No assessment criteria
The “Stress test“ specifications require descriptions of the plants’ properties and a justification of the chosen safety provisions. Requirements on the quality and the comprehensiveness of those descriptions are not defined. No assessment criteria or bench-marks are defined as a yard stick to assess whether reported provisions of the defence in depth reach a minimum quality.
In the case of earthquake, for example, the “Stress test“ specifications require a description of the up to date level of the assumed peak ground acceleration at the site and the justification of the chosen data.22 The procedure to determine the peak ground acceleration contains many assumptions in a complex field of investigation. There is no common opinion or commonly accepted rule to determine the basic data for the assessment. There is no common opinion or commonly accepted code to gain the results. The validity of data depends on many factors and the method of assessment depends on many sensitive benchmarks where key predeterminations can be hidden which influence the result significantly. One of many questions for example is what confidence interval the earthquake determination should rely on: it makes a difference whether one can trust the estimated probability of an earthquake of certain strength fully or whether it must be reckoned, that within, for example, a range of 50% the earthquake may be much stronger. The same complexity of finding out the strength and the probability of an earthquake is typical for most of the other assessments as for example the adequateness of the “provisions to protect the plant” against earthquake23 or flooding. So far as the “Specifications” rely on the licensed design24 and its safety case , it relies in many cases on out-dated criteria and methods that may have worked 30 years ago and that differ from plant to plant and from older to younger plant generations.
This insufficiency is characteristic for the whole set of the “Specifications”. Without clear and precise rules for assessment it is possible to generate arbitrary results. This is a major reason why the results of the country reports on the assessment of the 143 European plants will not be comparable.
22 Fn. 4, page 7 (downloadable document)
23 To know the loads that must be carried by the structures of the reactor building and its components at each relevant point and to know the load limits that can be carried it needs complex computer based methods of calculation and validated input data. It must take into account the different operation conditions and the interdependency of possible reactions.
24 Fn. 4, page 7 (downloadable document)