With little fanfare last week, the Chinese designed HPR1000 (previously Hualong-1), pressurised water reactor, cleared the first of four stages in the General Design Assessment (GDA) administered by the UK Office of Nuclear Regulation (ONR). China General Nuclear (CGN) proposes to build 2 reactors of this design at the Bradwell site in England, in partnership with French state-owned EDF that currently operates all UK commercial reactors.
This guest post by Andy Dawson gives a preliminary overview of the design focussing on safety systems.
Introduction
As almost all readers of the blog will be aware, a team of EdF and China General Nuclear (CGN) have proposed the construction of a Chinese designed nuclear station at Bradwell, in Essex. On Thursday of this week, the UK Office of Nuclear Regulation announced that the design proposed for the station -the “HPR1000”, originally known as the “Hualong-1” has successfully completed the first, preparatory stage of the Generic Design Approval (GDA) process. This appears to have been completed on time, or perhaps a few weeks early.
While we shouldn’t over-state the importance of this particular transition – GDA is a four stage process, in which stages 2 & 3 are where the great majority of the detailed evaluation of the design from a safety perspective is undertaken – it is important in that it’s the first point at which the developers have to publish reasonably detailed data on the design. That data is available here.
This piece is intended to give an overview of the design, highlighted particular strengths and weaknesses that may affect the GDA outcome, and giving a comparison against the virtues and vices of the other contenders for UK build.
Design overview
The starting point for this commentary is that this is a VERY conventional Light Water Reactor design, with a design ancestry going back to a highly successful 1970s Westinghouse Pressurised Water Reactor with evolutionary change thereafter. The direct precursor is the 900MW Framatome design (based on licensed Westinghouse technology) built at Daya Bay in the 1980s. There have then been several further design evolutions driven by the Chinese themselves, resulting in Hualong-1, several of which are in build in China and in Pakistan. The baseline plant for the UK GDA entry is Fanggchengang 3, currently about 2 years into build (inset image), and due on line in 2022. From a western perspective, this may appear unpronounceable. I can’t comment on why they didn’t select the other potential reference plant, which is running slightly earlier – at Fuqing.
It’s not a radical leap in the way that the AP1000 attempted simplification of the PWR concept, or that the EPR tried to address potential regulatory issues by delivering massive redundancy on key systems.
Gross electrical output is 1180MW – although not stated, net output is likely to be in the 1130-1160MW range, after losses to powering internal systems. Thermal output is 3150MW, giving an overall efficiency of 37% – respectable for a PWR, and entirely comparable with AP1000 and EPR, and somewhat better than ABWR. The standard refuelling cycle is 12-18 months.
The core is again, directly reflective of Westinghouse design practice; it consists of 177 fuel assemblies, each consisting of a 17×17 matrix of fuel pins. Cladding is conventional zircalloy (of which more later).These have been lengthened from the precursor designs to reduce linear power density, and improve neutron economy. The core design makes extensive use of “burnable poisons” to manage the neutron flux distribution, to this same end.
However, unlike most recent designs, it doesn’t appear that the core is surrounded by a cylindrical steel neutron reflector. This means that the fuel usage (aka “burn-up” is somewhat less efficient that in designs like the EPR. EPR is claimed to be capable of running to a burn-up of 60GWd/t (gigawatt-days per ton) against HPR1000s 47GWd/t.
This may also have some relationship to another striking feature; the entire main Reactor Pressure Vessel (RPV) is produced as a single forging, as opposed to the mixed forging/welding of other designs. This should serve to significantly improve the fracture mechanics of the RPV, and since neutron embrittlement tends to be more of an issue at welds than in forged metal, is probably integral to justifying the claimed design life of 60 years.
It’s a “three-loop” design, another design characteristic directly inherited from its Westinghouse (and later Framatome) precursors. That means it has three Steam Generators (SG), each with a single circulating pump. In this area, the other designs with, or in process for UK GDA differ. ABWR has, of course, no steam generators at all (steam raising takes place directly in the RPV). AP1000 has just two large SGs, with a single hot line from the reactor and a pair of cold return legs for each SG with a circulating pump for each cold leg mounted integrally with the SG. EP1000 has four SGs, with a dedicated pump for each.
Externally, it seems to be a compact unit, in appearance very similar to the stations familiar to anyone who’s spent time in France. It has a typical modern “dual containment” design, with an outer reinforced concrete “shield wall” providing protection against missile or aircraft impact, and an inner pre-stressed concrete pressure bearing “true” containment.
Safety Systems
In keeping with the “conventional/evolutionary” theme, HP1000 mixes redundancy-based approaches with some elements of passive safety design. The main systems are relatively standard – the three loops have completely independent power supply trains, and each is equipped for borated water make-up in the event of a pipe-break. In addition, there’s an independent system designed to allow for direct cooling of the core in the case of a loss of conventional circulation; this is fed by a large in-containment tank of borated water. The boron serves as a shutdown agent, as it’s a strong neutron absorber. This system (the “SIS”) is sized to allow removal of peak post-accident decay heat to atmosphere via an independent external heat exchanger. This system is utilised for decay heat cooling when the reactor is shut down, for example during refuelling.
There’s a back up to this system in the form of a passive injection system – tanks of borated water, pressurised to reactor pressures with nitrogen, also held in-containment. This can dump heat into another large internal water tank (the IRWST), which acts as a general reserve for post accident cooling.
There are three independent strains of back-up diesel generators (EPR has four and ABWR and AP1000 two), plus an additional two large “Station Black Out” (SBO) Generators (standard across designs except AP1000, where passivity renders one redundant). In addition, there are to be several mobile generators on-site at all times, and multiple external power hook-us to which they can be connected. This is conventional post-Fukushima design.
So far it’s all pretty conventional. There are however, three further systems worthy of note. The first of these is novel (albeit something similar is being built on very late model Russian VVER1200s). In the event of an accident that does NOT break the primary circuit, it allows for natural convection to circulate water from the RPV to the SG’s with a secondary additional natural convection circuit taking steam to external heat exchangers, and returning condensate in a closed loop. The external heat exchangers sit in a water tank, the contents of which will boil-off during operation. The tank has sufficient capacity for 72 hours of operation, and can be topped up using conventional fire pumps externally.
The second novel system is based on a system of internal heat exchangers within the containment. These are also connected to a small wet cooling tower, and can remove the full peak decay heat load – however, this system is not natural convection-based, requiring active circulation, and is hence dependent on station electrical power. This needs to be seen in conjunction with the third innovative system.
Designers of Generation III+ reactors have taken two distinctive design approaches to the issue of dealing with the implications of a full-scale core melt. One, as represented in the EPR and VVER1200 is to assume that melted fuel will cause a breach of the RPV – a “melt through” – and to add an actively cooled “core catcher” below the reactor. This obviates any possible penetration of the containment basemat, as was feared at Fukushima. The alternative approach is based on the concept pioneered by the AP1000 – that of flooding the reactor “pit”, and allowing boiling off the RPV walls as a means of heat removal from the melted fuel within. This has the obvious advantage of preventing any fuel release, but begs questions of how to remove the heat and allow condensation of the produced steam, and return of the condensate to the pit. HPR1000 uses this concept, with the internal heat exchangers acting as a heat sink for the steam, and (presumably) with return lines for the condensate cast into the containment floor.
An interesting sidelight on this particular design issue is emerging at Fukushima, however, and is also reflective of Three Mile Island. It’s looking increasingly likely, as on-site investigations continue at Fukushima that the feared “melt-throughs” didn’t occur. Nor, in the case of TMI did melt-through happen, despite the melting of most of the core. At Fukushima, although core material has exited the RPVs to a greater or lesser degree in each of the three reactors, it looks very much as though this was via the control rod and instrumentation penetrations, rather than as a result of gross failure of the vessel. So, while this poses a challenge to BWR designers, it’s moot for PWRs – as there are no penetrations in the bottom head of the RPV.
Summary and evaluation
The safety case makes a claim of a 10^-5/reactor-year core damage frequency, and a 10^-7/reactor year release frequency. Compared to the claims of the competing designs, these look if anything somewhat cautions – something I suspect reflects unfamiliarity with western probabilistic safety methodologies.
From a functional/operational perspective, there’s no reason to expect this to behave materially differently from LWRs/PWRs worldwide. There’s little that’s innovative in the design of the Nuclear Steam Supply System and downstream, and this is in essence a close relative of a design of which well over 100 are in operation worldwide.
From a safety perspective, it appears to be a considerable advance over the Sizewell B generation of plant – in fact, with the additional semi-passive systems, I suspect strongly that it will come out of GDA at least on a par with EPR, and in all probability somewhat better. Since it can’t manage heat removal passively post a primary circuit break, it probably sits somewhere below AP1000 in the pecking order; judging relative status to ABWR is harder to do, but I’d anticipate them being similar.
All of that is predicated, of course, on maintenance of manufacture and construction standards, and here I expect the ONR to apply particular scrutiny. Quite how CGN and its supply chain will react to that, I’m less than completely sure.
On the assumption that that issue can be addressed, and that the EDF-CGN tie-up will allow CGN to meet the documentation and analytical standards demanded by GDA, what I’m seeing makes me think it’s unlikely HPR1000 will suffer undue difficulty in passing GDA. And given CGN’s putative interest in the Moorside sight, that opens up interesting possibilities…













Alex. Good post. What are the “interesting possibilities” you foresee?
The potential purchase of the Moorside project takes us into the space of something approaching a series build, and justifying an increased UK participation in the supply chain. It’d mean a build of five UK units.
CGN is assuming at a cost of about $2600/kW for series build in China. While it won’t go that low in the UK, £3,000-3,250/kW should be attainable, which puts a strike price of well below £60/MWh in sight.
Assuming that were successful, I’d argue it should put other sites in play, or make an argument for further expansion at Bradwell. Add to that the fact that the relatively compact site requirements would make space limited sites like Heysham or Dungeness viable, at least for single units.
I’d not rule out eventually seeing 8-10 HPR1000s in the UK.
To elaborate on the strike price point.
Conventional wisdom is that 80-90% of the cost of a unit of nuclear-generated electricity is fixed cost, overwhelmingly capital and finance cost. Fuel (fully fabricated, not raw uranium) is 5-10%, and the remainder other variable cost (mostly working-class driven O&M.
If we apply this to the HPC strike price of £92.50, taking the mid point of the ranges we get
Fixed/Finance and capital related – £78.50
Fuel – £7
Other operational variable cost -£7
Assuming both projects have the same cost of capital – about 11% for HPC (and hence a very pessimistic assumption), and taking HPC at £20 billion for 3.2GW (£6200/kW), we get a fixed cost element of £41/KWh.
Operational variable costs should be similar, and fuel will be somewhat higher, down to the lower burn-up, at about £9.50/MWh.
Which gives approximately (41+7+9.5) = £57.50/MWh.
However, if CGN has a.lower CoC of 8% (and let’s face it, as part of secure build and with access to cheap capital, it should be) the financing cost element drops to £29/MWh, and total pro e to £45.50/May.
I have heard nuclear developers say “never underestimate the ability of regulators to add cost”.
We can hope that with a conventional design, cost increases will be limited.
But typically 60% of the value add is in the UK (which is good), but it does mean that every piece of steel has to be certified NG (nuclear grade), so there is scope for costs to rise.
If it can get down to below £50/MWh, then the logical reaction is to ask how many more can we have and where?
How many reactors are proposed for Bradwell?
Two.
And to be fair, with the sole exception of the separation of shutdown systems from the general C&I system on EPR, I’m not aware of any substantive design changes imposed by ONR in the AP1000 and ABWR GDAs. What change there’s been is at the level of increased fire isolation, etc.
What they have done is work the vendors hard to justify their designs. But at least in theory, that should give less grounds for design change in build.
I wrote the introductory paragraph and did a quick search on number and couldn’t easily find one. The picture Andy had sent me had two – hence I concluded that 2 was the number.
If the UK formed a nuclear consortium similar to the Finnish VVER project, pesent day interest rates should attract a fixed rate lower than 8%. Even the student loan rate of 6% is being called usury. If the UK requested competitive bids from the Korean, Chinese and Russian alternatives and offered 10 sites with finance at 5%, we could solve the problem in 10 years and forget the occasional wind solution.
Or go in halves with the French with EdF. Share the costs given we both needed quite a few new NPPs over the coming decades.
Too late for that now.
The single casting and 7 years build time must make this design or similar viable for the UK and French replacements. Wouldn’t it be wonderful if we could manage to have electricty at £60/MWh and standardisation like the French instead of every one a different design to maintain and approve. EDF must be wishing they could get off the EPR hook as the safety levels are similar. Any chance we have civil servants and ministers who can follow this Fuquing design logic?
My understanding is that once in Series build, build time is due to come down to between 5 and 6 years.
steve,
Standardisation sounds too much like common-sense. I mean, the whole point of “privatisation” is to have “competition” which automatically means different solutions.
As regards the “7 years build time”, I can only guffaw. I mean, even the Finns are having trouble sticking to their build time.
My introduction to the practical end of civil engineering was a short spell at Dungeness B. It was already 4 years late when I showed up and it took some 23 years to build versus the planned 5 years.
Like you, I worked on the AGRs – in my case, Heysham II and Torness. It’s safe to say we didn’t cover ourselves in glory on that programme – 7 stations built to four different designs, and worse, each pair building with only a six month lag for the second station. That made absolutely sure that the second station ended up with no time to apply learnings from the first.
In series build, times do come down radically, though. China averages about 5 years to build the precursor of the HPR1000, and the Japanese had construction time for ABWRs down to 39 months.
It’s worth thinking about what is the main virtue of standardisation, though. It’s the “learning curve”. And Bradwell B1 and B2 look like being perhaps the eighteenth and nineteenth HPR1000 units built – The Chinese are planning four units at Fangchenggang and two each at Changjiang, Fuqing, Ningde, Zhangzhou, and Huizhou, plus two at Karachi and one in Argentina, all breaking ground well ahead of the earliest possible GDA exit.
By the time ground’s broken at Bradwell, this will be a mature design.
I take that back.
Bradwell B1 and B2 will be the nineteenth and 20th HPR1000s.
http://www.world-nuclear-news.org/NN-Pakistan-China-agree-to-build-Chashma-5-2311177.html
The AGRs had their problems, but have they turned out all that bad?
I guess there must still be wrinkles else they’d still be being built?
AGRs were just expensive. The design was basically pretty good.
No, they weren’t.
They’ve all been derated well within design life. They were bastards to build.
And the real killer is that they can’t do the one thing that was supposed to be their particular advantage – on load refuelling.
The UKAEA were behind the AGR which was a British design. British designed gas cooled reactors had been sold around the world in the 1960s and the Government hoped the AGR would be exportable too. The lead AGR plant at Dungeness took too long to build and by the 1970s the world was more interested in LWRs.
The CEGB had wanted to build PWRs too. When Thatcher took over, the CEGB arranged for her to visit France and she saw their PWR’s being built to schedule. The rest is history.
BTW Dungeness is not refuelled on load, but the rest of the AGR fleet is, although at part load not full load as originally designed.
KEPCO allegedly charge the equivalent of $30/MWh for their nuclear power. NOw that really would be wonderful. We might even be able to restart Anglesey Aluminium.
Thanks Andy. Any thing on the instrumentation and control?
This will be a highly controversial area given the suspicion that China embeds back doors in their technology. There is a similar concern with Russian designs which is one reason why Rolls Royce is providing control and instrumentation on the new Finnish reactor.
The UK could gain some commercial advantage if the instrumentation was provided by Rolls Royce, as that would be the reference case for all other HP1000s sold outside China. It would be a bit “unfair” of the UK GDA to encourage this (and probably not legal), but it’s what other countries would do.
I believe the default is an Areva-Siemens system, although Chinese built units will gradually transition to a locally derived system.
The RR option is a good one; I’d not be surprised to see that “emerge”
What about general quality for equipment rather than generic design.
Poor quality control has dogged the reputation of Chinese manufactured goods, I’ve heard horror stories of low competence in China’s Engineering Service industry. My question is can we trust them with a nuclear power plant. After all the reason China wants to build a reactor in the UK is
for the prestige of a build in a strict regulatory environment and repair Chinas poor reputation for quality. Normally I would complain that the GDA takes to long but in this case perhaps not long enough.
From my own point of view this is another foreign design that will send all the engineering overseas.
Has anyone else noticed that very little Engineering work is being released to UK Engineering from EDF Hinkley Point. Starting to suspect we were completely mislead with the promise 60% of the supply chain and most of the work going to France. What makes me laugh when the industry talks about skills shortage in the nuclear industry.
How long does it take to build a nuclear power plant?
The answer is on average 7.5 years. There is no reason why with a serial build the UK should not achieve this. The UK government needs to treat all agencies that seek to hinder this objective as enemies of The State.
The above post has incidentally been resurgent this year getting 20 to 30 reads / day.
I also wrote a post called Nuclear Options where I reviewed the build options facing the UK. Somehow the Hualong-1 did not figure on my list 🙁
In terms of sites with space, I live in hope that Torness and Hunterston may one day make the cut. But implementation of Helm may have some interesting and unforeseen cross border consequences.
Average build time has got longer with time. The average construction duration of the early nuclear power reactors built globally (i.e. all countries) was: 3.5 years for the first three, 4.0 years for the first ten, 4.4 years for the first twenty, 5 years for the first thirty, and 5.4 years for the first eighty [22]. The first completed US power reactor was constructed and sending power to the grid in 1.8 years [23,24]. That was 60 years ago.
[22] IAEA (2016), http://www-pub.iaea.org/MTCD/Publications/PDF/RDS_2-36_web.pdf
[23] ASME, http://www.asme.org/getmedia/3663519d-0882-4b7e-ab6c-f036b080cfdd/128-vallecitos-boiling-water-reactor-1957.aspx
[24] IAEA (2016), https://www.iaea.org/pris/
I’d share your hopes re Hunterston and Torness (Torness certainly has lots of space. My memories of the site are that there’s bugger-all for miles…). And a quick look at google maps suggests plenty of space around Hunterston.
But remember, there are two English sites with new-build approval but no committed projects and one other that has strong local support and is a current site – respectively, Hartlepool, Heysham and Dungeness. I think any or all of those would be ahead of the Scottish sites in the pecking order.
At least the first two of those would be politically easier than Scottish new build -strong local support, plenty of space, and things like thermal impact will largely be offset by the impending closure of the current stations.
Hartlepool has huge amounts of space, and a strong local history of heavy industry. And given it’s (a) next to a graving yard that specialises in dismantling ships with heavy asbestos contamination, and (b) off “Zinc Works road”, there’d appear to be few ways the local environmental position could be worsened by new build
https://www.google.co.uk/maps/place/EDF+Energy/@54.6361804,-1.1865037,1339m/data=!3m1!1e3!4m5!3m4!1s0x487ef22546e4a235:0xc8916c6d5fb9ba37!8m2!3d54.6354913!4d-1.1811072
Heysham’s less attractive. Although it’s fully permitted, etc., the site is limited in space – like most nuclear stations, it’s surrounded by a nature reserve. The available land is also an inconvenient shape, which would preclude two units side by side as is usual practice. In reality, I could only see a single unit sited there.
Plus, there’s another issue. As a small site, it’s arguable the perfect placement for an SMR prototype, leaving space for additional subsequent units.
https://www.google.co.uk/maps/place/Heysham+Nuclear+Power+Station/@54.0279955,-2.9129716,679m/data=!3m1!1e3!4m5!3m4!1s0x487b6034293e6d0f:0x166b7bd36eefcfb0!8m2!3d54.0301526!4d-2.9153037
Dungeness is more problematic – it’s surrounded by a “Site of Special Scientific Impact”, plus (I gather) that cooling water is limited, unless intakes are run out a long way into the Channel. It was also omitted from the initial list of new build sites, on the grounds that the site is potentially subject to coastal erosion. However, there would appear to be plenty of space for at least one, and possible a twin unit there:
https://www.google.co.uk/maps/search/dungeness+power+station/@50.914636,0.9541772,731m/data=!3m1!1e3
Don’t build over the return loop of the Romney, Hythe and Dymchurch narrow guage railway! I had the pleasure of a return trip all the way from Hythe with a picnic on the very stony beach at Dungeness, with the reactor buildings for backdrop as a young schoolboy.
EDF recently bought the Dungeness estate, so own a fair part of the SSSI. RSPB have a lot more. The other big neighbour is the MOD live firing range.
The coastal errosion is a problem for the whole romney marsh and a developer ready to foot the bill for the miles of defences needed between camber and dungeness would probably get good support… it’s the MOD’s stretch.
I’ve been a regular at dungeness for many years and know the locals are divided.
The inlet screens get a lot of fish and the odd seal, but i wasn’t aware of thermal limiting.
Euan, the measure used to determine reactor build time in the west is the period between first nuclear safety concrete and reactor commissioning. First concrete means the pouring of concrete for the foundations of the reactor building. From that point to commissioning should take about 60 months for a trouble free PWR schedule.
All sites are different in terms the works needed before the first concrete date. As well as laying out the construction site and providing temporary buildings, some need extensive terracing work and earthworks, perhaps a seawall too. The time needed to reach first concrete is variable depending on the site so tends to be excluded from comparisons of build time from site to site. So the key comparator for PWR build time is the period from first concrete to commissioning.
Concrete batching plants needing to be set up years beforehand too. They have to proven and their product tested extensively before being used to supply concrete for nuclear safety critical construction. The batching plant at Hinkley was erected early and looms over the site currently.
ADDENDUM AND ERRATUM
Re-reading the above, there’s an item I need to add, and one I need to correct:
First, the addition. I’d said I’d comment on the use of (again conventional) zircalloy cladding.
Zirconium is wonderful stuff for cladding nuclear fuel rods. It’s got excellent neutron absorption characteristics (it doesn’t), good thermal properties (good conductivity) and attractive mechanical behaviour (reasonably strong, and ductile so it can accommodate fuel swelling). It’s massively better than the various alternatives, for example the Beryllium that was the original choice for AGRS (yes, honestly…).
There’s a downside, though. Heat it to about 850C in the presence of water and it oxidises, liberating hydrogen.
This was the cause of the explosions at Fukushima; it also caused some concern at Three Mile Island (and panic in Jimmy Carter), although in that latter case there were no real grounds for concern. Hydrogen did form a bubble within the RPV, but as there was no oxygen present, there was no explosion risk.
Fukushima was different. BWRs are designed to blow down pressure into the containment, which then has suppression mechanisms. Unfortunately, the containments had already been subjected to overpressures, damaging seals and allowing hydrogen to leak into the refuelling halls at the top of the reactor buildings. This is what exploded.
Fukushima did have hydrogen removal systems – unfortunately, though, they were electrically powered, and hence rendered inactive by the station blackout.
HPR1000 had hydrogen recombiners in the containment, should a primary circuit breach and core damage lead to hydrogen release outside the primary circuit. Unlike Fukushima’s, these are catalytic in nature, and hence entirely passive (meaning they work independently of power supplies)..
Now the correction.
In the week since I drafted the post above, there have been reports that robotic inspections of Reactor 1 at Fukushima have shown a “hole” below the reactor. The same reports state that there are “icicles” of fuel hanging below the control rod penetrations. Normally, I’d interpret that as an indication that the bottom head of the RPV had seen a melt-through, contradicting my comments in the article; however, when the reports first came n about the reactor 2 inspection, they conflated damage (a hole) to the working platform below the reactor to damage with the RPV itself.
Watch this space (or rather, watch Les Corrice’s excellent “Hiroshima Syndrome” website, where Les issues regular updates on developments at Fukushima)
Thank you for this very readable and informative article, Andy.
As one of the reactor designs innovative features, you note:
“This may also have some relationship to another striking feature; the entire main Reactor Pressure Vessel (RPV) is produced as a single forging, as opposed to the mixed forging/welding of other designs.”
How large and heavy is the RPV? I presume that it can be transported by sea/large river and then moved the short distance into the main building by means of a specialised vehicle and cranes. Or might it be forged on site?
For anyone interested in Beryllium:
“It’s (Zircalloy) massively better than the various alternatives, for example the Beryllium that was the original choice for AGRS (yes, honestly…).”
Beryllium metal is a low density and strong metal, with high thermal conductivity. It has a number of particular uses in x-ray equipment, and can be alloyed in low concentration with Copper to make non-sparking tools, amongst other things.
However, both the metal and its compounds are extremely toxic; perhaps surprisingly, they also have a sweet taste. It substitutes for Magnesium in the human body causing long term irreversible illness, even at low concentrations, and would rank as one of the least desirable materials to be ejected into the environment in case of an accident.
As to the pressure vessel, a typical PWR vessel and internals is of the order of 500 tons. Dimensions would be around 5 metres in diameter and 10 metres long. A large load, but not immovable.
All proposed new build sites, with the arguable exception of Oldbury are coastal and have reasonable good port access (or, in the case of Moorside, will have a jetty built specifically). So no, there’d be no issue of on-site vessel fabrication.
As to the beryllium, you missed out extremely expensive, and a bastard to work.
Think of it this way – would you fancy welding end caps on a beryllium tube in order to close up a fuel pin?
I forgot to add that the RPV forge would also need to be nearish a jetty at the other end. 500t on a road transporter would be slow but its small diameter of 5m would permit passage under bridges, if not over them.
“Think of it this way – would you fancy welding end caps on a beryllium tube in order to close up a fuel pin?”
Beryllium is one of the elements, although fascinating, that I would avoid even as a metallic sample. I would never work with the stuff.
Thank you for your reply, Andy.
Perhaps the process of the ONR actually doing some regulating on new design should raise the question as to whether its standards really are appropriate, or whether they constitute greenplating designed to make nuclear expensive unnecessarily in a bid to deter investment.
In the present circumstances, we should be considering South Korean designs. They seem to be consistently the cheapest to build among Western suppliers at around $2,500/kW for their APR1400 design being built at Shin Kori (and $24.4bn for 5.6GW or $4,360/kW at Barakah, UAE), and the troubles in the peninsula have resulted in political threat to their own domestic construction, albeit the programme of builds in the UAE continues. Perhaps we should offer to buy out their industry and re-acquire some expertise in the business.
Design is one thing and construction another, as you point out. We used to have industry capable of heavy nuclear engineering until our governments undermined it in every way possible – no domestic demand to cater to, expensive energy, selling off our expertise, lack of education of a new generation of nuclear engineers. Much the same applies to instrumentation and control systems that are reckoned to be a key vulnerability in letting China build in the UK: the idea that they could simply cause a shutdown, or worse an accident by remote control has hardly been lessened by the saga of the Stux virus.
http://www.world-nuclear.org/information-library/country-profiles/countries-o-s/south-korea.aspx
https://www.reuters.com/article/emirates-nuclear-southkorea-idUSL8N1CQ5HB
Just a few observations:
A lot of the risk occurs during shutdown, because many of the safety systems are designed for power operation and are disabled as the system is depressurised and also because the water level in the vessel has to be lowered to remove the upper head. I wonder how this is taken into account.
I would say that the 900MW Framatome design was based on technology that was pinched rather than licensed from Westinghouse. The first 3-loop PWR in France was built to the Westinghouse design, but Framatome had all the drawings from Westinghouse and subsequently said goodbye to Westinghouse and proceeded to build them without any Westinghouse involvement. To say that Westinghouse were upset by the French behaviour would be an understatement. However, Framatome subsequently had problems as the result of losing operational feedback from Westinghouse.
PWRs do have penetrations in the bottom head of the RPV. These are for in-core flux measurements, there being one penetration for each fuel assembly. Obviously the penetrations are of very small diameter.
Sizewell B was built in about 5.5 years.
Philip, Re the bottom head penetration.
No that’s not the case for almost all Pets. It seems to be a particular quirk of the Combustion Engineering design used at TMI.
The HPR1000 documentation makes an explicit statement that there are no penetration in the bottom head.
I’m all for free trade, but energy security is a bit different. If we can’t do it ourselves, strategic allies only. Basically, for NPPs, French, American or Japanese.
A NPP is for 60+ years. Would anyone here like a bet that when China finally becomes a dominant super power within that time, it won’t flex its muscles in a way the UK is opposed to?
China is an authoritarian state, with growing military ambitions and territorial ones if their behaviour in the South China Sea is anything to go by.
It just goes to show how far the UK has fallen, and how desperate our energy policy has become to sign up to all this. Way too risky.
The reactor is not controlled or managed by China, nor it it based on foreign soil.
Fair enough. I thought it was a joint venture between EdF and CGN. Both of whom are State controlled.
My mistake.
CGN wouldn’t have a foot in the door in the UK if EDF could finance 100% of HPC as they originally intended before the true costs emerged. When Centrica withdrew from EDF Energy’s new build programme, in part due to rising costs, EDF Energy had to look for other investors. Private investors were not interested. Also other EPR builds were having serious problems with cost overruns and delays. EDF and CGN are in partnership building EPRs at Taishan in China. EDF Energy had to bring in a new equity partner for HPC or abandon the project. CGN knew this and in return for promising to invest in HPC and SZC, CGN said they must have a site in the UK and EDF’s help to licence the Hualong design as a European showcase for their nuclear tech.
When EDF Energy bought British Energy they acquired land at Bradwell and then additional land was bought from the NDA through a site auction. Bradwell was seen as an alternative site for one EPR reactor in case only one EPR could be sited at Sizewell. EDF Energy now intend a twin EPR development at Sizewell.
Abdy: Great balance between comprehensible and technical. Thank you!
I agree especially on your price estimates.
This is what Hinkley SHOULD have been.
It is good to see this technology discussed by experts. As a non-expert in this field I would like to take advantage of this opportunity to ask two questions that concern me.
First, what is the technique/requirement for disposal of radioactive wastes?
Second, what do the experts think of the potential for thorium reactors, which are said to produce much less radioactive waste. Two experimental thorium plants have been or nearly are completed in India and the Netherlands. Is there any recent information on them?
Thanks,
David
Radioactive waste really depends upon who you ask. “High level” radioactive waste is basically the fuel rods after they’ve burnt up enough of their fissiles to no longer be usable. When you first shut down a reactor for refueling, these spent rods are ferociously radioactive and need pumped water to keep them cool inside the reactor vessel. After several days, the short-lived isotopes have decayed away and the rods can be transferred to “spent fuel pools” where they are stored underwater. They are still fiercely radioactive (exposure to a single vertical rod at 1m distance will deliver a dangerous dose in a couple of minutes) and have to be kept immersed or their temperature will rise to the point where the zirconium cladding will start to oxidise. After 5 years, they can be transferred to “dry cask” storage – basically big concrete cylinders with the rods spaced on racks inside. A nuclear power plant can have enough dry casks to store all the fuel it uses over its entire lifetime and you can happily play poker right next to them every lunch break with no radiation risk. Dry casks have a design life of about 90 years but you can always transfer from old to new.
After 40 years, a human could handle a single fuel rod for about 20 minutes with a very small risk increment (100 mSv dose.). After 150 years, this would extend to about 2 hours. After 600 years, you could store the things under your bed pretty safely!
So why is there a huge deal about radioactive waste being dangerous for tens of thousands of years? Well, there the standard being used is for the waste to return to the same level of ORAL biotoxicity as the natural uranium it started out as, which is frankly ridiculous. But that’s why people are building deep geological disposal sites.
Matt,
Many thanks to you and others for the replies to my questions. Can you recommend a book or other publication on radioactive disposal along the lines you outline?
David
This article in Dutch is not optimistic:
https://fd.nl/morgen/1227198/kernenergie-comeback-met-thorium
When can we expect the first power station?
That is pretty uncertain. The Nuclear Research and Consulting Group (NRG) has been conducting experiments in Petten since September, but it could take up to 30 years for a full scale plant to be constructed.
The research programme is quite lengthy, with the current experimental reactor due to be operated for a decade and more:
https://nlslash.nl/DTMI/170519_DTMI_&_DIMOS_Projectomschrijving.pdf
Thorium has gained a lot of publicity in the past few years. It shows some promise, but is possibly overhyped. Fission reactors currently come in two flavours, thermal-neutron-spectrum (most of them) and fast-neutron-spectrum (only a few left in the world, all in Russia as far as I know.) Thermal spectrum reactors have fissile U235 (about 5%) mixed with non-fissile U238. As they “burn” the U235, some of the U238 is converted to fissile plutonium Pu239 and also burned. However, less fissile fuel is created then burned and eventually the fuel rods need changing. Fast-spectrum reactors can actually create more Pu239 than the U235 they use up, and so are known as “fast breeders”. A fast breeder can therefore eventually burn ALL its uranium by converting it into plutonium.
Thorium is not in fact fissile. It’s strength is that it can be converted to fissile U233 in a thermal spectrum reactor. You can in fact mix some thorium into conventional fuel rods in pressurised water reactors and it will be converted and burned. This works better in heavy-water type reactors such as Canada’s CANDU, and the Indian reactors are also heavy-water types.
In theory then, you can run a thermal-spectrum reactor with thorium and a little fissile U235 (or U233, or Pu239) to start off, and ALL the thorium will eventually be converted and burned. Nothing’s quite so simple though: whether using uranium in a fast-spectrum breeder, or thorium in a thermal-spectrum breeder, you have to make sure you don’t end up with too much fissile material in your reactor, and you have to get rid of all the fission-product “ashes” or they’ll build up and interfere with the reaction. That means you have to be continuously cycling and reprocessing the fuel, which is messy and hazardous.
Thorium has also been publicised alongside an entirely different design of reactor, the Molten salt reactor. Instead of having your fuel as solid oxide pellets sealed in zirconium tubes, you have your fuel as a molten salt. This has some advantages – gaseous fission products bubble out and can be captured, operating temperature is higher, can run at atmospheric pressure, volatile fission products are chemically bound, meltdown is no longer a failure mode. Molten salt reactors don’t have to use thorium though, and so far only experimental reactors have ever been run. The Netherlands experiment is not a molten salt reactor. Instead they are irradiating a molten salt in a conventional research reactor to see how it responds. The Indian reactor, as already mentioned, is a heavy-water reactor that can use thorium.
The current news provide an Article by Prof. Philip Thomas, comparing the effect of a nuclear accident to the pollution in the city of London.
https://www.thetimes.co.uk/article/nuclear-disaster-fallout-would-be-no-worse-than-living-in-london-706w9xc6h
the problem with him is simple: his assessments in the past were very optimistic about nuclear and extremely pessimistic about renewables (he got it wrong by about 20 years).
https://www.city.ac.uk/news/spotlight-on-research/low-risks-for-new-nuclear-builds
But he is an interesting source and you can see his opinion about safety on video here:
https://vimeo.com/143845218
@ sod
It would be useful if you could post to the claims that you say are wrong so we can assess them.