Electric cars: What do you think?

[essenmein]

@essenmein: You are missing the point. A computer is more expensive than a brick of the same mass. A nuclear power plant is more expensive than a dam of the same mass. Same idea. The difference comes from the complexity of the project. R&D split over hundreds (nuclear power) to billions (some computer parts) of components, integration of millions to billions of elements, and more expensive raw materials.

I guess let's agree to disagree, our costing group hasn't been too far off estimating price of parts this way so far.

 

[mfb]

Your costing group hasn't compared dams with nuclear power plants.

 

[essenmein]

Your costing group hasn't compared dams with nuclear power plants.

Quite right, I work in neither a dam building company or a nuclear power reactor company.

I do however work in a field that makes power plants at an industrial scale, these are actually fairly complicated vs the power they make. Collectively as an industry we have power plant costs down to about $12/kw (you might notice this is about 400x cheaper than projected SMR costs).

The whole final product with many bells and whistles not even remotely related directly to making power comes to a retail cost of about $250/kw, retail! incl profit margins for re-sellers! Yet still somehow 15x cheaper than SMR.

I would say complexity per kw here is higher here than for nuclear. Then the complexity argument is actually a bit disingenuous, complexity itself does not directly translate to costs, what you might consider complex, when viewed by a good process engineer, might turn out to be really simple to reproduce. Take the humble photograph, highly detailed and complicated images are possible, yet somehow a printing press can make a copy in a fraction of a second. Or by extension a micro processor, millions of transistors, highly complicated, yet built by what is essentially a fancy screen printer. Our PCBs have 1000+ components, takes a tech about a week to hand assemble one board, yet one instance of capital gets you a line that makes a new board every 26 seconds, 18hrs a day, 365days a year. The capital and operating cost of that line add about 60 cents cost to that board, vs 25EUR purchase parts cost (components, PCB, solder, line sundries).

I stand by the fact that done right, nuclear power should be the single cheapest form of power production if we stop being chicken and properly industrialize.

 

[member 656954]

I stand by the fact that done right

@essenmein, this seems more faith than fact and 'done right' is subjective. I'd say the nuclear power industry's safety record says we've 'done right' but others disagree, suggesting it is 'over done'. A utility shareholder might think 'done right' is no regulatory oversight, while a legislative body might think 'done right' means more regulation.

nuclear power should be the single cheapest form of power production if we stop being chicken and properly industrialize.

You've emotive terms here - 'chicken' and 'properly' - in an assertion, not an evidence-based statement. A number of studies have found no evident basis for SMRs, which are intentionally 'industrialized' nuclear fission designs, to radically drive down the cost of electricity production. I like Jim Green's overview of the current state because he provides a nice summary of many reviews with links to the underlying references and that's my pointer to 'many studies'.

 

[essenmein]

@essenmein, this seems more faith than fact and 'done right' is subjective. I'd say the nuclear power industry's safety record says we've 'done right' but others disagree, suggesting it is 'over done'. A utility shareholder might think 'done right' is no regulatory oversight, while a legislative body might think 'done right' means more regulation.

By done right I mean simply DFM and process engineering.

You've emotive terms here - 'chicken' and 'properly' - in an assertion, not an evidence-based statement. A number of studies have found no evident basis for SMRs, which are intentionally 'industrialized' nuclear fission designs, to radically drive down the cost of electricity production. I like Jim Green's overview of the current state because he provides a nice summary of many reviews with links to the underlying references and that's my pointer to 'many studies'.

that wiseinternational.org link might be a little biased? Have a look at the front page, can you tell? It calls itself an "antinuclear webstop!", not sure, but they may not like nuclear. Perhaps for emotive reasons?

Then Jim Green is an anti nuclear activist:
https://en.wikipedia.org/wiki/Jim_Green_(activist)

Hardly someone who would actually look for positive signs, instead, far more likely he will cherry pick things that support his position.

 

[essenmein]

He did link a paper though that has some interesting detail in it:

https://www.scirp.org/journal/PaperInformation.aspx?PaperID=45669

Basically SMR has projected higher fuel costs ~$5-6/MWhr vs ~$3-4/MWhr for large reactors due to lower thermal efficiency and neutron leakage.

I don't know anything about detail of nuclear reaction process design, ie what are the parameters that change as reactor size changes. But would love to know more about this. I suspect the neutron leakage issue is probably due to something like path length through the core, but this is pure speculation on my part.

I don't understand exactly why thermal efficiency would be lower though? Lower operating temperature -> turbine less efficient? Other system losses?

 

[member 656954]

Hardly someone who would actually look for positive signs, instead, far more likely he will cherry pick things that support his position.

Well, I'm not throwing stones for that, most of us have confirmation bias. But even with that, I struggled to find data driven counterarguments, and that is the basis for my 'faith not fact' observation.

Even Chinese reactors experience overruns and delays, and they can minimize regulatory and societal roadblocks that democratic countries would find hard to do. On the evidence, a new nuclear plant of any size is a leap of faith with regards capital investment and reactor startup time frame, and should be is not proven in practice.

 

[sophiecentaur]

Even Chinese reactors experience overruns and delays, and they can minimize regulatory and societal roadblocks that democratic countries would find hard to do. On the evidence, a new nuclear plant of any size is a leap of faith with regards capital investment and reactor startup time frame, and should be is not proven in practice.

I strongly recommend the Film 'Chernoble'. It shows what can go wrong with absolutely any system when the human factor takes over for individuals and self preservation or fear of the regime changes priorities and decisions are made which can have terrible unplanned consequences. Even the most intelligent of artificial intelligence systems will learn to make humanly fallible decisions, however good willed the designers are.

 

[essenmein]

Cost estimates never seem to consider the lifetime costs and don't include proper estimates of de-commissioning and safe waste disposal - or the cost of dealing with malfunctions (despite the low probability that's built in). On a different tack, I was 'amused' that the cost estimate for HS2 (High Speed Rail link proposal in UK) did not actually (they admitted) include dealing with the Ground Conditions. Also, I heard some clown (government mister) saying that he was 'convinced' that large electrical powered aircraft are virtually just round the corner. Where do these guys get their blind optimism from?

I can't quite find the link to it, but I thought nuclear did have to pay for decommissioning as part of the upfront cost. That could be wishful thinking on my part.

Not considering the ground is hilariously inept when building a rail line!

I think one of the main differences between what are essentially large one off projects, vs mass produced products, is that in mass produced products we have put in many iterations to figure out how to produce them quickly and cheaply. IMO a lot of that knowledge is transferable, or at least the concepts, maybe processes have to change due to different materials. The mass production mindset I don't think is there, SMR concepts to some extent look like baby big plants. We should scale reactors smaller so the development costs are lower and build a more prototypes, lots of prototypes, test different technologies, learn, iterate. Then mass produce. I don't know a reasonable small size, russia has 12MWe plants running, is 1MW feasible?

If you did say come up with a cost effective 10MW brick that you can mass produce, that does not automatically mean you disperse them through out the population (while I can totally see the benefit of that for local heat), still place them in controlled sites that maybe house thousands of them.

China has the right idea, mind you on a bigger scale:
“China wants to test all the fourth-generation concepts before moving forward,”
“It’s still very cost-effective at the developing phase. Once you move to the commercial use, there is no turning back.”
https://www.scmp.com/news/china/sci...-play-leading-role-developing-next-generation

 

[essenmein]

I know it would be a hilariously unpopular concept, but IMO for mass produced nuclear reactors, somewhere during development you need to test to failure... and test the capability of the containment and clean up. Do it a few times to practice... heh

 

[member 656954]

I can't quite find the link to it, but I thought nuclear did have to pay for decommissioning as part of the upfront cost. That could be wishful thinking on my part.

Not wishful thinking, @essenmein, there is an official decommissioning trust for this, and PwC has suggestions on the NRC guidelines for how / what utilities can expense for decommissioning.

The fund had $64.7B at end of 2018, but with estimates of reactor decommissioning running at around $2B - $4B per plant ($3.9 billion for Diablo Canyon, $2.3B for Indian Point, £1.2B for Berkeley in the UK) and almost 100 operational reactors in the USA, the NRC cost estimate for decommissioning a nuclear power plant in the $280M - $612M range shows that the fund does not cover all costs.

Similarly, the World Nuclear Association decommissioning suggestion is 9-15% of the initial capital cost but the Indian Point reactors were about $3B to build (in current money) so that estimate seems low.

Interestingly, Japan allocated $7.7B to decommission 17 reactors at 9 plants, which if they can achieve, is more in line with the NRC costs.

I know it would be a hilariously unpopular concept, but IMO for mass produced nuclear reactors, somewhere during development you need to test to failure... and test the capability of the containment and clean up. Do it a few times to practice... heh

Hilarious! A nuclear plant failure (disaster) scenario costs a lot more. This summary of cost reviews for the impact of Chernobyl has tens and even hundreds of $B attributed against various health consequences. Fukushima's clean up costs estimate was revised up by the Japanese Government to around $200B. Three Mile Island was cheap by comparison, only $1B in early 1990's money.

 

[Rive]

Hilarious!

Not really. Just because it has 'nuclear' in it, it does not mean that you actually has to do a nuclear meltdown.

 

[essenmein]

Hilarious! A nuclear plant failure (disaster) scenario costs a lot more. This summary of cost reviews for the impact of Chernobyl has tens and even hundreds of $B attributed against various health consequences. Fukushima's clean up costs estimate was revised up by the Japanese Government to around $200B. Three Mile Island was cheap by comparison, only $1B in early 1990's money.

Its actually precisely this reason it has to be done. Look at any of the other products we build, they start of hilariously unsafe, till we have accidents to learn from. Modern product development has learned from those early miss steps, for automotive for example there is a massive amount of life time testing, test to failure to understand the mechanisms and failure rates. We crash hundreds of cars to make sure occupants are safe under a kaleidoscope of conditions. Its by far much cheaper to test out these failure modes in a controlled environment than have it happen in the field.

As an example we had a prototype 40kw inverter failure on a test bench, where the resulting electric arc was maintained by the DC rail and propagated through several power modules before the supply was shut down, this vomited molten copper and ceramic everywhere and shot 3ft of electric fire out of the thing (funny what a 80kW DC source will do). Needless to say, this type of failure mode is somewhat frowned upon in a car! The design was changed to mitigate this arc propagation, and other systems were implemented to prevent this from happening at all. We then repeatedly tried to recreate this failure mode till we were satisfied it no longer existed.

There has to be the pragmatic recognition that 100% safe is not possible. So given that there will be failures, its much better to know how it fails, and then design systems so that a failure is as inconsequential as possible.

 

[member 656954]

Its actually precisely this reason it has to be done. Look at any of the other products we build, they start of hilariously unsafe, till we have accidents to learn from.

I can only agree, noting that the consequence of serious failure with a Gen III/III+ nuclear reactor is at the extreme end of the scale, even if the likelihood is historically low. That will probably change if Gen IV SMRs are deployed in volume because while the consequence of serious failure may reduce due to their design, the likelihood goes up because there are more failure opportunities from having more units deployed.

Of course, this is supposition as we don't have much in the way to Gen IV designs to test against, but if history is a guide, the first serious accident or failure leading to death will obliterate trust in their design and we'll once again enter a nuclear sales winter as projects are indefinitely delayed or abandoned.

 

[member 656954]

0.3 to 0.6 cent/kWh assuming 80% average load of a 2 GW power plant over 50 years.

Yes, that's about the right percentage. But I'm not sure the point. Decommissioning a nuclear plant costs a bundle, no matter which metric you use to calculate it. Are you suggesting it's a reasonable equation, @mfb?

 

[member 656954]

0.23 cent/kWh distributed over ~10 TW*years, the approximate total electricity production from nuclear power so far. Let's round it up to 0.3 cent/kWh for western reactors. Other accidents came with negligible cost in comparison.

I'll admit I'm still struggling with this. If a plane crashes do we calculate the cost of all flights over the entire model of aircraft and say, "Only 7c per mile flown, that's not so bad"? Are you saying Fukushima's clean up costs make the other nuclear disasters good value?

 

[essenmein]

I'll admit I'm still struggling with this. If a plane crashes do we calculate the cost of all flights over the entire model of aircraft and say, "Only 7c per mile flown, that's not so bad"? Are you saying Fukushima's clean up costs make the other nuclear disasters good value?

Not sure why you struggle with this? Regarding the flight cost, you can bet that airlines and insurance companies are precisely making those calculations.

Its simply (benefit to society)/(cost to society). Improving this ratio is basically what our overarching goal is.

Then, we pre-pay decommissioning costs for nuclear, but for some reason, wind and solar get a pass, yet we are creating a serious environmental problem that people are just waking up to. If we start installing grid scale electrochemical cells needed to make renewable feasible, that will only make things far worse.

"Industry leader Toshiba Environmental Solutions can currently handle 44 tons of solar panel waste a month. It would take 19 years for the company to process even the 10,000 tons of waste expected in 2020. "

https://asia.nikkei.com/Business/Bi...on-ways-to-recycle-a-mountain-of-solar-panels

 

[anorlunda]

I made a major clean up of an off-topic subthread.

Actually, much of this thread drifted far from the OP ev topic. Rather than delete almost all the posts or attempt to split it into multiple independent threads, I'm going to close this thread.