• 0 Posts
  • 10 Comments
Joined 1 year ago
cake
Cake day: June 11th, 2023

help-circle


  • His figures are ridiculously optimistic for nuclear, $6000/kW and 6 year construction times.
    Flamanville-3 and Olkiluoto-3 were both 12 years over their 5 year construction schedules. They were supposed to cost €3.3B and €3B respectively for 1650MW. Flamanville is expected to end up somewhere over €20B (€12000/kW), and Olkiluoto is somewhere around €11B, only due to ‘not to exceed’ limits in the supply contracts.

    Hinkley Point C has gone from £16B to near enough £30B for 3200MW (£9400/kW)

    It was the same with Vogtle 3 & 4. The preliminary budget of $12B, was changed initially to $14B at the start of construction. It’s now somewhere around $30B and 7 years late. The two AP1000s have a combined output of 2200MW ($13000/kW).
    V.C.Summer 2 & 3 was a similar pair of AP1000s. Costs went from $9B to $23B when the project was cancelled mid-construction.

    Wind and solar are far faster to deploy, and typically on or near budget. The new, much cheaper redox flow batteries (100 MW/400 MWh for $266M Dalian, China) are capable of smoothing intermittency in areas without hydro, which can perform a similar function.

    Edit. I should add that as of 2021, the global average for onshore wind is roughly $1300/kW. Prices continue to fall as new designs are introduced.


  • That’s just not true. The Westinghouse AP1000 was given type approval in 2011. It’s what is referred to as a GEN3+ reactor. A lot of R&D was put into simplifying the design, reducing the number of pipe runs, valves, pumps etc compared to GEN2 reactors. It also used large sub assemblies that were factory built off-site then moved for final assembly.

    In theory they should have been cheaper to build, but they weren’t. Large assemblies that don’t fit together properly need a lot of very expensive site time for rework. There were other issues on top of that, which just compounded the assembly problems. It’s how Vogtle ended up going from $12B to $30B+, and V.C Summer went from $9B to an estimated $23B when the project was cancelled while under construction.

    The EPR units from Areva were similar GEN3+and received type approval in the early 2000s. They had similar cost overruns, for similar reasons.

    I have strong reservations about SMRs. So far the cost/MW is about on par with traditional reactors while the amount of waste increases by 2 to 30x traditional reactors depending on technology used.

    There are reasons why reactors moved from 300-600MW units to 1000MW+ in the first place. The increased output would cover what was thought to be marginal increase in costs. That turned out to be at least somewhat true.


  • Just like assuming a perfectly spherical cow, or a frictionless surface, you can completely ignore the economics, the massive cost and schedule overages to make nuclear work.

    Flamanville-3 in France started construction in 2007, was supposed to be operational in 2012 with a project budget of €3.3B. Construction is still ongoing, the in-service date is now sometime in 2024, and the budget has ballooned to €20B.

    Olkiluoto-3 is a similar EPR. Construction started in 2005, was supposed to be in-service in 2010, but finally came online late last year. Costs bloated from €3 to €11B.

    Hinkley Point C project is two EPRs. Construction started in 2017, it’s already running behind schedule, and the project costs have increased from £16B to somewhere approaching £30B. Start up has been pushed back to 2028 the last I’ve heard.

    It’s no different in the US, where the V.C. Summer (2 x AP1000) reactor project was cancelled while under construction after projections put the completed project at somewhere around $23B, up from an estimate of $9B.

    A similar set of AP1000s was built at Vogtle in Georgia. Unit 3 only recently came online, with unit 4 expected at the end of the year. Costs went from an initial estimate of $12B to somewhere over $30B.

    Note that design, site selection, regulatory approvals, and tendering aren’t included in the above. Those add between 5-10 years to the above schedules.