ClothSolver/Sopsolver Interaction

[sanostol]

unfortunately, the clothsolver right now does not offer any pressure capabilities, but combinded with the sopsolver and the fexternal force my first thought was that this should be at least be approximately doable. it works quite ok, for the first tests, but for the interaction of the cloth solver with the sopsolver there are still some questions.

for example I use the fexternal to transfere the pressure force to the clothsolver. the cloth solver has internal maximal substeps of 1024, but the sopsolver just gives one update per timestep. i attached a time attribute to check, and it just updates at every global timestep, so my guess is that the cloth solver works with the same external force for each timestep on all substeps. I tried to increase the dops "max feedback loops", but it seems to have no influence on the fexternal update to the cloth solver.

I attached a file, first I tried to use the normal of the object, but a volume gradient worked better for the forcedirection

if you increase the internal pressure, it does not result in the shape you would expect , instead it is not round

increasing the global substeps helps, but at the cost of time. would be cool to have something like a cosolver for the clothsolver so it can get updated data

clothinternalpressure.hip

Edited April 19, 2012 by sanostol

[TAkin]

unfortunately, the clothsolver right now does not offer any pressure capabilities, but combinded with the sopsolver and the fexternal force my first thought was that this should be at least be approximately doable. it works quite ok, for the first tests, but for the interaction of the cloth solver with the sopsolver there are still some questions.

for example I use the fexternal to transfere the pressure force to the clothsolver. the cloth solver has internal maximal substeps of 1024, but the sopsolver just gives one update per timestep. i attached a time attribute to check, and it just updates at every global timestep, so my guess is that the cloth solver works with the same external force for each timestep on all substeps. I tried to increase the dops "max feedback loops", but it seems to have no influence on the fexternal update to the cloth solver.

I attached a file, first I tried to use the normal of the object, but a volume gradient worked better for the forcedirection

if you increase the internal pressure, it does not result in the shape you would expect , instead it is not round

increasing the global substeps helps, but at the cost of time. would be cool to have something like a cosolver for the clothsolver so it can get updated data

Even thought this reply is coming several months later, I hope it can still be of help. I have experienced a similar problem with the fexternal attribute not working for me too. But infact, the fexternal attribute actually works...all you have to do is to read it in recursively from SOPs using a SOP Solver. If this is still of interest to you, let me know and I will post a scene file.

[anim]

there was a bug of fexternal not working at all prior to version 12.0.579

Tuesday, March 27, 2012

Houdini 12.0.579: Restored cloth-solver support for the "fexternal" point attribute. This attribute can be used to apply external forces to the points of a cloth object.

after that build it's as easy as feeding some info, like scaled normals, or custom vectors to fexternal inside sop solver and cloth will pick that up

note that sop solver is necessary for H12 cloth to update any attribute on mesh over time since they are no longer picked up from original geo after the initial frame