Case with SmagorinskyBGKdynamics diverges in v1.5 but runs fine in v1.4

[anandanandan86]

I’ve come across an issue trying to port a case I had set up for olb-1.4r0 over into olb-1.5r0. The case uses SmagorinskyBGKdynamics for the bulk flow. In olb-1.4r0 I set it up by the following constructor:

SmagorinskyBGKdynamics<T, DESCRIPTOR> bulkDynamics(converter.getLatticeRelaxationFrequency(), instances::getBulkMomenta<T, DESCRIPTOR>(), 0.1);

This constructor no longer works in olb-1.5r0 so for the new version, I set it up as a 2-step process that I gleaned from some of the examples instead:

using BulkDynamics = SmagorinskyBGKdynamics<T, DESCRIPTOR>;
sLattice.setParameter<collision::LES::Smagorinsky>(0.1);

T, DESCRIPTOR and all other details are kept the same between the 2 versions, but the olb-1.5r0 version diverges soon after the start of the calculation whereas the olb-1.4r0 version runs fine. Would anyone else have happened to encounter any similar issue, or have any idea as to how to resolve this?

[Adrian]

The basic approach of using the new defineDynamics call and setting the parameter seperately is correct.

In which order do you execute the calls? You need to first set the dynamics and then set the parameters (not just collision::LES::Smagorinsky but also OMEGA and so on).

If this doesn’t help I’ll have to see more of the code.

• This reply was modified 2 years, 3 months ago by Adrian.

[Author]

Posts