There have been two major developments in the progress of UK nuclear new build in the last two weeks or so – the announcement that KEPCO is now preferred bidder for the Moorside project in Cumbria, and the completion of the GDA (General Design Approval) process by Hitachi’s UK-ABWR design. It therefore seems a good time to set out a quick review of the key features of each design and specifically any adaptations to UK regulatory requirements. This article covers the KEPCO APR1400 and complements the article on the Chinese Hualong 1 design that was published on 14 November.
This guest post by Andy Dawson gives a preliminary overview of the design focussing on safety systems.
The KEPCO APR1400
As most readers of Euan and Roger’s excellent blog will be aware, KEPCO has been named as “preferred bidder” for NuGeneration’s Moorside nuclear project following the withdrawal of Toshiba-Westinghouse involvement. Most will also be aware that KEPCO apparently intend to utilise their proprietary “APR1400” technology, thus imposing a delay on any construction start while the APR1400 is taken through the UK GDA process.
Figure 1 Shin Kori 3&4 under construction
There are currently two operating APR1400 units, at the Shin Kori station in South Korea (Shin Kori units 3&4). Two more units are now on course for completion at the same site. The most well known APR1400 site isn’t however in its country of origin. It’s the 4 unit site at Barakah in the UAE. Barakah is something of an exemplar, with construction apparently completed on time and to budget (of which, more later).
Figure 2 Barakah 1&2 under construction in the UAE.
APR1400 and Moorside
KEPCO have yet to make a formal declaration on their plans for Moorside – the main aspect of which is just how many units they propose to build. The original NuGeneration/Toshiba proposal was for a station consisting of three AP1000 units with a total gross generating capacity of approximately 3360MW. The site is large – the various plans and artists impressions issued by NuGen show three units widely spaced with little or nothing in the way of common facilities, and while the AP1000 is a compact design, it’s obvious that the site has no real space constraint.
Figure 3 Moorside artists Impression.
APR1400 is not quite so compact as AP1000, although not requiring as much space as the EPRs in build at Hinkley Point C. From this I conclude that a three-unit project would be viable in space terms.
However, other constraints point to a two-unit build. First, 4,200MW would represent the largest single grid connection in the UK (the previous record holder being Drax’s 3,900MW (net). It’s technically entirely viable, but as part of the route requires undergrounding and undersea links, I suspect that this may be somewhat unattractive to National Grid.
A range of financial estimates have been bandied about, but (so far as I can see) nothing from KEPCO itself; most figures are of the order of US$15-18 billion. Various figures have been quoted for Barakah, ranging from $20 – 30 billion; it’s hard to pin down an exact figure since it appears at least part of the quoted contract value pertains to a 30-year fuel and operations contract tied to the main build. For the sake of discussion, let’s take the middle of that range at $25 billion, or $6.25 billion/unit.
It’s also quite hard to make direct comparisons between construction costs in the UK and the UAE; unskilled labour is cheap in the Gulf, however higher skilled labour is expensive, and logistics problematic. Land for Barakah will be have been virtually free. The best simplifying assumption is to assume broad parity.
We will review this in more detail below, but I think it is likely that there will be some incremental safety measures required to win GDA approval; the most obvious is a shield wall for the containment. I also expect some additional redundancy (see below). On a speculative basis, let’s assume around a 15% increment to give $7.2 million/unit.
On that basis, it seems unlikely a $15-18 billion budget will extend to a three-unit station. A two unit station appears more likely.
APR1400 – Design
APR1400 stands for “Advanced Pressurised Reactor – 1400MW”. It is (as the name suggests) a Pressurised Water Reactor (PWR), on the larger end of the scale for contemporary PWR design – only the EPR, at 1600-1700MW is notably larger, and most commercial designs such as AP1000, Hualong-1, ATMEA, VVER1200 are in the 1100-1200MW range. Despite the title, the design has an actual electrical capacity of 1450MWe – that, however is a gross figure with probable net output of perhaps 1420MWe
From a design perspective, the APR1400 represents an interesting intermediate. Its design represents a hybrid of previous Korean practice (the OPR1000) an advanced, but not cutting edge US design. It is broadly based on a 1990s Combustion Engineering design – the “System 80+”.
Rights to the System 80+ were eventually acquired by Westinghouse, and was used by them as the basis of the first real attempt to reduce plant complexity relative to designs such as Sizewell B. However, it was by no means as radical a simplification as was the subsequent Toshiba – Westinghouse AP1000; in particular, System 80+ remained dependent on “active” safety measures relying on redundancy of safety-related equipment as opposed to the AP1000’s move to passive systems.
The most obvious manifestation of this simplification is in the primary circuit. Instead of having four separate steam generators and associated piping loops, APR1400 has just two steam generators. Each steam generator is linked to the reactor by a single “hot leg”, and two “cold” (or return) legs. Each cold leg is fitted with a circulation pump (so a total of four). One leg is fitted with a “pressuriser” (a small pressure vessel which in operation contains water and a steam bubble; the pressuriser is fitted with electrical heating such that the primary circuit can be pressurised before the core is generating significant heat). The steam generators are thus the largest in use on any mainstream design at around 2,000MW (thermal) each.
Operating pressures and temperatures are typical for mainstream PWR design at around 180 bar/320C. As with most of the current generation of plant designs, design life is for a minimum of 60 years. Planned construction time is 48 months from first concrete pour (which appears to have been achieved at Barakah). Operational unavailability (including refuelling downtime) is intended to be less than 8%. The plant is designed around an assumption of running in a partially load following mode – 16 hours/day at 100% power, a 2 hour ramp down to 50%, 4 hours at 50% and ramping back to 100% over another 2 hours.
Figure 4 APR1400 schematic.
The core/fuel design differs slightly from its Westinghouse predecessors; instead of Westinghouse’s de-facto standard approach of 17×17 fuel assemblies, 16×16 assemblies are used, and the number of assemblies is somewhat larger that the power level would suggest, at 236. The design features a large number of control rods (both full-strength “black” and part strength “grey” rods) at 76 and 17 respectively. An additional 8 black rods are also fitted for redundancy. This suggests that the design offer the ability for very fine management of the flux distribution in the core, as well as load-following capacity. This is reflected in the high design fuel burn up (60GWd/ton)
The Reactor Pressure Vessel (RPV) is conventional in design, but is intended to use advanced metallurgical techniques (very low copper and nickel content to avoid embrittlement at high neutron exposures).
The Steam Generators (SGs) although large are also conventional u-tube types. They’ve been specifically designed to avoid problems encountered on some generation II plant of corrosion and /or “fretting” (rubbing of the tubes against support assemblies). There’s a 10% “spare” capacity in terms of SG capacity to allow for failure and plugging of tubes over time, or potentially allowing uprating later in life if tube life turns out to be good in operations. The SGs also feature a particularly large water inventory on the steam side; this guards against potential drying out during accidents. The Pressuriser (PZR) is similarly enlarged compared to earlier designs to the same end.
The primary circuit pipework has been designed around a “leak before break” concept in order to reduce potential for a full scale “Loss of Coolant” accident. This also has the benefit of reducing the need for pipe supports, and hence making the primary circuit more flexible and better able to ride out seismic transients.
Unlike AP1000, the coolant pumps are fully accessible through life, and are identical to those on OP1000.
Control Systems are fully digital, using a similar approach to that in aviation where there are four-channel redundant systems with different hard- and software suppliers for each. Reactor protection (i.e. safety and shutdown) systems are fully segregated from operational. Electrical Supply is quadruplicate – “normal” (in house generation and/or grid supply); standby grid supply with a separate transformer; on-site standby (two separated 100% capable diesel generators); and alternative standby (another 100% rated diesel generator set). I’d expect this to be added to for UK GDA requirements to include connection points for on-site mobile generators stored some hundreds of metres from the reactor buildings – Station Black-Out (SBO) tolerance is designed at 8 hours which will be required to be significantly increased (to 72 hours) for GDA.
Safety Systems
The overall objectives for the safety design are that core-damage events should be less frequent that 1 in 105 reactor-years, and significant radiation release from the containment at less than 1 in 106 reactor-years. That radiation release criterion is defined as a person at the site boundary should not receive more than 0.01 Sieverts in 24 hours.
The main safety systems are again conventional – a safety injection system of four independent trains (each capable of 100% removal of decay heat following a full-power accident) using either in-containment water storage or external tanks. There’s a rather interesting passive flow management device included to avoid wastage of emergency cooling water through pipe breaks which works to minimise injection rate when the core is covered and increase it rapidly if water levels drop. In-containment pumps are fitted which will return any surplus water spilled to the in-containment tanks. There is a two train containment spray system to supress any pressurisation through steam release. Post accident cooling is reliant on the steam generators to remove heat to outside the containment – however, this is not passive (relying up electric supply) and venting of produced steam (there’s a dedicated reserve feedwater system).
Hydrogen suppression is “state of the art”, using passive catalytic recombiners within the containment.
Protection against “melt-through” is provided by a system called the “External Reactor Vessel Cooling System” – basically a cooling blanket around the reactor vessel through which water from the Safety Injection System can be diverted to remove heat. It’s rated to remove full post-accident decay heat, from the lower parts of the vessel, thus preventing melting of the vessel bottom head.
This is further back up by a “cavity flooding” system. Cavity flooding is as simple as filling the cavity in which the RPV sits with water, which will then boil off, cooling the RPV and preventing melt-through. If, however, this is insufficient to prevent melt-through, it will directly cool any corium escaping the RPV. This includes specific protection to prevent the distribution of fission products through the containment.
In seismic terms the station is designed to ride out earthquakes producing up to 0.3 g lateral acceleration in any direction.
Containment design is for a steel lined single layer containment building with steel lining and pre-stressed concrete structure. The containment wrapped around with auxiliary buildings with only the hemispherical upper structure exposed.
The plant overall is designed for twin unit deployment with some shared ancillary buildings.
Future Developments
KEPCO has already proposed an evolutionary design based on APR1400 – the “APR+” This is enlarged to 1550MWe, and offers an order of magnitude improvement in core damage and radiation release probabilities, mostly through additional redundancy in safety trains
Summary and potential issues for UK GDA
As currently designed, I suspect this plant will require some enhancement to pass UK GDA, particularly in the light of post-Fukushima requirements. Some areas I’ve already commented on – the requirement to support connection of external mobile generators, for example. Others are obvious; there’s an explicit requirement that the pressure-bearing containment is protected by an external shield wall capable of resisting impact by a large commercial aircraft, so such will have to be added.
Most, however are more subtle; I don’t have sufficient detail to judge whether there is in fact sufficient redundancy in the key safety trains (although at face value it appears so). I’d also expect the current 8-hour SBO capability to be judged inadequate and require reinforcement, plus increased capability to top up “in containment” water storage from outside the containment.
What worries me most is what appears to be a lack of passivity in the safety systems (all appear to be reliant of electrical supplies), and the dependence on the SGs for post-accident heat removal. It may be the case that KEPCO can demonstrate sufficient reliability through probabilistic analysis, but I’d take some reassurance from at least one independent route for heat removal from inside the containment to atmosphere, and preferably one that wasn’t reliant on electrical power i.e. natural circulation based. On the upside, there appears to be plenty of space within the containment for such a system, and given that it should be passive, it should be cheap to build.
The APR+ would go some way to address these changes (albeit on the double containment and passivity issue); it does, however upgrade the shutdown cooling system to safety equipment categorisation, so at least introduces an alternate heat removal path from the containment to ambient air.
Other posts:
The UK’s Small Modular Reactor Competition

















Andy, one thing that strikes me looking at Figures 1 and 2 is how utterly identical ShinKori 3&4 are to Barakah 1&2. Lets hope the UK authorities recognise the advantage of cloning a design that is tried and tested and do not impose too many changes that may achieve little other than add costs.
Euan, have you seen this?
https://thoughtscapism.com/2017/11/17/the-right-price-for-saving-the-planet-depends-on-the-energy-form/
There’s a difference between changes requiring change to core systems and (what should be) relatively minor incremental capabilities.
From what I can see, by far the largest and most expensive change is likely to be the containment shield wall. Given the history of terrorism in the West, I don’t regard that as an unreasonable addition.
According to this report from 2002, https://www.nei.org/CorporateSite/media/MemberFiles/Backgrounders/Reports-Studies/EPRI_Nuclear_Plant_Structural_Study_2002.pdf?ext=.pdf , most existing nuclear power plants in the US are not likely to release substantial amounts of radioactive material in the event of an aircraft (Boeing 747) impact. So, maybe a containment shield wall isn’t necessary.
“most” and “not likely” aren’t really good enough….
Surround the plant with windmills to defend against commercial aircraft.
Would it not be possible to program the auto pilot of an aircraft to intervene if gets within a certain distance of possible terrorist targets, like NPPs?
Software is cheaper than containment buildings. Perhaps the builders of NPPs should bill the manufacturers of aircraft for the added costs of defending their reactors against such attacks?
Andy, the design of the aircraft crash protection shell is bespoke to each site. It adds significant cost to the civil works and construction time. Fortunately, not all the faces of the nuclear island need to be constructed to withstand the impact of a commercial size aircraft. For example, one side might be shielded by the conventional island structure, possibly others by adjacent reactors and there might be adjacent cliff faces (e.g. Flamanville) that would prevent a large jet gliding in and impacting the nuclear island from certain angles of attack.
Since 9/11, the doors to the flight deck of all commercial jets have been made virtually impenetrable and passengers are no longer permitted to enter except when on the tarmac. Hence it’s arguable that new nukes do not need an aircraft crash protection shell designed to withstand the impact of a commercial jet. Adequate protection against lighter missiles is however justified in my view.
Andy Dawson,
Thank you for this and your previous excellent posts. It’s great to have such expertise offering posts to Energy Matters, and a tribute to Euan Mearns that he has managed to attract such interesting and informative posts from people with expertise in such specialist areas as this.
You say:
I’d just point out that nuclear power has been the safest way to generate electricity for over 60 years – i.e. since the first reactor began operation in 1954 https://en.wikipedia.org/wiki/Obninsk_Nuclear_Power_Plant.
Hear, hear!
Peter,
The safest and lowest in cost/kWh, and with the highest capacity factor, 0.85 – 0.90, and a long life, up to 60 years.
At present, China, Russia and India are the main actors in the world’s nuclear sector. All three have major building programs. China will increase its nuclear plants from 45,000 MW in 2017 to 95,000 MW in 2026. Japan, the EU and the US have nearly abandoned the nuclear sector. Nuclear electricity generation increased in 2015 and 2016, after about a decades of decreases. See URLs.
http://www.world-nuclear.org/information-library/country-profiles.aspx
https://www.bloomberg.com/news/articles/2017-09-20/china-s-nuclear-power-appetite-to-offset-western-decline
The transformation of the world’s economies and build-outs of systems for 90% RE will never happen, unless massive nuclear plant capacity, at least 1,500,000 MW, is built by 2050, and that capacity would have to provide much of the world’s electrical energy to replace fossil fuels with synthetic fuels and other electrical consumption. This URL shows, the world is on track to have 500,000 MW of nuclear by 2040.
http://www.theenergycollective.com/todayinenergy/2416241/eia-forecasts-growth-world-nuclear-electricity-capacity-led-non-oecd-countries
Modern RE (wind, solar, hydro, bio, etc.) would provide part of the world’s electricity. At present modern renewables provide about 10%. See URL.
http://www.windtaskforce.org/profiles/blogs/the-world-making-almost-no-progress-towards-renewable-energy
Your point is irrelevant, as this is about the build-in systems in the reactor – not about nuclear power per se.
Planes are also safe, but made safer with each passing generation and so should nuclear also be. The Fukushima accident happened because there was no emergency power. If the plant had been able to passively cool itself after the shutdown, the accident would not have happened.
I am no expert in this area. However, from what I have read, the Fukushima reactors were cracked by the earthquake – and that is how they lost their fuel. The water came later.
“Fukushima clean-up chief still hunting for 600 tonnes of melted radioactive fuel”
http://www.abc.net.au/news/2016-05-24/fukushima-operator-reveals-600-tonnes-melted-during-the-disaster/7396362
I fail to see how it was caused by the power outage. Perhaps you can enlighten me? Are you saying that the hydrogen explosions – which happened elsewhere in the reactor – were responsible?
Mark Willacy has been writing anti-nuclear articles for years, and Greg Jazcko worked for anti-nuclear senators Harry Reid and Ed Markey, before they got him appointed to head the Nuclear Regulatory Commission, where he did his best to regulate the industry out of existence. He resigned after the other NRC commissioners, and his staff, complained about his bullying behaviour, and use of underhand tactics in trying to block the Yucca Mountain spent fuel repository, which Nevada senator Harry Reid had made a career of opposing.
The reactor pressure vessels at Fukushima Dai-Ichi were ‘scrammed’ automatically as soon as the earthquake struck –
the control rods went up between the fuel assemblies and absorbed so many neutrons that the chain reaction stopped immediately. The mains power failed at the same time , but the diesel generators kicked in and were pumping coolant through the plant. An hour later the second tsunami hit, flooding the basement diesels. Attempts to rig up alternative power failed, so the pressure and temperature in the three reactors started to rise. They should have been vented as soon as the pressure inside them reached the maximum allowable level, but the government in Tokyo would not allow this until residents within ten kilometres had been evacuated. If the pressure vessels had been vented, there would still have been radiation release and melt-down, but not as serious as occurred. At temperatures over about 700 C, the zirconium cladding of the fuel rods reacts with water, producing hydrogen and heat. The heat added to that from fission product decay, boiling off more volatile cesium, and the hydrogen ignited and blew the light roofing structures off the top of the reactors. If they’d been vented earlier, there would have been no water left to react with the cladding, and less steam to carry volatiles away.
Power plants further up and down the coast – Onagawa and Fukushima Dai-Ni – also were hit by the earthquake, but since their emergency cooling systems were not destroyed by the tsunami, they were able to cool down safely. Also it seems likely that the melted fuel did not go straight through the bottoms of the pressure vessels, but only through the openings where the control rods come up. PWRs like the Korean ones don’t have those openings in the bottom head – the control rods come down from the top.
Thanks jfon.
I had also put in a comment along your lines, although much simpler, but it has been “ethered” or something.
Anyway, a useful source of reliable information here, as opposed to the Aus ABC, The Guardian or Wiki, is:
http://www.world-nuclear.org/information-library/safety-and-security/safety-of-plants/fukushima-accident.aspx
Thank you. I guess I need a 3D model to understand properly what you have written. 🙂
Jfon,
The layout of the plant was flawed.
The plant dealt with the earth quake but not the tsunami
alfredmelbourne
If your “sources” of information are as you quote – the Aus ABC, The Guardian, Wiki – then your misinformation is to be expected.
The magnitude 9 quake did not “crack the Fukishima reactors” – all of them were unscathed by the quake, as severe as it was. Four other nuclear plants were operating in the region at the time of the quake and all shut down routinely and safely.
The resultant tsunami did the damage at Fukishima, overtopping the protective wall, flooding the plant and disabling 12 of 13 back-up generators on site and also the heat exchangers for dumping reactor waste heat and decay heat to the sea. One could argue with some force that the initial wall design, at least in height, was underestimated – I would agree, despite the hindsight aspect of that.
http://www.world-nuclear.org/information-library/safety-and-security/safety-of-plants/fukushima-accident.aspx
for a reliable and quite detailed account of the incident.
Alfred,
The plant was laid out to look pretty from the land side.
That meant the auxiliary transformers and emergency diesel generators were on the ocean side.
When the tsunami hit, that critical equipment was the first to go.
All else are consequences.
@AM
What happened in Fukushima: The earthquake made the plant shut down as it should, but active cooling was required to remove the heat afterwards. The generators started as they should.
The tsunami however took out the generators and since the grid was down, the plant lost power to cool the reactors in shutdown.
The plant did have seawalls, but they where not high enough.
If the generators had been at a higher level or there had been an alternate energy source, such as a generator, the accident had been recoverable at this point.
It is first when the hydrogen explosions that the plant is lost and structural damage occurs.
The plant did not suffer structural damage from the earthquake as far as I know, thou it was outside the design limit of 11 (the quake measured 14 as far I remember).
The accident was entirely preventable and the plant owner and the authorities doing the oversight has been judged to be guilty of negligence, and too close.
“What happened in Fukushima:”
No one was killed or ever likely to die as a result of radioactivity released from the nuclear power plant. On the other hand thousand dies as a result of the inappropriate response to the accident. The response is the result of regulations which them selves were a response to 50 year of anti nuclear rhetoric and scare mongering. Meanwhile, EPA estimate in USA alone, 15,000 to 25,000 people per year die as a result of pollution from coal fired power stations. Where’s the rational justification for the impediments that are imposed on nuclear power?
@ Peter
“You say:
What worries me most is what appears to be a lack of passivity in the safety systems
I’d just point out that nuclear power has been the safest way to generate electricity for over 60 years – i.e. since the first reactor began operation in 1954 ”
But that does not mean I will build a process plant to a 1954 standard based on one or even one hundred examples. Loss of power is a common scenario across CPI and power industries and needs to be mitigated against.
@ Andy
“however, this is not passive (relying up electric supply) and venting of produced steam (there’s a dedicated reserve feedwater system).”
Stick a diesel genny on it to supply the emergency power.
“Stick a diesel genny on it to supply the emergency power.”
which still leaves you dependent on a genny starting up and continuing to run – which isn’t passive. The whole point of passive is that it doesn’t need a power supply (i.e. it uses natural forces), and ideally (b) doesn’t rely on operator or control system intervention to operate.
Slightly different philosophy. But even a passive system will likely need a valve to open?
Which will entail a reliability to act on demand or avoid failure on zero demand which inherently is not passive, though almost (fail open valve on loss of power), and certainly simpler than a start of a diesel engine.
I am therefore guessing that a safety injection system is sitting there with a pressure accumulator behind it to force water in when circuit pressure drops or a gravity fed system? Or are we strictly talking about how the reactor acts i.e. loss of water = reduction of power? Which i guess means passive safety is different depending on the condition of the reactor i.e. running/shutting down versus shut down/residual.
“Slightly different philosophy. But even a passive system will likely need a valve to open?”
Indeed, but the idea is that they should be “fail safe” – that is requiring no positive actuation.
For example, all major valves on AP1000 move into their safety/shutdown position if a control signal was lost or an SBO. They require no external power (or gas pressure, or…) to actuate – they’d be spring or gas pressure actuated. The only major exception is the so-called “Squib valves” – explosive-actuated single use valves intended to isolate the RPV in the event of a major pipe break. Those are a bit “sh*t or bust”, and once used they write the plant off.
So strangely enough, they need operator intervention to trigger…
“I am therefore guessing that a safety injection system is sitting there with a pressure accumulator behind it to force water in when circuit pressure drops or a gravity fed system? ”
Absolutely correct. Or it can use other natural behaviours.
There’s a lovely example in the APR1400. The designers have come up with a device that controls flow levels in the safety injection system based on water levels in the RPV; I don’t claim to understand it fully, but it uses a vortex in the flow to create backpressure and restrict flow.
“Or are we strictly talking about how the reactor acts i.e. loss of water = reduction of power? ”
that’s inherent anyhow – cause voiding (i.e billing) in the core of a PWR, or increase the rate in a BWR, and power drops, probably to the point of shutdown. In some fast designs, you can use a phenomenon called Doppler broadening to reduce reactivity as temperatures increase.
And yes, that’s be part of passive safety – but would rarely be the totality.
EBR-II near Arco, Idaho in USA was entirely passively safe, without any dependence on power or valves. This was proven to an invited international audience in 1986. The operators turned off the automatic emergency response systems (but of course maintained access to manual ones). They turned off coolant circulation — the cause of damage at Three Mile Island and Fukushima. The fuel cladding temperature spiked from 1000 F to 1430 F within ten seconds. Sodium coolant boils at 1620 F. The operators sat back and watched. Fuel cladding temperature fell below operating temperature within seven minutes (not hours, days, weeks…). This worked because of immutable laws of physics, thermodynamics, properties of materials, …. The reactor was not damaged. No radioactive materials were released. The operators were not harmed. They restarted the reactor with current circulation but disconnected the steam generator. The same scenario ensued with very slight delay. Three months later the operators at Chernobyl repeated the latter experiment with a very different reactor and very different outcome.
@ Andy
Thanks. Seems like a more rigorous application of the inherent safety philosophy creeping into CPI. I have not reviewed the entire system but ti seems like belt, braces and then some.
Andy do you think problems with AP1000 are due to an overly complex design or poor construction performance. Also was the Korean APR1400 completion on time and on budget due to better construction practices from the Koreans or an improved design or both.
Another thing which concerns me is using another foreign design will mean fewer opportunities for UK engineering. It has been my observation that very few jobs have come from Hinkley Point even though the UK is supposed to get 60% of the supply chain makes me suspicious all the work has gone to Paris.
I was hoping for some work to come out of Moorside however the curse of the GDA means another 5 year delay. Why can’t the NDA go a bit faster surely 5 years is unreasonable to check a design. You did mention previously they might go a bit faster.
I’ve only the same information as everyone else, Rob; but from what I see, it wasn’t complexity that was the issue at Votgle or Sumner. It genuinely is an inherently more simple design.
It seems simply to have been that the design was adequately worked up (with contribution from a subcontractors who’d insufficient recent nuclear experience).
The biggest single issue seems to have been with the circulation pumps which were delayed by something like two years. That was (I think) a mixture of the fact that the characteristics of the pumps on ramp-down were different from design expectation – they stopped much faster than anticipated, restricting flow in the particularly challenging immediate post-shutdown period. Also, as they’re integrated into the SGs, they’re not accessible during the station’s working life; proving adequate reliability was difficult.
There’s one aspect where I think the jury’s out. AP1000 was meant to be built on a “modularised” basis. That’s not only about supply of major components factory built. The plan was to have an on-site “factory” where major assemblies could be put together and then craned into place. It’s routine practice in Japan (I don’t know about Korea). The Sumner contractors seem to have struggled with the approach.
That should of course be INadequately worked up…
I had read there were serious problems with weld quality and concrete emplacement, and some of the work didn’t actually comply with the blueprints.
The length of time to approve an existing design which has already been approved and working elsewhere is incomprehensible to most people working in other areas of building. We do have trouble with some building control departments who find any ridiculous reason to object, while sometimes missing obvious faults, but five years to understand and check an existing design seems like jobs for the boys. When it is obvious that we will have to abandon the CCA or have windmills, stationary electric cars and blackouts, why can’t some government minister tell them what the priorities are then decide on the best design and roll out sime at competitive cost. Just do it like the rest of the flippin world?
Steve,
One example of the complexity can come from the statement that containment has a leak before break design concept. Sounds good in simple terms but for verification needs a deep understanding and investigation of the design codes. One expert group may have an entirely different understanding than another so what one group thinks is safe, another may reject. Take this on board and a review of all critical piping and pressure vessels is required from the outset.
If this sounds a bit alarming there is at the moment a divergence in technical requirements between pressure vessel codes in uk and cen. The EN codes allow far more relaxed standards and some specialist technical analysis indicates the EN standards could give rise to failure.
Leak before failure and critical defect size has also caught out the cryogenics industry. Calculations based on the more conservative assumptions have proven wrong, albeit with extenuating circumstances.
I for one hope the authorities take an appropriate amount of time to approve these designs.
A problem in US is the NRC policy of license-then-test. This greatly delays new designs because all the analysis has to be done hypothetically. A test-then-license protocol has been proposed for a new design, whereby an exemplar would be built within or adjacent to a national laboratory, for example INEEL, within the (perhaps extended) security perimeter. The builder would have access to national laboratory expertise. Rather than hypothetical paper studies, NRC would inspect each phase of the construction before the next could commence. When the reactor is finally ready for service, the security perimeter of the national laboratory would be adjusted and an utility would take possession and operate the reactor. If it’s intended to be a sufficiently-small modular reactor it could be moved to a new site.
Great information all.
Wonder if Mr Dawson would review the recently “approved” Wylfa Reactors for North Wales & Bristol?
Happily – but this is Euan’s blog!
Gaz, Andy has already written and submitted the post which I will publish some time soon, over the holiday season.
Is there an overview of the UKs nuclear road map somewhere?
This is the closest I’ve found: –
https://wholefoodcatalog.info/food_img/x-large/1064.jpg
I didn’t want to presume 🙂
Maybe try again – comments seem to be disappeared.
Thanks jfon for the comment. I had put in one on similar lines, but far simpler, but it seems to have been ethered.
http://www.world-nuclear.org/information-library/safety-and-security/safety-of-plants/fukushima-accident.aspx
This supplies reliable information as compared to the Aus ABC, The Guardian, Wiki.
WordPress comment moderation has a mind of its own. A large number of commenter’s comments unintentionally go to moderation, including my own 🙁
No issues Euan. I really value your website here – polite, well informed and unafraid.
Ian’s
The best on the planet, by far.
@ Andy
Just to clarify
“The overall objectives for the safety design are that core-damage events should be less frequent that 1 in 10-5 reactor-years, and significant radiation release from the containment at less than 1 in 10-6 reactor-years.”
These are non fatality events (the latter sounds analogous to an environmental release)?
It’s actually related to a dose rate at plant boundary – “That radiation release criterion is defined as a person at the site boundary should not receive more than 0.01 Sieverts in 24 hours”
So, there should be a 1 in 10^7 reactor years probability that sufficient Bequerels are released such that that hypothetical person at the plant boundary gets that 0.01Sv/day (100mSv).
Timing is an issue, though – part of the Fukushima learning was that exposure rates peak at very high levels early in an incident and decline rapidly (as the very short lived volatiles like Iodine 131 decay).
For comparison, the maximum allowable short term exposure for emergency workers (set by the ICRP/IAEA) is 500mSv and 50mSv/yr. The highest six doses received in the first nine months at Fukushima ranged between 310MSv to 700mSv.
Set 100mSv/day against that 50mSv/yr, and this doesn’t look like a tight limit.
“Set 100mSv/day against that 50mSv/yr, and this doesn’t look like a tight limit.2
Indeed. I asked as to get a feeling of where it would sit in IEC 61511 and safety integrity level verification calculations.
I think 0.01 Sv/day is 10 mSv/day, not 100 mSv/day. Or do I not remember schoolboy maths from half a century ago correctly?
Just a thought on passive safety – one issue is providing cooling water independent of electrical supply. Small reactors can get round this by reverting to air for cooling, or drawing off some water at high tide into the basement silo, but larger reactors need pumped water.
Moorside has hills nearby with water storage opportunities. It should be possible to design in a water replenishment system that will maintain cooling water for ever. (Small reservoir filled with rain water in the hills, high pressure pipe, stop-cock style valve – as simple as a toilet cistern – tested (or “flushed” once per week).
Of course, such a system is in itself susceptible to attack or damage, but would be a cheap way of providing a fully independent, passive safety system in case the active systems go off line.
That’s triggered a bit of a thought…
Decay heat levels post-accident for a LWR drop rapidly after a shutdown from power; There’s a table in here:
https://mitnse.wordpress.com/2011/03/16/what-is-decay-heat/
that tracks the decline in decay heat in the Fukushima reactors – instantaneously on shutdown they’re at about 6.5% of full power, 3% after 1 minute, 1.6% after 15 minutes, 1% or so after an hour, a little over 0.5% after a day, after 5 days 0.3% and so on.
Turning that around for a 4,000MW(th) APR1400, that suggests we’ve got to remove about 60MW average over the first hour, 20MW after a day and so on.
Assuming a 20C water supply being boiled off, after that first day we need a supply of about 25 tons/hour. Which, when you consider that a single garden hose supplies about three tons/hour isn’t an unduly problematic amount to supply.
At typical UK mains pressure (4-6 bar), it’d need a pipe of about 3cm.
However…
the challenge in passive systems is less about how we get heat into the ultimate heat sink – it’s more about how we get the heat from the core to the outside world without carrying any potential contamination.
AP1000 used the most elegant mechanism – a water tank on the top of the containment shield building, and a steel containment where the walls are sufficiently conductive to pass the heat without excessive internal temperatures. Hualong and VVER 1200 have an additional passive circuit running from the SGs to external heat exchangers (not quite as good, as they rely on the primary circuit being intact). ABWR has heat exchangers that carry heat by passive circulation from the suppression pool to water tanks in the reactor hall which can be vented to atmosphere.
It’s that sort of mechanism that APR1400 lacks.
Here’s an architectural solution. A tall steel cone in a steel frame with a big water tank inside the cone. Self extinguishing protection from a plane strike and protected water storage. Extra water tanks on towers away from the reactor just in case a terrorist attack happens at the same time as an scale 7 earthquake in the UK.
Just a thought. If we need to be so ultra-cautious, why not close down Sizewell B and all the other existing nukes in the UK and France which don’t comply with new codes. You can’t be too careful.
“Extra water tanks on towers away from the reactor just in case a terrorist attack happens at the same time as an scale 7 earthquake in the UK.”
Which is why all the new designs feature large water tanks actually within the containment….
“If we need to be so ultra-cautious, why not close down Sizewell B and all the other existing nukes in the UK and France which don’t comply with new codes. ”
dependent on the shortfall, you’ll probably update them – for example, the ONR did a full post-Fukushima review of all the UK stations which resulted in a number of changes (enhanced waterproofing of external backup generators, mostly). And in France, Tricastin has been closed for the reinforcement of the walls of a nearby canal which could have caused flooding.
Any new nuclear plant build today faces two serious problems:
1. an economic problem: The plant will compete with ever falling costs of other renewables over a 50 year period. No current design is either cheap enough or flexible enough to survive in the business environment of the near future (main problem being solar and/or wind at over 100% demand during longer periods of time, complete loss of the daytime peak at least in the summer months).
2. the security problem: The plant has to withstand an enhanced 9/11 type of attack. It is not enough, that the hull an withstand huge planes. The plant must resist multiple hits by planes, for example taking out all sources of electricity. It must also be able to withstand a military assault of about an armed platoon. And it must be able to withstand sophisticated attacks, possibly using military grade equipment, which could be aquired from hostile governments or warzones (multiple huge truck bombs, chemical weapons, anti tank missiles, computer virus, …).
In many regions, it obviously also must resist real military attacks (ballistic missiles or bombardment).
we need a really really cheap 150% passive design. Good luck!
“(The plant must resist multiple hits by planes, for example taking out all sources of electricity….multiple huge truck bombs, chemical weapons, anti tank missiles, computer virus, …).”
I did some of the original blast modelling for the Heysham and Torness safety case.
We had to rework it when it became apparent that the MoD had contingency plans for berthing ammo ships running between Plymouth and Rossyth at Heysham Port during extreme weather.
The rework was on the basis of 1,000 tonnes of military grade explosives going off about 1/4 of a mile away from the plant boundary. There wouldn’t have been a lot left of the refuelling hall, but the reactors and all other significant bits of the plant were fine.
And when some of the more lunatic Swiss Greens fired anti tank missiles at the SuperPhenix containment, no-one even noticed…
What might interest a few here is that my employers (NNC, as was) were given a contract by the HSE to use the same blast modelling code to evaluate various other accidents – including what’d be now a medium sized LNG tanker suffering a breach and subsequent explosion of the gas cloud in Liverpool Bay.
There wasn’t much left of Liverpool. As to whether that’s a bad thing – well, I’m Mancunian.
So, Sod – would you argue that (for example) every chemical plant in the country with significant hard inventory were subject to that same condition? There’s the old ICI works at Runcorn – several thousand tons of Chlorine in single walled tanks. Or every waterworks and reservoir be comprehensively protected against someone dumping toxins or a biological agent into them?
It’s obvious what the tactic is – continually demand ever more unreasonable safety standards, while ignoring the impicaltions in terms of fatalities if we hit energy shortages during a cold snap.
Thanks for this detailed and interesting reply. I actually agree with most of what you wrote. Modern domes are not a relevant target for terrorist attacks, as they wont be able to penetrate them.
So the target will be the vital equipment outside the dome and it will cost a fortune to protect generators, power lines and spent fuel at the same level of protection offered by the dome.
i also agree with your assessment of chemical plants, which might actually be more dangerous targets for terrorists (and should be better protected). But this does not help nuclear power plant security.
with nuclear, we are talking about a 10 year building/planning period and any such period starting now or in the near future will most probably be interrupted by the first serious accident in a chinese or indian nuclear plant and by serious terrorist attacks. (possibly we will also see more nuclear facilities being targeted by military attacks in warzones and i would seriously worry about bad steel scandals as well and/or the total failure of one of the new reactor types).
Any new plant will need to be able to answer questions of security in each of these situations or risk delays or abandonment.
meanwhile in Australia the battery is demonstrating its ability to react to grid problems and the competition is getting worse and worse.
https://www.businessinsider.com.au/elon-musks-tesla-battery-south-australia-responded-in-record-time-2017-12
The famous South Australian battery, ‘ the world’s largest battery ‘, is capable of supplying the whole state’s average demand for about two and a half minutes, so peak demand for probably one minute ( though of course I’m talking energy quantity, not power rate, here.) So on a calm, cloudy day, or succession of days, how many batteries would you need to keep the lights, water pumps, operating theatres, elevators, freezers, etc ad infinitum, working ?
Let’s hope the Health and Safety ministry read the comment on LNG tankers and ban them, especillay now the Russians are sending their first delivery to help us out over Christmas. I live nearTilbury most of the time and it’s coming in there. On the other hand, I don’t want to freeze, so perhaps just be careful and ban vodka and smoking.
Happy Christmas.
present on way Euan
Steve, the hse are well aware of the risks and the lng terminals are top tier comah sites that are highly regulated. The only gas ships going into tilbury are small co2 ships as far as i know. The fuel terminal further up river has other plans, but they wouldn’t involve the big lng ships.
Buncefield was a bit if a wake up call to the hse because the blast was substantially more energetic than had been assumed possible. Hse have published a research report looking at incidents at large hydrocarbon facilities; you can easily find on-line by searching hse rr1113.
I’m more scared about the green evangalists touting hydrogen fuelled vevicles. 600 bar storage tanks are the current state of play. 1200 bar is in development. The pressure alone is bad enough, but add in a gas with the widest flammability range, a propensity to auto ignite and stick all this on the public roads and the only word i can think of to sum it up is madness.
@ 1. an economic problem: The plant will compete with ever falling costs of other renewables over a 50 year period. No current design is either cheap enough or flexible enough to survive in the business environment of the near future (main problem being solar and/or wind at over 100% demand during longer periods of time, complete loss of the daytime peak at least in the summer months).
Why would you favour a generating source that can only provide demand for a few specific hours in a day over a specific time of the year? Economically you certainly would not.
i does not matter whether i favour anything or not. Home owners will switch to solar PV when it is cheaper than electricity from the grid. They will switch to solar and battery, when batteries get cheaper. They will switch to diesel/gas backup when the grid decides to punish them for their partial grid defection.
i simply think, that the economic reality is hostile towards nuclear plants. my opinion on nuclear itself is basically irrelevant to this line of argument. Making a 50 years decision is simply difficult in an environment that is changing fast.
At least in California, solar and wind are artificially inexpensive because of subsidies. US Federal subsidies for wind in 2013 were 3.53 c/kWh, and 23.12 c/kWh for solar. California added 30% more for solar. Hiding an expense in your tax bill doesn’t actually make anything cheaper. In Illinois, Excelon is being indirectly subsidized out of business by being required to buy wind-generated electricity at more than their cost of producing electricity themselves.
Andy Dawson
Many thanks for a most interesting and detailed read.
I wonder if I may raise a ticklish question which has been floating around in the Scottish media in recent months, that of C14 getting into marine mammals around the Solway and South Ayrshire coasts, with presumed origin from Cumbria (?). Studies produced do show very low concentrations getting into the environment, though shellfish consumers (not human, ie species which eat the shell as well, and so the carbonates which have incorporated the isotope) will tend to concentrate the C14. The fact that dosage is still very low, even in this route, augurs reasonably well. Nevertheless, the hysterical sections of the media latch on to this kind of thing very quickly.
My question is about how the C14 from nuclear plants may be getting into the environment, assuming that it does not result from unplanned releases or system accidents (this may indeed be the route), or historic incidents or indeed historic nuclear testing. Since C14 has to be produced by irradiating Nitrogen, one would be inclined to think that radioactive sources placed in air or gases containing Nitrogen would be the immediate route, assuming the resulting gas is not processed before release. Coolants containing ionised Nitrogen or irradiated (though protected) containment materials, themselves placed in air, might be other routes. Open ponds containing radioactive waste might be another, worrying, route, through any number of means, including seabirds or even insects. Such features ought to be relatively easily dealt with.
If indeed any of this is the case (?) then an obvious (though cumbersome) safety precaution presents itself: ensuring sources are never placed in air/gases/liquids containing Nitrogen.
I realise this is a very sensitive matter, so you may prefer not to answer. I fully understand if you choose not to. However, an authoritative view may help to allay anxieties, particularly if the problem is well understood and with measures well in-hand. And this all assumes there really is a current problem, as opposed to a legacy question.
Presumably, that’s a reference to this paper
http://www.sciencedirect.com/science/article/pii/S0025326X17305842
It at face value makes a case that there is 14C contamination detectable in marine mammals – although I’d take issue with the claim that bio-accumulation is occurring, as 14C levels in the prey species tested are much the same as those in the mammals.
A note on scale, though. Brazil nuts are about 50% higher in specific radioactivity (Bq/kg) than the levels quoted in the various mammals – and the emmissions quoted amount to 42 grammes/year of 14C!
I’d need to do some digging/research on where the 14C is produced, but I’d be pretty confident it’s not because of the use of N2 as a coolant – simply because N2 isn’t used as a coolant. It’s a strong neutron absorber, which renders it utterly unsuitable – to the degree that N2 injection is actually the tertiary shutdown mechanism for AGRs.
I don’t believe that it is possible for 14C to bio-accumulate. We know that 12C is slightly preferred to 13C in photosynthesis but animal metabolism treats all carbon the same; never read anything to the contrary.
Wikipedia gives a pretty good explanation.
https://en.wikipedia.org/wiki/Carbon-14
I imagine that measures may already be taken to reduce the N content of coolant.
“This is expected as 14C is transferred through the food web without any bioaccumulation or concentration effect.” Tierney et al, 2017.
Efforts to reduce the N content of coolants would indeed seem desirable. While the concentrations of C14 are pretty low, and as Tierney et al. say pose no radiological risk to humans, doing something about the released coolant would help to make nuclear power more sellable to more emotionally susceptible sections of electorates.
Alistair,
it’s noting to do with coolant.
See my posting of a few minutes ago (and when we think about it, if it’s a coolant issue, it wouldn’t be released at Sellafield, it’d be at the reactor sites).
So wylfa and hunterston bracketing sellafield would have increased the apparent exposure area, which according to the research they didn’t. One less irrational argument.
Euan,,
I really don’t think that N2 in coolant is a likely source – in AGRs it’s very definitely be a no-no, due to corrosion and neutron absorption effects – they run with a very pure CO2 coolant (the only significant non-CO2 component is a little methane, which reduces corrosion of the graphite cores).
Similarly, LWRs run with completely de-aerated water as coolant/moderator – I see some claims of low levels of 14C production at LWRs, but I’m struggling to see a path other than what’s discussed below…
Plus, the emissions are happening at Sellafield, not at reactor sites. Sellafield doesn’t have any running reactors.
Having had to do a bit of digging, there’s only way that 14C gets made -absorption of a neutron by 14N and subsequent decay. The tricky part is that the neutrons need to be at relatively low “thermal” energies. It happens in the upper atmosphere by absorption of “secondary” neutrons resulting from various interactions with high energy cosmic rays (far too high energies to happen in earthly reactors). but we do get lots of thermal neutrons anyhow in any non-fast reactors.
So, where does the nitrogen come from? I didn’t know this, but there’s always a small amount of nitrogen (in the form of nitrides) in oxide fuel – it seems that the assumption is that as it’s at a low enough concentration to have no significant nucleonic impact), it’s not worth using especially pure O2 in the production process.
So, curable as an issue (and frankly, the cost implications would be negligible). But in all honesty, the dosages involved are irrelevant.
The company i work for supplies the G to the AGR’s. I’d have to have a root around but spec for N2 is probably of the order 10 or 20 ppmv. Much less CH4 in the raw product, so unless this is added on site for graphite protection it is not down to us.
My former boss was cegb and cut his teeth on the agr’s. He goes a lot further than saying it’s a crying shame they got the design and build right then stopped the programme.
The CH4-is indeed added on site (again in the order of tens of ppm).
Your boss would have had a point if we had indeed got the design right by Heysham and Torness. But we hadn’t .
They’ve all been downrated (corrosion around the top end of the boilers) They can’t be refuelled on decent power levels – and there’s no way they’d pass modern safety requirements, or could credibly be made to do so.
Apart from that, they’re fine…
Nuclear science makes the energy future so easy and US have the great scientist in this field. Still US try to rely on fossil. Going even more backwards adding solar wind and burning wood! This due to US political corruption and ignorance. But it will get a lot better the next 10 years. Nuclear will dominate. Not hard to predict. Just neccessary.
Would have been best for the UK form a company like EdF and tender out a specification for a NPP with a view to building 20 of them.
Seem like we’ve chosen the worst way to do it.
As usual Benjamin.
Can’t be worse than US NRC.
If South Korea and other countries can design and build nuclear reactors then why can’t we?
I’m not ignoring the very significant costs but we do have some very good engineers.
Is it because of the general market, government policy, risk aversion or what?
I’d have far more confidence in a nuclear economy than a renewable one an the risk of military use of the radioactive material is presumably pretty low other than being used in a dirty bomb.
“The site is large – the various plans and artists impressions issued by NuGen show three units widely spaced with little or nothing in the way of common facilities, and while the AP1000 is a compact design, it’s obvious that the site has no real space constraint.”
Large nukes need spacious construction laydown areas around the plant. EDFs EPR-NM design will be using larger buildings and modular construction techniques as the stick built EPR takes too long to build. The way to decrease construction time, hence costs, is by using modules assembled close to the work faces that can be craned into place. Note that is how our new carriers were built from modules using a goliath crane spanning the dock to lift them into place. Interestingly goliath cranes were used in the 1950s to build the Magnox nuke plants in a hurry. Would be far better than the forest of cranes trying to work around each other seen on PWR construction sites to date.