"C Knowles" wrote in message
.. .
This is hardly news- another study to confirm what we have known for
years.
The point here is what are we going to do to improve? What I find
interesting is that the "level off" they discuss is coincidental with the
arrival of two-man cockpits and increased automation. Unfortunately humans
do not do too well at monitoring gages for hours at a time; we need to
stay
interactive with the machinery. The 757 crash in Columbia was attributed
to
crew error, and in fact the crew did make multiple errors. There were
several points where they could have recovered. But the sequence started
when the airplane's FMS turned torward the wrong navaid and the crew did
not
recognise it.
No Curt, the sequence began when the validation rule: "remove all doubles"
was written for the FMS design. In fact, Columbia uses ROZO twice as a
waypoint indetifyer for navigating to the runway in question. Something
both operators should have known.
The second error occured when Honeywell sent a Service Bulletin to American
Airlines and it was allowed to set on dock for 180 days. The 180 days is
signifigant, in that the Columbians released the runway to 757 traffic in
that regulatory time period.
A third error occured when the first officer accepted the runway change and
used the "direct to" menu, in place of selecting the approach. Thereby
discovering the removal of the double ROZO.
John P. Tarver, MS/PE
"Dave Kearton" wrote in message
...
Quoting without comment, I guess they'll come later
Cheers
Dave Kearton
http://www.theadvertiser.news.com.au...55E911,00.html
By ANDREA STYLIANOU
19jul03
THE actions of aircrew may be the major factor in most aircraft
accidents,
research shows.
The crew's actions in the technological surrounds of the cockpit and the
impact of external factors such as flying conditions have been studied
by
University of South Australia engineers.
"On examining accident statistics, it can be seen that in about 70 per
cent
of recent aircraft accidents pilot error has been cited as the major
contributing factor," said Professor Stephen Cook, the director of the
university's Systems Engineering and Evaluation Centre.
"Aircraft failures make a low contribution to the overall accident
rate,"
Professor Cook said.
"As a consequence, the significant improvements in air transport safety
must
address this area," he said.
The safety of large passenger aircraft was at a high level, but fatal
accidents continued to occur around the world at a rate of almost one a
week.
"Data analysed shows that the rate of fatal accidents per flying hour
has
decreased from the beginning of aviation up to around the 1980s,"
Professor
Cook said.
"The number of fatal accidents has since levelled out and has been
almost
constant for more than two decades," he said.
Practical studies of air crash data are hampered by the extended time
frames
needed to make proper assessments from the overall low rate of
accidents.
The university engineers overcome this by studying computer models. "The
use
of modelling as a tool for improving safety levels is one way of
improving
aircraft safety," Professor Cook said.
"Computer-based modelling is essential and has been used in engineering
studies for decades.
"We need to pay more attention to the interaction between humans and the
systems they control.
"There is some resistance to the idea of modelling the behaviour of
people,
such as pilots, who undertake complex functions, but given useful
information we can successfully model whole populations."
Although in the early stages of development, the computer model used in
the
study has shown an initial capability to produce believable data on the
factors affecting pilot behaviour.
"Aviation is seen as the benchmark in safety performance. It is studied
and
emulated by others from hospitals to railways," Professor Cook said.
"While aviation has a lot to teach, it also has a lot to learn if it is
to
break through to the next level of safety."
The results of the study will be presented at a Brisbane aerospace
conference later this month.