Tuesday, April 17, 2007

TXU, Mitsubishi and the Benefits of Standard Designs

Going through my clips yesterday, I came across the following editorial from the Boston Herald concerning TXU and its plan to build new nuclear plants in lieu of coal-fired electric generation. Though I don't normally do this, I'm going to print the entire editorial:

TXU Corp., formerly Texas Utilities, says it wants to build nuclear plants instead of the 11 large coal-fired plants - so scorned by environmentalists - that it announced earlier this year. It’s another welcome sign of the return to favor of nuclear power, which generates no greenhouse gases said to warm the earth.

The problem is that TXU, one of more than a dozen utilities considering nuclear projects, wants a new plant design. But one of the reasons nuclear fell from favor was that U.S. utilities vastly overestimated their ability to cope with the new designs they kept ordering throughout the 1960s and 1970s. Of the 105 nuclear plants operating in 1986 (there are 102 today), there were 62 designs.

One reason for the success of nuclear power in France is that nearly all the 59 reactors there share a common design.

U.S. reactors steadily got bigger as companies sought to spread the costs over more megawatts of output. The reactors became dauntingly complicated faster than understanding of them could spread.

The two major U.S. reactor suppliers, General Electric Co. and Westinghouse Corp. (now controlled by Toshiba Corp. of Japan), spent more than a decade coming up with simpler, safer designs and winning approval of them from the Nuclear Regulatory Commission. That means they may be used as an off-the-shelf item without further approval, and most nuclear-planning utilities have chosen one company or the other.

TXU had said earlier it expected to build a Westinghouse design of 1,150 megawatts. Now it says it wants two to five 1,700-megawatt plants from Japan’s Mitsubishi Corp. - for the old familiar reason that it can spread costs over more output.

This could risk a rerun of all the problems of the 1960s, 1970s and 1980s, from vast construction cost overruns to mysterious plant disturbances and shutdowns. The Mitsubishi design has not been studied by the Nuclear Regulatory Commission and approval will take a long time.

Going with Mitsubishi could risk a failure that would seriously set back the nuclear revival.
Now, there is a kernel of truth in this piece. Non-standardization was certainly a problem with the first generation of nuclear reactors -- one that was substantially mitigated by an industry consolidation that concentrated plants into fleets with similar designs.

But to suggest that TXU's plan could possibly derail a revival for the American nuclear industry seemed a little far-fetched. So I sent a note to Adrian Heymer, our new plants guru, asking what his reaction to the editorial was. Here's what he wrote in reply:
America is the land of free enterprise. There are specific laws relating to restriction of trade and business, such as the Anti-Trust Act.

At present there are essentially 103 different designs and two main owners' groups. There are now 16 companies preparing license applications for as may as 30 reactors and only five designs. The commitment to standardization within a design-centered group is high, between 70 and 90 percent. The differences are linked to different site specific features.

While the Mitsubishi APWR has advanced design features, it is also very similar to plants under construction in Japan. As a result, TXU will benefit from standard construction techniques and practices that have been refined in overseas projects. The same is true for GE's ABWR which is the selected design for NRG, which is preparing a license application for two units near Houston, Texas.

TXU has evaluated more than one design, as have the other fifteen companies. TXU made their choice based on their own set of circumstances, which included the benefits of standardization within their own company and within the industry, both here and overseas. What is encouraging is that Mitsubishi is willing to make substantial investment to obtain a US design certification for the APWR and market that design in the US. This emphasizes the strength of interest in the potential growth of new nuclear generating stations in the US.
Thanks to Adrian for the assist.


KenG said...

The editorial repeats a misleading generalization relative to the standardization of plants. Although you can say that the 103 US plants have 62 different designs, this is true only in the details. There are less than 10 basic designs which is why Standard Tech Specs and Owners Groups have been so successful.

Although I'm not expert on the French designs, I doubt they really have 59 units of a standard design. It seems more likely there are evolutionary design series with the 59 units.

As long as there are enough plants of a given design series to develop economies of scale in fabrication and to allow formation of an effective owners industry group, new plant programs will be effective.

sefarkas said...

The differences in plant designs mentioned in the editorial happened over 15 years or so as the plant sizes were raised from the 500MWe range to 900MWe to the 1100MWe range. Four companies, GE, Westinghouse, Combustion Engineering, and B&W took various approaches to achieving those three power levels. When you group the plants by electric power production size and reactor vendor, they are remarkably similar. The French use the Westinghouse designs for the 900 and 1100 series plants.

A similar situation is developing today with the GE ESBWR, the WEC AP1000, the AREVA (B&W) EPR, the Mitsubishi APWR. With the legal structure of the US economy, it would be unusual for these four companies to have the same design. Only in a socialist system could the government limit the breadth of commercial offerings. The licensing simplification today is that each of those designs will be set before construction begins, thus checking the ability of opponents to challenge the "safety" of the plant once it is already built in conformance with the certified design.

Starvid, Sweden said...

The French have essentially the same design, but scaled up and improved as time passed.

30 reactors are 900 MW, 24 are 1300 MW and 4 are 1450 MW.

And the EPR in Flamanville is supposed to be 1650 MW.