FOAKE costs, learning curves, carbon taxes and nuclear plant economics.

[NNadir]

The University of Chicago has put a very detailed study of the economics of new nuclear plants in the United States.

Nuclear power plants that are fully depreciated, which constitutes most of the 103 nuclear power plants now running in the United States, are very cheap to operate. The Catawba nuclear power plant, in South Carolina, which to be fair is the cheapest nuclear power plant to operate had and operating and maintenance cost of 0.011 dollars/kw-hr, including plant costs and fuel costs in 2002.

http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/08-13-2003/0002000237&EDATE=

Using my new conversion factor for gasoline equivalent price, as described in another thread, this is the equivalent of producing gasoline at a price of 0.41/gallon or having oil available at about $17/barrel. Note this is the fully loaded cost, not the price of fuel, as the link I will produce below will show.

In the future, the cost of nuclear fuel is expected to rise significantly. In the worst case, the fully enriched and fully assembled cost of nuclear fuel is $5.53/MW-hr which has the gasoline-equivalent price of $0.41/gallon, the current operating cost at Catawba. The best case has a gasoline price equivalent of $0.13/gallon for fully assembled fuel rods ready for use.

The University of Chicago study, which is the product of economists, and not physical scientists and engineers, although such scientists were clearly consulted, touches on some economic issues that extend beyond the obvious fact that nuclear energy is the safest and cleanest form of continuously available scalable energy.

First, although the United States generates more nuclear energy than any other country, the United States has, for all purposes, no practical experience building nuclear power plants. All of the nuclear power plants that now exist were built by an earlier generation. For better or worse, nuclear construction was stopped here in the early 1980's on the grounds that objections to nuclear power were plausible. The actual experience has proved to have been much better than what opponents argued, but in 1980, at the close of the first era of nuclear power plant construction, it certainly seemed possible that fatal nuclear accidents would occur regularly, and that spent fuel would be an intractable problem. It is experience that proved otherwise. Still, the unfortunate decision was made to halt new construction. Thus there are few highly experienced engineers who have the ready experience of nuclear power plant construction.

Thus every new nuclear power plant will experience what are called "FOAKE" costs, "First of a Kind Engineering" costs. This will raise the cost of new nuclear construction. From the summary study:

First-of-a-kind engineering (FOAKE) costs for new nuclear designs could increase capital costs by 35 percent, adversely affecting nuclear energy’s competitiveness.

I will touch on the issue of how much FOAKE costs will effect competitiveness later.

The study also posits the existence of a "learning by doing" parameter which is a quantitative estimate of the improvement in plant economics. This parameter has an interesting structure. It estimates the percentage by which the experience of building nuclear power plants will reduce the cost of subsequent plants if the number of plants built in the second round is doubled.

Here is an example of what they mean:

According to the world nuclear website table of world nuclear reactors, http://www.world-nuclear.org/info/reactors.htm, 13 new nuclear reactors have been proposed in the United States, i.e. are undergoing evaluation. Let us assume for a moment that all 13 reactors are built, for a total capacity of 17000 MWe. Let us also assume that all 13 reactors are built simultaneously and moreover that each of the four types of reactors contemplated for construction here, the Toshiba Advanced Boiling Water Reactor, the Advanced European Pressurized Water Reactor, the Westinghouse API-1000, and the Canadian Advanced Heavy Water CANDU type reactor are represented. Thus all of the reactors will experience FOAKE costs. Let us assume the worst case, and find that the FOAKE costs drive the reactors up to a price of $1,800/kw. Thus the cost of of the 13 new nuclear power plants will be about $31 billion dollars, or roughly half of the annual cost of the war in Afghanistan (yes we have two wars going). The average plant will thus cost about 2.4 billion dollars.

According to the study, the "learning by doing" cost reduction is scenario based. If, for instance, the construction of nuclear power plants proceeds at a breakneck pace (as it should if we intend to survive global climate change and oil depletion), and if the construction time is shortened by the elimination of unnecessary delays, and if reactor design is standardized, then the "learning by doing" savings will be 10%. So while the first series of 13 reactors would cost $30 billion dollars, the second set of 13 reactors (the doubling) will be $30 million - 10% = $27 billion dollars and the third set of nuclear reactors will about $24/billion dollars, and so on.

The capacity for the production of electricity by the use of coal - which is the primary form of continuous electrical generation in the United States as well as the most egregious environmental source of electricity is, in winter, about 315,000 MWe. Thus we can estimate the cost in the "learning by doing" cost scenario of eliminating coal fired power generation in the United States by building nuclear power plants (which is in fact, the only way to eliminate coal since no renewable strategy is actually continuously available on scale.) Since the 13 plants will have 17,000 MWe, we see that 18 construction cycles (rounding up) will be required, involving 240 new nuclear power plants. The first cycle will cost $30 billion dollars, and carrying through the math, we see the fifth cycle will cost about $20 billion dollars. (Here a minimum cost, about $1200/kw is reached.) The study compares the cost of nuclear power plant building with the costs of coal and natural gas plants on an internal cost basis alone, that is, as usual, ignoring the environmental costs.

Thus we can estimate the cost of eliminating coal through nuclear power plant building. Roughly the cost is about $400 billion dollars, less than the cost of 4 years of war.

Of course the actual costs of power generated in a power plant varies from factors that are independent of capital costs (the bulk of the nuclear costs) and operating costs like salaries and materials including fuel. One important factor is the capacity loading, i.e. the percentage of time that the plant can be considered to operate at its name plate capacity. Another factor is how long the plant can operate before it is shut down. If you have to invest, 2.3 billion dollars in a plant and it operates for 10 years, it's economics are very different than if you operate it for 60 years.

Similarly if you operate it flat out at maximum capacity and sell everything you generate, you will generate more revenue than if you operate the plant 20% of the time. Of course, in the electricity industry, one cannot always sell everything one generates. There are peaks in electrical demand. This is why some plants, like gas plants, are designed to operate for only 20% of the time, like solar plants typically do, although in the gas case, the power is available on demand, where in the solar case, the demand of the user has no bearing whatsoever on the availability of the power.

Nuclear power plants are designed to be flat out, constant demand, fully loaded devices, fitting the niche now occupied by coal.

The study breaks down the nuclear costs according to the effects of three parameters which is shown in table 7 in the summary report. The first shows construction time at 5 years, and construction time of 7 years. (Toshiba builds Advanced Boiling Water Reactors in Japan in a little over three years, but they're Japanese.) Next it considers three capacity load cases, 85%, 90% and 95%. Most modern nuclear power plants operate at better than 90%. Then it offers the cost of capital construction as $1800/kw (FOAKE intensive) to $1200/kw (No FOAKE costs involved.) Finally it considers the plants to operate at either 40 years or 60 years. Almost all of the plants in the United States were originally licensed for 40 years. Some less successful plants shut down well before reaching the licensing period, but most operating nuclear reactors are now planned to reach a 60 year lifetime. Under these conditions, the busbar cost of nuclear energy is said to be between $47/MW-hr to $70/MW-hr. This is the equivalent of gasoline priced at $1.72/gallon (low end) to $2.56/gallon. In these gasoline equivalent prices I've been using an important distinction needs to be made: It is treating the costs as if the cost of gasoline were the only cost of driving. This treats the situation as if the cost of the car were free. Thus I am treating the situation as it would be treated as if you already owned a fully paid for electric car with no maintenance costs and were charging it at your home, using the electrical outlet.

The study shows the range of costs for coal and gas fired plants is lower than the cost of nuclear power plants. The cost of coal and gas plants are said to range between a low of $33/MW-hr to a high of $45/MW-hr. However some important caveats apply. First, the study was written in 2003 when the price of natural gas was less than $4.00/MMBTU. After Katrina last year, the price reached close to $14.00/MMBTU or more than 300% higher. The study also explicitly notes that these assumptions exclude any carbon tax.

If a carbon tax is applied, the cost of gas fired plants goes to as high as $91/MW-hr and gas to as high $68/MW-hr. Thus if only one external cost, that of global climate change, is applied to these fossil fuels, they are immediately uncompetitive with nuclear power. We ignore for the moment the cost of mining of transporting these fuels, the cost of abandoned facilities, and the cost of other pollutants such as sulfur dioxides, heavy metals and radon. (Radon of course also applies to nuclear energy.)

A note on the cost of fuel: I have treated the cost of nuclear fuel for fully assembled ready to used fuel, i.e. fuel that has been mined, refined, enriched, and built into assemblies. Under these conditions the cost of fuel is as I've shown, gasoline equivalent to between $0.13/gallon and 0.41/gallon. These are of course, trivial costs. However uranium (and more importantly thorium) itself is gasoline equivalent to crude oil at less than a penny a gallon, much less, in fact.

Another interesting aspect of the analysis is the inclusion, for reference, the cases of recycled fuel and once through fuel. For the record, I favor the fully recycled option, although it raises, for the short term, the busbar cost of energy.

The summary report is found here:

http://www.anl.gov/Special_Reports/NuclEconSumAug04.pdf

A more detailed version is found here:

http://www.anl.gov/Special_Reports/NuclEconAug04.pdf