Formal–statistical methods - Probabilistic Safety Assessment for Optimum Nuclear Power Plant Life Management (PLiM) - page 73

situations taken into account in PSA are outside the range of the failure
conditions considered in the project.
Analysis of the possible impact of ageing on modelling of failure
consequences, the formulation of success criteria for safety functions, and
erroneous actions of the personnel is given in Ref. 32.
Input parameters in the PSA model used at the frequencies of initial
events averaged out over a specific period of time (e.g. 1 year) and factors
of unavailability of equipment. In most cases, these average values are
calculated assuming that the time to failure or probability of failure 'on
demand' are random variables distributed according to the exponential and
binomial laws, respectively. The parameters of reliability, which are required
for the calculation of the average unavailability in the given period of time,
failure rate and probability of failure 'on demand', do not depend on the
age (operating time) of equipment and are constant. Similar assumptions
are used for calculating the frequency of initial events.
The use of these assumptions is explained and justified by the following
considerations:
• simple models : the use of exponential and binomial distribution laws
greatly simplifies the process of calculating risk (unavailability of systems,
the frequency of melting of the reactor core, etc.) and, in fact, of reliability
indicators;
• lack of representative statistical data : given the high reliability of
safety systems, the number of equipment failures is extremely small so
that when developing the first PSA it is not possible to perform reliability
analysis taking into account the time of operation;
• adjustment of the negative impact of ageing on the reliability of
equipment by organisational and technical measures : from an engineering
point of view, for the optimum control of ageing (the technical and
operational measures and maintenance), equipment failure rate should
remain constant for at least the design life of the plant.
In addition, PSA practice implies a periodic review of models, input
data and results of calculations of risk, taking operating experience into
account. Typically, such audit is conducted once in 10 years in the periodic
reassessment of the security unit.
Nevertheless, possible errors in the management of ageing and/or
extending the life of plants (i.e. beyond the limits originally set by the
design) may cause deterioration of equipment due to ageing and as a
consequence the above assumptions are baseless.
Thus, accounting for the effects of ageing in the PSA is primarily related
to the revision of the basic assumptions made in modelling.
For nuclear power units approaching the design life and working outside
of original design life, it is necessary at least to verify the assumptions
made in modelling failure sequences, the formulation of criteria for the
success of safety systems and constant failure rate (the probability of failure
on demand) of the components over time.

￿ ￿ ￿ ￿ ￿