Hacker News Reader

#3 is the big one for me. Solar and wind can't be simultaneously "the cheapest form of energy available" and yet also require grid-scale battery technology that doesn't exist--and even if it did exist, would not be economical--and require massive government subsidies (read: taxes stolen from people to fund boondoggles). Why is it the externalities are always ignored by proponents of wind/solar but then discussing other forms of power are nitpicked to death about said externalities? It's almost like wind/solar fanatics hold religious opinions.

> compare that with the lifepsan of 80-90 years that nuclear often attain and you see that the cost that people usually refer for solar panel must be doubled.

One thing I am wary about is what the lifespan of any newly build plant would be that uses modern technology.

If you take a random 8bit home computer from 1983, all the non-mechanical parts on it likely still work today.

If you take a random computer from 1993, well most bits probably work.

If you take a random computer from 2003, if you are lucky some things work.

2013? Odds are it won't work.

The differences in manufacturing between everyday consumer electronics and high reliability electronics has grown larger and larger, for two reasons.

One is progress of Moore's law, smaller feature sizes have made electronics more susceptible to many different types of failure[1][2].

The second is RoHS[3], removing lead from solder has basically given all consumer electronics a 10 year life span.

Now of course you can get non-RoHS components, but they are not the norm, and to make things reliable you need to ensure everything is not RoHS.

Want a reliable storage medium? That is going to cost an insane multiple. Want a reliable computer? Another insane multiple, you are going to need to have something built on an ancient manufacturing process, and because you aren't getting the benefits of economies of scale you are paying more, and you are basically getting a bespoke machine so you now need to test it to heck and back to ensure reliability.

One problem with how we have chosen to advance our electronics industry is that the types of goods that are made en-masse have steadily separated from what types of goods need to be made for long term engineering projects.

An example of this in action: when Tesla first came out they realized existing automotive infotainment systems basically sucked and that they couldn't provide a good UX with anything available off the shelf at the time, so they instead plopped new consumer non-automotive gear. A touch screen that didn't suck! But it also melted. Everyone yelled at them "should have known better!" but their choices were "unreliable consumer" or "crappy behind the scenes".

Another example, RoHS doesn't have exclusions for home appliances. RoHS is also why the control board on washing machines gives out after ~10 years, necessitating purchasing a new washing machine even though the mechanics may have another decade left in them! (RoHS needs a crap ton more exclusions for long lived appliances...)

[1] https://en.wikipedia.org/wiki/Reliability_(semiconductor)#Fa... [2] https://en.wikipedia.org/wiki/Electromigration [3] https://en.wikipedia.org/wiki/Restriction_of_Hazardous_Subst...

That's a lot of FUD about RoHS - especially for the consumer space. Don't you think that maybe everything built to be as cheap as possible to barely survive the warranty period is maybe a slightly larger effect than tin whiskers? Also old appliances typically contained less things that can break in the first place: What doesn't exist cannot break.

It's just a function of cost, do you want to engineer your device/machine to last or not, and are more features more important to you than reliability.

> Don't you think that maybe everything built to be as cheap as possible to barely survive the warranty period is maybe a slightly larger effect than tin whiskers?

YouTube recently recommended me some tear down videos of washing machines, for whatever reason. The video compared 1990s to modern machines.

The insides are remarkably similar, the main failure point is the electronics, flat out.

And in my experience (n=1), what fails on modern appliances is the electronics! Not the pumps or motors, sure sometimes those go out, but most often it is the control board.

Look at every car built since 2010. Sure you have some models with known manufacturing defects on a given part, but replace that one part and mechanically most cars are reliable. But how many cars have you come across that run just fine, but the check engine light is stuck on and the mechanic is like "well I could fix it but it will cost a lot"?

We have hundreds of years of knowledge about making reliable machines in massive quantities, and then cost reducing them (ok maybe 200 years tops), as a species we haven't even bothered trying to cost reduce reliable electronics! We just accept that they will fail.

I also had my fair share of repairing home appliances (trying to anyway) and in my N~5 sample size I was not only limited to failed control boards. But having seen those control boards I am honestly surprised that they even survive the warranty period. They are that cheaply and badly made. From the circuit board itself to the components used. And of course not designed to be repairable at all. They have more in common with a cheap calculator than something that controls kWs of power. But allowing the manufacturer to use leaded solder would have made zero difference in their reliability.

People absolutely can build reliable electronics these days, but it will be slightly more expensive than not quite so reliable electronics. Higher profit margin has a higher priority than making sure the product survives long after the warranty period has ended. Especially if the consumer only looks at the initial price without taking into consideration how long it will last. Which is also hard to do without reviewers tearing the machine to bits. Being expensive does not mean it is actually any good and will last a long time.

The changes that drive the reduction in lifespan of modern electronics is a very good topic, however it is an obvious issue that is taken into account in nuclear designs. The latest VVER-TOI has a life expectancy of 60 years, upgradeable to 100 years. It is possible that they have to change the CPUs every 10 years as routine maintenance though. I don't think they are full analog nowadays.

> How come people do not see what I see?

I guess that's because you are missing part of the picture.

> For starters the economical costs are accidental bureaucratic contingencies.

Yes, everything says that it's possible to create safer nuclear reactors. But our current ones aren't. Safe reactors are on the same level of "nobody does this because of market and bureaucratic constraints" as cheap open-storage fuel cells. Except that the constraints are harder to change for the reactors.

We should be trying very hard to change both. But until we do, the reality will be that both nuclear reactors and batteries are expensive. And our current reactors are only safe because of the safeguards put on them, remove the safeguards, and they won't be safe anymore.

> renewable is increasing the cost of nuclear operation (underusage of maximal throughput) for nor reason

That's a pretty bad way to frame it. The problem here is that all non-dipatchable (on the economical sense, not the technical one) sources compete to get full utilization. That's not a problem when most of the power is dispatchable, but none of the options you are considering (renewables - except for buifuels - or nuclear) are. Adding nuclear capacity will increase the nuclear operational cost even more than renewables (because they have the same generation profile).

The only solution here is storage.

> solar panel have a short lifespan (20-30 years) with diminishing returns of efficiency ... the cost that people usually refer for solar panel must be doubled.

You noticed that you took the manufacturer guarantee and cited it as expected lifetime, right? AFAIK, we don't have a good number for how long our current tech of solar panels will last, because after just a few decades on the field, nobody catalogued enough failures to measure it. But we know it's way longer than that.

Anyway, doubling the cost of solar panels is both the wrong way to go (interest rates exist), and changes about nothing when comparing with the current costs of nuclear. (It would make them lose to coal, but then, both solar and coal are normally built by private business with proper accounting in practice to compare their costs.)

> Solar and wind economics are absurd

Well, you clearly didn't dig into those. Storage seems to be viable even with current tech (not using batteries). But it will of course always lose to natural gas generation.

Anyway, natural gas generation at night is a pretty good stop-gap.

> not even talking about the immense stress that would have on lithium resources which are seriously needed for the future of mankind

I don't expect grid storage to use lithium at all, but the fact that you have gone over lithium¹ reserves and are ignoring uranium reserves shows a very biased research. TLDR, none will be a problem soon (or probably ever for lithium).

> In the end, once renewable reach a major percentage of energy and heat production, the use of gaz and fuel as pilotable energies will skyrocket making the point moot.

Well, eventually we will need storage. That's true for nuclear too.

But anyway, renewables don't increase the consumption of fossil fuel based sources in any way, so the only way they can push the prices higher is by reducing economies of scale, what is actually the goal.

1 - Of any element you could choose, you decided to go with lithium, one of the most available ones on the Earth's surface? But well, once in a while somebody pops up talking about iron or silicon shortages, so you could have done worse.

> Yes, everything says that it's possible to create safer nuclear reactors. But our current ones aren't.

1) extraordinary claim 2) completely miss the point that the 10000 bureaucratic safety rules could be trimed by half while actually improving security and reducing costs. 3) completely miss the point that current models from Russia and China are cost effective and comply with aforementioned unoptimized rules.

> remove the safeguards, and they won't be safe anymore

most of the reactors from the 70s and 80s are still in use. they have much less safeguards and are still safe enough as in you'd better invest the saved money in saving lifes/medecine research than in those hypothethical safety gains that only drive marketing and hypocrisy.

> Adding nuclear capacity will increase the nuclear operational cost even more than renewables (because they have the same generation profile).

completely miss the point that not developing renewable and going for 100% nuclear increase overall economics. completely miss the point that nuclear plants can contrary to renewable decrease their output in a very fine grained way. Why is that interesting? The plant cost almost the same, but does not add a significant cost by overproducing electricity. When renewable overproduce electricity (which happens all the time given enough share) the excess electricity will break the electric grid and cause a nationwide shutdown. To prevent that, the Grid operator has to sell electricity at a loss (AKA pay people to consume the electricity). Given some level this startegy won't even work and will need a special costly infrastructure optimized for burning (throwing away) excess electricity.

> But we know it's way longer than that

who is we? Where is the evidence? Also newer reactors have a lifespan extendable to 100 years hence you need to triple or quadruple the price. And wind, which is generally much more prevalent than solar last in most cases 20 years. In 2016 they have started (in germany) installing twice as big turbines which are likely to have shorter lifespans.

> The only solution here is storage. > Well, eventually we will need storage. That's true for nuclear too.

nuclear doesn't need energy storage.

> Storage seems to be viable even with current tech (not using batteries)

please source, enligthen me, I am unable to find reliable information supporting this.

> Anyway, natural gas generation at night is a pretty good stop-gap

Please let's not be a joke, the discussion is about a long term fossil-free future. Not a thesis that only works until it doesn't.

> grid storage to use lithium at all

wtf, energy storage without batteries is niche. What are you refering to? Water, hydrogen? The majority should be with electric batteries AKA lithium. This should be obvious.

> renewables don't increase the consumption of fossil fuel based sources in any way

Renewable needs pilotable energies. Solar in january and december is almost non existent. How much is an empirical and country dependent question but bewteen 20-30% average share of fossil energy needed seems probable. Meanwhile 100% nuclear is perfectly doable. 70% renewable 30% nuclear is also doable but would be economically inefficient.

> both solar and coal are normally built by private business with proper accounting in practice to compare their costs

Wrong assumption, solar and wind in germany are extensively subsidized.

Anyway the most salient point about this discussion is the storage energy cost and feasability, of which you provide zero data.

regarding lithium availability: https://www.reddit.com/r/AskScienceDiscussion/comments/q78xv... Enough to build 8 billions cars and then zero cars for the next generation LOL. Now add that in addition of one car per habitant we need a similar or even bigger battery for his regular electricity uses, including house heating.

Doing policies that last for just the current generation and not the next centuries is madness. Yes we've all heard there are HUGE reserves in the oceans. Which pointless to say. There are a looot of gold in the oceans too. The density and extractability of it is the key question. The fact this is not currently used means it cost more than current methods. How much more? x10 ? x100? x10000? Who knows? Sharing technical evidence about this issues would improve the depth of the discussion but until then we are manipulating a huge existential risk.

about uranium reserves: right, the current estimate of reserves is of a 230 year supply at current consumption. The needed consumption for 100% would be at least a x10 so let's say we ran out of uranium in 23 years if massively adopted. 1) A major point is that contrary to lithium, the mining industry of uranium is much smaller and therefore if it received mass funding, the reserves might significantly increase. 2) like lithium there are huge oceanic reserves and like lithium, this is not a solution until proven otherwise. 3) reactors that reuse MOX fuel (like e.g. France does) can significantly (how much ?) reduce uranium consumption. Note that lithium can be recycled too but setting a global systematic recycling industry is yet to be done. 4) Indeed thorium reactors (not uranium) are the ideal solution since we have thorium for the next millenium. Thoriums reactors are not just a theoretical concepts, some have been made so it is feasable. It has not been developed because the energy efficiency/economics are inferior (how much ?) although since most of the cost is on the time to build the plant anyway, this should still be a viable solution although not necessarily as cost effective as renewable here (but with an expensive energy storage grid and pragmatic designs (like russian/chinese ones) and taking into account lifespan, and renewable need for fossil fuels/gaz, it very well could be) As for the ocean argument it seems easier to solve for uranium though which is needed in much less quantity (because of its ernergy density) and hence needs less regional density. Note however that water filtration methods for lithium are often similar to existing desalinisation plants. Which ironically are a great specialization for nuclear power plants.

> 1) extraordinary claim

So... The claim safeguards created after major accidents happened are necessary for safety is extraordinary?

They can probably be improved, but then, you keep comparing with Russia and China as if those two countries cared about the safety of their population.

(Anyway, there was a recent paper here on HN that looked into the issue and discovered the costs are more of a consequence of the small size of the industry than the safety constraints. That kind of problem is hard to fix, but if you want to try, it's a worthy cause.)

> completely miss the point that not developing renewable and going for 100% nuclear increase overall economics.

You seem willfully ignorant of how power generation economics work. Or are you talking about the opportunity costs of choosing the cheapest option?

>> But we know it's way longer than that

>who is we? Where is the evidence?

Hum... The first Google result I get seems to be using this data:

https://www.nrel.gov/docs/fy12osti/51664.pdf

TLDR, the worst kind of panel they measured lost less than 2% of it's capacity per year, so about half of its capacity in 50 years. Some slightly more expensive ones lost less than 1% per year.

> > Storage seems to be viable even with current tech (not using batteries)

> please source

Oh, there's a comment on this thread about how hydro-storage can hold about 10 times the requirements for the world. There was one recently about how hydrogen is perfectly sufficiently, as flawed as it is. And of curse, there is always some weird design appearing once in a while, that nobody invests in, of course, because nobody has ever invested on energy storage.

> The majority should be with electric batteries AKA lithium.

What a lack of principled thinking, this equating batteries with lithium. Lithium isn't that great for stationary storage in any way other than the batteries factories already existing.

> Renewable needs pilotable energies.

Not more of them than we are using now. Anyway, no, 100% nuclear doesn't work well either, it would be incredibly expensive. None work without storage. (By the way, you seem to have an incorrect model of why renewables don't decrease their production when they overproduce. Most of them can do it perfectly well if the overproduction becomes too severe, it just brings costs.)

> regarding lithium availability

And then we go into the "what is the meaning of that "reserves" word" treadmill... You may want to look into it in detail, as both lithium and uranium have a problem with that word. None has a problem of availability.

Anyway, about sea extraction, people have done it economically for lithium, but for uranium it's many orders of magnitude harder.

Also, about this:

> reactors that reuse MOX fuel (like e.g. France does) can significantly (how much ?) reduce uranium consumption.

In theory, recycling can reduce consumption around 1000 times (varies with every detail on the lifecycle). Breeding can do much more (that's why thorium is interesting).

Except, thorium has been tried already. It costs more. Cost is already not nukes' strong suit.

Yep, thorium brings several engineering problems. Uranium breeding brings a lot of them too.

Overall, breeding reactors are currently way too expensive. They are interesting on the long term, once those solvable problems get solved. I do agree they have no place at all in a discussion about current energy crisis or global warming.