40
$\begingroup$

This question might come off as crude because it involves people's salaries, but I do not understand why new designs of large turbofans cost many billions of dollars to design. For example, the Pratt & Whitney geared turbofan apparently cost 10 billion USD.

Why? Last I checked, most scientists and engineers do not make millionaire salaries. I think it's more around 100 to 250k tops. Even if you had 100 of them working for 10 years on it, that would be 250M, or a quarter of a billion dollars. Instead, jet engine designs seem to cost multi-billions.

I don't understand. If it's just designing a jet engine, then raw materials cannot possibly be a major factor even if it's something like titanium or composites. How many prototypes could you possibly need? I mean I hope it's not all trial and error.

The other thing is computer software, which I thought would make things easier and cheaper to design. Granted you cannot just input the desired thrust and press a button, but surely there is decent fluid dynamic software out there that can help you design something much easier and faster than before.

So what makes it so expensive? Is there some super costly certification process? Or do you somehow need more than 100 engineers collaborating? 1,000? 10,000? If so, why? Cannot imagine that.

Maybe someone can explain the general process of jet engine design in the first place because I'm sure that would be helpful. They way I imagine it, you just go through stage by stage and try to get each blade shape and diameter right.

Or maybe I have some misunderstanding and the 10 billion includes the costs to put up the factory?

$\endgroup$
3
  • 12
    $\begingroup$ You might want to read a bit about what goes into designing the materials & structure of a turbine blade: en.wikipedia.org/wiki/Turbine_blade $\endgroup$
    – jamesqf
    Nov 12, 2019 at 18:34
  • 10
    $\begingroup$ "computer software ..." - extremely correct programs for very small market are not exactly cheap... Even standard non-customized AutoCad will run more than 1K/person/year (even with all enterprise discount I doubt you get anywhere close to 3.99 Apple AppStore prices :) ) . $\endgroup$ Nov 13, 2019 at 1:33
  • 1
    $\begingroup$ Comments have been cleaned up. Any attempt to answer the question in the comments will be deleted on sight $\endgroup$
    – Federico
    Nov 14, 2019 at 8:04

11 Answers 11

59
$\begingroup$

Even if we ignore the salaries of everyone involved; engineers, sales people, management, Q/A teams, manufacturing teams, more sales people, and then a few spare engineers...

I don't understand. If it's just designing a jet engine, then raw materials cannot possibly be a major factor even if it's something like titanium or composites.

While the raw materials alone are not necessarily costly, the materials, as well as the machining and processing, can be quite costly. This can be further compounded if the engine uses some kind of new material for which there is no effective production method yet. This was the case for the entire production of the SR-71 Blackbird which needed to figure out how to work with titanium before actually building anything. Even once the bespoke parts are produced for a new engine the company will then need to figure out how to produce the parts in sufficient quantity to produce the engines for market.

How many prototypes could you possibly need? I mean I hope it's not all trial and error.

It's not, but the FAA may require various demonstrations where they destroy the engine and you can bet the manufacturer is going to try that out before running any official tests. Once passed initial tests flying prototypes will need to be built and tested on real airframes that cost money, running jet fuel that costs money.

The other thing is computer software, which I thought would make things easier and cheaper to design. Granted you cannot just input the desired thrust and press a button, but surely there is decent fluid dynamic software out there that can help you design something much easier and faster than before.

It makes some things faster, like FEM, and surely makes complex routing easier, but like a good wrench, CAD software is a tool that makes things faster and easier. It does not do the work for you.

$\endgroup$
10
  • 2
    $\begingroup$ Numbers for fuel cost to run a modern engine at various power levels for a day of testing may be useful to reinforce this answer. - Also further consideration on costs like real estate and potential 'legacy costs' like paying out for shareholders or maintaining pension funds could be useful to highlight. Jet engine projects aren't typically done out of a lone engineer's garage after all... As they say, costs add up, and then you eventually have to pay people to add up those costs... $\endgroup$ Nov 13, 2019 at 0:11
  • 32
    $\begingroup$ Never underestimate the costs to build the tools required to build a product. Non-aviation example: the tooling required to manufacture an integrated circuit chip can cost more than a million dollars, and that's assuming you already own all of the manufacturing equipment. Any time your tests find a flaw that requires a big change, you get to pay those costs all over again. $\endgroup$
    – bta
    Nov 13, 2019 at 2:11
  • 2
    $\begingroup$ I see your 'raw materials' and I raise you titanium smithing. $\endgroup$
    – Mazura
    Nov 13, 2019 at 2:46
  • 1
    $\begingroup$ Here is a cool link where it shows how engine are tested... they're tested on a plane built for testing engines... $\endgroup$
    – Nelson
    Nov 13, 2019 at 14:30
  • 29
    $\begingroup$ Worth noting that because modern CAD makes things faster and easier, engineers are able to design more complicated things. You often don't end up spending less time designing things, you design things that just weren't feasible before. $\endgroup$
    – Seth R
    Nov 13, 2019 at 15:19
35
$\begingroup$

Jet engines are some of the most complex machines ever created. They have to be as light, efficient, safe, and reliable as possible. There's a reason that most new airliners recently have been affected by delays from engine manufacturers. This is a hard balance to engineer when on a schedule and budget.

Jet engines could certainly be cheaper to develop and buy. You can get them at relatively "affordable" prices for remote controlled aircraft. But cost certainly increases with scale, and an aircraft owner expects an engine to run for thousands of hours with minimal maintenance while burning as little fuel as possible and not hurting anyone. Each new generation of engines has been more efficient than the last, and those improvements do not come for free.

If it's just designing a jet engine, then raw materials cannot possibly be a major factor even if it's something like titanium or composites.

It's not just the raw materials, but the processing involved. Modern engines push materials to their limits and beyond. Advanced manufacturing technologies have to be developed.

Let's say you have a new material or process you want to use. It can easily take at least hundreds of thousands of dollars just to develop one, and a new engine could include many of these. Even for a cheap raw material, the amount of labor required to create test articles, set up tests, run them, and document the results grows very quickly. You want to be sure you understand how the new material or process will work before moving forward with it. If things go wrong, you create big problems for your customers (aircraft manufacturers, and their customers).

How many prototypes could you possibly need? I mean I hope it's not all trial and error.

"Trial and error" is sometimes also called "science" which is what you need to develop new technologies. Obviously as testing progresses and the risks increase you'd like the "error" part to keep decreasing. But the trial part is very important for understanding how things will actually work (or not). This means not just full scale prototypes (which will go through several design iterations, even through airplane certification) but also subsystems and components. And you need to do enough tests to have statistical confidence that the results can be reliably reproduced.

The other thing is computer software, which I thought would make things easier and cheaper to design.

This is certainly true and these technologies have decreased the amount of physical testing that has to be done. But either way it's going to cost you money.

With products like jet engines, better tools does not generally mean "how cheap can we make this process" but "how much more performance can we get for the same money."

So what makes it so expensive? Is there some super costly certification process?

Yes. People like to fly on planes with engines that keep working and don't explode. This means rigorous regulations and certification. For the FAA, 14 CFR Part 33 covers the certification requirements for jet engines, to try to make failure events as rare as possible. Here are just some of the tests required by regulations:

  • Vibration
  • Overtorque
  • Calibration
  • Endurance
  • Overtemperature
  • Full operating range
  • System and component tests
  • Rotor lock
  • Full teardown
  • Blade containment/rotor unbalance
  • Rain, hail, and bird ingestion

Some of these tests are going to be destructive, either by design or by accident. Some of them are going to take a lot of time and effort. Just the paperwork involved with understanding all these requirements and documenting to the regulators that you've met them could easily take a good chunk of your 100 people.

Maybe someone can explain the general process of jet engine design in the first place because I'm sure that would be helpful. They way I imagine it, you just go through stage by stage and try to get each blade shape and diameter right.

It sounds like you have the basic idea. But engineering is about the devil in the details.

First, modern engines could have 20 or more stages, attached to 2 or 3 separate spools. The engineers have to decide the optimum number of stages and spools for the engine design. This means analyzing many different configurations, complexity tends to increase exponentially, as each stage affects the rest of the system.

Yes, the process is relatively simple if you're given static conditions to analyze. Of course it's important to optimize fuel consumption at cruise. But the engine still has to operate across a huge range of conditions. Then there are the dynamic conditions of acceleration and deceleration. The engine has to start and be stable in both crosswinds and tailwinds. It has to be able to start on the ground or in the air after getting extremely cold. Weird things can happen as things expand and contract with temperature.

If you're looking at simple analysis of how pressure and temperature change through a jet engine, there's probably a lot of hand waving about a stage called the "combustor" where you magically get an increase in temperature. The process of burning the fuel in the extreme conditions of a jet engine is extremely complex. The air rushing in the front has to be compressed, then slow down enough to not extinguish the flame. The flame has to be contained in the combustor section throughout operation, and not overheat the turbine stages behind it.

Higher temperatures and pressures provide better efficiency but materials are pushed to their limits. New superalloys and manufacturing techniques have to be perfected to create materials able to withstand extreme temperatures while spinning at thousands of RPM. They have to put small holes and passages into the blades to force out cooling air that covers the surface of the blade so it doesn't directly contact the extremely hot air in the turbine.

Then you also have mechanical energy being extracted by a generator, and pneumatic energy being extracted for the aircraft's bleed air system. The engine has to be able to cope with varying demands of these systems.

There's also the problem of various spools rotating and thousands of RPM and not causing too much friction heat or prematurely wearing out. Engineers need to understand the temperatures, aerodynamics, and rotational stress on each part, through the whole operating range of the engine, and how it affects the rest of the engine.

And it's not just enough to get something that works. Someone will always be asking the question, "How can we make this more efficient?" Modern engines are pulling many different tricks to squeeze out every bit of efficiency that they can. Air is bled off and vanes can be adjusted to make the engine stable in all operating conditions. New concepts and technologies are developed. Modern turbofans have the problem of a low pressure turbine in the back that needs to spin as fast as possible to be efficient connected to a fan in the front that needs to spin much slower to be efficient. For the Pratt & Whitney example you give, their solution was a gearbox to allow the two to turn at different speeds. This was a very difficult challenge that took them decades to finally get into an end product.

All of this complexity has to be managed by software that monitors an array of sensors throughout the engine and continually adjusts the many parameters to maintain stable and efficient operation. This software has to run on computers that will operate across a huge range of temperatures and under constant vibration.

You also have to keep in mind how all of these thousands of parts will be manufactured and then assembled, and then maintained through the life of the engine. You need people planning to ensure that a mechanic will have access to the right components with the tools they need, and what processes have to be followed to assemble and disassemble the various parts.

Then there are also collateral effects like noise and pollution. There will be engineers tasked with understanding how these are generated and how they can be reduced to acceptable levels with as little cost as possible.

This is just an overview of the many areas involved in designing a jet engine. There are certainly more, and each detail here could easily require a specialized team working on it.

$\endgroup$
8
  • 3
    $\begingroup$ Naturally, I just calculated that we need about 1,000 of those 50 lbs thrust RC engines (5,000$ each) to make an A320 fly. :-) $\endgroup$
    – PerlDuck
    Nov 12, 2019 at 19:23
  • 6
    $\begingroup$ @PerlDuck Have you included in that calculation the higher fuel consumption of the engines, the additional weight of the engines and the additional fuel, and the higher fuel consumption due to that additional weight, and the additional weight of that fuel and the higher fuel consumption due to that additional weight...? $\endgroup$
    – Alexander
    Nov 13, 2019 at 7:46
  • 1
    $\begingroup$ This is a great overview of the complexities of a modern 'jet' engine but I didn't see anything that showed how it all added up to US $10B. Estimating the number of people required for the specifications, design and redesign, test, facilities, etc. would help this answer (for me.) $\endgroup$
    – CramerTV
    Nov 13, 2019 at 18:06
  • 2
    $\begingroup$ @Alexander Of course I have not. It was merely a joke. I was just curious how many of those little ones we would need. Made me think of a fly with its thousands of tiny eyes compared to other animals with just two larger ones. $\endgroup$
    – PerlDuck
    Nov 14, 2019 at 10:15
  • 2
    $\begingroup$ @CramerTV, number of people involved? Lots. For example, if you're using a new superalloy, you need to quantify the properties of the material: a mechanical-test technician to measure the strength, hardness, toughness, and so on, a chemist to confirm the composition, a machinist to convert bulk metal into test specimens, and a laboratory supervisor to coordinate things. That's four people for just one aspect of one part of designing a jet engine. $\endgroup$
    – Mark
    Nov 14, 2019 at 23:09
15
$\begingroup$

Last I checked, most scientists and engineers do not make millionaire salaries. I think it's more around 100 to 250k tops. Even if you had 100 of them working for 10 years on it, that would be 250M, or a quarter of a billion dollars.

The salaries in aerospace average less than 100k, it ain't IT, but they're not the issue.

You can develop a basic experimental jet engine with 100 engineers and scientists. The thing is, you can't design and build a mass-produced turbofan engine with such a team.
You can with 1,000. But airlines and authorities want it reliable, and you can't build a reliable high-bypass turbofan with only 1,000. That takes thousands because of how thoroughly everything has to be validated and double-checked.

Building a competitive airliner engine is even harder. It would be about realistic today with 10,000 employees, but still a feat. The job isn't just engine parts design, most of the hard work is in trying out hundreds of materials in R&D, machine design, technology development, QM and QC development. All the things that contribute to producing good engines, and then producing them efficiently.

Saturn, one of the smaller jet engine builders today, has ~23,000 employees.
Pratt&Whitney, the smallest of the big three in the West, has ~40,000.
Rolls-Royce, which mostly makes aerospace engines (the car brand was sold off long ago) has ~50,000 employees.
Also ~50,000 for GE Aviation, with another 200,000 in General Electric overall.

Not all of these are engineers and scientists, but more than half the staff in such high-tech industries is in research, design, engineering, management, and other jobs that contribute to design cost.

The actual design team for a modern jet engine will be under 1,000 people. But that's just the people doing the high-level work, the flow diagrams, the FEA calculations, the design models.
They'll rely on thousands to supply them with the data. From their models, thousands more will produce detailed drawings and CNC programs for each individual part. Then, for each individual part, a separate QC program has to be developed.

You can't simply copy-paste from design drawings to CNC programs. Nor can you copy-paste from those to measurement machine programs for QC. The measurement bases are different, so the tolerances are different, it's a different level of detail. Make that mistake just once. for a tiny and not especially critical part, and the consequences can be noticeable.

$\endgroup$
12
$\begingroup$

The answers are all very good in that they detail potential costs, but let me give a different angle for looking at these types of questions. In a highly competitive environment, companies will throw as much money at a problem as it is worth for them to have it solved. In economic speak: "marginal cost equals marginal gain".


When designing a new engine, one starts with all the changes that give a large performance improvement at a low cost. Over time, those "trivial" changes are explored, and if there is enough "gain" from continuous attack at the problem, more complicated changes with smaller expected gains will be attacked.

Now, think about how high the gain from an improvement of an engine is: Over the many thousand hours, how much fuel is being saved? What is the expected future market value of that fuel in a world with increasing scarcity and expected co2 taxation?

Now, consider that you will implement this new engine not in a single plane, but a large fleet of hundreds, perhaps thousands of planes? Any improvement to the engine that you make has just such a large market value. Finally, consider that many improvements to the next generation of engines can be taken over to newer developments later on, something that is referred to as "standing on the shoulder of giants".


An example Take an engine that costs around 30 mio USD. A performance improvement that increases the value of each engine by only 1% will be worth 300 mio USD if that engine is sold 1000 times. If that performance improvement can be reused in the next 10 generations of engines, it is worth 3 billion USD. This simple example shows you that the marginal value of R&D can get really high very quickly, and that the companies are therefore willing to throw a lot of money at these problems.

$\endgroup$
8
$\begingroup$

Adding to other excellent answers, I'd like to focus on the nature of the research.

The work involved in developing and exploring ideas that aren't just capable of resolving via computer models, are huge.

As other answers note, jet engines develop at the leading edge of theory and new ideas,as well as pushing existing ones.

Concrete typical example #1

Suppose we believe that a fan blade can be made stronger if it's cast in a way that it grows without certain crystalline flaws, or with a certain crystalline structure that in theory should be possible. Call it "delta form titanium-carbon crystalline matrix", or "delta-TCCM" for short. This would allow 1.7% thinner and lighter blades, without loss of strength or safety, or blades that can run 1.5% faster without increase in stress. If correct, this could be a big deal as part of the next generation of the current engine.

The problem is, that's as far as a model takes you. Now you need to actually achieve it reliably as a material science problem. You need to

  • Design a process to develop delta-TCCM reliably in a lab, which may be a huge challenge. You may need to explore multiple techniques, consider how they scale, their susceptibility to flaws and risks. The conditions for reliable low-flaw-rate delta-TCCM production may be very precise and hard to maintain for the time it takes. This can be a huge problem, far from trivial. If you don't want it to take years, you might need to throw 600 people just at the delta-TCCM research, to turn it from a concept to a usable material with verified properties.

  • The properties may only be somewhat predictable by theory. You may need to add trace amounts or tiny process changes, on gut feel, to resolve the issues. Each of these is a mini project in itself.

  • The material may be difficult to shape once formed, so you may need to go back to your lab to not only devise equipment to reliably create it, but to reliably create it to shape. Perfect shape.

  • You need to scale up from laboratory to industry scale. Meaning, to create enough of it to confirm properties, and to ultimately build blades. That's also far from trivial. Industry is littered with things that are easy to create in tiny quantities for research, but incredibly hard to produce at scale to the same conditions. Going reliably from 2mm2 samples without detectable crystalline/atomic-structure flaws, to curved 1.5m fan blades without detectable crystalline/atomic-structure flaws, is exactly as difficult as it sounds in many cases.

  • You need to test and assess ten thousand samples in a thousand ways - in isolation and in a thousand scenarios in an engine. This is a very intense process. What is its atomic structure, how does it fail (what are its modes of failure and safe limits), how does its atomic structure respond to a hundred thousand combinations/kinds/patterns of stressors, both short and long term - gaining enough understanding of the actual properties to be able to rely on them for jet safety. Perhaps go back to basics if something isn't as needed. After all, if just one engine fails and the flaw is traced to a fundamental issue with the material, your entire reputation and product range is at risk, and refunds will be due for any sold so far, plus litigation. Your entire $500 bn business could be put at risk, in that sense, in the worst case.

  • You could well also build 2 or 3 entire prototype production plants (factories) in different locations, just for delta-TCCM, to confirm that you can in fact reproduce your delta-TCCM quality control reliably over time and in different establishments/sources.

  • Fan blades are usually made of a combination of different materials. For example the GE-9X, currently the largest turbofan engine made, uses a carbon fibre composite with steel leading edges and glass-fibre trailing edges for bird strike protection. Just making and shaping delta-TCCM isn't enough, it also needs techniques that reliably allow it to be part of a composite blade, tightly enough to retain its unified structure under all the stresses, and heating/cooling cycles, and vibrations, that are part of jet blade life. If the components don't move, shrink, and expand together, the blade may eventually weaken.

  • If it works, you may need to build an entire toolchain just for delta-TCCM. Machining tools, production tools, blade casts (maybe they're destructively cast and you need a new mold for every part), specialist laser or other welding, development of delta-TCCM coatings and adhesives capable of enduring in an engine environment, which are all their own independent projects. The works.

And that's just the project to commercialise delta-TCCM. $20-50 million easily, off the top (total guesswork on my part, but gives an idea). You might have 50 or 200 such projects in progress and others coming up in your R&D flow, all related to concepts you're going to explore for your new generation of engines - and all of their costs needing to be recovered by sales of the engine when finally complete.

Concrete example 2:

The space shuttle had to resist intense heat on re-entry. Far more heat than any material could possibly withstand. The idea reached was simple: ablation. The coating would burn away rather than melt, gradually exposing layers below, but not degrading as a whole.

Creating the material was a major effort. There wasn't much of a theory of such things, just a goal to create such a material. Huge research. And every time, "Good. Now find a way to do the same but making it 20% less weight".

$\endgroup$
1
  • $\begingroup$ Add to that the cost of all the ideas that sounded great until they hit a setback in one of those steps that just couldn't be overcome. All that time, money & effort for a shelved project and time to start over on a new one. $\endgroup$
    – FreeMan
    Nov 15, 2019 at 17:04
5
$\begingroup$

In addition to the other answers:

Jet engines are not only complex, they operate on the edge of what's physically possible. For example, modern jet engines run at internal temperatures that can be higher than the melting point of the metals used.

When you design a new jet engine, in order for it to be successful in the market, it has to be better than the engines currently available: it has to have more thrust, lower noise levels, lower fuel consumption, higher reliability, lower running costs or a combination thereof.

This means that every design moves "the edge of what's physically possible", i.e. it advances the state of the art. It's not just a new engine design, you have to develop new materials, new construction methods etc. Then you have to prove that these new developments of yours are safe to use. This is where the cost goes: scientific research (which always carries the risk that your new idea won't work as well as you'd hoped), development of the new technology to a consumer-ready level, and certification.

$\endgroup$
4
$\begingroup$

The problem is not fundamentally about jet engines, but about building complex things in general.

The reasons are the same as for building a complex software. There are only gradual differences.

The question can be seen as "Why does it cost surprisingly much to create complex systems of high quality?"

The main issue is complexity. The design of existing jet engines is complex, and we know that designing an alternative is a process complex more complex than that. The same again for actually building one in series.

We want to create a complex artifact, let's name it "new jet engine".

To do this,

we need a design for it.

As base of that, we need a design specification.

To verify it, we need to build at least one instance.

In practice, we want to be actually able to create multiple exemplars for limited cost per instance.

That means we need to create multiple other artifacts too:

We need to create one or more prototypes without limiting the cost.

We need to create a complete set of tools to produce multiple instances of the artifact.

We also need to build tools to test the artifact.

We need to test one or more prototypes, and multiple produced instances based on the design specification.

We need to make an external organisation test the design specification based on certification rules.

We need to make an external organisation test instances based on certification rules.

We need to create documentation, including reliable instructions for multiple variants of service.

Note that all this is independent of the complexity of the thing we want to create. It not even depends on whether we build a physical artifact, it applies just the same for building a simulation of it, producing instances by integrating it in airplane simulation of the customer.

The many of the steps are somewhat complex in itself. When steps interact, the complexity tends to multiply instead of adding up. For example, a minor error in the design specification causes minor changes in most of the steps, and each of them has a significant overhead. Changing one screw size and the strength of one weld requires practically the same effort as only changing the screw size, because the overheads are dominating.

If we are building something complex, there are some counter intuitive aspects in terms of complexity. An important one is that the complexity and effort of testing increases very quickly for increasing quality requirements. That is in part because there are many more smaller errors than larger. It means that many more single errors need to be handled, requiring more prototypes. The overhead for handling a small error is about the same as for a large error.

To illustrate the effect of increasing quality requirements, think about building an airplane based on a plan specifying the shape and size of its parts. Compare that to an additional requirement of the total length with tolerance of a few centimeters. Now, you need to take the variation of component connections into accounts, like the distance of screws to edges of parts, and also the thermal expansion of parts. Now, refine the requirements to specify the length with a tolerance of some millimeters according to a temperature curve. Now, some tests need to be done multiple times, after finding out how many times are enough. And the differences in thermal expansion of different materials and parts from different suppliers becomes relevant. You get the point. And just in case it seems irrelevant to care about thermal expansion: The Lockheed SR-71 Blackbird actually leaked fuel when cold on ground, but did not when flying at at Mach 3.2 and about 300 °C hull temperature, based on accepted limits of precision. The Concorde became 17 cm longer in flight at about 100 °C. They had great fun when placing hydraulic lines.

Basically, adding individual parts is much more complex than the intuition would expect. Adding a part to a jet engine does not only involve stability of the part, but stability and change in shape over heat cycling, and determination of acceptable number of cycles before service.

Note that all this, apart from illustrating examples, has nothing to do with jet engines, not even whether we want to build something physical.

The specific design elements for a jet engine can be found in other answers, and a complexity estimate of them can be used to derive the over all effort here.

$\endgroup$
3
$\begingroup$

I think most of the answers address the points quite nicely, the teams are huge and there is a lot of expensive kit involved. I'd add three more points:

  • There is risk involved that needs to be priced in. It's not like in Pharma, but not all engines sell equally well, so you need to manage the costs across different engines and designs.

  • These are highly specialised machines, so together with a new engine your developing new tooling, new measurement techniques and new software. (There are lot's of spin offs and resultant benefits from these programs e.g.: touch trigger probe)

  • Just to illustrate the point of materials and manufacturing costs, these engines would be cheaper if made out of solid gold.

I happen to know the guy that designs the profile for the fan for one of the large manufacturers. He is just the academic that contributes to that design, and that is the only problem he works on. But that involved developing new software to compute the flow.

$\endgroup$
1
$\begingroup$

The existing answers do a great job at answering why jet engines are expensive to develop: because they are so complicated. Let me try to answer, why are jet engines so complicated? To understand that, we have to examine the economics of jet engines, and it comes down to fuel efficiency.

Let's say that you go shopping for a new car, and one car gets 1% better gas mileage that the other. Maybe 30 mpg and 30.3 mpg. You'd say those are so close that it barely matters. Might as well be identical and you start to look at which one has the better sound system or the most stylish seats. But when the airlines go shopping for new planes, 1% difference fuel efficiency is HUGE.

Airlines are large, and jet fuel is expensive. A modest sized airline (say Jet Blue size) will spend \$1 - 2 billion per year on jet fuel alone. And, when you buy a new plane, it will generally last for 30 years. So over the lifetime of the fleet, the airline is spending something like \$45 billion in fuel. If one jet engine is 1% worse in fuel efficiency, that's going to cost the airline ~ $450 million over 30 years. That's for one percent difference in fuel efficiency.

Now there's other things that might make up for that, like the purchase price of the engine, the cost of service and spare parts, etc. So a jet engine that's 1% worse in fuel efficiency might still be competitive overall if it makes up for it in other areas. But beyond a few percent, the difference is so great that you couldn't even give them away.

So what you end up with is this intense "arms race" competition between the major jet engine OEMS. One company makes their engine a little bit more complicated such that they can improve fuel efficiency by a tiny fraction, and then all of the others race to catch up. This continues year after year, engine model after engine model, and before you know it what started as a fairly simple machine has been extremely complicated, and thus expensive to make.

$\endgroup$
4
  • $\begingroup$ A modest sized airline (say Jet Blue size) will spend $1 - 2 billion per year on jet fuel alone. ...Billion? Citation needed. A little math refutes this. A budget airplane, maybe 4 flights per day, 150 ppl per flight, 60 bucks per ticket. The gross revenue of that is 13.14 million per year. No way it could ever approach 1 billion. Did you mean 1 - 2 million? $\endgroup$
    – DrZ214
    Nov 15, 2019 at 12:11
  • $\begingroup$ @DrZ214 no I meant billion. See for example: businessinsider.com/… key quote "...fuel and related taxes rose to \$515 million for the quarter". So \$500m / quarter is 2 billion per year. Also, Jet Blue is definitely more than 4 flights per day, it's more like 1000. Maybe you are confusing them with a different airline? $\endgroup$
    – Daniel K
    Nov 15, 2019 at 12:27
  • $\begingroup$ Oops, I read "airliner" and saw it as airplane, as in 1 single plane. Yeah the entire airliner could have thousands of flights per day. I'll leave these comments here in case others misread it. $\endgroup$
    – DrZ214
    Nov 15, 2019 at 12:44
  • $\begingroup$ @DrZ214 I quoted it that way because airlines don't just buy one engine at a time. When they make a buying decision, they are usually committing to 10s if not 100s of engines at a time. So the fuel costs involved in a single buying decision are large. $\endgroup$
    – Daniel K
    Nov 15, 2019 at 23:47
0
$\begingroup$

This is only part of the answer, but I didn't want to post it in a comment after the moderators said to keep answers out of the comments:

Look at the accounting concept of a "burdened rate." It helps to remember that you don't just need 100 engineers collaborating. You need 100 engineers in a building that has to keep the lights on and heat and/or air conditioning, with janitors and admin assistants, and all the other wonderful people who keep the engineers productive. When you factor in all these other costs of business, the hourly rate the company has to pay (as opposed to what the engineer receives) is quite different.

$\endgroup$
0
$\begingroup$

This is a great question! You hear this kind of question a lot on different subjects. the closer you look, the more you see the answer.

There are some great comments here already, so I'll just add a couple of points.

If you read through the design of the GE9X (here), there are pointers to the huge design cost.

There's no point in doing a redesign unless it's better. In this case, 10% more fuel efficient. This requires new knowledge, new materials, new design...

It took 8 years for deign/engineering from start to certification. They will need to run that team some years after it goers full-production to cope with any issues, which are bound to arise.

So, your core design team (could be 100 people) is looking at 10 years work. Capital equipment and support will add a lot in that time.

This will quickly swallow up $150m in-house. That's pay, not cost. Probably add 50% to house and feed, medical etc them. Each of those engineers will spend money on their part of the project. Some of them will spend a lot of money. Add on external contractors, and that could quickly looks like $$1billion.

They're using new materials which all require learning and certification. Certification is very expensive. And that's for each part!

Read the article and you'll see a single stator vane failure took 3 months to redesign, required 4 more test engines. Even at production cost, those are 45 million USD each. You can't know these things in advance. They will be built in to contingency budgets.

GE had to modify a 747 to test the engine as it didn't have the ground clearance. The had to produce 8 test engines to get full certification. That must be $800 million.

When you start adding it all up, that 10 years work is quite easily explained and divide the design cost across 10 years and it doesn't look so huge.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .