|
Post by Ed Ireland on Oct 11, 2012 10:17:39 GMT 9.5
I have been wandering around this website for some time now and it has just occurred to me that I have a possible, albiet potentially absurdly expensive, solution to a question that has been dogging me since I made the decision to advocate construction of as many nuclear reactors as possible as fast as possible: that of the supply of sufficient large pressure vessel, and perhaps steam generator, forgings to do the job.
As many people here will already know, modern BWRs and PWRs both require very large quantities of very large forgings of types that are only manufacturable in a handful of facilities globally. These sites are now approaching saturation in capacity terms and ti will take several years to expand this capacity. (The recent binge of Gen-3 reactors in China and South Korea is partially to blame).
These are the aforementioned pressure vessel parts and the steam generators of PWRs.
Although older Gen-II plants require fewer very large forgings, the amount required is quite considerable in the present industrial climate with large quantities of the industrial infrastructure for the production of said forgings under considerable strain.
Gen II reactors are also hamstrung by potential safety issues, capital cost issues and the fact that they are relatively low conversion designs, which will increase uranium production, and with the extended plant lifes we are now faced with that could prove to be a significant issue down the line.
The alternative has come to me however: a modern Advanced Sodium Cooled reactor.
Reading various studies that I found on this and other sites, abusing my academic priveledges, I discovered the concept of using a loop-type sodium cooled fast reactor that uses "printed circuit heat exchangers" to directly heat helium gas which generates electricity using a multiple reheat brayton cycle. Such a design has no water passing through the same heat exchangers as metallic sodium and would indeed need to get no closer to the core than the turbine hall (where it would pass through the intercooler equipment). This allows the intermediate heat transfer loop to be removed from the design and removes the need to shut down the plant at the slightest leak in the heat exchangers as small leaks can be dealt with using a gas separator in the loop without even having to shut down the plant.
It also drastically reduces the cost of construction of the plant for reasons that I am sure are clear.
These reactors would require no large forgings whatsoever as far as I can tell, as they would not operate at significant positive pressure (beyond perhaps ~1atm relative to outside to ensure any breaches in the containment do not allow air in) and because the PCHE heat exchangers are made out of large numbers of small stainless steel plates an do not require forging at all.
The largest parts would be the pressure cases for the helium turbomachinery and huge numbers of similar pressure cases are produced every year for the gas turbine industry. (as several could be arrayed in parallel for a large reactor).
Adopting such a design would allow the use of metallic fuels bearing plutonium, or ~15% enriched uranium to allow reactor start ups to be accelerated beyond the available plutonium supply, to be used, providing access to a wide range of reduced cost technologies such as a simplified FLUOREX or pyroprocessing.
Finally, since large numbers of sodium cooled reactors have been operated, and there is operating experience with cores in the ~1000-1500MWe range (SuperPhenix), it may be possible to skip the demonstration plant step and proceed directly to the first run of commercial plants. The new balance of plant design could be tested during site preparation works on the first run of commercial reactors and provision made to revise the design to include conventional steam plants hould the helium plant prove to be too troublesome prior to actually starting to pour concrete.
It might not actually be neccesary to even complete detailed core design work before construction, if the designers are confident they can provide the required core characteristics in a defined space that could be sized relatively conservatively, such a design would provide opportunity for later adaptions of the reactor to different purposes.
Overall, I propose to throw the current rule book out of the window and proceed directly to construction of commercial sodium cooled plants, trusting our accrued fast reactor operation knowledge and our access to large numbers of supercomputers to win the day.
This might allow us to get the construction time down to 3years from deciding to build the plant to first electricity production, especially after the first "tranche" is completed and the second "tranche" commences work, although the components would be manufactured and stockpiled in advance to maintain continuous production lines.
(And yes, the objective requires such steps as I want to go all nuclear in the 1st world as fast as possible, my primary interest is in the UK as that is where I am from).
So anyone got any opinions as to whether I am made or simply insane?
|
|
|
Post by David B. Benson on Oct 11, 2012 12:36:50 GMT 9.5
Ed Ireland --- The GE-Hitachi PRISM is already designed, proposed to the UK, and incorporates the most important of the features you mention, although it still uses a Rankine cycle steam turbine to turn the electricity generator. (There are several earlier threads here on fast reactors, usually termed IFRs.)
US DoE is sponsoring a carbon dioxide Brayton cycle turbine, but it is not yet ready for immediate deployment.
The biggest factor in fast construction is the planning process. Note the UK difficulties on The Dilemma thread. As an example, although not an NPP, it took Idaho Power 10 years to finish the planning (for the route) of a transmission line through a lightly populated 500 km.
The next is that, at least in the US, it takes a minimum of 2 years to do the site preparation once the permits are in place. The PRISM will almost surely require at least a further 2 years to construct and test before enabling power to the grid. This (mere) 4 years assumes no (significant) court actions delay matters and that transmission is already in place or, at least, but short and not subject to court actions.
So ideally, about 4 years is the minimum one can hope for in a construction-friendly legal and regulatory atmosphere.
|
|
|
Post by Ed Ireland on Oct 11, 2012 18:58:54 GMT 9.5
Ed Ireland --- The GE-Hitachi PRISM is already designed, proposed to the UK, and incorporates the most important of the features you mention, although it still uses a Rankine cycle steam turbine to turn the electricity generator. (There are several earlier threads here on fast reactors, usually termed IFRs.) The problem with the S-PRISM is that it is rather small for a power reactor, which makes me worry about the number of components you require to generate each gigwatt. I'm of the opinion that a reactor in the 1000-1500MWe range is required for economic power generation. It also has the intermediate sodium loop and as you say, conventional steam cycle equipment. That makes me worry about capital cost escalation and the prospect of the problems which killed all previous attempts at commercial fast reactors rearing thier heads. US DoE is sponsoring a carbon dioxide Brayton cycle turbine, but it is not yet ready for immediate deployment. Well that is an interesting question, how soon is intermediate? If we really have 2 years of site preparation work before the reactor itself begins construction, could you develop the cycle completely in that time? Ditto the helium cycle, as the supercritical carbon dioxide brayton cycle has its own problems with sodium coolant (it is even more reactive than water with metallic molten sodium). You would have two years to build and run a full scale model of one of the cooling loops, with the sodium electrically or by natural gas, and then a further two years to tweak it even after construction starts and large changes become impractical. The biggest factor in fast construction is the planning process. Note the UK difficulties on The Dilemma thread. As an example, although not an NPP, it took Idaho Power 10 years to finish the planning (for the route) of a transmission line through a lightly populated 500 km. It luckily doesn't take the National Grid quite that long, although that is because our power grid is highly developed so most of the time new 400kV lines follow the routes of existing 132kV lines that are dismantled. I suppose this is why single circuit 400kV lines are unheard of in the UK and single circuit 132kV routes are extremely rare outside of the Scottish Highlands. The next is that, at least in the US, it takes a minimum of 2 years to do the site preparation once the permits are in place. The PRISM will almost surely require at least a further 2 years to construct and test before enabling power to the grid. This (mere) 4 years assumes no (significant) court actions delay matters and that transmission is already in place or, at least, but short and not subject to court actions. So ideally, about 4 years is the minimum one can hope for in a construction-friendly legal and regulatory atmosphere. Well any crash nuclear programme would be authorised by some appropriate act of Parliament which would short circuit any court challenges they could make. That is if the government was willing to risk all its political capital on pushing through a massive nuclear newbuild programme, which sadly it isn't.
|
|
|
Post by David B. Benson on Oct 12, 2012 11:28:52 GMT 9.5
Ed Ireland The PRISM is 633 MWe so a pair works out quite nicely, sharing the pyroprocessing unit.
A new Brayton cycle turbine requires vastly longer than a mere 2 years to bring to fit-for-service state. You rush too fast.
While site preparation is in progress, sometimes even sooner, the turbine is ordered and constructed. I don't know about now, but in 1955 I watched the shaft for a large steam turbine being prepped for its final trimming; all the suits at GE Lynn River works were in attendance. Clearly a great deal of time and $$ had already gone into this shaft.
Once ready, the many blades are precisely affixed and possibly the generator is wound onto the same shaft. All that looks time consuming to me.
With all the pre-fabricated parts in hand, including the turbine and generator, then the on-site construction takes only 2 years for a combined cycle gas turbine and I don't see what it should take longer for the comparably sized PRISM.
Other units which should also take 4 years or possibly a bit less are the small modular Gen III reactors (SMRs). These are discussed on another thread here.
|
|
|
Post by LancedDendrite on Oct 12, 2012 13:46:09 GMT 9.5
The problem with the S-PRISM is that it is rather small for a power reactor, which makes me worry about the number of components you require to generate each gigawatt. I'm of the opinion that a reactor in the 1000-1500MWe range is required for economic power generation. If you look at modern gas turbines, you'll find that they're usually in the 150-500MWe range for each turbogenerator. Given that gas-fired plant is the type being most heavily deployed in Western countries it would not be too much of a stretch to say that is around the right size. I think that the PRISM design is in the right tradespace given that it doesn't need to be scaled up in terms of power output as much as modern LWR designs to get good economies of scale. As for total power station size, it depends on the area. A large city (2 million+ inhabitants) could get by with a few 1-1.5GWe plants, but in more decentralised/lightly populated areas of the world it would be overkill.
|
|
|
Post by Ed Ireland on Oct 12, 2012 23:21:59 GMT 9.5
The problem with the S-PRISM is that it is rather small for a power reactor, which makes me worry about the number of components you require to generate each gigawatt. I'm of the opinion that a reactor in the 1000-1500MWe range is required for economic power generation. If you look at modern gas turbines, you'll find that they're usually in the 150-500MWe range for each turbogenerator. Given that gas-fired plant is the type being most heavily deployed in Western countries it would not be too much of a stretch to say that is around the right size. I think that the PRISM design is in the right tradespace given that it doesn't need to be scaled up in terms of power output as much as modern LWR designs to get good economies of scale. As for total power station size, it depends on the area. A large city (2 million+ inhabitants) could get by with a few 1-1.5GWe plants, but in more decentralised/lightly populated areas of the world it would be overkill. Unfortunately the political situation is rather different for nuclear power than it is for CCGT plants, attempting to build large numbers of small scattered power plants would be a nightmare even if the government wished to go all out for nuclear power. At best you would be able to drive through power plants at every site that has or has had a nuclear power plant. Since we have potential water shortage issues we have to exclude the two inland sites (Chapelcross and Trawsfynydd) and due to the environmental problems in the area we have to exclude Dungeness. That leaves only 12 sites, and two of those are in Scotland and construction there is likely to be frustrated by interference from the Scottish Government, so we are looking at 10-12 power plant sites. I calculate that if the UK wanted to become a zero-net importer of energy something on order of 100GWe average output would be required, which would be something on order of 110-120GWe of plant. This means that every one of those sites is going to have to be one of the largest power plants ever constructed, probably larger even than Kashiwazaki-Kariwa. As I understand it, 1500MWe units are not much larger in area terms than 900MWe Gen II units, and as such a 6-unit 9GWe plant could fit on the area occupied by the plant at Gravelines (1km by 500m). I am a little skeptical that the 15 power blocks (30 cores) required for a 9GWe site using the current PRISM design could fit in a similar space. As to the efficiency of such plants, the UK is so small as to make it potentially possible to concentrate all our generation in one place and still economically deliver power to every part of the UK, although that would not be the cheapest solution. (For reference, one of the Nelson River HVDC lines would stretch from the South Coast to the north coast of Scotland). Concentrating the supplies into the smallest possible number of sites reduces political trouble and does not significantly increase the cost of power. It also possibly allows for pooling of pyroprocessing resources to increase reliability, although I am not myself sold on pyroprocessing compared to the various FLUOREX style processes.
|
|
|
Post by David B. Benson on Oct 13, 2012 12:45:09 GMT 9.5
I don't know of any 1500 MWe units; the Areva EPR is a bit larger and the South Korean unit is a bit smaller. While the reactors might not require much more area, the power block does in proportion to the turbine/generator size. If 300 meters is sufficiently far away for the sodium/water heat exchanger then 30 PRISMs ought to be able to line up nicely within 1000 meters. I didn't leave any room for the pyroprocessor, but there ought to easily be room of a line 15 PRISMs, the pyroprocessor and then the other 15 PRISMs. In reviewing Processing of Used Nuclear Fuelwww.world-nuclear.org/info/inf69.htmlit seems a major advantage of pyroprocessing is the simplicity of the design (which does not involve any chemical processing except of the waste, to separate the Cesium-137).
|
|
|
Post by Ed Ireland on Oct 14, 2012 13:13:03 GMT 9.5
Indeed, there are no units I know of that are exactly 1500MWe exportable power but I normalise my preferred thermal reactor (the ESBWR) to that size.
At this point I am beginning to believe that going straight to fast breeders would be an engineering nightmare, it might be better to build BWRs now and get to work on a true fast breeder programme to manufacture start-up plutonium for the BWR fleet's expansion
Also build Hitachi's proposed 180MWt test reactor to prove the RBWR concept which would allow the BWR fleet to be retrofitted into breakeven capable reactors as required. (Or even to relatively high ratio breeders using exotic fuels like uranium nitride).
The forgings issue can be overcome by building as many plants as posssible using the slack capacity available after Fukushima order cancellations and by simply welding up smaller forgings, although that does measurably increase plant operations costs.
Also perhaps by building a 14,000t press domestically... do we have any information on how long those take to comission anyway? As I understand it a single press could probably manage 6 reactor vessels per year, perhaps more if it was used solely for RPV work.
That would make a big dent in the requirements for the programme I envisage (something on order of 70 reactors in a burst and then a drastically reduced rate thereafter).
|
|
|
Post by David B. Benson on Oct 14, 2012 14:05:26 GMT 9.5
Ed Ireland --- (1) I fail to see anything approaching an engineering nightmare. (2) By following World Nuclear News, I fail to detect any order cancellations following Fukushima, only some non-replacement policies in some countries. (3) New large forges could be built; I suppose this would take about 6 years. This would be vastly better than attempting to through-weld smaller pieces to fabricate the large one called for in the original design; at least in the USA the NRC would never allow that. After reviewing Advanced Nuclear Power Reactorswww.world-nuclear.org/info/inf08.html(some of which information appears to be out-of-date) I, at least, cannot see much favoring one design over another as all appear quite good.
|
|
|
Post by LancedDendrite on Oct 14, 2012 14:22:49 GMT 9.5
Indeed, there are no units I know of that are exactly 1500MWe exportable power but I normalise my preferred thermal reactor (the ESBWR) to that size. I would go for a real, in-production reactor design instead of the ESBWR. For instance, the AP1000 is essentially a lower power output PWR version of it - same safety features essentially, and it's already got manufacturing and construction knowledge based around it. At this point I am beginning to believe that going straight to fast breeders would be an engineering nightmare, it might be better to build BWRs now and get to work on a true fast breeder programme to manufacture start-up plutonium for the BWR fleet's expansion Most sane plans call for upgrading existing nuclear power stations with Gen III+ LWR designs. LMFBRs and MSRs are better off being built at greenfield or brownfield sites. As an example, the PRISM proposed to the UK NDA is to be built at Sellafield - this could be done at other such sites if they have access to sufficient cooling water. Also build Hitachi's proposed 180MWt test reactor to prove the RBWR concept which would allow the BWR fleet to be retrofitted into breakeven capable reactors as required. (Or even to relatively high ratio breeders using exotic fuels like uranium nitride). I don't think we should be looking at new variants of LWRs, especially light-water breeders. Solid oxide fuels + water coolant isn't the way to go in the long term; there's too many reprocessing and safety issues. Molten salt breeders (along with DMSR burners) and liquid sodium fast breeders are excellent choices for Gen IV buildouts. I suspect that there won't be better designs that will come along. The forgings issue can be overcome by building as many plants as posssible using the slack capacity available after Fukushima order cancellations and by simply welding up smaller forgings, although that does measurably increase plant operations costs. Where is this slack capacity from cancelled orders? China is trying to build reactors as fast as it can still and the US hasn't really stopped either. Also perhaps by building a 14,000t press domestically... do we have any information on how long those take to commission anyway? As I understand it a single press could probably manage 6 reactor vessels per year, perhaps more if it was used solely for RPV work. Sheffield Forgemasters has a 10,000t press and was looking at building a 15,000t press, although I have no timetable on when they were going to commission it. Funding for the project appears to have dried up as of present. Also, Japan Steelworks was/is looking at increasing their throughput to 11 RPVs per year in 2013, up from the 6 that they can do at present. Feasibly if you reduced the size of the RPV then you could get by with more existing presses like that one. My opinion though is that if you're concerned about development hurdles and forge press capacity then there's a simple solution - the PHWR, otherwise known as the CANDU. The Enhanced-CANDU 6 is a minor upgrade to existing CANDU 6 reactors and uses a 'calandria' composed of many small pipes to carry water through the reactor core instead of a monolithic RPV. It's only a ~700MWe reactor, but built in pairs it can substitute for a single EPR/ESBWR/APR1400 reactor. Serial production of those would be an excellent way of getting down costs, at a small cost to footprint.
|
|
|
Post by Ed Ireland on Oct 14, 2012 15:16:12 GMT 9.5
Ed Ireland --- (1) I fail to see anything approaching an engineering nightmare. (2) By following World Nuclear News, I fail to detect any order cancellations following Fukushima, only some non-replacement policies in some countries. (3) New large forges could be built; I suppose this would take about 6 years. This would be vastly better than attempting to through-weld smaller pieces to fabricate the large one called for in the original design; at least in the USA the NRC would never allow that. After reviewing Advanced Nuclear Power Reactors www.world-nuclear.org/info/inf08.html(some of which information appears to be out-of-date) I, at least, cannot see much favoring one design over another as all appear quite good. Read more: bravenewclimate.proboards.com/index.cgi?action=display&board=energy&thread=346&page=1#ixzz29FP69ET31) While there is no forge bottleneck to consider, there are other critical production bottlenecks that would cause massive issues with regards a very large scale programme. One that springs immediately to mind is the fact that each reactor would consume a significant percentage of a year's total global metallic sodium production, and there is only a very limited capacity currently available that will produce reactor grade sodium as most is relatively impure sodium used in a variety of chemical syntheses. 2) JSW supposedly claims they have had order cancellations, remember the very restricted forging capacity has caused reactor vendors to place orders for reactor pressure vessels well ahead of any projects actually being announced since the RPVs can be used in any reactor of the same design. 3) Assuming the Forge was simply a carbon copy of one of the Forges already built relatively recently, I would think six years is probably very pessimistic, before the funding was pulled Forgemasters planned to go from the drawing board to turning out product in roughly three years as far as I can tell, and that was with an all new Bespoke press with a capacity unmatched anywhere. Indeed, there are no units I know of that are exactly 1500MWe exportable power but I normalise my preferred thermal reactor (the ESBWR) to that size. I would go for a real, in-production reactor design instead of the ESBWR. For instance, the AP1000 is essentially a lower power output PWR version of it - same safety features essentially, and it's already got manufacturing and construction knowledge based around it. Well ESBWR supposedly shares most of its major structural components with the ABWR, which is also a proven reactor design in production today. BWRs also offer the possibility of conversion to light water breeders using uranium (not thorium as occured at Shippingport) through the installation of an apparently simple to install RBWR "pancake core" at any time during the reactor's life, and when we are dealing with 60 year life plants this is a useful feature IMO. Most sane plans call for upgrading existing nuclear power stations with Gen III+ LWR designs. LMFBRs and MSRs are better off being built at greenfield or brownfield sites. As an example, the PRISM proposed to the UK NDA is to be built at Sellafield - this could be done at other such sites if they have access to sufficient cooling water. Well technically the PRISM power block at Sellafield would not be the first reactors on the site, as Sellafield is the site of the Calder Hall facility which operated six 50MWe Magnox plants, they were one of only two civil nuclear power reactors operated with cooling the UK, the the other being Chapelcross, the third reactor at Trawsfynydd simply directly cycled water from a lake. All reasonably proposed sites for nuclear newbuild in the UK are coastal and thus have potentially unlimited cooling water potential (well three are on the Severn Estuary but it is tidal at those sites so it still counts as the sea I think). Therefore limitations in plant capacity are essentially based on how much land you can commit to the facilities. I don't think we should be looking at new variants of LWRs, especially light-water breeders. Solid oxide fuels + water coolant isn't the way to go in the long term; there's too many reprocessing and safety issues. Molten salt breeders (along with DMSR burners) and liquid sodium fast breeders are excellent choices for Gen IV buildouts. I suspect that there won't be better designs that will come along. The advantage to building LWRs is that we can build large amounts of capacity very quickly and we have acquired vast amounts of operating knowledge with them. Operating costs for a Light Water Breeder are likely to be similar to operating a 100% MOX core conventional LWR and are thus probably lower, in the near term, than for a sodium cooled fast reactor. Additionally capital cost has been highly optimised in Gen 3+ reactors like the ESBWR thanks to passing through several dozen design iterations. The design developments that will allow fast reactors to compete, like the aforementioned helium brayton cycle, are years away from being ready for deployment and we simply cannot wait. To support the estimated 1% growth in British energy demand, less than 5% of the thermal output of the British nuclear park would have to be fast breeders in the PRISM mould, the rest could be LWBRs that could simply recieve start up plutonium from the FRs or from enriched uranium if the fast reactor programme is stalled. As to the problems with oxide fuels, there are much interesting work being done on stabilising UN fuels for light water use and there is always the possibility of Zr-U or even Zr-U-Pu fuels like those being worked on by Lightbridge. (All metal RBWR fuel would solve many of the problems in this thread by allowing the RPV to be shrunk down rather significantly). Where is this slack capacity from cancelled orders? China is trying to build reactors as fast as it can still and the US hasn't really stopped either. As I said, JSW apparently claims to have had orders for RPVs cancelled, but remember how long the waiting time is, many vendors apparently order the pressure vessels before they know if they will get any orders. Also perhaps by building a 14,000t press domestically... do we have any information on how long those take to commission anyway? As I understand it a single press could probably manage 6 reactor vessels per year, perhaps more if it was used solely for RPV work. Sheffield Forgemasters has a 10,000t press and was looking at building a 15,000t press, although I have no timetable on when they were going to commission it. Funding for the project appears to have dried up as of present. [/quote] The sheffield forgemasters monster press is dead, the incoming Conservative government decided that giving any sort of loan to private industry was tantamount to the hated "picking winners" and was thus socialism and must be destroyed at any cost. After it caused massive political damage they reapproved a new loan for far less money for other more minor improvements, but the dream of having an RPV capable forging press in Sheffield is dead. Also, Japan Steelworks was/is looking at increasing their throughput to 11 RPVs per year in 2013, up from the 6 that they can do at present. Feasibly if you reduced the size of the RPV then you could get by with more existing presses like that one. I've thought about that too, apparently the reduced core height of the RBWR enables the RPV to be shortened which would enable atleast some of the forgings to be eliminated without reducing core output power, but otherwise you are limited by how small you can make plants before you have to cut output power and fall into a region where you reduce the overall economic competitiveness of the programme. My opinion though is that if you're concerned about development hurdles and forge press capacity then there's a simple solution - the PHWR, otherwise known as the CANDU. The Enhanced-CANDU 6 is a minor upgrade to existing CANDU 6 reactors and uses a 'calandria' composed of many small pipes to carry water through the reactor core instead of a monolithic RPV. It's only a ~700MWe reactor, but built in pairs it can substitute for a single EPR/ESBWR/APR1400 reactor. Serial production of those would be an excellent way of getting down costs, at a small cost to footprint. Unfortunately the CANDU seems to have cost control issues that make the EPR's current difficulties look minor, everything seems to run massively over budget and behind schedule, but I don't know if there are any unusually massive roadblocks being thrown in front of the Canadians. Although the Chinese did manage to turn out the Qinshan CANDUs early and under budget so its probably not inherent to the design. Unfortunately the CANDU falls into the trap I mentioned earlier about plant footprints which are very important for political reasons in the UK. It effectively has to be ~1000MWe+ plants. Development hurdles might be able to reduce the neccesary size of the RPVs and the forgings required but I don't think we have the time to wait, we have to start a massive serial reactor programme right now.
|
|
|
Post by David B. Benson on Oct 14, 2012 16:08:16 GMT 9.5
Expanding en.wikipedia.org/wiki/Sodium#Commercial_productionwould be relatively simple and fast; I don't see this as a serious obstacle. It seems that big forges can either be finished more rapidly in Britain than in the USA or the schedule was too optimistic. The different might lie in the length of the permitting process. An alternative to seriously consider adding are blocks of SMRs at alternate sites. The one I know the most about is the Nuscale 45 MWe (Gen III+) unit, designed to be blocked in units up to 12. The footprint is quite modest and the design offers three obvious advantages: (1) add units as (local) demand grows; (2) rapid construction (44 months start to finish); (3) little additional transmission lines need to be constructed, up to the limit of existing distribution. Admittedly none of the SMR designs have been NRC licensed yet, but Nuscale seems still hopeful they'll have their type license in 2018.
|
|
|
Post by jagdish on Oct 14, 2012 19:14:03 GMT 9.5
Two types of solution are possible:- 1. Fast reactors, not using water or any moderator. 2. Multitube design, used at present in heavy water reactors. This will require a large number of lighter forgings. The progress in the fast reactors has not advanced much due to fires in the sodium coolant. The solution is to use Lead coolant as in some of the Russian reactors, or a salt eutectic as coolant as suggested in pebble bed advanced high temperature reactor design. Multitube design has further possibilities. It could use separate moderator and coolant, as in the Indian AHWR design. Coolant could even be non-moderating but low vapor pressure like sodium, lead or salt. Multitube fast reactors could even be simpler like an industrial boiler with fire tubes separated from a lead or salt coolant. Heat exchangers are in any case only boilers with a different heating agent. Sufficient industrial capacity of boilers is not a problem.
|
|
|
Post by David B. Benson on Oct 15, 2012 11:41:43 GMT 9.5
There is extensive experience with the sodium cooled Russian BN-600; fires are not a problem because any such fire quickly extinguishes itself.
|
|
|
Post by Ed Ireland on Oct 15, 2012 11:55:32 GMT 9.5
Sodium leaks in the BN-600's heat exchangers have dogged the plant since it first opened, that is one of the reasons it has a 75% availability when ideally we would want something on order of 90% or more.
Until some sort of helium reheat cycle is available I do not believe fast reactors can compete economically with light water reactors for energy generation. Since it now appears that BWRs and perhaps PWRs (using UN type fuels) can be made into breeders, it further weakens the case at the present time.
|
|
|
Post by David B. Benson on Oct 15, 2012 14:49:44 GMT 9.5
|
|
|
Post by David Walters on Oct 19, 2012 0:36:25 GMT 9.5
A small FYI. while all gas turbines (Brayton cycle) are "single shaft" they are not built as a single shaft. The generator is wound on it's own shaft and the blades of the GT are fitted on it's single shaft. They are VERY simply bolt together between the two when they are separately delivered to the construction site.
The bolting creates a 'single shaft'.
David
|
|
|
Post by edireland on Oct 20, 2012 12:45:56 GMT 9.5
The low capacity factor is indeed by design, the whole idea is that minor equipment failures, which are apparently still depressingly regular, will not result in a reactor scram as the cooling system consists of numerous parallel trains which means that heat exchangers and cooling loops can be shut off with only a loss of a small amount of reactor output.
That keeps the CF up even in the face of equipment problems.
Based on the above Sheffield Forgemasters schedule and the schedules listed in the ESBWR documentation I estimate 91 months from project start to the first reactor being available for commercial service with reactors then arriving once every 2 months after this time.
Unfortunately that is very slow and it could do with atleast a couple of years cutting off the start time.
EDIT: Does anyone know if the reactor construction time assumes a single shift working 8-12 hour days or if it assumes multiple shifts working 24 hours a day? The reactor building speed seems rather low considering it supposedly contains rather less concrete and steel than Gen II PWRs do.
This would seem to be a way to rapidly accelerate construction although it would likely require a large itinerate work force that would have to be housed at significant cost.
|
|
|
Post by David B. Benson on Oct 21, 2012 9:58:34 GMT 9.5
edireland --- The construction schedule is paced according to laws governing labor at each construction site. Naturally when capital costs (finance costs) accumulate without the ability to (yet) pay there is an incentive to hurry the schedule versus the time-and-a-half for overtime required in the USA. Westinghouse has paced the construction of AP1000s to use some overtime so as to lower capital costs, but not so much as to raise overall costs (according to their economists).
But in China, with a different labor rate system, AP1000 construction proceeds much more rapidly. While I'm not sure about the AP1000s, the Chinese CP-1000 (completely different design, only Gen II+) requires only 44 months, start to finish, once site preparation is complete. The South Koreans build their units in 48 months. Etc.
AFAIK nobody uses a 24 hour schedule.
|
|
|
Post by edireland on Oct 22, 2012 1:54:37 GMT 9.5
The ESBWR General information schedule maintains that the ABWRs in Japan took 36 months from first concrete to fuel loading (and that ESBWR would be broadly comparable).... is there any chance that could be cut to 24 months if you were to go for a "war emergency" schedule with three shifts running around the clock?
I assume you wouldn't want to rush the 6 month commercial shakedown period at the end of that but you might be able to manage a 3 year overall schedule? (Ignoring the lead times neccesary to source long lead items which would tkae it up to approximately 4 years probably)
And Americans still get time-and-a-half for overtime?
|
|
|
Post by David B. Benson on Oct 22, 2012 6:48:59 GMT 9.5
edireland --- For construction work, yes, 1.5 time for overtime.
|
|
|
Post by edireland on Oct 23, 2012 19:58:11 GMT 9.5
So this potentially cuts construction time by half to two thirds? Assuming no limits such as maximum rate of concrete pouring are hit? That would be rather impressive and allow nuclear to actually approach coal plant construction time.
|
|
|
Post by David B. Benson on Oct 24, 2012 11:40:19 GMT 9.5
edireland --- The Westinghouse AP1000 builds in the USA use some overtime. An example might be after the rebar is put up and inspected, the concrete pour continues, without pause, until the unit is finished. Large concrete dams typically do not have much rebar, if any. The last one constructed around here, Dworshak Dam, took several years of (nearly) 24/7 concrete pouring. en.wikipedia.org/wiki/Dworshak_DamI'm impressed by the 36 month build time in Japan. Perhaps there are more details about that construction period that you could find and share, an example being the hours per day actually worked.
|
|