Ship bouncing with Active Body

[david_slachta]

Hi! I'm working on a simulation of a large ship going through a rough sea, and I've set the simulator together with massive wave force and the ship as an active body. However, the movement of the ship over the waves looks a bit strange. It seems to bounce up and down a lot. I did check the guides and tutorials, and I believe I've set the center of mass and density of the ship properly. The ship's density is actually quite low (200kg/m3), but with a higher density, it gets submerged a lot.

Here's a video of the sim:

Should I perhaps keep the ship still for some frames to subside the bouncing and only then start the wave force? Is it possible to animate the wave force so that it's 0 at the start and only picks up after some frames? The attached video is 1500 frames long and the bouncing seems to subside after 500 frames or so, but then it picks up again towards the end, so I'm really not sure...

By the way, the ship's movement forward is not animated but handled by the thruster force. Is that perhaps making the bouncing worse?

Please let me know if I should post any screengrabs of the settings.

Thanks for any pointers!

[georgi.zhekov]

Hello,

How about if you add a few more steps per frame in the Dynamics rollout? How many are you using at the moment?

Yes, you can animate the Wave force to be 0 at the start and to pick up afterwards.

Alternatively there are two other approaches that you can try:

1 - Use a Phoenix Axis lock node - this will allow you to restrict the movement of the ship in certain axes. Info on how it works here - https://docs.chaos.com/display/PHX4M...Body+Axis+Lock

2 - Another approach would be to hand animate the ship (the origianl body) and in the right-click Phoenix properties menu of the ship to set the Original Animation Influence to something like 0.1 or 0.005. This way the Active Body ship will follow the animation of the original, but not too strictly.
The higher the Influence - the more the Active Body will follow the original path. You can find more about the option here - https://docs.chaos.com/display/PHX4M...ode+Properties

[david_slachta]

Hi Georgi, and thank you for the response,

Regarding the steps per frame, I'm using the default 1 step.

I found that if I increase the resolution of the simulator the bouncing and the overall movement of the ship is much better and more realistic. With the higher res simulator, I was able to set a more realistic density of the ship. Before I had to use a value of 200 (which seems to be quite low, when in the preset the cork is 270), but now I can use a value of 310. It's still on the lower end, but the behavior of the ship looks much better. It does take quite a bit of time to simulate though. In the video I posted above I was using a simulator with 5 milion vortices (voxel size 0.5m) and now I'm working with a 50 milion simulator (voxel size 0.22m). Do you think I can use a low res simulator with more steps per frame or should I stick with a higher res simulator? I'm wondering what would simulate faster and give more realistic results, as I currently need to wait a full day to see the result of one test. My workflow is that I first simulate only for the ship movement, and then I'll use the simulated ship and place it into a very high res simulator (200 million or more) and do the simulation of the water, foam, mist, etc.

Regarding the axis locks, I don't really want to lock the ship's movement, as it should be moving up and down along the waves. Same about the hand animation influence. I would really love to get a realistic ship behavior on the big waves. If I do my own hand animation, the animation influence will dampen the physical reactions with the water, right? But it's something I will explore if I can't get it right with no hand animation.

[georgi.zhekov]

Ah, indeed - higher grid resolution will yield better results. Honestly there is no specific and right way - adding more steps might solve the issue even with the lower resolution, but still the simulation time will increase progressively with each additional step, 2 steps will simulate for double the time and so on..so in the end it's about what looks good enough for you. The benefit of the higher res simulation is that you also get better looking liquid.

My suggestion is to try the different approaches and see what looks best for your case and pick that.