[shereez234]

Dear OpenFOAM users.


Good afternoon and hope every one is staying safe and well.

I would like to know if any one had any experience/sucess in implementing the dual-time stepping method in OpenFOAM. I saw one implementation in Unofficial Foam-Extend Github where the implementation looks like:

 scalarField rDeltaTau = scalar(1.0)/(beta*deltaTau.internalField());     
 scalar coefft   = 1 + deltaT/(deltaT + deltaT0);              // = 3/2     
 scalar coefft00 = deltaT*deltaT/(deltaT0*(deltaT + deltaT0)); // = 1/2                           
 scalar coefft0  = coefft + coefft00;                    // = 4/2                                               

  fvm.diag() = (rDeltaTau+coefft*rDeltaT)*mesh().V();                        
   if (mesh().moving())               

        {        
    fvm.source() =  (                     
       rDeltaTau*vf.oldTime().internalField()*mesh().V()  + 
coefft0*rDeltaT*vfOld.internalField()*vfOld.mesh().V()   - 
coefft00*rDeltaT*vfOldOld.internalField()*vfOldOld.mesh().V()                        
         );              
     }                    

  else         
  {                          
     fvm.source() = mesh().V()*         
    (                         
      rDeltaTau*vf.oldTime().internalField()   + 
      coefft0*rDeltaT*vfOld.internalField()    - 
     coefft00*rDeltaT*vfOldOld.internalField()  
         );                     
    }                                   

   return tfvm;    scalar rDeltaT = 1.0/deltaT;

Just to be more clear, for the implemented dual time-stepping method, deltaT is the physical time step size, and deltaTau is the pseudo time step size.

The transfer of this source code to OpenFOAM 7 is not clear for me. Is this the right dual time stepping method as implemented by Jameson for large time step external flow CFD cases?



I hope to get some insight from some of you. Any help is appreciated.





Kind regards




ps: References of the code
References: Venkatakrishnan V. and Mavriplis D. J. , "Implicit Method for the Computation of Unsteady Flows on Unstructured Grids" Journal of Computational Physics, (1996), 127, p. 380-397, equation 10
Blazek, Jiri. "Computational Fluid Dynamics: Principles and Applications", 2001, Referex Engineering, Elsevier (Chapter 6.3)
Arnone, Liou and Povinelli, "Multigrid Time-Accurate Integration of Navier-Stokes Equations" AIAA 1993-3361, NASA-TM-106373

[lpz456]

Hello! Are there any development? Recently, I encountered the issue of excessively high residuals within a single physical time step while simulating unsteay compressible flows.

[shereez234]

If you want unsteady dual time stepping for compressible use the HISA code

[lpz456]

Thanks! But my case is related to cooling film with low speed about 0.07 Mach number and strong heat transfer. The Hisa solver is suitable for compressible transonic and supersonic flow.

[shereez234]

Mostly yes. But they have preconditioning for low mach number flows.