Wednesday, March 16, 2011

The systematic error in the licensing process of nuclear power plants or why the meltdown is more likely than promised

In the development and approval of nuclear power plants was and is the probability of a catastrophic accident at a key size. It is measured by the so-called residual risk that must be accepted by the population.

To make general statements about the reliability of a component, in nuclear power plants only techniques are used that have been used for a long time. A nuclear power plant is at its start a technology museum.

Is then determined which properties are necessary to ensure safety. This is quite simple. Primary is that the gaseous, liquid and solid reaction products are released not so long until their radioactivity has decayed so far is that they do not affect the biosphere. This results in secondary demands, such as the control of the fission chain reaction for the fuel may never get out of control or generated heat to be dissipated reliably. It is of special weight to the after shutdown of the reactor, thus stopping the chain reaction and thus the splitting of atoms, yet still 7% of the thermal power of the nuclear power plant released. This heat comes from the radioactivity of the fission products. After one day, this performance has decreased by about one order of magnitude. The waste is then more slowly until it merges into the long-term radiation of nuclear waste.

How to go now to the necessary safety equipment? Assume we have a component who fail more or less 3 Years, this are approximative 1000 days. Such a part would be alone is not useful for a nuclear plant. But if two independent parts only one is needed, one can apply a mathematical theorem. That theorem says that the probability that both parts are broken can calculated by multiplying the probabilities of fails of the two parts. It follows then that the combination of two parts that are individually defective in the middle every 1000 days is on average one million days defective. This mean, one problem every 3,000 years. That's not enough. But if an additional independent part is added, we get 1,000,000,000 days mean time between fails. This are round about 3,000,000 years. Then, in order to carry out maintenance work on the fly, add another part, which is one of four independent components.

Now you need not only a part of type A but also other of type B. Assume if A or B fails, we get a release. The the same calculation is used. Instead of the probability of failure the provability of working well is used. If function A fails in mean every 1,000,000,000 days and function B fails every 2,000,000,000 Years, so we get one fatal error every 666,666,666 days. These calculations are for a nuclear plant, of course, much more complex because it has a lot more parts than A and B.

Now where is the error, the release of radioactivity is more common, as promised. The secret is the word "Independent". If the parts are independent in the calculation, then the parts must designed truly independent, to be built independent and maintained independent. Why? Suppose we have a maintenance technician who maintains all emergency generators. This includes, in an internal combustion engine, for example, an oil change. If the man but now fills in an oil that is unsuitable, so the engine fails after an hour, the units in case of need would turn out one by one systematically. In that regard must be maintained for the different branches of the redundant system of different technicians. But that may also not come from the same company, because the boss could make a mistake, and maintain documentation wrong. This may take effect, that the independent mechanics make the same error.

It may also not identical units are used, because a design problem would result in a specific reason for a simultaneous failure of all components. This is what happened in Chernobyl, now Cronobyl. There was a design problem in the control rods. From a control rod is expected that the retraction means less energy release and retraction means an increase in energy release. Due to the circumstances of a failed experiment in the reactor was an unusually high reactivity nuclear status. The control rods were moved out abnormally. It has reached an area where further out of the control rods pull back has a braking effect. This was caused by a cap of graphite at the tips of the control rods, which has an energy-releasing effect. Then, when the reactor had to be throttled back, was in fact given more energy-release. So much that is changing the reactor is not only good-natured his performance. The reactor was prompt critical. In terms of popular science, it is the working in the mode of an A-bomb.

Going Nuclear Harrisburg there was the same hose connections for water and compressed air. As result from an Operator error there was water in the pneumatic power network. There was a big number number of pneumatic components who are not working properly. Forsberg in Sweden, there was again a warning because the emergency treatment after an electrical problem was not even been switched on. Again the same problem. Components who are identical act in the same evil kind. In this respect designer must be take that redundant devices don't contain the same identical subcomponents.

The Japanese, it has now caught more evil. There you have six pieces of apparently identical reactors built on the same beach. These reactors have identical emergency generators.The tsunami has now put all these units out of action. The earthquake had destroyed before the regular power supply. Thus, it now looks like we in Japan, a further increase of the Chernobyl accident, the multiple super meltdown. The youngest of Germans nuclear power plants, the so called convoy plants, are identical. Fortunately, at least not in one place - but would the above-illustrated problem with the oil and the maintenance technicians could also draw on all sites, which then would be worshiping, then when a blackout happened in the electricity grid.

What does this mean for the nuclear power plants around the world are the calculations for reliability in compliance with the requirements of independence. Check it If components are not independent, so must the reliability calculations needs to be corrected. As far as I know my way around on the construction of nuclear power plants, there is extreme need for correction, for example because the control rods in nuclear power plants are usually all identical in construction. The same goes for feed water pumps and emergency generators. I have no doubt that such inspection will result in the revocation without compensation of all permits for nuclear power plants!

1 comment:

  1. Thanks for this well written article. Would be nice if you can add some inline links with your sources.

    Rene Schulte