[nnprashant]

Is there any way to adjust the porous domain parameters iteratively using cel to reach the target pressure loss ?

For example:
Target pressure loss = 20 mbar
Assumption: Pressure loss only due to linear coefficient (Porous domain)
To avoid user input , can one use a cel script inside a porous domain to iteratively change the linear coefficient used in Direction/Loss model ?

Any input would be helpful! Cheers !

[ghorrocks]

You cannot do iterations in CEL.

I would do this by solving it at several coefficient values and plot the function to estimate the correct coefficient. It won't take long to do and it is simple - your suggestion will introduce new stability problems and is potentially quite difficult to keep stable.

[nnprashant]

If the geometry was simpler, i could have done what you have suggested. I am trying to compute the flow behaviour inside a fuel cell stack (around 300 plates). The steady state simulation is computationally intensive and it takes about 48 Hours for a simulation to finish.

Currently, there is no testbench data for the simulated model. Rather a pressure drop is measured for a certain volumetric flow.Trying to do a parameter study is not feasible.

If there is a way to change the porous coefficients dynamically during the simulation, the periodicity of change would only occur every 50 iterations or more and certainly not every iteration

[ghorrocks]

nnprashant - You appear to be forgetting that iterating for your porosity parameter is not a trivial exercise. You will need to make it numerically stable and converge. And you say you can do that inside a model which appears to be large and takes 48 hours to converge to steady state? You and making a simulation which is challenging because it is large and making it challenging squared by trying to iteratively solve a numerical problem inside it.

If you do it the simple way I recommend then yes, you are going to have to run it a few times. But you are going to have to run it MORE with your approach, so your approach will be slower. And that is assuming you can get your approach working - which is challenging in itself.

Why not just do the simple way and you can be confident you will have a result in a week or two, compared to your complex approach and I guarantee you won't have it finished in a month