36 ms·

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

systemvoltage 6y ago
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.
- redis_mlc 6y ago
  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 https://www.youtube.com/watch?v=m1jl6IlywAk
- tantalor 6y ago
  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?
  - wlesieutre 6y ago
    It’s been used for several rocket engines
    - WalterBright 6y ago
      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.
      - wlesieutre 6y ago
        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 https://en.wikipedia.org/wiki/SuperDraco [2] https://3dprint.com/10830/spacex-3d-print-oxidizer-valve/ https://3dprint.com/10830/spacex-3d-print-oxidizer-valve/ [3] https://www.nasaspaceflight.com/2016/10/its-propulsion-evolution-raptor-engine/ https://www.nasaspaceflight.com/2016/10/its-propulsion-evolu... [4] https://en.wikipedia.org/wiki/Rutherford_(rocket_engine) https://en.wikipedia.org/wiki/Rutherford_(rocket_engine)
tpmx 6y ago
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.
- robocat 6y ago
  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.

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.

redis_mlc 6y ago
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 https://www.youtube.com/watch?v=m1jl6IlywAk
tantalor 6y ago
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?
- wlesieutre 6y ago
  It’s been used for several rocket engines
  - WalterBright 6y ago
    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.
    - wlesieutre 6y ago
      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 https://en.wikipedia.org/wiki/SuperDraco [2] https://3dprint.com/10830/spacex-3d-print-oxidizer-valve/ https://3dprint.com/10830/spacex-3d-print-oxidizer-valve/ [3] https://www.nasaspaceflight.com/2016/10/its-propulsion-evolution-raptor-engine/ https://www.nasaspaceflight.com/2016/10/its-propulsion-evolu... [4] https://en.wikipedia.org/wiki/Rutherford_(rocket_engine) https://en.wikipedia.org/wiki/Rutherford_(rocket_engine)

tpmx 6y ago

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.

robocat 6y ago
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.