[Ramwi]

Dear all,

when writing a custom solver, I found a piece of code in tractionDisplacement boundary condition in solidDisplacementFoam that got me thinking. Specifically,

1. In readMechanicalProperties.H in the solidDisplacementFoam, the user-suplied Young modulus is loaded as rhoE and then it follows,

Info<< "Normalising E : E/rho\n" << endl;
volScalarField E(rhoE/rho);

Info<< "Calculating Lame's coefficients\n" << endl;

volScalarField mu(E/(2.0*(1.0 + nu)));
volScalarField lambda(nu*E/((1.0 + nu)*(1.0 - 2.0*nu)));
volScalarField threeK(E/(1.0 - 2.0*nu));

Q: So, the volScalarField registered as E is already normalized?

2. In tractionDisplacementFvPatchVectorField.C, there is

const fvPatchField<scalar>& rhoE =
        patch().lookupPatchField<volScalarField, scalar>("E");

and it follows,

scalarField E(rhoE/rho);
scalarField mu(E/(2.0*(1.0 + nu)));
scalarField lambda(nu*E/((1.0 + nu)*(1.0 - 2.0*nu)));
scalarField threeK(E/(1.0 - 2.0*nu));

3. So, it seemed to me, that the looked up field rhoE is already normalized and it is "normalized again" by scalarField E(rhoE/rho);.

Q: Did I make any completely obvious mistake in my train of thoughts?

Thanks for anyone kind enough to explain this to me.

Best wishes,
Martin

[Ramwi]

OK, I got lost in the different names of variables and registered objects. Henry Weller was kind enough to explain it to me. The problem for me was the fact that the variable rhoE in the code is registered as E and the variable E as "(rhoE/rho)".