hills)?

(Warning: I have zero expertise on this)

Apparently there's an argument about that, since there's a plan to upgrade the Snowy Hydro pumped storage system, but it'll cost $2B and it couldn't be done in 100 days like the Tesla installation. The other advantage seems to be that the battery can respond very quickly (0.14s), so it's still useful to avoid instability in the network.

They're not using it in the same manner as pumped storage. They're using it for millisecond level FCAS response. Turbines can't do that very well.

Life is hugely politicized

We think we’re boxing it off from that but if you look closer, through the abstract nonsense people peddle, you’ll see it comes down to a political yes/no process

You’re welcome to pretend otherwise. I’d rather just dispense with the notion life is anything but dialog back and forth, jockeying for acceptance and access into some clique or the “right” to build useful infra

You comment seems to suggest there is some rational decision making emerging from an open back-n-forth of facts and ideas.

I wish that were the case, but the simplest explanation of political reality both in AUS and USA is that our leaders owe favors to a few very rich people, in fact work for them.

Can you explain in this model, how any useful bipartisan investment happened, in the past wartime expediency aside?

I think its true this has always been a dimension of things, but the scale of this dimension set against other dimensions has substantially changed in the last two decades. Before this, there were some things that both (notional) sides were prepared to express some unanimity on.

For example: "soviet power plus electricity == socialism" did not actually stop anyone of a more rightist persuasion from investing in electic generation.. did it?

Bipartisan consensus is also a political process.

Yes. One which is increasingly absent from western electoral democratic process. Oppositional politics is currently the preferred model even though it wedges capital investment in infrastructure. Are you bringing anything to this conversation or just trying to win an argument here?

krl02 made the point that all these things are political processes. You brought up the example of past bipartisan investments as a counterargument; my understanding of your point was "we've had bipartisan consensus on investments before, therefore there wasn't a political process going on around that". I was pointing out that this also is a political process, and that krl02's model of economic decision-making as a political process still stands.

Were you trying to make some other point instead?

All these things are political processes, but one is a trend which has swept the world. Oppositional politics has swept the world, and is impeding capital investment in public works, because it reduces to a high-tax/low-tax argument and favours low-tax reductionism, over state capital works.

I was trying to make the point that if (like me) you want capital works, and national utility project investment, you don't want oppositional politics, you want the other model.

I am not in denial that there is another, or other models.

We had more overt bipartisan politics over a number of important things like housing, defence policy, capital works. We have gone to a more oppositional model even though this means less public utility outcomes happen.

I’m confused. Is it 90% savings vs regular power (awesome) or 90% vs “emergency” power (ok)?

It's 90% cheaper bids into a frequency stability service. It corrects for drift from 50hz. It doesn't supply power in that sense, it supplies a slightly out of phase AC to refocus the frequency back to where it should be. It does it inside milliseconds of a drift and prevents overreaction by other generators with synchronous power, and can do it more often for less "cost" (synchronous flywheel systems lose most of their power reserve reacting, and have to be recharged over a longer and more costly period. A battery is really ideal for this frequency adjustment service)

So tesla power station acts like a nationwide super cap ?

In what sense doesn't it supply power? If demand for power exceeds supply, then the frequency starts to go down. "Maintaining frequency" and "supplying the demanded power" are two different expressions of the same problem.

If the frequency goes down, the battery supplies power to the grid as you say. If the frequency goes up, the battery absorbs excess power from the grid. On net, it doesn't supply anything.