fix a bug in the latent heat material & add min and max viscosity #1430
Conversation
source/material_model/latent_heat.cc
Outdated
| @@ -208,6 +208,7 @@ namespace aspect | |||
| // this means, that there are no actual density "jumps", but gradual | |||
| // transition between the materials | |||
| double phase_dependence = 0.0; | |||
| double viscosity_phase_dependence = 1.0; | |||
There was a problem hiding this comment.
Please update the comment above for this variable
There was a problem hiding this comment.
Also there is a comment in line 181 saying that this part is only computing density, maybe extend that to mention also the phase dependence of the viscosity
source/material_model/latent_heat.cc
Outdated
| @@ -228,6 +229,7 @@ namespace aspect | |||
| i); | |||
|
|
|||
| phase_dependence += phaseFunction * density_jumps[i]; | |||
| viscosity_phase_dependence *= 1 + phaseFunction * (phase_prefactors[i+1]-1); | |||
There was a problem hiding this comment.
why do you use phase_prefactors[i+1] but density_jumps[i]? Maybe there is a reason, but this asymmetry looks strange.
source/material_model/latent_heat.cc
Outdated
| @@ -449,6 +453,14 @@ namespace aspect | |||
| "viscosity for each phase. " | |||
| "List must have one more entry than Phase transition depths. " | |||
| "Units: non-dimensional."); | |||
| prm.declare_entry ("Minimum viscosity", "1e19", | |||
| Patterns::List (Patterns::Double(0)), | |||
There was a problem hiding this comment.
Should this really be a list? Or just a double?
source/material_model/latent_heat.cc
Outdated
| "Limit for the minimum viscosity in the model. " | ||
| "Units: Pa s."); | ||
| prm.declare_entry ("Maximum viscosity", "1e24", | ||
| Patterns::List (Patterns::Double(0)), |
source/material_model/latent_heat.cc
Outdated
| } | ||
|
|
||
| // as the phase viscosity prefactors are all applied on top of each other, we have | ||
| // to scale them here so that they are relative gactors in comparison to the phase above |
There was a problem hiding this comment.
I would rephrase this a bit (it took me some time to understand what you mean). What about: "as the phase viscosity prefactors are all applied multiplicatively on top of each other, we have to scale them here so that they are relative factors in comparison to the product of all phase above"
source/material_model/latent_heat.cc
Outdated
| @@ -228,6 +229,7 @@ namespace aspect | |||
| i); | |||
|
|
|||
| phase_dependence += phaseFunction * density_jumps[i]; | |||
| viscosity_phase_dependence *= 1 + phaseFunction * (phase_prefactors[i+1]-1); | |||
source/material_model/latent_heat.cc
Outdated
| @@ -242,6 +244,7 @@ namespace aspect | |||
| phase_dependence += phaseFunction * density_jumps[i] * (1.0 - composition[0]); | |||
| else if (transition_phases[i] == 1) // 2nd compositional field | |||
| phase_dependence += phaseFunction * density_jumps[i] * composition[0]; | |||
| viscosity_phase_dependence *= 1 + phaseFunction * (phase_prefactors[i]-1); | |||
jdannberg
force-pushed
the
latent_heat_material
branch
from
9cd7c3e
to
189be1f
Compare
|
Okay, I think I addressed all of your comments, and I also added a test. |
tests/latent_heat_viscosity.prm
Outdated
| @@ -0,0 +1,122 @@ | |||
| # A simple setup for for using the GPlates interface in a 2d shell. See the | |||
There was a problem hiding this comment.
can you update the comment for this test please?
jdannberg
force-pushed
the
latent_heat_material
branch
from
189be1f
to
0f2fed8
Compare
This pull request fixes a bug in the latent heat material model: there was an input parameter
Viscosity prefactorsthat was documented and could be set in the input file, but was never used in the code and hence had no actual effect. With these changes, the parameter now does what the documentation says. In addition, parameters for setting a min and max viscosity is introduced. This brings the material to the state we used in our manuscript about numerical methods in ASPECT.I also want to see if this changes any tests, if not, I will add one.