Getting realistic water surface representation in Flow Simulation
Posted: Tue May 14, 2024 4:09 pm
Hi, I am learning SW Flow Simulation with my ongoing boat projects (nothing too critical, but still educational), mainly external studies with immiscible air/water mixture, and I'm having a hard time setting up/visualizing studies to get an accurate and realistic representation of the air/water boundary. As I was advised by a colleague, in Flow Sim there is no such thing as a distinct boundary, but instead each cell of the mesh gets assigned mixture ratio of the fluids, which is why visualizing it with Cut/Isometric Mass/Volume Fraction plots makes it really difficult to set a value that matches the actual boundary between these fluids.
For example, I have this simulation where I'm trying to measure how much time it would take to drain water from boat's cockpit through a drain pipe at the back, and to get a realistic animation of the water surface. This is a side cut plot showing the volume fraction of water vs air:
As you can see, there is no distinct boundary between water and air as there would be in real life due to water surface tension; the transition is faded. I thought setting Isosurface Plot with Volume Fraction of Water at 50% should give me a good approximation of that boundary, but apparently it's not even close. For example, this is what I get when I try to visualize the water coming out of the pipe... I really doubt it would look like that in real life.
The best I've been able to do is to just set the mesh really fine, which reduces the faded area and somewhat makes visualization more realistic. Problem is, where this study would normally take ~1 hour to solve with default mesh, will take around 3 days to solve with mesh fine enough to look like in that screenshot above, which isn't great either. I did try localizing mesh refinement to the really important areas, but since the water level drops as the cockpit drains, I basically need a fine mesh everywhere... Which means studies solving for days.
So assuming that my goal is reasonably fast solution time and realistic visual representation of the water surface (meaning - clear and distinct, without fading), how should I set up my study and viewing the results? Attaching the file in case you want to take a look, the simulation is already set up.
For example, I have this simulation where I'm trying to measure how much time it would take to drain water from boat's cockpit through a drain pipe at the back, and to get a realistic animation of the water surface. This is a side cut plot showing the volume fraction of water vs air:
As you can see, there is no distinct boundary between water and air as there would be in real life due to water surface tension; the transition is faded. I thought setting Isosurface Plot with Volume Fraction of Water at 50% should give me a good approximation of that boundary, but apparently it's not even close. For example, this is what I get when I try to visualize the water coming out of the pipe... I really doubt it would look like that in real life.
The best I've been able to do is to just set the mesh really fine, which reduces the faded area and somewhat makes visualization more realistic. Problem is, where this study would normally take ~1 hour to solve with default mesh, will take around 3 days to solve with mesh fine enough to look like in that screenshot above, which isn't great either. I did try localizing mesh refinement to the really important areas, but since the water level drops as the cockpit drains, I basically need a fine mesh everywhere... Which means studies solving for days.
So assuming that my goal is reasonably fast solution time and realistic visual representation of the water surface (meaning - clear and distinct, without fading), how should I set up my study and viewing the results? Attaching the file in case you want to take a look, the simulation is already set up.