Multicomponent entropy method - average the looked-up material properties #5622
Conversation
| /** | ||
| * The thermodynamic lookup equation of state. | ||
| */ | ||
| EquationOfState::ThermodynamicTableLookup<dim> equation_of_state; |
There was a problem hiding this comment.
So here you introduce a new member object into the EntropyModel class. But I dont see where you actually use this member. Is it necessary? The ThermodynamicTableLookup duplicates a lot of the functionality of the EntropyModel class, and eventually we may want to replace some parts of EntropyModel by using ThermodynamicTableLookup instead, but that doesnt seem like it happened in this PR.
| @@ -163,6 +164,9 @@ namespace aspect | |||
| const unsigned int projected_density_index = this->introspection().compositional_index_for_name("density_field"); | |||
| const unsigned int entropy_index = this->introspection().compositional_index_for_name("entropy"); | |||
|
|
|||
| std::vector<EquationOfStateOutputs<dim>> eos_outputs (in.n_evaluation_points(), material_file_names.size()); | |||
There was a problem hiding this comment.
I think you do not need a std::vector of EquationOfStateOutputs. If you check the for-loop below, you only ever access eos_outputs[i] at a time, you never need all of eos_outputs. So you may as well make this simply a EquationOfStateOutputs<dim> eos_outputs(material_file_names.size()) and then access that single object in the loop below. This is different for the Steinberger material model (it accesses the whole vector of eos_outputs in line 278).
The same is true for the volume_fractions object in the next line.
| for (unsigned int j=0; j<material_file_names.size(); ++j) | ||
| { | ||
| eos_outputs[i].vp[j] = entropy_reader[j]->seismic_vp(entropy, pressure); | ||
| eos_outputs[i].vs[j] = entropy_reader[j]->seismic_vs(entropy,pressure); | ||
| } | ||
|
|
||
| seismic_out->vp[i] = MaterialUtilities::average_value (volume_fractions[i], eos_outputs[i].vp, MaterialUtilities::arithmetic); | ||
| seismic_out->vs[i] = MaterialUtilities::average_value (volume_fractions[i], eos_outputs[i].vs, MaterialUtilities::arithmetic); |
There was a problem hiding this comment.
So this is the place why you introduced vp and vs into EquationOfStateOutputs. But you should be able to write this just as well as:
| for (unsigned int j=0; j<material_file_names.size(); ++j) | |
| { | |
| eos_outputs[i].vp[j] = entropy_reader[j]->seismic_vp(entropy, pressure); | |
| eos_outputs[i].vs[j] = entropy_reader[j]->seismic_vs(entropy,pressure); | |
| } | |
| seismic_out->vp[i] = MaterialUtilities::average_value (volume_fractions[i], eos_outputs[i].vp, MaterialUtilities::arithmetic); | |
| seismic_out->vs[i] = MaterialUtilities::average_value (volume_fractions[i], eos_outputs[i].vs, MaterialUtilities::arithmetic); | |
| std::vector<double> vp (material_file_names.size()); | |
| std::vector<double> vs (material_file_names.size()); | |
| for (unsigned int j=0; j<material_file_names.size(); ++j) | |
| { | |
| vp[j] = entropy_reader[j]->seismic_vp(entropy,pressure); | |
| vs[j] = entropy_reader[j]->seismic_vs(entropy,pressure); | |
| } | |
| seismic_out->vp[i] = MaterialUtilities::average_value (volume_fractions[i], vp, MaterialUtilities::arithmetic); | |
| seismic_out->vs[i] = MaterialUtilities::average_value (volume_fractions[i], vs, MaterialUtilities::arithmetic); |
Can you give this a try?
| eos_outputs[i].densities, | ||
| true); | ||
|
|
||
| out.densities[i] = MaterialUtilities::average_value (volume_fractions[i], eos_outputs[i].densities, MaterialUtilities::arithmetic); |
There was a problem hiding this comment.
Yes, this is exactly what you want and should work just fine. Can you already add a test case that shows the functionality is working, or are there any pieces missing? I would guess that in the current state the entropy to temperature conversion would only happen correctly for the first composition table, but all other material properties are already averaged correctly?
RanpengLi
changed the title
RanpengLi
changed the title
|
This pull request is ready for another review. Thank you! |
RanpengLi
force-pushed
the
read-multi-table-eos-output
branch
2 times, most recently
from
038b214
to
731d9c3
Compare
|
This pull request is ready for another review. Thank you again! |
|
You need to run |
RanpengLi
force-pushed
the
read-multi-table-eos-output
branch
from
4333d13
to
52e41df
Compare
|
Done! |
| @@ -161,7 +164,18 @@ namespace aspect | |||
| MaterialModel::MaterialModelOutputs<dim> &out) const | |||
| { | |||
| const unsigned int projected_density_index = this->introspection().compositional_index_for_name("density_field"); | |||
| const unsigned int entropy_index = this->introspection().compositional_index_for_name("entropy"); | |||
| //TODO : need to make it work for more than one field | |||
| const std::vector<unsigned int> entropy_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::entropy); | |||
There was a problem hiding this comment.
| const std::vector<unsigned int> entropy_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::entropy); | |
| const std::vector<unsigned int> &entropy_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::entropy); |
| //TODO : need to make it work for more than one field | ||
| const std::vector<unsigned int> entropy_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::entropy); | ||
| const unsigned int entropy_index = entropy_indices[0]; | ||
| const std::vector<unsigned int> composition_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::chemical_composition); |
There was a problem hiding this comment.
| const std::vector<unsigned int> composition_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::chemical_composition); | |
| const std::vector<unsigned int> &composition_indices = this->introspection().get_indices_for_fields_of_type(CompositionalFieldDescription::chemical_composition); |
…mass of different chemical compositions
RanpengLi
force-pushed
the
read-multi-table-eos-output
branch
from
52e41df
to
e0e6d8b
Compare
Following #5615 which allows the entropy model to read in multiple files, I make the entropy model loop over look-up tables for all components, then average the material properties according to the components' volume or mass fractions.