CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Pseudo-transient Lagrangian approach (https://www.cfd-online.com/Forums/openfoam-solving/227190-pseudo-transient-lagrangian-approach.html)

 ASimonsen May 20, 2020 08:47

Pseudo-transient Lagrangian approach

Hi all

I'm carrying out a Lagrangian simulation. I've previously used Ansys Fluent, where I was able to solve the Lagrangian phase in an unsteady manner, while the continuous carrier phase was solved using steady-state methods. Basically, the solution process was as follows:
(1) Inject and advance the Lagrangian phase by "dt".
(2) Update source terms "seen" by the continuous phase
(3) Carry out N steady-state iterations in the continuous phase.
(4) Go back to step (1)

This means that the solution approaches steady-state much faster compared with a simulation, where both the continuous and Lagrangian phases are treated in a transient manner - i.e.:
(1) Inject and advance the Lagrangian phase by "dt"
(2) Update source terms "seen by the continuous phase
(3) Advance the continuous phase by "dt"
(4) Go back to step (1)

In OpenFOAM 7.0, I have used the "reactingParcelFoam" solver, which supports LTS (Local Time Stepping) methods for solving the continuous phase - great! But when I'm enabling this using the "localEuler" ddtSchemes, the solver insists on tracking the parcels "completely" each time the parcels->evolve() method is called - i.e. the parcels are injected and tracked until they either evaporate or exists the domain. This is not what I want, as it takes a long time to carry this operation out. Instead, I'd like to track the parcels in an unsteady manner (i.e. advance by "dt" for each parcel->evolve()), and solve the continuous phase using the "localEuler" scheme (Or steady-state ideally).
I know this is possible, but I can't seem to get it to work. Is there any simple fix to force the parcels/Lagrangian phase to be treated in a unsteady manner while using the LTS scheme?
And how come that this feature isn't available in OpenFOAM already?

Kind regards Anders

 All times are GMT -4. The time now is 02:29.