So @Flyer59 we are saying that an auto engine with even lighter (so presumably less durable) parts has better than turbine ETOPS 120 reliability?
No, i am not qualified to say that. I would say that the cylinders, valve train etc. are more reliable in a Mercedes based aircraft engine. Other components (gearbox) create more problems.
But a turbine is definitely more reliable – but of course it depends what we’re talking about. The MTBF of a PT-6 is around 150.000 hours, if i remember correctly …
I wasn’t thinking of a specific turbine but rather specific standards. The early ETOPS 120 aimed for 5 IFSD’s per 100k hours. This after 40 odd years evolution in jet engine technology. The MB “off-the-shelf” engine is down to 3.3 IFSD per 100k hours in spite of “not being designed for aviation” and “not being able to produce more than 40% MCP because it’s a car engine”.
Hopefully eventually people are going to realise that nowadays avionics (just like powerplants) in spite of not being certified according to ca. 1960es requirements can be as good if not better when installed :-)
In the context of applying ICAO/EASA risk analysis (see below) to GA SEP operations, which level of probability applies to mechanical failure leading to immediate and total loss of propulsion?
Background:
In its GA policy framework consultation response the UK CAA states: “we will use the EASA criteria to assess risks to third parties whilst allowing GA participants to assume higher levels of risk”.
By way of a baseline for such allowable higher levels of risk, the UK CAA chief executive Andrew Haines wrote in his introduction to CAP 1123 “For example….you can fly at night in the UK in a single engine aircraft, providing you have the appropriate rating on your license; and you can land at night without runway lights – that is for you to assess the risk.”
Knowing the probability of power plant failure, we could quantify one or both of Mr Haines’ examples so as to see just where they sit on the ICAO/EASA risk matrix depicted in Annex C of CAP 1188.
From here
Fuji_Abound wrote:
If you selected two different populations, one, let us say where the engines had been blue printed, only the best engineers employed to service and check the engines, and the engines always operated within recommened parameters are you suggesting this populartion would have the same mtbf as a radomn selection?
Well, MTBF is an engineering parameter used in engineering. The premise IS the bathtub curve. Max MTBF is valid only along the flat bottom of that curve. This is “assured” through:
Obviously this is an idealized setting in a way, but it’s not that difficult to achieve. The problem is more that there are no good value for the MTBF for these engines. I have only seen guesstimates, or perhaps numbers made from NTSB reports, which are still only guesstimates because the operating time is unknown. It’s reasonable to assume it’s above 2k h and below 100k h somewhere, but closer to 2k than 100k IMO. According to this, the failure rate of a PT6 turboprop is 12 in a million flight hours. The MTBF then becomes 1/F = 83k hours. I refuse to believe that a Lycoming/Continental/Rotax is more reliable than a PT6, so more than 100k h is pure nonsense IMO.
Also, only a small fraction of failures, even compete failures where the engine stop 100%, cause fatal accidents. Most pilots do after all fly (mostly, lets say 99% of the time) in a manner so that engine failures are survivable. This is IMO the essence of the discussion. What is the risk of flying in a way so that an engine failure is not survivable? The accident statistics won’t really help you all that much, it doesn’t include those numbers. The only thing you “know” is that the MTBF is somewhere between 2k and 100k hours.
As I have shown, even if you assume 100k hours (which must be above the actual number IMO), the risk of flying in a way so you rely 100% on the engine, is way above most other stuff you can possibly do. That is, do on an hour per hour basis.
JasonC wrote:
The average does not describe the shape of the curve or imply a particular outcome for an individual member.
Irrelevant. Max MTBF is what you get when everything is working properly, and the only failures to occur are random failures. You cannot do maintenance to prevent random occurrences. An engine is not a fail safe device. I have said nothing about the average. I have only assumed an MTBF for a well working (top condition) engine. If the engine is not well maintained, you can be sure the MTBF is much closer to 2k than 100k, it could be much lower than 2k in fact.
What you mean is perhaps about probabilities, exposure and risk? Base jumping is highly risky, but you stand no risk if you don’t do base jumping. You stand no risk of dying due to engine failure if you fly so that an engine failure is survivable. But what happens if you don’t?
AeroPlus wrote:
How much you feel comfortable with depends for a large part on your personality.
Exactly. But also age. For instance, the MTBF of a 70 year old male is what? a couple of weeks? I mean seriously 
At a certain age, you stand a much higher risk of your heart stopping, or some other acute and deadly thing happens than the risk of the engine stopping. The risk of flying is small compared with the risk of being so old. I think we unconsciously feel this fact. The risk of dying at any minute increases for each day, so the problem becomes more about having a great time while alive, than reducing the risk associated with that time. The risk of living (the risk dying at any time due to no particular reason other than age) becomes so large it will eventually overshadow everything else.
Accidental death is one thing. We can control that to some extent. But age, we cannot do anything about.
I can’t really follow the numbers, @LeSving.
The number of engine failures (other than fuel starvation) in a Lycoming or Continental is around 1 per 10,000 hours (see here, for example local copy ). But that is across the entire “bathtub curve”, with real-life maintained and operated engines, not a theoretical MTBF in the flat bit of the curve,
So a “top condition” engine will have a failure rate quite a bit lower than 1 in 10k, and of course a badly maintained one a higher one.
Cobalt wrote:
But that is across the entire “bathtub curve”, with real-life maintained and operated engines, not a theoretical MTBF in the flat bit of the curve,
There are good arguments that a well-maintained engine doesn’t have a bathtub curve. Statistics show that the failure rate is highest when the engine is new or newly overhauled. After about 500 h the failure rate as stabilised and is essentially constant until the engine is overhauled again – even if this is done well past the official TBO.
So a “top condition” engine will have a failure rate quite a bit lower than 1 in 10k, and of course a badly maintained one a higher one.
Could be right. It’s roughly 10 times the rate of a PT6 turbine (also across the board by the looks of it). They are certified though. It means the MTBF is around 10k, and the PT6 has 88k h.
MTBF is not a meaningful statistic if the object is very reliable during its design life (as humans are) but has a fairly hard life limit (as humans have). It doesn’t take long to find this calculation, for example:
There are 500,000 25-year-old humans in the sample population.
Over the course of a year, data is collected on failures (deaths) for this population.
The operational life of the population is 500,000 × 1 year = 500,000 people years.
Throughout the year, 625 people failed (died).
The failure rate is 625 failures / 500,000 people years = 0.125% / year.
The MTBF is the inverse of failure rate or 1 / 0.00125 = 800 years
In a part of the UK, from here, you get 87 years for a 77 year old
so we can expect to see the usual great contributions on EuroGA from Maoraigh until I am around 71
