Hitachi ABWR

From ScienceForSustainability
Revision as of 15:17, 16 September 2020 by Sisussman (talk | contribs)
Jump to navigation Jump to search


The Hitachi Advanced Boiling Water Reactor is a boiling water nuclear reactor marketed by a consortium of Hitachi and General Electric (GE) based on a design originally by GE developed in collaboration with Hitachi and Toshiba. Toshiba-GE and GE itself offer other variations of the design.

Advanced Boiling Water Reactor Wikipedia

The following discussion of the ABWR, by Andy Dawson, was published on Euan Mearns' "Energy Matters" blog on 5th Jan 2018: The Hitachi Advanced Boiling Water Reactor

Horizon Nuclear Power, a subsidiary of Japan’s Hitachi, has proposed to build advanced boiling water reactors (ABWR) at Wylfa on the Isle of Anglesey, N Wales and potentially at Oldbury too. The design, similar to the reactors destroyed at Fukushima, has just passed the UK general design acceptance process, clearing the way hopefully for construction to begin in the near future. In this guest post, Andy Dawson reviews the ABWR design concept, once again focussing on safety systems.

On the 14th of December the UK Office for Nuclear Regulation (the ONR) announced that the UK-Advanced Boiling Water Reactor (ABWR) design submitted by Horizon Nuclear Power had successfully completed the UK general design assessment (GDA) process. While this isn’t the end of the regulatory process for the UK-ABWR (it still needs site level permits for Wylfa and potentially Oldbury), it does mean that the overwhelming majority of design issues have been signed off.

Nor does it mean that funding is in place, or that the construction of UK-ABWRs is a done deal. That remains an issue for the (hopefully near) future.

It does make this a good moment, however, to review the design; the changes and developments to the baseline design that were required to pass GDA may throw some light on the challenges faced by designs following UK-ABWR into GDA. We’ll look initially into the basic design of BWRs (including some aspects of its evolution), then the ABWR itself, and then the specific changes needed for GDA.

Boiling Water Reactors

Boiling Water Reactors are generally somewhat overshadowed by their Pressurised Water (PWR) cousins – however, in the US, Japan and a few other jurisdictions, they form a substantial part of the operating fleet. There are around 33 operating BWRs in the US out of approximately 100 nuclear stations in service, and they form the majority of the Japanese fleet – the global total is around 90 units

The most obvious difference between a BWR and a PWR is apparent in the name. Although both operate at similar temperatures, a PWR operates at high pressure (150-180 bar) and hence suppresses any boiling in the primary circuit. A BWR operates at lower pressure (usually c 80 bar), and raises steam directly in the core. PWRs transfer heat from the primary circuit into the secondary steam circuit that supplies the Turbine via large steam generators. By contrast, a BWR requires no steam generators, the steam raised in the core going directly to the turbine circuit.

This apparently simple difference gives rise to major differences in the overall design of the reactor.

The first of these is in the Core and the Reactor Pressure Vessel (RPV). As the majority of the core of a BWR isn’t submerged under liquid water – instead, it’s immersed in liquid at the base, with a rising column of “bubbly” steam and water of declining density we move up the core – it operates at somewhat lower power density than a PWR. Cores thus tend to be taller for a given power output.

Why does this effect come about? Both designs are reliant on the water coolant to fulfil another role – that of “moderator”. In order to operate with relatively low enrichment fuel, these “light Water Reactors” need to adjust the energy spectrum of the neutrons emitted by fission events. They need to slow the neutrons to much lower energies, at which they are much more likely to be captured in the fissile nuclei of the fuel, and hence cause further fission events. This slowing occurs as a result of the neutrons colliding with atoms in the coolant/moderator – and obviously, if that coolant/moderator is at lower density, fewer collisions will take place in any given volume of the core.

Early BWR schematic

There is another effect as a result of the steam/water column in the core. BWRs are, in general, very much less dependent on pumping to maintain water flow through the core due to strong natural convection effects. – to the extent that the most developed version of the design, the “Economic Simplified BWR” or ESBWR does away with forced circulation entirely.

The result is that BWRs need a rather larger RPV than do PWRs, albeit somewhat less thick walled due to the lower operating pressure. They’re typically about double the height (for equivalent power). About half of this is due to the longer core; the rest is due to the need to reduce the water content of the steam before it leaves the RPV. There is thus a steam separator and a steam dryer in the reactor vessel head which occupies considerable volume.

The presence of the separator leads, together with the fact that most of the power is generated in the lower part of the core, to something that many people find one of the weirdest aspects of BWR design.

The separator & dryer obstruct access from above into the core (it must be lifted out for refuelling, for example). Further, for maximum effectiveness the control rods need to be inserted into the lower part of the core, where most of the reactivity occurs.

The Separator and dryer make insertion of control rods into the core from above (a la PWR) all but infeasible. So, they go to the other end – they insert from below, from where they are also (fortuitously) most effective. Many people find this intuitively difficult to accept – they feel that, as an arrangement, it’s likely to lead to leakage. In fact, given that the reactor is pressurised to 80bar, there’s little difference between top or bottom entry in those terms – and sealing is easier in a bottom entry BWR arrangement at 80 bar than it is in a top entry PWR arrangement at 160bar.

The use of a “direct cycle” with steam going straight from Core to Turbine, although attractively simple, does have implications. Obviously, a breach in the turbine casing, or condenser or in the supply and return pipework constitutes a breach of the primary circuit. It’s therefore necessary to ensure a rapid isolation of the reactor in those circumstances. All steam and return lines must be fitted with fast-acting and multiply redundant isolating valves, preferably with “fail safe” operation and no reliance on external power supplies.

Direct supply also tends to result in somewhat greater operator radiation exposures than on a PWR; there is production of 16N, which undergoes beta decay, meaning that the entire primary circuit is radioactive. The saving grace is the 16N is very short lived – a seven second half life – which means that any hazard is removed within minutes of shutdown (unless there’s other contamination from failed fuel pins). Exposures can be managed down with a mixture of shielding and minimising operator entry into the turbine hall during operation, while the rapid decay means maintenance on shutdown plant suffers no real obstacles.

So, we now have our “Nuclear Steam Supply System” or NSSS. It’s remarkably compact – delivering around 4,000 MW of heat from a hemispherically ended cylinder of about five metres diameter and twenty or so metres in length. This contrasts markedly with a PWR – where we’d have a RPV of similar diameter but only 10-12 metres high, but several large steam generators each similar in size to, or larger than the BWR RPV. It also needs more ancillary plant such as a pressuriser (BWRs self-pressurise).

This comparative compactness drives the other major departure in design between PWRs and BWRs – that of the Containment.

PWR containments must be large to accommodate steam generators and so on. This has some virtues; the volume can be utilised for emergency cooling water tanks, and large contained volume means that even in the worst case, pressurisation in an accident is low – perhaps 2 bars. BWRs offer an alternative design choice. Apart from the RPV, they do require some space for access below the Reactor for maintenance work on control rod drives and circulation pumps, but that’s small by comparison with a PWR’s volume requirements.

Although some BWRs (the “Mark 3”) have used large volume containments, the economic choice is usually to go for a small volume, close fitting containment. These are designed for higher pressures than PWR containments, but even so, to avoid over-pressurisation, they are designed to suppress steam pressurisation. By this means, design pressures are acceptable at around 3-5 bars.

Mark 1 BWR containment

The exact means by which this is done has evolved with the development of the BWR. Earlier designs, like those in Fukushima were based on the main containment being the shape of an inverted light bulb, with the reactor occupying the “neck”. The spherical part of the bulb sat in a torus to which it was connected by pipes. The torus was part full of water, and the pipes opened below the surface of this water; any release of steam from the reactor was thus directed down thought the water and (in theory) condensed. Later versions evolved this concept. The early containments were welded steel, and the later versions used steel lined reinforced concrete.

Concrete, by contrast makes it easy to build a robust containment – and, at the relevant design pressures, don’t require expensive pre-stressed construction. They can be built using relatively simple reinforced concrete techniques (loads are shared between the concrete and the welded steel liner).

The ability to condense steam is vital. Both PWRs and BWRs are designed to be depressurised following shutdowns. Ideally this is a gradual process, with heat being dumped through normal route to the Turbine condenser. However, in an emergency, it can be necessary to depressurise rapidly, especially if there’s a loss of power to the high-pressure emergency cooling system with it becoming imperative to cool via lower pressure systems. These cooling systems must inject water against the prevailing pressure in the RPV – and obviously, with an RPV at 80bar, at 20 bar low pressure system isn’t going to be overly effective.

Some BWRs feature a rather elegant response to this requirement. Steam can be blown down from the reactor to the suppression pool via a small turbine which drives a high-pressure injection pump.

The early design was shown, at Fukushima to be problematic. The loss of heat removal functions led to the water in the tori becoming heated to 100C, at which they could obviously no longer fulfil their condensation role. That led in its turn to increased pressure in the containments leading to failures in various seals (which is how contamination came to be released). And, of course, as we now know the presence of control rod penetrations in the lower head seems to have provided a path for corium to escape into the containments. Also, in the cases of two of the reactors, the turbine-driven injections systems were closed down and couldn’t be reactivated due to loss of station power to drive valves.

The ABWR

The ABWR is the seventh major iteration of the design. It arose from collaboration between General electric (GE the originator of the BWR concept) and its two Japanese partners, Hitachi and Toshiba. As a result, the landscape is quite complex, with all three partners offering variants on the main design, while remaining in alliance. The UK-ABWR is an offering from Hitachi-GE; a not dissimilar, but slightly enlarged variant designed around common EU requirements is offered by Toshiba-GE; and the derivative ESBWR is marketed by GE itself.

(As an aside, in the course of writing this article, I’ve become aware that an earlier ABB BWR design, as built at Olilkuoto in Finland included most of the design improvements that differentiate ABWR from it’s direct precursors. OL1 and OL2 are amongst the best performing reactors anywhere, regularly achieving capacity factors in the middle 90s)

The first ABWRs built were units 6&7 and at Kawarishi-Karima in Japan; two further units have either been completed there, and another two were in construction when caught by the post Fukushima moratorium. Two units are similarly on hold in Taiwan.

The design of the core, RPV and steam system has seen only limited change as opposed to ongoing refinement. The coolant circulation system has been refined considerably.

BWRs utilise a concept where a certain proportion of the coolant bypasses the core – so called “recirculation”. This was originally achieved with separate piping loops into which the circulation pumps were welded. In ABWRs the recirculation path is entirely within the RPV, and the pumps are mounted in the RPV, submerged around the bottom head with external drives accessible for maintenance outside the vessel from below.

Advanced Boiling Water Reactor

These same pumps have a secondary function – by varying speed and the proportion of coolant flow recirculated, the height of coolant in the core can be varied, providing a useful additional fine control mechanism for load-following.

The only other notable change from earlier generations is a move to electrically driven drives for the control rods; specifically, some of the rods have “fine motion” drive allowing precise tuning of power output for load following

The containment is constructed from reinforced concrete with a welded steel lining; the suppression function is supplied by a square section low mounted toroidal tank. There is a large reserve of additional cooling water held above the reactor in more tanks, and there are natural circulation cooling loops arranged to remove heat from the suppression pool to the upper tanks which are, in their turn vented to ambient air.

UK-ABWR

The UK proposed design from Hitachi incorporates further refinements beyond those of the core ABWR, many reflecting learnings from Fukushima.

The most obvious of these is major reinforcement of the reactor fuelling hall (the space above the RPV and containment); this has been strengthened to the point where it is accepted as being proof against the impact of a large commercial aircraft.

Much safety critical equipment has been relocated to higher levels in the plant, or to flooding proofed rooms. External supply points for power and cooling water have been incorporated so that on-site generators and water pumps can directly supply equipment in the case of a loss of external power and inability to use “hardwired” generators. Several key valves have been relocated and provided with alternate actuation methods (including those for the blowdown turbine system), following issues at Fukushima with inability to operate such valve due to loss of electrical supply.

There is added protection against containment overpressurisation, and “hardened vents” incorporates (these are simple bursting discs which vent through extensive filtration systems).

One aspect I have been asked about previously is worth mentioning. Why, given its apparent technical superiority, was ESBWR not put forward as opposed to ABWR? The answer is simple. ESBWR exists only on paper, and at a level of design resolution that’s some way short of what’s needed for UK GDA. Add to that a fully worked up ABWR design was available, with all electrical design based on 50Hz equipment (as opposed to US 60 Hz practice).

Construction and Operational Experience

The simplicity of a BWR reflects directly in the build time; of those stations which have NOT had their construction interrupted by political interventions, average time from first concrete to first power is around 39 months. The author recalls a discussion with a GE representative at a WNA conference in which it was claimed that the critical path is equally defined by the construction of the nuclear island and the turbine island; notably, this time is much the same as that required for large CCGT units like West Burton. It should also be noted that this is using Japanese construction site practice with a single day shift; some improvements should be possible given UK “double-day” shift working.

By contrast, construction of ABWRs in Taiwan has been more problematic, mainly due to regulatory interventions reflecting shifting Government policies (including a two year moratorium on construction work).

It’s harder to extract an objective assessment of the operational performance. A number of anti-nuclear groups have made claims of low availability/capacity factor, most notably at Kashiwazaki-Kariwa – capacity factors between commissioning (late 1996) and 2011 are around 70% for both reactors. However, these include periods of extensive imposed outages in the aftermath of earthquakes in 2004 and 2007, and in the ongoing aftermath of Fukushima. A 16 month outage was imposed after the 2007 quake, and several months after 2004. Correcting for this pushes up capacity factors to above 80% – prevailing Japanese refuelling and inspection practice tends to limit capacity factors to around 85%, so in fact the performance is respectable.

Notably, the second of those earthquakes exposed the reactors to a 6.6 magnitude earthquake centred within 12 miles of the plant, resulting in lateral ground accelerations of around 0.45 gravities; neither reactors suffered damage.

Conclusion

It’s true that their ability to supply very high temperature dry steam is poor, and hence so is their thermal efficiency; however, this matters less in a nuclear context, where fuel costs are a very minor fraction of the overall. The reliability and cost implications of doing away with steam generators and much of the primary circuit are very attractive – and the control possibilities of manipulating recirculation rates are a bonus. In the specific context of ABWR, the safety systems aren’t fully passive, but they do represent an impressive step in that direction with natural circulation heat removal but powered actuation of the associated valvegear.

The construction performance of these plants has been exceptional – to the degree that, if I might make a challenge to enthusiasts for SMRs – it’s hard to see the advantages on deployment of 300 or 400MW units, when a 1600MW unit can be built in 3-4 years, with the entire NSSS being assembled in the factory and shipped to site. Note that the proposed Rolls-Royce SMR has a five year construction period, and requires on-site integration of RPV, Pressuriser, four SGs and circulation pumps.

All in all, this looks like an excellent design. Fingers crossed that funding issues can be resolved and construction started as soon as possible.

Further Reading

The high level overview of BWRs from the US NRC

The UK NNL’s high level assessment

On the ABB “ABWR”