Adding Flip Fluid to cached auto fracturing. Need Help

[tricecold]

Please Help me out here;

Is this the right approach;

Fill the glass with water and write out a rested initial state of the liquid.

Break the glass with a projectile with auto fracture.

Bring back the initial state of the liquid, and run another simulation just for the liquid to collide with pre-simulated breakign glass.

Now I tried to simulate with both flip fluid and auto fracture at the same time but it mostly either crashes or simulation takes forever.

I tried to write out bgeo sequences both with rop output driver and with file node only for the glass geometry yet the liquid does not collide even if I add active or passive rigid body to the bgeo files.

Please help...

[ikarus]

check the collision sdf on the geo sequence. the glass is probably too thin to be evaluated at a proper resolution so there is no collision volume for the fluid to interact with.

if thats the case then you need to either make the glass thicker or increase the resolution of the collision volume a good deal.

[Spyrogif]

Instead of increasing the substep (leave it at 1) it can be a good idea to reduce the time scale. It's help a lot to detect and understand collision problems in a simulation.

what did you mean by "simulation takes forever"....is it forever or very long ?

Do you have a lot of fractures pieces ? 10, 100 or 1000 ?

and the most important. Do you have check your normal, is it logical ? for the outside part the normal needs to point to outside and for the inside they need to point inside.

[tricecold]

Does using houdini 11 on x64 ubuntu makes a difference compared to win7 64bit for simulations...

[tricecold]

Instead of increasing the substep (leave it at 1) it can be a good idea to reduce the time scale. It's help a lot to detect and understand collision problems in a simulation.

what did you mean by "simulation takes forever"....is it forever or very long ?

Do you have a lot of fractures pieces ? 10, 100 or 1000 ?

and the most important. Do you have check your normal, is it logical ? for the outside part the normal needs to point to outside and for the inside they need to point inside.

Reducing the time scale sounds like a great idea.

I am terribly new to Houdini, I can write out .sim and .bgeo files with ROP dynamics driver but Honestly I dont even know where to plug them back in another DOP solver.

What I did was step by step;

Created a glass from revolved curve.

add ground object

add RBD to glass

create a bullet

add RBD to bullet

add initial speed with the right speed and direction to hit the glass

glass gets hit...

add auto fracture to glass

hit play and it simulates fine.

on the object level choose autodop and at the cache area I choose the most bottom setting to write out the .sim for given frame range, I guess this can also be done with adding a ROP Dynamic Output driver and hit render.

Then create another file without the bullet, add flip fluid in the glass to a geometry ( smaller )inside glass simulate the liquid settling down and write out an initial state.

My problem here is how to join these sim data, should I be writing out .bgeo files for fracturing glass. If so which is the right way to it in order to get the velocity data for liquid collision and motion blurring.

Like I said I am very very new to houdini, I guess I cant even explain my problem properly.

I downloaded the original glass breaking file, But it seems like it is way too complicated for my current level.