A couple of Workflow Questions

[Ivaylo Katev]

yes it would help, but this will decrease the speed. if you are satisfied by the simulation details, the dummy object is better. no need to be low poly one, you can use the original mesh but without internal walls.

[ben_hamburg]

Since the original mesh was a pretty mess CAD import, we already had to fix it and I created a simulation dummy right there, with no internal walls. I guess we´d have to check if we have enough time to increase grid resolution.
I´ll also recheck the dummy for any unweld vertices or stuff like that.
I´ve set the boat up to start in the air above the surface level of the liquid, lowering it into the water over 50 frames and I can see the water penetrating the boat right there, so I´ll do another test with higher grid resolution and see if it helps. Another problem might be, that the grid moves with the boat right from the beginning, maybe it also helps to start moving the grid only after the boat is lowered into the water?

[Ivaylo Katev]

you mean the grid is moving with the boat but without the vertical movement?
well, this explains a lot, the boat occupies place initially filled with water. there is a new option "clear inside", see the per object phoenix properties. it allows you to clear the overlapped water.

[ben_hamburg]

Yes, I rigged the boat to be orientation constrained to a helper for up and down and rotation keying and position constrained to another helper that moves at constant velocity on a path constraint. I linked the grid to the Position helper.
Because I was worried that the initial fillup would mess with the boat collision, I keyed the boat to slowly sink in the liquid after initial fillup.

Where to I find the "clear inside" option? Maybe I´m not on the newest nighlty build? I´m currently on 24130...

[Ivaylo Katev]

right click over the geometry, see for "phoenix fd properties" in the menu, it must bring up a dialog with per object properties
if you start from position above the water only because the initial fill up, this option makes this not needed anymore.

[ben_hamburg]

Ok, thanks. I was on the wrong build on the Computer I was running the fluids on and didn´t see the option. It looks better after the first few frames no, I´ll report back with the results!

[ben_hamburg]

Yes, it seems to work way better now!!

One more question though for better understanding:
My simulation is set up so that the grid moves at constant velocity with the boat. Enabling inertial forces makes the boat react with the grid in a way, that it seems to plow through the water, even if actually both move at the same speed.
The set back is, that in the beginning there is sort of a little "shockwave" going through the grid, which resolves after a certain amount of frames.
But as far as I understand, this setup, which allows me to simulate the boat going through a huge distance without having to have a huge grid that covers the entire distance, only works if the boat and the grid move at a constant velocity.
If the boat (and the grid linked to it) would for example speed up or slow down during the simulation, due to the inertial forces, I would get another shockwave each time a slow down or speed up would occur.
Am I correct in my understanding of the process here?
Also it only works if the boat is going straight, if the boat and the grid would rotate, I would also get unwanted effects, right?
Not sure if you already answered the question, this thread is getting bigger and bigger...

[Ivaylo Katev]

If the boat (and the grid linked to it) would for example speed up or slow down during the simulation, due to the inertial forces, I would get another shockwave each time a slow down or speed up would occur.

the initial shock wave is just the pressure convergence, you have it even in a static grid, but is less visible (the surface goes up and down).
the detailed description of the "inertial forces" parameter is force that moves you from internal to external reference frame. this is very confusing for a man that goes deep into the physical details (including me) because in the physics changing the reference frame turns the acceleration into forces, the centrifugal forces are the most popular example. in the official release the inertial forces were REALLY exactly like this. however, about six months ago we tried to make a boat that follows curved path, and this was a disaster, the liquid started to move in outward direction, that was apparently wrong. i will not bore you with technical details, just will say that we reconsidered the meaning of the inertial forces and now you can accelerate the container and get the expected result.
and if you are afraid that fixing the problem we lose the ability to have centrifugal forces - we don't the inertial forces fix is not a cheat, it's physically correct and produces the results that should happen in the real world.

[ben_hamburg]

Yeah, I´m really glad I don´t have to fully understand whats going on in the physics department, as long as I get a firm grasp of what buttons I need to press...

But to make sure that I understand this correctly: It´s possible to accelerate or decelerate the grid now with correct results?
But having the grid follow a curved path still won´t work, right?
Because so far I could reuse the same sim over and over and with clever camera work and cuts make it less obvious, that the boat was at a constant speed on a straight trajectory.
But no I´ll have to do a curved path and I´m still figuring out how to make the grid as small as possible to avoid extraorbitant sim times...

[Ivaylo Katev]

no problem to accelerate the grid or to follow curved path, we had this problem but was fixed half year ago.

[ben_hamburg]

And another update on the collisions... I´m still having trouble at the front of the boat, now it looks like liquid is spilling on top of the boat at the very front and then kind of sticks there and builds up, I think this is probably also something that has to do with the inertial forces: the splashes in front of the boat, that land on the boat itself are probably now moving at the same speed as the boat and don´t fall of, because the top is flat.
Kind of hard to explain without seeing it in motion, but I attached a screenshot.
My solution now is to extrude the topfront of the dummy, so the splashes don´t reach the top of the boat anymore and hope that the collision-barrier that of course isn´t there in the render mesh won´t be visible during animation...

[ben_hamburg]

no problem to accelerate the grid or to follow curved path, we had this problem but was fixed half year ago.

That is great news!! So now I´ll only have to experiment to figure out how wide I need to make the grid, so it captures the waves from the curved path.

[Ivaylo Katev]

And another update on the collisions... I´m still having trouble at the front of the boat, now it looks like liquid is spilling on top of the boat at the very front and then kind of sticks there and builds up, I think this is probably also something that has to do with the inertial forces: the splashes in front of the boat, that land on the boat itself are probably now moving at the same speed as the boat and don´t fall of, because the top is flat.
Kind of hard to explain without seeing it in motion, but I attached a screenshot.
My solution now is to extrude the topfront of the dummy, so the splashes don´t reach the top of the boat anymore and hope that the collision-barrier that of course isn´t there in the render mesh won´t be visible during animation...

[ATTACH=CONFIG]18440[/ATTACH]

well... not sure how to understand this. you mean the liquid just refuses to leak once fallen over the geometry?

[ben_hamburg]

Yes, thats exactly how it looks like...unlike with the Borg it seems that resistance is NOT futile.