Hi I have been using F3D this fall in my master thesis, simulating free-falling objects in water. Most of my time has been spent on evaluating the terminal velocity of such objects. My results are very consistent with theory, and the GMO-model seems to work very fine. The reason I am posting is some strange results with the hydraulc force. The force is experiencing extreme oscillations, spikes of up to 300% in only one timestep. I have tried adding limited compressibility (IMP=2), without much change. This does not seem to impact the velocity of the object, but I still find it quite strange.

Peter,

what kind of objects the GMO represent? Are they impacting free surface or are they always fully submerged?

Michael

Have you tried implicit GMO? It may work better.

Hi I have tried both explicit and implicit, without any difference.

I have run simulations with the object just below and just above the surface, with and without an initial velocity straight down. No differences are seen.

How about convergence when the object contact water surface? If it cannot converge, the force will be wrong.

You mean at water entry? I have done simulations with fully submerged objects as well as ones with water entry. The force doesn't seem to start oscillating until after some times, when the object has travelled som distance down. Here is output of General History->Hydraulic Force i z-direction, space system

http://folk.ntnu.no/ellevset/resources/oscil.jpg

This is an extreme case, and the oscillations start very early.

Wow! The force looks crazy. This needs a closer look at the input file. Have you tried contacting support?

I sent them an email today. The strange thing is that the kinetic energy of the same object looks quite nice, you can see some small oscillations, but nothing like these. The previous result was very dramatic. It was only run with 160 000 cells. This one is using approximately 1.1 million. http://folk.ntnu.no/ellevset/resources/oscil2.jpg

I have gotten an explanation for the violent oscillations in my situation. When a GMO-object enters a computational cell the source term from the GMO increases the pressure by this introduction of mass. This causes fluctuations in the pressure, which causes these spikes. This also happens when the object leaves cells. However, since the spikes last for a very short time their effect on the velocity is very small, perhaps even none. If one were to smooth the curve the actual force is found. These spikes may then be viewed as just noise.

Sounds about right, Peter. The sources on the leading and trailing edges of the GMO are for volume, not mass. They are used to account for the change in the open part of the control volumes that contain the GMO.

In is interesting that despite those oscillations, the predicted velocity is good. THis is truely noise!

Michael