13
$\begingroup$

Before the first jetliners appeared, all airliners used piston (reciprocating) engines, which were astoundingly prone to failing in flight, to the point where an inflight engine failure was an everyday, expected occurrence. To quote Wikipedia:

Engine failures were considered fairly routine events on piston-engined airliners in the 1940s, so the crew elected to continue the flight to Dallas, and Captain Claude announced to the passengers that they would switch to another airplane upon arrival.

In contrast, in-flight failures of modern turbofan engines are so rare that most pilots will go their entire careers without ever encountering one. Even in the 1950s, jets were already more reliable than piston engines, which was one of the reasons they quickly seized centre stage for long-distance operations. For low-altitude short-haul flights, where jets are inefficient compared to propellers, airliners still switched away from piston engines, moving to turboprops en masse despite piston engines having better fuel efficiency than turboprops. All of this is despite turbine engines placing far greater thermal and mechanical stresses on their components than any piston engine.

Nowadays, piston engines are generally used only on very small general-aviation airplanes... and are no less reliable than the turboprops and jets on larger aircraft.

What is it about piston engines that makes them so unreliable on large aircraft, yet extremely reliable on small aircraft?

$\endgroup$
  • 6
    $\begingroup$ There are A LOT of moving parts in a two-row radial engine (R-1830). Some of the bigger engines had up to four rows (R-4360) but not many, if any, popular passenger planes used the biggest engines. With more moving parts, more opportunity for metal-on-metal wear, and the general complexity of the valves/turbos/carbs...there was a lot to go wrong. A turbine has fewer moving parts and creates MUCH more power for its weight. The monster -4360 made 4,300hp and weighed 3,800lbs dry. The Allison T56 makes roughly the same power and weighs less than 2,000lbs dry with fewer moving parts. $\endgroup$ – acpilot Dec 20 '18 at 23:57
  • 1
    $\begingroup$ Plus any more than two rows of cylinders was extremely difficult to cool properly, which was the main reason for the problems with 3 and 4 row engines. $\endgroup$ – John K Dec 21 '18 at 1:54
  • 9
    $\begingroup$ I think you're overlooking the great improvement in the reliability of all engines since the era of piston-engined airliners. Compare a 1940s auto engine to one from the '60s, and to a modern one. Then too, I think turbines actually place a lot less stress on their components. Turbines basically just go around and around, while pistons, valves, and so on have to reverse direction a couple of thousand times every minute. $\endgroup$ – jamesqf Dec 21 '18 at 5:35
  • 5
    $\begingroup$ "Nowadays, piston engines are generally used only on very small general-aviation airplanes... and are no less reliable than the turboprops and jets on larger aircraft." - As someone who commutes daily in a single-engine piston airplane across an urban areas, I wish that were true. Light aircraft piston engines, including modern ones, have a comparatively terrible reliability record. $\endgroup$ – pericynthion Dec 21 '18 at 6:24
  • 1
    $\begingroup$ Nowadays you don't have to change the points on your distributor, that doesn't exist anymore, or clean the carburetor and change the fuel pump every 40k miles. $\endgroup$ – Mazura Dec 22 '18 at 16:56
2
$\begingroup$

This answer does not need to be complex. Piston engines have many more parts, and they are reciprocating parts. They are much more complex machines than turbofans. That's the answer.

Turbofans are fairly simple machines. Yes, there are a lot of parts, but many of the parts are static, or move very little. Their development and construction is complex, but their operation is relatively simple, and therefore more reliable.

For example; an 18 cylinder radial engine requires approximately 54,000 ignition events per minute of flight. A turbofan requires only one ignition event for the entire flight, and it happens on the ground.

Similarly, that same radial requires two valve operations per ignition event. That's 108,000 valve operations per minute, or about 19.4 million valve operations per engine on a 3 hour flight.

This means that a Douglas DC-7 would have 78 million valve operations on a 3 hour flight, and 39 million ignition events. Ignition systems used to be largely mechanical, with many moving parts. (Electronic ignition and distributorless ignition has changed this a lot in modern piston engines.)

Combine with this, the fact that they had devices like the turbo-compound, which you could think of as a turbo-shaft drive that helps drive the crankshaft. Each engine on a DC-7 had three of them. Incidentally, they caused reliability problems because they raised exhaust temperatures, and were known to cause exhaust valve failures.

3 hour DC-7 flight example at 350 mph: Denver to San Francisco

Reference: math, and engine knowledge

The Wright R-3350 on the DC-7 is geared at 16:7, or about 2.29:1. Use 2,600 propeller RPM to get 6,000 engine RPM.

$\endgroup$
21
$\begingroup$

This statement is not entirely true. the R2800 engine was considered at the time to be quite reliable in comparison to the R3350 and the R4360. Those engines represented the evolutionary dead-end of big piston power, where pushing to more power began making the engines less reliable than their lower-power predecessors had been.

Part of this was simple statistics, where the probability of failure of a machine scales with the number of parts it contains. In crude terms this means doubling the number of pistons doubles the opportunity for something to go wrong with one of them in the engine. Note also that those extra pistons do not provide redundancy, in the sense that a failed piston will fill the oil system with broken metal, causing the other pistons to fail not long after, or start the engine on fire.

Turbine power plants provide more uptime and longer TBO's than the piston engines they replaced as well as more power for less engine weight. These attributes offset the greater replacement costs and higher fuel consumption in applications where uptime and power were required to enable a viable business model.

Cost-sensitive, low power applications still make use of pistons, with the breakpoint being around ~300HP. This is at least partly due to the fact that as turbines are scaled down in power output, their efficiency usually suffers, and their manufacturing costs do not fall fast enough to compete with pistons in the 250HP arena.

$\endgroup$
  • $\begingroup$ One would think that the cascading-failure problem with big piston engines could easily be solved with a simple oil filter... $\endgroup$ – Sean Dec 21 '18 at 3:44
  • 14
    $\begingroup$ well, once the piston bits fall down into the crank shaft, it's game over... $\endgroup$ – niels nielsen Dec 21 '18 at 4:15
  • 4
    $\begingroup$ It's not like turbines would be more cascade failure resistant.... :) $\endgroup$ – rackandboneman Dec 21 '18 at 19:27
15
$\begingroup$

You in-part answer your own question

piston-engined airliners in the 1940

compared with

in-flight failures of modern turbofan engines

Simply put we have learned a lot in the past 75 years about manufacturing and construction, as well as how to test for, identify and prevent fatigue. This knowledge has helped piston engines just as much as it has jets. As others have noted there are also far more moving parts in a piston engine that all have to operate in sync for everything run.

Since piston engines are no longer used on the big aircraft the best analog is those that are still flying on GA airplanes. FAA analysis from the twenty years between '84 and '04 show that only 26% of accidents were mechanical and of those only 11% were reciprocating engine failures. In other words aviation piston engines are actually pretty reliable when used properly. Other trend reports note that failure is often related to misuse of the engine control system. In the 40's there was no FADEC, there may not have even been rev-limiters, operational parameters were based on decades of data not nearly a century.

...so unreliable on large aircraft, yet extremely reliable on small aircraft?

The piston engines on small aircraft are not by any means new the Lycoming in the archer I fly is pretty much the same as what it was when it came out in 1955. Its reliable because our machining, casting, manufacturing and related functions are better than they were in 1955. A lot of people are flying around with full engine monitors which is a lot more than the single temp gauge the airframe originally had. A case could be made that an extra 50 years of development on large piston engines would have made them just as reliable as their GA counterparts.

Turbines are not without their issues either.


Another thing to consider is that a lot of the hanger talk (or whats left from that era) of large piston multi flying has heavy military roots. Engine servicing in the field was good but not great, and when heavy daily use, live fire, and not always "by the book" operations are accounted for reliable operation should never be expected. This has lead to a lot of stores of failed engines, gliding home (or into a field) etc. This helped to fuel the twin market for a long time and helped to create the notion that large piston engines were unreliable.

$\endgroup$
3
$\begingroup$

While it doesn't relate to reliability, operating piston engines at enough power at both high and low altitudes requires the use of superchargers (which add mass, decrease the efficiency, requires additional components, ...).

Also, not strictly related to reliability:

Hughes HK-1 'Spruce Goose'. Eight R-4360s making up a total of 448 spark plugs for one aircraft.

Avro Lancaster (4 engines of 12 cylinders, two spark plugs per cylinder). ... and it's real fun changing them. Mind you its not as bad as checking and re-setting the valve clearences, come on how many, 48 X 4 = 192 of the bloody things....

$\endgroup$
  • 2
    $\begingroup$ 4 engines x 12 cylinders/engine x 2 spark plugs/cylinder = 96 spark plugs... $\endgroup$ – xxavier Dec 21 '18 at 14:18

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.