Tuesday, 3 November 2015

The Hinkley Point deal is prohibitively expensive

So say the Guardian, again, today. It's not "prohibitively expensive" but it is "too expensive". Hinkley Point deal is too expensive is the common doxa, among both nukes, anti-nukes, and neutrals. The question of expense has interested me for sometime. Why so expensive, what can be done about that?

The AREVA EPR is the only modern reactor design currently approved by our regulator: Office of Nuclear Regulation (ONR). If we want a reactor right now, this is the only one we can build. Had we more choice, we could build nuclear power plants for much less than this. Probably about half the price. [KEPCO build their APR1400 reactors in UAE for £3¼ billion each] So why don't we? Given the EPR design is the only one with a current UK GDA, the obvious question to ask is: what's holding up the UK GDA process? The obvious answer is often: lack of resources. But what resources?

Government framed the UK nuclear reactor regulatory process to discourage vendors from applying for reactor approval and spent no effort encouraging vendors to apply to get reactors approved. In effect, that's how current regulation works in the UK. There was no design nor malice in this, just the habit government have of copying what seems to work elsewhere. UK government, basically, copied the US model (how US funds their NRC : Nuclear Regulatory Commission). NRC funding works reasonably well in the USA because it's a big country. US electricity demand is 11 times UK's. Because US population is 5 times ours, and per capita electricity use just over twice ours. After a US reactor GDA application is approved, the reactor vendor has reasonable confidence that many will be built. That allows Americans to justify the up front costs of gaining NRC approval. Not so in the UK. A vendor trying to build reactors here can't expect many of their reactors to be built here at all. The market for new builds is just not that large.

Electricity requirement of UK / US
Average demand (GW)Per capita use (kWh)Population
US46313,010320 million
UK425,95864 million

UK Government framed legislation such that a large fee, probably about £40 million, is charged up front for reactor approval. This is large enough to discourage vendors putting their designs forward for GDA validation. That, and the 5-year approval period. The UK GDA process began in 2007 with 4 competing designs: the AREVA EPR (PWR), the Toshiba-Westinghouse AP1000 (PWR), the GE-Hitachi ESBWR (BWR) and the AECL ACR1000 (Candu). The ESBWR and the Candu designs were withdrawn, leaving only 2. The AP1000 application was later suspended because the submission was not in S.I. units. That left only 1 reactor: EPR. Our system of taxing vendors beforehand led directly to GE-Hitachi pulling their ESBWR design out of the approval process despite spending £20 million on British approval. It partly led to AECL (Candu) not really trying. That might not be so bad in a world with lots of vendor competition. There are only about 7 such vendors in the world: AREVA (France), KEPCO (South Korea), Westinghouse (USA-Japan), GE-Hitachi (Japan), Rosatom (Russia), CANDU (Canada), Various Chinese. Some argue we should not consider Russian and Chinese designs due to security issues. CANDU reactors have positive voids, so many will not want them due to safety issues. That narrows the field down to 4 nuclear reactor vendors! So what other reactors could reasonably have been submitted for UK GDA? KEPCO have perfectly good designs they build at low cost (e.g. the APR1400 in UAE), but they've never been encouraged to apply for a GDA. That despite Britain and South Korea signing a special free trade deal in 2012. AREVA understood that their EPR design was too expensive at a very early stage: nearly 20 years ago. That's why they increased it's capacity from an initial 1100 MWe to 1650 MWe - to mitigate the expense by trying to gain an advantage in scale: in theory the cost of building and running 2 giant reactors being lower than 3 large reactors. It never panned out that way. Yet AREVA always had other designs it could've submitted for UK GDA. They had the KERENA and ATMEA1 designs too (since 2009 and 2007 respectively).

A better system would've seen the approval fee mostly paid for by a tiny tax on nuclear power, with an up front fee of, say 10% of approval costs. That would force immediate costs on the vendor of ~ £4 million. I think that would've encouraged KEPCO to apply for a GDA on a purely speculative basis. Imagine that it costs £40 million to approve a reactor for UK use. 10 different reactor approvals would cost £0.4 billion. A lot of money. The argument goes: the tax payer shouldn't have to cough up this money. Yet over a 35 year period, during which the Hinkley C contract for difference (CfD) applies, the electricity bill payer will cough up about £35 billion more than they would otherwise pay had their electricity been sourced at the current electricity wholesale price. We pay a tax anyhow. It's just that we're paying a tax at least 50 times larger than need be. Not quite: nearly all new nuclear power plants will need some kind of CfD but we could reasonably expect the difference in CfD and wholesale price to be half what it is for the EPR. So, instead of paying £35 billion more over 35 years, we'd pay £20 billion more over 35 years for a different reactor design.

  1. Guardian Live: should we say yes to nuclear power?
  2. KEPCO are building their APR1400 reactors in UAE for £3¼ billion each
  3. The costs of UK GDA for the AREVA EPR were said to be £35 million in 2012
  4. In 2007, GE-Hitachi Nuclear Energy Submitted ESBWR to UK Regulators for Generic Design Assessment (GDA)
  5. CfD: contract for difference. A CfD means that the customer pays more for electricity than they would were they paying the market price. It allows new electricity capacity suppliers to recoup their capital costs. Think of it like a mortgage. You buy the house now, live there but it takes you 25+ years to repay the cost.
  6. The saga of Hinkley Point C: Europe’s key nuclear decision
  7. vendor: I'm using the term vendor here to mean `nuclear power plant design and construction company`
  8. GDA = Generic Design Approval. This is like a licence and safety certificate allowing one to build several nuclear reactors in one country.


  1. I wonder how much the Tories' obsession with cutting the government deficit (when it's really the current account deficit that we need to worry about) has contributed to bad PFI deals like this?

  2. "The AP1000 application was later suspended because the submission was not in S.I. units."

    Oh, FFS!

    1. The AP1000 generic design assessment has will take over 10 years (2007 to early 2017). It should've taken up to 5 years.

      I'm not convinced Westinghouse (AP1000 owner) were entirely at fault. It seems to me the regulator were biased against the AP1000; presumably because the EPR was full of safety features; indeed overladen with such.