Hacker News new | past | comments | ask | show | jobs | submit login
The US can reach 90% clean electricity by 2035 without increasing consumer bills (berkeley.edu)
184 points by MilnerRoute 17 days ago | hide | past | favorite | 112 comments

Exciting. But here are some problems.

- Unfortunately, they are averaging the entire US, which is made of several grids, so the bad days are not that bad in their analysis. To fix those bad days, you'd need hella more storage, hella more capacity. e.g. 3-4x the renewables capacity than the normal need.

- Such a system has a very low resiliency because it depends on national grid. If Covid has taught us anything, it is the need for resiliency in the face of rare events. In this case unfavorable weather, broken transmission, etc.

- Cost projection is rosy... The panels, batteries, and storage are already very close to the material cost. Need some pretty fundamental tech transitions to improve further.

- If there is no policy change, nukes and nat gas will not be competitive 99% of the time because solar/wind are nearly free electricity and bid extremely low power prices that nuke/gas can't compete with. So even though they will provide the backbone and dispatchibility needed for on-demand power, they do not currently have the financial incentive to do so. We need resiliency and dispatchibility payments.

Report is Berkeley and GridLab, mostly policy wonks

A much better report from MIT: https://www.sciencedirect.com/science/article/abs/pii/S25424... or http://energy.mit.edu/research/utility-future-study/

Both are harder to read and give worse news, so probably won't be part of the policy discussion...

The report itself: http://www.2035report.com/wp-content/uploads/2020/06/2035-Re...

And the data explorer: https://www.2035report.com/data-explorer/?hsCtaTracking=aefa...

I think I came out a little strong and biased here. My apologies. Some more thoughts on this article.

- They group nukes in the clean %, so today we're at 37% clean already. not bad. Going to higher clean % with mix of wind/solar expansion + new storage + dispatchable nukes is very possible and the political momentum to throw a lot of money at this is there.

- I think it's unlikely costs can be reduced much in most parts of the US if we still want 100% dispatch capability.

- It's not just about cost. We should consider paying a bit more for carbon considerations, within reason and without forgetting who gets punished by higher prices in the short term, and without compromising the 100% dispatchable standard.

- Wind/solar LCOE is approaching very low numbers, so it should be used as much as possible, but we shouldn't throw a trillion dollars from the government, as the article suggests, to shove wind/solar down the grid's throat. It's likely to break the grid and waste a load of money. Government should just ensure the markets can properly deal with higher renewables without compromising the grid. They can do this via establishing capacity markets, resiliency markets, R&D for storage tech, etc. They might also do a carbon tax to try to set the higher cost we are willing to pay from above.

- There are significant resilience, material, and regulation challenges. It's not just the energy storage for a few hours, but longer term for days and weeks. If we are a responsible society and expect smooth operations and we want to do an extreme renewables expansion, we need to pay the price to keep an idle fleet of fully dispatchable gas/nukes in addition to the renewables fleet or establish multi month storage.

> I think it's unlikely costs can be reduced much in most parts of the US if we still want 100% dispatch capability.

I disagree, if there was a desire to transport molten salt tanks and retrofit coal plants with supercritical steam turbines and build large molten salt storage reservoirs on such sites, we could do it today for under 5 cents/kwh for 24/7 availability.

See comments [0] and [1]

[0] https://news.ycombinator.com/item?id=23588691

[1] https://news.ycombinator.com/item?id=23581910

The key enabling technology is to tie all the separate grids together with High Voltage Direct Current (HVDC) power lines. This technology is flying under the radar but it's getting cheaper very quickly and it allows for continent size electric grids.

To give a sense of how cheap it's getting consider that there's a proposal to send solar energy from North Australia to Singapore using this tech. That's about 4000km distance total with about 3000km of that under the ocean. 4000km is very close to the distance from Seattle to Miami. If Singapore can afford this then for sure the US can dig around in the couch and come up with enough spare change to build this out.

"But here are some problems."

The opportunities are insurmountable.

"...a very low resiliency because it depends on national grid."

Grid will become massively decentralized. Transmission costs are already greater than production costs in many systems, a trend that will continue.

We'll get resiliency "for free".

Others have already commented on demand shifting. Additionally, producers will innovate, creating new "products" to meet unique customer demands. Like colocation.

"...nukes and nat gas will not be competitive 99%"

Current nukes must be propped up for as long as possible, stopgap until renewables can overtake.

We'll definitely need new nuke options, like small modular reactors (micro reactors) and reprocessing. (My future perfect is traveling wave reactors, but that's a long ways off.)

We need to phase out natural gas asap.

Instead of dispatchability payments, why not demand response? Thermostats, electric cars, water heaters, dryers, smelters, and data centers can all turn on a dime if an incentive is there.

Also, what's wrong with policy wonks?

Most data centers needs non-stop 24/7 reliability, and there are very few use cases where it makes sense and is even possible to delay your workload waiting for lower energy prices. Smelters have complicated processes, and they cannot really turn on a dime. Lastly, data centers and smelters are expensive capital investments, and typically the economic calculation requires them to be run non-stop to recoup the capital investment.

> Most data centers needs non-stop 24/7 reliability, and there are very few use cases where it makes sense and is even possible to delay your workload waiting for lower energy prices.

Perhaps your parent was thinking of EC2 spot instances?

> Lastly, data centers and smelters are expensive capital investments, and typically the economic calculation requires them to be run non-stop to recoup the capital investment.

I suspect this is true for almost all commercial operations.

Residential electric car charging is the only thing I can think of that's

1. A relatively large draw

2. Relatively time insensitive

In general, people aren't going to delay their dinner or wash by a few hours to save a few cents in electricity fees. They might be ok with car charging being delayed as long as it's full when they leave in the morning.

Perhaps your parent was thinking of EC2 spot instances?

Maybe, but the whole reason cloud providers offer ephemeral instances is precisely because letting the machines sit idle is a waste of capital.

I suspect this is true for almost all commercial operations.

I'm not sure if that's the case. I think it depends on the efficiency of production, ability to stockpile the product and the demand. Not all factories operate around the clock, because round the clock labor is also very expensive.

Residential electric car charging is the only thing I can think of that's (...)

I think a relatively easy development here would be to install large water tanks in people houses, and heat them up/cool them down when energy is cheap, and reuse stored energy when energy is more expensive, to save on heating/AC. But yes, energy these days is so cheap compared to labor and to value added that a complex schemes of idling production when energy prices are high is probably a non-starter in developed countries.

A system that heats water in tank for bath by heat pump at idle time is popular in Japan since mid 2000s (named "Eco Cute"). It expects unused power is found in midnight because of nuclear plants keeps generating power all time. It looks works well until 2011 earthquake. After that, not much nuclear plants working so the assumption isn't very good. Modern systems has connected to HEMS to heat water when power is generated by solar panel. AFAIK currently not connected with smart power grid.

I think a relatively easy development here would be to install large water tanks in people houses, and heat them up/cool them down when energy is cheap

Literally already happening. I have a demand response control box retrofitted onto my electric tanked water heater. It shuts off during high rate times.

Yes, but demand response for thermostats, EVs, water heaters, even overnight ice freezing for daytime cooling use - all those are dispatchable loads.

it's just difficult to coordinate all these things top down and there's a bunch of extra infrastructure and IT, and more regulation and rules and complexity. People telling you when you can and can't do things, etc. Eh. Seems unAmerican.

But given the total enthusiasm and commitment, I bet it could be made to work. But maybe we have to accept that it's gonna cost more, and that we'll probably lose on other sides. This renewable world just might not be such a nice place. Power would be less taken for granted because it would not always be available. a lot more people would have to work in the industry to keep it going (low EROI for renewables) and they wouldn't be working in other industries. Ironically, that's a plus in the article, which touts immense job creation. lol - not a very good objective. The landscape near populated areas would be dominated with wind turbines and solar plants.

> it's just difficult to coordinate all these things top down and there's a bunch of extra infrastructure and IT, and more regulation and rules and complexity. People telling you when you can and can't do things, etc. Eh. Seems unAmerican.

I honestly read that as "it requires people to communicate and co-ordinate in a way that is a net benefit to everyone, seems un-american" and thought yeah that seems about right.

For what it's worth national grids with co-operation agreements and a strict regulatory regime absolutely do work when well implemented, the UK has had one for decades.

People always nay-say demand management, but here in the UK the National Grid already does this. They have contracts with many large industrial companies who presumably get lower rates in return for being able to turn down their demand on request.

They do, agreements are typically in place with companies that use a lot of power but can live without it for a fair few hours.

aluminium production/smelting uses a colossal amount of electricity but they can shut down quickly for a few hours (as long as they don't let things get too cold and freeze which is expensive as it destroys equipment and takes a long time to restart - which is what happened in the ransomware attack against norsk, they had to switch to manual line processes to keep it moving).

For anyone interested the people responsible for monitoring our grid have a good website with a tonne of info.


Say you drop it to 80%, would that work?

Drop what to 80%?

Our economy works really well because electricity is 100% dispatchable. You flip the switch at any time, things work. If you go below that, the economy will follow the weather, which I guess is just a different way to operate. But it's not so fun. I have some friends in South Africa where power rationing is in place. You basically have a few dead hours every day because the grid isn't run well and they haven't built enough capacity. They hate it. You're cooking a nice pie for grandma and then BAM, powers out. You're on your treadmill, and BAM - you hit the wall. You're playing CoD, and BAM. Sucks to suck.

Headline says: "US can reach 90 percent clean electricity by 2035" and you gave your analysis.

GP is asking how your analysis would change if we drop the target to "US reaching 80 percent clean electricity by 2035".

The SPP region (small piece of Texas and then the heartland around Oklahoma and Kansas although there are many states) has already hit > 70% renewable penetration during some hours of the day with very little storage, but a LOT of wind (> 20 GW).

The problem is sustaining that 70% for a longer time period. Wind is intermittent, so firming it up with additional solar (also intermittent, but generally at different times) and a lot of storage would have to be the next step if you wanted a sustained > 70%.

I'd guess with very high probability that this region (very geographically large), but not too many large population centers will hit 80% renewable penetration within 3 years based off of current growth. It is still a guess as to how long it will take before you'd see 80% across peak and not in the morning. Still, this has all happened in a few years.

Short answer: anything above 20% is unproven and will cost significantly more if you expect to have 100% uptime/reliability. If you're okay with 95% uptime, wind/solar can get the job done, and I believe cheaper based on https://www.sciencedirect.com/science/article/abs/pii/S25424...


The flaw with renewables is they are so dam cheap and so dam finicky. Basically, renewables force nukes and nat gas to sell power at a loss or sit idly by for most of the time. This becomes worse and worse as the renewable's market share increases and they eat away at the profitability of the dispatchable sources. The limit is something like 20% at which point nukes and nat gas just can't make any money and so a free market system would just get rid of them. But yet, they need to be there, just in case, because they are dispatchable and you want shit to work at all times. Unfortunately, prices for consumers don't go down because you need the super dispatchable power sources which just sit there for a lot of the time and cost just about the same if you have em on or off, and you also just built a bunch of renewables which are cheap and make money for the guys who built them, but hurt the rest of the grid because you have a bunch of idle equipment (gas/nukes during the day and most of the year, panels at night). Lower cost storage is a way to address that, but has to be really low cost, especially when you increase wind/solar market share. Keep in mind, there's not yet a good solution for long term storage (days or weeks or months).

To go beyond 20% renewables you need a power market that rewards dispatchable sources, essentially preventing solar and wind from taking all the money and bidding everyone into oblivion. If the market is fair, there will be little incentive to push renewables because they just make things more expensive. There's a bunch of ideas on how to do this like capacity payments. Might work, but it's gonna cost more because of all the unused equipment. Everything is possible, you just need to be willing to pay for it.

In an 80% renewables grid, most of the time, renewables will be meeting 100% of demand and producing much more than needed. 20% of the power is delivered by nat gas. But they will mostly be called to do that only on the bad days where they will have to be able to match closer to 100% of the demand. So you've basically got two nearly complete fleets. Who's gonna pay for the idle equipment? The solar power producers should right?

One data point: Germany is 30% renewables and pays 50% higher than EU average. That's old prices in a country not ideal for renewables. It's no Arizona or Washington.

> One data point: Germany is 30% renewables and pays 50% higher than EU average.

The price of energy in Germany was already high before the switch to renewables, for reasons unrelated to renewables. The renewables surcharge only is 21% of the price per kWh: https://www.statista.com/statistics/1009248/household-power-...

Edit: I feel it's worth mentioning that the price per kWh only tells half the story. High energy prices also incentivize buying efficient appliances and having well-insulated houses: https://www.researchgate.net/figure/Annual-household-electri...

As anecdata, our German household of five only pays about 80€/month for electricity (cooking with electricity, heating and warm water with gas).

Every serious plan of conversion of an economy towards renewables has reduced energy consumption through reduction of waste as a core element.

> Short answer: anything above 20% is unproven and will cost significantly more if you expect to have 100% uptime/reliability

CPS Energy, serving the San Antonio, TX metro is already above 20% renewable[1] and its rates are very low, to the point San Antonio is a preferred data center location.

[1] https://cpsenergy.com/en/about-us/programs-services/energy-g...

So CPS energy and it’s grid are disconnected from the rest of the national grid and have no reliance upon external sources to ensure dispatchability? Cool. Would love to hear more about how they managed that trick.

> Cool. Would love to hear more about how they managed that trick.

The grid isn't one big national national. Our national grid is a bunch of separate grids. Since San Antonio is on the middle of Texas it is on the Texas Interconnection manged by ERCOT [1].

Here's information on the overall grid generation where if you look at the fuel mix report for 2020 YTD you'll see 26% is Wind so overall it is also above 20% [2].

[1] https://en.m.wikipedia.org/wiki/Texas_Interconnection

[2] http://www.ercot.com/gridinfo/generation/

So they rely upon the external grid and pass the externalities involved in being being able to dispatch power when you flick the switch to some non-renewable plant elsewhere. Gotcha.

That's interesting, I didn't know of these limitations.

I wonder if we could end up doing something like we do with farming and pay farmers not to grow crops

Apparently the battery plant Tesla built in Australia only powers 30k homes which is a tiny amount. They are working to 'double' the capacity, which I assume goes from 30k to 60k.

It seems to be making net-positive amount of money and Tesla built it in record time but sourcing that much lithium to build it for millions of people across the country sounds like a tall order, not to mention the 2-5x costs/time any major public works infrastructure project would add based on recent history of all of them in western countries (for ex: both of Obama's clean coal and solar project went bust, before roll out, after billions spent each and many years, so outright failure of even top-down federal-level projects must also be considered).

I'd be curious to hear an expert speculation on the matter for what a practical battery farm deployment would look like. Which would likely need to exist, assuming we're abandoning backup energy sources like nuclear and natural gas.

The Tesla battery only powers about 30k homes for about and hour and a half. Dont get me wrong, it is an amazing piece of tech for stabalising the grid and reducing the dependence on the gas peaker plants but it is a very long way from mass storage for want is a fairly small Aussie state. SA still regularly pulls a lot of power from the Victoria grid which is still dependent a lot on dirty brown coal.

Hopefully it is just the first steps towards a clean grid and not an example of a Concorde like moon shot failure.

The battery plant capacity is growing in hand with decentralised battery storage, an interconnector to New South Wales which is building hydro storage capacity, and demand management. Apparently generation and transmission costs are falling nationwide too, but not to the extent that my bills are going down.

That's good to see, it has been about a year or so since I last checked in any detail. That plus from what I have heard even conservative numbers put Snow hydro 2 as being a big win in terms of storage capacity.


Digging into someone's comment history to paint an agenda is a very reprehensible behavior. I was a Covid-19 skeptic as well, I changed my opinion when I saw that the overall death rate YoY - the excess deaths painted a very bleak picture. To think that someone would go down my comment, post, or tweeting history to comment on my bias makes me wonder if it's worth being on social media at all.

It's ironic a bit because if you dig into spenrose's comment history they're critical of nuclear.. and seem to mostly comment on articles like this one.

I think this person has an agenda they want to push and they want to shout down anyone with ad-hominems who doesn't share their world view.

Is this not against hacker news guidelines?


> It's ironic a bit because if you dig into spenrose's comment history they're critical of nuclear

Huh? spenrose's most recent comment about nuclear power starts with:

> I am rooting for lots and lots of new nuclear

Even that comment chain is in fact not pro nuclear. If you read further they're basically arguing against nuclear saying that why even bother because it's not economical. And the only nuclear they want is "safe nuclear"

> Safe micro-nuclear plants are an excellent start on the "economy" side of "political economy." Now y'all need to build a political coalition—and pro-variable-renewable peeps such as myself must be part of your core coalition if you are going to get anywhere. So start treating us as allies.

It's a lot of double speak so I can see how it would be confusing. But definitely take a closer look.

Argue the argument, not the arguer.


You are not arguing with the person you reply to. You are arguing in front of an audience of the rest of us.

Pointing out the ad hominem's goes further in discrediting the argument, than throwing your own insults and trolling through their history.

All he did was discredit himself in front of the rest of us.

Nuclear energy, so far, is the only proven way of producing more than 3/4 of a nation's electricity from carbon free sources other than geographically limited options like hydroelectricity. It's a proven technology that has served as France's main electricity source for close to 4 decades. As compared to plans to turn our power grid into a Rube Goldberg machine and dealing with power outages when wind and solar output don't conform to projections, it's not hard to see why people favor nuclear.

I don't really agree with this, even though I am an avid nuclear proponent.

I do think France has it right, but there's some other things to consider. France is phasing out some of its nuclear and does not plan on replacing it. No they aren't going to 0% or even 50%, but they are doing it and this is a good idea. They are replacing old plants with renewables where they can and replacing and building new plants where they can't. This is part of why France has been doing so well in terms of CO2eq production (one of the lowest in Europe, which is especially crazy considering it doesn't have the hydrodynamic opportunities that Sweden does, who also uses a fair amount of nuclear).

Remember that no matter what you do, you want a diversified portfolio for power generation. You don't want to rely on a single resource.

> France is phasing out some of its nuclear and does not plan on replacing it

A statement they keep making to placate anti-nuclear advocates but keep postponing and postponing. Because there's no way to reliably generate the majority of electricity using wind and solar, and only small parts of France are suitable for hydroelectricity.

Diversification doesn't matter for energy that comes from heat engines. Just like how many countries successfully grew their economies with > 80% fossil fuels plants, there's nothing wrong with powering a country entirely on nuclear power. Sure it's good build alternatives where it's cheaper. But the main draw of nuclear power is that its output entirely controllable. Diversification is much more important in renewables where you don't control the weather.

> A statement they keep making to placate anti-nuclear advocates but keep postponing and postponing. Because there's no way to reliably generate the majority of electricity using wind and solar, and only small parts of France are suitable for hydroelectricity.

This is really confusing. Me, nor France, said that they are going to go to 0% nuclear. In fact, I remember saying

> No they aren't going to 0% or even 50%

Maybe I should be more clear and say "Less than 50%".

So I'm confused because I'm reading your response as disagreeing what IS happening (decommissioning IS happening) and taking it to mean that ALL reactors are being proposed to be decommissioned. Which is definitely not true.

> Diversification doesn't matter for energy that comes from heat engines.

What? We've always had diversification. So does every other country. They also always have. Different energy sources have different advantages. It would be dumb not to use the one that has the most advantage for the specific situation. And of course there's the old saying "don't put all your eggs in one basket." But you seem to be disagreeing with this but also agreeing by saying

> Sure it's good build alternatives where it's cheaper

A big part is that France doesn't produce its own nuclear and so renewables do provide some self resilience. That does have some advantages, but obviously France does want to keep a mostly nuclear system (recycling is self resilience too).

> But the main draw of nuclear power is that its output entirely controllable.

Which this isn't really true. You CAN throttle nuclear but not infinitely and surely not like you can with coal or natural gas. France just sells the excess and this "throttles" a lot for them. It is very efficient.

> there's nothing wrong with powering a country entirely on nuclear power.

The issue is easier for France than other countries, simply because momentum. One thing you have to consider with building reactors is that they take longer to develop and build than other sources. Which I'll agree that we should have started yesterday, but let's be realistic. If you really believe this then you should also be a proponent of renewables because they are a better stop gap to a nuclear based system than coal and natural gas are.

> Maybe I should be more clear and say "Less than 50%".

> So I'm confused because I'm reading your response as disagreeing what IS happening (decommissioning IS happening) and taking it to mean that ALL reactors are being proposed to be decommissioned. Which is definitely not true.

They keep delaying their supposed nuclear cutbacks: https://www.pv-magazine.com/2018/11/20/frances-decision-to-d...

The reality is that wind and solar can't replace nuclear or Fossil fuels without some insanely good energy storage that is currently the stuff of science fiction.

> Which this isn't really true. You CAN throttle nuclear but not infinitely and surely not like you can with coal or natural gas. France just sells the excess and this "throttles" a lot for them. It is very efficient.

Incorrect. Reactors can be modulated in terms of thermal output. You can throttle the electrical output even if the thermal output remains the same by cooling the water more aggressively. Furthermore excess energy is not a hard problem to solve. Desalinate water, or other energy intensive tasks are good energy sinks.

> One thing you have to consider with building reactors is that they take longer to develop and build than other sources. Which I'll agree that we should have started yesterday, but let's be realistic. If you really believe this then you should also be a proponent of renewables because they are a better stop gap to a nuclear based system than coal and natural gas are.

France built up it's nuclear generation from 10% to 80% in the span of 15 years. By comparison, solar and wind are still waiting for an effective energy storage system to come into existence.

Renewables are not capable of generating more than ~40% of a country's electricity due to the fundamental mismatch between the time that energy is generated and the time that energy is consumed. Look at Germany to see the failure of solar and wind to produce a carbon free energy system. The solution exists, and has existed for over half a century now.

You seem to be taking my statements to well past a good faith misinterpretation so I am not going to continue this further. There is no point discussing something when the other person doesn't respond to what you're saying but something completely different.

I am not making any misrepresentation. Yes, you did not say that France is going to reduce their share of nuclear power generation to 0%, or even 50%. Their reduction plans are more modest than that. I understand this, contrary to your insistence that I am misrepresenting your claims.

And my point, is that France has repeatedly postponed even these much more modest plans for denuclearization.

Something else I really like: flipping the switch and seeing things turn on.

Regarding the immense capacity required for wind/solar: see Figure 10 of https://www.sciencedirect.com/science/article/abs/pii/S25424.... They essentially optimize the renewable mix to minimize cost while considering different scenarios for storage costs. The optimization must meet 100% of the demand at 100% of times for the grid in a few different states over the last decade of real data. Plot shows the levelized cost of power or LCOE (the color) in the storage energy vs storage power capacity space. In Arizona, it's no problem, renewables are for sure gonna work out even without a storage miracle. Not so much everywhere else, although I bet most people are so climate woke they wouldn't have a problem paying double for power. I don't want to though.

There are flaws here as well. First, how will they enforce the optimal mix of solar and wind? Will they set quotas of how much of each can be built? As soon as you fall out of the optimum, prices go up. And then there's the material/engineering issue. This kind of storage production is out of this world. Yes, Tesla is going to make 100 $/kWh batteries and they will have dozens of giga factories, but we'd have to produce more than we have ever made. That scares me, and even then, the cost is not lower than natural gas. Gentlemen, natural gas is king - it's so cheap, and easy, and dispatchable.

For resiliency, the key is distributing your power and not depending on others to get things to work. So don't rely on solar plants in another state, or gas pipeline, or whatever. Get it done locally. The paper averages the nationwide demand and production which requires a national grid, and this is not resilient. Grid failures tend to cascade and everyone is affected. Some literature on terrorist or weather based grid vulnerabilities (might not be accessible since this enters sensitive info):

- https://www.wsj.com/articles/assault-on-california-power-sta...

- Lanza, R. Fragility of Energy Networks. in INTERNATIONAL SEMINARS ON PLANETARY EMERGENCIES AND ASSOCIATED MEETINGS- 47thSession (2014). doi:10.1017/CBO9781107415324.004

- Wang, S. et al. An Optimization-Based Adaptation Framework for Coastal Electrical Infrastructure Resilience To Climate Change.

PS. Love the character attack with insinuation of conspiracy theorist. Nicely done. To be clear, I get paid by nuclear lobbyists to troll sites like these.

Natural gas costs more than solar power under 50% utilization. As in the fuel it’s self costs more than solar power plants that waste half of all power produced at current prices. This means a vast over abundance of supply is part of any rational minimal cost strategy. Storage costs are so high it further makes solar tracking arrays cost effective for mitigation of the duck curve.

Also, hydro is a massive part of the energy mix going forward and excluding it gives meaningless results.

since hydro is basically constant, any analysis can just subtract it from the total demand. the dynamic analysis can proceed without considering hydro.

Reading the description on Wikipedia, I would say it can be part of the dynamic analysis. "With a dam and reservoir it is also a flexible source of electricity, since the amount produced by the station can be varied up or down very rapidly (as little as a few seconds) to adapt to changing energy demands." https://en.wikipedia.org/wiki/Hydroelectricity

Exactly, large scale hydro is extremely flexible, shifting production not just on a hourly basis in many cases across months. Hydro’s rapid response is why pumped storage was initially developed.

Environmentally, rivers need variable flow rates to minimize silt buildup so doing so is actually good for the environment. Comparing the costs of constructing a dam with the cost of extra generating capacity, this flexibility is also extremely cheap. They do maintain a minimum flow at all times it’s generally a low fraction of average let alone maximum capacity.

A few things stick out to me:

* This study assumes that we can ramp up purchases of solar and wind generation to match this graph: https://imgur.com/lMU252r while costs continue to decrease. That's not how markets work. Rapidly buy solar panels and the cost of solar panels go up. Put solar panels in all the places where it's easy to install and now you have to pay extra to put them in places where it's not so convenient.

* This study thinks we can achieve this with only ~200 MWh of storage. This is an optimistic projection. To put this in perspective, daily US energy use is 11.5TWh. We're giving solar and wind generation only ~15% allowed variability before we're dipping into gas. Yeah, it does say that it's only 90% clean electricity, but even reaching this figure is optimistic.

That's far too optimistic about storage. For small off-grid systems, you need ~5 days of battery capacity to ride out lack of wind/sun. On a large scale you'll need less because you can move power from place to place.

But still. Sun is down half the day and wind speeds can be low in areas hundreds of miles wide. Its obvious they will need at least a day of battery capacity, probably more. Not hours.

Solar and wind are great. Many areas in the US have hours long periods during the day where power is 100% renewable. But batteries are way too expensive and I don't like when politicians make laws based on predictions 15+ years in the future.

How did you get 200 MWh? From page 20 of the report:

"In summary, retaining existing hydropower capacity and nuclear power capacity (after accounting for planned retirements) and about half of existing fossil fuel capacity, combined with 150 GW of new 4-hour battery storage, is sufficient to meet U.S. electricity demand with a 90% clean grid in 2035, even during periods of low renewable energy generation and/or high demand. Under the 90% Clean case, all existing coal plants are retired by 2035, and no new fossil fuel plants are built beyond those already under construction. During normal periods of generation and demand, wind, solar, and batteries provide 70% of total annual generation, while hydropower and nuclear provide 20%. During periods of high demand and/or low renewable generation, existing natural gas plants (primarily combined-cycle plants) cost-effectively compensate for remaining mismatches between demand and renewables-plus-battery generation—accounting for about 10% of total annual electricity generation, which is about 70% lower than their generation in 2019."

150 GW of batteries with 4-hour discharge is 600,000 MWh.


The graph displays 150 GW, with no duration specified. 4 hours of 150GW storage, or 600GWh of storage, is still only 5% of the nation's current daily electricity consumption.

And that's probably going to be an even smaller share in 2035 as more vehicles go electric.

> This study thinks we can achieve this with only ~200 MWh of storage.

Agree, very unrealistic.

One thing that always disappointing to me about these conversations is that solar == photovoltaics, batteries == li-ion electrochemical… luckily the world isn't limited to these asumptions and developments have been and continue to be made, even if they are mostly ignored/overlooked by the HN crowd.

Things like kinetic storage are geographically limited. Most places with available hydroelectric storage already generate electricity from dams.

Chemical storage in the form of natural gas produced by the Sabatier process is very inefficient. Currently it's at 55% efficiency to go from power to gas. Even the most efficient combined cycle plants are currently 65% efficient. End to end efficiency is ~1/3.

Also, using synthetic natural gas requires a means to capture the exhaust of gas plants to actually make it carbon neutral. Converting the fraction of a percent of carbon dioxide in the atmosphere is not currently possible.

The conversation is mostly about batteries because that is the only known option at the moment, other than geographically limited ones.

There's also many kinds of thermal energy storage systems in use (I think molten salt systems are the most promising[0], esp if you design salt mixtures that have have higher delta k's[1] and more efficient HEX and steam turbines to about ~60% efficiency of converting heat to electricity[2]), not limited by geography.

Going of the example powering 100MW turbine with a working temp range of ~278C[0] and using an eutectic mixture of calcium chloride and magnesium chloride (ionic halide salt)[1] with a working temp range of ~1440C, can make the "tank of about 9.1 metres (30 ft) tall and 24 metres (79 ft) in diameter to drive it for four hours by this design." last for nearly a day, though with higher upfront cost than a typical NaNO3 + KNO3 system, but nowhere near nuclear costs (except for the similarities with dealing with supercritical water for electricity generation).

[0]: https://en.wikipedia.org/wiki/Thermal_energy_storage#Molten-...

[1]: https://gist.github.com/cinquemb/015f8b0775db2381513f0e20186...

[2]: https://www.solarpaces.org/vc-funded-thermal-battery-based-m...

Solar thermal plants [1] effectively already use this system. They heat a model salt container, which drives a heat engine. The molten salt remains hot enough enough to keep producing energy through the night. However this scheme requires both solar collectors (in the form of mirrors) as well as the molten salt tank plus a heat engine. The resulting levelized cost is about the same as nuclear power [2].

If this startup does manage to produce energy storage at the efficiencies and costs as advertised that'd make intermittent sources more viable. Ultimately building solar and wind is betting that we'll be able to build efficient, cheap, and massively scalable energy storage.

1. https://en.wikipedia.org/wiki/Concentrated_solar_power

2. https://www.lazard.com/media/450784/lazards-levelized-cost-o...

Yes, I had CSP in mind, however most CSP systems today use the kn03+ nano3 eutectic mixtures, and the LCOE is based on that assumption, as well as assuming you can have no baseload from the storage, and cannot distribute storage separately from collection. And even with their assumptions the range for solar thermal capital costs/kwh is about 20% less than nuclear on the top end, 41% less on the bottom end. One also has to consider the political costs that nuclear comes with and that CSP doesn't.

I seriously doubt America will buy solar panels quickly enough to outstrip supply. Much more likely that production will increase and costs will go down due to economies of scale.

Also there is no shortage of places to put solar panels, especially in America.

Are you predicting that solar panel production won't continue to scale to meet demand?

The limiting factor of producing solar panel at these immense scales is the cost of raw materials. As other posters pointed out this study is already assuming that the cost of solar panels and wind will be barely greater than the cost of materials.

And yes, if the US tries to increase it's solar capacity by a factor of 50 in the next 15 years (and especially while other countries try to do the same) these materials will become more scarce and thus more expensive.

In my county in California, the county purchases the electricity while PG&E provides the infrastructure. The 50/50% solar/wind plan is only around $4/month more than the standard plan.


This plan is rather misleading. You're buying energy from the same grid, but paying a markup to say it's coming from solar and wind. Say the company generates 100Wh of solar, 100Wh of wind and 200Wh of natural gas. Bill buys the 50/50 solar and wind plan and "buys" 100Wh of solar and 100Wh of wind at a marked up rate. Charlie buys the normal plan and gets his energy from the remaining sources. If a meteor hits the gas plant at night, Bill's electricity is still going out.

California still generates 45% of its energy from fossil fuels. And another 10% from nuclear which some don't consider renewable. Solar and wind account for 21% of electricity generation: https://www.energy.ca.gov/data-reports/energy-almanac/califo....

California is sunnier than most parts. That makes a big difference.

California also charges people up to .48/kwh so there's a good incentive for people to get solar.

Moving to clean electricity is not strictly about money. Even if it is for you, factor in the difference in health costs of people around a bunch of burning coal, vs solar or wind. also environmental cleanup. etc.

Not convinced.

I’m using utility solar for about 20% of my electricity needs, precisely to experience what’s claimed. Average cost is 50% above regular sourcing, and rising significantly.


I installed 10kW on my home DIY in AZ. Cost was $1.00W and $0.70W after tax incentive. It took me 3 weekends all by myself. It works out to like 1.7c/kWh over the life of the asset. Next step is to add batteries and go off grid. Right now, the utility pays me 9.6c/kWh so I’m not in a rush until they further sabotage net metering. My cost pre-solar was 14-16c/kWh. Also, the cap on grid connected was 10kW and it’s about to go to 20kW.

Oh, and I had never done it before. Read a lot and watched videos. Technology is making it plug and play.

That's a wonderful achievement. I've browsed www.sunelec.com for years because they sell pallets of panels for less than 0.50/watt and I think ... someday DIY.

I bought panels from https://www.santansolar.com/. Enphase micros and Ironridge racking from https://tandem-solar-systems.com/. Then I made permits using https://get.solardesigntool.com/.

All-in-all, it was fun and easy.

That's pretty cool, I wonder what your $/kwh would jump to with battery storage… probably a significant jump compared to 1.7 cents/kwh over the lifetime (esp consider degradation of electrochemical batteries over charging cycles)

SolarReserve bid under 5 cents/kwh in Chile in 2017 for a grid scale 24/7 CSP/molten salt storage system… we'd probably need a safer way to transport molten salt tanks from high Irradiance areas to local grids to give 5 cents/kwh for most people though.

So 2 Powerwalls should cover me. That’s 27kWh but only 14kWac peak. It’s the peak that’s really the issue. If I’m running a furnace and my stove, it may go over.

But currently 2 PW is $17,000 with 10 year warranty. That’s $142/mo. If they work for 20 years, that’s half. Enphase is coming out with its own system.

Anyways, it already makes sense but for peak usage. I only use 15kWh max a night but my peak can be up to 20kWac for a few minutes at a time.

If battery makers can increase peak of home stationary batteries - Tesla cars peak much higher, like 400kW - and Tesla’s battery costs are under $100/kWh, the end is nigh for electric utilities. They need to start stringing fiber.

End is even closer if Tesla enables vehicle2grid soon.

You'll definitely have chemical degradation within those 10 years, but even assuming $71/month over 20 years + $0.017/kwh, that's still pretty expensive for an urban environment, though may work out for suburban and rural (I pay about 11-12c/kwh in my 2br apt in Jakarta), I still think there is still plenty of future opportunity for electric utilities (esp for industrial applications that will require peaks alot higher than residential).

I really don't see costs coming down that much in the next 10 years for electrochemical batteries.

Yes, utilities seemingly will be all industrial. Their poles in residential neighborhoods could be put to good use offering other high value services.

I wonder if Tesla should buy some of these large publicly traded utilities, like APS.

I expect with grid-tied and reasonable utility policies, you can reach zero and batteries would just be an added expense.

It's just when they start "regulatory capture", batteries can take you out of the drama.

We did a PPA for full replacement at a time in our life when we were probably at max usage (2 teenagers at home, all electric appliances) and haven't looked back. Our monthly bill immediately dropped and stayed down. As far as I can tell, the only thing I need to worry about is the risk of deflation, and there's not much I can do about that.

It GREATLY depends on where you live. Where I live, solar is not a good option. Terrible yield on many cloudy days - horrible yield in winter.

Is that a price of power rise or a price of transmission rise?

That’s the curious bit: power generated per flat panel-rental fee is decreasing.

Rental cost is consistent.

Are people willing to pay a green premium for it?

We have a company near me that has some of the cheapest power sources from wind, but they used to market with the slogan “pay more for power.”

In many places, it's the "fees because we can" rise.

We have that here too. Plenty of “fees.”

The UK has recently ended its coal-free streak for power generation; up until a few days ago we had been coal free for 67 days, the longest time since the start of the industrial revolution.


The plan is for the national grid to close coal plants by 2024 leaving just natural gas (the gas, not the American car fuel) as a high carbon source in our power grid.

I wonder if there is a way to quantify the health benefit of a solar installation on a roof, and/or associate battery storage as a dollar value on a per roof / per home basis.

I was thinking about going big and buying about 100 panels for my 12 acre property. Anyone have any success in small scale solar farming with a utility?

The questions are always "How much will it cost?" "Is it worth it?" "Do we have the political will to do it?"

My opinion is that it doesn't matter how much it costs; it is worth it at any cost, because continuing down the path of making our planet uninhabitable for humans is not a winning strategy.

However, I don't think we have the political will to do it, which is really the only thing that matters in the end.

not in that order.

But it won't.

One more reason to not re-elect Mr. I Love Coal.

Ever since watching "Planet of the Humans" I find myself scrutinizing anything that says "renewable energy" because it might be burning trees.

This article did mention solar, wind and batteries and carbon-free so it might be ok.

People are so focused on $$ when it comes to (green or other) energy. The real problem is not $$ but Joules. Building any type of a power plant costs energy itself, and unfortunately unlike capital, which is just a concept that you can argue with and win, there is absolutely no arguing with physics: https://en.wikipedia.org/wiki/Energy_return_on_investment#ER...

The energy amortisation time of green power generators (the time it takes to pay back the Joules) is orders of magnitude higher than any dirty energy. This problem is well visible once you count all of the world's energy usage and just plot it. The rollout is anything but fast: https://en.wikipedia.org/wiki/World_energy_consumption

Interesting chart, but it does not seem to consider the cost of the installation, just the fuel, or it seems to assume a very long amortization period for nuclear/fossil fuel generation plants.

With wind and solar, all you have is the plant. There's no EROEI on the fuel because there is no energy investment in the fuel for wind and solar.

For a very long time solar panels would cost more energy to manufacture than they would output through their lifetime. This is no longer the case, but the amortization period is still large.

Fuel is not the problem. The power plants are the problem.

They show the payoff period Germany and in Southern Europe. I wonder what the payoff period is in Texas.

It's probably better, but don't forget that this is inherently a global problem. Scotland is running 100% renewable - so what, who cares? Not only it's not enough (the country is small), they also import a lot of goods with manufacturing energy footprints that released co2 elsewhere.

Even if the US switches to 100% renewable the problem is far from solved. Look at the population pyramid of Nigeria, for example. Those people will want homes (cement, steel) and cars and flights.

I think I understand where you're coming from, but I would say 'don't let the perfect be the enemy of the good'.

Grand jumps in capability rarely (if ever) happen in the real world.

Gradual change can feel like a capitulation, but it's often the safest and best overall.

If you try to force huge change and fail or things go wrong or the system is deficient in some way, there can be a big backlash.

The CO2 released in production and deployment of solar cells is regrettable, but as soon as that solar cell comes online it starts displacing fossil fuel.

What is the significance of this?

Well you can do the maths.

Let's assume it takes 1 year to pay back the energy usage, on average, for the green projects (look up that Wikipedia link). Let's also assume that we can dedicate 5% of the world's energy budget towards new power plants and that the world stops growing. So: 5% energy recession and zero global energy demand growth. How long would that take to roll out? 1/5% = 20 years, all of those years will be of serious austerity. Multiple countries (especially those with "good", that is growing demographics) would blow up. Up to you to decide if it's feasible to execute, and those are very optimistic numbers. What if the energy usage grows at 2% and we can dedicate only 3%? Well, that would give us 1/(3% - 2%) = 100 years to finish the project.

These are not happy things to look at. Personally I'm just trying to avoid the impending climate doom. Whatever happens - this problem is here to stay with us for many decades.

That's some weird kind of fear mongering. You assume instant replacement of fossil fuel plants. If you don't do this, but keep them running, and add renewables on top, it looks different.

Using your numbers to demonstrate it of 1 year EROEI. Let's say we make a starting investment for one year where we take 5% of the world's energy output to invest into new renewable power. After that year is over, we have capacity to generate 5% of the world's energy output per year, but we don't replace any of the existing fossil plants, but put it to work for new energy plants. So the second year we spend that energy on another 5% of renewable energy. Now we have capacity of 10% of the world's energy budget at start of year 1. If we continue doing this, we have 20% after the third year, 40% after the fourth, and 80% after the fifth. By the end of the sixth year, we can generate 160% of the energy demand of the world at start of year 1. We can now turn off the fossil fuel plants and enjoy a 60% increase in energy output, most likely more than the world has grown in that time.

Now you wonder, won't it cost tons of money to use all of this energy to build new plants? Sure it will, but that's already priced into things. That's what we have banks for.

Your fears are unfounded.

You're right, I did not compound the green energy (or, as you say, assumed parallel termination of dirty plants). It would work if the world could collectively put aside 5% of the energy budget. Not sure if that's realistic, and also will there be enough commodities (though those are at multi decade lows in prices right now), but if so it should work.

Let's see how the world looks like post covid, as we're very likely to experience a huge "energy recession". Something that would be similar to that 5% energy austerity rule.

We have tons of unused capacity. Energy shortages would be indicated by high energy prices, but they are low. China had been building coal plants before actually having demand for them (creating a great place to mine BTC in the process), then expanded to build coal plants all over the world. The US is swamping international markets with fracking oil. In OPEC, the hate for each other still seems to be stronger than the desire to increase prices.

Nuclear is the answer.

You could start planning now and you would be lucky to get shovels in the ground by 2035.

France turned its whole grid nuclear in 15 years with 60s technology. The utter absurdity that has happened with nuclear energy is crazy. Its like the whole climate change disaster could have been avoided in the first place.

Nuclear is like anything else, if you build one bespoke reactor, its expensive. If you build 20 that are the same its gone end up cheap.

The problem is every country has its own government trying to build its own reactors and only have like one project going. Its like the rocket industry just 10x worse.

And the way regulation was managed after Three Mile Island was insane, and has utterly stopped nuclear innovation essentially impossible (specially new reactors) and makes nuclear building beyond expensive.

Its really a sad situation, humanity should be in the nuclear age right now. That we are still grappling with replacing coal power is beyond mind-blowing.

too true for most nuclear projects out there.

Well noted. Here's a paper: https://www.sciencedirect.com/science/article/abs/pii/S03605... The energy returned on invested, EROI, has been evaluated for typical power plants representing wind energy, photovoltaics, solar thermal, hydro, natural gas, biogas, coal and nuclear power. The strict exergy concept with no " primary energy weighting" , updated material databases, and updated technical procedures make it possible to directly compare the overall efficiency of those power plants on a uniform mathematical and physical basis. Pump storage systems, needed for solar and wind energy, have been included in the EROI so that the efficiency can be compared with an " unbuffered" scenario. The results show that nuclear, hydro, coal, and natural gas power systems (in this order) are one order of magnitude more effective than photovoltaics and wind power. =

Even if we assume that's correct, those sources have massive impacts under other categories. In contrast, anaerobic bacteria are terribly inefficient with their energy use, yet vastly outperform all animals in terms of biomass. So it's unclear to me that the EROI metric actually matters all that much. If I can produce solar panels with a solar-powered factory at a profit, I don't care about the price of coal or its EROI. I'm printing free money.

Of the 4 options you mention, I'm entirely a fan of hydro and nuclear. Especially nuclear. But FUD makes nuclear hard politically, so while I wouldn't campaign against it, I wouldn't waste my rep campaigning hard for it.

You can make while society falls around you. You want to make the most efficient use of time, energy, land, and materials. Why? So you can do more with it, give more people more time to think about things unrelated to survival, so you can support more people, whatever. Efficiency is key to civilization and progress. Solar takes 10X the land and 15x the material with no plan to deal with it after it has reached the end of life. Wind is worse.

I see what you're saying, but if you're arguing for advancement of civilization, then carbon should be priced accordingly.

End-of-life for solar is still a young field, but maturing rapidly. Regulations are already being formed in the US at a federal level. I guess gas plants can just close the well. But worst case,solar panels are small and can be buried (after we deal with the cadmium telluride)

I agree with end-of-life for wind though, not sure how it's gonna paly out.

What gets my goat is people don't bat an eye for granite countertops at home for 20-25k, but then solar roofs or battery storage is considered an expense.

Seriously? Trying to fearmonger what as dangerous crystaline silicon and fiberglass materials lying around? There are natural minerals more dangerous out there without disposal plans.

> Solar takes 10X the land and 15x the material with no plan to deal with it after it has reached the end of life.

Where solar in this case is photovoltaics/li-ion batteries and not CSP/molten salt?

edit: nvm see what they say here, not really impressed with a few of their assumptions of a optimal system, and using SEGS[0] is probably the worst one they could consider. Doesn't help that CSP tech is still in it's nascent stages compared to Nuclear,Hydro, Coal, and alot of system assumtioms for transmission are applied to CSP from them where CSP has other ways it can achieve even higher efficiencies with lower line losses (i.e. vacuum storage of molten salts and moving that to local grids to use in DSG systems) as well as being from 2013 (a lot more efficient systems have come online and global CSP capacity has doubled from 2013 to 2019).

"Only one work [30] has been found that provides values for invested energy and materials for different CSP technologies in a sufficient manner. The study which is part of a report from the German Centre for Aerospace (Deutsches Zentrum für Luft- und Raumfahrt) is a life cycle assessment for a hypothetical plant called “Sokrates”. Three different techniques were analyzed: Parabolic through with phenyl (SEGS) or steam (DSG) as coolant and a steamcooled Fresnel power plant, whose plane mirrors are roughly arranged to big parabolic mirrors. Plane mirrors are easier to manufacture and maintain, but have a lower concentration capability compared to parabolic mirrors, reducing the plant’s efficiency. Extremely high temperatures will reduce heat transportation, which also reduces the efficiency.

To achieve the high solar concentration, relatively big parts (mirrors) are necessary which are only usable in big plants. Therefore, due to economical aspects small or even individual plants are never considered. It should further be mentioned that, contrary to photovoltaics, the output of a CSP plant is not a linear function of the solar radiation intensity, so that only a deployment in sunbelt regions is in an economical scope. For regions with lower solar radiation like Germany this means additional power trans- portation energy demands and energy losses. The results shown in Table 4 are only for SEGS and Fresnel type technique, DSG is not tested yet. The location was assumed to be Ain Beni Mathar (Marocco, 34.17 N, 2.12 W) with a solar radiation constant of 2340 kWh/m2. The plant size is scaled to an annual output of 145 GWh (525 TJ), or 15,660 TJ over its adopted 30 year lifetime.

The higher demand for maintaining SEGS is caused by coolant losses due to the dominant phenyl energy inventory. The energy demand does not include the efforts for daily energy storage to provide electricity in the night hours, which is not possible with steam coolant as used by SEGS. The estimations for mirror replacements are very optimistic, doubling this rate will reduce the EROI by 20%, so the given values are the upper limit. Another significant reduction of roughly 30% (buffered) occurs when connecting this CSP plant to the European grid instead of using the output nearby the plant due to the very large copper demand.

It should be mentioned that the authors of the report [30] subtracted the phenyl maintaining demand from the output rather than adding it to the demand which can only be done if the used phenyls are directly produced by the CSP plant, on its site (see Sec. 5). This is not possible with the described CSP plant and would lead to wrong EROIs, making it infinite if all energy inputs are subtracted. Further corrections due to material inventory corrections lead to EROIs given in Table 4."[1]

[0] https://en.wikipedia.org/wiki/Solar_Energy_Generating_System...

[1] https://sci-hub.uno/10.1016/j.energy.2013.01.029

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