[wildfire230]

Hi all,


I am using a custom solver based on the simpleFoam and scalarTransportFoam solvers. Essentially, I use simpleFoam to calculate the steady state velocity/pressure, and then I use scalarTransportFoam to update a transported quantity, in my case a concentration. I update the concentration for several time steps and then once again calculate the pseudo-steady-state velocity/pressure and repeat. The issue I am having is that my viscosity is a field variable that is a function of the concentration. I have implemented this successfully using a new viscosityModel, but now I am a little concerned about the solver itself. For example, in the Navier-Stokes equations, there are extra terms that are needed in the case of variable viscosity. For example, gradients in viscosity can potentially drive fluid motions on their own. Does simpleFoam already consider these terms in the way it is implemented, or do I need to modify the equations being solved in order to account for the possibility of variable viscosity?


Thanks so much

[decah]

Sounds to me like you need a transient solver? Is the concentration is changing the viscosity in certain regions only or is it something continuous?

I work with the drift flux solver which uses user-defined viscosity models with the PISO solver, maybe that could work for you if you are working with concentrations of dispersed particles

[LuckyTran]

Variable viscosity is already handled. That's because all the solvers that use turbulence models already implement it as muEff/nuEff and it's quite obvious that the eddy viscosity is not spatially uniform. Even if it were not due to nuEff...

There are no extra terms in the Navier-Stokes if you are familiar with how transport equations look like in FVM. The viscosity sits inside the divergence, there is no gradient of viscosity.

[wildfire230]

Ok great, thanks! I was thinking in terms of expanding out the equations and not in terms of FVM. That is helpful. That's what I was hoping for.