Background - Using Event-B for Critical Device Software Systems - page 21

2.5.2 Risk Assessment and Safety Integrity
A risk assessment is simply a careful examination of the past data related to the
hazard's analysis for the similar systems; from the reliability assessments of com-
ponents of the system being developed; and other sources. The outcome of the risk
assessment presents some kind of gradation and may be expressed in terms of what
constitutes a tolerable and intolerable risk. This outcome results help for regulat-
ing industrial risk, and to determine whether a risk is unacceptable, acceptable or
somewhere in between. Lots of factors are used for determining the risk based on
quantitative and qualitative analyses [ 8 ]. Using a risk classification of accidents ac-
cording to the frequency and severity usefully serves as a relatively simple basis for
its determination.
Assessment of a risk can decide a necessary level of safety that can be achieved
from various functions of a system. This is an issue of safety integrity, which is de-
fined as, “ Safety integrity is the likelihood of a safety-related system achieving the
required safety functions under all the stated conditions within a stated period of
time ”[ 108 ]. The system activities are contributing to the integrity may be charac-
terised by two kinds of requirements:
1. Generation of the new safety requirements of a system is resulting from the de-
sign and development.
2. Ensuring that what is being built meets the requirements that have already been
specified.
Here, the first requirement is related to the requirement analysis and hazard anal-
yses of a system. The second requirement is related to the reliability engineering
techniques, whose consideration may have to be sustained throughout the develop-
ment as the design evolves with modification to interfaces, rearrangement of compo-
nents or other kinds of changes. To apply the several techniques like FHA, HAZOP,
FMEA and FTA for the fault prediction, fault removal, fault avoidance and fault tol-
erance, and to achieve the system integrity require together with methods and design
of the system, are the main resources for measuring the system reliability [ 103 ].
A safety of a system may be simply characterised by a process of reducing risks
to appropriate effect. The main objective of a qualitative or quantitative risk assess-
ment is to establish the level of tolerability for any identified risk. If a risk falls in
between the states of 'intolerable' and 'acceptable' then any risk must be reduced to
'as low as reasonably practicable'. This is known as the ALARP principle as illus-
trated in Fig. 2.6 . The width of the triangle is proportionate to the level of risk and
thus also to the amount of resources that can be justified to reduce it. A comprehen-
sive survey of risks and safety integrity is provided in [ 8 ].
2.5.3 Safety Integrity and Assurance
Finally, there is always a question in the development of critical system, “What is
the assurance level according to the certain level of integrity of the system?”. In