Hacker News new | past | comments | ask | show | jobs | submit login
World Energy Hits a Turning Point: Solar That's Cheaper Than Wind (bloomberg.com)
785 points by Osiris30 on Dec 15, 2016 | hide | past | web | favorite | 301 comments

I have high hopes for ARES [1]. just a heavy train on a hill with a regenerative break. Add lots of tracks, multiple trains per track, solar above the rails. Avoids needing all the water for pumped hydro storage. none of toxic stuff to deal with when batteries have reached their end of life. Zero carbon emission (aside from whatever is required for manufacturing) is just hugely appealing.

If solar prices continue as they have for another 3-5 years, the question is going to be pretty clear, how do we store all of this insanely cheap power. I'm a little mystified we're not taxing carbon emissions and subsidizing storage. But hey, there are clearly powerful forces at play, that don't agree with me.

[1] http://www.aresnorthamerica.com/

I'm extremely skeptical that ARES will ever be cost efficient. Potential energy is proportional to mass * g * d_height - any reasonable grid storage option will need to be moving truly enormous amounts of mass.

Back of the napkin here:

California consumes 9.54 * 10^17 Joules of electricity / year, or 1.09 * 10^14 Joules / hour.

Suppose we wish to hold in storage enough energy to supply to grid for twelve hours, since the sun does not shine at night. Assume we have a 100% efficient method of recovering potential energy stored on a train. Suppose we find a suitable mountain with a 1000 meter height differential. Then according to e = mgh, we would need to shift a load of 1.33*10^11 kilograms to supply twelve hours of electrical demand.

Let's assume we use the cheapest rock we can to fill this mass. I assume you can get wholesale pea gravel for $8 a metric ton.

Then the cost of gravel alone is $1.07 billion.

That doesn't sound so bad, but remember, this project will need ~1333500 100-metric-ton freight cars. Suppose they cost $120,000 / ea. Then suddenly, the cost of the train cars is 160 billion.

This is on the order of magnitude of the entire California yearly state budget. And also ignoring a bunch of other cost factors that will increase the total price by another 1-2 orders of magnitude.

Pumped storage hydroelectric is viable precisely because the reservoir holds your water for essentially free. Train cars are far, far more expensive.

It's not a forest of Sisyphean gravel funiculars which your calculation presumes.

According to the brochure on the site the Ares system seems to be an automated freight yard on a slightly sloping ground. Only a fraction of the total mass is on wheels at a time. The scaling issues might be a bit different.

>It's not a forest of Sisyphean gravel funiculars

I strongly feel this expression belongs somewhere. I just have no clue where.

Song lyrics, or poetry.

I personally like the idea of turning really deep mines into literal potential energy wells. Use renewable energy to lift loads (water, gravel etc) and then power generators at night by dropping the load back down. Basically enormous cuckoo clocks.

While I like the idea for its steampunk appeal, looking at potential/kinetic energy delta, you have two ways really: build overcapacity for solar, immovable, uses rare earth minerals, but other than dust, maintenance free and then that capacity will sit idle at night or you build capacity like these lifts at that capacity is going to sit unusued during the day, but it does require a lot more maintenance due to moving parts and potentially more resource heavy to build, but no exotic materials. I think the best way forward is to, of course use the least amount of energy possible - insulate from both heat and cold, make efficient CPUs and run everything you can in software and anything physical make it run during the day on solar. Once environmental costs are accounted for, it makes no sense to run things like cars and shops and washing machines at night.

Paul Simon song.

That system needs to be called Sisyphos.

A more apt figure from Greek mythology than the god of war, though, might be Sisyphus, who was condemned by the gods to push a rock to the top of a mountain, only for it to roll back down again so that he had to repeat the punishment for eternity. ARES does indeed push rocks uphill, only to let them roll down again. - See more at: http://www.aresnorthamerica.com/article/9232-sisyphus%E2%80%...

You would need a lot less rail cars than that. Rails cars move the weights between the low and the high point, but they are re-used multiple times in a cycle, going back and forth constantly.

It wouldn't be very efficient to have a mobile piece of tech sitting idle for 99% of the time.

The main constraint would be to make the maximum use of cheap technology that is standard today. 100t freight cars are pretty standard.

Edit: nevermind, I was wrong

That's not how the proposal works. Each car picks up a box of rocks on a hydraulic jack and hauls it up the hill and then sets it down again and gets another one. Maybe the railcar makes 10 or 100 trips during the day.

But you still need many cars to a supply high volume of energy per unit of time. You could do the whole project with a single train, but then it would be very slow at storing and retrieving energy.

Yes, you're both right. My fault – I misunderstood the ARES concept.

All three of you are right: Ares advertises two different concepts, only the one they call "grid storage" uses more ballast than they can fit on their rolling stock.

The one they call "ancillary services", intended only for short term grid regulation/stabilisation", uses fixed ballast that is never removed from the trains during regular operation. This is what the first prototype will be.

Everybody wins!

Just a few minor nits, you are off by a bit in your calculation of the cost of the train cars. A 5800CF open-topped hopper car will hold a little more than 200 metric tons of gravel and will cost less than 50K for the car and the gravel to fill it. Now you need 600000 cars at a cost of 50K each so the cost has dropped to around 33 billion. If you were buying rail cars at this sort of quantity you can probably also get a much better deal :)

I also seriously doubt there are going to be cost factors that add one or two orders of magnitude to the price here, but it is possible. The rail itself is very cheap and land to site the project is similarly easy to find since the rail line does not actually need to go anywhere and you can run lines in parallel instead of needing a long distance.

Definitely not cheap, but pumped hydro storage is in no way "free" and there are really no places you can do pumped hydro in any large scale that are not already used for existing dams and rivers. The lack of significant environmental impact is also a win for something like ARES.

What's the maintenance cost of the lines and equipment vs the hydro ?

Probably much much cheaper for rail lines. Rail doesn't leak. The entire infra is open to direct maintenance, and you can easily take segments of the system out of production for repair without having any significant impact on the productivity of the whole. You can extend or lay parallel track while your main system is still operational without too much bother so expansion is also easier.

I wouldn't be so quick to dismiss the maintenance cost for a ARES-type rail system. For pumped hydro, your ongoing maintenance is largely going to be maintaining a pumping system (to send water up) and a generator system. Hopefully you build your reservoir properly so that it lasts for 30 years. For rail, you have to have to maintain a large number of moving parts. It won't be cheap to insure all the parts are greased and good to go. I think it's a lot easier to maintain a set of pumps and generators than it is to maintain what is essentially a rail fleet filled with gravel on a mountain.

So for pumped hydro you have a massive up front cost and potentially smaller maintenance. Maintaining a large number of moving parts for rail might seem daunting, but these rail cars are made to operate for very long periods of time in unpleasant conditions with minimal maintenance; topping up lubrication systems and visual inspection is basically covering monthly maintenance for the system. I am not sure what the ongoing maintenance cost for most hydro installations is (pipe inspections, anti-silting and other water maintenance, inspections of the system, etc.), but I had to guess I would say that it is probably quite similar to rail.

Standard rail cars are sturdier than they'd need to be for this application - if you were going to buy that many you could pull out a lot of the cost by making them fit for purpose.

True, but all your math assumes that energy is used equally all 24 hours of the day. A large chunk of energy is consumed during the day, when the sun is shinning. Much less is used at night when the trains need to supply energy.

If anything I think my estimate is too conservative. Electrical storage must be adequate to meet demand in the worst case. The minute somebody's grandmother freezes to death because of brownouts due to wintry, overcast skies in December, and there will be riots in the state legislature.

One could argue that, if people freeze to death because of brownouts due to wintry overcast skies in December (weeks of freezing weather probably would be something different), there's something wrong with the way houses are built.

Getting rid of oil addiction doesn't solely mean replacing oil with wind and solar; it also means thinking rationally about why we need so much energy.

Even the most energy efficient home needs a heater in freezing conditions.

A modern passive house could keep you warm in the winter with a heat source as minor as a candle. It's actually amazing how poorly our homes are insulated.

Don't you need souther exposure too?

Passive solar homes with heat storage may not.

Once you're talking about tearing down all existing houses and rebuilding them based on new and alien philosophies, your plan is starting to become a bit impractical.

> Getting rid of oil addiction doesn't solely mean replacing oil with wind and solar; it also means thinking rationally about why we need so much energy.

The prominent people who say this sort of thing usually have enormous mansions with heated pools and fly private jets everywhere.

No, we are long over designing systems for the worst case. This is what got us where we are today. The smart grid is about keeping grandmother alive tonight while drying your clothes during the light cycle tomorrow.

And pray that worst case scenarios don't happen.

Most people work during the light cycle.

Would this smart grid require that people only ever wash clothes on days they don't work?

You can wash clothes while you're not there. It's even possible to schedule the washing so that it's done near the end of the light cycle so they don't sit in the machine for long after they are clean.

The desire to be able to do whatever you what, when you want, is part of how we got to this climate-changing situation in the first place.

Good luck winning elections with that.

There is a very good chance i've screwed up my math, but my silly back of the envelope calculation says,

9.54 * 10^17 Joules/ (3.6 *10^9 Joules/mwh) = 265000000 mwh. Levelized cost, [1] (operating + construction + financing) works out to around $72 per mwh, over 30 years. (going with the cheapest average option) So if we were to throw away everything today, and be required start fresh, I get about $570 billion.

That's clearly a stupid thing to do.

However, my big point is, if solar falls another order of magnitude (which is a pretty big if), we can afford to spend a lot on storage. I think solar is now competitive with coal, but it's not dispatchable.

I'd argue solar would probably need 4x the amount of production of a similar coal plant, it's only going to work really well part of the day, and there are losses in storage (ares claims 20% loss, which is comparable to batteries)

there's not much weight difference between gravel and packed dirt. I figured they'd just use whatever mass is handy when they're building the track.

I don't know if $120k is a fair price. the cars have to have regenerative breaks, which are likely expensive. Just building something strong enough to support 100 metric tons can't be cheap. I don't know if there are any special features of regular freight cars that can be abandoned to cut down costs. On the other hand, if you're building a million of them, the costs would likely come down a bit.

Perhaps solar has reached the end, and any improvements will be slow. I don't believe that, and i do believe some sort of efficient storage will be the big focus in energy going forward.

[1] https://en.wikipedia.org/wiki/Cost_of_electricity_by_source#...

The further you build your renewable energy grid over peak consumption, the less storage you need, because if, let's say, an input of 50% electricity of installed capacity is 100% of peak consumption, the days where you actually need to draw from the storage will be less and the days where you can fill the storage will be more, in simple terms. There are studies on this.

IF solar falls another order of magnitude, it will be far cheaper (which only has the conversion efficiency problem) to go for power-to-gas technology that has the additional advantage of being able to use gas (not gasoline) powered cars with normal range.

IIRC one of the big ideas behind ARES is to make money by supplying peak power for consumption and possibly even by taking peak power production off the grid.

Basically a grid scale AC capacitor equivalent.

I only have education in electronics and no other experience with High Voltage or big grids so I might be very wrong.

If power-to-gas means methane, it's bullshit. You'd need a source of carbon dioxide, and despite all the hype, it's only a trace gas in the atmosphere. (Personally, I think power-to-gas is a propaganda piece invented to link renewables and natural gas in the minds of gullible people.)

How does that compare to other energy storage though?

Favorably. 84% is 16% loss. batteries are 20-30% loss. hydro can compete, but the southwest has lots of sun and hills. not so much with huge supplies of cheap or free water.

the southwest is nice, because 300+ days of sun. cloud cover may reduce efficiency, but might actually improve it. clouds reflect more light from the ground.

There are many options for storage. Trains are very very good. You need to beat 80% to beat a train. Some things can do that, but they're not cheap, like land in the SW.

For reference, 133,350,000 metric tons is 1.81x the mass capacity of the US strategic oil storage, another grand energy storage scheme.


Continuing the back of napkin calculations, to compare train cars with batteries:

Rail cars costing $10^5 can hold 100 metric tons or 10^5 kg each. Cost is $1/kg. This will be amortized over the system's useful life. 1000 meter height differential at 10 m/sec^2 stores 10^4 J/kg. System stores about 10^4 J/$.

Tesla's Powerwall stores 10 kWhr for $3500. That's about 10^4 J per $ (since an hour is 3600 seconds).

What surprises me is that these order-of-magnitude costs are about the same -- 10^4 Joules per dollar.

A 5800CF open-topped hopper can hold 200 metric tons of pea gravel (at least according to the people selling them) so you just doubled the J/$ of the Powerwall. This is also using new rail cars, if you buy used then the price for the same rail car drops down to the 20K range according to online rail car brokers. Now we are looking at 1/4 the price for a train car solution.

> will need to be moving truly enormous amounts of mass.

Use the extra energy to increase the spin of the earth, then harvest the spin to pull off energy.

This is what every flywheel already does (unless it is balanced by a second flywheel going the other way). Conservation of angular momentum: as a flywheel gains angular momentum, it must be compensated by the Earth gaining the same amount of angular momentum in the opposite direction.

Usually this isn't referred to as "spinning up the Earth" though :-)

If that was easy and cheap to do, our problem wouldn't be climate change due to greenhouse gases - it would be climate change due to the earth rotating slower since we'd been using it as an energy source for 200 years!

You'd speed it back up in times of energy surplus. It's a giant flywheel.

Rotational energy is already harvested using tidal power and wind power.

There are band type "windmills" it's a taught band with magnets on each end when it flaps it moves the magnets to generate power.

>> the reservoir holds your water for essentially free

Once you build a dam, that is.

Friendly calculator if you want to play with the various assumptions:


As others said, you can reuse cars (if a car does 100 1km journeys per charge interval the materials cost for cars and gravel become comparable) and a custom built car can be cheaper, too.

The volume may be another challenge though: 1.33*10^11 kg is roughly 0.1 km^3 of gravel/dirt. So we need to move a cube of dirt ~450m in size up and down 1km each day.

Not impossible, but presents bit challenges on its own imo.

I suspect the whole 'needs a hill' and amount if land taken becomes a problem as this cant be taken for granted as available. There are other solutions that seem effective and take so much space. Maybe something like this? http://www.gravitybattery.info/

Unless the state budget pays for the entire energy industry, that seems like a curious comparison to make.

You started with the total energy (I presume electrical) usage of the state. About how much do Californians pay for all that electricity per year?

I don't know anything about freight car pricing, but these cars should be 1-3 orders of magnitude cheaper than regular high speed for expensive cargo.

The ideal one is basically a rock with wheels.

The Flintstones model of energy storage.

I would think the idea has merit if rather than purpose-built rails, the existing rail infrastructure is retrofitted (and built out) to act as a storage mechanism with freight that already has to move anyway. If that's done though, a proof of concept would need to be built first. So I wish these guys all the best, despite the fact that the economics look kind of silly on the surface.

In this case you are talking about electrifying the entire US rail infrastructure. Definitely not cheap...

And then you compare these costs to atomic power and it becomes clear that it is the only real option.

Actually you made it sound pretty attractive. It's only $160 billion (plus a few for yearly maintenance why not) to get California entirely on renewable energy. That's amazing and gives me hope.

And now you do the same with pumping water back and forth... I get it that trains are handy when you have plenty of space and lack water, but many, many places around the globe lack the space for the hughe amounts of train tracks running up a 1000 meter high hill.

In addition, take into account the enormous amount of moving parts. Maintenance everywhere! Compare the tracks with carts to a large pump/generator and pipelines with water I guess the latter is much simpler to maintain and better looking to the eye. While train tracks should be kept clear of people you could easily see people using the lakes up & down the mountain for recreation.

You still need to generate the energy in the first place.

California state budget is $167B.

As neat as this sounds, in a Rube Goldberg sort of way, batteries seem more reliable without the moving parts. The new Lithium Titanium batteries are amazing. 10k cycle before loss of ~20% capacity and fast charge/discharge cycle of 6 min. Since they last a long time, less e-waste. And lithium is supposed to be relatively recyclable when the time comes.

We used to think building heliostats made more sense, but it turns out cheaper cells won out to complexity. I suspect this will be repeated in the energy storage market.


I would like to point out that moving parts are what keep our civilization running. Intricate moving parts.

To feed the intuition:There are machines that have run with minimal maintenance for hundred years, such as the original newcomen engines used to pump water away from mines.

Each time an assault rifle is fired lots of small parts work together in a delicately timed dance. Now, take an AK-47 - due to it's massive parts and loose fitting it can be filled with dirt and sand and happily continue operation.

And, it's quite hard for me to imagine the turbines in electric plants and all the "piping" around them are any less "rude goldbergish" than the Ares system would be.

Machines are quite durable if they are designed that way.

> Now, take an AK-47 - due to it's massive parts and loose fitting it can be filled with dirt and sand and happily continue operation.

Not quite. It is statistically more reliable in such conditions in a sense that it will have a higher mean rounds to failure number compared to most everything else. But MRTF won't be infinite. In fact, it'll still be low enough that you will definitely notice.

Then there's the reason why. Loose fitting and massive parts, yes - but that also requires overgassing to move those massive parts (and then it also moves them faster than they actually need to be moving to cycle). So you're basically sacrificing energy efficiency for the sake of some extra oomph to clear problem. On any cycle that doesn't actually have a problem to clear, the sacrificed energy is wasted.

Then there's the normal metal stress and fatigue when running something 24/7 non-stop. Here's what happens when you do it to an AK - the guy is running a full auto rental range in Vegas for tourists, and shared his experiences about how soon guns fail, their failure modes etc: http://www.ar15.com/forums/t_4_64/159106_AK_abuse__update_on...

It's fun to read about a guy who's clients pay for him to shoot an AK-47. Where I'm from a male goes to jail (or other service) if he refuses to train for it (well, a local copy https://en.m.wikipedia.org/wiki/Rk_62). Another man's hobby is another man's burden :)

Based on what the guy said, his clientele is mostly young men who play video games like Battlefield or Call of Duty, and want to see how the same guns there work in real life. So they literally come asking for "that gun that was in CoD", hence why he mainly stocks things that are trending in games and movies.

Apparently, it's also popular with foreign tourists; especially from Asian countries, with more extensive civilian gun control than in Europe, where many wouldn't even see a military (or military-looking) firearm outside of service. It got to the point where some ranges added signage in Chinese and hired Chinese-speaking instructors to cater to the new audience:


Good points. I doubt any rifle can be fired continuously for hundred years :). A boring combustion engine would have been a better example of mechanical robustness all together.

I would tend to agree that solid state is better than moving parts for reliability, but we have been building trains for like 150 years and the parts here that move seem to be just really big weights. We also don't have batteries that can do city scale storage economically, but a series of trains seems intuitively do-able. This is just intuition so could be wrong, but trains seem so simple.

Trains are not reducing in price like lithium batteries are. Lithium battery prices are coming down yearly. If it continues at this rate, I think we'll have a story like this about energy storage in 5-10 years where the price makes renewable storage the goto choice for general energy needs. That would be awesome, planet earth could use some good news like that.

South America produces most of it recycling may be clean but production seems dirty.


I've just read through the article: what do you see as 'dirty' about the salars? The fact that workers wear masks?

Like any mining operation you have to move a lot of earth use a lot of energy and dig up large regions.

The processing of the slurry to get the lithium after it's mined. All that water needed in a very arid region.

It was in contrast to lithium being easily recyclable compared to the vast scale of mining it.

It is a bummer though that, from what little I know, the process to manufacture batteries is not good for the environment. That said, probably the process to manufacture all the parts of ARES is not either. And, probably, both of those issues pale at the moment in comparison to the need to shift to renewable sources of electricity.

The problem: Gravitational potential energy isn't a very efficient storage format. 1kg elevated 1m is only around 10j. A 10,000kg mass elevated 1km would be 98Mj- only 27kWh. And only with perfect conversions, which you won't get.

Each Tesla powerwall holds 14kWh for around $8000 each, and that price may fall over the.

Just to put this in perspective - a gallon of gasoline holds about 33 kWh (about 10 kWh of electricity after thermodynamic losses). So lifting 10 tons a km vertically only gives you about 1-3 gallon of gas, give or take (however motor efficiency could be 95% which helps).

There is merit to lifting large loads like this, but IMHO they should be concentrating on weight instead of distance. Which basically takes us back to pumped hydro because we just don't have anything on the order of a million tons or more that can be moved cheaply and easily. I could maybe see a large carpet rolling a layer of sand or stones up into a cinnamon roll so it's self supporting. Probably not the safest place to stand under though!

What about using old shipping containers just filled with rock and dirt found on site?

If someone started with a hill or mountain at an almost suitable grade, then what they dug out to grade the slope properly could be put in second hand shipping containers at the top of a hill. Bring the containers up or down one at a time to get the power in or out.

Maybe have some one site solar or wind generation to try to bring and extra container up every once in a while.

People generally don't want steep slopes for much anyway so even the property ought to be cheap, it could even be municipal property for maximum cheapness.

The hardest part of any of this is the all electric trains with efficient regenerative brakes.

Water is worth it in very specific locations, because we can build a dam and let the natural shape of the Earth do the heavy lifting.

But a dry system, relying on weights? There isn't a material that is dense enough to make it worthwhile. Even a boxcar full of granite would need to be orders of magnitude heavier than it is for gravity storage to be economically useful.

It can be efficient in $$ terms since ballast is cheap.

Ballast is > 100x cheaper than powerwalls by the kWh in your equation[1]. 10 tons of rock just 5 m^3, you can put hundreds or thousands of this size of rock in a train[2]. You load/unload ballast instead of having all of it statically bolted onto rails and cars, so you get by with 1/50..1/100 carrying capacity[3] per trip vs total ballast amount.

The amount of raw materials (lithium mining and refining etc) required by the powerwalls is probably much higher too - good lithium deposits have ~ 1/100 of lithium content and turning ore into batteries is a heavy process consuming a lot of water, electricity, fossil fuels (ore transport/mining) etc. Not to mention limited battery service life.

[1] 2x powerwalls = $16000, even building material grade gravel is < $300 per 10 tons (50x cheaper) but here you could use sand or something from the site.

[2] big ore transport trains carry 10000-40000 ton loads, so 1000-4000 of our 10t rocks per train. You would probably use a smaller train for energy storage purpouse since the distance is small and you can have quick load/unload turnarounds

[3] if aiming for 1-day charge capacity, you can fit 100 15-min round-trips in 24 hours, or 50 30-min round trips

But it costs far more than just the ballast. The rails, the car, the huge system for pulling and releasing the cars, the dozens of km of land needed to have a train car be elevated by any significant amount.

Lithium systems are going to wind up being cheaper overall because of all of those costs, plus operating costs.

We already know how to cheaply and efficiently build rail, rail cars, huge systems for pulling and releasing the cars (they are called engines and overhead electrification :) and in fact you can pick up a lot of the materials used for less than half the cost of new. Dozens of km of land are damn near free in places in the southwest US where you can also dump lots of solar panels. Lithium may eventually be cheaper, but not within the next decade it won't.

Remember, the entire mass of the Earth @ 5.972 × 10^24 kg (which is 1.317 × 10^25 lbs) i.e. 5 septillion kilos or 13 septillion pounds) which is under you is enough to push you down with something like 150 pounds of force. A magnet with 150 lbs of holding power is able to fight all of that and takes up this massive amount of room:


Yes - that's a penny for comparison.

Gravity is soooo weak.

Your "Earth vs. Magnet" fight is not fair. For that magnet to hold up something of 150 pounds, the magnet has to be quite close to the object.

The Earth in that fight, on the other hand, is doing it with its center of mass (which is the relevant distance) about 6.7 x 10^6 meters away.

Squeeze the Earth down to the size of the magnet so you can move it close to your object too, and it won't seem so weak. The magnet would still win, I believe, but not by nearly as much.

Ah, good point! I'm not sure that it would still win versus point-sized Earth - according to this [1], shrinking Earth to half its size would increase gravity 4x. And a point is obviously much smaller than half.

But still, to see how incredibly weak gravity is consider that you don't even have to consider the gravitational pull of the magnet. Even if you walk next to a mountain you don't have to think twice about falling toward it. Gravity is negligible at even monstrously huge scales! It's just not a very large force at all, compared with magnetism, electricity, or e.g. solar radiation. It's truly weak. I've never had to think about the gravitational pull of any object other than literally the Earth or maybe occasionally the moon (which has a mass of 7.35x10^22 kg, or 1.2 percent of Earth's mass -- i.e. the Earth and the 5 septillion kilos I quoted is only 81 times as large.) The everyday pull of objects (for example the pull by which the building next to you right now is attracting you) is so negligible. It's weak!

[1] https://www.quora.com/If-the-earth-shrinks-half-of-its-radiu...

Yeah, I expressed myself poorly when I said shrunken Earth at the same distance from the object as the magnet would lose.

What I wanted to convey was that if you could get the mass closer to the object, you wouldn't need anywhere near 5.972 × 10^24 kg anymore. For instance, if you had a sphere whose center of mass was 1 m from your test object, and wanted it to exert the same force that the Earth exerts 6.7 x 10^6 meters from the object, you would only need your sphere to have a mass of 1.33 x 10^11 kg.

If we were using a sphere small enough that we could get its center of mass 10 mm from our test object's center of mass, it would only a mass of 1.33 x 10^7 kg.

So, the magnet is still winning in the sense that it still takes a lot of mass to counter the magnet, but we're talking 10^7 kg, not 10^25 kg.

>but we're talking 10^7 kg, not 10^25 kg

which, let's not forget, is still 10,000,000 kg!

But it gets worse. The thing I quoted, which generates 150 lbs of magnetism -- only weighs 0.3 ounces!! (0.01875 lbs or 0.0085 kg).

were your figures for 150 lbs or for 0.018 lbs of gravitational force from a point source? :)

I wonder if something using hydrogen as the buffer would work - it has a density of 37 kWh/kg (or 2.4 kWh/litre).

Instead of fuel cells (which are still expensive and relatively low capacity), using excess electricity to electrolyse water.

The oxygen and hydrogen can be stored in tanks, and when you need the power it can be combusted, generate steam and drive turbines.

There isn't really any new technology here, so it should be relatively cheap and quick to setup too. What am I missing?

The volumetric energy density of hydrogen is very low, so either you need enormous tanks or very strong tanks plus extreme compression. One way of making enormous "tanks" that can tolerate fairly high pressures is by solution-mining caverns out of natural underground salt formations. But this is geographically limited to where such salt formations occur, like parts of Texas and Louisiana.

The other problem is that the round-trip energy efficiency isn't great. You might expect 75% efficiency for electricity-to-hydrogen from electrolysis and then 60% hydrogen-to-electricity if you burn it in an efficient combined cycle gas turbine for 45% efficiency overall. Oh, and you irreversibly lose energy as heat when you compress the hydrogen, for further losses. And it's pretty expensive to buy big industrial-scale electrolysis units but run them only during the few hours each day when there's excess renewable energy you want to store (though it becomes a more attractive idea as the penetration rate of renewables, and corresponding excess generation times, increase.)

The conversion factor is not that important because without storage the excess energy would be lost anyway if we assume electricity demand doesn't shift from night to day.

The lead-acid battery in my car holds 660Wh (12V, 55Ah) and costs less than $100. That puts lead acid at a quarter the price of the Tesla powerwall (admittedly without an inverter).

If you tried to use it like a PowerWall it would fail before it had reached 1/4 of the PowerWall's rated cycle life. Lead-acid is not a good chemistry for grid tied storage.

If it's cheap to store the energy, and you can store it at large scale, the conversion factor starts to look less important.

If anyone had looked at the amount of energy that goes into an ICE car versus what makes it to the rear wheels, would we have autos today?

Yeah, especially with the front drive ones!


Put the powerwall on the train as well and use both :)

What about a global power grid? Whichever side is receiving sunlight would power the other side of the Earth. We've already got communication cables running worldwide, couldn't we add cables for transmitting power? Perhaps, electric cables will weigh orders of magnitude than optic fiber comm links, but would it be infeasible?

You can't transport power that far without massive transmission losses. Communications cables use repeaters and fiber optics, so they don't have nearly the same problem.


There's also the political issues to consider. You'd have to really, really trust the other countries you're relying on for the supply of electricity not to do something to mess with your economy. I could imagine it working politically in an area like the EU[1] or between the US and Canada, but across Asia or the Middle East there's likely to be too little trust.

An example with pipelines, Russia turned off gas pipelines into Ukraine in 2009 in winter: https://en.wikipedia.org/wiki/2009_Russia%E2%80%93Ukraine_ga...

[1] http://europa.eu/rapid/press-release_MEMO-15-4486_en.htm

Because they're very expensive and require liquid nitrogen cooling? Which itself represents energy consumption, so instead of losing energy as resistive losses you're spending it on cooling.

What's more likely is some form of energy-to-liquid-fuels process that runs during times and places of excess energy. The fuel can then be stored, pumped, trucked etc to other locations. This already exists, it's just not price-competitive with fossil fuels.

Cooling benefits from insulation - the energy loss does not scale as an exponential factor, and levels off very dramatically (the deep sea is already very cold).

We're trying to avoid the deep sea getting hot, remember?

What about electrolysis of water into hydrogen and oxygen during the day?

Wow, that was an amazing article. I learned a lot reading it. Thanks.

One way to deal with night time electricity demands is to increase the cost of it. This will push demand to the daytime.

There's no reason to have uniform power rates 24 hours when the cost of providing electricity will vary dramatically from day to night.

In my country (Poland, so not much renewable energy unfortunately) night-time (and weekend) rates are significantly lower (in some cases over 50% lower). You may choose uniform rate but it's not worth it (just remember to do your washing at night).

I worked in a company that sold energy-trading systems for many countries in Europe, and it was similar there (usually individual customers may choose uniform average rate, or much better offpeak rate and slightly worse peak rate, while the big energy market participants trade with different peak and off-peak rates almost always).

Is this a real issue? My understanding was that there is a significantly lower demand for electricity at night. This is reflected in my energy tariff at least (daytime ~3x cost per kWh over night time).

Traditionally, the oversupply due to keeping coal and nuclear running at their optimum is available at night, when the demand is lower.

Many locations, like yours, already account for this by charging lower prices at night to incentivize people to shift their load to these times e.g. putting the dishwasher on a timer so it runs overnight. Many pumped hydro stations store energy at this time too, often they were built specifically to pair with nuclear stations to better map a steady supply to a variable demand.

In areas that are ideal for solar, and roll out solar heavily, they may end up in the opposite situation, with the supply and demand mismatch reversed and so try to push usage to the middle of the day by dropping the prices then. Since most usage is during the day due to air-con load and the fact that people sleep at night, this may mean lower prices overall.

Once you add the internet into the equation, rather than the blunt instrument of a block of hours at a lower charge you can do things like have specific plugs for electric cars or an irrigation pump on a farm that can respond on a second by second basis to over or undersupply of electricity by using more or less in exchange for a cheaper rate. This is called "demand response" and while it happens already with big industrial users, it's getting cheaper and easier for it to be used by more and more users. One neat product that already exists is an internet connected car charger that checks the carbon content of the grid and charges your car overnight only when the wind energy is a big part of the mix.

Electric rates are the same around here 24/7. If it's more expensive to provide baseline load at night, increase the night rates.

Elastic uses of power, like charging car batteries, running the dryer, heating the water heater, etc., can be shifted to when power is cheaper.

I.e. reducing demand for night power is a cheap and effective means of dealing with the sun not shining at night.

Okay, so we ask people not to charge their EVs during the day and now we won't want them to charge them at night? I know EV penetration isn't high but many utilities have good rates at night!

The best way to fix demand is to get all usage to be as efficient as possible, not restrict its use

Potentially relevant, with much less moving parts: A bigger kind of "Hubspeicherkraftwerk" (not sure about the english name, wikipedia does not offer one ;)) - carve out a large cylindrical block of a mountain and pump water below.


(In German, though the pictures should be enough)

What a fun system. I've added it to my notebook with the German 'hydraulic rock' (http://heindl-energy.com/) system. The latter doesn't need a hill.

What I like about the train system is that it should be cost effective and can be built without water. Cost effective because steel and concrete seem to be the bulk of the cost. The lack of water makes it more useful in places like the desert where water tends to evaporate over time and is hard to come by. It would interesting to understand the maintenance burden relative to the renewable energy source it is paired with.

Another, less talked about, issue is replacing peaker plants. Here is an article from 2015 on why that is a good idea (http://www.aiche.org/chenected/2015/04/battery-storage-takes...) using batteries. Even a "small" ARES system it would seem could address some of this need. The challenge being its hard to build something like that in an urban area.

Wow, I thought I had seen it all when it comes to energy storage, but this really takes the cake though. That piston is enormous! It might be the largest moving machine ever constructed if it was built.

I agree with you on solar. From my perspective, any energy we don't capture and harvest from the Sun as it hits Earth is wasted. Now allow me to offer you a different view on my buddy; Carbon. Carbon itself is a commodity and should be handled as such. We need it to make graphene, synthetic diamonds and carbon nanotubes which will in turn become the new silicon for the next decade or so.

What if all those things can be as accessible as what is common today? We must harvest every last molecule of carbon for this end. That include the crap we have overloaded the atmosphere with. I've already come up with a system for this that needs some fleshing out. With it we can make better solar panels, water filters, textiles of immense tensile strength, relativistic electrical conductivity and insulation.

Carbon is the fourth most common element in the universe or at least from what we can see. We must keep it that way if we want complete proliferation of this next generation of desktop and mobile quantum computers in every part of human society. So we need to capture and reclaim it from every corner of the Earth because the end goal is totally worth it.

Why not fill the rail cars with batteries as the ballast?

We should tax carbon emissions but we shouldn't subsidize storage. Anyone who developrs efficient storage will be able to make money on the energy markets directly (buy low and sell high). Let solar+storage compete on a level playing field with zero-carbon sources that can offer baseline power without requiring storage (hydro, nuclear, geothermal...)

Many governments (state and local) are in a position where taxing carbon isn't a possibility but subsidizing low-carbon forms of generation is.

They should not be handicapped.

They shouldn't be subsidizing one form of low-carbon generation but not another. That will distort the market and lead to buying subsidized forms that are more expensive overall.

I agree in principle, but given the huge proportion of energy which comes from fossil fuels, a carbon tax would have to be huge (and hence unpopular) if it were to match the market competitiveness of subsidies.

That is, to subsidize solar 33% (our current subsidy in the US), you would have to make a carbon tax equivalent to 50% of the current price of fossil fuels, or about $1.50/gal. This is obviously politically untenable in the United States.

The happy medium is probably to do a bit of both, and work toward the middle. Smart, targeted subsidies for carbon-neutral sources, and small, gradually increasing carbon taxes.

You can get away with a higher carbon tax if you return the money to the population.

In fact, there's a fair amount of popular support for basic income. Meanwhile a leading idea for carbon pricing is a flat fee per ton with the money returned to the population, equal amount per capita, which essentially is a small basic income funded by the carbon tax.

British Columbia has a revenue-neutral carbon tax (in their case by reducing payroll taxes accordingly) and reportedly it's popular. The opposition party ran against it and lost.

Washington State rejected one, unfortunately, one significant downside of the state solution is that some demand can just easily shift to another location. I wonder if there's a scheme where on state will pass it but only have it come into affect if neighboring states do as well.

A few states' actions wouldn't help much anyway. The real benefit of state carbon pricing was that the fossil industry was starting to come around on a national carbon price, because they'd rather deal with that than a bunch of different state systems.

Between the WA defeat and Trump win, that's probably pushed back a few years.

What about Cryogenic storage? http://www.bbc.co.uk/news/science-environment-37902773

I haven't looked into it much, but from the little I've seen it looks promising.

Build a proper electricity grid and sell power from sunlit areas to dark areas. You need the grid anyway. At the same time insist that all new dwellings are built to passivhus standard. Also arrange for all plugged in electric vehicles to be able to supply power to reduce peak demand from the net.

All this is scaleable and mostly going to happen to some degree anyway and avoids the need for massive capital projects that are susceptible to corruption.

Sounds like a good idea, but i do worry about how much overcapacity in energy storage would be required to handle the worst case scenario (a significant, extended drop in wind and solar output).

You can use natural gas for backup on demand energy generation. It isn't all or nothing.

And who pays for the backup on demand energy generation when it isn't in use?

The same people who pay for it now, the consumers.

Electricity grids have massive 'problems' with balancing themselves as demand varies throughout the day. Currently there are e.g. open cycle gas turbines on standby - they are paid a retainer in return for always being available, plus are price per MWh when they are generating. For the UK you can see them http://www.gridwatch.templar.co.uk, also see https://en.wikipedia.org/wiki/National_Grid_Reserve_Service

That works fine for a daily cycle. The backup plants still get paid every day.

But will it work for a yearly cycle, where solar/wind output varies from year to year?

Can a backup plant survive a year or two when it isn't needed and therefore isn't paid?

The problem is high start up cost and high real estate cost. Batteries are small and you can purchase and deploy them in small quantities.

This is a really great mini-documentary about solar from May this year. They talk about the tumbling price and interview investors and manufacturers in China. The eye opening moment for me was the fact that they already have 2M people in China manufacturing panels, and production is still ramping up.

It looks like solar is starting to become an inevitability.


That's a great documentary for data nerds and quant-types -- it's basically a bloomberg interview with the market analysts and the manufacturing plant operators talking about why the trends are exciting and how the finance markets are starting to see real return in this field. Stuff like experience curves (https://en.wikipedia.org/wiki/Experience_curve_effects) and cost parity.

My favorite was the guy who noted that in resource-based industries, things get more expensive the more you use; solar and wind are manufacturing-based industries where the more you use, the less expensive things get because you get better at making it.

Great documentary, coolest part is the small town that plans to have 200 cars that can connect to their local grid to charge themselves or even cooler to give power back to the grid at night if the car does not need it!

Is using your car's battery for grid storage a good idea? You're shortening its life, which you've paid a premium for to put in a mobile format. I suppose you'd have to figure out how much a charge/discharge cycle is worth and that will tell you how much you'd need to be paid to make it economic.


"The single thing that will shift people off high carbon energy into lower carbon energy is going to be the price...When that situation is broken and free market decides to go for solar it will break like a dam is broken."

The dam just broke.

... until import tariffs with China end the solar revolution here.

Lots of places around the world not the US. Those buyers will continue to drag the price of solar down.

The US election doesn't mean a thing for the energy market. We'll be dragged along with the rest of the world by market forces.

Furthermore, there is a world market for other energy sources, which are all somewhat fungible. So if solar takes off around the world, fossil fuel demand will slack and prices will continue to drop making their extraction less viable.

Hopefully we also start to see some significant replacement of fossil fuels used for transportation.

We need to use solar to pay for the extraction and transportation of fossil fuels.

LOL, very good point. Maybe for the back of the classroom you could clarify?

It is clear that the calculation or real EROEI is very hard when you start counting all externalities and dependencies.

Do you have any good sources of anylsis on this that you could recommend?

What do those terms mean?

I don't get that. Could you elaborate?

I'll answer the implied and straight up admit that I was pissed off by the presidential election in the USA this year.

But the doom and gloom about it is almost as bad. The great thing about the USA is that in 4 years, if this guy sucks, we can change our minds again and get back to business.

This too shall pass.

In principle, yes. But...

...that assumes elections work, which is becoming more and more dubious. Remember that he actually lost this election by popular vote, just like Congress would be in Democratic hands if it weren't for gerrymandering. And I don't think it's entirely unreasonable to assume that this president will be even more brazen about undermining the election process than anything that's come before.

The popular vote does not matter. 1-2000000 of the popular votes could have just moved from NY, SF and LA to the rust belt and this would have made all the difference.

The big corporations with democratic leaning CEO could have moved investment and jobs there.

You could have moved a lot of federal jobs there too.

And with them will come the people that will guarantee better demographic distribution.

The US election system is on the stupider side IMO, but a candidate cannot lose the popular vote, because such thing as popular vote does not exist.

The big corporations with democratic leaning CEO could have moved investment and jobs there.

If a big corporation with a democratic leaning CEO moved to a state where the population was mostly republican, then the company would have a workforce that is mostly republican.

Few people align themselves with the company's CEO's values, and even if the company was made up of mostly democratic workers, few people will follow their company to another state. If my company changed states, I'd get a job with a new company, my life just isn't that portable that I can pick up and leave with ease.

There are already large tariffs between China/US and China/EU for solar gear:


But then it will be cost efficient to mine coal, and America will be great again!

Not really US electricity demand is going down in the first place. And the major energy demand is going to come from Africa and Asia. On avg 2-3 billion people in Asia and Africa use less than 5% of electricity per person compared to the US and Europe. When you consider the fact that is almost 5 times the population of US and europe combined.

So the solar revolution is more important for the rest of the world as they will be the major users not the US.

There is still the rest of the world...

Good thing Tesla is going to manufacture its panels in the US.

Trump is going to open those coal mines anyway, so don't worry. US is unlikely to lead the new energy revolution in, at least the next 4 years, I guess.

Trump is going to cause US fossil fuels production to skyrocket. One consequence will be oil prices falling much lower than they already are. And that will make his good friend Putin very unhappy. I wonder if Trump has thought that out.

The good thing about solar is that it leverages off the semiconductor industry, whereas other renewables do not. For example, in the case of smartphones, before they existed accelerometers and GPS IC's were crazy expensive, but now they are cheap. Solar probably has years and years to go before it is as cheap as possible. I would not be surprised if solar hits $0.10/watt in 10 years. Basically because as the demand picks up and money is being made, more companies will be saying, "if we can just increase our efficiency by 20% or so, we can lower the price and win huge contracts.." so they will be trying things like multilayers, changing the structure to improve efficiency as the panel heats up, coatings, lenses, etc that can all be done in low cost ways. Definitely would be interesting to compare how much it costs to run/maintain a solar farm vs. a natural gas power plant. Basically you have to deliver the fuel to the plant, maintain all the stuff, pay for the workers to make sure stuff doesn't blow up, etc. Seems like with a solar farm you basically need a guy with a truck and a leaf blower.

The big difference between solar PV and microelectronics is that there's very little miniaturization achievable in PV. More power requires more area; light conversion efficiency increases only slowly over time. The sunlight-to-electricity conversion efficiency of common solar cells has not even doubled since 1980. There's less than a factor of 4 left between the efficiency of panels you can buy for your rooftop today and the maximum efficiency permitted by the laws of nature. That's quite a difference from the past trajectory of microelectronics, where Moore's Law doubled efficiency over and over.

That's also the short explanation for why companies with deep experience in microelectronics manufacturing, like Intel, TSMC, or Texas Instruments, did not come to dominate solar PV manufacturing. PV is a very different game. It's about producing huge quantities of doped silicon wafers with little patterning across their surfaces.

However, it is true that PV has made enormous strides in reducing manufacturing costs, though aggressive miniaturization was not how it was done. In sunny regions PV now has the lowest O&M costs of any new-build electricity source.

interesting.. would be interesting to listen to your explanation while listening to some of your music mr. glass. :-)

:-) Philip K Dick meets Philip Glass: I combined two of my favorite Philips for my nom de plume.

Are you in the renewable energy industry in the US? If yes, send me a message on the e-mail in my profile please.

> Seems like with a solar farm you basically need a guy with a truck and a leaf blower.

You don't even need that! Automated, water-free robots can already clean panels, even in dusty desert areas: https://www.revolvesolar.com/5-robots-that-are-revolutionizi...

The really interesting thing here is what this means for all the coal-warriors making their resurgence (or thinking they will): they're not going to find investors for new plants.

If solar is cheaper, and scales well (i.e. you can just keep deploying it, pretty much anywhere, and have it get cheaper the more you do) then all the smart money is going to go to building as much PV as quickly as possible. There'll be no one willing to invest in coal-plants because they'll be looking at the on-going costs, looking at the up-front costs, placement issues, build-times, risk of actual action on carbon pricing and saying "you know what, let's build out solar instead".

This scenario is already playing out and it's largely due to market forces. More people are employed in the solar industry now than are employed by the coal industry. Simply put, coal in the US is dead because it can't compete with gas, solar, and wind. That is why it is silly when people say we need to prop up the coal industry. We would much better off by just training those workers to install solar panels or become wind turbine technicians because there is actually a growing market there.

Coal working jobs are more dangerous but pay much more than solar panel installation. The issue the coal workers and uneducated middle class have faced recently is a loss in earning potential (with existing skill set) moreso than absolute unemployment.

The internet suggests the average hourly pay is $23 for coal to $16 for solar installers. That's a big hair cut, plus solar installers likely have to live in or closer to cities, generally increasing cost of living and providing one more obstacle that makes it hard to switch careers.

... and solar panel installation (and maintenance) cannot be outsourced.

Solar and wind are so different that comparing them doesn't really make any sense. Generally a combination is optimal. The news in this article though is that solar prices have tumbled in the last few years - whereas wind prices have declined at a more expected rate.

Comparing the photo-voltaic capacity installed in 2016 with wind capacity is a bit misleading, as wind typically has a much higher capacity factor than solar - so the 59GW of wind will almost certainly produce more electricity than the 70GW of solar.

Utility scale solar projects, the majority of which in the US now use single axis tracking, attain significantly higher capacity factors than rooftop installations. There's still a wind advantage but I wouldn't say its CF is "much" higher. The EIA measured the capacity factors for US utility scale wind and PV respectively at 32.2% and 25.8% in 2015: https://www.eia.gov/electricity/monthly/epm_table_grapher.cf...

Wind still has a capacity factor advantage, but at 59 GW and 70 GW, multiplying by those above capacity factors, wind and large scale PV would be nearly matched for real output (18.998 real GW and 18.06 real GW respectively). Somewhere like Germany the real-output gap between wind and solar would be greater while in India I'd expect it to be lesser, due to relatively poor wind resources. (In fact single-axis PV might actually beat wind on capacity factor in India.)

Keep in mind wind CF's are still increasing as the world moves more to offshore wind farms with taller turbine heights.

See for example the new Burbo Bank extension off Liverpool. Dong Energy tweeted me this when asked about their Vestas CF's there: https://twitter.com/DONGEnergy/status/775692082497351680

"In general it is fair to assume a capacity factor of app. 50% for new wind farms at good sites."

One of my favorite charts:


For the 4.5GW of wind installations that came online in 2014, the average capacity factor is above 40%. For 1/6 of the projects, they're above 45%. When we go to 140+ meter hub heights, NREL expects the best projects to have capacity factors of 60-65%:


Nuclear obviously kills that with 92+% factors, but coal and natural gas plants are right there near 60%.. we're really on the precipice in the best possible way.

That's a good point too, and the happy surprise of recent low prices for new offshore wind projects indicates that wind will remain competitive in many scenarios.

> 59GW of wind will almost certainly produce more electricity than the 70GW of solar.

What do you mean? These are not some peak figures, they are expected energy production figures.

Capacity factor means "take the amount of MWh actually produced by the plant, divide by the number of MWh that theoretically could be produced if it generated its maximum output 24/7/365." 70GW of solar refers to the maximum instantaneous production (note that it is measured in GW not GWh). So while in full sun, all of that solar is generating 70GW of power per unit time but at night, it generates 0W. Average that over a year and it generated about 28% of it's maximum power over time.

Hydro plants have huge ranges of capacity factors depending on other restrictions on the dam like if it is used for flood mitigation, if it is operated as a peaker plant, and if it snowed a lot the past winter. Nuclear Plants generally have capacity factors in the 90% range to account for swapping out fuel periodically. Fossil fuel plants have capacity factors of >90% unless they are used as peaker plants.

pilom's answer was good, but a good metaphor is to compare light bulbs;

Does a 100W light bulb use more energy than a 75W light bulb? It depends on how long each bulb is turned on.

The same relationship exists between energy sources. Does 60GW of wind power create more energy than 70GW of solar power? It depends on how much the wind blows and how long the sun shines.

If you divide the actual energy production by the ideal amount of energy production (aka if the wind blew steadily every minute of every day), that will give you your capacity factor.

If perovskite pans out -- and it is starting to look like it will -- then solar is about to get a lot cheaper still.

See https://www.sciencedaily.com/releases/2016/10/161020142037.h...

Just couple of days I saw this documentary by national geographic about world's largest solar power plant in India completed in ~8 months.


i am really happy that the "green" energy production is in such a price war. It's Solar vs. Wind not Solar vs. Coal. I don't care what's cheaper as long as it's not Coal!

There is sadly a fix for that, which is to lift carbon tax and relax requirements for having carbon sequestration systems - and coal would be quite cheap again.

We don't have a carbon tax in the US and we don't have requirements for carbon sequestration systems. Yet, the coal industry is still contracting. Now you could remove the regulations that require limiting emissions of things like mercury and NOX and SOX and it would make coal cheaper. But people really don't like acid rain or mercury in their water. Coal is already phasing out and it's because natural gas is cheap and renewables are getting cheaper every year.

Why would we ever do that?

Remember that the untaxed price of coal isn't a real price for it. Without a tax the market does not take the externality of its environmental effect into an account.

Economically speaking, a high tax on coal/other fossil fuels is simply forcing the market to consider its full and actual price.

I'm very confused by this conversation. What taxes are you guys talking about? They don't exist as far as I'm aware. I am aware of some subsidies, but not taxes on specifically carbon/coal/etc.

I'd guess he's talking about this: http://ec.europa.eu/clima/policies/ets_en

P.S. I agree about the comment regarding negative externalities not being accounted for unless something like this is in place.

> Why would we ever do that?

I don't know, but there is soon to be a new president who says he wants to support the coal industry.

Where are these carbon taxes? The only one I was aware of was in Australia, but that got repealed.

Do solar panels produce more energy that the amount required to create and operate them? (including mining the materials)

Yes: http://www.mdpi.com/1996-1073/9/8/622/htm

Energy pay-back time (EPBT) results for fixed-tilt ground mounted installations range from 0.5 years for CdTe PV at high-irradiation (2300 kWh/(m^2·yr)) to 2.8 years for sc-Si PV at low-irradiation (1000 kWh/(m^2·yr)), with corresponding quality-adjusted energy return on investment (EROIPE-eq) values ranging from over 60 to ~10.

If you want to see other publications on the same topic, try hitting Google Scholar with search terms "photovoltaic" "life cycle" "energy payback". For the most accurate results, restrict your search to recent years. Manufacturing processes are revised rapidly and studies from the turn of the millennium are now badly obsolete. (Though even in 1990 PV systems were net energy positive on average; the worry that PV system manufacturing consumes more energy than the lifetime output appears to be a holdover from the very early (1970s) days of terrestrial PV systems.)

I'm skeptical of your numbers. With such a quick payback time you'd expect to see solar power going up like crazy. (Like how wind was being installed quite fast for a while - although only in the best spots.)

Except you don't, you just see limited installations here and there.

If solar power has such a quick payback period why aren't more investment groups funding it?

Are the panels not available in sufficient quantity?

Solar power is going up like crazy. The US just had its highest-ever quarterly solar installation rate, up 191% over Q3 2015: https://www.greentechmedia.com/articles/read/in-its-largest-...

Global installations for 2016 are also on track to set a new record, surpassing 2015 installations by 48%: https://cleantechnica.com/2016/11/30/global-solar-installati...

And despite these records, the manufacturing side of the industry is currently demand-constrained rather than supply-constrained.

When you say you there's just "limited installations here and there," are you estimating from whatever you personally see in your local area?

I have to echo the statement that PV is going up like crazy -- at least where I live. Admittedly I'm in one of the sunniest places in Japan (and at 34 degrees latitude to boot). They are clear cutting forest to put up installations everywhere. Virtually every single apartment building already has solar panels on top of them. Every house builder offers solar panels on new houses and while I've seen a couple of new houses without them, they are definitely the exception. Every electronics store sells solar panels and usually it is the thing that they are pushing hardest (weird to see huge solar panel displays next to the refigerators and TVs). I've done the math and with the price of electricity we pay here, it is a complete no brainer to get solar panels. If I lived in a house, I wouldn't even think twice.

Thats great to hear but Im not a fan of the idea of clear cutting forests in place of solar.

Energy payback time is a different concept than money payback time. Investors are interested in the latter, First one is a rudimentary requirement for long term feasibility without subsidies.

Also, global solar is going up like crazy. Global sustained growth rate of around 40% is crazy fast. If that keeps on for a bit more than ten years, 100% of current global electricity demand is covered by solar. Just for comparison, there exists no two subsequent full years of iPhone production where Apple would have managed to achieve 100% growth.

The main problem with solar is that it is not a full replacement for current power plants, since it does not produce power at all times. Until we have the capacity to store energy to use at night and on cloudy days, solar will be limited in potential.

I think his numbers didn't take into consideration the expensive backup power plant that you have to have ready even if you have adequate solar capacity.

The term is EROI (energy returned on energy invested), for renewables its currently roughly equivalent to fossil fuels and getting higher as the technology improves. As oil and natural gas get more expensive and energy intensive to extract, their EROI is going down.

Yep. Many forms of petroleum are also very convenient for energy storage. So even an EROI less than 1 they would still have utility.

1 L of diesel has about 39 MJ of energy. That's about 10 kWH (a Tesla Powerwall) in the size of a Nalgene bottle.

Most definitely, chemical energy storage is fantastically dense.

I think that as intermittent renewable energy sources get cheaper and cheaper, in addition to time-shifting of arbitrage from batteries, energy storage in chemical bonds may make sense. Methods to create synthetic fuels from electricity are in their very early days, and all of them are terribly inefficient; most go through hydrogen and that step alone results in a huge loss of energy.

So for applications where energy density is needed, e.g. jets, the fuel costs will just be that X% higher than being able to use straight electricity + battery.

I can definitely imagine a world where creating synthetic fuels from excess grid energy is cheaper than fossil fuel extraction, but it involves tons of research and development in those synthetic fuel methods, and many decades of improving renewable technologies at their current rate.

From what I understand, for electric generation storage, at present the technology with the best promise seems to be flow batteries.

Going down from a fantastically high return to a slightly less fantastically high return.

Yes, but it's going to continue to go down. According to Wikipedia [1]

  EROI    Fuel
  ----    ----
  35.0	Oil imports 1990
  18.0	Oil imports 2005
  12.0	Oil imports 2007  
   6.8  Photovoltaic (wikipedia)
  14.4  Photovoltaic [2] (from recent analysis of post-2008 data)
The trend lines are clear and stark. The technologies for photovoltaic, wind turbines, and for fossil fuel extraction are all changing. It's certainly not the case that fossil fuels will disappear, but over the coming decades the sector will shrink dramatically due to market forces; investments will plummet.

And once you start looking at the slow and steady pace of improvement in storage technologies, it looks highly possible that around 2025-2030, only specialized energy applications ever make new investments in fossil fuel based technologies. And that's without accounting for the negative externalities that fossil fuel users (all of us) make everybody else subsidize. Existing infrastructure will continue as long as its economical, of course, but there's most definitely a huge turning point around the corner in energy technology.

[1] https://en.wikipedia.org/wiki/Energy_returned_on_energy_inve...

[2] https://www.researchgate.net/publication/311170056_Solar-PV_...

I believe those EROI numbers also don't include converting the energy into the useful output that you need and so are a a pointless apple-to-oranges comparison

PV output is electricity already, to get electricity from coal you'd need to throw away 2/3rds of the energy as heat, about half for gas, reducing their EROI by that factor in a correct comparison. So if it's electricity you want then PV is already winning over fossil fuels I think.

Similarly, for powering cars with gasoline they're not as efficient in converting that energy "well-to-wheel" as EVs, as the losses in battery storage and electric motors are much smaller, giving you a "miles driven return on energy input" measure about 3x higher for a PV charged car vs ICE.

You can avoid the generation loss on natural gas by piping to the home and burning it directly for heat in areas that need heat, getting close to 100% efficiency with modern central heating systems, but then in those same situations you can use an electric heat pump and get 3-500% efficiency, since you're only moving the heat around, not generating it.

Don't heat pumps only work in temperate climates though? The low in Denver tomorrow is going to be 7 below...

At very low temperatures an air source heat pump does not have an advantage over simple resistance heating. But -7 F is not too cold for leading edge ASHP technology (or rather, this was about the leading edge 5 years ago): http://www.franklinenergy.com/wp-content/uploads/2014/10/Air...

"Independent research has verified the ability of air source heat pumps to maintain energy efficiency well above other electric heating systems, with coefficients of performance (COP) of between 2 to 3, in temperatures as low as -15⁰ F. Multiple manufacturers have developed ASHPs designed for cold climate operation, incorporating features such as two-stage compressors and advanced defrost capabilities.

Where data is available to evaluate the energy savings impacts of [Western Minnesota] customers moving from less efficient electric heating systems, the energy savings associated with the heating demand served by the hybrid ASHP systems is found to be in the range from 10% to more than 40%, with median savings of around 22%."

It looks like the Western Minnesota region where they ran this trial has colder winters than Denver, so Denver-area results would be even better. The best air source heat pumps should be able to save energy over resistance heating for the vast majority of American households. They will also save energy over direct combustion of natural gas in a household furnace if the natural gas is instead burned in a combined cycle electrical plant. They will not reduce energy consumption or emissions vs. a home natural gas furnace if your winter-time grid electricity mix has a substantial coal component.

for climates with colder winters, you can get high efficiency with ground heat pumps instead. It costs more to install since you have to drill a borehole, but the efficiency doesn't drop with colder weather.

In doing a quick google search on this question, it seems to be "YES". Payback is in the 4 year range where the service life of the panel is in the 20-plus year range.

Was also discussed (more in terms of carbon footprint) in The Economist this week:


Don't forget the costs to dispose of them once they are not producing energy (or enough energy), typically after 15 to 20 years, and the overcosts for replacements due to "natural" events, like hailstorms.

There could actually be an interesting secondary market for solar cell recycling. My understanding is that the primary degredation is due to the contacts, which could likely be reapplied or annealed in order to reach their original efficiency.

No one ever really likes to provide these numbers. Hopefully they are a net gain, but I'm a bit skeptical

This does not include storage costs to offset the cyclic nature of solar.

For smallish amounts of solar, you need stabilization but not storage -- when it's sunny, there's more air conditioning usage.

For large amounts of solar, sure, there's a lot of work to be done on storage.

Sun doesn't shine for ~50% time everyday.

Solar and wind make up 2-3% of energy produced in the US. We can still make huge improvements just doubling a few times to bring it to 33% before talking about a 100% renewable grid.

(Though some people have said that a 100% renewable grid is still achievable within reasonable economic conditions.)

That would only be relevant if demand curves had a duty cycle of 50%. As OP mentions, in the warm parts of the world demand peaks during sunny times (thanks to AC), so there is good alignment.

Yes but peak demand is 9PM in the evening. Wind is down at that time as well.


Solar+wind don't have to provide 100% of the mix. They has to reduce the daytime fossil fuel generation.

That doesn't line up with the CalISO graphs, for today, the peak is around 6pm - 7pm (still late in the day for Solar, but solar does ramp up in time to help with the daytime peak at around 8:30am):


But if you look at the August 1st graph here, the summer peak is around 4pm:


In any case, the peak solar output is around 4600MW while overall demand peaks at 29000MW, so solar still has a ways to go before we have to figure out what to do with all of the peak power.

Any idea why wind is up or down on a particular time? I would've thought it's extremely random.

No, it's definitely not random.

Off shore you get 24 hour constant wind, the synoptic wind. When we race to Hawaii, first you HAVE to get past the Farallones the first day and into the synoptic wind. If the inshore wind shuts down, you're bobbing all night with 0 wind. Get to the synoptic wind, set the spinnaker and it's a downhill sled race to Oahu.

At the coast, you get marine layer driven winds in the afternoon, dying off in the evening. Stockton heats up midday, creates a low pressure and sucks in the marine layer. It cools off in the evening, shutting it down.

Now, as to the wind chart, that's CA as a whole. Dunno. But wind is definitely not random by geography and season. Basically, it's solar driven.

Demand for electricity is not at peak for 90% of the day.

This does not include externality costs to offset the harmful emissions of coal and the environmental consequences of mining coal.

Thank god, this can't come soon enough.

Saying that solar is less expensive than nat gas is misleading. It's and apples to oranges comparison. To make a valid comparison you need to compare solar + batteries to natural gas. There is value to electricity production that has an on/off switch because it helps keep the grid stable.

At modest levels of penetration it doesn't matter much that solar is non-dispatchable. There's already enough spare capacity in the system that you don't need to build backup specifically for solar. Consider, for example, the tens-of-gigawatts CAISO demand gap between 3:00 AM in April and 6:00 PM on the hottest day of August. The system already had to be ready for the peak-demand situation. Adding solar to the grid doesn't require new dispatchable backup capacity until penetration rises significantly. The backup capacity that was there before solar is a sunk cost.

You might as well say that calling nat gas cheaper than solar is misleading because a valid comparison would require carbon capture and storage on the fossil plants.

Solar did not help meet peak demand today in California: http://www.caiso.com/Pages/TodaysOutlook.aspx Nor does solar help meet peak demand for 2/3rds of the year because peak demand comes after the sun goes down.

Solar can help during some summer days, but for much of the year it makes the system less efficient by requiring more ramping up and down of other generation resources.

Even in summer in California solar peaks mid-day while demand peaks in early evening. Nevertheless, it appears that the additional costs and efficiency loss from cycling fossil plants is dwarfed by the fuel-consumption savings, at least up to modest (33%) penetration of solar/wind: https://www1.eere.energy.gov/wind/pdfs/55588.pdf

In this study, we found that up to 33% of wind and solar energy penetration increases annual cycling costs by $35–$157 million in the West. From the perspective of the average fossil-fueled plant, 33% wind and solar penetration causes cycling costs to increase by $0.47–$1.28/MWh, compared to total fuel and variable operations and maintenance (VOM) costs of $27–$28/MWh. The impact of 33% wind and solar penetration on system operations is to increase cycling costs but also to displace annual fuel costs by approximately $7 billion.

The modeling assumptions in that study look pretty reasonable to me. The one significant change I'd make is plugging in a lower natural gas price, e.g. the $3.66/MMBtu from Henry Hub a week ago, compared to $4.60 assumed in the study. Even with that change it looks like fuel savings dwarf cycling costs.

This is such a frustrating misconception. The annual peak is far more important than the daily peak for capacity planning.

There are two capacity issues being confounded here:

(1) Generation capacity. The idea here is that if solar doesn't generate during peak loads, then building solar will not displace power plant capacity. But solar does generate during peak loads, so building solar does reduce the number of natural gas plants you need. At higher levels of PV penetration this will be a concern, but we're not there yet.

(2) Ramping capacity. The idea here is that solar generation is correlated and non-dispatchable, so dispatchable generation must be able to match the swings in solar output. Again, this will be a concern at higher penetrations, but right now the grid can handle it.

Philipkglass's replies are excellent.

You've already got a couple of good replies, but I just wanted to be clear about how far you've missed the point of what you're replying to.

His point was that the absolute peak of the system is in the afternoon during the summer. Which is when solar is most effective, so it shaves this peak quite well. It's like profiling your application and finding the hotspot, that's the place you most want to fix. It'll not only generate some energy that's needed, it will save you building a whole extra plant that's only needed for a few hours a year, if it's distributed then it'll save you upgrading transmission for a peak that's only needed a few hours a year.

You talk about today, which is in mid December, which in places that use air-con means the peak load today is about 1/3 or so of the peak in the summer. California has more than enough power plants to supply the load in the winter even if all solar was unplugged suddenly. That was the whole point of what you replied to.

Solar peaked at 8% of total production today, which looks like it matches the comment you're replying to: "At modest levels of penetration..."

Once we get our Tesla Powerwall batteries, that should help with peak demand, no?

Only if they sell power back onto the grid when there's more demand than can be met. Which is a doable thing but we don't have the framework quite in place yet.

Localized storage on households and businesses can help even if it never feeds back to the grid, because it means that batteries can charge via off-peak electricity and switch to discharging to fulfill local demand during grid peak times. It has pretty much the same system effect as discharging into the grid during peak demand but it's more efficient to use the electricity at the same point it was stored.

It's not all that misleading when they actually take the time to make the same point and the article does this in at least three places.

There won't be many "batteries" when renewables are supplying most power. We have fairly cheap solutions for large-scale energy storage like hydroelectric with pumps.

You're far from correct. Large-scale energy storage is one of the biggest unsolved problems today, and there are numerous startups in the space. LightSail is an interesting one, and is using compressed air combined with micro-droplets of water to ensure stable, efficient energy transfer.

Renewables success is heavily tied to battery technology. The wind isn't always moving and the sun isn't always shining, so you need to store that energy. Right now with existing battery tech, most of the energy is lost in a storage process due to heat or battery wear/tear.

If energy storage was solved today, the oil/gas industry would be on their knees.

China is installing an 800MW-h vanadium flow battery:


The cost claimed for a small installation coming online in Washington is $0.05 per kilowatt hour. That might not be cheap enough to beat out fossil fuel generation, but it is cheap enough to run the world on.

If one runs the numbers, Pumped Heat Electrical Storage should work fine for time shifting energy at scale. None of the technology needed is at all novel or untested. All of the technologies needed scale up very well. Basically you store electric power as heat using motor/generators and axial flow compressors/turbines. Storage media is 'a big steel tank of hot rocks'. One can quibble but the end to end efficiency is going to be better than 60%. Could be as high as 80%.

Far as I understand the reason such doesn't exist is there is no current market need until you have excess peak solar or wind power.

It is incredibly difficult to site new pumped hydro in places like California today. It might work in other places, but it's almost impossible in California.

this source claims that it would not be feasible to rely only on pumped hydro, due to the scale. [1]

1. http://physics.ucsd.edu/do-the-math/2011/11/pump-up-the-stor...

That source does thing like assume

>I argued that we need 7 days of storage for it to be invisible to the end-user

That said even with more reasonable numbers pumped hydro storage isn't going to work because the energy density is too low.

I think roughly to store 1kwh your need to raise ~3.66 t water to 100m. Not very energy dense. However 1.8 kg of stone at 1000 degC will store about 1kwh of energy. That's a lot better.

What's curious is Tom Murphy studiously ignores thermal storage.

with grid tie-in your grid is the battery.

"The average monthly cost of living in the United States for a single adult with two children is $4,820. This adds up to an average annual cost of living of $57,851."

What would the cost of living fall to if energy costs were $0?

A decent rule of thumb is that the cost of residential electricity is half for transmission and distribution, half for generation. So save 50% on your home power bill if electricity became free to generate but still had to be delivered from centralized plants.

If household solar and small-scale storage get ridiculously cheap, grid defection may be cheaper than buying large-scale renewable output from the grid, but utility scale solar plants can generate more real output per nominal watt and cost less to build per nominal watt. Unless your local electricity system's history is weird (unneeded grid capacity built in Australia being paid for by households, Energiewende that shifted costs away from big industrial customers and toward individuals in Germany) then grid defection probably won't be the lowest-cost option in the next 5 years.

What you call grid defection might very well happen in currently developing countries though, just as they never adopted telephone lines or stable postal services.

In most developing countries the grid is the problem, not the production. They often lose more than 50% of the electricity before it gets O the point of use. They also need to make large scale coordinated upfront investments, which i difficult with weak governments.

In a way, solar + cheap storage could become an economic and social democratisation force.

On the topic of grid defection, in some areas where gas heating for homes is common, there is a move to stop hooking new houses up to the gas grid, instead having all heating and cooking done via electricity. That adds another interesting angle on these calculations, since not hooking up in the first place is obviously cheaper than hooking it up and then not using it.

I think you can reasonably estimate this simply. If you imagine that in any good or service you buy, there is a percentage that goes to salary and taxes and a percentage that goes to some physical thing. Now, assume that the physical thing's cost is 100% due to energy. It will be less than this, but even with this upper bound, it's not very high. For example, imagine a pair of shoes. You pay $100, but materials cost maybe $10. My guess is somewhere between 10-20%. My guess is that if energy costs were zero, the cost of living would still be at least $50K (in this stated scenario -- I actually have a hard time believing the cost of living is so high in the US).

It'll never cost $0, even if we could satisfy all the demand of energy with solar power, electric grids still involve costs, same as roads.

I think $0 cost in this context is near-zero. If my utily bill (electric + NG for heat) went from ~$150/mo to $15, I would think of as practically free.

Energy is free? I'm going to turn my (apartment)living room into a grow op for all my veggies. That should help reduce the cost of living some.

I'm not sure, but that will never happen unfortunately. The support infrastructure always comes with cost.

Perhaps the commentor is thinking rooftop solar.

My electric bill barely gets to $200 in the hottest summer months. So, it would not change by a lot.

Part of the cost of most things you buy goes towards the energy costs invested in making that thing. Lower energy costs would lower the prices of most things you buy.

Gov't and business will never let costs go to zero

Cost of living would go to zero, because material cost for stuff would be zero.

But you might wonder, what about salary? But salary would go to zero as well because you wouldn't need much money since everything is basically free.

The problem will be motivating people to do stuff.

You might expect an additional problem that certain items might be scarce, but the solar system is very large, and with free energy there will be no limit to getting things.

If you think I'm wrong, try this thought experiment - what are the costs of things? To get started ignore salary, and think what do you have to pay for?

Shipping: Free obviously.

Mining: Free since you don't have to spend energy getting stuff out of the ground,

Machines: Free, since they are made of metal and other resources that are in turn free.

All you are left with is assembly and design i.e. wages. But as mentioned, those will go to zero since everything will be so cheap people won't need much money.

(Note, I'm assuming unlimited energy. If you meant limited energy, but it was free, then you have a contradiction since if it's free no one will limit themself. Maybe you could do quotas.)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact