JimL:"Just to correct the basis of the statistics: the figure for engine failures (power plant, not core failures) comes out consistently at 1:100,000/flying hour. The probability of failure of an engine in a single is therefore 1 x 10**-5; the probability of an engine failure in a twin (one or the other) is 2 x 10**-5; and the probability of two failures in a twin (from unconnected causes - i.e. one and then the other) is 2 x 10**-5 and (times) 1 x 10**-5 - that is 2 x 10**-10"
Jim you are absoloutly right that that is the basis used BUT Jim this is arithmetic nonsense, it is faulty maths, incorrect and wrong:
This is why:
The 1:100,000 is per flying hour . So the result you have calculated is the chance of both engines failing within an hour of flight. So to get the chance of the second failing within the subsequent 10 seconds (say) you have to divide by another 360, making the chance of this occuring about 5^10-13. That is once per 50,000,000,000,000 hours (that's once per 50 thousand billion hours !!). Since the chance of losing 2 engines within 10 seconds has happened at least 10,000,000 times more frequently than that there must be something wrong with some other assumption in your statement. The only thing that is likely to be is the (ridiculous) assumption that the systems are independant.
What you say is of course true(ish) if (as you say clearly) we assume (wrongly btw) that the engines are independant.
1 There are many reasons why the engines are not independant:
They share a common drive chain.
They have the same pilot (trying) to control them.
They share fuel system components.
They share common fuel (impurities, wrong grade etc)
Theyt are located near eachother (so they would have to theoretically never interfere?)
The engine controls look alike and are located in a similar place.
They have had a similar history having flown together for possible all of their life.
They consume the same air (or water, snow, ice, flame, dust, chemical or volcanic ash!)
They are both maybe fixed by the same engineer (who puts the wrong oil in both etc etc)
The second engine is expected to do more than it has ever done before and operate to new maxima, which are previously untested (for that engine).
Same electronic control components subject to RadHaz, software problems.
The heat seeking missile (which might have an easier job!) doesn't know it is 'only suppose to blow the bloody' port engine.
2 And not to be ignored there are other factors where the risk is increased:
Reduced margins of other critical components.
Increased risk from engine explosion/fire.
Increased system risk from increased complexity gearbox. (More critical cogs, and bearings)
Higher risk from freewheel unit faults.
Complexity causing more pilot confusion and more scope for error.
3 Not to mention that to shift the same payload the reduced payload of a twin may result in more trips (exposures) being required.
If the capacity reduction in the NS required twice the number of trips we might expect twice the number of accidents, all else being equal.
People are totally screwed in their use and understanding of statistics: If you increase a risk in the 10^-6 realm in order to improve a risk in the 10^-9 realm then you've shot yourself in the foot (maybe a thousand times!).
(Incorrect) Theory doesn't deliver, that's why legislators should not design soloutions, design engineers should.
There were no losses asociated with banning onshore SE helicopter PT at night. Just unfounded instinct.
The loss rate (of night twins, not neccessarily PT) has been pretty high since.
Last edited by AnFI; 25th Mar 2014 at 22:58.
Reason: error correction and missed missile, add reason