Impossible mass generation at resistances!

[Anonymous]

I have got a problem with the mass flow/volume flow/velocity (whatever you want because of course it corresponds to each other) at resistances.

The geometry is a column where the entry is at the bottom and the exit is a pipe at the top. In the column there are two resistances which are in reality wired meshes. (look attached screenshot) The whole thing is symmetric in the x,y-area, so that I composed it as a half with the symmetry/slip boundary condition.

I simulated the resistances with the permeability method.

The flow and pressure loss seems to be allright, but at both of the resistances the mass flow increases at the beginning and decreases over the length of the resistance to the real value of 1000 m³/h I gave in the boundary conditions. (next screenshot with summary planes distributed over the resistance; 0 means beginning of resistance, '+' values are mm inside the resistance, '-' values are before the resistance; to present everything more clearly I only put planes in one of the resistances). This is of course not possible because you have to keep the mass balance.

The problem does appear in other simulations with resistances, too.

Why does it appear? And how can I fix it?

[Jon.Wilde]

Hi Florian,

Can I ask what your mesh looks like? We need a excellent mesh on resistances (4-5 elements from inlet to outlet and a uniform mesh).

Are you also using ADV5? This is probably the best Advection Scheme to use here.

Kind regards,

Jon

[apolo_vanderberg]

Can you check these same planes using the Bulk Calculator and not the Summary output?

The Summary output does not distinguish different regions such that if you hide (or had suppressed ) the steel wall, the Bulk calculator will present you with a mass flow for both region 1 and 2 (on either side of the wall), whereas the Summary file will SUM the regions making this comparison a little different than what you might be expecting.

[Anonymous]

Hi Jon,

I have got a mesh of about 3000000 total elements. The attachment shows the mesh at the resistance. I think that the must-have of 4-5 elements from inlet to outlet of the resistance is fulfilled.

Indeed I am using ADV 5. The number of iterations is more than 4000.

Kind regards,

Florian

[Anonymous]

I tried a closer look at the bulk function results of the planes. For my attached example I took the planes in scenario 1 and summarized them in an excel file.

As you can see the problem is still there. While the area which is in the stream without resistance shows quite constant flow rates, the other one with resistance varies strongly (about 20%).

Nice to see in the table of the excel sheet is that the increasement of volume flow is maximised at the beginning of the resistance and after it decreases exponentially to the right value.

[Jon.Wilde · Accepted solution]

Your mesh looks good, I agree that this is not the issue.

Have you tried the DRsmoothing flag set to -2? We often use this when there is recirculation near a resistance, it may help you here.