[ConvDiffApp] Adding Substepping thermomechanical solver #12126
Conversation
There was a problem hiding this comment.
Though this is a long-demanded feature, I think we should discuss it carefully. At first glance I see two important things in here:
- Time advance has been so far handled at the
AnalysisStagelevel. Changes in this PR go against this. - Substepping strategy must be the same across applications.
|
Hi! Let me comment your questions
|
applications/ConvectionDiffusionApplication/python_scripts/coupled_structural_thermal_solver.py
Outdated
Show resolved
Hide resolved
...nvectionDiffusionApplication/python_scripts/coupled_structural_thermal_substepping_solver.py
Outdated
Show resolved
Hide resolved
...nvectionDiffusionApplication/python_scripts/coupled_structural_thermal_substepping_solver.py
Outdated
Show resolved
Hide resolved
|
|
||
| super(CoupledThermoMechanicalSubsteppingSolver, self).__init__(model, custom_settings) | ||
|
|
||
| self.var_utils = KratosMultiphysics.VariableUtils() |
There was a problem hiding this comment.
| self.var_utils = KratosMultiphysics.VariableUtils() |
this should be instantiated in place
| @@ -46,6 +46,8 @@ def CreateSolverByParameters(model, solver_settings, parallelism): | |||
| # Coupled mechanical-thermal solver | |||
| elif (solver_type == "thermo_mechanically_coupled" or solver_type == "ThermoMechanicallyCoupled"): | |||
| solver_module_name = "coupled_structural_thermal_solver" | |||
| elif (solver_type == "thermo_mechanically_coupled_substepping" or solver_type == "ThermoMechanicallyCoupledSubstepping"): | |||
There was a problem hiding this comment.
| elif (solver_type == "thermo_mechanically_coupled_substepping" or solver_type == "ThermoMechanicallyCoupledSubstepping"): | |
| elif solver_type == "thermo_mechanically_coupled_substepping": |
new solvers should only use the json style
…pled_structural_thermal_solver.py Co-authored-by: Vicente Mataix Ferrándiz <[email protected]>
…pled_structural_thermal_substepping_solver.py Co-authored-by: Vicente Mataix Ferrándiz <[email protected]>
…pled_structural_thermal_substepping_solver.py Co-authored-by: Philipp Bucher <[email protected]>
|
|
||
| solid_is_converged = self.structural_solver.SolveSolutionStep() | ||
|
|
||
| self.RemoveConvectiveVelocity() |
There was a problem hiding this comment.
the database does not go back to the original status for the past steps, is that correct?
|
state of this? |
❄️ |
Ironic, because is thermomechanic... |
|
Is this in the TO-DO list of the @KratosMultiphysics/technical-committee ? |
It is since long ago. |
|
@rubenzorrilla and @RiccardoRossi will do some brainstorming about this and come back with a concrete proposal |
📝 Description
In this PR I am adding a substepping version of the thermomechanical coupled solver.
A coupled one-way thermo-mechanical solver in which the time step of the two solvers is independent. It is assumed that the dt of the structural part is equal or lower than the thermal one.