Hacker News Reader

Comment by beat | Hacker News Reader

I hope this is a solid study and not just wishful thinking. I've been looking at pricing, and right now, fracked gas plants are running around $40/mwh, with wind at under $50/mwh and PV solar (no battery storage) at around $60/mwh. Coal and traditional nuclear are running around $100/mwh (these are all rounded numbers from memory, don't get unhelpfully pedantic!).

These numbers are obliterating both coal and nuclear. Coal plants are shutting down as fast as replacements can be built, and nuclear plants are shutting down a decade or more before official end-of-life, because they're losing major customers to cheaper alternatives that are available right now.

The cost of PV solar has dropped over 80% in the past decade, and wind like 50%, due to engineering improvements in operation and economies of scale in manufacturing. But it's a reasonable question how much farther their prices will drop. It seems unlikely to me that we'll see another decade of such radical price drops.

Gas, on the other hand, is vulnerable to markets and availability. How long will it stay so cheap? And what happens if a politically progressive government gets serious about carbon taxing? On the other hand, it's a lot cheaper up-front to build a gas plant, and a lot less painful if it closes early.

View on HN

Numbers for wind and solar are now about $40/MWh unsubsidized according to current LCOE analysis: https://www.lazard.com/media/450784/lazards-levelized-cost-o... ...as you say, this is not the whole story as you still need storage (etc).

But the floor for (unsubsidized) solar may be a lot lower than that. About $14/MWh: https://www.greentechmedia.com/articles/read/the-floor-for-u...

And we're already seeing bids nearly that low: https://www.pv-tech.org/news/brazil-auction-sets-record-low-...

Wind is also super cheap.

I really hope so! I think we need to see wind/solar at half the cost of gas to really break gas as an energy source. And battery storage is quite expensive for now - hundreds of dollars per mwh. Of course, there are other interesting alternatives, too - thermal storage, gravity, compressed air, etc. But conversion losses will probably always limit those systems to 25-35% efficiency at best, plus much higher construction cost relative to solar panels.

> But conversion losses will probably always limit those systems to 25-35% efficiency at best

Maybe I'm misunderstanding, but the "alternative" energy storage systems you mentioned have about 80% or more round-trip energy efficiency: https://www.eesi.org/papers/view/energy-storage-2019#2

I quoted alternative because pumped-storage hydro accounts for 95% of grid energy storage in the world.

And it's not like the world lacks places to put new pumped storage: https://nationalmap.gov.au/renewables/#share=s-oDPMo1jDBBtwB...

25-35% efficiency? Pumped hydro is generally over 70% roundtrip.

Yes, we are seeing very low bids. $26-$30/MWh for solar doesn't even turn heads anymore.

And battery costs are coming down too, and I think that curve looks more linear, but its coming down.

One of the issues is that these costs are so capital intensive that the financing costs drive a lot of the price drop. With near zero or negative rates in so much of the world, I think its hard to tell what the real rates would be.

Capital costs are a legit concern. If something happens in international markets that causes interest rates to spike globally, it will set back the pace of conversion tremendously.

“Fortunately” we appear to be heading into a recession, so I doubt high interest rates are on the way any time soon.

Any way to leverage negative interest rates from developed countries into renewables? Very interested if someone knows how to facilitate such a bond issue.

> and nuclear plants are shutting down a decade or more before official end-of-life, because they're losing major customers to cheaper alternatives that are available right now

How? My understanding is that most of the cost of nuclear power comes from building the plant and decommissioning it. Running existing power plants for as long as they can seems like the most economical choice.

Indeed. See page 2 (or 3, counting from the pdf) of this analysis by Lazard. In particular, the tan diamond symbol by nuclear (with a footnote). https://www.lazard.com/media/450784/lazards-levelized-cost-o...

Nuclear that's already built is very cheap AND has a very high capacity factor (i.e. doesn't need storage--although its value can be extended with storage).

Nuclear also has high operating costs, just not fuel costs.

Solar can generally just be left alone for weeks or months at a time. Nuclear however needs security guards, regular maintenance, a highly trained workforce, large insurance premiums etc. On top of that it also needs some fuel.

What’s happening with Nuclear is these older power plants occasionally need a significant overhaul and it’s at that point where they are considered too expensive.

And a nuclear plant costs about the same to run whether it's running at 100%, or 20%. Fuel costs are negligible relative to capital and operational costs. So when lower-cost competition appears and eats into the demand for nuclear power, it actually increases the per-unit cost of nuclear power, which makes alternatives more compelling, which eats into demand, which turns into a cost death spiral.

Nuclear plants are great for base demand for these reasons (ie, the power that is always needed 24 hours round the clock, 365 days a year).

Alternative energies can provide some of that, meaning that plants need to shut down because outside of providing base demand, they're incredibly expensive to make power from.

You are totally correct. US, France, Japan and UK all decided to extend the filetime of their nuclear power plants.

https://www.reuters.com/article/us-usa-nuclearpower/u-s-to-e...

Wind is cheap because when the wind blows, everybody is trying to sell their wind produced energy at the same time and the country often overproduces. In Germany, we even see negative prices when the wind blows; The price is low not because it does not cost a lot, but because it is heavily subsidized (wind is 60% of subsidies in Germany, for only 18% of the production) and needs to sell fast otherwise the fines and costs for overloading the network are massive.

Negative prices? You'll pay me to take electricity off your hands that I can sell later?

That's the best argument for the battery storage market I've seen yet... (And really, it makes sense. Generate surplus when the sun is shining or wind is especially strong, run from storage when demand exceeds supply, build both until there's enough. As long as the costs are less than current baseline systems, which seems possible, then it wins.)

They won't pay you, unless you're literally running part of the power grid.

> Generate surplus when the sun is shining or wind is especially strong, run from storage when demand exceeds supply, build both until there's enough.

That's the idea, but the battery capacity required would be enormous:

https://www.oulu.fi/blogs/how-much-storage-does-the-energiew...

> As long as the costs are less than current baseline systems, which seems possible, then it wins.

Batteries don't seem likely for a cost-effective replacement, but alternatives may develop down the road.

It is unfortunate that link didn't pursue non worst case scenarios, like being able to import some power, or have some non renewable generation used just enough that it was economic to exist at normal energy prices, with the biggest payoff being the avoided cost of energy storage for October and November.

that link is kind of unrealistically dismissive.

The question is, can you really go fully renewable with battery storage? If power can be sourced elsewhere but it comes from nuclear or fossil fuels (as in the case of Germany and France), it would be a sham. If everyone went renewable, everyone would also have the same problem of seasonal variance.

If you want to go "mostly renewable" and do the buffering e.g. with natural gas as it is currently done, that's of course a far more plausible scenario.

Negative electricity prices happen in Australia fairly regularly. The last time was just a few days ago:

    https://www.abc.net.au/news/11489188

Interestingly is isn't generally caused by renewables. It's caused by those "base load" generators: coal and if Australia had any, nuclear. Base load actually means "can't vary their output quickly". If you can't vary your output and you ramp down your output too early, you missing out on selling power. Too late, and the price plummets and have to sell it at loss - or worse pay someone to absorb the energy released by your boiler cooling down.

But renewables have made this worse by unpredictably flooding the market with cheap power. They don't really care if the price drops to near zero as the wind and sun cost them nothing. But jeeze it must hurt to be selling electricity for less than it cost you to dig the coal out of the ground. I'm pretty sure that's one of the things driving the early retirement of coal and nuclear.

Essentially you don't get paid to take the energy of their hands; you get paid to use energy so the grid doesn't overload. Crashing the grid is expensive as it means you won't be able to sell power (at 52Hz or so the grid operators start disconnecting larger power sources like nuclear power, above and the grid may have to be collapsed). So it's in the best interests of the power generator sites to sell energy at a loss if there is too much in the grid to push consumers (ie, big production plants) to burn more energy.

In the other direction, if there is not enough, the grid operators will disconnect power consumers, so in that case it's in the best interest of generators to prevent people from consuming power, thus raising the price dramatically.

Negative prices are not uncommon. Happens nearly every spring in the Pacific Northwest when the mountain snowpack melts, flooding the rivers and surging the output of many hydroelectric dams. The negative prices are factored into the annual revenue, and consumer prices are sort-of averaged over the year which is why you never see it on your electric bill.

I live in Portland. Last time I toured Bonneville, no turbines were operating on the East side. The ranger told us there was sufficient solar and wind at the time and they shutdown to protect aquatic life.

Assuming that's true, I assume they'd also shutdown to protect thier pocket book.

In a word, price. As a consumer, do you want a bill for 2 cents a kWh? Or one for 15 cents a kWh?

This example explainsa lot of what's happening: [1].

By way of a brief explanation, wind is expanding so fast in a lot of places that the combination of wind+LNG is just too cheap to compete with. Iowa is particularly bad, because wind is not "unreliable" in Iowa. There's a lot of wind and it's almost always blowing. So they aren't even using those LNG plants that much. Which only lowers the price to consumers even further.

In some states the government has stepped in to protect the nuclear industry. Sometimes with things like carbon credits, that are given for generating power without generating carbon. Then the nuclear plants sell those credits on the free market to raise funds. But in some places government has had to be a good deal more draconian. In Georgia, they've basically made it illegal for anyone in Vogtie's service area to switch away. That's the only way they could get the math to work. (And it still doesn't work. The federal government is still pumping billions in cash into that project.)

[1] - https://www.washingtonexaminer.com/policy/energy/iowas-only-...

An interesting factoid I ran across when researching this stuff a while back is that no one has ever built a new reactor in a deregulated electricity market. Period. The time frame of nuclear reactors is very long - like 50 years. So they're very sensitive to capital costs, and capital is sensitive to risk.

Now, if you were doing the risk analysis on financing for a multibillion-dollar nuclear project, and you just saw the cost of a major competitor (solar) drop 80% in a decade... well, how would that factor into your risk analysis? How much would your interest rates increase?

The only safe way to finance a reactor is government-based guarantees for price regulation - the nation will consume X megawatts for Y dollars, no matter what the market thinks the price should be. And do it in such a way that some government two generations down the road can't just throw it all out and screw you.

The complexity of reactors needs to go down and they need to be smaller. Big reactors can only be cooled if they are near the ocean, otherwise they have to be shutdown when the flow rate of the river is too low, this usually happens during heatwaves. It's easier to achieve economies of scale with 16 identical small scale reactors instead of designing a custom reactor for each plant and then building only 2.

I don't see this happening. Relying on new reactor designs to save nuclear power isn't going to solve the problem in time. Right now governments are busy with massive power plant designs that suffer from multi billion cost overruns. After that experience they will probably stop building more power plants and funding for new reactor designs will be reduced.

Yeah. A lot of my attitude about pricing turns into pro-nuclear advocates treating me as Treehugging Enemy of Progress, and one of the things they point to is the always-in-the-future dream of small, simple thorium molten salt reactors. But the best I've seen on that is a company that promises 30 cent megawatts in 2030, but doesn't really have a working prototype yet. And it would take, what, two decades past that to really get that going at scale? We can't wait 30 years to do something about greenhouse gasses just because we really really really want to do it with nuclear.

A nuclear plant in Arizona uses treated sewage for cooling. In theory, any nuclear plant near any sufficiently large city would have access to enough water for cooling, unless the city was unusually water efficient.

Problem I see with nuclear, and gas and oil pipelines is the real risk of a complete default on it's bonds. Plant will either not get built or the plant may get shutdown[1] before paying off it's bonds. I think San Onofre was shutdown because it's brand new replacement steam generators showed excessive wear. And new nuclear plants seems to be costing more over time not less.

Vs solar or wind farms where default is much more likely to be 'soft'. Maybe the plant defaults on it bonds, but it still produces power. Which is preferable to both defaulting and shutting down.

"The time frame of nuclear reactors is very long - like 50 years. So they're very sensitive to capital costs, and capital is sensitive to risk"

50 years doesn't sound like all that long a timeframe when people are talking about 100 year bonds, and the idea that capital is scarce and expensive doesn't quite seem to apply when $15 trillion in global debt pays negative yields.

And that's one of the major arguments for mass produced small modular reactors.

No one-off construction, flexible siting due to no water cooling, underground for security, and meltdown proof. Just walkaway safe electricity at five cents a KWH or less. :-)

terrapower.com thorcon.com x-energy.com

And maybe someday, they'll be an actual alternative that actual power companies can by on an actual market.

But they aren't, not today, and they won't be for at least a decade. I don't want to wait, and even if I did, the market definitely will not.

> As a consumer, do you want a bill for 2 cents a kWh? Or one for 15 cents a kWh?

If that energy produces zero carbon then I'll take the later. But we're also only talking a difference of 7c for PV and 10c for nuclear and competitive prices for coal and natural gas. I'll gladly pay more for the nuclear than pay less for coal and natural gas.

I fully agree. Unfortunately most wouldn’t.

They're shutting down because it's cheaper to shutter them (even including the loss of interest on money set aside for decommissioning) than to continue operating them.

Whatever seems "obvious" to us, the hardnosed managers responsible for the actual financial decisions are pulling the plug on dozens of reactors. If it seems irrational to us, it's because there are other cost factors we don't understand. That's why I like to look at the end, rather than make assumptions on what I think I know about the means.

> If it seems irrational to us, it's because there are other cost factors we don't understand.

Since they are shielded from almost all liability by law (something like $150 million liability cap + $12.5 billion industry fund; plausible disasters could reach $1 trillion+ depending on winds), it seems like there would probably be even more shutdowns if they had to pay for real insurance.

Don't get fooled by simple price comparison. It's comparing apples and oranges. Gas, nuclear, coal, reversible dams, geothermy... are all controllable energy sources: they produce when we need it. Wind and solar are intermittent.

Equip fully the world with solar and wind and the trains will have to leave their stations when the wind blows and not at any other moment.

What's happening is that countries migrating towards wind turbines like Germany also adds coal plant to the same capacity in order to compensate for when the wind doesn't blow. For each GW of wind one can find a GW of coal or gas being installed.

Then, if we are talking about adding batteries into the mix to make it work, solar and wind become far from carbon neutral, and endangers us way more than nuclear.

I suspect that if variable electricity pricing for consumers, a great deal of the need for "batteries" will evaporate. Heavy consumer electric uses (charging the car, HVAC, hot water) can easily be time shifted. It isn't time shifted today because consumer electricity rates are constant 24/7.

Electricity price is already adapted hour per hour on the European Market. When the wind blows in Germany, price sometimes even go negative: it costs operationally less to energy producers to pay people to use the energy than stopping equipment. No change has been noted on the electricity usage.

Electricity bill is on average $100 per household per month. People won't shift their usage because electricity is too cheap anyway. 1.6% of your revenue === 1.6% of your saving effort.

Note that the periodicity of wind blowing is days and weeks, so it's too long and too intermittent to align daily routines on it. Forget about charging your car or hot water aligned on it, the only solution we have scaled to use this wind energy is reversible dams / potential energy, and of that we have no more unexploited geography to support wind turbines. Feb 2019 we produced in Germany 10,000GWh with wind versus 3,000GWh in July 2019. And between 2 consecutive weeks, it can be a factor of x10 of difference. http://energynumbers.info/capacity-factors-at-danish-offshor...

> to align daily routines on it

That's not necessary. The utility can make the current price available via an internet API. The water heater can be connected to the internet and can read that data. Heat the water to 130 or so when electricity is cheap, heat it only to 100 or so when it is expensive. From personal experience, the water remains hot enough for a comfortable shower 3 days after the power goes out. This is more than a workable system. The hot water itself becomes the "battery".

And that's only the hot water heater. You can do it with the HVAC system (the thermal mass of the house becomes the "battery"), and with charging your car.

Right there it's a huge chunk of one's electricity consumption, all without needing grid storage batteries.

While there is already variable pricing (Time-of-use) in some jurisdictions, I don't think that's what you're thinking of. "Transactive Energy" is more towards the concept of running the distribution system as a market akin to the transmission system. The company I work for has deployed a pilot program of this concept in New York. https://www.opusonesolutions.com/news/national-grid-launches...

The price for electricity already varies hour by hour. Maybe not as much as solar would, but most people (including me) are not paying a pegged price for electricity 24/7.

Not in the Pacific Northwest (Puget Power).

Also, to make this work best, appliances need to be connected to the internet and the power companies need to have an internet ABI where the current power price can be extracted by those appliances.

At last, an actual use for internet connected appliances!

Pacific Power (Oregon) also has fixed rates.

I don't think forcing people to use electricity only when it's available is a practical solution to the availability problem with wind and solar.

Not forcing... just having market forces encourage good behavior, instead of hiding the underlying cost pressures from them.

Look at it this way. With enough variation over the day, you create a market for storage. Let's say stored electricity costs twice as much as direct, but the price of direct varies by 3x over the course of the day. You've now created an arbitrage market for buying cheap electricity during the day, and selling it at night.

Of course, the more storage gets built, the more the load (and thus cost) increases during the high-availability times, and the lower the selling price during demand spikes. And thus the market attains equilibrium.

The law of supply and demand is a marvelous thing sometimes.

I understand your point, but GP is talking about avoiding the need for batteries by pricing people out of the market instead. If you're not using batteries, supply during dark or windless hours is inelastic, because there's nothing being generated - in other words you're not solving the problem and you can't really charge anyone anyway because there's nothing to deliver.

> GP is talking about avoiding the need for batteries by pricing people out of the market instead.

Nope, I talked about incentivizing consumers to shift their demand based on a volatile price. I pointed out that much of this can be accomplished without need of batteries.

For example, have the hot water tank heat up water higher than normal when electricity is cheaper, and lower when electricity is more expensive. Hot water tanks lose heat very slowly, this could result in never needing to heat water during the expensive times.

It's a heluva lot cheaper than having a battery do it.

The water in the tank is functioning as a battery in that case.

Yes, of course.

"Windless" is greatly overrated, for two reasons. First, utility-scale wind is built hundreds of feet in the air, not on the ground. Wind is much steadier there. And it's built in carefully selected locations for steady behavior.

Second, it's not a single location. Wind may be lower in one location, but strong in another location 50 miles away. The idea that wind goes completely dead all at once across statewide areas is not how wind actually works.

"The idea that wind goes completely dead all at once across statewide areas is not how wind actually works."

The idea that wind currents are uncorrelated across geographical areas, even on the scale of states, is not how wind actually works.

> HVAC

Yes, the demand for HVAC is very high when the winds are low.

> Equip fully the world with solar and wind and the trains will have to leave their stations when the wind blows and not at any other moment.

Wind just doesn't stop blowing across the world all at once. When one place is in the doldrums, another is likely windy. With a sufficiently connected grid, a large deployment of wind turbines can smooth out production in ways smaller, regionally focused deployments cannot. I posit that a global distribution of wind power is _more_ available than a regional one, not less, as your comment seems to imply.

Gas and hydro are highly controllable, but nuclear, coal and geothermal aren't really - just in the opposite way to solar and wind.

Those sources are slow to ramp up and down - that's why the time-of-use power prices I face right now are at their lowest overnight (so I currently time shift as much load as I can to those times).

> Wind and solar are intermittent.

True, but the story is actually worse than they paint it. Electricity from renewables + storage is now cheaper than electricity from a new coal or nuclear plant. As those old plants close, they won't be replaced. https://www.abc.net.au/news/11495558

In the tussle between renewables vs fossil fuels it's all over bar the shouting, which is to say all we will have to put up with the coal and gas industry screaming how unfair it all is until the last of them die off.

But it's is just as well renewables do cost less because if someone had to plan on losing money to save the planet it would be a gonner.

The $/mwh for renewables is only part of the story. The intermittency costs they impose on the system as a whole is higher than the direct cost of producing the energy, and the only way to solve this is through improving battery technology by orders of magnitude. Additionally, the intermittency costs per mwh will increase as the share of renewable energy increases.

In general, renewable energy investments right now are completely irrational in terms of the environmental impact, as they cost up to 10x more than the total social cost of the carbon emissions they are displacing. (And also much more than _current_ carbon sequestration tech!) [0]

[0] https://bfi.uchicago.edu/wp-content/uploads/BFIEPIC_WP_20196...

Do those prices for solar and wind factor in storage?

Depends. But wind doesn't really need storage... a broad-based wind grid delivers a pretty consistent base load, as we've learned from early adopters like Iowa and Texas that have over a decade of wind data at scale.

edit: There's enough variation in wind farm output to maybe make some storage financially beneficial - production costs could actually go negative when there's enough wind.

As a Texas resident who lives in the ERCOT region, I am going to have to vehemently disagree regarding the "consistent base load" that should allegedly be supplied by our wind farms. We had a day no fewer than 4 weeks ago that broke every record in terms of energy cost - for over 90 minutes our electricity cost 9 entire American dollars per KWh. We were <500MW of capacity away from rolling blackouts. The cause is the fact that we had bet a certain % of the base load would have been supplied by all the new wind farms, but in reality we had to fire up every peaker plant in the state, and even request assistance from other grids in order to meet demand. The wind simply wasn't blowing. If more of our mix had been natural gas, even just 5-8%, that event would have been completely mitigated.

Adding just 10%, to Nat Gas plants, would be 5,000 MW on the Texas grid. Given that wind farms were operating at 1/5th of capacity during this Texas grid events, you would need something like double the current installed wind farm base to be able to deliver 5,000 MW at that 1/5th capacity utilization. It has taken decades to get to the current, rough, 24,000 MW nameplate capacity for Texas to get to where it is.

So help me with the costs here. Adding 25,000 MWh (5,000MW * 5 hours emergency supply) of lithium battery storage would cost... what? As compared to constructing and operating new natural gas plants? I can't get clear numbers that I trust.

The largest battery farm on earth is 100MWh and it cost about 90 million dollars. For a 25GWh battery installation, you'd be looking at 22~23 billion dollars using basic linear extrapolation, and I still question if this is enough reserve capacity for our summer peak demand.

Wow. That's quite a lot. I've seen price estimates as low as $387/kWh, which seems to come from NREL [1]. At that price it comes to something like $9.6bn. That's huge, but maybe doable. Alternatively, couldn't Texas just build some HVDC lines to other parts of the US grid to get more stability, as China is doing?

ETA [1] https://pv-magazine-usa.com/2019/01/02/utility-scale-solar-p...

There's still a margin of soft costs that can be reduced (more in the residential than commercial/utility sectors, admittedly): https://www.nrel.gov/docs/fy19osti/72399.pdf

So cost of PV solar has dropped 80% over past decade - it it reasonable to expect another 40% drop over the coming decade?

A quick Google search suggests that Solar PV efficiency is 11-22%. Could it be 20-35% in 2029?

Reply on news.ycombinator.com