[hrushi.397]

I am getting initial residual of zero. What does that mean exactly?

[fumiya]

I have no idea about what problems you are solving
but I think it is possible depending on your boundary
and initial conditions and geometry.

Which solver are you using?
Whose initial residual is zero? (I guess it is one component of the momentum equation.)

Fumiya

[hrushi.397]

Originally Posted by fumiya

I have no idea about what problems you are solving
but I think it is possible depending on your boundary
and initial conditions and geometry.

Which solver are you using?
Whose initial residual is zero? (I guess it is one component of the momentum equation.)

Fumiya

Hi Fumiya,

Thank you for the reply.

I have attached the files herewith. I am trying to solve a diffusion reaction equation (Allen Cahn equation). I modified laplacianFoam because the equation also has a laplacian term. I modified dambreak case to generate 64x64 grid which is intitalized randomly to 1 or -1. Boundary conditions are cyclic. I fail to get any residual at the beginning.

I am also printing laplacian of phi, field which I am trying to solve, and it turns out to be unusually large. But I donot understand why this is happening.


I have tried every solver. Currently it is diagonal. I have implicit discretization and explicit discretization for laplacian.

One more thing: T tried to solve same in Matlab and it works fine, but not in Openfoam. Could you please help me?

Thanks and regards,

Hrushi

[hrushi.397]

Sorry, Dambreak is not small enough to upload. Uploading files individually.

[hrushi.397]

Hi All,

I found the reason why laplacian results were unphysical. The phi^3 term in the implementation was the problem. When I used an implicit source term as fvm::Sp(phi*phi,phi), I could get rid of zero initial residual and segmentation fault. Although I have got the results, they are not physical though. What could be the reason? Should I add fvm::Sp(3*phi*phi,phi) in accordance with whatever linearization Patankar has suggested?

Thanks