Create multiple FEM over time?

[ParticleSkull 57]

Hey guys,

I need to fill an object with some FEM spheres but:

- I need the spheres to be a copy of the same FEM sphere;
- I need it to be generated at the same place on every 10 frames or so (and fall due to gravity);

Does it make sense? 

Here's the basic idea but with 1 sphere only, I'd like to keep generating them until the beaker is full



File is attached

Cheers,

Alvaro

beaker.v1a.hiplc

[ParticleSkull 57]

I've been trying this since yesterday and now that publish it i almost found a solution by myself
(it happens ofen when I publish things here)

I realised that if I animate the Creation Frame parameter (DOP / Solid Object) I can generate just what I need.



Now I just need a way to animate this parameter procedurally. Does anyone have any idea on how to procede?

I just need the number to scale up by 10 on every 10 frames.

- between 10F and 19F the value must be 10

- between 20F and 29F the value must be 20

- between 30F and 39F the value must be 30

and so on...

Any ideas? 

multiple.fem.v2a.hiplc

[Atom 717]

Would something like this work...

floor($F/10)*10

 

[ParticleSkull 57]

simple as hell! Thank's again, Atom!

[loudsubs 33]

More efficient to leave the creation frame parm alone, and just create one solid object at the start of the sim. Then keep merging new geo into it with a sopsolver

fem_createmultiple.hip

[ParticleSkull 57]

Yeah Arrev, that's what I was looking for!!  Thank you

Do you know a good way to get rid of these interpenetrations?

My scene is a bit faster and I could make it better with a lot of Sub Steps (about 30) but it still having some interpenetration.


 

Edited June 7, 2017 by ParticleSkull
it's not integration steps, my bad

[Atom 717]

I have had to go as high as 100 substeps to get collision to work correctly. Maybe try something like adding more triangles to your sphere. You would be adding weight to the scene but it might offset the time you lose in adding substeps..?

Edited June 7, 2017 by Atom

[ParticleSkull 57]

Good to know! I guess I'll have to go around 100 as well.
I'll try some different things and let you guys know about the best result I get

[Farmfield 308]

12 hours ago, loudsubs said:

More efficient to leave the creation frame parm alone, and just create one solid object at the start of the sim. Then keep merging new geo into it with a sopsolver

fem_createmultiple.hip

If your DOP doesn't contain a SOP solver, you're doing something wrong.  #thetruthisoutthere

[anthonymcgrath 65]

oh this is super cool and something i was looking for a while back - lots of good options here ta chaps!

it reminds me of trying to spawn ragdolls in a similar way the first raggy was fine and fell to the floor in a heap but the others just broke wildly - my guess is the sop solver constraints needed updating to allow it to spawn new ragdolls with collision constraints on that ragdoll or something but i lacked the win to sort it out lol

[loudsubs 33]

Regarding the interpenetrations, try using higher res geo, and more tets.  I had a clip sop in thst scene to visualize the inner tets of the sphere.  Experiment with the sphere resolution and "quality" tab on the tetrahedralize sop.  That's what I'd try before getting too crazy with substeps

[tar 600]

23 hours ago, ParticleSkull said:

Do you know a good way to get rid of these interpenetrations?

Where is this happening for you? the only ones I see are birthed like that. i.e. sphere gets created within another sphere.

[ParticleSkull 57]

Hey Marty, it's happening when the spheres have a lot of pressure from other spheres, check it out

This simulation does have a lot of tets and I'm using 100 sub steps, I'm losing my hopes here

Guys, is there a way to get the older spheres out of the simulation, make it rigid?
Let's say, by age. If the sphere is 24 frames old it becames a static/rigid object?

I'll try to think on something else

Edited June 8, 2017 by ParticleSkull

[tar 600]

can you share your scene?

[ParticleSkull 57]

Yes! Thank you, Marty.

I've just decreased the sub steps back to 1. At 100 it took around 8 hours to sim

 

flask.v1e.Odforce.Edit.hiplc

[Noobini 267]

very basic question: wouldn't you just increase Collision Passes rather than increasing Substeps ?

[tar 600]

I've simmed to frame 170 with the ABE2 integrator and the only spheres that are merging are the ones created that were overlapping at the start - can you double check your file to see if they were overlapping. 

Edited June 8, 2017 by marty

[ParticleSkull 57]

Hey Noobini, yes, I did tried the Collision Passes but it didn't help. Thx though

Marty, thank you very much! I'm running a sim right now with ABE2 to see how it goes. 
I wasn't really aware of this. For some reason it was completelly out of my sight!
Have you increased the substeps or collision passes?

[tar 600]

Oh - I did set it to 3 and I also used a tetrahedralizeSop instead of solidembeded - sorry forgot to mention - just an inbuilt reflex!

[ParticleSkull 57]

Alright  
With 1 substep I got the interpenetrations. Now I'm simulating it with 3 substeps + tetrahedralize Sop.

Again, thank you very much, Marty. You saved me hours of tests and research that could lead to nothing