How should rockets flex?

[Vl3d]

The poll is very simple - what is your preferred amount of rocket flexibility and how should rockets behave when under joint stress?

 

[modus]

Flex a little but also autostrutted

[Sea_Kerman]

I want it to make sense to have to reinforce certain things, like struts at the tops of boosters and to hold large payloads in place in fairings

[Periple]

4 minutes ago, Sea_Kerman said:

I want it to make sense to have to reinforce certain things, like struts at the tops of boosters and to hold large payloads in place in fairings

Joints don't need to flex for that, they just need to break under stress. 

[Mutex]

I have to say I've never been a fan of auto-strutting. I've used it, but it feels like a band-aid over a problem that should have been fixed better. Maybe by keeping joints pretty rigid, and having them just break rather than flex if the forces are too great. That's how I'd expect it to work in real life. And if your ship still isn't stable, then it's a bad design and needs more internal supports.

[magnemoe]

Tanks should bend very little, before breaking, but lots of oscar tanks on top of each other would able to bend quite a bit. 
Rigidity should increase with diameter so an 50 meter tall 5 meter stack would be rigid but an 1.25 meter one of tanks with similar length would start to flex a bit. 
Obviously stuff like balancing an space station on an 1.25 m decopler would be unstable unless strutted. 

[KSK]

My go-to real-life comparisons here are:

SpaceX’s Starship (the orbiter not the Superheavy booster)

- demonstrably made KSP style from lots of small tank sections stacked on top of each other.

- capable of flying sideways.

- does not visibly bend, flex, wobble or break up. At least not during flight.

SpaceX’s Falcon 9.

- High fineness (long and spindly) rocket.

- Wider payload fairing.

- Does not visibly bend, flex, wobble or break up during flight - at least due to aerodynamic problems.

Noodle rockets are an artifact of KSP’s joint model that bears little resemblance to real life. Rockets can also be designed and built to withstand aerodynamic forces acting off their long axes.

The only in-game tool we’re given to deal with noodle rockets or joint flexing in general is to stitch them together with unphysical magic struts. I fail to see how unphysical magic struts are an improvement over unphysical magic unbreaking joints. Plus ‘adding moar struts’ is just tedious.

Conclusion. Noodle rockets, and the joint model that causes them, adds nothing to realism and nothing to gameplay. Get rid of them/it.

 

Edited March 2, 2023 by KSK

[Vl3d]

35 minutes ago, Mutex said:

I have to say I've never been a fan of auto-strutting. I've used it, but it feels like a band-aid over a problem that should have been fixed better.

I agree. I also do NOT like autostruts. It is a hidden game mechanic that is just not necessary if joints are implemented correctly.
I do NOT like rockets that bend more than 5-10 degrees. The kerbals are goofy enough - rockets in the game should have realistic behavior.
Rockets breaking apart should have been a game mechanic from the start, even in KSP1. If you fail the launch profile, you try again. I don't like it when a rockets has a speed of hundreds of m/s in the atmosphere and then it can just flip like nothing is happening. Aerodynamic forces should have a much bigger impact on gameplay.

Rockets should break apart like in the cinematic trailer!

Edited March 2, 2023 by Vl3d

[NutellaSandwich]

Don't planes at least break apart if you pull really hard turns? I've had planes explode mid-air while trying to turn really aggressively.

[Mutex]

There are G-force limits on the parts, but they seem to be quite generous.

[Vl3d]

15 minutes ago, KSK said:

- [Falcon 9] Does not (...) break up during flight - at least due to aerodynamic problems.

Yes they do. Any long rocket booster will break up if going at high speed sideways in the atmosphere. The mass needed to prevent that would be huge and doesn't make sense. At 01:55 in the video they say that Falcon 9 broke up.

"The Falcon 9 rocket exploded in a brilliant fireball soon after, breaking apart due to increasing aerodynamic forces."  https://www.space.com/spacex-crew-dragon-abort-elon-musk-reaction.html

[Drakenred65]

Falcons flying sideways is only really a thing when hovering or at low velocity. They will still break up when they lose lateral or longitudinal controll. I think it's happened once, in a payed launch and in a test launch that was deliberately flown to destruction.

 

Edited March 2, 2023 by Drakenred65
Clarifying my point

[sumeee]

16 minutes ago, Vl3d said:

Any long rocket booster will break up if going at high speed sideways in the atmosphere.

Not quite any rocket, most of them but not all. Here is an example:

At around 5:30 the rocket start spinning out of control but it didn't brake (except for the fairing). The explosion was caused by the flight termination system.

Also, since KSP is a game it should go for fun gameplay over realistic physics. At least in my opinion a spinning rocket is more fun than no rocket.

[tstein]

41 minutes ago, NutellaSandwich said:

Don't planes at least break apart if you pull really hard turns? I've had planes explode mid-air while trying to turn really aggressively.

Modern planes made for maneuvers are way sturdier than the pilots flying them.

The key word is LITTLE. Little is really little. Most  things humans  design break above 2-4 degrees of bending...

[Drakenred65]

10 minutes ago, sumeee said:

Not quite any rocket, most of them but not all. Here is an example:

At around 5:30 the rocket start spinning out of control but it didn't brake (except for the fairing). The explosion was caused by the flight termination system.

Also, since KSP is a game it should go for fun gameplay over realistic physics. At least in my opinion a spinning rocket is more fun than no rocket.

My understanding is that most explosions are range safety kicking in, followed by failure of fueling and then structural  failure of tanks.

[KSK]

30 minutes ago, Drakenred65 said:

Falcons flying sideways is only really a thing when hovering or at low velocity. They will still break up when they lose lateral or longitudinal controll. I think it's happened once, in a payed launch and in a test launch that was deliberately flown to destruction.

 

Which is why I didn’t say anything about Falcons flying sideways.

1 hour ago, Vl3d said:

Yes they do. Any long rocket booster will break up if going at high speed sideways in the atmosphere. The mass needed to prevent that would be huge and doesn't make sense. At 01:55 in the video they say that Falcon 9 broke up.

"The Falcon 9 rocket exploded in a brilliant fireball soon after, breaking apart due to increasing aerodynamic forces."  https://www.space.com/spacex-crew-dragon-abort-elon-musk-reaction.html

1. Did I say anything about Falcon flying sideways?

2. Well duh. If you deliberately shoot the nose cone off a rocket at an altitude where it’s not normally supposed to come off, that’s going to cause problems. 

I can recall exactly one Falcon flight coming apart in mid air and that’s because it’s upper stage overpressurised and blew up.

My point was that Falcon is a long, spindly rocket but somehow manages to avoid flying like a KSP style wet noodle. And the one time that one did explode in flight had nothing to do with bendy joints.

[tstein]

12 minutes ago, KSK said:

My point was that Falcon is a long, spindly rocket but somehow manages to avoid flying like a KSP style wet noodle. And the one time that one did explode in flight had nothing to do with bendy joints.

Because steel  tubes are MUCH MUCH more resilient than what we get in game. No need to  even look so far. See  military missiles, they do  30 G turns, in some cases 50 G turns and they do not break! It is EXTREMELY easy for an engineer to   make such a structure rigid enough. The wobbleness of KSP rockets is pure goofy  physics.

[Vl3d]

41 minutes ago, sumeee said:

At least in my opinion a spinning rocket is more fun than no rocket.

A rocket breaking apart and exploding is more fun than a rocket spinning.

[VlonaldKerman]

Can anyone explain why joint rigidity is so low? Seems like an unnecessary thing. Unlike other problems, it seems like this isn’t something that needs a change implemented, but rather, I implemented. Am I missing something?

[tstein]

41 minutes ago, VlonaldKerman said:

Can anyone explain why joint rigidity is so low? Seems like an unnecessary thing. Unlike other problems, it seems like this isn’t something that needs a change implemented, but rather, I implemented. Am I missing something?

Usually it is due to numerical stability on some physics engines that cannot handle large rigidity between elements with vastly different masses. Some engines have problems with any mass  much larger than 1 (of whatever base mass unit used). Less rigid constraints can alleviate some numerical issues. In proper engineering simulation that is usually solved by  in real time  collapsing and creating joints when certain thresholds are reached (so when you have not enough force to justify  ANY dynamics,  you handle the object as one). But  without access to the specific type of engine they are using and some details of the implementation it is hard to guess what style of solver they are using. I am out of the industry loop for some 18 years (jumped to healthcare tech ), so maybe K^2  has a better idea of the usual candidates uses nowadays.

Edited March 2, 2023 by tstein

[VlonaldKerman]

3 minutes ago, tstein said:

Usually it is due to numerical stability on some physics engines that cannot handle large rigidity between elements with vastly different masses. Some engines have problems with any mass  much larger than 1 (of whatever base mass unit used). Less rigid constraints can alleviate some numerical issues. In proper engineering simulation that is usually solved by  in real time  collapsing and creating joints when certain thresholds are reached (so when you have not enough force to justify  ANY dynamics,  you handle the object as one). But  without access to the specific type of engine they are using and some details of the implementation it is hard to guess what style of solver they are using.

Understood, makes sense, thank you! Learn something new every day.

[Sequence]

I don't like wobble, but it does give you a good visual clue as to what went wrong with the design. As long as the failure screen is clear in explaining which joins broke I'd rather have minor flexing but not wobbly. 

[EvelynThe Dragon]

I am very displeased with the floppy noodleness of KSP rockets. It's quite ridiculous. I heard that there is a means to tweak this in config files, I tried finding a thread about that without any success, but if anyone can point me to it, I will appreciate it.

[sumeee]

1 hour ago, EvelynThe Dragon said:

I am very displeased with the floppy noodleness of KSP rockets. It's quite ridiculous. I heard that there is a means to tweak this in config files, I tried finding a thread about that without any success, but if anyone can point me to it, I will appreciate it.

I'm to lazy to find the actual thread, but the file you need to modify is: %USERPROFILE%\AppData\LocalLow\Intercept Games\Kerbal Space Program 2\Global\PhysicsSettings.json

In there add a few more zeros to "JOINT_RIGIDITY".

[EvelynThe Dragon]

Just now, sumeee said:

I'm to lazy to find the actual thread, but the file you need to modify is: %USERPROFILE%\AppData\LocalLow\Intercept Games\Kerbal Space Program 2\Global\PhysicsSettings.json

In there add a few more zeros to "JOINT_RIGIDITY".

Thank you. Much appreciated.