Hacker News Reader

This is also not true - solar + battery is a perfect use case for distributed deployment

This is simply not feasible. Do the math. To store 3 days worth of current US energy use, you need to build $50 trillion worth of batteries. If you assume that the battery installation lifetime is 20 years, and amortize it over that period, you’re looking at over 10% of US GDP spent just on building and installing the batteries. Seriously, do the math.

> To store 3 days worth of current US energy use,

There is absolutely no need, of any kind, to store 3 days of US energy use. It's a complete strawman.

So where do we get energy from in case we get a windless week in winter, if not fossil fuels or nuclear?

The wind is always blowing somewhere.

What are the transmission power losses over large stretches of land?

2/3rds of the electricity produced in the US is lost in transmission: https://www.eia.gov/energyexplained//us-energy-facts/

EDIT: this is actually wrong, I misinterpreted the diagram, see the comment below.

Go look at https://www.eia.gov/totalenergy/data/monthly/pdf/flow/electr...

I believe you mis-quoted the figures. You rolled up total conversion costs and bloated them into transmission losses.

Tell that to someone stuck in a lift because of a power cut 1000km from where it blows.

Have you actually found a confirmed reliable meteorology backed statement on a continent-wide, windless week? Compared to e.g. the current likelihood of catastrophic failure of parts of the net, Aurora effects on powerlines, Ice, Fire..

you're floating a strawman which has never been floated as a serious risk analysis: there isn't a single US distribution network, it is not a single integral thing, parts fail all the time, and more to the point, you haven't even demonstrated one weather system can produce no wind and no solar output for a week.

Winter does not == the entire united states is black for 24h/day. The Rockies exist, and weather on the west coast is not like weather in the mid-west, or the east coast, all the time, for sustained week-long periods.

Of course it will provide some output. The issue here is that if you want the diminished power output in winter to cover the energy use in winter, you need to either store the summer energy in some way, or overbuild by adding additional generation capacity, so that in summer you have some good multiple generating capacity of what you actually need, especially if you account for transmission losses, which are 2/3rds of current electricity generation. The result is that the spot price of energy in summer will be extremely low, and so the capital costs of building solar panels will have to be recovered by selling winter energy, precisely when solar panels generate least power. This will push the cost of electricity from renewables through the roof: right now they are quite competitive, because they can sell it for good price in summer.

You don't seem to understand either pumped-hydro, or economics. Using hyperbolic language like "through the roof" when we cannot magically wave a wand to make a 100% renewables world in one go, but are approaching it asymptotically means that the pricing function, the costs, and the outcomes are not 'through the roof' but a function of where we are in a migration.

I don't know how to predict the future reliably either. I just think the argument here is ridiculously reductive.

We aren't on mars. The current systems exist, and won't be turned off overnight, and we don't have a single global network in the economy under discussion. Pricing and distribution and storage have to be approached, not arrived at suddenly, And have upside consequences like increased employment, which means the spend is not 'hole in the ground' but actually has net beneficial effects on the economy.

To repeat what I said elsewhere you have mis-represented the 2/3 losses figure.

https://www.eia.gov/totalenergy/data/monthly/pdf/flow/electr...

(same source as you) is clear transmission losses are significantly different. You rolled up energy conversion and production losses from all sources which includes Coal and Gas and Oil and Nuclear.

> You don't seem to understand either pumped-hydro,

No, you don't seem to understand pumped-hydro. The largest pumped-hydro plant in the US, the Bath County Pumped Storage Station, can store enough energy to satisfy US energy use for 250 milliseconds. Yes, milliseconds. Suppose you want to have enough pumped-hydro to store US energy use for, say, 1 hour. There are 3600 seconds in an hour. You need ~15 000 pumped hydro plants. Where are you going to build them? Where are you going to get enough water from?

> or economics. Using hyperbolic language like "through the roof" when we cannot magically wave a wand to make a 100% renewables world in one go, but are approaching it asymptotically means that the pricing function, the costs, and the outcomes are not 'through the roof' but a function of where we are in a migration.

Of course it will be "a function" of something. The above sentence is devoid of content. The real question is, what will be the shape of the function? I argue that if we depend on renewables on majority of our energy, the cost per kWh produced will go through the roof. I argue why it is so. Do you have any argument? Show me your math, like I do in my every comment.

> And have upside consequences like increased employment, which means the spend is not 'hole in the ground' but actually has net beneficial effects on the economy.

This is just a broken window fallacy. Having lots of people spend lots of effort on something that before didn't require so many people will be greatly detrimental, not beneficial to economy.

> To repeat what I said elsewhere you have mis-represented the 2/3 losses figure.

Ah, thanks. I misinterpreted the figure.

Where are you getting 250 milliseconds and needing 15,000 from?

Wikipedia claims US yearly electricity consumption is 4,090 TWh, and Bath County stores 24 TWh, with a generation capacity of 3,003 MW. That comes out as enough storage for a couple of days worth of US electricity use. Obviously you'd need 3 more to cover generation capacity.

Ah, I rechecked, I actually lost 2 orders of magnitude. It's 25 seconds, not 250 milliseconds. However, you lost 3 orders of magnitude: Bath County is 24 GWh, not TWh.

My calculation was done under assumption of 101 quads of annual US energy use, but I'll redo this calculation with assumption more charitable to your argument.

US uses only 4090 TWh worth of electricity, which is about 13 quads, but it also uses about 70 quads of energy in other forms than electricity. See the breakdown at [1]. Let's assume that moving to all-electric world reduces the transportation energy use by half. In all electric world, the 25 quads we waste to generate electricity also vanish. Thus, the annual US energy needs is 101 - 25 - 14 = 62 quads, assuming we can increase efficiency by replacing majority of heat engines with electric engines. Putting this in Wolfram Alpha[2] gives you about 40 seconds.

So, now we need about 100 pumped hydro plants of the same size as largest one energy storage facility in existence, to cover 1 hour of US energy use (I'm ignoring of course the problem of not enough generating capacity to convert that storage to electricity fast enough as needed). Where are you going to build them? Where are you going to take water from?

[1] - https://www.eia.gov/energyexplained//us-energy-facts/ [2] - https://www.wolframalpha.com/input/?i=%2824GWh%2F62+quads%2F...

> So, now we need about 100 pumped hydro plants of the same size as largest one energy storage facility in existence

There are about 70 > 1MW pumped storage stations around the world already, with 40 more planned (https://en.wikipedia.org/wiki/List_of_pumped-storage_hydroel...), mostly in China. These are just glorified dams and as far as dams go they're tiny, so scaling up (and maybe down too) isn't an issue, there are an estimated 84,000 dams already in the United States. At ~$4 bilion dollars to build 100 of them would come in at the cost of about 1/3 the F-35 program and deliver much more national security.

So it's entirely feasible to have enough pumped hydro storage to cover this mythical scenario of a continent having no wind for an hour.

Your comment doesn’t answer any of my questions. Where are you going to build 100 dams with enough water to store gigawathours worth of electricity?

The problem with massive projects like that is that you simply cannot build them just about anywhere.

You can build them just about anywhere, all you need is a tiny amount of land area for the reservoirs enough of a vertical gradient. There are large mountain ranges running down both coasts with more than enough space to accommodate this. Looking at the surrounding geography of the bath county installation alone looks like it could easily be scaled to 10 times it's current size.

Oops, seems my split second skim of Wikipedia was a shortcut too far. :) I had to look up a quad (a short-scale quadrillion BTU). Haven't come across British Thermal Units in Britain for decades!

Anyway. 100 of that scale might be better or easier as 1,000 smaller capacity, but even then we can reduce further. Some percentage of the grid is nuclear (roughly 20% of US electric) which would not need underwriting nor would 100% of continental wind power ever be becalmed. I would further assume that any concerted national effort to electrify carbon neutrally would come with a matching efficiency drive - insulation, tax inefficient users, take first steps toward a national grid (that TIL the US apparently does not have), etc to allow you to use the power you have rather better.

Yet ignoring all that, the US must have thousands or tens of thousands of disused mines and quarries, probably many open ones with an accidental lake in their place that could be repurposed. If there's adequate volume, you don't need much head of water. There's no shortage of hills and mountains with rainfall or snow melt that should give plenty of potential spots, population and environment permitting. The one thing America does not lack is space - of all terrain types.

Plus with a lot of intra day seasonability.

Lets just review. You continue to ignore that there is no single weather system US continent wide, demanding an ability to simultaneously manage electricity supply from storage in one instance.

You insist I "do the math" but your premises has twice to your own admission had fundamental flaws in the source data and in the maths.

You really want me to "put up my numbers" ? I think you're trolling.

There is no compelling "all US from storage" problem. There is no "week long no-wind no-solar" problem. There is no immediate current TWh supply from renewables.

There are undoubtedly problems: They're just not the phantoms you are floating.

Go away please.

There is no US wide electrical network that can actually survive big failures, much less one that would allow reliably sending power from renewables across the continent.

These actually are insurmountable problems. Even just taking the time to deploy the required capacity wiggle allowing a generous 5% compound speed scaling gets us to required capacity in 50 years of constantly building renewables. For just the USA. And this ignores the expansion of the grid, attempts to actually secure best land for the deployments and so on.

A great solar system has peak 300 W/m^2 efficiency or thereabouts. You need 1400 W per capita on average, which means with generous 140 W/m^2 about 10m^2 of solar power per person plus support wiring and storage. Wind requires 20+m height and big rotors to match this average, making it require some 100 m^2 per. Neither completely infeasible, but the millions of square km to cover should show the magnitude of the task.

Hardcore California plans envision 60% renewable at 2030. This means at increased growth rate of still 2 billion tons. Essentially at the high endgrowth of renewables the 20 years would match last 10.

Conservative climate predictions at this level of CO2 equivalent give over 5 C increase!

It is too slow. Way too slow.

Solar panels? Even on the most heavily clouded days mine generate over 50% of nominal. And on that very rare windless, deeply overcast week, maybe we adapt and just don't run the tumble dryer or aircon for a week? Most house energy usage is pretty gratuitous and if you cut it back to lighting, refrigeration and entertainment you could get by on a few kWh a day.

If you absolutely must have 100% uptime then $1000 will get you a diesel generator which would keep the batteries topped off if you ever need it.

And if your argument is that someone living inside the arctic circle could find themselves becalmed for too long during the long night... then maybe back it up a notch and realise that 99% of humans don't live in those conditions and maybe the ones that do can still use some fuel oil.

> If you absolutely must have 100% uptime then $1000 will get you a diesel generator which would keep the batteries topped off if you ever need it

Ah, great, so every house that has food in a freezer or likes to stay warm in winter now has to buy and maintain a diesel generator. Welcome to the great new renewable world.

"Ha! It's not perfect in every conceivable future! This proves that it is worthless forever!"

There are other potential adaptations.

We've developed a food distribution network dependent on a cold chain -- not only the ability to refrigerate or freeze product, but to maintain constant refrigeration from the process point to retail to the home until final consumption. This is not an inevitability, and it need not remain standard practice into the future.

There are other methods of preserving food. They sacrifice convenience, freshness, and nutritional quality, but are tractable.

Canning, fermentation, drying, and other preservation methods can make many foods long-term stable.

Certain foods, especially fruit and vegetables, may need to be sourced fresh. This is where local production can be appropriate, including home or neighbourhood gardens.

If meat is still to be consumed, the traditional method of keeping it fresh was to ship it live. Canned or salt beef and port were common prior to widespread refrigeration.

How widely or enthusiastically such changes might be adopted becomes an interesting question. But as a long-term adaptation, this could well be a potential path.

One key mindset to looking at radical changes to environment and resource regimes is to recognise that changes in underlying costs, abundance, and/or suitability may be radical changes in products, practices, and behaviours.

Is there any evidence that this ever happens? Or at least happens more frequently than say, a 1979 style fossile fuel shortage?

> So where do we get energy from in case we get a windless week in winter, if not fossil fuels or nuclear?

It isn't actually unreasonable to just say fossil fuels here. If we had to burn natural gas for a week every two years it would still represent a 99% reduction in carbon emissions.

And on top of that, suppose we keep a week's supply of biofuels. They're not cost-competitive on a normal day, but you're not using them on a normal day.

A week's supply of biofuels, given our current energy use of 101 quadrillion BTU/year, and assuming 100% efficiency, is 2 quadrillion BTU. At the peak of our production, we produced around 0.2 quadrillion BTU worth of biodiesel. This means that a week's supply of biofuels is around 10 years of production.

Of course, in practice, if you overbuild renewable energy generation to a large enough multiple that will greatly increase the average cost of produced kWh, and if you keep most of the fossil fuel infrastructure on standby, then at the immense expense, and greatly increased energy prices, you could reduce CO2 emissions by switching to renewables. The real question is, why would you do so, if it would be much, much cheaper to just build enough nuclear plants to cover your needs. France managed to do it in the 70s., why can't we do it 50 years later?

> A week's supply of biofuels, given our current energy use of 101 quadrillion BTU/year, and assuming 100% efficiency, is 2 quadrillion BTU. At the peak of our production, we produced around 0.2 quadrillion BTU worth of biodiesel. This means that a week's supply of biofuels is around 10 years of production.

We produce around ten times as many gallons of ethanol as biodiesel. It has about half as many BTUs, which implies we would need around two years of production. Not really a problem if we only need that much once every two years.

It's also assuming we couldn't produce more than we do now, or do something more efficient like convert existing coal fired plants to burn biomass from algae.

> The real question is, why would you do so, if it would be much, much cheaper to just build enough nuclear plants to cover your needs. France managed to do it in the 70s., why can't we do it 50 years later?

Because we do whatever costs less. Nuclear plants are expensive. On the other hand, so is storage. It's not actually clear which one is going to cost less in the long term, so what we should probably be doing is building both at the same time.

Which has the additional advantage of replacing fossil fuels even faster, because the two technologies have very little overlap in terms of either labor expertise or materials use, and complement one another in that there is more electricity demand during the day when solar is generating but there is still non-trivial demand at night which can be handled by nuclear. They're not coming at the expense of one another, they're both coming at the expense of fossil fuels, so the more the merrier.

> Because we do whatever costs less. Nuclear plants are expensive. On the other hand, so is storage. It's not actually clear which one is going to cost less in the long term,

Right. It's really worth asking why exactly nuclear plants are so expensive to build, and why they were so much cheaper in the past. The plants that French built in the 70s and 80s were produce much cheaper electricity than fossils and renewables, in terms of LCOE. Why is US so terrible at building things today? Would it even be possible to build, say, Hoover dam today? How many more times it would cost now than it did in the 1930s, in inflation adjusted dollars, despite great advancements in the available technology?

> so what we should probably be doing is building both at the same time.

If you build nuclear, there's no need to build storage, really. Energy storage is very expensive to build, per stored MWh, much more expensive than nuclear is today, even at today's greatly inflated prices.

> Right. It's really worth asking why exactly nuclear plants are so expensive to build, and why they were so much cheaper in the past.

I tend to think it's a combination of factors that mostly boil down to bureaucratic inefficiency and various forms of political corruption.

The problem is there are multiple groups who profit from it being expensive.

First you have the people actually receiving the money. Contractors and unions. If it costs ten times as much they get a ten times bigger contract with ten times more hours.

Then you have the competitors. The fossil fuel industry is committed to ensuring that nuclear reactors are "sufficiently" regulated.

And the problem is there's nobody on the other side. The people paying the money are the utility ratepayers. They're not organized.

Maybe the key to fixing it is recognizing that the two groups in favor of higher costs aren't actually aligned. The contractors and unions want lucrative construction work, the competitors don't want construction at all. But if nobody's building then you don't get a big fat contract, you get a big fat nothing.

But at the root it's a political problem. Existing laws make construction expensive. Either fix them or expect it to continue to be expensive.

> If you build nuclear, there's no need to build storage, really.

Solar is cheaper than anything when the sun is shining, and there are a lot of things -- charging electric vehicles is a big one -- that can be demand-shifted into that time period with price incentives. But as soon as you're using it you've got an issue. Sunlight doesn't perfectly align with demand. You need something to get you over the demand peak in the evening. But storage enough to cover the base to peak load differential for two or three hours a day is a whole lot cheaper than needing to cover the full load for the full night.

Interestingly, some of the newer nuclear reactor designs have built-in storage. A molten salt reactor can use hot salt as thermal storage. Let solar handle more of the load during daylight while you put heat from the reactor into a vat of molten salt, then have the extra heat to spin more turbines during the load peak in the evening.

My understanding from listening to conferences and podcasts is that it is a combination of a lack of mass production (every new reactor is rare and apart, France had economies of scale by building the same design all over the country), unnecessarily tight and ever changing regulations (having to change a design in the middle of construction is always going to derail a project) and lack of private financing (given the upfront nature of the investment with no positive cash flow for 10 years).

What you've just demonstrated is that it is feasible if we devote 10% of our resources to it.

Humans have run civilisations with 80% of the effort dedicated to gathering agricultural produce. 10% devoted to energy storage is workable. I don't like the idea but it is clearly doable.

It is doable, but clearly makes no sense when we can just build nuclear plants instead for tenth of the price.

> solar panels don't last "a few years more" [than turbines]

This is not a point against solar, but the "lifetime" of wind turbine installations is misrepresented in the article. It states "the life expectancy of a wind turbine may be less than 15 years." and it links one 7 year old article to support that idea, which is known to be false except in the sly sense that for sure a windturbine may need repairs at any time.

Modern windfarms are commonly financed to be built and make a profit from investment by selling electricity for 15 years. After that contract is up, however much they require new blades and/or refurbished or modernized generators - extending their life is due to be much cheaper than putting entirely new ones in their foundations. The relatively short supply contracts are very desirable to investors, not a sign of unreliability.[1]

[1] https://www.intechopen.com/books/stability-control-and-relia...

Breaking out energy demand by country doesn’t work, because the climate doesn’t care whether a ton of CO2 is emitted by an American or a newly middle class Indian. There is no strategy based on renewables and conservation that even plausibly allows the third world to reach the development of the first world while averting climate change.

However, the developing locations have the much larger populations.

But as they are still developing (with less lock in/sunk costs) perhaps they will be able to avoid some really poor emissions pitfalls and move into renewables more easily when the economics make more sense

I would talk also about different ecosystems. Some are more exigent in therms of energy than other.

If you run short of energy in january in many boreal areas (like Moscow or Bismark), and suddenly can't heat your home, you could face a serious danger of death. Gaz and electricity are a must. Same for Mojave or Death Valley if you can't use electricity to keep the temperature in acceptable limits.

There is a limit on how much the developped countries could lower their "lifestyle" (needs of energy) to get closer to the energy requirements of developping countries, just because winter here can be really harsh.

Hot areas (in developping and developped countries) will face the same problem but many developping countries in tropical areas would not have this problem (even if they could have other serious problems to solve, like the monsoon period)

The energy requirements can be very different in the different periods of the year also. If we talk about average requirements of energy that would be wrong.

Even 25-30 years is starting to look pessimistic.

https://www.greenbuildingadvisor.com/article/testing-a-thirt...

https://global.kyocera.com/news-archive/2009/0805_jeao.html

My feeling is that most significant degradation comes from problems with assembly weather sealing, connectors, and/or electronic components. The cells themselves should last for a long time with modest degradation.

Inverters on the other hand seem to be closer to consumables in PV installs, but even those should last 10-20 years depending on the model, and refurbishment is possible.

> renewables deployment will continue to be slower than new energy demand.

Does renewable deployment cover new energy demand as well as the fossil fuels we are supposed not to use anymore?

> renewables deployments are offsetting dirtier production methods

Renewables take their toll on the environment too. Batteries and solar panels need fossil fuels for their production, and the mining of rare metals is an extremely polluting industry (and often delocalized in countries with little regulations).

> If you look at the charts

The charts I looked at suggested that there will be both an energy and an environmental crisis.

It's very hard to argue over this topic as all we can do is cite experts. Maybe I'm attracted to the most pessimistic authors? where to look to have the scientific consensus on this matter?

And, of course, to build significant acreages of solar farms, you need to destroy ecosystems already existing there. Consider, for example, the case of the desert tortoises that were blocking the construction of the Ivanpah Solar Plant. The company building it had to spend over $20 million to have people literally walk over the area to find and relocate every desert tortoise living there. This was in the middle of the desert a few dozen miles south of Las Vegas.

That’s not the demand curve. That’s the increase in demand curve.

The increase in fossil fuel demand, however “lumpy”, is more than current total renewable output.

(Who is downvoting? this is a statement of fact, from the article.)

Good point, but I don't see why that matters. Exponential output increase will overcome linear demand increase in pretty short order.

Unfortunately true exponential growth of a physical system cannot exist - it'll always be some sort of logistic model that looks exponential near x=0.

I’m intrigued. Why cant true exponential growth of a physical system exist?

Do you mean it can’t be maintained forever? I can think of lots of physical processes that have exponential behavior for some amount of time.

If it is exponential. There are a lot of curves that look exponential at the base of the curve and then turns out not to be exponential