Varying pCoeff: DATopoCHTFoam

[Ananth-Narayan-IITM]

Hello all,

Based on the tutorial of Channel_TopoCHT, for constant pCoeff, I would like to vary from 0.01 to 1 exponentially based on Major iteration limit (for SNOPT). Now question is what is the variable that tracks the major iterations such that I can vary the pCoeff vary linearly or exponential.

If not, is there any better method to keep track of this.

Thank you

[friedenhe]

This is not possible with the current DAFoam setup. If you really want to implement this, a dirty workaround is to hard-code a counter in the adjoint solution and use that as the "major iterations". That would need quite some modifications to the code...

[Ananth-Narayan-IITM]

Hello Prof. Ping,

I did some workaround in modifying class DAOPTIONS by modifying dictionary in runScript.py which is then transferred to OpenFoam solver.

I also modified mphys.py code where I found solution_counter tracks this major iteration such that the code definition updates pCoeffs based on solution_counter.

Now I have the code that can change that pCoeffs after certain major iterations.

The codes are under testing and development. However I do get some descent result but will consolidate them and raise PR for Topology Optimization once I feel things are stable.