Skip to content

cyclicAMI requires up-to-date neighbourPatch field

Summary

cyclicAMI in non-blocking mode (#2963 (closed)) produces different results

Steps to reproduce

Run incompressible/pimpleFoam/laminar/mixerVesselAMI2D/mixerVesselAMI2D tutorial even for one time step. Initial residuals on U of 2nd pimple are different from v2306.

Example case

See above

What is the current bug behaviour?

The stress term (nuEff*dev2(T(grad(U))))) is not consistent across the cyclicAMI. In the old formulation it would be calculated - in the new one it is stored/cached and updated during the evaluation. However this intermediate calculation is not evaluated.

What is the expected correct behavior?

Same as v2306.

Relevant logs and/or images

Environment information

  • OpenFOAM version : v2307
  • Operating system :
  • Hardware info :
  • Compiler :

Possible fixes

This is can be solved by

@mark