Re: Re[2]: ETOPS

From:         cp@panix.com (Charles Platt)
Organization: PANIX Public Access Internet and UNIX, NYC
Date:         21 Oct 96 13:31:39 
References:   1 2 3 4
Next article
View raw article
  or MIME structure

Woodhams (mdw@ccu1.auckland.ac.nz) wrote:
> The basic argument is that a tri-jet that looses two engines is just
> as dead as a twin that looses two engines, but assuming equal engine
> reliability and that engine losses are independent, loosing two
> engines on a tri-jet is three times as likely as loosing two engines
> on a twin. (There are three ways to lose two engines on a tri-jet:
> loose 1 and 2, 1 and 3 or 2 and 3. There is only one way to loose two
> engines on a twin.)

Since there are well documented cases of 3-engine aircraft limping home on
1 engine (a friend of mine was on such a flight!) I can't accept your
basic premise that a tri-jet that loses two engines is "just as dead" as
a twin in the same state.

Since a single-engine failure is the more likely scenario, I prefer to
look at it this way:

For the sake of simplicity, suppose each engine has a 9 in 10 chance of
remaining reliable over some fixed period of time. Thus the probability of
NO engine failures during the fixed period will be 9/10 x 9/10 or 81
percent if we have 2 engines running, but 9/10 x 9/10 x 9/10 or about 73
percent if we have 3 engines running. (Naturally the actual figures would
be far, far lower if we reduce the fixed time to one average flight, but
the principle remains the same.) By subtraction, we get a 27 percent risk
of one engine failing if there are three running, and a 19 percent risk if
the aircraft only has two engines. Thus the risk of single-engine failure
is almost 50 percent higher when we go from two engines to three.

Since the loss of one engine is much more serious on a two-engine
aircraft, I prefer the 50 percent extra risk on a three-engine aircraft.
Of course I realize there are numerous other factors that should be taken
into account in real-world situations--e.g. the accident record of the
airline, most obviously.