Obama threatens to veto defense bill over F-22/F-35 engine

[erik_t]

Uhhh... because our predictive capabilities in high temperature materials, heat transfer and reacting fluid flows is a bit more advanced than it was 40 years ago? What part of this isn't making sense to you?

Most of that advancement seems to be applied to pushing the limits of performance, not making existing engines more reliable (at least, in military applications). I'm sure if the F-135 was designed for the same thrust rating as the F110 it could be made much more reliable, but it puts out about 50% more thrust in the CTOL version, plus there's the additional complexities of the VTOL version.

I strongly disagree with this statement. The most visible changes have been to performance, yes. However, the ability to predict performance a priori is the thing that renders this conversation vastly different from F100/F110. Forget forty years ago, let's just consider ten years ago. In 1989, this (google books warning) was a fairly substantial calculation. 70,000 cells, non-chemistry, 10 hours on a Cray. This calculation would have been heroic in 1985, and intractable in 1980. I've got a 22 million cell calculation, with chemical nonequilibrium, running at work right now; it's taking up about a tenth of one of our machines. At a university, not some crazy NSF supercluster. Finite element analysis has undergone a similar revolution.

Even in 1980, most design work was done with experimentally derived approximate laws, often without an intrinsic understanding of the underlying physics. We can directly model an entire engine if we so desire, from first principles, without ever cutting metal, and we can have remarkably high confidence and small error bounds. Suggesting that design failures of forty years ago are directly applicable to today's industry is simply not an informed opinion.

[raptor3x]

Even in 1980, most design work was done with experimentally derived approximate laws, often without an intrinsic understanding of the underlying physics. We can directly model an entire engine if we so desire, from first principles, without ever cutting metal, and we can have remarkably high confidence and small error bounds. Suggesting that design failures of forty years ago are directly applicable to today's industry is simply not an informed opinion.

Realistically, much of engine design is still done this way; in industry using physics based methods is still a little ways into the future, especially for turbine durability design.

[Dendrobius]

There's no such thing as a composite turbine blade, unless you're talking about a kazoo.

Composite might be the wrong word, but the blades aren’t solid cast pieces of machined metal , it’s a bunch of powder baked together. The plane has already been significantly delayed by F-135 engine problems too.

I cannot speak for every turbine blade currently manufactured as there are a lot of different casting techniques employed but I can tell you for sure that the blades and vanes of the high turbine on the f135 are indeed single cast pieces of metal, that are later machined to create the fir tree structure and film cooling holes.

I think what Sea Skimmer's trying to say is sintering as opposed to composites. He's got the description down right for that at least. Either that, or he was thinking of carbon composite fan blades for turbofan engines, and confusing their application from the fan stage to the compressor or turbine stages.

If he's talking about sintering, well, to my knowlege there has been/is some experimentation with compressor blades (I THINK low pressure) being manufactured by hot or cold isostatic pressing and sintering technique, but I didn't think that was in any production engine in the world at this point in time. IIRC most HP turbine/compressor blades are cast monocrystalline superalloys which are manufactured via investment casting.

The main engine for F-35 uses completely new technology including composite turbine blades, not to mention the highest compression ratio ever in a service jet engine

Well, first point's been addressed, there are no composite OR sintered turbine blades in the F135. As for the compression ratio, overall pressure ratio for the F135 is said to be 30:1...a Rolls Royce Trent 900 as used in the A380 is at a MASSIVE 39:1. A GE CF6-50, powering the "advanced" DC-10-30 of 1969, has an overall pressure ratio of...wait for it...29.3:1 Even if you talk only of military service jet engine, the 1970s vintage F100-PW-229 has an overall pressure ratio of 32:1.

Just thought these points were worth highlighting for everybody.

[raptor3x]

[
Well, first point's been addressed, there are no composite OR sintered turbine blades in the F135. As for the compression ratio, overall pressure ratio for the F135 is said to be 30:1...a Rolls Royce Trent 900 as used in the A380 is at a MASSIVE 39:1. A GE CF6-50, powering the "advanced" DC-10-30 of 1969, has an overall pressure ratio of...wait for it...29.3:1 Even if you talk only of military service jet engine, the 1970s vintage F100-PW-229 has an overall pressure ratio of 32:1.

Just thought these points were worth highlighting for everybody.

Thank you for pointing this out, I thought that claim was also dubious but, as I've pointed out in the past, my knowledge is mostly focused around heat transfer augmentation in turbines so I didn't want to make any claims.

[erik_t]

Realistically, much of engine design is still done this way; in industry using physics based methods is still a little ways into the future, especially for turbine durability design.

Think so? All my work at Honeywell and ATK was CFD/FEA. Maybe very very preliminary sizing is done with historical methods, but I didn't see any of them in detailed design work. It doesn't really change my point, though, that predictive capabilities are far in advance of where they were decades ago. You're unlikely to have some component fail fleetwide at 20% of the expected service life.

[raptor3x]

Realistically, much of engine design is still done this way; in industry using physics based methods is still a little ways into the future, especially for turbine durability design.

Think so? All my work at Honeywell and ATK was CFD/FEA. Maybe very very preliminary sizing is done with historical methods, but I didn't see any of them in detailed design work. It doesn't really change my point, though, that predictive capabilities are far in advance of where they were decades ago. You're unlikely to have some component fail fleetwide at 20% of the expected service life.

I guess I should clarify; structures, heat conduction, and external flow are done with physics based methods in my experience. The areas that are still dependent on historical methods seem to be internal flows and anything involving conjugate heat transfer; primarily due to the massive turnover time necessary for a single iteration. Companies very much would like to move to fully physics based methods but for these types of simulations where one design iteration takes ~1 week to calculate using around 50 cpus, it's not practical just yet when historical methods can give fairly good answers in less than a minute on a business grade laptop.