View Full Version : Mean time between failure

22nd May 2018, 15:13
Anybody have any idea of the probability of engine failure on take off in an average club trainer?
I ask this as one place I visit with students try to insist on flying runway track to 500' agl before turning. This is just possible safely only if the takeoff and initial climb is flown to an accuracy beyond most students and ppls not in current practice. The cfi wants me to follow this even if, as normally happens, you have a short period where I doubt you'd be able to glide clear if the engine quit. I've refused and insist on an early turn as soon as I feel I need to, usually through about 20-30 degrees to the left. I'm on my own on this as other instructors don't see the problem.
Thoughts please.

Captain Jock
22nd May 2018, 16:35
A few years since I stopped instructing but I would have been pretty disappointed if a student could not climb to 500 ft agl in a straight line. Surely it is not a good idea to encourage a student to make a 30 degree bank turn below 500 ft?

22nd May 2018, 17:26
MTBF is normally expressed as a failure rate per specific time, usually per every thousand flying hours.
You can either use a know failure number or if this is not available a generic number. This can be a high failure number for early failures such as infant mortality, a lower random failure number for the bottom of the bathtub shape and a higher age related or wear out number.
Next is the likelihood of such a failure occurrence per unit of time. If you assume a typical unit of time is one flying hour then you need to consider (in this case) takeoff. If you assume the takeoff phase as two minutes, which would include the climb, then MTBF can be calculated.
As an example the engine failure characteristic is one per thousand flying hours then the MTBF just for the takeoff phase would equate to one thirtieth of one per thousand flying hours, assuming a single engine case.
This is a fairly simplistic case as to be more accurate you could consider that the engine might be at maximum power/maximum stress during that two minutes. Having said that the engine should be designed for the maximum power situation.
I hope that this helps.

22nd May 2018, 18:53
You are looking at the wrong number. MTBF is 1/failure rate, which is the number of failure expected per unit of time at some specified confidence interval (typically 95%). This tells you nothing about whether your engine is about to fail. The number you want is the probability of failure at some specific time t - this is called the "Hazard rate". Engine failure is a "complex" event - one which can result from multiple cuases each or which has a different probability density function. In such cases the Central Limit Theorem tells us that the overal failure probability density function will approximate to an exponential distribution (the more the number of failure modes the closer the approximation). From the point of view of your question this is bad news, because the Hazard Rate for an exponential distribution is constant.

What does this mean? Well it tells you that unless there is a specific cause which has a higher probability just after take off (eg carb icing or birdstrike) the probability of a random failure occurring just after take off is no more or less than at any other time.



22nd May 2018, 19:03
You don't need to know the MTBF - just fly it like it could go at any minute. Of course you absolutely should fly a profile that gives you forced landing areas. Just flying straight ahead to 500' with nowhere to go (when you have better options) because the numbers sound good?

Ascend Charlie
26th May 2018, 10:46
Most airfields have a relatively clear area on the extended centreline, making it a little safer to go straight ahead if the engine stops.

Do not ever attempt to turn back to the runway below 500'.

Even in a jet, it was an ejection if we were below 800'agl - above that, we could look at a turnback, but in an unexpected failure, the delays in realising and assessing usually meant it was too late to turn back. Convert speed to height, eject. Use the Martin-Baker Departure Lounge.

26th May 2018, 21:20
Where I used to work there was an immediate turn on both ends of the runway - nothing wrong with teaching a student to turn at point X provided they’re taught properly.

However, if that is the standard and there are no clear benefits to making earlier turns then you have an issue. Behind safety, standardised teaching from all the instructors even in any training environment should be paramount.

Big Pistons Forever
27th May 2018, 04:41
If you look at the aggregate engine failure rate you will find that around 80 % of all engine failures are caused by the actions or in actions of the pilot with carb icing and fuel contamination and mismanagement the leading causes. A total engine failure with no warning is statistically the least likely scenario

The best way to deal with an engine failure is to not have the engine fail in the first place. The way to avoid the EFATO is do a complete and effective DI, effectively checklist use and understand what the engine is telling you when you do the run up.


I personally have elected not to takeoff after observing the following issues. Can you tell me why I aborted the flight ?
1) There was is no Mag drop when the Mags were cycled

2) At runup RPM both the oil temp and the oil pressure are at the bottom of the green

3) During the runup the ammeter constantly showed a + 20 amp indication

4) There was no drop in RPM when carb heat was applied

5) The fuel tanks were dipped and were almost full. The left fuel gauge showed just below full and the right gauge was pegged at the far right past the full indication

6) Assuming you are flying a fixed pitched prop aircraft what RPM would you expect to see at the beginning of the takeoff run at the moment you have the throttle full in?

30th May 2018, 20:35
A few years since I stopped instructing but I would have been pretty disappointed if a student could not climb to 500 ft agl in a straight line. Surely it is not a good idea to encourage a student to make a 30 degree bank turn below 500 ft?

Hmmm - have a look at the departure on 06L at EGTB, for example!

31st May 2018, 12:35
Big Pistons Forever, I think I know the reasoning behind most of the points you mentioned but I'm wondering about 5). I've had quite a few occasions where the fuel level sensor on a C172 would be misbehaving but I've always been happy to go, provided the fuel was dipped before departure and we had ample reserves (meaning an extra 2 hours of fuel or so, usually) for the planned flight. I'm always eager to learn, what was your reasoning for not accepting this particular issue?