Formal–statistical methods - Probabilistic Safety Assessment for Optimum Nuclear Power Plant Life Management (PLiM) - page 86

• the date of replacement and reconditioning of equipment;
• strategy for periodic testing and maintenance.
Availability and reliability of the additional data is a major obstacle for
the practical application of reliability models of ageing.
Generalised data on the reliability of nuclear plant equipment, taking
into account the effects of ageing, are extremely limited. The data set
TRIGALEX, proposed in Ref. 49 for the linear model of the failure rate is
referred to most frequently. Another example of the generalised data can
be Weibull model parameters given in Ref. 33 for various types of safety
equipment.
Direct application of the parameters of the generalised data in a specific
PSA model is impossible because of the lack of important characteristics
affecting the reliability of ageing equipment such as the types and intensity
of the load, the strategy of testing and maintenance, etc. The generalised
data can be used for sensitivity analysis or Bayesian estimates 51,52 .
Thus, the methods of ageing data analysis discussed in this chapter:
graphical analysis and non-parametric methods for testing the trend - are
easy to implement and do not require special software.
If the original data are accurate, these methods make it possible to obtain
the primary data quite quickly and provide information about important
attributes such as the homogeneity of the sample, the presence of running-in
failure, the level of uncertainty and importance of the estimates.
However, the main factor limiting the use of all these proposed models
is the lack of representative and accessible baseline data. It is recommended
to choose the simplest two-parameter reliability models and perform detailed
analysis when calculating the parameters of the model and extrapolating
the results. This approach is applicable mainly to the active elements of
safety of the nuclear power plant.
￿ ￿ ￿ ￿ ￿