In recent months, as the large Hinkley Point nuclear project has hit a succession of problems (see the recent blog by Phil) there has been increasing attention to the prospects of small nuclear reactors, most often small modular reactors (SMRs). These are seen as either a complement to or, increasingly, a substitute for, very large units like Hinkley. Disquiet over the high cost and delays to Hinkley – some of it from within the nuclear community – as well as signs of a faltering global nuclear renaissance – have led to questioning whether the long-established conventional wisdom that bigger units are cheaper than small reactors is any longer true.
The UK’s National Nuclear Laboratory (NNL) has produced a ‘feasibility’ study which argues that SMRs might eventually prove cheaper than Hinkley-sized units, and the House of Commons Select Committee on Energy and Climate Change has urged Government to spend public money to develop a demonstration small reactor in the UK. The Committee suggests, implausibly as we shall see later, that ‘SMRs…are a viable proposition for future deployment in the UK in the next decade.’
Over the last forty years, the size of nuclear reactors has consistently risen, so that the main designs now being offered by vendors are well over 1000 MW – the apogee being the French company Areva’s 1650 MW EPRs being built at Hinkley. Nuclear engineers have always argued that bigger means cheaper per unit of installed capacity, and per unit of power produced. This seems to make sense – there are irreducible overheads for nuclear plants plus classic engineering economies of size. Thus if volume doubles, the quantity of materials needed is considerably less than double. This does not mean that newer, larger units are cheaper than the older smaller ones, as other forces have caused reactor costs to rise over time and across the board. So a consequence of generic cost increases and ever-larger unit sizes means that the investment costs of new plants are very high – Hinkley is expected to cost £16 billion before financing charges (and £24.5 bn. after taking them into account, and before any revenue is earned).
These large and financially hard-to-digest costs added to large commercial and political risks have helped reinvigorate an older hope in the nuclear community that small (less than 300MW) units might prove viable. The idea here is that factory assembly and mass production economies could overcome the cost disadvantage of small unit size. Or even if this is hard to envisage, then at least the up-front financial commitments are lower, so the financing issue might be less onerous.
The increasing enthusiasm for the ‘rescaling’ of nuclear power in the nuclear community may signify recognition of of a wider energy transition underway towards more decentralized forms of provision more generally. As Cooper points out, advocates of SMRs suggest that reduced power demand due to the recession and increasing drives to energy efficiency makes SMRs appealing. Their lower total capital commitments, reduced construction times, and smaller unit size seem to make them more flexible and thus better suited to current trends in many energy systems. Thus SMRs are sometimes seen as playing a potential role in adapting to the possible contexts of future electricity markets where nuclear will have to be more accommodating of intermittent renewables.
So what are prospects, both in the UK and more widely? None of the designs, including the most credible, which are based on scaled-down versions of currently deployed PWR technology, is yet ready: NNL speaks of ‘detailed technical challenges’ not yet resolved. It is therefore no surprise that no-one has yet built a single SMR let alone the up-front commitment to large unit numbers that would be needed to make the economic case remotely credible. And the safety licensing process that will need to follow design completion would, according to the Chief UK nuclear inspector, take up to 6 years in the UK. The Select Committee’s 10-year vision already begins to look wildly optimistic, especially since it is already nine years since the UK Government gave an enthusiastic go-ahead for much more mature nuclear technologies than any of the SMRs – and no financial deal has yet been reached. SMRs are a classic case of supply-push technology development – no potential user of SMRs, mostly electric utilities, has expressed any serious interest in them.
Nevertheless commitments to low carbon technology options in the UK and elsewhere mean that Governments, including the UK’s, are willing to intervene in markets and require deployment of technologies with substantially higher costs than would be chosen commercially. So could SMRs, even on a rather longer time-scale than a decade, become deployed in the UK and elsewhere?
First, do we have any idea what SMRs would cost? The answer is clear: we have virtually none. NNL canvassed the main potential vendors of SMRs for their views and, somewhat unsurprisingly, they all (as eventually did NNL) suggested that under conditions that have to be seen as extraordinarily optimistic, SMRs might eventually prove cheaper, per unit capacity and power production, than large reactors. Historically, vendor estimates of nuclear costs have proved almost grotesquely optimistic and there is no reason to believe, given the obvious self-interest of vendors in promoting their potential products that this has changed. Given the uncertainties involved, the conclusion is that costs are essentially unknowable at present, but that there are powerful reasons to suppose that they will prove much higher than the industry now promotes.
This uncertainty makes doing a world ‘market survey’ very difficult, but NNL nevertheless conducts one. In two scenarios, ‘niche’ and ‘parity’ (of cost) it concludes that the world market would be in 2035 only just over 5 GW in ‘niche’ but 65-85GW in ‘parity’. It then suggests a potential UK market of between 7GW and 21GW in 2035, the latter number being frankly not credible under any conceivable circumstances. These hoped-for UK markets are also linked to the idea that the UK could become a major technological player in SMR technology, a view that seems tinged almost with fantasy, given that all significant SMR development to date has been outside the UK:. In the USA for example the Obama administration has pledged a further $217 million to the SMR company NuScale, following substantial earlier Federal funding for two SMR designs. Over a much longer period than a decade, some SMRs might nevertheless be deployed, in the UK and elsewhere. Whether the subsidies necessary for this would be a good use of public or consumers’ money in terms of carbon emission reductions must however be open to major doubt.
Professor Gordon MacKerron, Professor of Science and Technology Policy, SPRU, Co-Director of Sussex Energy Group.
Dr Philip Johnstone is a Research Fellow at the Science Policy Research Unit and a member of The Sussex Energy Group
Follow Sussex Energy Group
There is a typo in the final paragraph: ‘It then suggests a potential UK market of between 7GW and 21GW in 2015’. That should be 2035, not 2015.
Thanks Matt – absolutely right – we have updated the text to 2015. Thank you for flagging this up to us.
The important point to emphasize is modularity. Small is overemphasized. Why would modular reactors cost more than non-modular reactors built and tested on site? It’s a bit like saying that a custom-made wardrobe will be cheaper than one from IKEA. If the SMRs are running at atmospheric pressure with liquid metal or salt coolant they’ll do away with the massive pressure vessels and bizarre methods of countering the problems when using covalently-bonded coolants such as water. Namely: 2H2O + gamma rad -> 2H2 + O2 -> big explosion. Many of the proposed modular reactors are intrinsically safe with much of the hardware easier to source. There’s every reason to believe that world-wide take up of modular reactors will cut capital costs by at least half.