1.1 In the case of traditional craft, it has been
possible to specify certain aspects of design or construction in some
level of detail, in a way which was consistent with some level of
risk which had over the years been intuitively accepted without having
to be defined.
1.2 With the development of large high-speed craft,
this required experience has not been widely available. However, with
the now broad acceptance of the probabilistic approach to safety assessments
within industry as a whole, it is proposed that an analysis of failure
performance may be used to assist in the assessment of the safety
of operation of high-speed craft.
1.3 A practical, realistic and documented assessment
of the failure characteristics of the craft and its component systems
shall be undertaken with the aim of defining and studying the important
failure conditions that may exist.
1.4 This annex describes a failure mode and effects
analysis (FMEA) and gives guidance as to how it may be applied by:
-
.1 explaining basic principles;
-
.2 providing the procedural steps necessary to
perform an analysis;
-
.3 identifying appropriate terms, assumptions,
measures and failure modes; and
-
.4 providing examples of the necessary worksheets.
1.5 FMEA for high-speed craft is based on a single-failure
concept under which each system at various levels of a system's functional
hierarchy is assumed to fail by one probable cause at a time. The
effects of the postulated failure are analysed and classified according
to their severity. Such effects may include secondary failures (or
multiple failures) at other level(s). Any failure mode which may cause
a catastrophic effect to the craft shall be guarded against by system
or equipment redundancy unless the probability of such failure is
extremely improbable (refer to section
13). For failure modes causing hazardous effects, corrective
measures may be accepted in lieu. A test programme shall be drawn
to confirm the conclusions of FMEA.
1.6 Whilst FMEA is suggested as one of the most
flexible analysis techniques, it is accepted that there are other
methods which may be used and which in certain circumstances may offer
an equally comprehensive insight into particular failure characteristics.