Recent nuclear power cost estimates – separating fact from myth

I’m about to head to Madrid for two weeks, to do some research with a colleague at the National Museum of Natural Sciences — but do not fear, I’ll still be on the ‘net in the evenings to run the blog!

I see, happily, that I leave behind an ongoing debate on the need for nuclear power in South Australia. In the news today it was reported that the SA Liberal Party has supported a motion to debate the use of nuclear power technology to reduce carbon emissions. Good on ’em — I trust (more like vainly hope) that the Labor party, Federal and State, will soon also be willing to open the floor. They’re certainly now getting serious prompting from their traditional supporter base.

However, something in that news story on the SA Libs particularly irked me (no, it was not the potshot from Kevin Foley; that was unsurprising political wedging). It was this:

Greens MP Mark Parnell says he does not think Ms Redmond’s stance will damage her credibility and the debate is futile.

“We are not about to have nuclear power here any time soon,” he said.

“South Australians don’t want it; it’s too expensive; it’s too dangerous; it’s not the solution to climate change; and we don’t know how to dispose of nuclear waste yet.”

Now Mark knows about Integral Fast Reactors — he’s been in the audience when I’ve talked about them. So why would he continue to claim that we don’t know how to dispose of nuclear waste? I find it hard to fathom. But anyway, this blog post is not about that statement. It’s about this one: it’s too expensive.

To address this critique, I use the example of the recent claims of Joe Romm on Climate Progress and Tyler Hamilton on Clean Break. They claim that the cost of new (Gen III+) nuclear power in Ontario, Canada, based on the competitive bids of companies AREVA (for the European Pressurised Reactor) and AECL (for the ACR-1000), would be “anywhere from $7,400 to $10,800 per kilowatt, depending on your appetite for risk“. Wow. That’s pretty gobsmacking. So what’s the deal? I decided to ask a few of my email contacts on nuclear matters, some of whom also write blogs.

Dan Yurman from Idaho Samizdat: Nuke Notes directed me to his superb post on this topic. You need to read this. Some excepts from Dan’s analysis:

Observers of the political turmoil now underway in Ontario over the media reports that AECL bid $26 billion to build two new ACR1000 reactors (2,220 MW) are in good company trying to make sense of these figures.

The news media, notably the Toronto Star, had a field day with the numbers sticking provincial politicians like they were morsels on a shish-ka-bob skewer. The problem with all the fire, smoke, and spit from the grill is that the numbers are undoubtedly wrong and wrongly reported in the news media.

First, $26 billion is an aggregate number that includes two reactors, turbines, transmission and distribution infrastructure (power lines or T&D), plant infrastructure, and nuclear fuel for 60 years as well as decommissioning costs. The most important number in the whole controversy has gone largely without notice and that is the delivered cost of electricity from the plants is in the range of five cents per kilowatt hour.

… the widely reported figure of about $11,000/Kw for AECL’s reactor package is also wildly inaccurate because it wraps the entire cost of the package into a single bundle and then allocates all of them to a figure designed to inflate the cost of the electricity generated by the reactor alone.

The other bidder at Darlington agrees this kind of analysis is nuts. In a conference call with nuclear energy bloggers on July 17, a spokesman for Areva declined to provide exact numbers, but did not specifically dispute a report in the Toronto Star on July 14 which pegged the cost of two 1,650 EPR reactors at $7.8 billion. Doing the math, that comes out to just under $2,400/Kw which is a very competitive price.

For comparison purposes, AECL bid two of the new ACR1000, which is an 1,100 MW reactor, which comes to 2,200 MW. At a reported bid price of approximately $6.0 billion, the price per Kw/Hr of the reactors is $2,700/Kw or very close to the price reported in the news media for the Areva reactors.

David Walters, of Daily Kos, said this in reply:

First, it’s about 10,000,000 to 20,000,000 man hours to build a new generation III AP-1000. Charles Barton got that number from some place. We figured that at the high end…a total wage and benefit package of $60/hr average would bring you to $600 million to $1.2 billion for all overnight labor costs. Obviously everywhere else, even Europe, it’s probably a lot less.

So the question is raised “where is the rest of the money going”? Obviously profit for one, but seriously, if they get 15% over their costs, I’d be surprised. The other obvious price is the cost of the reactor components from pumps, to pipes to stainless steel, to conrete, to the turbine to the generator to the controls and meters and so on.

What do we know for sure? We know, for sure, that the costs of the material for an AP-1000 is is less than 1.4 billion dollars. How do we know that? Because this is what the Chinese are saying it is going to cost to built their AP-1000s, take or add $100 million or so. This means that Westinghouse is charging, at most, about $1000/kW installed components. Interesting, yes? We also can assume that Westinghouse is charging the Chinese the same as they are the dozen or more new builds submitted to the US NRC.

Obviously things you note from grid access, grid upgrades, balance of plant, licenses etc etc all boot up the prices a hundred or more million dollars but not to the $10,000/kw Romm weaved from whole cloth in his hit piece. Which is not, btw, quoted extensively as “a study” on nuclear power plant costs!

I made the point your blog just now that what Romm, et al, ad nauseum, are afraid of is what is going on in Asia, specifically China and to a lesser extent S, Korea. If these countries do bring in their plants at the $1,400kW installed, Romm, et al, ad discustous, will have a hard time pitching their anti-nuclear vodoo economics to an educated public and policy makers.

Lastly, your point was excellent. France is now signing papers, as Dan has reported, for more and more EPRs. I’m not a big fan of the EPR which I won’t discuss here, but…you point that their costs WILL come down with expertise in production.

Early this year I met with a EDF nuclear physicist, far-left member of their union at his nuclear plant in southern France and ask him about this. He said at Flamesville, they are already developing new concrete techniques (some, believe or not ,dervived from French wine barrel making techniques!) to speed up and help in building the plant. They expect, still, that the plant will be over budget but not nearly as much as the Finnish version. People at EDF, both in management and in the unions are very optimistic.

The Aug. 4th blog entry from The Next Big future offers a detailed analysis of Westhinghouse’s modular construction design, and their eventual 36 month construction target. It goes toward smaller reactors in method, as well and has profound implication as perhaps half of all reactors in the next 11 years are going to be AP-1000s. The Next Big Future covers this extremely well.

Rod Adams, of Atomic Insights, said something similar:

The problem I have with the “bottom line” number provided by AECL is that determining the cost of any project is a complex, iterative effort that can only be done with a spreadsheet model that has a variety of stated assumptions and inputs. The model can then be used by the negotiation participants to run “what if” scenarios and reach agreement acceptable to all involved parties. Allocation of risk, schedule terms, charges for changes, loan terms, ROI, and many other factors come into play. I have been involved in many project negotiations and planning processes ranging in size from a few thousand to tens of billions; the bigger the project the larger the error bars and risk allowances.

The included options will also have a huge impact on the total cost – imagine how scary one could make the predicted cost of a new football stadium if the bid had to include the cost of the road network, the transmission lines to supply the stadium with electricity, the trash trucks to haul away the residue from each game, the concrete plant to provide the construction materials, the college athletic infrastructure to train the players, and the cars and buses required to deliver fans and players.

Here are some particulars about the terms and conditions under which AECL’s published bid of $26 Billion was computed. Excerpts:

The Ontario bid process required the risk of cost overruns to be assumed by the would-be vendors, which initially included French-based Areva Group, and Westinghouse Electric Co.

The Harper government told AECL that its bid must provide a commercial rate of return, and that its price must recover all costs, rather than spread them out over future sales. The federal officials who prepared the briefing documents predicted Ontario would be knocking on Ottawa’s door for a better deal.

Provincial Energy Minister Jack Keir said Monday he remains optimistic the project will go ahead, and that Ottawa will help shoulder some of the risk. But he drew a line in the sand: “Not on your life is the province going to take on the risk,” he said.

Under those conditions, the bid is not surprising, but I think that negotiations are going to continue, especially since there are other interested parties. The price per unit for any complex project is absolutely dependent on the number of units that will be built without changing the design. In the meantime, there is plenty of fodder for speculation and for dire predictions of gloom, doom and cost overruns. The only real answer would be successfully completing a project with full transparency, something that I doubt is going to happen soon enough to quiet the opposition.

Of course, there is also the option of taking smaller bites at a time.

As an aside, when people throw cost overrun accusations at you regarding Finland’s first EPR, you might be interested in having some useful numbers on off-shore wind power to throw back. As the crow flies, this project is pretty close to Finland. It is also being run by a similarly qualified professional organization.

Rod’s post on offshore wind, which he refers to in the last sentence, points out that it is 1 year late and $85 million over budget. He says:

The project was delayed for more than a year partially due to bad weather (on second thought, that might not be just a first of a kind issue for off-shore wind farms). The projected total cost will be $357 million, approximately $85 million more than the initial estimate. Considering the size of the array and its capacity factor, that cost overrun is comparable to exceeding the budget for a 1600 MWe Areva EPR by more than $5 Billion dollars.

Assuming a generous 40% CF for the 60 MWe peak capacity wind farm, the total cost of $357 million is equivalent to paying $19 Billion for a single 1600 MWe nuclear power plant that can operate at a modest average capacity factor – for a nuclear plant – of 80%.

However, please do not cry for the investors in the project. Their profitability will be assured by the rules of Germany’s feed in tariff laws which will guarantee that the project owners will receive $0.18 per kilowatt hour for their generated power.

The obviously implication is that first-of-a-kind cost overruns are not peculiar to, or indeed amplified in, nuclear power. New renewables are just as vulnerable, perhaps more so. Charles Barton has also written on the topic of recent nuclear power costs on Nuclear Green, here.

So please Mark Parnell, Joe Romm, and various commenters on this blog — spare me the hype and spin that nuclear power is somehow the expensive low-carbon energy option. It isn’t. Get over it.