Army Dissects Black Hawk Helicopter, Scans Parts for 3D Printing | Hacker News Reader

Army Dissects Black Hawk Helicopter, Scans Parts for 3D Printing

breakingdefense.com

69 pointsdfsegoat66 comments

Media loves talking about 3D Printing primarily due to its click ratio and viewership. As an mechanical engineer, I've had experience developing aircraft stages and frames - the whole premise of 3D printed parts going into an aircraft from airworthiness perspective is flat out insane.

I believe you, but please go on and tell us why

This was 10+ years ago but the process is roughly as follows: Requirements of loads and flight envelopes are first provided to us (I was employed in a stress analysis group). The parts that I worked on weren't "new" per se, these were modifications to existing parts to comply to new loads or new materials, fasteners etc.

Once envelopes + safety factors are set, we literally sketch out basic FBD (free body diagrams) along with shear-moment diagrams depending on the type of the part. You know the kind of diagrams you had to do in Statics class in University. Most of us knew the appendix formulas and many other empirical results by heart, and some of the guys at work were total nerds. They knew every possible cross sectional moment of inertia formulas, etc. It was really humbling. You'd think that aircraft industry some slow moving folks, but it couldn't be more false. The discipline was astounding. Then we consult existing parts directory, fastener books and materials for conceptual design. During the first conceptual design review (CDR), Principal engineers (depending on how critical this part is) get involved and provide us feedback.

Post CDR, we send out parts to our FEA dept (mostly part time workers, some of them even commercial pilots) that get the models from the drafting department (Catia V4 at the time) and run some discretization scripts, they are pretty versed on how fine the mesh should be, but everything must be documented. Sensitivity study against mesh size needs to be proven. Converged models are sent back to us, we test it against intuition, and theoretical predictions (but not always).

Detail design: Flange thickeness to thread pitch is set in stone, about 1800 fastener choices are narrowed down to a few, cost implications are non-existent, so most of the times weight is king. Coatings if any are specified. Prototypes are made by our supplier, GD&T red marks are fixed. Goodman diagram is consulted for fatigue studies, and sometimes even simulated in the FEA model. Same with creep and crack propagation studies. Detail Design review takes place and we are blessed by principal engineer. If it is an external part with any sort of interaction with the airflow, flutter dynamics to CFD guys are called, I honestly never worked on any farrings or anything affecting flight dynamics.

Then comes the guaging (where to put rosette and stress gauges) and ground test certification process. Prototypes are tested against real world conditions and failures. Depending on the part, various tensile strength tests are conducted and sometimes with simulated internal cabin pressure. Not all parts were tested though. All in all, it takes forever to make a part in an airplane. My memory is fading so I can't recall everything. It was a lot of paperwork and documentation. We took that seriously.

Having designed parts in an airplane, I just can't take 3D printing seriously. You wouldn't believe the kind of work that goes into designing parts from composites - that was a new thing at the time when I was in the industry.

Consumer 3D printing results in bathtub toys.

An Apache helicopter is not made of bathtub toys.

Metal is awesome compared to filaments and composites:

Mike Busch - "Making Metal Behave" https://www.youtube.com/watch?v=m1jl6IlywAk

I don't think you have answered the question even remotely. What about 3d printing prohibits using parts in aircrafts?

Are you aware "3d printing" can produce many types of materials, including metal, and not just plastic?

It’s been used for several rocket engines

Interesting. Are they man-rated engines? Which parts? Since rocket engines tend to be single-use only, fatigue strength wouldn't be much of an issue.

SuperDraco on SpaceX’s crew dragon capsule aren’t used in the launch, but it’s the launch escape system. [1]

SpaceX also apparently prints the main oxidizer valve in Falcon 9’s Merlin 1D, which isn’t a single use vehicle, but who knows what components get replaced between launches. [2]

As of 2016, the Raptor engine being developed for ITS (now Starship) was 40% printed by mass, but I have no idea how much of that was for faster prototyping and if they might be using less in later production. [3]

Rocket Lab’s Rutherford engine for their Electron rocket. [4]

More in the works from Relativity Space, Launcher, NASA, and others I’m sure. We’ll see what pans out.

[1] https://en.wikipedia.org/wiki/SuperDraco

[2] https://3dprint.com/10830/spacex-3d-print-oxidizer-valve/

[3] https://www.nasaspaceflight.com/2016/10/its-propulsion-evolu...

[4] https://en.wikipedia.org/wiki/Rutherford_(rocket_engine)

The block diagram of the Raptor engine is just sick - it's brilliantly simple! I suspect that parts of it are 3D printed because of the difficulties in machining the shapes necessary. That may make it worthwhile to use lower-strength 3D technology and compensate by having more weight, but I'm just speculating.

I'd love to just set one of those engines on a bench and take it apart.

It's interesting to see von Braun's turbo-pump innovation is still the go-to technology after nearly 90 years. Also the fuel pre-heater used as the nozzle cooler.

It's not the first full-flow staged combustion engine ever built, but it is the first to get off the test stand and fly.

Looks a lot more complicated in real life compared to the block diagram https://en.wikipedia.org/wiki/SpaceX_Raptor#/media/File:Spac...

And the Super Heavy booster is supposed to have at least 28 of those. At one point it was 37.

> a lot more complicated in real life compared to the block diagram

Sure, jet engines are like that, too. Steam locomotives, too, they are festooned with pipes and things. I once spent some time at a museum tracing the tubes trying to discern their purpose. (I've seen many engines in a museum, simply none of them had any sort of detailed explanation of what you're looking at. Too bad, so sad. Just a couple sentences "This is an XYX motor from a GHZ." Making it boring.)

It would be fun to see what all the tubes on the raptor engine are for.

Thanks for your co tribute is to this discussion, it is a well reasoned opinion.

In response to your question, there are a few articles easily found including a wiki page in 3D printed space craft. No manned rocket as of yet, but orbit has been achieved.

https://additivemanufacturingtoday.com/spacex-uses-dmls-to-3...

The issue is going to be whether the 3d printed parts can fit the requirements for the conditions the prior parts were used in. Can laser sintering produce a bolt or pin that can withstand the same forces as other manufacturing techniques? Or will there be weaknesses in it. There are some parts that shouldn't be a problem, but many others where serious study is going to be required to verify that the aircraft won't break in a catastrophic way while in flight.

As a non-3d printing example: A drone aircraft used a COTS product as its drive shaft. The same part was bought for years, no problem. Suddenly the drones started literally falling from the sky. Turned out the COTS product had changed manufacturing source and the quality of the steel was inferior, leading to frequent failures.

3d printed parts need to be reliable enough to replace things like that, or we need to tone down our enthusiasm and realize most of the 3d printed parts are going to be door handles and similar non-stressed or non-critical parts.

Composites are the way to go in Airframes, the amount of effort poured into this area is incredible. I think the new 787 (haha "new") is primarily composite and so is the A350.

> Composites are the way to go in Airframes

Well, not really - that's mostly an HN fanboi comment.

- composites are often the same weight as aluminum when all is said and done

- composites normally are only good for subsonic flight because of skin heating

- large-area composites require multiple large autoclaves, etc. when normally there's only one.

- composites are more difficult to repair than metal. If a baggage cart hits your composite fuselage or wing, how do you fix that? How long will that take?

Metal is high-tech to me, not composites.

Wouldn't 3d printing potentially make it easier to flatten the supply chain and help guarantee predictability?

I haven't but I guess the goal was to give a perspective of what it takes to design a normal part from 6000 series aluminum and see what has to happen to certify it.

You can extrapolate from there. We didn't have 3D printing as an option back then and I don't think it makes any sense even today. Sure, for prototyping it is valuable but even then we would make prototypes from the actual process (CNC) because then we can test the mechanical properties.

Most people think prototyping is holding the part in the hand and seeing how it looks. Perhaps if youre doing ergo work.

If 3D printing can go through rigorous process of understanding variation, void forming, failure modes, repeatability and reproducibility tests, sure it can be used. My question is - why? Perhaps a part can be made so that it saves so much weight with its shape that is not easily machinable. Then it makes sense. There are so many hurdles.

In some ways, 3D printed parts are considered "composite". It would need to go through some rigor as some of the theory/proceedings laid out in a book like this: https://www.amazon.com/ICAF-2019-Manufacturing-Proceedings-I...

It is very much under study and scrutiny.

I suspect 3D printing aircraft parts becomes more viable my the day, especially as the materials, accuracy and reliability improve. I can appreciate that not _everything_ is a candidate for some form of printing, but I can imagine quite a few parts are. Some examples:

1. Highly specialized missions that require custom parts. The mission parameters themselves might be quite volatile, imagine some engineering team cooking up some mission critical sensors at the last moment to be used for this mission only.

2. Small incremental improvements that can be rolled out fleet wide. Usually these aircraft will have at least _some_ issues with the hardware, and the military answer is typically to write more SOPs. Instead they could 3D print a clip to stop that darned wire hitting the hot pipes when you open the panel.

3. Broken parts in remote locations could potentially be fixed whilst in those locations. You don't have to rely on some distribution network to fix your aircraft when you can build a patch and get it through one or more trips safely until a proper service.

4. The ability to semi-break away from the need of manufacturing.

You’re not even wrong. Parent is explaining past the level you’re working at. Closest you can get is their mention of composites. I’ll flip the script: show us some 3D printers that can output specific phases of steel, for example, and then we can talk.

Hey, I would also like to learn if these things exist and whether anyone in aviation is using them. More curious about how they would certify these things - machining is so well understood over many decades that the risk tolerance against anything else would pretty high.

Weight is the implacable enemy in aircraft design. This means the minimum amount of material for each part. This means the strongest / weight materials, and manufacturing methods that reliably achieve it.

For metal parts, this means forgings, which are 3x stronger than castings. The forging process is also pretty reliable compared to castings. Forging is also expensive, the only thing more expensive is a hogout from a billet.

In order to be usable for critical airplane parts, a 3D printed part would have to have the strength and reliability of forgings - and of course be cheaper.

BTW, on my hot rod engine I replaced all the moving parts in it with forgings, and the bell housing is a steel forging, too (stock is cast aluminum), because I like my feet.

Did a bit of looking at the USAF’s interest in this, it sounds like it’s not about structurally critical parts so much as avoiding supply chain problems on obscure little bits. They recently made a metal anti-ice gasket for a turbofan engine: https://www.flightglobal.com/fixed-wing/us-air-force-produce...

For non-metal parts, this looks cool too: https://www.sciencedirect.com/science/article/abs/pii/S13598...

I think you're too dismissive. I work in aerospace now. For the majority of parts, 3D printing is not the best process, whether it's due to strength or tolerance requirements, a standard profile or sheet being more suitable, and most importantly it often not being the most time or cost effective method. But some parts are well-suited to being 3D printed, and all of the design processes, review, analysis, testing and documentation you mentioned can be (and is) applied to 3D printed parts just like any other.

A few examples I've seen:

- 3D printed engineering plastic to transition honeycomb in a composite panel to another shape

- 3D printed engineering plastic housings for connectors, etc

- 3D printed titanium and aluminum component mounts. Designed using topology optimization, and therefore not really achievable using machining. Really useful when you've got a part mounted on an fixed interface plane, and after a few design changes, the part needs to be mounted 50cm away in a mass-efficient way.

- 3D printed blanks which then go to final machining. For metals this is not much different than casting + machining, but better suited when you need 10 parts rather than 1,000 and a gamechanger when you need 100 slightly different version of a part

- Jigs, moulds, test fixtures, prototype parts. 3D printing is a gamechanger here.

The problem is in the media, and often times in the maker community and therefore general public, "3D printing" is described as the one process that will replace all others. You don't need to take this notion of 3D-print-everything seriously, but 3D printing is here to stay in virtually every industry including aerospace.

I should note that I work in space, not aviation. Right now we are seeing more 3D printed parts on spacecraft every year. Most spacecraft are made in quantities of 1-2. Most aircraft are made in much larger quantities and so can benefit from all kinds of mass manufacturing techniques, meaning 3D printing is probably suitable less often. But not never. And a quick google search shows Boeing and Airbus are already using 3D printed parts for passenger flight today.

Curious how you do quality control and machine-to-machine variation studies? Do you have CSAM inspection or some kind of X-ray inspection techniques for finding voids/abnormalities?

That was really interesting to read. I shouldn't be so dismissive, sounding like an old record. I would however be dismissive if proper analysis isn't done on the parts and generally deteriorating standards on quality of parts. People's lives are at stake and we were very conservative in building the frames I worked on. Not just FAA but other agencies were watching us.

X-ray inspection is indeed done for critical parts. I found reference to cosmetic interior 3D printed parts being used in production to test the waters.

I don't work in anything related to human safety, and I think very few parts on an airframe would be suited to 3D printing (surely something though). If you're coming from that background it makes sense to be very skeptical. But as another commenter mentioned, some aircraft engine parts are already 3D printed.

The article does cover that, in a very brief manner:

> So far, he said, they’ve evaluated 31 percent of the parts – over 30,000 items – and found 252 that are “potential candidates for additive manufacturing.” That’s less than one percent.

I think that was for a prototype for the Lower Tier Air and Missile Defense Sensor radar.

For the helicopter: “We expect to probably get about 20,000 structural parts, -ish, out of that,” Royar said. “We’ll take a look at it and evaluate every one” to see whether it can be safely and economically replaced with a 3D-printed version.

So I am guessing they might find say 100 suitable non-critical parts (say plastic covers). 3D scans of critical parts could still be very useful for rigs, emergency field use, or other non-flight purposes.

> [GP]: the whole premise of 3D printed parts going into an aircraft from airworthiness perspective is flat out insane.

The article responds that observation very well IMHO.

I was a structural engineer for Sikorsky. I think it's obvious we will not be replacing a machined titanium main rotor hub with a 3D printed part soon. The article touches on this. It will probably be non-structural components like buttons on an MFD.

The super draco combustion chambers made by space x are 3d printed.

I'm not saying they're the best fit for everything, but saying they'll be relegated to button on an MFD is probably too far on the other end of the spectrum.

Form is function in high performance manufacturing (including most of aerospace). Just switch materials is a huge deal. If you set out with "use additive manufacturing" as one of your requirements/design goals, you can probably use additive manufacturing in a lot of aero. Using additive manufacturing to drop in replace existing parts that were made with different materials and techniques is very unlikely however.

Oh absolutely. There's no way you can just drop a scan in to solid works, send it to the printer and call it a day. Some serious engineering needs to occur to change manufacturing processes.

That's probably not a deal breaker in the grand scheme of things is all I'm saying.

I was going to write a reply saying that I don't see why we can't use it in many places as long as we understand the parameters of the product, but as so often happens I found an excellent source that actually explains your stance in details.[1] The TL;DR is that there are lots of imperfections in 3D printed parts, which means they typically fail at much lower stresses. It actually includes some pretty good video illustrating this at a microscopic level.

That said, there are apparently new techniques being researched and discovered that have promise to increase the quality of 3D printed parts. One way to thing of it is that we have hundreds to thousands of years of experience in material science with forming metals in other manners, and we've just started with additive engineering, so there's a lot of room for improvement.

1: https://www.youtube.com/watch?v=fzBRYsiyxjI

I know a bit about turbofan engines in aircraft.

3D-printed parts are already going into them right now, as we speak: https://www.ge.com/news/reports/airbus-gets-1st-production-j...

Also, to elaborate on which parts can be 3D-printed. The latest and greatest breakthrough is using EBM (electron beam melting) for 3D printing second stage turbine blades from titanium aluminide in GE9X engine.

Those parts are continuously experiencing mechanical and thermal loads far in excess of forged or machined parts in any airframe. There is no added safety barrier for them unlike for fan blades. If turbine stage flies apart, its parts will go straight through everything in their path, including passengers.

To further explain my point but also support the root comment.

Aluminium/steel/titanium airframe parts are usualy machined or forged. It is not possible to replace them with 3D printed parts just yet.

But! All of jet engine turbine blades are cast, not machined. Those parts can be safely 3D printed, as various metal sintering/remelting methods produce similar or better structural quality. Same goes for those magnesium castings used in aircraft landing gear - can be replaced with 3D printed ones.

Parts that only under hydrostatic load like fuel nozzles can be 3D printed. Fatigue is not an issue there.

Parts that do not experience significant structural loads, but high thermal loads (combustors) can be also 3D printed.

I believe that's a better answer to the topmost comment.

I also worked on Turbine blades. Specifically did studies on trailing edge dump and pressure side bleeding holes, the air is bled from 7-10th compressor stage for cooling.

3D printing would massively benefit it because more than structural loads, the part is at its thermal limits and structurally it becomes weak as a result. If they can just cool it by 200 degrees, it would be groundbreaking. The trailing edge of the blade is by far the most critical - thicker and aerodynamic efficiency will take a toll; make it razor sharp and it will just melt away.

I think industry is moving to CMC materials for first stage blades, because nickel superalloys hit the limit of performance. Not sure if those can be 3D printed reliably.

That's insane. During my tenure in the industry, we had ceramic coatings on the first stage. We were hitting some 1250 C temperatures with these things and they were barely holding up. There was a huge push for efficiency in that era and we would shorten the life of the blades in favor of efficiency, there would be regular inspections and coatings were such a pain in the ass.

A lecturer and later colleague of mine did CMC blade research as a postgraduate. Taught some of it to me as an undergraduate, alas I forget most of it. He's retiring now.

Not sure what to make of that. I do hope the idea finally has its day.

"make [the trailing edge of the blade] razor sharp and it will just melt away"

Why would this happen on the trailing edge?

Not enough cooling. A turbine blade has a web of channels and internal cavity for cooling.

Here is a good diagram: https://www.semanticscholar.org/paper/Numerical-study-of-tra...

> Those parts can be safely 3D printed, as various metal sintering/remelting methods produce similar or better structural quality.

Wait, really? If that's true, colour me impressed. I would have put getting the creep resistance, fracture resistance, and tolerances necessary for turbine blades from 3D printing as just about impossible.

I'm conflicted by agreeing with systemvoltage's point, that it's overhyped, with knowing it is empirically wrong. Still, turbine blades? Wow.

I would like to add that additive manufacturing is still in its infancy, and that there are structures that pay large strength dividends which are impossible to manufacture any other way, such as 3-dimensional honeycombs, which haven't made their way into standard engineering practice yet. But there may come a point where the available design space is so much better that it's worth sacrificing some strength in the material itself.

Derailing a bit, but my absolute favourite HN threads are ones where people say something could never be done, and get immediate replies saying "it's being done now! here's sources!"

In civilian aviation, or in military training and peacetime operations, sure, but in a fly-or-die situation, like the one described in the article where supply lines are compromised, the ability to print and use a part for emergency use could be a literal lifesaver. It's analogous to replacing fuses with copper bars before entering combat (in some scenarios).

Incredibly strange view, because additive 3d metal printed parts are being used in high powered automotive and aviation right now.

Perhaps check out some of the leading contenders in the field and reconsider that appeal-to-authority stance?

https://www.ge.com/additive/industry/aerospace

https://youtube.com/watch?v=NKr_zCCkEKs

https://youtube.com/watch?v=a6SSw7FGky0

https://youtube.com/watch?v=cGGLSN6CTwI

Exactly.

The issue of developing aircraft parts is how to make it as light as possible and still perform reliably.

Machining the part is least of the problem. Any other method will be inferior if it produces heavier part to get same reliability or will be less reliable for the same weight.

It might be now, but metal composites with electron beam sintered 3d printing will get us there eventually.

What strikes me is that we are reaching "mass manufacturing" limits (or possibly internal combustion engine limits. Possibly both. I have a nice modern car, and it stopped working the other day. And I had absolutely no idea what was wrong. No way to diagnose and no conceptual framework.

Military hardware (I am guessing) cannot be as fragile as my nice car, so it must now be lagging behind the cost benefit curve of commercial car production - all the efficiency gains Ford makes must get lost as they cannot risk the fragility.

Additionally military hardware is really a special snowflake - lots of highly specific not really available commercially parts.

So military hardware must be desperate for actually working 3d printing / one off manufacturing - it solves the cost problem and the special snowflake problem.

So I am guessing the DoD will be a heavy investor in all things additive manufacturing for a decade ? Am Inway off base ?

I'm curious why they don't make 3D models from the original drawings. Maybe they're lost or kept by the manufacturers? By only measuring existing parts, they wouldn't know the tolerances or loads and would have to kind of re-engineer that missing information.

With a $693,058,000,000 annual budget, might as well just make everything out of titanium.

I'm assuming the original drawings weren't used for the same reason you can't trust most contractors documentation. Emergency bugfixes get pushed to production.

Then how do they make conventional replacement parts? Somebody must know the dimensions, tolerances, materials, etc.

Not to mention wear and deformation. Are the landing gear of that particular UH60 bent in the slightest? Has the throttle lever raceway worn? How much damage did the door slamming on some GI's M4 do to the sheet metal?

I saw this making the rounds today, even though it's evidently old news, about the b1 bomber being disassembled/scanned. Here's an article on the subject from late April 2020: https://www.thedrive.com/the-war-zone/33151/air-force-sends-...

I wonder how this relates to "right to repair" topics.

I hope "3D scan" means "use CMMs, CT capture, x-ray spectral alloy analysis, profilometers, etc. to reverse engineer the components."

This will be helpful for any army that wants to manufacture them.

Inherent in every part is tolerances, details of finishes, requirements for strength, etc. You can't find those things through 3D scans.

At best this is publicity stunt.

That is some valuable data.

This seems like a surprisingly bad idea. I would think the vendors' engineers would be able to identify the parts that qualify to be 3d printed, given how well the expected material characteristics match what the printer could produce.

If nothing else, they could provide the cad files.

These arn't irregular shapes. You can draw them from the part if you need to

From the article:

“But digital models simply don’t exist for vast inventories of older parts, some dating back decades.”

I presume that this has something to do with their approach.

Seems amazing to me that this method would be more cost effective or even just plain effective vs scanning and interpreting paper drafts/blueprints.

Because actually finding blueprints that reliably match what the helicopter is actually using can be a real problem that can take far longer, will __still__ require you to cross-check against the part because of human error :-)

That would be the most economical, but it might be bureaucratically easier for a general to tell a unit to figure this out all in-house than deal with the vendor which may need new contracts to ask for these details.

My guess is that this might be a way to get spare parts for a helicopter in case they break. A family friend who worked on antique cars would often run into the problem where a specific piece broke and he needed to find it replaced. Finding the exact part for a 1970s Porsche is not only difficult, but also incredibly expensive. Having a 3D model of the specific part and just printing it would have been not only far faster but orders of magnitude cheaper. The exact material could just inferred from the broken part or a stronger material could be selected out of caution.

I hope their plan isn't to 3D print safety critical parts like the rotor blades, that would certainly end poorly in the way you mention.

Reply on news.ycombinator.com