The bad news and good news about engine failures

You’ve probably said it to a nervous passenger: “Don’t worry, airplane engines almost never quit.” It’s only in World War II movies that engines cough and pilots have to save the day, right? This is mostly true for turbine engines, which have a stunningly good reliability record. Unfortunately, it’s far less true for piston engines.

The bad news

While the general aviation industry doesn’t exactly advertise engine failures, the numbers aren’t hard to find. According to the NTSB and FAA, there are somewhere between 150 and 200 accidents per year that are caused by power loss. Roughly a quarter of these are fatal, which makes this the second leading cause of fatal accidents, behind the much-maligned “loss of control-inflight.” For twins and experimental aircraft, the accident rate is even higher. Perhaps worst of all, such events seem to be holding steady, even though overall accidents have declined somewhat in recent years. Here are the statistics for 2014, the most recent year for which data is available and right in line with the five-year trend:

NTSB accident graph

Data from AOPA paint a similar picture. Here is a chart from the most recent Nall Report, showing the cause of mechanical accidents. This excludes pilot error, but note that powerplant failures easily beat out other system malfunctions:

An interesting study from the Australian Transportation Safety Board (which does some first-rate safety research), offers more supporting evidence: there were 322 engine failures or malfunctions between 2009 and 2014. That’s over a dozen per year, in a country with fewer than 10,000 “sport and recreational” aircraft. While the poor reliability of Jabiru engines jumps out (with over 40% of the engine failures), the study still found 86 failures of Lycoming and Continental engines.

It’s worth pointing out that all these statistics almost certainly undercount the actual number of engine failures, since they only appear in these reports the result is an accident. An engine failure that leads to a safe landing at an airport will never show up in the statistics, and may not even make it to an insurance company’s figures.

The good news

Those numbers may be surprising to many pilots who have thousands of hours flying behind the latest engines from Lycoming and Continental, which seem to be smooth and trouble-free. Those numbers might even shake your confidence just a bit.

The good news is that “System Malfunction (Powerplant)” hides an awful lot of important details. That NTSB phrase merely defines the event, not the root cause. And by far the most common reason piston engines quit is because they don’t receive fuel, either due to fuel starvation (the airplane has fuel but it doesn’t make it to the engine) or fuel exhaustion (the airplane truly ran out of it). These two causes account for over one-third of engine failure accidents, but they are completely under the control of the pilot.

Pilots will probably always find ways to run out of fuel, but it bears repeating that a few good habits can dramatically reduce your chances of such an engine failure. Having a hard one-hour minimum is a great place to start – under no circumstances can you still be flying with less than one hour of fuel in the tanks. Next, spend some time understanding the fuel system so you can always get that fuel to the engine, especially in twins and older airplanes with complicated fuel systems. Take a little time away from practicing a rare emergency and instead discuss different fuel scenarios that might pop up. Finally, always know how much fuel was in the tanks at engine start and know your real-world fuel burn rate. Depending on those, and not as much on the gauges, will lead to more realistic decision-making.

Engine stopped in a Cessna
The engine quit. But why?

Behind those mental mistakes on the list of engine failure causes is a surprise (at least for me): fuel contamination. While I’ve been fortunate to never find bad fuel in 20+ years of flying, even in some quiet parts of the Caribbean, it does happen and it can have serious consequences. Again, though, this is almost totally preventable, in this case by performing a thorough preflight every time and staying with the airplane whenever it’s fueled to verify you get the right type.

Next comes caburetor icing, which is either impossible (fuel injected engines) or preventable (by using the carb heat). The penalty for pulling that knob is fairly small in most airplanes, so when in doubt you should use it – even if the conditions seem inhospitable for icing. A carburetor temperature gauge is a good idea too, especially for some Continental models.

By the time you get to real mechanical failures such as a failed magneto or a broken connecting rod, the numbers are fairly small – less than 20% of all powerplant problems. Some of these are simply bad luck, with the pilot dealt a bad hand by a sudden part failure, but it’s not all up to fate. A decent number of these mechanical failures were due to faulty maintenance, typically soon after major repair work or overhaul. This argues for high quality maintenance, but it also supports Mike Busch’s theory of Reliability Centered Maintenance, where overhauls are completed on-condition, not based on an arbitrary time limit.

The right habits

What’s the takeaway for pilots and aircraft owners? First, the easy stuff. Develop good habits regarding fuel management and maintain the discipline to follow them every time. Sample fuel before every flight, buy from reputable FBOs, and make sure your fuel caps seal tightly. Use carb heat (if applicable) on every flight to prevent icing, not just when the engine starts to run rough.

Beyond those everyday basics, a wise owner will seek out high quality maintenance, but perhaps only when it’s really needed. The right balance will keep the engine under close supervision, with regular oil analysis and borescopes, but avoid added risk by doing invasive part replacement too often. When maintenance is performed, pilots should be skeptical on the first flight after overhaul or parts replacement. In other words, don’t make that first trip a hard IFR trip over the mountains.

Of course some traditional advice also helps a lot: fly the airplane regularly, avoid cold starts, and operate the engine conservatively (especially with respect to CHT). The biggest payoff from these habits is in longevity, especially when it comes to preventing corrosion, but there is certainly some improvement in reliability to be had as well.

Following these rules can reduce your chances of an engine failure by over 75%, which should make you feel a lot more comfortable on your next flight over remote terrain. But that still means the fan out front can stop turning. In that situation, all you have to fall back on is good training and realistic planning. Based on the numbers above, practicing engine failure scenarios as a part of your regular training is time well spent, and continuously thinking about forced landing sites in flight doesn’t hurt either.

Engines can and do fail, but a little preparation and a little paranoia should keep you from adding to those statistics.

22 Comments

  • I have to take issue with the generic statement that “turbines are known to be more reliable than pistons”. These statements are just based on global statistics of engine failures vs time. They do not take into account the differences in owner/operator groups or operational environments.

    The airlines did experience much greater reliability of turbines vs pistons but the piston engines they operated were a whole ‘nuther animal.

    • I think you have a point, although even excluding airline operations, the turbine record is awfully good. Look at Piper Meridians or TBM 700/850s vs. Cessna 210s or Piper Malibus. The statistics are quite different.

      • John, At an NBAA meeting a few years ago the main man for FAA statistics had the data to prove the following statement that he made:
        “The single most improvement in ALL aviation safety since the Wright Brothers was the introduction of the turbojet engine”

        just sayin’

        Kent Ewing, Captain, USN, Ret
        ATP, CFII, MEI,
        18500 hours/6500 flight instruction hours
        v/p Bonanza/Baron Pilot Training, INC

      • John, no question about that. But where I see the big disparity is in the operator/operations factor. How many Meridian/TBM owners let the engine run past TBO or don’t bother complying with manufacturer’s Service Bulletins? I would guess very few versus the typical piston owner that will commonly run the engine way beyond manufacturer’s recommended chronological TBO (most don’t know there is a chronological TBO) or the operating time TBO. Engine Service Bulletins? Most don’t even know they exist. I believe it is reasonable to speculate that turbine aircraft maintenance is a large step above typical piston maintenance. And that, I believe, could skew the statistics.

        • I don’t think running a recip engine past TBO is a major contributor to accidents. I don’t have the data right in front of me, but I recall reading that reciprocating engines are more likely to fail catastrophically in the FIRST 100 hours after an overhaul. I would ascribe the differences in reliability to complexity (turbines tend to be mechanically simpler in the high wear/high temp parts), and cost of manufacture (a turbine engine is 10x the cost of a comparable reciprocating engine and designed to much closer tolerances). My A&P (he exclusively works on engines in spite of his rating) agrees with the idea that a piston engine is more likely to grenade just after/during break in than later down the road.

          Our club’s 172 had a 2100 hour o-300 in it that leaked oil, but otherwise ran like a top. The newly overhauled engine runs like a swiss watch, but I’ll feel a lot better about it once it makes it to 100 hours.

  • Maybe they should make “mechanical failure of engine” a separate category.
    The original statistic was much more alarming until I read the article.
    As a turbine pilot who hopes to be able to afford a little GA plane someday that really got my attention.

  • Overall the article is very informative.

    The total number of accidents is a useful metric, but it would be helpful to put these numbers in context such as engine failures per flight hour or engine failures per flight. Can the US, Nall and Australian values (generated over a 6 year period) presented this way?

    • Estimating flight hours is really tricky. The best guess comes from the FAA’s annual activity survey, and even that is not much better than a guess.

      Just doing some quick math, the 2014 FAA survey says the GA/air taxi fleet flew 11.97 million hours on piston engines that year, or 13.21 million hours if you include experimental and LSA. That leads to a rate of about 1.6 powerplant failures per 100,000 hours.

      Again, this is super rough math, but that’s compared to a total GA accident rate of about 5 per 100,000 hours and a fatal accident rate of about 1 per 100,000 hours.

  • I believe the numbers are actually much worse than the FAA’s data indicate. My family and I experienced an engine failure in a Piper Warrior several years ago. We were quite lucky and there was an airport nearby. We glided in for a normal landing, turned off mid-field, and got out and pulled the airplane the rest of the way to the ramp. This incident was not recorded – because there was no accident!

    • I agree. I also don’t think the multi engine stats are as bad as reflected in the studies. I had an engine failure in a Cessna 340, landed safely with one, not in the stats at all.

  • John, I am in the 20%..youtube: Bonanza Crash Thanksgiving. #2 connecting rod bearing failed. Can you get statistics on which mfg fails more…Lycoming vs Continental? I have a theory /unsubstantuated..that #2 rod on continentals fails more often than any other lower end part. …just comments from engine overhaul folks.

    Will use your excellent statistics at my next Bonanza/Baron clinic in Concord, CA

    Kent Ewing, Captain, USN, Ret.
    V/P BPT Inc
    now flying a 55 Baron

    • I had a dual connecting rod failure back in June in a P206B with a CMI IO520 FCA. The #2 let go first, tearing off the left mag followed by the #4 at 53 hrs. since overhaul. Fortunately, I landed unscathed in a waist high corn field. The airplane was flying again by mid August.

  • I worked as a fire investigator, we would visit the local flight schools and we were told that most all the aircraft had Lycoming engine’s some with three to four thousand hours without o/haul or mishap, I’ve yet to see a Continental engine with only a fourth of these times without o/haul. I owned a C-210 Centurion almost made it to one thousand hours before I started to replace cylinders, just saying. I also worked as a machinist for a aircraft o/hauling Co. most of the failures and o/hauls were on Continental engine’s and I might add they were all well before o/haul time. I’ve owned Comanche’s with IO-540 size engine’s and never a problem and I will say,” very dependable”, I remember Continental engineers not wanting to answer direct questions on there reliability of some Continental engines. just saying. I like Lycoming engine’s.

  • Being an Engineer, a pilot for ~60 years, and worked for Pratt & Whitney for ~40 years; there is no doubt that turbine engines are much more reliable than piston engines. Turbine engines have essentially NO vibration; but Piston engines are ‘pounded’ with vibration inherently. That’s one of the major reasons Turbine engines have a TBO that far exceeds the TBO of Piston engines. Amen!

  • Couple of comments:

    I’ve had a total failure of the Lycoming O-360 in my P172D (Avcon conversion), shortly after I bought the airplane about 14 1/2 years ago. Probably caused by lack of use by the previous owner who had flown it only a couple of hours between annuals for the last several years, a bearing apparently spun cutting off oil circulation, leading to a thrown rod. I landed in a field, with no damage to the airplane. Best soft field landing I’ve ever made!

    Just because fuel injected engines can’t get carburetor ice doesn’t mean that they can’t get induction ice. On a winter flight in a Mooney 231 with the OAT well below zero, the engine lost all power at 12,000′ in IMC, with no visible icing on any surfaces. I declared an emergency and was vectored toward Pueblo, CO, which had weather at minimums of 200′ and 1/2 mile. After losing several hundred feet, I remembered the manual alternate air knob hidden above my right knee, pulled it, and the engine returned to life. Rather than land in minimum conditions, I elected to continue the flight. Over the San Luis Valley, we broke into VMC which continued all the way to Durango, our destination. After landing, I looked the airplane over. There was no ice on any of the flight surfaces, but there was a thick coating of ice all over the front of the cowl, blocking the air intake. My theory has been that running the heated prop caused ice crystals to melt to stick to the cowl. Whatever the reason, it taught me that even a FI engine could be affected by icing.

    I’m no super pilot, but keeping one’s cool in the event of an engine failure is pretty important to a successful outcome. That helps to stay out of the NTSB data base, but it also warps the engine failure figures. No accident means no record.

  • Another great article John. You are stepping into Richard’s shoes. Look at the data, and tell us what we can do to improve our odds. We all know that flying has risks, especially with 70 year old technology. Managing that risk improves our chances. I do recommend Engine Out Survival Tactics by Nate Jaro. It is a bit wordy but there are a number of exercises and tactics that can improve your chances in the case of an engine failure.

  • Very interesting article, with good reminders about the things we as pilots have control over. I am having a bit of trouble however following the statistical analysis cited in the article. The NTSB category “Fuel Related” covers all the events of fuel exhaustion, starvation, contamination and carb ice. I assume its meant to be a separate and distinct event from “System Malfunction (Powerplant).” If so, how do you get to the conclusion that only 20% of the engine malfunctions are caused by things like failed magnetos as opposed to fuel management?

    Thanks

Leave a Reply

Your email address will not be published. Required fields are marked *