Skip to content

rhoCentralDyMFoam + maxwellSlipU = Crash

Summary

Using rhoCentralDyMFoam to simulate a case with dynamic mesh and maxwellSlipU boundary condition results in immediate crash at the first time step.

Steps to reproduce

Take the movingCone tutorial case for rhoCentralDyMFoam and change one of the slip boundary conditions for U to maxwellSlipU. Start the simulation.

Example case

Here movingCone.zip is said movingCone tutorial case with maxwellSlipU boundary condition. Just create the mesh with blockMesh and use rhoCentralDyMFoam to start the simulation.

Environment information

  • OpenFOAM version : v2006, v2012
  • Operating system : Linux
  • Compiler : GCC

Possible fixes

The reason for the crash is as following: at the beginning of the time loop, mesh.update() is called, where the velocity boundary field is updated: U.correctBoundaryConditions(). There, the maxwellSlipU boundary condition is called looking for the volTensorField tauMC. However, this volTensorField is created inside the time loop of rhoCentralDyMFoam after (!) mesh.update() is called. Therefore, it cannot be found.

A simple fix would be to create the volTensorField tauMC in createFields.H and then update it at its current position. Therefore, it would be present when mesh.update() calls the maxwellSlipU boundary condition at the very first time step.