The Ergosphere
Tuesday, November 07, 2017

Don't mess with the engineers when physics is on the line

This is an open letter to Brandon Schollenberger, also sent by e-mail.

I received a pointer to your essay.  No, lying is not OK.  Neither is spouting off in utter ignorance.

TL;DR  You're not right.  You're not even wrong.  You wrote a mish-mash of conceptual nonsense that is cringeworthy to every student who scraped out a passing grade in AP Physics, and quite a few who never made it that far.  I'd say take it down NOW, but... it's already too late for you; someone archived it 2 days ago.  The web never forgets.

At its core, you do not understand the difference between energy and power.  Gallons of gasoline are a measure of energy.  Horsepower are a measure of power.  You can burn 5 gallons of gasoline making 5 horsepower for 20 hours, or making 100 horsepower for 1 hour.  It's vastly different amounts of power, but the same amount of energy.

On to specifics.  You wrote:
To understand what the lie is, you need to know the purpose of a pumped-storage hydroelectric (PHS) power station. Like any power station, a PHS station produces electricity for consumers. It does so by converting the kinetic energy of flowing water into electricity. However, it has an additional purpose as indicated by the phrase "pumped storage."
Perhaps if you had looked up the Wikipedia article about pumped hydro storage sites, you would find that there are precisely 9 of them in the USA with a total nameplate capacity of 13,612 megawatts.  (This will go up by a few hundred MW as Ludington is upgraded with more advanced and efficient pump-turbines.)  This is not what they can store (equivalent to gallons of gasoline); it's their maximum instantaneous output (equivalent to engine horsepower).  Average US grid demand is on the order of 450,000 MW.  All the PHS plants in the USA, running flat-out, can serve roughly 1.5% of average US grid consumption.

Obviously, there's a lot more than 13,612 MW of total hydro capacity on the US grid (Jacobson lists 87,480 MW).  Just as obviously, the vast majority of it is NOT pumped hydro.  The only energy they can store is the energy delivered by the rainfall feeding into their reservoirs.  This is limited by many factors, including minimum river flows for ecological reasons.  One thing they can't do is reach down and pull water back out of their outflowing rivers to store energy again.

When more electricity is generated in the electrical grid than is necessary, it needs to be disposed off somehow. It can be burnt off, but a better solution is to find a way to store that electricity until it is needed. In a PHS, electricity is stored by using the extra electricity to pump water into a reservoir at a higher elevation. Later, that water can be released to produce electricity. It is basically a battery you can charge when you have extra power and discharge when you need more power.
You can't generate more power than the grid needs, not for a significant time or fraction.  The reasons why involve BSEE-level mathematics which you obviously don't have, but the point is that there are NO significant stores of energy in the grid proper except for the sheer mechanical inertia of its large synchronous rotating machines (both generators and motors).  If you pump in power over consumption, those machines speed up past their rated speeds; a power deficit causes an underspeed.  Too much of a deviation trips generating plants off the grid and causes a blackout.  Generation must match load to a very high degree instantanously, and even more closely over time.

PHS systems can function as both generators and loads, but... 1.5% of average grid consumption.  That's a handy ±1.5%, but it's still only 1.5%.

If a power station produces 100 GW every hour, no more, no less, would we say it is impossible for it to output 1,000 GW in a single hour? I would hope not. If the station's electricity wasn't needed for 10 hours, it might store up 1,000 GW.
Aside from impoundment-fed hydro stations (both pumped and otherwise), the only stores of energy in powerplants on the grid are:
  1. Coal stockpiles at coal-fired plants.
  2. Oil tanks at oil-fired and dual-fueled steam and gas-turbine plants.
  3. Uranium in the cores of nuclear power plants (by far the largest of all).
Nothing else stores significant energy.  A steam-turbine powerplant can cycle several million pounds of steam per hour from boiler to turbine to condenser and back to boiler.  There's nowhere in the plant to store millions or even thousands of pounds of steam, and the turbine can only accept it so fast.  The alternator which converts the turbine's output to electricity has its own instantaneous and sustained power limits, as do the transformers which put it out to the grid and the wires themselves.  It doesn't matter HOW many gallons-equivalent of gasoline you've got, you can only USE it as fast as the horsepower of your engine.

I will stop rubbing your face in your embarassment here.  I hope you have learned a lesson. 
The main thing Brandon did was butcher basic terminology, falsely equating nameplate capacity with average capacity. This allowed him to think that a PHS facility with a nameplate capacity of 100 MW, can magically discharge at a rate of 1000 MW as needed - because the 100 MW is just the 'average.' Rod's car analogy was fantastically crafted and worded, in that a 100 Horse Power car cannot temporarily summon 1000 HP. Additionally, other commentators explained this to the blog author, multiple times, in extraordinary detail - he just brushed aside any explanation. I am doubtful your letter will provide any further revelation.

Specifics of PHS generation output are unclear and are discussed little. As water is discharged and water volume/height is reduced - so is the discharge rate. Using Ludington specs as an example, what is the discharge rate for the 1st hour when the reservoir is full, and what is the discharge rate during the last hour when the reservoir is nearly drained? Just curious how much more water would be needed on hour 1 vs. hour 13 during 13 hours of constant generation of 1,300 GW, as modeled by Fakeobson.
"Fakeobson".... that's going to leave a mark!

Schollenberger wrote back to me twice, the second time after I opined that he'd never passed even an intro physics course and told him that nature has the only vote that counts.  He did not contradict me on the former point, and told me that writing letters to him is pointless.

He's an idiot, and I'm going to take him at his word.
Thanks for writing this article. Reminds me somewhat of this one in which capacity may get mixed up with actual generation, but the claim here is that IEA counts primary data for fossil fuels 3 to 1 compared to wind and solar and therefore, he seems to claim intermittent renewables are getting short-changes, I suppose. If you have time and could provide commentary, that would be appreciated. If you're on twitter
Sorry, not on Twatter.  Never liked the idea and it's pointless now that they're shadowbanning and deleting accounts left and right.
Fair enough about twitter, it was just a suggestion or option if you wanted to see some comments from the author. Any commentary on the original article linked to in my other comment (I suspect a similar author like Brenden)? And this IEA commentary may counter it?
I sympathize slightly with energiogklima (the IEA and also UK energy figures for electricity are given in MTOE which is totally opaque and I have never found what conversion factor they use) but they're outright lying themselves.  They're trying to spin things the other way, counting e.g. solar input to a 20%-efficient CSP plant the same as the NG input to a 60%-efficient CCGT plant.

IOW, a pox on both their houses.
Hi EP - completely off topic but I couldn't find a comment I think you made recently about pool reactors for district heating. I was wondering whether spent fuel could be used for this ? Your swimming pool reactors sound very similar to a spent fuel pool. The DUPIC concept involved the use of spent PWR fuel in Candu reactors, as it would still contain a higher percentage of fissile than the natural uranium which can be used in heavy water reactors. However, DUPIC still involved decladding the highly radioactive fuel rods, grinding up the 'meat', and recladding them in the short, shuffle-ready assemblies used in Candus. My idea was just to put them straight into one of the swimming-pool type reactors you mentioned, as is, except that the water directly round the fuel assemblies, and up to the printed circuit heat exchangers, would be heavy water. Diaphragms or free pistons would equalise the pressure of the D2O with the ten metres or so of H2O above it, and the total depth could be changed gradually to control the boiling rate, and so the rate of fission. ( A Chinese paper I looked at on swimming pool reactors claimed that printed circuit heat exchangers can now control the temperature of the product within two degrees C - I'm not sure what allowance would have to be made for heavy water boiling at 101 C at standard pressure.) Maybe the fuel assemblies could be put in upside down, since in the PWR they would have been under-moderated at the top by the hotter water flowing past them, compared to the bottom of the rods. The fuel assemblies would have to be checked for any leakage before use, but operating at only 100 C or so, instead of the 300 C plus in a PWR, would not put much stress on them, and the heating season only lasts for about four months, compared to the five years they'd already spent cooking. Any leak into the primary circuit would soon starve the chain reaction of neutrons - as an extra safety measure, you could have a neutron poison dissolved in the main pool. The first couple of months of fission product heat would become an advantage, not a liability. PWR refueling is often done during autumn, when power demand is low, so the used assemblies could be taken straight to the heating plant. I can see North Americans objecting to freshly used fuel being moved to the outskirts of their cities, but the Chinese would probably be happy to get their district heating without the smog.
John, I think that's a fascinating notion you've got there but there do appear to be technical gotchas.  Note that IANAnuclearengineer so there are probably more than what my poor knowledge leads me to.

One technical gotcha that comes to me right off the top of my head is that fuel lifetime is limited by fuel swelling.  If you burn the fuel too much, pellet swelling might rupture the cladding and leave you with a nastily contaminated primary loop.  This is why the burnup limits are burnup limits.

As for trying to pressure-balance a D2O moderator inside an H2O pool without physical barriers... turbulent mixing is going to homogenize that in no time.  Simply don't bother.  The whole primary circuit should be D2O.  What kind of restrictions the NRC has on D2O I don't know, and I know the stuff is anything but cheap, so that's going to hit you with a whole bunch of costs for financing and regulatory compliance which might well make "free" fuel too costly to think about.  However, I'd say to work the numbers and see what you get.

Note that the boiling point of H2O under 1 bar gauge pressure is about 120°C.  It'll be slightly higher for D2O.  I've seen ΔT numbers for PCHE's in the 5°C range, so that seems reasonable; you could still get output temperatures of 100°C, maybe 110.

To avoid cross-contamination of the water loops, an intermediate loop is required.  A heat pipe between PCHEs would do.  Something like n-butane might do as the working fluid, having enough pressure to keep water out at either end and being immiscible with water.  This also gives you 3 walls between the fuel pellets and anything that leaves the plant.  Hmmm, butane might also be a good working fluid for the engines which power the circulating pumps.  That's definitely worth looking at.

You should do a back-of-the-envelope analysis of the potential of this idea.  I'd suggest digging up some numbers on total US spent fuel inventory, picking some nice conservative figure for the power density that this elderly fuel can sustain, and comparing the total power output with natural gas consumption for heating.  Feel free to post here.

Note that D2O production is typically done by exchange of hydrogen between water and H2S... which is driven thermally.  If the production plant is cheap enough (I know the Canadians had fits getting theirs to work), the winter heating plant might be employed in the production of D2O in the off-season.

On the regulatory side, if this thing requires full-time on-site NRC inspectors it's a dead letter.  It would have to be designed to operate as a black box that isn't opened except to change fuel.  Given the number of ways things can go wrong, that's going to be a tall order.
Your probably right that it's a non-starter, even in China, where they're actually building lots of reactors, lots of district-heated apartments, and have a lot of smokey coal to replace. I didn't propose any mixing of the heavy water with the top of the pool - the light water would only be there as radiation shield, heat sink and fail-safe pressure control. The heat exchangers would have to be far enough above the fuel to avoid neutron activation of the secondary circuit water. If that was deemed not safe enough, you could have a second set of heat exchangers at the top of the pool. China has roughly thirty gigawatts of PWRs. About a third of the fuel will be changed out every eighteen months, and the thermal efficiency is about a third, so if the spent fuel could still put out power at the same rate, you'd have about thirty gigawatts of heat. The idea though was to run them at very low power, to eke a bit more use out of them without too much risk of bursting open the fuel rods - I'd heard that with current fuel technology, breaches are very rare. However, looking at it again, I see the inlet temperature of the calandria water in Candus is only 70 C, so heavy water at 120 C wouldn't moderate as well. The fuel bundles in Candus are also much further apart than the fuel assemblies in a PWR - since heavy water is about ten times worse at slowing neutrons than light ( though a thousand times better at not absorbing them ) there has to be more room for that to happen. That would lead to very uneven burnup. So you'd probably have to take the spent tubes out of their assemblies, at least, and probably reclad then somehow as well. By this time, as you say, your free fuel is starting to look a bit pricey.
"it's a non-starter, even in China, where they're actually building lots of reactors, lots of district-heated apartments, and have a lot of smokey coal to replace."

We'll see if they actually do it.  But that's different from your LWR-fuel-reuse proposal.

"I didn't propose any mixing of the heavy water with the top of the pool - the light water would only be there as radiation shield, heat sink and fail-safe pressure control."

That wasn't clear to me immediately, and it does make the idea far more practicable.

"The heat exchangers would have to be far enough above the fuel to avoid neutron activation of the secondary circuit water."

I'm not sure how much of an issue that is outside of core-level neutron flux, and activation of light water is a remarkably innocuous thing.  That is part of why I proposed n-butane as the intermediate heat transfer fluid.  It would work well in a heat pipe at those temperatures (very low ΔT given the phase-change nature of heat transfer) and any hydrogen that was activated would be very unlikely to get into the tertiary loop.

"If that was deemed not safe enough, you could have a second set of heat exchangers at the top of the pool."

That's precisely what I proposed with the PCHRs connected by the n-butane vapor loop (with the n-butane also functioning as the working fluid for the circulation pumps).

"China has roughly thirty gigawatts of PWRs. About a third of the fuel will be changed out every eighteen months, and the thermal efficiency is about a third, so if the spent fuel could still put out power at the same rate, you'd have about thirty gigawatts of heat."

Math fail.  That would be 90 GW of heat.

TBH, you couldn't even approach the same heat production that you could get under PWR conditions.  A barely-pressurized HWR would be under boiling conditions at a much lower volumetric heat power, and boiling causes repeated thermal cycling of cladding due to the production and separation of steam bubbles.  I am not a nuclear engineer and don't know the full details of the difference between PWR and BWR fuel, but I know the cladding of the latter is much more robust and I suspect that thermal cycling under boiling conditions is a large part of the reasons why.

"The idea though was to run them at very low power, to eke a bit more use out of them without too much risk of bursting open the fuel rods"

As I said, "some nice conservative figure for the power density that this elderly fuel can sustain".  Let me show you how this is done, since you don't seem to grasp it intuitively.

Typical figures for LWR fuel burnup are given in the neighborhood of 35000-45000 megawatt-days per metric ton heavy metal.  Figuring 3 17-month cycles before replacement and 30.5 days per month, this comes to an average thermal output power between 937 kW(th)/ton and 1.205 MW(th)/ton (excluding oxygen, cladding, etc.).

Figure that you might be able to get 10% of this power level in unpressurized HWR operation, and maybe another 10,000 MW-days per ton before something requires that the fuel be retired permanently or reprocessed.  10k MW-days/ton at 93.7-120.5 kW/ton is 22.7-29.2 YEARS of full-power operation before exhaustion.  This is close to eternity in political terms.  Note that if heavy water isn't viable as a moderator/coolant due to thermal cycling or corrosion, graphite works perfectly well as a moderator and CO2 as a coolant.

The more interesting figure is what we could get out of our existing SNF inventory.  If memory serves, the US has on the order of 80,000 tons currently being held.  At those figures, the available thermal power is somewhere between 7.5 and 9.6 GW thermal power.  This is not very much on the national level, but it's enough to heat all of NYC and several other cities without further assistance.  That should be enough to get the attention of people who matter.  And of course, we add considerably to this available power every year.
Post a Comment

Links to this post:

Create a Link

<< Home
Talk largely about energy and work, but also politics and other random thoughts

Mail Engineer-Poet

(If you're mailing a question, is it already in the FAQ?)

Important links

The Reference Library

Blogchild of

Armed and Dangerous

Blogparent of


The best prospect for our energy future:
Flibe Energy

January 1990 / February 2004 / March 2004 / June 2004 / July 2004 / August 2004 / September 2004 / October 2004 / November 2004 / December 2004 / January 2005 / February 2005 / March 2005 / April 2005 / May 2005 / June 2005 / July 2005 / August 2005 / September 2005 / October 2005 / November 2005 / December 2005 / January 2006 / February 2006 / March 2006 / April 2006 / May 2006 / June 2006 / July 2006 / August 2006 / September 2006 / October 2006 / November 2006 / December 2006 / January 2007 / February 2007 / March 2007 / April 2007 / December 2007 / January 2008 / May 2008 / June 2008 / August 2008 / October 2008 / November 2008 / December 2008 / February 2009 / March 2009 / April 2009 / May 2009 / June 2009 / July 2009 / August 2009 / September 2009 / October 2009 / November 2009 / December 2009 / January 2010 / April 2010 / May 2010 / June 2010 / July 2010 / August 2010 / September 2010 / October 2010 / November 2010 / December 2010 / January 2011 / February 2011 / March 2011 / April 2011 / May 2011 / July 2011 / August 2011 / September 2011 / October 2011 / April 2013 / November 2013 / December 2013 / January 2014 / February 2014 / March 2014 / April 2014 / July 2014 / August 2014 / September 2014 / October 2014 / November 2014 / February 2015 / April 2015 / October 2015 / March 2016 / April 2016 / May 2016 / June 2016 / July 2016 / November 2016 / December 2016 / February 2017 / May 2017 / June 2017 / September 2017 / October 2017 / November 2017 / March 2018 /

Powered by Blogger

RSS feed

Visits since 2006/05/11: