Why are all engines identical in typical commercial airplanes?

Question

I come from the IT world where redundancy is in place to cover several cases, among other:

• a statistical failure of a device (so we use more of them expecting that the event is independent)
• a rogue usage of a device (hacking into it though a vulnerability)
• a specific context where a bug is triggered

In order to account for the last two ones, redundancy is sometimes (or usually - depends on the criticality of the system) done with different devices (form different vendors). This is typically the case with firewalls, for instance.

I was wondering why this isn't the case with aircraft engines?

It would be "easy" with planes having four engines (so two of each kind), somehow less "easy" with two engines (where the differences between their regimes would need to be compensated - the pilot states that they want 80% of power, which, in the specific conditions they are in, means 78% for the left engine and 83% for the right one).

Having two redundant engines which are equally susceptible to be impacted by, say, ashes is probably not interesting, compared to having two different ones which would each have its strong and weak points (but different between the two)

---

Note 1: "easy" above is used as a placeholder only. I am aware that aircraft software and hardware is complicated, I mainly wanted to highlight the difference between 2 and 4 engines. Having two firewalls is a simple task compared to that.

Note 2: it may be that engine failure is not a big deal, statistically speaking compared to other issues (or in an absolute sense), and therefore the extra maintenance, structural compensations because of the differences etc. are not worth it. Thanks for pointing this out if this is the case

Answer 1

Engines don't fail, on average. And if they do, it's a very low probability that two engines will fail at the same time.

Modern jet engines are extremely reliable, with failure rates on the order of 0.01/1000hrs. And if you do have a failure, you have redundancy, as you have two (or more) engines. A plane can stay aloft and land with a single engine.

It would add a lot of complexity and maintenance costs, with no real benefits.

If the failure is external, such as water in the fuel, lack of fuel, or ash ingestion, two different engines are unlikely to help, because all jet engines works in a very similar fashion. The external problem would simply affect both, maybe at slightly different rates, but ultimately shutting them down anyway.

So it adds cost and complexity, without providing any real improvement. Jet liners simply does not crash because all engines stops due to internal problems.

The one case where it would help is if there's a design fault with the engine, e.g. manifesting in turbine breaking after x hours. But engines are extensively tested before being put into serial production, and this occurs very rarely on production aircraft. In addition, the likelihood of the same problem occurring on two engines in such a sport span of time as a flight, but not at all during testing, is a rather low likelihood.

Answer 2

In IT more hardware is redundancy, but in aviation a liability

Short answer

• In IT, having redundancy and diversity is low cost and no liability, but high gains in service reliability.

• In aviation, redundancy and diversity is high cost and a liability, for no increase in reliability of the service.

The long answer

The big difference is that in case of a hardware failure in IT, you just lose a bit of capacity. The service stays up, the customers do not even notice, you revenue stays intact. In IT, a failure is all right, and is allowed to happen.

In aviation, failure means you lose the entire service. All operations are centered on the philosophy that failure is not an option, in the sense: we can deal with it... but it costs a lot and we really do not want to do it.

Let us go through all the consequences of a failure...

Injury, loss of life

• IT: a server blade goes down. So what? No-one is injured, no-one dies.
• Aviation: An engine goes to pieces. That. Is. Not. Good. Losing/injuring colleagues and/or clients on the job must not happen.

Disruption to service, loss of revenue

• IT: a server blade goes down. So what? The performance of the service is slightly degraded. But customers are not turned away. You do not lose revenue over it.

• Aviation: an engine comes apart. You are now rapidly losing money, because as of that moment that aircraft is a cost, not a source of revenue.

The key difference here is that in IT you are only losing a percentage of the service. In aviation, as soon as an engine failure happens, you lose the entire service. This means more engines are a liability, because they mean more possible points of failure.

First the flight in question is instantly cancelled/interrupted, because you do not keep flying as if nothing happened. You are setting the aircraft down on the ground and you are doing it now. This means accommodating all the passengers, either hosting them in a hotel or rescheduling them on other flights... fares you have to pick up the tab for.

Second: the aircraft is now out of commission for unscheduled repairs. This is a big deal because airliners earn all their revenue while flying; they earn nothing while standing around in a repair-shop. An airliner spends about 2/3 of its service life in the air. That is to say: 16 out of 24 hours every day, for 20-30 years, an airliner is supposed to be in the air making money.

Third, you can bet your bottoms that someone had their phone up — even though you told them to turn it off — and filmed the whole thing happening. Then they call their tabloid and email over the link to their YouTube clip of it happening. Your airline's logo is now all over the evening news, your failure published nation- and world-wide. This means loss of customer confidence, which means more loss of revenue.

Replacing the busted hardware

• IT: the cost of the broken hardware and the manhours needed to replace them are rarely counted in excess of 10 000 USD. Now granted this happens a whole lot more often than an aircraft engine failue, but in general, this is a very small cost.
• Aviation: The cost of an engine is in the range of 5 000 000 USD and up. A new Rolls Royce Trent 1000 is 15 000 000 USD. Add to that hundreds of manhours to get the replacement engine out of storage, ship it to the location of the wounded bird, remove the busted engine, inspect the aircraft for damage and certifiy it fit enough to fly, put in the new engine, get the aircraft back in service.

Maintenance

Others have already gone through this so I will just mention it briefly: maintenace costs is counted per unit in aviation. Twice as many engines = twice as much maintenace cost. Such is not the case in IT.

Also — and this is the crux of your question — you ask why in aviation one goes for commonality instead of diversity. This is because in aviation knowledge and experience of a system are commodities. Double the number of types of systems (such as different engines), and you double the amount of people you need to hire, along with double the amount of experience you need to accrue.

Also the support infrastructure needed to deal with one type of system is different from the next. Again: you multiply the maintenance cost for every type of system you add to your organisation.

Conclusion

The only reason airliners have not gone down to using only one engine per aircraft is for safety reasons... since the only thing more unacceptable than having one engine quit in mid flight is to have all engines quit.

In IT, having redundancy is low cost and no liability, but offers high gains in service reliability. In aviation, redundancy is high cost and a liability, for no increase in availability of the service. The same goes for diversity: there are no gains to be had from it.

So in conclusion: failure is not an option. We can deal with it but — unlike in IT — any such failure is an expensive and very disruptive event. We just do not want to have to deal with it. Since more hardware, and diversity of hardware, increases both the risk of failure, and the maintenance and support costs for it, there is nothing to be gained from diversity and redundant redundancies.

Answer 3

Maintenance costs are a big deal. Maintenance costs across multiple engine types for a fleet of the same airplane would be a big deal - training, parts, etc. Worldwide weather tracking & reporting lets pilots avoid ash, thunderstorms most of the time, and other severe weather which is rough to fly in and can leave customers rattled.

I don't think having an engine that would be strong in one thing and weak in another compared to the other engine would end up being the most fuel efficient scenario, and the airlines are all about fuel efficiency. Anything that raises costs will not fly, as it were, unless the gov't oversight agencies make it a mandatory requirement.

Answer 4

One thing not yet addressed is the failure mode of computer systems versus mechanical systems.

In computer systems, redundancy may not be enough if the failure modes are not independent of each other; that is, if there is a design flaw. If both systems contain the exact same bug, and perform the exact same calculations, both computers will have the same error. The solution is to have two independently programmed computers working on different hardware. This is done in aviation for the most critical flight computers.

Mechanical systems are much less likely to be exactly the same across two components. While a design flaw may lead to an eventual failure of a component, this is unlikely to happen across multiple components at the same time. Most extreme conditions are addressed during the initial testing, so most 'unnoticed' problems are gradual failures like fatigue. Since fatigue relies on random defects in the material, it is extremely unlikely that two systems fail at the same time.

To give a bit of an idea of how unlikely a double engine failure could be prevented by different engines, you can check Wikipedia's List of airline flights that required gliding, which I suppose is a reasonable compilation of dual engine failures. The vast majority is due to fuel exhaustion, shutting down the wrong engine, or extreme conditions that would have disabled any engine. While not contained in this list, I think the only accident I can think of that would have benefited from two different engines is British Airways Flight 38 which crashed short of the runway due to similarly clogged fuel/oil heat exchangers.

Answer 5

@CrossRoads touches upon the upkeep implications nicely

Beyond that keeping engines all the same greatly simplifies operation and prevents potential user error. Keep in mind that many of the common airframes out there are of aging design (>20 years old) and generally come from a time when the mentality about computers was not what it is today. FADEC has made engine control on aircraft much simpler and potentially would make your situation possible. But in a time prior to that there are lots of parameters that would vary between engines, a pilot would be thus responsible for knowing double the amount of critical numbers just to fly the plane. For example

• Fuel burn: Even similar engines on the same airframe will have different fuel burns. This is important for trip planning, efficiency, and reserve calculations. Generally in a multi engine aircraft you can make certain assumptions about all the engines being the same.

• Thrust parameters: Back in the piston days this was manifold pressure in jets its N1, N1 and EGT. If you have one type of engine you should see pretty much the same numbers across the panel for a given power setting. If you introduce a situation where the engines are different you would need to know all the combinations for each independent engine. Quick panel scanning would also yield lots of confusion.

• Engine operating parameters: Not all engines have the same operation parameters so you now create a situation where you must know the parameters of both engines and essentially the resulting set of operating parameters that will allow you to do something. Lets say engine 1 has a 5 minute max thrust limit and engine two has a 4 minute 30 second max thrust limitation you can now only do max thrust climbs for 4 minutes 30 seconds due to the lower number. You may also have differing emergency procedures which would make high stress situations even worse.

• Connections: You also run into (on some level) a simple connection issue. Different engines may require different physical mountings as well as different electrical/control mountings meaning you are going to need some confusing mix of components.

---

Failure is a big deal but its just handled differently than you describe. In aviation the notion is you mitigate the engine failure by just having a second engine. There is not really an added benefit (in practice) by mixing up the engine types on an aircraft. Ignoring situations where the engines failed for external reasons (fuel run out, volcanic ash cloud etc) in the more recent engine failure incidents we have only seen a single engine fail even though both are of the same type.

---

FWIW some piston twins solve the counter rotating issue by having engines that actually rotate (at the crank) in opposite directions. While they are almost always the same engine design with cams and cranks made to spin in the opposing direction they are, strictly speaking, different engines.

Answer 6

In the IT world, redundant hardware is typically on standby to be used as failover. This comparison does not fly [sic] when it comes to airplanes.

• The engines on an airplane are in continuous use all the time, so no redundancy there. A plane cannot lift off with just one engine.
• Like you stated you want similar performance between left and right engines, which would need to be compensated all the time if the engines were of a different type from a different vendor.
• You actually have double the risk to have a grounded airplane due to unforeseen problems with a certain engine type (think of airworthiness directives)
• There also is the maintenance issue that CrossRoads highlighted. You'll need spare parts and trained technicians for both engine types, which will be much more complicated and costly.

Answer 7

This answer is from a maintenance perspective:

Aircraft Maintenance seems so easy until the plane is loaded full of passengers and/or freight, there is a pushback tug hooked up, it’s raining or snowing and everyone is looking at YOU to do something quick to get the plane out on time. Why is this relevant?

Imagine you have two different engines installed on an airplane. The crew reports over the radio that the right engine has an issue, but the transmission was kind of garbled. Your lead/foreman dispatches you and lets you know that the right engine has something wrong. You get to the airplane and the crew says, “Yeah, my left engine widget thingy doesn’t work.” You say, “Ah. OK. I was told it was the right engine.” The crew says, “No, I stated on the radio the left engine was bad, the right engine DIDN’T have the problem.” OK.... No time to waste on a slight miscommunication, you go down to look at the left engine. You open it up, and based on some preliminary troubleshooting, you decide the “Widget Switch” is bad. No such part exists, but we will call it that.

You, being in a rush, call your lead/foreman and say, “Hey boss... Send me a widget switch. Oh by the way,” and just then someone from the ramp butts into your conversation thinking they are the most important person in the world, and asks, “Is the plane going to leave?” You respond, and then the pilot is standing over your shoulder wanting to know what you think it is. You explain all of this and then you go about your business removing the old part while your lead (who still thinks it’s for the right engine gets one on order for you).

The new part shows up, just as you have the old one removed. For all intents and purposes, both parts look identical, but one is capable of internally withstanding much higher pressures than the other. You pull it out of the box, and throw the new one in. Everything looks good and goes in great.

You run upstairs to do a quick engine run to make sure everything works. That’s all the manual says to do anyway. You fire up the motor, but because there are two different motor types installed, there are different EGT limits, oil pressure limits, and fuel flow limits. There are also different “widget” limits. In the rush, you fail to notice that your widget limits are different than those of the opposite engine. They are, after all just a few widget detection units different from one another.

You decide it’s good, and sign the logbook off. You depart the airplane, get back to the shop, and begin the process of starting your paperwork. As you’re processing the part, you notice that the part numbers don’t match. Just then you hear your jet roar overhead on takeoff. You get that sinking feeling in your stomach. After 5-10 minutes, another call comes over the radio. The plane you just “fixed” is turning around with one engine shutdown.

The investigation is concluded with you as the sole proprietor of the blame because you failed to check you were putting on the right part. Maybe that’s true, but when all the holes in the Swiss Cheese line up, this is exactly what happens. Bad weather, garbled radio transmissions, interruptions, everyone believes in a sterile cockpit environment but no one seems to believe in a sterile maintenance environment, confusion of parts interchangeability, etc... They all play against you.

Answer 8

The technical term for this problem is common-mode failure, and it's a central focus of modern aircraft design. Where safety is assured by the use of redundancy, common-mode failures are a threat to the effectiveness of that redundancy, so are eliminated or mitigated at every turn.

When it comes to engines, common-mode failures virtually always originate from outside the engine itself.

All engines attached to the same aircraft will generally fly through the same environmental hazards (ash, wildlife) at the same time and are liable to the same modes of failure as a result. Using different engine types would only enhance safety if one of those types was shown to be less susceptible to those failure modes than the other - but in that case, why not make them all of that type in the first place?

All engines are fed by the same set of fuel tanks, so if they all run dry (which is usually due to pilot error rather than mechanical failure), they'll all flame out at nearly the same time - unless the engines are so different as to require different types of fuel. With that said, in practice the two sides of the aircraft operate independent fuel systems under normal circumstances, so that a leak that goes unnoticed for a while only has a chance of draining half of the fuel tanks and disabling half of the engines. This is the type of mitigation usually applied to common-mode failures.

Aircraft have existed with dissimilar engines, however.

The most obvious and famous example would be late versions of the Convair B-36 Peacemaker, which had six "pusher" propellors and four turbojets, hoping to obtain the performance benefits of both types rather than reliability:

Beginning with the B-36D, Convair added a pair of General Electric J47-19 jet engines suspended near the end of each wing; these were also retrofitted to all extant B-36Bs. [...] The jet pods greatly improved takeoff performance and dash speed over the target. In normal cruising flight, the jet engines were shut down to conserve fuel. When the jet engines were shut down, louvers closed off the front of the pods to reduce drag and to prevent ingestion of sand and dirt.

As engine fires occurred with the B-36's radial engines, some crews humorously changed the aircraft's slogan from "six turning, four burning" into "two turning, two burning, two smoking, two choking, and two more unaccounted for." This problem was exacerbated by the propellers' pusher configuration, which increased carburetor icing. [...] Three engine fires of this nature led to the first loss of an American nuclear weapon when a B-36 crashed in February 1950.

As an experiment, a VC-10 airliner in military service had one pair of its original turbojet engines replaced with a single, much larger turbofan. The object of the exercise was to assist with the flight test program of the new engine. Reportedly, the VC-10 flew well in this configuration, but was subsequently found to have suffered severe airframe distortion due to the asymmetric thrust loads, and was retired from service forthwith.

The de Havilland Hornet, essentially a heavy-fighter derivative of the better-known Mosquito light bomber, was fitted with a pair of Rolls-Royce Merlin engines - but with two different Mark numbers (130, 131). In almost all respects these engines were identical, except for the final-drive gearbox; one of these reversed the direction of drive to the propellor, while the other did not. This made the aircraft equally easy to turn to either side (much more important in a fighter than a bomber), and also made it more forgiving to fly in an engine-out situation.

Arguably, aircraft in push-pull configuration, such as the Cessna Skymaster, could be described as having this property. Although both engines are nominally similar, their mountings are sufficiently different as to induce different failure modes under different adverse circumstances. Nevertheless, the individual failure rates per engine will be comparable with conventional single- and twin-engine aircraft, and any engine failure will be treated as an emergency requiring an expeditious return to solid ground.

Answer 9

All commercial airplanes have at least two engine types. This is true even for single engine aircraft.

The first engine type is of course whatever the airplane uses. The second engine is gravity.

As part of passing airworthiness tests, all commercial airplanes must be aerodynamically stable. For an airplane to be aerodynamically stable it must have the property of having a positive static margin. Static margin is one aerodynamic measure of an airplane's stability.

In layman's term, static margin is often called "can glide". Airplane that cannot glide are simply not stable enough to be controlled by a human. Note that airplanes that cannot glide can be controlled by computers. In fact a lot of modern jet fighters are deliberately designed to be unstable (cannot glide) in order to improve areal combat performance.

Because of this, there is nothing to be gained by using different engines on an aircraft while incurring significant disadvantages in terms of engineering, maintenance, drag and fuel management.

Note: for a reference on how far commercial airliners can glide, google "Gimli glider". A typical airliner is almost as efficient as a purpose-designed glider or sailplane.

Also note: I used the word "airplane" instead of "aircraft" because this does not apply to helicopters or (in the near future) multirotors. For helicopters there is a similar reasoning due to autorotation. For multirotors most designs cannot survive total power lost so will just drop like a rock. Some designs are based on helicopter collective pitch so they can also do autorotation

Answer 10

This principle of safety through using different designs for the same goal is actually used in real planes, just not for the engines.

For some of the most critical system in a plane (like for example everything involved in making the flaps move when the pilot decides so), there is often a fully electronic system supplemented with a mechanical fallback. Redundancy is just one of multiple ways available to plane and systems manufacturers to prove that their plane is safe to the certification authorities. The norms currently used in most of the world actually define quite precisely what is expected when two systems are redundant. One such expectation is that the two systems come from two different companies and that the engineers didn't meet for lunch (I'm paraphrasing). The point is that common modes of failure caused by design defects are avoided by ensuring that the engineering of the two systems was done independently.

Safety is all about tradeoffs and experience (and sometimes, the traditions of the authority...). Not that many companies make jet engines... I can guess that engine problems almost always occur due to the environment (and thus would impact all engines in the same way) and not a fault in the way the engine was made. Engines are tested with incredible rigor. Considering all this and the logistic/piloting problems that using dissimilar engines would bring, there is little point in not using the same engines.