HydroMechanics non-equilibrium initial states. #2561
Conversation
| @@ -0,0 +1,3 @@ | |||
| Non-equilibrium initial state variables which are not explicitly in equilibrium | |||
|
Will the changes in the local assembler be in another PR? |
|
Yes, the F^{oob} (out of balance forces) formulation will follow in other PR. |
The processes might change the solution, which is now passed as a non-const reference.
| @@ -188,13 +188,45 @@ std::unique_ptr<Process> createHydroMechanicsProcess( | |||
| config.getConfigParameter<double>( | |||
| "reference_temperature", std::numeric_limits<double>::quiet_NaN()); | |||
|
|
|||
| // Non-equilibrium variables | |||
| ParameterLib::Parameter<double> const* nonequilibrium_stress = nullptr; | |||
There was a problem hiding this comment.
Just out of curiosity: Why not ´unique_ptr`?
There was a problem hiding this comment.
It's a non-owning optional parameter.
| @@ -97,6 +101,12 @@ struct HydroMechanicsProcessData | |||
| ParameterLib::Parameter<double> const& porosity; | |||
| /// Solid's density. A scalar quantity, ParameterLib::Parameter<double>. | |||
| ParameterLib::Parameter<double> const& solid_density; | |||
|
|
|||
| /// Optional, non-equilibrium stress parameter. | |||
| ParameterLib::Parameter<double> const* const nonequilibrium_stress; | |||
| <Collection> | ||
| <DataSet timestep="0" group="" part="0" file="into_initial_state_pcs_0_ts_0_t_0.000000.vtu"/> | ||
| </Collection> | ||
| </VTKFile> |
|
Are the gml files needed? I didn´t see references in the prooject files. |
|
I guess that Tests/Data/HydroMechanics/InitialStates/pressure/Docu_HM_Non_Equilibrium_Initial_States.pdf is for the documentation of one benchmark. However the web content file, which should refer to that PDF is missing. |
| @@ -269,11 +302,11 @@ void HydroMechanicsLocalAssembler<ShapeFunctionDisplacement, | |||
|
|
|||
| // pressure equation | |||
| local_rhs.template segment<pressure_size>(pressure_index).noalias() -= | |||
| laplace_p * p + storage_p * p_dot + Kpu * u_dot; | |||
| laplace_p * (p - p_neq) + storage_p * p_dot + Kpu * u_dot; | |||
There was a problem hiding this comment.
I think that every simulation starts from the previous status, whether which is equilibrium (may from the previous simulation) or non-equilibrium (specfied). Therefore p_neq should be p0, and local_rhs should not be changed. This is a suggestion for future change. The PR itself is OK with the current consideration.
|
This should be no longer needed with the CompensateInitial.... |
|
That's true, but I still didn't redo the test cases... |
|
OpenGeoSys development has been moved to GitLab. See this pull request on GitLab. This issue / pull request was open before the GitLab migration. |
Adding same non-equilibrium initial states as for the SmallDeformation process ##2501.
Meanwhile the theoretical ideas were developed further and this is only for the comparison with the next new implementation and also provides the project files/tests. The new idea is to include the non-equilibrium in the residuum.