|
Tim,
Just one problem with your theory of duplicative
redundancy in airplanes, engines. The common placement of engines is
on the wings and in this configuration you are 4 times more likely to die in a
single engine out in a twin engine airplane than the only engine out in a
single engine airplane. With two engines the extra weight is very
important, which translates into extra fuel usage (also weight) as well
as asymmetrical thrust and human reaction issues on failure.
Instead of extra reliability you actually INCREASE the odds of dying in a fatal
crash by adding redundancy.
In electrical systems, redundancy does make
great sense and is becoming the long overdue norm in certified airplanes,
Diamond, Lancair, Cirrus, etc... Two plugs, two mags, and electrical/vac
redundancy has worked well in the past. I just spent 4 hours behind the
Garmin G1000 yesterday shooting 12 approaches and I will tell you that the
Chelton/Garmin/Avidyne systems are amazingly less complex than the wavy steam
gage needles, once you learn how to use them. Having a reliable approach
coupled to a good autopilot with great situational awareness at all
times is a great example of the benefits of the advances made possible by
reliable electrical system redundancy.
Stuart Seffern
Madison, WI
----- Original Message -----
Sent: Monday, May 16, 2005 9:32 PM
Subject: [LML] Airplane Construction
Philosophy
Hello
all,
Working in the computer field I was wondering why so much time in the
aircraft industry is concerned with minimizing the mean time between failure
(MTBF) on a single component. For example in the computer field we have moved
extensively to a concept called RAID (Redundant Array of Inexpensive Disks),
in the “old” days a mainframe disk drive with an average of a one MTBF every
million hours cost tens of thousands, versus a PC drive had an MTBF of a
thousand hours and sold for a few hundred. With RAID you would mirror the data
between two PC drives, and the chance then of a both failing at the same time
was back up to a million hours. Two PC drives would cost significantly less
than the equivalent mainframe, the result is that the cost of disk drives for
servers has gone down significantly. In addition, computer staff has become
used to a failure mode for disk drives resulting in reduced data loss
and better recovery procedures. In the aircraft industry we have continued to
engineer for the MTBF of a million hours, with two consequences. One,
everything is very expensive, two pilots are not used to any failures; so when
a failure occurs the pilot does not know how to effectively deal with it.
Therefore, why do we not accept a lower MTBF and have two complete avionics
systems, fly by wire controls, engines…. The point could continue to
everything except core structural elements.
Tim
|