[gabrieltunoda]

Hello everybody.

I'm trying to analyze the flow on the shell side of a heat exchanger, however, due to the difficulty of modeling the mesh near the various tubes of the model, the idea of ​​representing the region with tubes as a porous domain appeared.
First, to define the permeability as a function of velocity, I created a periodic model of the finite region to verify what the average velocity would be by varying the pressure difference.



Once the average velocity was obtained (as a function of the pressure difference), I created a model of a porous domain equivalent to the periodic one and found what would be the required permeability to obtain the same average velocity with the same pressure difference.
Therefore, a function of Kx (u), Ky (v) and Kz (w) was obtained.



Now, I'm having difficulty in applying these functions to the complete model. How can I apply this function so that the permeability varies in each mesh element according to local speed?
Note: For this analysis I am disregarding the thermal exchange effects.

Thank you in advance.

[evcelica]

I would think a loss coefficient would be applicable as well. Permeability is the linear part of the loss equation, Kloss is the quadratic part.

Look up the function for pressure drop in the documentation, it has velocity in it.


Something like:

dP/dL = (mu/Permeability)*V + (Kloss/2)*rho*V^2

There are directional multiplier model you can use for changing the characteristics in each direction.

[gabrieltunoda]

Thank you very much for your attention. It worked here. Initially I was concerned about how to represent the permeability "K" and did not check the other input methods. The linear and quadratic coefficients were directly obtained in the first periodic model already based on the "Vmed" vs. "dp" curve.
Thanks again.