Learning Forum

[jil260]

It's a water entry problem, and I simulated the case followed by tutorials. But the results diverge after the wedge falls into the water a few milliseconds. I checked the mesh, time step (0.0001s), and iteration numbers(100). I guess they are good enough. Can anyone tell me what should I be careful about?

[Amine Ben Hadj Ali]

The mesh quality metrics checker is even showing errors in pre-processer. Add all necessary information regarding your case and start by having a good quality mesh.n

[jil260]

The fluid domain is 800mm*500mm*650mm, the default mesh size is 0.02m. The wedge is 200mm*200mm (length and width), and the angle is 25 degrees. As shown in figure 2, the only refinement is used. I did not see any error or warning and the info said The mesh translation to Fluent was successful.n

[Rob]

Look at the skew reports: a max of 0.8 is OK, 0.95 is bad, 0.9857 is awful. n

[jil260]

When I delete the refinement, the maximum skewness is under 0.8. Do I need to refine the mesh near the wedge boundary like inflations used for flow past a rigid cylinder?n

[jil260]

Here are most settings of the case, could anyone find what is the problem? The result still cannot converge after the wedge impacts water 0.02 seconds.nThank you in advance.n

[Amine Ben Hadj Ali]

Perhaps worth using Overset here to have component and background high quality mesh without any remeshing. Give it a try.nn

[jil260]

I reduced the dimension to 2D, this time, there was no convergence problem or negative cell volume. But the wedge bounced up with a high velocity after impacted the water, which is an impossible result. Maybe this is also the reason causing errors in 3D. Please give some advice.nn

[Rob]

What are you using to calculate the motion? How far does the wedge move in a time step? In our first models with MDM back in Fluent 6 we bounced a block off water because we'd got too big a time step and forgot about the mass. n

[jil260]

I used 10e-5 as timestep with initial wedge velocity -1.6m/s. In 3D, the simulation works but the result seems weird. For example, the pressure contour is not varying gradually in a continuous timestep, positive v.s. negative. May I infer that the timestep is not small enough?nn

[Rob]

With those conditions how many time steps does it take for the wedge to cross a cell? If the free surface is also moving how quickly does everything come together?n