CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   Water impact of cylinder on free surface (http://www.cfd-online.com/Forums/star-ccm/114389-water-impact-cylinder-free-surface.html)

Espen March 10, 2013 09:18

Water impact of cylinder on free surface
 
2 Attachment(s)
Hi everybody,


For my thesis, I'm trying to use STAR-CCM+ for simulating a constant velocity water entry of a cylinder using VOF. For simplicity, the cylinder is kept fixed and the relative velocity between water and cylinder is simulated by applying a constant velocity inlet condition on the bottom of the domain (see figure).

The problem is that when I use small time steps, to ensure CFL<1, I get unrealistic and fluctuating force records, whereas if i choose higher time steps, the results are more realistic. I believe this migh be because the HRIC scheme is used for CFL<1 and UD for CFL>1 and UD is more stable than HRIC. Another aspect is that the UD scheme "smeares" the free surface compared to HRIC.

Also, I'm starting the simulation a short time before the impact, to reduce any unphysical transient effects, this means that there is some air between the bottom and the cylinder at the simulation start. As the deadrise angle is zero at the bottom of the cylinder, air gets trapped under the cylinder, but I'm not sure if this should affect the solution (force is calculated from pressure only) and both air and water are treated as incompressible.

Some extra facts:
Cylinder radius=1m, relative velocity=1m/s, trimmer mesh + prism layer. Prism layer 3cm with approx 20 elements (depending on case). The problem seems to be dependent on CFL number, i.e. the problem is the same for mesh size 2x2mm and dt=0.0001 as for mesh size 4x4mm and dt=0.0002. The problem is found for both turbulent (k-eps), laminar and inviscid flow, both 2d and 3d.

This is really frustrating, as I can't get any further on the thesis before I have solved this, since the latter parts of the thesis will be based on the recommendations I find for the simple case described above...

Do any of you have ideas for what could be wrong or suggestions for changes to be made?

liladhar July 16, 2013 11:54

you can do same problem in another way..............
instead of water u can drop cylinder on water by using dfbi model.........
in dfbi u arrest all the motion except downward motion of cylinder and u can get result......
all the best..............:)

jcamilleri March 30, 2014 15:08

Hi Espen,

I am trying to simulate a very similar problem using Star CCM+; a wedge drop on calm water surface. I am facing a very similar problem. I am using an overset mesh with mesh refinement around the wedge. Pressure is being monitored at 6 different positions on the wedge surface. The pressure plots show a lot of fluctuations when I use a very small time step (to get CFL<1). Also, the normalized residuals are very high especially the water residual, on the order of 1 mil. How did you manage to solve this problem?

Thanks and regards,

Josef


All times are GMT -4. The time now is 01:39.