Physical damping for the solid in solids4Foam
 

Dear all,

I’m currently investigating wave loading on flexible structures with solids4foam.
I was wondering if in the solid model/solution is included any kind of damping coefficient or effect. I’m using a linearElastic consitutive law and a
linearGeometryTotalDisplacement solid model.


Thanks,
Tommaso

By default, there is no physical damping included.

Depending on the temporal discretisation, however, numerical damping may be significant. For example, 1st order Euler typically requires very small time-steps to keep numerical damping small.

I am now wondering if the numerical damping may depend also on other factors, for example the spatial discretisation or the mechanical properties of the structure (e.g. mass, sfiffness..).

Primarily numerical damping for transient-dominant cases comes for temporal discretisation.

There are, however, other types of numerical damping, like spatial damping (spatial smoothing of the solution). For example, the Rhie-Chow smoothing term does this and in some cases it may be too large (particularly for elasto-plastic analyses, it may need to be scaled). The spatial discretisation in general will also contributed.

I guess the easiest way to check for your case is to try out different settings.

Philip

Hi all,

I’m trying to use a dynamic mesh for the solid to see if it also has a certain effect on the final displacements.

I modified constant/solid/dynamicMeshDict:

dynamicFvMesh dynamicMotionSolverFvMesh;

solver displacementSBRStress;

diffusivity quadratic inverseDistance (interface);

frozenDiffusion off;
distancePatches
(
interface
);

And I included pointDisplacement in 0/solid and the celldisplacement solver in system/solid/fvSolution.
The code runs without any errors.
However, the mesh of the solid still looks static and actually pointDisplacements are zero during the simulations, so it seems they are not solved.

Hi evxta3,

As most(/all) of the solid models currently included in solids4foam use a Lagrangian approach (as opposed to Eulerian), the mesh motion is dependent on the displacement solution field. As a result, it is incorrect to use any mesh motion with the Lagrangian solid models.

If you did want to implement an Eulerian (or arbitrary Lagrangian Eulerian) solid model with arbitrary mesh motion, you would need to reformulation the governing equation to include a convection term as well as advecting all fields needed to calculate the stress (e.g. the displacement gradient or deformation gradient).

Kind regards,
Philip

Thanks Philip.