Physico–statistical models based on the residual defectiveness of structural materials - Probabilistic Safety Assessment for Optimum Nuclear Power Plant Life Management (PLiM) - page 169

and using the same technology as the product. Artificial defects of three
types are made in the test sample:
• defects of the critical size or close to it;
• defects whose size is in the range from the size permissible in
operation to the critical size;
• defects whose size is in the range from the size permissible in
manufacture to the size acceptable for service.
The defects (discontinuities) should simulate the defects of operational
nature, i.e., defects that may develop from technological defects or nucleate
and evolve under the influence of service loads (fatigue cracks, stress
corrosion cracking, etc.). The embedded defects can also simulate the
technological defects (if necessary).
All embedded defects should be hidden from NDT operators, i.e., should
be internal (subsurface), or, if it is a surface defect, it should be located
in the area where it cannot be detected by visual examination (or have
dimensions that cannot be detected visually).
It should be noted that the defects are distributed randomly in the
sample, for example, using tables of random numbers.
The minimum allowable distance between the defects is determined
based on the conditions for the existence of single defects (if single defects
are produced) or less - for a group of cracks (the conditions of mutual
influence are known, for example, Ref. 92).
The number of defects of each type should be sufficient for statistical
processing of the results, for example, at least nine defects (if the number
of defects is smaller the results are less reliable).
Any defect can be modelled conservatively by a crack, and any crack
can be described by an ellipse with the minor a and major c semiaxes.
There are various options for distributing defects in the test sample:
• in the form of ellipses with the ratio of the axes taken from the
condition of maximum growth rate of the defect in the service
stress field;
• in the form of ellipses with an arbitrary ratio of the axes, and the
quantity characterising the size of the defect is the area of the planar
defect or the area of projection of the volume defect on the plane
of the probable development of the defect;
• in the form of ellipses, with the number of defects and the ratio of
the axes chosen using mathematical methods of experiment planning,
based on the condition of minimising the number of embedded
defects. 93
If the defects produced int the test sample do not have the shape of an
ellipse, then they are schematised by ellipses.
After making a test sample, it is inspected using the same inspection
means and methods and the operators with the same skills which will then
be applied in inspection of products, and the results are compared with the
actual defects embedded in the test sample.

￿ ￿ ￿ ￿ ￿