Thursday, July 23, 2009

A Concluding Unscientific Preface to the Keys, Revised

I began my thinking about reactor manufacture some time ago. There are enormous drawbacks to on site manufacture. First it is a crafts-labor intensive form of manufacture. Craftsmen swarm over the site performing all sorts of tasks. We have pipe fitters installing pipes, welders welding, electricians installing miles of wiring, carpenters building forms. Rebar being installed, forms placed, and concrete poured. Workers are constantly shifting within job sites; they have no assigned station at which tools can be kept handy. Workers require frequent assignment and direction in performing their tasks. Supervisors must constantly move back and forth from their superiors to their subordinates relaying messages and assignments, and monitoring work for progress.

Workers are assigned tasks for which they are expected to use their general repository of work skills, but the tasks may not be performed up to speck, and the supervisor does not have time to notice it, because he or she is too busy moving back and forth, between project managers relaying assignments and information.

Lets look at what can happen. Take the Olkiluoto-3 reactor construction project in Finland. The story is well known: “Flawed welds for the reactor's steel liner, unusable coolant pipes and suspect concrete in the foundation.” The massive construction project is already more than two years late and 25% over budget. An investigation revealed ”The power plant vendor has selected subcontractors with no prior experience in nuclear power plant construction to implement the project. These subcontractors have not received sufficient guidance and supervision to ensure smooth progress of their work.”

This was hardly surprising. Any construction project of that size is an opportunity for human error to flower. Investigators reported:
“[T]he specific quality requirements for constructing a nuclear power plant were not clearly brought up when inviting tenders on concrete supply. The process of designing the concrete composition, concrete manufacturing and the respective quality control measures involved problems in accountability and communication because there were many subcontractors. There was no manager on site with an overall responsibility for the preparation of the base slab and an authority to issue binding orders to all parties. Problems had arisen during minor concreting performed on site, but they did not result in measures for ensuring the smooth implementation of the main concreting. The approved concrete composition was altered during concrete mixing. Deviations in the concrete composition and in concrete pouring were not addressed openly and without delay.”
The root cause then was that no reactor had been built in Finland since the 1980’s, hence no one had experience building a big reactor, and no one really knew what he or she was doing. Furthermore the project clearly got out of control of its managers. The history of reactor construction in the United States and Canada has been one of a consistent pattern of construction taking longer than scheduled, and costing more than planned. Consider now that in the United States, with the exception of TVA and its contractors, no electrical utility has experience with reactor construction in a generation, and the prior performance of utilities was, to say the least, not encouraging. Since the NRC licensing will not be completed for the first reactor project before 2012, the earliest any of the current proposed reactors construction project can be expected to come on line is 2015. Hence we might expect to begin reaping a practical benefit from reactor construction experience about 2015. Hopefully, short, or at least shorter construction periods will teach begin to teach everybody involved by actually doing the work of constructing reactors and learning something from the experience.

Even if accomplished perfectly on site. construction of reactors is expensive and probably always will be. Reactor construction methods are in some cases use technologies that were known to the Romans. Such construction relies on skilled human power, rather than mechanized manufacturing approaches. Thus in an era of increasing construction costs, it become increasingly expensive to build reactors as construction rather than manufacturing projects.

Westinghouse estimates that the construction of AP-1000 reactors require between 16 and 20 million man hours of labor, most of which will be performed by skilled laborers. The most skilled workers for a construction project of this scale would be the engineers and managers who are organizing the work schedules, and assuring a constant flow of up to spects materialsonto the job site. Thus a 1 GW reactor is a massive construction project that requires enormous amount of skill at every level of the project from the utility management to the welders and pipe fitters. In order for a construction plan to work both those who make the plan and those who carry it out must have a precise understanding of what they are doing. Even with perfect planning and organization, the skill level required of workers, the amount of labor involved, and the relatively low degree of mechanization will insure that reactor construction labor costs will remain high.

John Geesman, a former Commissioner and Executive Director of the California Energy Commission, recently wrote a series of essays on the Congressional Budget Office recently released study of nuclear power. While the CBO's study appears flawed, nuclear advocates ought to pay attention to its conclusion, and Geesman's essays (found here, here , here, and here.)

In fact the CBO report reflected a great deal of uncertainty about reactor construction costs. The report noted a range of projected reactor costs from 1.2 to $4.8 million per installed megawatt. and stated: “The breadth of that range reflects the uncertainty associated with the cost of building new nuclear plants in the United States, and is wide enough to capture plausible further increases in construction costs.” Thus inflation was factored in. Critics of nuclear power have ignored the inclusion of inflation in CBO reactor cost estimates, never the less the CBLO report highlights the need to take major steps to control the cost of nuclear power.

It is my view then that the custom, on site manufacture of large reactors by teams of craftsmen under divided and multilevel supervision is an expensive manufacturing system that invites human error. I have not even considered her the possibility for corruption entailed in this construction system.

One of the major arguments against nuclear power as a major replacement of fossil fuel generated electricity, is simply that the difficulties of manufacture, the slow pace, the labor requirements and the cost all conspire against the construction of enough reactors to quickly enough to effectively replace fossil fuel power plants by 2050. It might beaded, that the prospects for windmills replacing fossil fuel generators, are dismal, and the prospect of PV and CSP replacing fossil fuels rely on the same sort of massive construction projects that raises questions about the future of nuclear power.

The question then if I start with a blank piece of paper, or in my case a blank computer screen with the word “Word” in the upper left menu bar, and a self imposed mandate to replace fossil fuel power with post-carbon generated electricity by 2050, how should I proceed?

I felt that several goals were important:
1. Decreasing reactor cost
2. Rapid manufacture
3. The production of as many reactors as were required to replace the use of fossil fuels as energy sources for our society.
4. The highest possible level of reactor safety
5. Resolution of the problem of nuclear waste.

Dr. David L. Goodstein the vice provost, and a professor of physics and applied physics at the California Institute of Technology, published a book titled “Out of Gas: The End of the Age of Oil” in 2004. Goodstein argued that 10,000 (1 GWe) nuclear power plants to replace all the energy we are currently getting from fossil fuels for all purposes, and expresses a justifiable skepticism about our society’s ability to accomplish this goal, at least justified given the way we do business now.

Environmentalist Joe Romm has also recently argued that there would be serious problems replacing fossil fuel power plants with nuclear generated electricity. Romm argues that problems like high reactor construction costs, manufacturing bottlenecks, the slow pace of reactor manufacture, the limited supply of key reactor parts, and alleged water shortages, an alleged shortage or Uranium, and difficulties associated with the supposed problem of nuclear waste, are going to limit the futuire of the nuclear industry.

It is my view that a rethinking nuclear power is required to answer Goodstein's and to a certain extent Romm's arguments. The rethinking assumed that we would need 10,000 GWs worth of electrical generating reactors by 2050, and that goal is achievable is we take certain well defined steps to attain it. In order to accomplish the goal we must change the way we do business in order to accommodate our new circumstances. We must move beyond business as usual if we are going to replace fossil fuels as sources of energy. Changing the way reactors are built is both possible and necessary.

Update July 23, 2009: Some concluding observations

Using Data from the DoE's Energy Information Administration, the Institute for Energy Research has concluded that the 2016 levelized cost of renewable electrical generation sources will carry higher levelized cost than conventional nuclear generated electricity. Thus if the cost of nuclear generated electricity is unacceptably high, the cost of renewable generated electricity is even more unacceptably expensive.

I have chosen to not disagree with the critics of Nuclear power about the various flaws in the current system of nuclear generation of electricity. I do disagree with the contention that the flaws in current nuclear generation technology are fatal, or justify a rejection to current nuclear technology. Rather the flaws of current nuclear technology justify the development of an improved nuclear technology, one whould overcome current flaws. I have pointed to LFTR technology, a radical nuclear technology that might be nothing less than the silver bullit, which energy experts allege does not exist. In the "Keys" series i pointed to approaches to LFTR implementation that would lead to a levelized cost for LFTRs that is far lower than the levelized cost of new conventional nuclear plants. The actual cost of LFTRs will not be not be known for some time to come, but in my "Keys" series, i have demonstrated that there are good reasons for the hope that the levelized cost of LFTR electrical generating facilities will be far lower than the levelized cost of conventional nuclear power pants. Further research will help to establish the accuracy of this picture.


David Walters said...

Profoundly troubling...but also a positive challange for all of us.


jimcrea said...

Yes, the problem is human error. There has been much effort devoted to methods and procedures that minimize or eliminate it. Two disciplines that come to mine are aerospace and information processing. They owe their very existence to fault free operation. The reactor construction endeavor can learn game changing lessons from these two fields.

As an observer who has scene how the miracle has happened, I will try to paint a picture of how a successful product in these two fields evolves.

A designer first plans the project in as much detail as possible in layers as he goes. When he is absolutely sure that the design is perfect, preliminary development begins. The programmer develops his modules until he is certain that they are perfect. Then preliminary tests begin.

Low and behold, a thousand errors are found. He codes and tests again; only 750 errors now. Redevelopment and retest: many times, he builds on what has gone before until the package is fault free. Then integration with the other packages; more errors are found.

Throw out the bad, keep the good. Build a baseline of error free modules. Test the implementation of all the modules in stages. Always test from the beginning, take nothing for granted, make no assumption, don’t trust anything, build … build … build… test, … test, ….test.

Always follow the same procedures that have been successful and have proven to be error free. Any deviation can let in that devil of human error. No laziness, like a robot or a computer, follow the steps, and test. You can’t do too many tests.

When you are absolutely sure that the package is perfect, you give it to the user. Errors come back from him, … so many errors, … and you redesign, always improving the baseline, never restarting from zero, constant improvement, until after many years, the project works to a bearable and useable level.

But the feed back loop never, ever stops, until the project is trusted without question. But you know there are still 100 bugs in there somewhere, and you hope that they stay asleep for a little while longer.

This methodology could be enforced in a manufacturing plant, with the ultimate in quality control. This is why I think small modular reactors are such a good idea. Given enough blood, sweat, and tears, a production line can produce reactors that can work perfectly every time, well almost.


Blog Archive

Some neat videos

Nuclear Advocacy Webring
Ring Owner: Nuclear is Our Future Site: Nuclear is Our Future
Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet
Get Your Free Web Ring
Dr. Joe Bonometti speaking on thorium/LFTR technology at Georgia Tech David LeBlanc on LFTR/MSR technology Robert Hargraves on AIM High