Energy From Thorium Discussion Forum

It is currently Feb 20, 2018 2:16 pm

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 94 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7  Next
Author Message
PostPosted: Jan 04, 2014 10:13 pm 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
Cyril R wrote:
Quote:
the net electricity revenues for the NACC plant would be 50% greater for NACC than for a conventional base-load plant under typical U.S. market conditions.


Bit apples and oranges here, Dr. Peterson. We should be comparing a nuclear baseload plant with independent load follower/peaker gas turbine, versus the combined NACC plant. Since you're using natural gas for the peaking, not nuclear heat. If the baseload plant is cheaper, that capital could be spent in a peaker gas turbine as well (the latter are very low in capital cost and not being nuclear would be easier to build on the side). The only advantage for the NACC plant, then, would be a reduction in the gas usage for peaking compared to an inefficient seperate gas peaker plant. It is likely that a seperate gas peaker would be cheaper than when integrated in a nuclear plant, due to massive quality control issues (costs) with any nuclear build (even for nonnuclear components in nuclear plants).


You make a reasonable and debatable point. The devil is in the details.

We need to go back to the important point about wind and solar, which is that the value of electricity which is intermittent is lower than the value of electricity which is constant with scheduled outages (e.g., nuclear with its current <2% U.S. forced outage rate). So the use of "levelized electricity cost" is incorrect to compare intermittent and base load energy sources. To make an accurate comparison between wind and base-load nuclear, or between base-load nuclear and NACC, one must integrate production and prices over time, using logical rules for how sources are dispatched.

As you point out, an alternative to NACC is to combine a nuclear base-load plant with a cheap natural gas peaker turbine, which would consume more natural gas than NACC but would capture at least some of the peaking revenues (it would capture less of this revenue, because the low efficiency gas turbine would be dispatched less of the time, and would spend more on fuel).

For the current U.S. situation with $3.5/MMBtu natural gas, for a 38% efficient peaking turbine the cost of fuel is ($3.5/MMBtu)(3.4MMBtu/MWht)/(0.38MWhe/MWht) = $3.0/MWhe, and for a 60% efficient natural gas combined cycle plant the fuel cost is $2/MMBtu. Nuclear cannot compete at these natural gas prices. We do not agree on the reasons, but we agree that natural gas prices in the U.S. will rise from these current levels in the future.

In Europe, Japan, South Korea and China natural gas prices are 3 to 5 times higher and nuclear power is generally economic for base load compared to natural gas. (As a side note, if the U.S. wants to be a significant player in changing the future trajectory of nuclear energy technology and nuclear fuel cycles, it must improve its ability to gain significant access to these international markets.)

For "high" natural gas prices, NACC is clearly preferable over conventional base-load nuclear plants combined with conventional gas peakers, because NACC will provide lower-cost peaking, frequency-regulation and other grid support services, as well as providing base load generation.

The key question is, at the natural gas price threshold where nuclear becomes competitive, whether the added revenues from NACC make it more attractive than conventional nuclear base load with separate gas peaking. We're unlikely to see this question answered, because the added costs of first-of-a-kind deployment will drive new reactor deployment to occur primarily in markets that already have higher natural gas prices. NACC will be of greater interest for these markets.

Answering the question of whether NACC, or conventional base-load nuclear combined with gas peaking, is more competitive at intermediate natural gas prices requires modeling both the relative capital costs of the two types of plants, as well as their net revenues under realistic market conditions. What we know for sure is that anyone who builds NACC plants can be confident that they will dispatch before the natural gas plants, not just for base load but also for peaking, and that they reduce net carbon dioxide emissions more than conventional nuclear can achieve.

This does not obviate the argument that steam cycles might be the best near-term option, even though in the longer term NACC provides more valuable electricity. But the argument for steam rests on the assumption that it can have significantly lower capital cost that NACC, and that remains debatable.


Top
 Profile  
 
PostPosted: Jan 04, 2014 10:28 pm 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
Cyril R wrote:
Quote:
With a pressure differential of 18 bar and with the salt supplied and removed by manifold pipes, it turns out that the tube sheet is quite thin (2-3cm). The joint is loaded in compression. We're working on options to design and fabricate these joints, as it is clearly a key design issue. But at least they are inspectable.


How did you get such thin tubesheets? ORNL's tubesheets for the low pressure primary HX were over 12 cm. Baffles would be similar thickness.


Ansaldo solved this problem for the GE SBWR project. The original SBWR isolation condenser design had a conventional flat tube sheet, which ended up needing to be very thick due to the high pressure difference between the 70bar steam and the 1 bar water in the water pool the condenser sat in, and was calculated to have unacceptable thermal stresses due to the slow thermal diffusion into this thick tube sheet (similar to the major problem one must overcome with PCHE's). The isolation condenser that they built and tested used cylindrical, thin-walled inlet and outlet manifold headers (ratio of wall thickness to pipe diameter similar to the value for the tubes), with the tubes emerging perpendicular from the manifold pipes. The NACC CTAHs do the same thing.


Top
 Profile  
 
PostPosted: Jan 04, 2014 10:47 pm 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
Cyril R wrote:
Quote:
Inability to perform in-service inspection in PCHEs is a source of technical and licensing risk. NRC General Design Criteria require that the reactor coolant boundary be designed to be inspectable, so one needs a strong motivation to take on the testing and qualification program needed to demonstrate that the PCHEs will operate with sufficient reliability that in-service inspection is not needed.


Tubular HXs must be inspected because of the many degradation mechanisms that exist. Vibration, fretting, differential expansion issues from great relative thickness (between tubes and tubesheets/baffles/shells). None of those are present in PCHEs. Also the smaller size of PCHEs opens up more options for nondestructure inservice inspection without opening the HX. Ultrasound resonance techniques and eddy current techniques appear feasible.

The PCHEs are so compact that they can fit inside the reactor vessel so your coolant boundary is the reactor vessel. Similar to PWRs, there would be isolation devices on the secondary loop so that in the unlikely event of a plate through-leak the leak could be isolated. In case of pressure equalized salt-salt exchangers however, a major leak is not plausible. The design approach is to eliminate the drivers for leaks in HXs.


PCHEs do not perform as well as tubular heat exchangers with respect to thermal stress transients and fouling. Simplistic models greatly under predict the stresses induced by transient changes in flow rate of the fluids in PCHEs, which is why after developing a detailed, multi-dimensional, multi-scale model for PCHEs to evaluate their use for high temperature reactors, we chose to focus on tubular heat exchangers as the most practical approach for high temperature nuclear application. PCHEs have very small flow channels, and the experience with fluoride salts is that corrosion products and noble-metal precipitate in the cold parts of heat exchangers.

The lack of ability to perform any PCHE in-service inspection is a problem, because one will be required to monitor for their degradation due to low-cycle thermal stress fatigue and corrosion product deposition. If MSRs are required to meet similar general design criteria as SFRs, they will be required to have their intermediate salt operate at higher pressure than their fuel salt. Perfect pressure equilibrium in IHXs is an idealization that one is likely overly optimistic to make, and because regulators will want you to prove that fuel salt will not leak into the intermediate system (or design one's safety systems to manage injection of fuel salt into the intermediate loop), one will end up with a positive pressure difference anyhow and any leak in an IHX will inject intermediate salt into the fuel salt.

Tubular heat exchangers have their problems too, so it's legitimate to have opinions for and against the technology options.


Top
 Profile  
 
PostPosted: Jan 04, 2014 11:43 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1954
Location: Montreal
Per Peterson wrote:
Ansaldo solved this problem for the GE SBWR project......
I don't know about Ansaldo, but this type of "tube sheet" design has been part of VVER technology forever.....


Attachments:
Temelin_SG_details.jpg
Temelin_SG_details.jpg [ 285.05 KiB | Viewed 2725 times ]
Temelin_SG2.jpg
Temelin_SG2.jpg [ 289.34 KiB | Viewed 2725 times ]
Top
 Profile  
 
PostPosted: Jan 05, 2014 12:23 am 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
jaro wrote:
I don't know about Ansaldo, but this type of "tube sheet" design has been part of VVER technology forever.....


Very good point. The SBWR/ESBWR isolation condensers are a lot smaller,

http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/25/067/25067131.pdf

and with their thinner wall cross sections are pretty good at handling thermal transients.


Top
 Profile  
 
PostPosted: Jan 06, 2014 4:40 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Thank you dr. Peterson for your high quality responses.

Tube-and-header exchangers are indeed very attractive, doing away with thick tubesheets and in stead going for headering.

They aren't new though; they have been used for decades in industry and even ORNL had a tube-and-header design for a steam generator design decades ago.

I'm surprised Jaro didn't mention CANDUs; these are essentially tube-and-header machines, though they don't have to transfer heat through walls.

For me though, the main point against tubular HXs still stands, even for tube-and-header exchangers. They are big and involve lots of thin fusion welds that can corrode or otherwise leak, and there will be vibration issues, the recent experience with SONGS is not encouraging (a team of design specialists got it all wrong and the steam generator failed due to vibration). Designing against vibration tends to go against mitigating thermal stress from differential expansion, as you can imagine, so its a hard tradeoff with tubular exchangers.

The ESBWR ICs are compact because they are high pressure condensers in a pool of excellent coolant (cold water). There is no shell, so a bare tube-and-header configuration is a clear winner here.

In contrast, the ESBWRs PCCS condensers have much higher volume/MWt, due to lower pressure steam condensing. And that's still condensing, which is very efficient compared to heating medium pressure air. The ESBWR PCCS are only rated at 7.8 MWt/piece. So not compact at all. Typically tube-and-header (or tube-and-drum) require greater tube pitches than tube-and-shell. Very tight pitches are possible with tube-and-shells. It remains to be seen how compact these exchangers will be. This is unimportant in the ESBWR, since the rating is small and they don't care about expensive salt or expensive high temperature design and they don't have a shell. Also there is no tritium issue with BWRs PCCS since there just isn't much in the primary coolant.

The AHTRs CTAHs will be more similar to the ESBWRs PCCSs. They will be big and cumbersome, they can't fit in a reasonably sized reactor vessel especially for MSRs where size is more important, that means no integral design space. An MSR has significant activity from delayed neutrons in the exchanger and secondary loop, unless FLiBe is used, even then there will be some activation in the coolant. This makes it more important to go for compaction.

If we look at the micro electronics industry, and also in other areas like recent downsizing of car engines, home furnace boilers, and such, we see that compaction is an important cost cutter.

Having talked at length with a Ph.D. employee from Heatric, I believe that a PCHE can be made to withstand transients without failure, by designing for a significant deformation allowance in the plates and passages, though some of this is proprietary and I shouldn't talk too much about it here. Suffice to say that while more advanced computational software like COMSOL can calculate thermal stresses, its not very good at calculating the mitigative effect of local deformation in a complex geometry, so its peak stress calculations tend to be gross overestimates. In fact, one of the reasons why ductile steels and alloys are used so much is that they are very forgiving materials; local peak stress buildup will strain itself out quite readily with ductile alloys, rather than causing failure as predicted by the multiphysics software. A test stand will be needed to prove this as PCHEs are difficult to model via software.


Top
 Profile  
 
PostPosted: Jan 06, 2014 4:58 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Quote:
Perfect pressure equilibrium in IHXs is an idealization that one is likely overly optimistic to make


Agreed. It doesn't have to be perfect, though; compared to a direct air heater with a dif pressure of 18 bars, a 1.8 bar dif pressure would already cut this particular stress item by a factor of 10. With a low pressure drop PCHE, and a salt-salt exchanger, we can do a lot better than 1.8 bar pressure differential. We could probably have an operating band of +-0.2 bar so very close to pressure equalized design.


Top
 Profile  
 
PostPosted: Jan 06, 2014 5:03 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Quote:
PCHEs have very small flow channels, and the experience with fluoride salts is that corrosion products and noble-metal precipitate in the cold parts of heat exchangers.


Not sure I agree on this one. In low velocity natural circulation corrosion loops operating in very oxidating environment (or even uncontrolled redox conditions), these effects are observed. In real, high velocity heat exchangers like the MSREs, with low corrosion from redox control, plating and such were not observed to any meaningful extent. There were very few corrosion products at all, and they were largely dissolved. Noble metal plating would not be a major issue as there's so much surface area in PCHEs that offsets the smaller passages, and noble metals have high thermal conductivity so their heat transfer fouling factor is low. If a PCHE gets saturated with deposits, it could simply be cut off and replaced with a new module.

The most stagnant (lowest flow) areas occur inside the reactor vessel, where we have a large acceptance for deposition. Filters or sponges will likely be added somewhere in the more stagnant areas, to effectively compete for particulates.


Top
 Profile  
 
PostPosted: Jan 06, 2014 5:21 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Per Peterson wrote:
jaro wrote:
I don't know about Ansaldo, but this type of "tube sheet" design has been part of VVER technology forever.....


Very good point. The SBWR/ESBWR isolation condensers are a lot smaller,

http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/25/067/25067131.pdf

and with their thinner wall cross sections are pretty good at handling thermal transients.


I'm curious though. The ESBWR ICs have no shell (just a pool of water) so obviously there is no differential expansion issue here! How do you mate a high temperature shell (holding FLiBe I presume) to the tubes and drums/headers? Will you still get the advantages of handling thermal transients when you consider the shell? Will you simply place everything (including drums) inside a really big shell and hang the whole tube/drum assembly from that shell? What is the temperature of the shell? Is it cooled by salt, if so what kind of transient temperature can you get?


Top
 Profile  
 
PostPosted: Jan 06, 2014 1:57 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1954
Location: Montreal
Cyril R wrote:
I'm surprised Jaro didn't mention CANDUs; these are essentially tube-and-header machines, though they don't have to transfer heat through walls.
There are only 380 feeder tubes on the reactor face of a typical Candu reactor (C6), welded to the inlet/outlet headers -- by contrast, the VVER SG has 11,000 tubes. Hardly the same situation.


Attachments:
CANDU 6 inlet-outlet headers.JPG
CANDU 6 inlet-outlet headers.JPG [ 297.82 KiB | Viewed 2683 times ]
Top
 Profile  
 
PostPosted: Jan 06, 2014 4:03 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Hmm, looking at the images some more, it appears the VVER SG is a sort of hybrid, with a thick tubesheet for the primary circuit inlet cone, and shell-and-drum design for the secondary steam side. It's hard to tell from the images though.


Top
 Profile  
 
PostPosted: Jan 06, 2014 4:23 pm 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
Cyril R wrote:
I'm curious though. The ESBWR ICs have no shell (just a pool of water) so obviously there is no differential expansion issue here! How do you mate a high temperature shell (holding FLiBe I presume) to the tubes and drums/headers? Will you still get the advantages of handling thermal transients when you consider the shell? Will you simply place everything (including drums) inside a really big shell and hang the whole tube/drum assembly from that shell? What is the temperature of the shell? Is it cooled by salt, if so what kind of transient temperature can you get?


It's a bit difficult to describe without presenting a CAD model, and because we are still developing details in the CAD model we are not distributing it outside the review group, mainly for version control reasons (it's not helpful to have a lot of different drawings out that are not the same). We're distributing the CAD model showing the isometric of the reactor coupled to the CTAHS and the modified GE 7FB gas turbine, because for these elements the design is now finalized and no major further changes will occur.

We expect to have a final report published later this year. So it may make sense to pick this question back up when this is available. But basically the CTAHS run immersed inside the air; the technology for ducting high temperature air around using internally insulated steel vessels is well developed for heat recovery steam generators; the ducting that supplies and returns air differs in that it is cylindrical due to the higher internal pressure than one has in a HRSG but the materials and insulation system are similar.


Top
 Profile  
 
PostPosted: Jan 06, 2014 4:29 pm 
Offline

Joined: Dec 31, 2008 12:09 am
Posts: 236
Location: Berkeley, CA
jaro wrote:
Cyril R wrote:
I'm surprised Jaro didn't mention CANDUs; these are essentially tube-and-header machines, though they don't have to transfer heat through walls.
There are only 380 feeder tubes on the reactor face of a typical Candu reactor (C6), welded to the inlet/outlet headers -- by contrast, the VVER SG has 11,000 tubes. Hardly the same situation.


Might you have info on how the tube to header joints are fabricated for the CANDUs? The key difference for the CTAHS is that the fluid in the tubes and headers is at a lower pressure than the outside fluid, so everything is loaded in compression rather than in tension. The most important effect is that the pressure differential wants to drive the tubes into the header holes, like corks, rather than pull them out, so we expect that it may be beneficial to machine a small taper on the ends of the tubes (this also helps remove the aluminized metal cladding used to generate a Al2O3 tritium diffusion barrier, at the tube sheet joint). We've not found any other analogous tube-to-header heat exchanger design that is loaded in compression this way.


Top
 Profile  
 
PostPosted: Jan 06, 2014 4:50 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5057
Yes, its quite uncommon to have the high pressure in the shell and low pressure in the tubes. I guess it makes sense for this application, where its easier to have air than FLiBe in the shell, especially in terms of inspections.

LWRs have the cladding in compression for much of the operating life, they are simply welded.

Cold rolled joints may also be possible, if welding is not desireable for whatever reason. Compression would force the tube into the rolled joint extension (that is part of the header, say a die/punched hole) so it should never fail under operating compression. Not sure if such small tubes can be rolled, though...


Top
 Profile  
 
PostPosted: Jan 06, 2014 8:15 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1954
Location: Montreal
Per Peterson wrote:
Might you have info on how the tube to header joints are fabricated for the CANDUs?
According to the technical documentation,
Quote:
All fittings are the butt weld type, and meet ASME material specification SA-234 (grade WPB), SA-181 (grade II), and SA-105.
Dimensions are as per ANSI B16.9.
The reactor inlet and outlet headers are designed and manufactured from ASTM A106 Grade B carbon steel to Class 1, Section III of the ASME Boiler and Pressure Vessel Code.
The header nozzles for the feeders, boiler inlet lines and pump discharge lines are cold extruded from the header wall.


Attachments:
header nozzles for feeders.jpg
header nozzles for feeders.jpg [ 353.92 KiB | Viewed 2661 times ]
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 94 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7  Next

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group