probability of multiple engine failure
In response to The Real Slim Shady's post, a little excursion into probability theory. Assume the risk of one engine failure is x (say 1 basis point or 1 in 10,000 flights). If the events of engine number 1 failing *and* engine no. 2 failing are independent, then the probability of 2 engine failures at the same time is 1bp x 1bp which is 10^(-8) or 1 in 100 million.
The assumptions of both events being independent is incorrect IMHO for several reasons:
1. Birdstrikes: self explanatory really. Flock of birds has high likelihood of damaging both engines, confirmed by NY and Ryanair Ciampino incident.
2. Fuel system: BA038 comes to mind. Let there be a central systemic failure in the fuel supply and the probability of both engines being affected is again quite high.
3. Pilot error / human performance: If one thing goes wrong and you have systems failing, the probability of further errors down the line increases dramatically. (Stress, shock, unknown or rarely practiced situation, sudden loss of instruments etc.) Example here is pilots shutting down the good engine after engine fire and letting the other one burn. They say it has happened...
4. Volcanic ash - another sweet piece of anecdotal evidence (and excellent airmanship).
To make a long story short, you have to use *conditional* probability, i.e. the probability of you losing engine no.2 *given* that you have lost engine no. 1. I would argue that when using the concept of conditional probability, the chance of having multiple engine failure is suddenly much higher in quantitative terms.
For example, I would estimate that given you have lost one engine due to a flock of birds, the probability of losing the other engine as well is somewhere between 10 and 50%. (Which would be astronomical dimensions greater than the chance of having a birdstrike in the first place.)
Three events of multiple engine failures in less than two years (BA, Ryanair and US Airways) - those are only the ones I am aware of - *is* statistically significant. Especially given that all of them were in critical phases of flight and rather close to disaster...
I would suggest every airline and aviation authority and sim instructor should include multiple engine failure (especially at low altitude, after t/o and before landing) into their standard training program.
I think the different air forces around the world have done so for decades already...