Guest Post by Andy Dawson who is an energy sector systems consultant and former nuclear engineer.
The UK government has launched a competition to select a design of a small modular reactor (SMR) for future deployment in the UK. The idea behind SMRs is that they can be factory built and stamped out like aircraft and transported to location on the back of a truck. With thirty-three companies / designs on the shortlist, this looks like the process could take a while to complete.
In this post, nuclear engineer Andy Dawson provides an overview of SMR technology together with descriptions of the leading contenders.
A few weeks ago, Euan was kind enough to publish a piece I wrote on the possibility of an all-nuclear UK generation system, supported by pumped storage. In the responses to that, I was asked several times to comment on the virtues (or otherwise) of various reactors types either currently intended for construction in the UK, or other possibilities. Subsequent to that, it’s been announced that there have been 38, now cut down to a “shortlist” of 33 (we’ll leave aside for the moment the thought that only a Civil Servant could consider 33 a short list…).
On the basis of which, it occurred to me that a short exercise to run through the “runners and riders” likely to be in the competition – I should stress that since the listing is still “commercial in confidence”, I cannot state with any certainty that any of the firms and designs I’ll go through are definitely “in play” – these are simply the expectations of what I’d like to think is a reasonably well-informed observer.
It’s worth spending a moment going through the background to the process, albeit I don’t intend here to debate the positives and negatives for the case for SMRs; I will say, however, that I have some doubts that they’ll necessarily be quite so transformational as some proponents believe; my views are close to those of Admiral Rickover, the creator of the US “nuclear navy”:
“An academic reactor or reactor plant almost always has the following basic characteristics: (1) It is simple. (2) It is small. (3) It is cheap. (4) It is light. (5) It can be built very quickly. (6) It is very flexible in purpose. (7) Very little development will be required. It will use off-the-shelf components. (8) The reactor is in the study phase. It is not being built now.
On the other hand a practical reactor can be distinguished by the following characteristics: (1) It is being built now. (2) It is behind schedule. (3) It requires an immense amount of development on apparently trivial items. (4) It is very expensive. (5) It takes a long time to build because of its engineering development problems. (6) It is large. (7) It is heavy. (8) It is complicated…”.
Although the process was announced formally by the Chancellor in last November’s Autumn Statement, it has in reality been running for some time – the National Nuclear Laboratory was instructed in 2014 to prepare a feasibility study on UK engagement in an SMR market, and to review then available designs, which concluded that such was viable, and that there were a number of viable designs – that viability being based on a combination of:
- Technical maturity and viability
- Maturity of Safety Case and certification
- Key strengths and areas for development
- Programmes to address development needs
- Available resources – people, capability, facilities and funding
- Economic viability.
I’ve seen nothing in the process subsequently to make me think that the competition assessment will be conducted on any other basis (with the possible exception of the addition of a large UK role in the supply chain being mandated). The questionnaire submitted to test the firms making Expressions of Interest is consistent with that view, placing heavy stress on the readiness to enter Generic Design Assessment, stability and technical resourcing of the consortium/company, and willingness to make an early commitment (for example, pre-siting decisions). To be clear about one thing – the intent is to select a SINGLE design.
The broad classes of candidate designs:
Small Modular Reactors (SMRs), like larger reactors, fall into a number of major classes. A short run through these is in order, before moving onto the specifics:
Light Water Reactors (LWRs):
LWRs, for good or ill, form the background of the global fleet – and hence represent the class of which there’s by far most design and operational experience. The common characteristics are:
- Light water acting as coolant and moderator (all LWRs use the thermal part of the neutron spectrum)
- Steam raising – either directly as in Boiling Water Reactors (BWR), or via an indirect circuit as in Pressurised Water Reactors (PWR)
- Using low enrichment fuel (~5%) clad with zirconium or other low-absorption alloys
- Operate at significant pressure – c. 80 bar in the case of BWRs, c. 150 bar for PWRs.
- Low to moderate thermal efficiency, due to temperature limits imposed by the use of water as coolant.
LWRs require a “containment” capable of bearing significant overpressure in the case of a failure in the “primary circuit” (those parts of the reactor containing pressurised coolant); containment designs vary from large volume/moderate pressure typically used for PWRs( ~2 bar), to close-fitting, low volume designs for BWRs (~5 bar), relying on steam-suppression systems.
PWR containment volume is largely determined by the need to accommodate several steam generators, which are considerably larger than the reactor vessel itself, plus other systems like coolant pumps, the pressuriser and various safety related systems. BWRs do not require steam generators, operating “direct cycle”. Ancillary systems are also somewhat simpler, and things like emergency make-up tanks are usually installed externally to the containment.
LWRs typically operate an 18-month to 2 year refuelling cycle, requiring shutdown and depressurisation, although some military versions run as “Fuelled for life” using high enrichment fuel.
Liquid Metal-cooled Reactors (LMRs)
Liquid Metal cooling is almost invariably associated with “fast” reactors – as such they need no moderator, and hence can be considerably more compact that thermal designs. The use of liquid metals brings its own challenges, however – the two main candidates are Sodium and Lead. Both have many remarkably benign characteristics as a coolant – extremely good heat transport characteristics, and very high boiling points (and in the case of lead, heat capacity). The challenge with them has historically been chemical – Sodium reacts, of course, with water, and the behaviour of any oxygen in lead coolant is surprisingly complex and corrosive of many engineering materials. Both allow reactors to operate at high (600C+) temperatures, hence with high thermal efficiency; Lead in particular offers opportunities for extremely high temperature operation (800C+), potentially usable for process heat applications.
Liquid metal cooling obviates the need for pressurisation, albeit it remains necessary to isolate the coolant from atmospheric contact (achieved by blanketing with inert gas); Lead coolant, through its sheer mass, can pose challenges to design against seismic events. In recent years, LMRs have been very much a Russian specialism, having brought a new unit into operation (“BN800”) as recently as last year. Western experience has been chequered; some systems worked well (most notably “EBR II” in the US) and others less well (PFR at Dounreay, Superphenix at Creys-Malville, France).
High Temperature Reactors (HTRs)
HTRs are usually – but not exclusively – thermal in spectrum. They’re designed to operate at 800C+, using helium coolant. In many ways they can be regarded as a descendant of the Magnox-AGR design concepts, using graphite moderator and gas coolant. However, the fuel concept is very different – it’s based on a concept called “TRISO” (TRIstructural iSOtropic), which a ceramic form – typically pellets of Uranium Carbide, embedded in a matrix of pyrolitic Carbon, and the whole element enrobed in a layer of Silicon Carbide. It’s EXTREMELY tough and tolerant of very high temperatures. It basically integrates fuel and moderator into either a single prismatic block, or into “pebbles”.
HTRs do operate at some pressure, albeit not overly high – 40bar would be typical. Although a number of prototypes were built in the 1960s and 1970s, the field have been pretty much moribund
in recent years. However, it’s been reopened by China’s construction of a pair of 250MW modular prototypes at Shiadowan
Probably the biggest downside to HTRs (beyond immaturity) is that the fuel is just about completely un-reprocessable. That can be argued as a virtue, for proliferation reasons, but it may hurt plant economics in the long run.
Molten Salt Reactors (MSRs)
I enter this topic with trepidation… within the pro-nuclear community, there’s no topic so polarising between enthusiasts and doubters. I’ll declare my hand up-front – I suspect that the concept may have virtues (but perhaps no more than LMRs), and I can see significant issues ahead, not least because most designs are “academic” in the Rickover sense. That confusion is not helped by the sheer proliferation of MSR concepts – ranging from this with discrete fuel and molten salt coolant (which, contrary to popular belief, is the nature of the ongoing Chinese work), to those operating with fuel dissolved in coolant, to breeder concepts based on either Uranium or Thorium. On that latter note, let’s be clear – there are no Thorium fuelled designs. The Thorium is “bred” into U233 in order to be fissionable.
Whatever your views on MSRs, one thing should be common; the designs that are out there at the moment are immature at best (most are best described as conceptual), and many of the companies proposing them are under-resourced and lack nuclear sector experience. On that ground alone, I anticipate that most will struggle with processes like GDA and in particular Probabilistic Safety Analysis.
Assuming a design can be brought to maturity, MSR can have design advantages – lack of pressurisation, high temperatures, etc. On the downside, there are needs to demonstrate the durability of structural materials, radiolytic problems with moderators (or the need to validate assumptions about core-replacement options), handling of fission products that are outgassed or extracted from fuel salt, and so on.
In more than one sense, “Nuclear Batteries” do not represent a distinctive technology – in fact, the various designs that have been announced cover a range of technologies, fast and thermal, various coolants, unusual control mechanisms and so on. However, I think it’s worth separating them out – not least because they have potential applications beyond power generation. Some designs have potential to form the basis for propulsion power units for shipping, for example.
They’re typically small – the largest design that I think legitimately falls into the category is would have about a 10MWe capacity; They’re designed to be “sealed for life”, installed into remote sites or as local power sources and swapped out when the fuel is exhausted. They’re designed for zero operator intervention in normal operation, and extremely high levels of inherent safety.
Runners and Riders
Having run over the basics of the various technologies, let’s now start to look at who’s definitely in the competition, or is suspected to be there. One note I should add is that if 30+ designs remain in, there are obviously some very under-developed designs in there, or vendors who’ve not broken cover yet.
The leading pack
Both from looking at the NNL report, and seeing who’s being mentioned in the press from leaks, we can pick out a small group of designs that are either recognised as “in”, or are so unlikely not to be there that I’m willing to risk embarrassment by publically assuming that they are:
The Westinghouse SMR –
Probably one of the most mature designs in the market, this is what I suspect to be the most probable winner. The design makes heavy use of technology derived from the AP1000 full sized PWR, which is currently nearing the end of its GDA process and is planned for the Moorside site. At 225MWe, it’s one of the larger designs (the upper limit is approx. 300MWe), which is a good match for “off the shelf” steam turbines used in the back-end of CCGT units.
Westinghouse (owned by Toshiba), have moved early to establish a strong potential UK supply chain – they’ve aligned with the Nuclear Advanced Manufacturing Research Centre (NAMRC) of Sheffield University, and through that are already working in collaboration with firms like Sheffield Forgemasters, Weirs and so on. The design utilises the same fuel as the AP1000, which will be manufactured at Springfields, near Preston.
It’s what’s called an “Integral” PWR – that is, all the major components are incorporated into a single vessel, factory manufactured and assembled on site. In broad terms, the upper half of the vessel (which bolts off for refuelling) contains pressuriser and heat exchanger, the lower half the core, pumps and control rods and actuators (it uses the same submerged actuation system as the AP1000, which has great virtues in terms of things like vessel integrity).
It’s a relatively conventional design (albeit integral reactors are a new concept), heavily based on proven or near-proven technologies, and of all the probable vendors, Westinghouse has most recent experience of the GDA process. Physically, the Reactor Vessel is approximately 25 metres tall by 3.5 in diameter. The Containment is close fitting, but still about 27 metres tall by almost 10m in diameter – which makes me think it’ll have to be constructed on-site from factory-supplied modules. It’s designed to be installed below-ground level, and with the containment submerged in a water tank, acting as a heat sink in the case of accident.
If we’ll allow divergence into personal opinion, rather than fact, the rating of this is close to a “sweet spot” – not so large that it represents a massive addition to a grid, with the associated cost and risk, but not so small that it would struggle with scale economies, numbers in deployment and so on.
NuScale is another integral PWR, but of notably smaller capacity – 50MWe/unit. Conceptually, that aside, there’s little major difference from the Westinghouse unit. Surprisingly, the Reactor Vessel is little difference in size from Westinghouse’s – quoted as 20 metres tall and 2.75 metres in diameter.
The containment is 23.4 metres tall and 4.6 in diameter. I understand the plan is to ship a combined Reactor/Containment unit as a single assembly.
Perhaps the main delta between the two designs is the NuScale’s use of an entirely natural circulation model – the design has no coolant pumps. The Westinghouse unit uses such passive flow in an accident situation, but is actively pumped in normal operation. Again, the containment is submerged and below ground. The intent is for units to be deployed in cluster of up to about a dozen, feeding one or two turbines. Unusually, for refuelling an entire unit would be lifted out of place and craned to a separate refuelling pond, which the remainder of the plant continued operation.
NuScale looks likely to build a demonstration plant at the US DoE’s Ohio site, and has won Federal support funds.
Again, at the level of personal opinion, I have some issues – 50MWe is probably just too small, and it’s hard to see why a 50MWe NuScale module would be built at ¼- 1/5th the price of a Westinghouse module. It’s inherent safety is good – but that applies to all the “leading pack” designs I’ll be discussing.
What Rolls are proposing is, at the moment, something of a mystery. All that’s emerged is the capacity – 220MWe; and approximate dimensions. The “Economist” quoted 16 metres tall and 4 in diameter. If those are correct this is almost certainly NOT an integral design – it’s simply too short to accommodate a heat exchanger/pressuriser above the core.
That would be consistent with Rolls’ submarine experience (integral designs are hard to fit in the space constraints of a submarine).
Non-integral designs will struggle somewhat in that they cannot be fully factory assembled – the installation of separate steam generators entails more site-works, and mitigates against a compact containment; however, I gather that Rolls do have some experience with passive designs (PWR3, destined for the Trident replacement boats is reputedly completely natural-circulation based).
So, given the uncertainty, why have I rated Rolls as a leading contender? Simple – it’s almost certainly the only credible bid led by a UK company.
There are several designs which look likely to make it into the latter stages of the contest, but are so far “below the radar”.
Few observers of the nuclear scene have failed to be impressed by Korea’s delivery of the Barakh project in the UAE on time and at a VERY competitive cost (under half the cost/MW of Hinkley C). Similarly, anyone who’s had cause to look at the APR-1400 PWR used in that project will have concluded that it’s anything other than an extremely competent design.
KAERI stands for the ”Korea Atomic Energy Research Institute”. In addition to supporting their industry on LWR development, and being probably the world leaders in electrolytic/pyrolitic reprocessing, they’re in the midst of designing their own mid-sized modular fast reactor.
Somehow, they’ve found the time, while doing all that, to produce the first SMR design that’s passed a national regulatory accreditation (their own, in 2012), at a spend well north of $300m. It’s a 100MWe design, again integral, and fully passive post-shutdown (for 20 days). There’s an AP1000 style vessel submergence system for passive heat removal.
The downside is a conventional, large volume containment, which is above-ground.
As you’d expect from having passed accreditation, this is probably the most thoroughly worked up design likely to be in play. Having said that, KAERI in particular and the Korean sector in general has never taken a design through a Western accreditation process, so it would probably need a local partner. It has, however, been confirmed a that a Korean design has been entered, and I’d be staggered if it were anything else (having said which, there’s a rather clever Nuclear Battery design – discussed in that section).
I’ll start by declaring I’ve no confirmation at all that this design has been put forward – notwithstanding that, this is a design with major technical virtues, and has the backing of an experienced and competent vendor. Mitsubishi HAS taken a reactor design through US certification, which is similarly onerous to GDA, but it’s unclear how far the design has advanced into detailed design.
Once more, it’s an integral design. Where it’s unusual is that it’s a BWR-PWR hybrid – that is, it permits boiling in the core and what’s described as a “bubbly” flow upward to the steam generator. That generator is somewhat unusually sited, in a flared section of the upper Pressure vessel. The steam-water column rises, and descends through the steam generator before returning to the core. As such, it needs no pumping in either normal operation or post-accident. It uses in-vessel control rod drives, thus has minimal vessel penetrations, and is designed for 100MWe in normal operation.
It’s claimed that a conventional Emergency Core cooling system can be done away with, as a result of the vessel design, the natural circulation cooling and external vessel cooling (presumably flooding) – although I think there’d be some resistance to that concept from UK regulators.
Inevitably, it seems there’s a Chinese design in play (see the link provided under the KAERI SMART section above).
CPR100 is a scaled down equivalent to the Hualong design that’s undergoing preliminary processes for the UK GDA at the moment. It’s undergoing a preliminary IAEA review which should complete sometime soon:
I believe, although I can’t confirm that it’s the same design that’s mooted for the “floating power plants” intended for deployment in the South China Sea; capacity is about 100MWe. Overall, there seems little to separate it from other designs (Integral, semi-passive, above-ground large volume containment), except for one thing.
CNNC has already committed to building a demonstration plant.
Where I suspect this will suffer is finding space for a significant UK share in the supply chain work (perhaps an RR control and instrumentation system?)
This is the design of which China is currently building prototypes – a pair of 250MWe modules of a pebble-bed High Temperature Reactor. It’s design that has much to commend it (although it should be noted they’re intended as demonstrators for a 600MWe commercial design, rather than being the end destination in themselves). The design has received quite a lot of interest worldwide – for example, collaboration agreements between China and Saudi Arabia, presumably as a potential process heat source.
It’s another design that, no matter how promising, will probably not make the final cut on grounds of lack of ability to drive through a GDA process, and lack of UK involvement in a supply chain. I’d also have some questions about the modularity of the design; our past experience with gas-cooled designs hasn’t suggested that it’s a technology suited to factory build.
This is truly the “left field” offer, but offers the chance of a leap ahead in terms of technology.
PRISM is a derivative of the US “Integrated Fast Reactor” design, closely based on the highly successful EBR-II; it’s a metal-fuelled fast reactor, utilizing a “pool” design with a secondary sodium circuit to steam generators. In capacity terms, it’s at the upper end of SMR designs – 300MWe, intended to be deployed in pairs driving a single shared 600WMW turbine.
The design extremely safe – it has intrinsic self-stabilising features demonstrated on EBR-II; there, it was demonstrated that control rods could be withdrawn, and coolant pumps turned off witihout damage. The reactor shuts down when overheated, as a result of a phenomenon called “Doppler broadening”. Readers can feel free to think the operators had the proverbial “balls of steel”, and it’d not be for me to contradict them…
The design is already in one competition for UK build – it’s been proposed as the solution to the “plutonium burner” requirement for disposing of the stocks held at Sellafield; to many accounts, the late David McKay was a great proponent within DECC, and GE-H have offered to fully fund the build on the basis is a “per kg” disposal charge for the plutonium burned.
We can but hope…albeit non-UK sales of something like the Westinghouse unit are more probable.
There are a number of other designs mooted – however, I suspect unlikely to progress, mostly because they offer little differentiation, and seem to be further back in the design process, or because of political/economic considerations.
Terrestrial IMSR / MSRs
Probably the furthest advanced of any of the MSR designs, Canada’s Terrestrial Energy have this design currently in process with their own national regulatory bodies, with the construction of a large-scale electrically heated mock up in mind. In reality, it’s probable that the design is 5-10 years away from being ready for submission to an accreditation process.
Other MSRs are even further behind, and in most cases the organisations proposing them are “mom and pop” shops with insufficient capacity to convince DECC that they have a high probability of progressing through GDA.
This is yet another integral PWR, of 180MWe capacity. It’s had a somewhat troubled gestation – BWXT is the nuclear operating unit of the former Babcock and Wilcox, post their Chapter 11 restructuring. BWXT has cut back work on the design to just $10m/year, until Bechtel partnered with them.
The most notable difference to the Westinghouse and NuScale designs is that it doesn’t use steel containment – it uses a more conventional pre-stressed/reinforced concrete design, of large volume relative to the Reactor vessel size.
It’s dimensionally similar to the Westinghouse and NuScale designs; the core Reactor module would be rail-shippable, but would require in-situ build of the containment.
Unless Bechtel were to be able to pull a rabbit out of the hat re a UK supply chain, it’s hard to see why this would succeed over-and above the front-runners. Bechtel is, however, part of the consortium named to build Wylfa Newydd, so it’s fair to assume it has some exposure to UK GDA process.
There are a considerable number of Russian SMR designs, mostly but not exclusively LWRs. It’s most unlikely, given current relations that any of these will proceed to the later stages of the competition.
Only one Nuclear Battery design seems to have advanced beyond the conceptual design stage – the Toshiba “4S”. That is a radical design, a sealed fast reactor with sodium coolant, of about 10MWe output. As with other designs, that small scale almost certainly weighs against it.
There are obviously acceptance issues with the idea of a “neighbourhood nuclear reactor”, worthy as some designs are – given this, and the fact that many hundreds of such units would have to be constructed to make a significant dent in the UK’s electricity supply gap, it seems unlikely that there’s an attractive international market or a strong domestic case. As things stand, I expect these designs to be more acceptable in future attempts to decarbonise large-scale shipping.
At least one DECC advisor has already commented that this is not going to be a quick process – and in that context, the failure to apply a more aggressive cull at this early stage is probably a missed opportunity. DECC will also have to decide whether it’s going to continue with its current stance that the objective is a near term deployment (for example, GDA entry in 2018/19), and whether this is in parallel with, independent of, or one and the same as the Sellafield Plutonium Burner proposal. Alternatively, if this is to be an exercise of support for “blue-sky” technologies, then we won’t see a UK supply chain for a couple of decades.
Having said that, the leading players look extremely credible, and near “shovel ready”. Were DECC to get things right – a first, if it happened – I think a construction start on a first unit could be as soon as 2020-21, assuming that the Westinghouse unit is chosen, and GDA can be accelerated on the grounds of commonalities with AP1000.