Understanding Solar Energy

The bit how about incredibly quickly PV has grown is a figurative slap in the face to Vaclav Smil. He had just ten years earlier said PV wasn't going to grow quickly because historically energy replacements took a long time.

https://vaclavsmil.com/wp-content/uploads/2024/10/scientific...

This retrospective on Smil's predictions four years ago is notable:

https://www.quora.com/Is-Vaclav-Smil-right-in-his-criticisms...

"To get 1 PWh/year of electricity you need to install about 450 GW worth of solar panels. You need dozens of years to acomplish such task. Reality check: 3 years in current speed, in the future probably faster."

Indeed, as the thread top link shows in 2024 the world installed 595 GW of PV.

As John Kenneth Galbraith said, "If all else fails, immortality can always be assured by spectacular error."

What people tend to forget is, that coal, oil and gas are all restricted by mining or drilling as the old are consumed, and it gets harder to access new oil wells etc. For PV there is no such limit (only copper basically, but this is recyclable and aluminum can do many tasks.) For batteries, there is lithium (lifepo4) and even that is questionable (sodium batteries) and again there is the potential for recycling. Hence, I do not see anything stopping the exponential growth of PV and batteries.

The IPCC & IEA grossly underestimates PV (and Wind) by any metric for years. Many scenarios assumed costs for 2050 that are already outdated today.

In the same time they overestimate Nuclear Energy and carbon capture by any metric (debatable). It’s getting so bad that there are numerous studies about that problem.

https://www.carbonbrief.org/guest-post-why-solar-keeps-being...

https://www.pv-magazine.com/2021/03/31/solar-still-largely-u...

https://www.theenergymix.com/leading-climate-models-underest...

https://climatenexus.org/climate-change-news/iea-historicall...

You are right.

But one misconception I often read is that everyone focuses on batteries. It would make more sense in general to talk about energy storage instead of just batteries. Like Kinetic, chemical, thermal and so on.

Batteries cannot be solely responsible for back-up. You need different types of storage: short term, medium term and long term storage.

There are different concepts for each application. Batteries, compressed air storage, pumped storage, kinetic, thermal storage as well as power-to-X systems are able to absorb the increasing summer power and provide the energy again in the medium term or seasonally shifted.

https://doi.org/10.3929/ethz-b-000445597

There are only three energy storage forms that are relevant for the next decade. All the others looked promising, but the learning curve on batteries has rendered them irrelevant. Your link is from 2020, it is out of date.

The best energy storage form is "final form". Some energy products can be stored. For example if you are using the energy to create heat, you can store heat for use in the future. Heat storage sucks as a way to store energy destined for electricity, but is a great way to store energy destined for use as heat.

The utility of batteries for daily storage is obvious and well proven.

Thirdly, the best annual storage is pumped hydro. It's the cheapest and it can be used pretty much everywhere -- all you need is water at one end of an elevation change and a way to build storage at the other end.

All the other forms that you'd think would fit in between the two are being quickly subsumed by the rapid price drops in battery pricing. The cutover points are rapidly shifting -- batteries are now cheapest for biweekly-ish.

And the primary sources are getting so cheap that overbuilding is an alternative to storage. Rather than storing for the reduced amount of daylight in the winter, just overbuild. More overbuilding and a few days of storage will let you handle a stretch of cloudy, windless days in January. No annual storage required.

I think this is going to turn out like the exotic panel chemistries: batteries are simple and have powerful continual improvement in performance and price, while the others turn out to be more complicated. In particular solid state wins over mechanical anything almost every time.

There is a lot of stuff that people said about this solar that got overtaken by reality. And some of those people were proponents even.

People have underestimated economics, learning effects, and the effects of increased scale. Mostly the exponentials were actually pretty clear to some investors as early as 15 years ago. And the success those investors have had, has driven more investment.

The thing with exponential trends is that doubling a little bit results in a little bit more. It doesn't add up to something people notice until suddenly it jumps from fractions of a percent, to full percents, to double digit percentages in the space of a few years. That threshold got crossed a few years ago and people started to notice. And that's now leading to further price drops and more adoption. Of course, it's not a real exponential but an s-curve. But until the curve flattens, you won't be able to tell the difference.

Back of the envelope calculations can be misleading because they tend over simplify and make silly assumptions. Like assuming we are going to move 100% of energy to solar all at once. In reality, what we're doing is a decades long transition where most of the decision making is cost driven and the energy supply is coming from mixed sources.

We don't have just solar. We have existing nuclear. Existing deployments of coal and gas, which like them or not are not going to disappear overnight. And a lot of onshore and offshore wind. And a rapidly growing amount of batteries and cables which give us the ability to time shift supply and demand and move energy around over large distances.

The world's electricity consumption is about 30 PWh per year and will probably grow to 35 or 40 soonish. Most of that growth (>90%) will be powered by renewables. It's outgrowing everything else by a large margin. And because they are cheaper, there is also pressure to replace existing generation with renewables. That basically happens based on cost and age of plants.

This is another effect that people keep underestimating. The reason coal generation is rapidly disappearing from many markets (and is completely gone in some of them) is that replacing them with cheap renewables is cheaper than continuing to operate them.

That same effect is going to affect gas generation. Anyone building gas plants with the expectation that they'll have a 60 year life span is dreaming at this point. These investments should be considered as under water at this point. By the 2050s, most currently new gas plants will have probably have been mothballed (maybe kept around as rarely used peaker plants) or demolished. They are simply too expensive to operate relative to renewables. Some places keep gas prices low via subsidies (the US for example). But even there gas plants are going to face a reality check. And for a lot of countries, gas imports are a drag on their economy. Germany is a good example.

Worth observing what investors do here. They tend to have long term outlooks.

I think a lot this comes down to huge cultural biases. And the two cultures are "hard energy" and "soft energy" folks. Coal, gas, fission, fusion, etc. are all hard energy. Coupled GDP and energy consumption was a core assumption. Renewables, energy efficiency, technological advancement via learning curves all fall under "soft energy".

Most of the energy industry was hard energy because that's what paid everyone's bills. Any estimates that did not cater at least a bit to those biases would just be completely ignored.

But there's another effect too: solar just completely outperforms even the most optimistic assessments. There's one famous solar financial analyst, whose name I'm blanking on, who continues to underestimate even though she knows the effect.

Don’t get me wrong—I’d be all for batteries ruling the world if they were both affordable and technically advanced enough to meet various demands. That means they shouldn’t degrade too quickly, for example, when capturing and releasing wind energy in milliseconds. Or they shouldn’t lose too much energy over time due to self-discharge. Or they should be able to supply large amounts of energy instantly. Overbuilding is also a valid approach, especially in connection with a smart grid spanning multiple countries. All of that is fine.

However, the point of the study is different, and that makes it still relevant today: The barrier to expanding energy storage isn’t a technical one—it’s a political one. The study also shows that there is a great deal of variability, and the often-used argument that there’s not enough lithium or rare earth elements doesn’t hold up. More recent studies validate different storage technologies depending on their specific use case, showing that they can complement batteries in a meaningful way—also from a financial perspective.

Another perspective is that we still have a long way to go before full electrification. Right now, batteries are used in suitable scenarios, but many other areas haven’t been electrified or optimized at all. Other storage technologies might still become relevant. Building a house around a 20,000-liter tank to store energy for heating in Alaska over six months might already be financially and technically viable. But whether the logistical challenges of such solutions will ever make them truly feasible—that’s something I neither want nor can predict.

I am on you side (but not all are more complicated and there are mechanical variations that are better than batteries for some scenarios) but the takeaway of that study is described here: https://news.ycombinator.com/item?id=43425560

It was also underestimation of China. Outright chauvinism there.

PV doesn't require much copper, either. Maybe for front contact wires, if silver gets too expensive? But if it can afford to use silver now, copper won't be a huge ask (basically just need to deposit a barrier layer to keep the copper from reacting with the silicon.)

The cables connecting PV to the grid, as well as the grid itself, can all use aluminum conductors. Even large transformers can be designed with aluminum if copper gets too expensive.

> Thirdly, the best annual storage is pumped hydro.

I strongly dispute this. E-fuels like hydrogen would be much superior to PHES for annual storage.

https://x.com/iain_staffell/status/1722544993179504965

iCal them „simple“ and „complex“ power. For someone who isn’t truly informed, a „Simple Energy“ solution seems much simpler than one based on renewable energy. With „simple“ power, solving climate change appears straightforward: just build more nuclear plants, which conveniently replace coal and gas on a 1:1 basis since they are baseload power generators.

Renewable energy, on the other hand, is (for now, the transition time) complex. It requires a better, smarter, and much larger interconnected grid, as well as intelligent management of supply, demand, and storage. It means considering and understanding multiple aspects at once. This complexity often leads people who are convinced that more simple power is the answer to dismiss the idea of renewables too quickly—because nuclear seems so much simpler.

I understand the appeal of simple energy. The sad part is that many people likely believe this is the scientifically correct position. And they are often so convinced that, even when presented with current studies and reasonable arguments against new nuclear plants, they quickly assume that the other person is just an irrational, biased anti-nuclear activist. After all, the simplest solution must also be the right one, right?

Being informed in this context doesn’t just mean knowing the pros and cons of nuclear, wind, or solar power. It requires a deep understanding of what is technically and financially feasible today—including energy forms, grid transformation, storage solutions (not just lithium-ion batteries), follow-up costs, sustainability (mining, waste disposal), as well as political, economic, military, and social implications. And how all of these factors interact.

But none of that is necessary if you just want to build more simple power plants.

The transition to 100% renewable energy is as complex as the development of the internet. If we were still relying on letters, telephones, fax machines, newspapers, radio, and TV, the idea of transitioning to a globally available, instant multimedia internet would have seemed just as utopian and impossible.

Only if we have more than enough renewable energy to spend making hydrogen. Hydrogen storage has a round-trip efficiency of 40%-50%, leading to significant energy losses. Partly by: Electrolysis requires 50-55 kWh to produce 1 kg of hydrogen, which only contains about 40 kWh, resulting in a 20%-30% energy loss upfront. It’s low energy density requires high-pressure or cryogenic storage, increasing costs and energy use, while leakage further reduces efficiency. Limited pipelines and refueling stations make hydrogen adoption costly and complex. Highly flammable hydrogen demands a lot of safety measures adding even more cost and complexity.

At the current exponential growth rate, PV will reach the point of supplying the entire world primary energy demand in a decade and a half.

Yes, hydrogen has low round trip efficiency. But it comes out cheaper than PHES. The "cost of inefficiency" is proportional to the number of charge/discharge cycles. For annual storage, efficiency is 365x less impactful than it is for diurnal storage. What matters for annual storage is capex of storage capacity.

> What matters for annual storage is capex of storage capacity.

Which is exactly why PHES wins the cost comparison for annual storage. Open air water storage is ridiculously cheap compared to hydrogen storage.

Jenny Chase perhaps:

> On Friday my colleagues suggested I get a tattoo reading "COWARDS", to save me time saying it in solar forecast calibration meetings.

Smil was just bullshitting though, really poor quality arguments made for rhetorical effect with a side helping of smug fake reasonableness.

He's a cranky old academic propelled to fame because he said what the establishment wanted to hear like an energy Jordan Peterson.

I dispute this as well. From what I see, the very best case per kWh cost of just the reservoirs and waterways for PHES is about $10/kWh. Hydrogen stored as compressed gas in solution mined salt caverns would be an order of magnitude cheaper. For storage of liquid e-fuels in tanks, tank capex would be another order of magnitude cheaper still. This assessment is consistent with the link I posted earlier.

If you want something that may compete with hydrogen for annual storage, consider bulk thermal storage (using artificially injected heat, not naturally occurring heat). The thermal time constant of a very large object increases quadratically with radius, if everything is scaled proportionally, and can easily reach many years. This is why geothermal works at all -- there's plenty of heat stored in the near crust ready to be mined.

That's her! She's done really amazing work for BloombergNEF.

You're comparing using an existing reservoir for hydrogen to building a new reservoir for PHES. There similarly exist dry lake beds that could be used for water storage. But generally they're not in suitable locations, which is the same problem that salt mines will have.

You're also comparing hypothetical costs to historical costs. Hypothetical costs put out by industry are usually out by about an order of magnitude.

There's a reason that PHES is the only one with historical costs.

> Thirdly, the best annual storage is pumped hydro. It's the cheapest and it can be used pretty much everywhere -- all you need is water at one end of an elevation change and a way to build storage at the other end.

Pumped hydro is primarily used for short term storage. The vast majority of pumped hydro installations around the world operate on an intra-day cycle.

For storage systems generally (not just electricity), profitability is a linear function of capacity, the possible price arbitrage AND how frequently you charge and discharge. Nobody is going to build a pumped hydro storage facility with the intension of operating a single charge/discharge cycle per year.

Nor are pumped hydro facilities cheap to build and certainly cannot be deployed everywhere as they require particular geographic and geologic conditions and mostly locations suitable for pumped hydro are few and far between and those locations that are suitable are generally far away from population centers where the demand for electricity is.

Batteries are often cheaper than pumped hydro, they can be located near demand, they scaled down as well as up and can be distributed around the grid to provide "virtual transmission". They are quick to deploy and require little maintenance or staffing.

The solution for "long term" storage will be massive over-provision of wind and solar and more grid interconnections. Batteries will take care of everything else.

No, I was describing the cost of constructing a new hydrogen storage reservoir in a salt formation by solution mining. Of course existing natural gas storage caverns could be repurposed; that would be even cheaper.

These are not hypothetical costs. Construction of these caverns is state of the practice for natural gas storage. Vast volumes of gas are stored in these things, allowing steady production of natural gas and constrained pipeline capacity to serve seasonally unsteady consumption patterns.

The reason PHES is the only one with historical costs is that, historically, PHES has been used for diurnal storage, from the days when baseload plants were cheaper. There was never a market for long term storage via hydrogen (although some hydrogen storage has been constructed and used to help steady the hydrogen input to ammonia plants); why bother for the grid when just varying the use of fossil fuels would serve that function just as well?

I'd be curious if either of you would have a link to actual projects (projected or realised) with their respective costs?

Here's a presentation of a comprehensive NREL study from 2018, but I don't know the source of the numbers. It finds hydrogen and flexible generation (that is, natural gas turbines) are best for long duration storage. Notice the slides on page 13. PHS is way out of the running for the storage case being discussed here; it's not close.

https://www.nrel.gov/docs/fy21osti/77833.pdf

Nuclear can’t (cheaply) replace gas.

Indeed. Particularly in the US, fracking was an extinction-level event for new nuclear construction.

https://pubs.aip.org/physicstoday/article/71/12/26/904707/US...

“The cost of new nuclear is prohibitive for us to be investing in,” says Crane. Exelon considered building two new reactors in Texas in 2005, he says, when gas prices were $8/MMBtu and were projected to rise to $13/MMBtu. At that price, the project would have been viable with a CO2 tax of $25 per ton. “We’re sitting here trading 2019 gas at $2.90 per MMBtu,” he says; for new nuclear power to be competitive at that price, a CO2 tax “would be $300–$400.” Exelon currently is placing its bets instead on advances in energy storage and carbon sequestration technologies.

Thanks. I mean it’s from 2018 and that is ancient history as far as storage costs are concerned. But yeah those $/kWh numbers for PHS are orders of magnitude higher. Thanks for the link, I’ll try to find the final study tomorrow