What Is Prudent?
 Red Flags Clearly Ignored
 In Vogtle 3&4 Project

Georgia Power is in the midst of a prudence review of its spending at the Vogtle 3 and 4 nuclear project—a review that undoubtedly will be lengthy, comprehensive, and mind-numbingly dull, turning on such issues as whether given decisions were “reasonable given the facts and circumstances which were known or reasonably should have been known at the time the decision was made.”1The real question though isn’t whether Georgia Power has spent customers’ money (and believe you me it is customers, not the utility, that are paying for this long-delayed, much over-budget project) prudently, but where the hell the adults where when the decision was made to go ahead with construction in the first place.

A close review of Georgia Power’s own documents (the filing can be found here) in the case shows two things: First, there were red flags aplenty when someone, anyone in the decisionmaking process would have been justified in standing up and saying, ‘Hey, wait a minute, what are we thinking?” Second, executives at Georgia Power apparently are color-blind and can’t see the color red, and continue to insist that everything—and I mean everything—they have done during the past decade to build the two new Vogtle units has been appropriate. For example, writing in the introduction to the company’s 885-page filing to the public service commission, Paul Bowers, Georgia Power’s chairman, president and CEO, offered up this classic: “Every dollar, and every day, that has been invested has been necessary to complete these new units safely and correctly. Our reports will establish that the new units could not have been built for less money or in less time than it has taken.” That may be, but that kind of logic can justify almost any expenditure. If the utility had paid attention to the red flags hanging everywhere it might have more accurately estimated the project’s cost and required construction time in the first place, which in turn might have led to a different decision by the commission.

The problems with the project go back to the very beginning. For starters, what were Georgia Power and Westinghouse executives thinking in April 2008 when they signed an engineering procurement and construction (EPC) contract for the two new nuclear units that was essentially a fixed price affair—even though detailed design drawings for the reactor’s construction were still years from completion, meaning, for the clear-eyed, that the contract price was little more than an estimate scribbled on the back of a cocktail napkin.

To be fair to Georgia Power and Westinghouse, it is easy to understand their mutual interest in signing this type of deal. For Georgia Power, it represented cost certainty—a key issue following the debacle surrounding construction of the first two units at Vogtle, which ended up costing billions (yes, billions) more than projected and taking years longer to construct than initially forecast. For Westinghouse the motivation was even stronger—the contract with Georgia Power represented an opportunity to prove that its new AP1000 reactor was commercially viable.

So the deal was signed, even though a cursory review of past nuclear construction projects—projects that executives at both Westinghouse and Georgia Power undoubtedly were familiar with—would have raised stadium-sized red flags.

Fixed price contracts in the nuclear power business had essentially ceased to exist by the 1980s [the last period of major nuclear plant construction in the U.S.] because suppliers had lost so much money on previous construction projects, wrote Charles Huston, one of the nuclear power consultants hired by Georgia Power to evaluate the prudence of its Vogtle 3 & 4 spending in the current PSC review. The problem with fixed pricing, he pointed out, was that design changes and schedule delays frequently ran up the costs with no potential for recovery.

Nevertheless, donning the rosy-colored glasses apparently available for free in Georgia Power’s executive offices, Huston dubbed the utility’s ability to obtain a fixed price contract for Vogtle “a major achievement”—even though design changes and schedule delays for the new reactor were almost inevitable given its first of a kind status, absence of detailed construction drawings and lack of nuclear construction infrastructure. The only thing the fixed price contract really did was kick the cost can down the road, setting the utility and the construction team up for failure when the issues ignored in 2008 reared their ugly head, as they inevitably did. As Huston points out:  “Accordingly, certifying that a design is complete for issuance of the COL [combined operating license] does not also mean that the design is complete from a construction point of view.  All major construction projects experience continual design review as construction proceeds, which impacts schedule and cost.  This risk is heightened with a FOAK [first of a kind] design in an industry that has been dormant for a generation with respect to construction.”

But no one at either the utility or Westinghouse said anything.

A key factor in the construction savings expected by Westinghouse was its modular design approach, whereby key components would be manufactured offsite and then transported to the construction site for assembly instead of being built one at a time at the site itself. It was all supposed to run like clockwork, except the clock quickly broke down.

In Westinghouse’s initial schedule, for example, construction of what it dubbed CA01, the most important module in the containment building was supposed to begin in June 2011 with the fabrication of the first of its 47 sub-modules. All of those sub-modules were supposed to be complete three months later in September and then CA01 and two other modules could be set in place and welded together by April 2012. Done and done in well under a year.

Unfortunately, it didn’t quite turn out that way.

As Huston points out in his review, no work at all had been done on the CA01 sub-modules as of December 2013—18 months after it was all supposed to be completed. Worse, he wrote, “I received a site tour of the nuclear island for Vogtle Unit 3 on February 1, 2016.  Work had not been completed on module CA01 in the containment building to allow the setting of modules 2 and 3.” [Emphasis added, although perhaps not necessary.]

Going on four years and counting, so what happened? The obvious, and the predictable, if anyone had wanted to look. A warning about potential problems was even included in the original Georgia Public Service Commission reactor certification process by Dr. William Jacobs, the expert hired by the PSC to evaluate Georgia Power’s proposal. “As no AP1000 units have been built, it is likely that problems will be encountered during the construction process that will require redesign and rework,” Jacobs wrote.

The problems showed up almost immediately, as Huston pointed out: “The main problem was that SMS [Shaw Modular Systems, the manufacturer responsible for building the modules that would then be assembled onsite] had difficulty finding management personnel with the required level of expertise in quality control of fabricated modules for a nuclear plant.  These personnel did not exist because no fabricator in the United States had ever fabricated modules for a nuclear plant on this scale.”

But no one at Westinghouse or Georgia Power could have foreseen that being a problem?

Similarly, the modular design, by definition, would require a lot of welding to put the various pieces together. Just to stitch CA01 together, Huston said, would require 47,000 linear feet of welding.  “This is a staggering amount of welding that clearly was not fully appreciated by the consortium when the decision was made to use a modular approach for the construction for Vogtle Units 3 and 4,” Huston wrote.

How could Westinghouse and its construction partners not have appreciated how much welding would be involved in their own damn design?

And then, surprise, surprise, the partners found that it was hard to find welders capable of meeting nuclear standards—even though construction on the last nuclear reactor had been wrapped up at least 15 years earlier. Don’t you think that someone, anyone at Georgia Power or Westinghouse or one of the other partners might have thought to ask: Where are we going to find the specialized construction workers needed to pull this off?

But no one asked, because no one wanted to know.

Regardless of the red flags, Westinghouse and Georgia Power went forward with the 57-month construction timetable for the new reactors, a schedule that Huston noted was “extremely aggressive…based on the actual schedule durations of the nuclear plants completed in the 1980s at the end of the previous generation of nuclear construction.” For comparison’s sake, Huston pointed to Wolf Creek, a project that he worked on, that entered commercial operation in 1985. Construction on that plant took 98 months, he wrote, even though it did not encounter any significant licensing delays and it was a standardized nuclear design.

Why, after 20-plus years of essentially no construction activity, Westinghouse assumed it could slice more than three years from the new reactors’ timeline is a good question. Why, after having been through the painful Vogtle 1&2 construction process, no one at Georgia Power asked Westinghouse about its schedule is an even better question.

Sadly, there undoubtedly will be more questions to answer in the years to come. Huston’s review notes that “no significant mechanical, piping, and electrical work had started in the containment building” as of his February 2016 site visit, and there had been only limited progress in the auxiliary building. The original schedule called for piping and electrical work to begin in the containment building in April 2012—meaning this part of the project is now 48 months behind schedule. Whether some of that time can be made up is uncertain, but past history would say no, as Huston points out:  “The previous generation of nuclear plants generated a large number of fabricator and construction design changes for nuclear island pipe, pipe supports, cable tray supports, conduit supports, and instrumentation fabrication and erection.  Since the erection of bulk pipe and electrical work has not yet begun in the nuclear island, the amount of additional design changes for Vogtle Units 3 and 4 is unknown at this time.”

What is known, Huston added, is that the “nuclear islands for Vogtle Units 3 and 4 have very tight working spaces that are conducive to interference and design changes.” In addition, there is that welding issue, which Huston said is “likely to continue to be a problem when the welding of nuclear island piping and pipe supports increases in the future. “

Despite the questions he raises about the project, Huston comes down squarely on Georgia Power side: “I find that GPC’s actions in managing the project and the cost of the work performed by the EPC contractor have been prudent and as a result, all costs on the project have been prudently incurred.”

In regulatory speak Huston may well be correct, but that misses the larger question: Where were the adults willing to ask the hard questions and challenge the easy assumptions?

–Dennis Wamsted

1 Georgia Power Company’s supplemental information report, page 26, citing an earlier decision by the Georgia Public Service Commission concerning prudence.

16 thoughts on “What Is Prudent?
 Red Flags Clearly Ignored
 In Vogtle 3&4 Project

  1. Dennis, I know it’s difficult for many to think long-term, but that’s not only what’s required for energy – it’s what’s required for the environment.

    Vogtle has the potential to be there long after you and I have escaped our Earthly bounds, generating clean energy for our ancestors. You can snipe all you want about a $billion here or there – it’s pennies in the long-term scheme of things. Cheaper than solar, cheaper than wind, cheaper and less dangerous than the fossil fuel generation solar and wind require to make them even possible. When the disconnected Ivanpah mirrors are still frying birds in the California desert in 2060, Vogtle will be generating clean energy 24/7/365 (assuming fossil fuel schemers don’t determine otherwise).

    So get over it. These are not major issues compared to climate change, and trying to turn them into major issues only makes dealing with climate change that much more problematic.

    1. Vogtle has more potential to be closed prematurely for economic reasons, following other such nuclear plants such as 1.3GW GrafenRheinfeld, VY, etc.

      Because continuation of the cost decreases of renewable during next decades are widely predicted by experts. The expected cost decreases are substantial: wind (3%/a), PV-solar (8%/a), battery storage (10%/a) and Power-to-Gas for seasonal storage (10%/a).

      Recent study by Agora, Germany’s scientific think tank regarding these matters, concluded that in 2050 unsubsidized PV-solar will produce for 2-3cent/KWh in Germany, despite its poor (near Alaska like) insolation.

      1. The only reason renewables are being built at all is because of massive subsidies from your federal and state tax dollar and mandates for preferred purchase agreements for renewables. Otherwise they would still be a pipe dream.

        1. As the costs of wind & solar decrease fast, those subsidies are reducing fast. They will be near zero before 2030.

          Consider Germany as an example (as it has a consistent policy):
          In 2005 the Feed-in-Tariff (FiT) for a 40KW rooftop solar was 51cnt/KWh.
          In 2015 the FiT for same rooftop solar was 11cnt/KWh.*)

          As experts predict a continuation of the costs decrease in next decades, you may assume that it will be ~3cnt/KWh in 2030. Far below the cost price of new nuclear as well as most new ff generators.
          Worse, those solar installations deliver direct to the end user, so one should compare with end user prices which are >10cnt/KWh…

          Furthermore, last year ~40% of new rooftop solar installations in Germany also got a battery covering the evening hours.
          *) The FiT is fixed during the first 20years. Thereafter no guarantees. So the owner has to sell his electricity on the market.

          These regulations support the development of Virtual Power Plants (VPP’s) that deliver 100% renewable electricity to it’s customers, cheaper than the big incumbent utilities can. An example of such VPP:

    2. Ivanpah in 2060? The design life of Ivanpah is 30 years. But, I doubt it will last that long. It can’t even generate the de-rated capacity (377MW vs 418MW) they said it would do, and they are burning a significant amount of natural gas to make up the difference.

  2. One of the first delays occurred when it showed that the sub-contractor poured the concrete for the base, not level. It had an unevenness far beyond the specs.

    That unevenness was not detected by inspectors of the principal during the process of pouring the concrete, but had to be detected by inspectors of the NRC afterwards.
    It wasn’t corrected, but some compromise with NRC was made after a delay of ~5months (probably thanks to political connections of the governor)…

    Allegedly the sub-contractor got paid anyway after deducting a small amount (seems he was related to the governor).

    So even such simple job (there are lasers since the nineties) wasn’t done correctly!

    It shows that the conditions are wrong. While that is needed for such project, top-management doesn’t do it’s utmost.
    They are covered as rate-payers have to pay and invest anyway…
    So more significant delays and cost overruns can be expected.

    And the PUC cannot reverse emotionally, especially not the more is invested.
    Only if the PUC & Topmanagement is completely replaced with capable people, there is some chance.

    So the new NPP will produce for >10cnt/KWh while market prices are decreasing towards 3-4cnt/KWh level under the pressure of the decreasing costs for wind, solar and storage.
    Which imply far more unnecessary extra costs for Georgia’s rate payers in the future.

    1. Bas, these errors are routine – not only for nuclear plants, but for coal plants, wind turbines, and solar farms.

      They happen. All. The. Time.

      Do you know how competent contractors deal with mistakes? No, they don’t shut down the project and go home, crying in a hankie. They fix them ASAP. Because it’s not only the public’s safety, but their job if they don’t.

      Sometimes the mistake is deemed insignificant, and they’re granted a permit exception. They leave it, and move on to more important problems (like ejecting antinuclear protestors who might hurt themselves, then file a lawsuit).

      1. Bob, these errors do not happen with wind turbines and photovoltaic solar farms… because the technology is so much simpler that there is no opportunity to make such errors.

        They’re manufactured mass-produced products, not one-off special buildings. That’s a huge difference, and a revolution in energy generations.

        (Ivanpah is an irrelevance, since it’s not photovoltaic.)

      2. So may be the construction sector in USA is rather backwards now.
        May be they never heard about doing things “first time right”.
        Or didn’t learn from Japanese quality assurance (the reason Japanese cars are more reliable).

        To my dismay, it seems to fit with the lousy quality of electricity supply in USA (8 times less reliable than in Germany) and the frequently backwards regulation environment which often prevents sharp competition.

    2. Where do you get your numbers?
      As Ivanpah was mentioned, its supposed to generate 377MW for $2.2B with a 30 year design life.
      VC Summer is also an AP1000, 2200MW for $9.8B with a 60 year design life.

      Taking into account solar’s 25% capacity factor and nuclear 90%, to recover the construction costs of the design life with these capacity factors, Ivanpah will have to bring in about $100/MW-hr, whereas VC Summer will be about $11/MW-hr.

      The only way Ivanpah can turn a profit is though huge tax breaks and government subsidies.

      The fact that solar and wind costs are decreasing is a fallacy based upon government intervention and price supports.

      In 2013, the feds pumped $15B into renewables and for that their percent of the total US generation went from about 1% to 5% of the total. While nuclear with about $1.6B in subsidy and closing of 5 plants, maintained essentially the same percentage of our total generation.

      Which was the better investment?

      1. The Ivanpah technology was out-dated from the start.
        Spanish utilities operate a few similar installations for decades but decided not to expand because of the high costs and the poor prospects for cost reduction.

        PV-solar became much cheaper, so Spanish utilities changed to PV-solar installations (the panels often mounted on a one or two axis turning construction) .

        I’m puzzled why the investors in US decided for the old backwards technology despite its much higher costs???

    3. Actually, I don’t think your assertion is correct. The first delay was related to the pad, but not for that reason. There are standards that are used for installing everything. In the case of the pad, there are a couple specs. The installation was done by one spec instead of the one cited in the standard design of the AP1000. Both specs afford the same amount to of strength, use different methods. That required a license amendment and delayed the concrete pours.

      1. Leonard,
        The first significant delay concerned not the concrete pour, but the fact that the sub-contractor did not level the concrete according to the specs (which is easy with present laser technology).
        That caused ~6 months of delay as NRC rightly refused to approve.

        A compromise (thicker next layer) was accepted only after political intervention of Georgia’s governor. Which creates a dirty feeling as a level concrete floor is important in case of radio-active leakages; so the fluid moves to the drain and no radio-active fluid stays at the floor.*)
        So the compromise implies a lot more problems to restart, in case of leakages…
        *) Japanese NPP had that problem. It was one of the reasons it had to stay off-line during years.

Leave a Reply

Your email address will not be published. Required fields are marked *