[chinaduck]

Hello All,

I have met a problem regarding the dam break simulation. I am developing a VOF code based on the PISO method. When I simulate a dam break flow, the water ftont becomes a "step" shape as shown in the attached figures. I have checked the code, nothing found. If you have any suggestion, do let me know. Thanks for your instructions.

Best,

Peter

[Parisa]

Hi,

I don't know really anything about how your code can be wrong. But from my VOF experience you usually need a finer mesh resolution where the interface is. Just a first thought that comes to my mind, can it be that your mesh on the interface is too coarse and it is difficult to capture the physics in this region and it is therefore you see this effect?
As I can see at that part of the domain it is very little water and mostly interface, I can imagine that it is hard to capture it differently. I recommend that before you change your code, try creating a mesh with a finer grid where you expect the interface to be.

Hope that helps!

//Parisa

[Alex C.]

1) Have you fully tested your N-S solver for a single phase? If your flow solution is not correct, you can't expect that the VOF solution will be good.

2) Have you performed a convergence analysis on your solution? Such an analysis needs to cover both spatial and temporal discretization.

3) By integrating your solution, you should be able to calculate the mass of each phase at each time-step. Each phase's mass should remain constant. (Assuming there is no flow across the boundaries, if there is a flow across your boundaries, you can integrate the normal flow across the boundaries and add this term to the conservation equation)

[chinaduck]

Thanks for your response. I have tested the single phase flow, it works well. And also, I have tested the mass convergence property. it also is reasonable.

Now, I am thinking about the method to compute the density on the control surface of the control volume.

I directly use the averged density from the Cell P and its neighbour Cell N.

I am continuing to check the code, once I fix it, I will tell all of you!

Best,

Peter

Originally Posted by Alex C.

1) Have you fully tested your N-S solver for a single phase? If your flow solution is not correct, you can't expect that the VOF solution will be good.

2) Have you performed a convergence analysis on your solution? Such an analysis needs to cover both spatial and temporal discretization.

3) By integrating your solution, you should be able to calculate the mass of each phase at each time-step. Each phase's mass should remain constant. (Assuming there is no flow across the boundaries, if there is a flow across your boundaries, you can integrate the normal flow across the boundaries and add this term to the conservation equation)

[chinaduck]

Thanks for your response. Actually, I have tested the fine mesh case, the problem is eliminate to some extent, while it still has some similar shape of the water front as shown in the attached figure. I am continuing to work on this. Once I fix it, I will let you know. Thanks again for your help and instruction.

Best,

Peter

Originally Posted by Parisa

Hi,

I don't know really anything about how your code can be wrong. But from my VOF experience you usually need a finer mesh resolution where the interface is. Just a first thought that comes to my mind, can it be that your mesh on the interface is too coarse and it is difficult to capture the physics in this region and it is therefore you see this effect?
As I can see at that part of the domain it is very little water and mostly interface, I can imagine that it is hard to capture it differently. I recommend that before you change your code, try creating a mesh with a finer grid where you expect the interface to be.

Hope that helps!

//Parisa

[virendra_p]

Does this shape persists over all transient vof contour plots?..if yes..then you can switch to higher order intetface tracking schemes...