[cshsgy]

Hi all,
I have recently encountered a strange problem that troubles me a lot. I am running a two-phase case the involves particles and air. The geometry is simply a 2D slice of a cylindrical body.
The simulation is done with variable time step with Courant Number smaller than 1, so the problem is discovered that deltaT suddenly drops to an extremely low level after simulating for a while, which means the Courant Number actually blows up at that position.
To simplify this problem, I forced the volume fraction of particles to be 0, so what is simulated is actually a single phase flow.
The problem happens at around 1.5s of simulation, and the velocity field is attached: before and after the blowing up of Courant Number, nearly all quantities behave quite well, including k, epsilon, velocity, etc. The only observable irregularity is observed with the pressure field. Pictures of the velocity field and the pressure field are attached.
I appreciate it a lot if you could be helping me, since it has really troubling me for quite a long time.
Thanks a lot guys!

[piu58]

In most such situations the b.c. are not correct. I would check them first.

[cshsgy]

Originally Posted by piu58

In most such situations the b.c. are not correct. I would check them first.

Hi piu58,
Thanks a lot for your reply!
However, I have checked my boundary conditions and changed them several times as well. It seems that they are not that problematic.
Since I have observed in my openfoam that the cells are affected from the outlet, I hereby provide my b.c. for the outlet:

velocity: zeroGradient
pressure: fixed value (uniform, atmospheric pressure)
p_rgh: prghPressure
k: inletOutlet
epsilon: inletOutlet

Is there any problem in the B.C. here?
(To be honest, for the boundary conditions, I mostly followed the fluidised bed in the tutorial cases)
Another important thing is that I have actually run the same case before, using the icoFoam solver,which gives quite good results.
Thanks again!

[piu58]

You need a pressure reference point somewhere. I don't know whether atmospheric pressure gives it.

[cshsgy]

Originally Posted by piu58

You need a pressure reference point somewhere. I don't know whether atmospheric pressure gives it.

Actually what I am doing is simply setting the atmospheric pressure as internal field as an initial condition. I am not quite aware whether this gives so-called reference point

[cshsgy]

Hey guys
Update of this problem: I have consulted my professor, and he told me to focus on p_rgh instead of p alone. But the problem is that I could not find the suggested boundary conditions for p_rgh. Is there anything that we can refer to? Thanks guyz

[sharonyue]

Are you simulating a pipe flow? You mesh is same with mine.

I would suggest you switch off the turbulence to check if the problem comes from turbulence. The small value of deltaT implies your velocity's too large, which probably comes from a wrong prediction of turbulence fields.

Just a guess.

[cshsgy]

Originally Posted by sharonyue

Are you simulating a pipe flow? You mesh is same with mine.

I would suggest you switch off the turbulence to check if the problem comes from turbulence. The small value of deltaT implies your velocity's too large, which probably comes from a wrong prediction of turbulence fields.

Just a guess.

Hi Dongyue
Thanks a lot for your suggestion. Actually, just now I turned off the adjustable runtime, and make it fixed to a low enough but still acceptable fixed deltaT, and let it run from the beginning. Strangely, this seems to be working for the simplified single-phase case. I am now running it for two-phase, and I think it would be OK.
BTW thanks a lot for your website and your papers: it really helps me a lot to get more familiar with CFD and OpenFOAM.