Formal–statistical methods - Probabilistic Safety Assessment for Optimum Nuclear Power Plant Life Management (PLiM) - page 65

Engineering - It relies on the statistics of breakdowns and accidents,
on probabilistic safety assessment (PSA): construction and calculation of
the so-called 'event trees' and 'fault trees' - a process based on oriented
graphs. The former are used for predicting the state to which one or another
technology failure may develop, and the fault trees, on the contrary, help
to trace all the causes that can cause some adverse effects. When the trees
are constructed, the probability of each scenario (each branch) is predicted,
and the overall probability of a failure at the facility is then calculated.
Model - construction of the models of the impact of harmful factors on
humans and the environment. These models can describe both the effects
of normal operation of enterprises as well as damage from the accident to
them.
The first two approaches are based on calculations; however, in most
cases there are insufficient reliable baseline data for such calculations. In
this case, the third approach - the expert approach should be used:
the probabilities of various events, relationship between them and the
consequences of a failure is determined not by calculation but by a survey
of experienced experts.
There are many uncertainties associated with risk assessment. Uncertainty
analysis is an essential part of risk assessment. Typically, the main sources
of uncertainty are information on equipment reliability and human errors as
well as the assumptions of the models of emergency processes. To correctly
interpret the magnitude of risk, we should understand the uncertainty and
their causes. Uncertainty analysis is the translation of uncertainty of initial
parameters and suggestions for assessing risks, the uncertainty of the results.
Sources of uncertainty should be identified. The main parameters to
which the analysis is sensitive must be included in the results.
It is important to emphasise that the complex and expensive calculations
often give risk values with very low accuracy. The accuracy of calculations
of individual risk for complex technical systems is not higher than one order,
even in the presence of all necessary information. The full quantitative risk
assessment is more useful for comparing different options (e.g., placement
of equipment) than for conclusions about the safety of the facility. Overseas
experience shows that the largest volume of safety recommendations
is produced using high-quality (engineering) methods of risk analysis
which enable us to achieve the main objectives of risk analysis using a
smaller amount of information and labour costs. However, quantitative risk
assessment techniques are always very helpful and in some situations they
are the only acceptable techniques, in particular, when comparing risks of
different nature or examining high-risk, complex and expensive systems.
The quantitative risk indicator is the numerical value of the likelihood
of an undesirable event or (and) the results of undesirable consequences
(damage).
The risk can be quantitatively defined as the frequency (dimension - the
inverse time) of occurrence of hazard.

￿ ￿ ￿ ￿ ￿