Validation Carreau Yasuda Model

[sherin-joseph]

Please note that this is more of a steady-profile preservation check than a full convergence-from-rest validation, because the default and high-resolution runs use analytical initial conditions. . For a stronger transient validation, we need to use initial_condition_mode = zero or newtonian and verify convergence of the velocity profile toward the analytical solution.

The relative L2 velocity-profile error remains below 3% for all tested Carreau-Yasuda indices at ny = 200. The largest error occurs for n = 0.25, the strongest shear-thinning case, while the Newtonian limit n = 1 is recovered with about 1.4% relative L2 error.

The reference Poiseuille validation from Section 3.1 of Coclite et al. (2020) is shown below for comparison.

[svchb]

For validation cases we want to have the following structure:

1. example file that is trixi_include'd into the validation_run file
2. the validation run file contain some kind of error bounds
3. Additionally a plot file to plot the results of the validation run.

So you just need to create this additional example file and figure out an error bound for the validation to get to that structure.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.50%. Comparing base (02fc220) to head (5a18c37).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1207   +/-   ##
=======================================
  Coverage   90.50%   90.50%           
=======================================
  Files         141      141           
  Lines       11409    11410    +1     
=======================================
+ Hits        10326    10327    +1     
  Misses       1083     1083           

Flag	Coverage Δ
total	90.51% <100.00%> (+<0.01%)	⬆️
unit	77.76% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sherin-joseph]

@svchb
The PR now includes:

examples/fluid/poiseuille_carreau_2d.jl as the example file
validation/poiseuille_carreau_2d/validation_poiseuille_carreau_2d.jl as the validation runner using trixi_include
relative L2 error bounds in the validation runner
plot_carreau_comparison.jl for plotting the validation results

I hope this follows the expected structure now. Please let me know if any further changes or corrections are needed.

Thanks for the guidance!

[efaulhaber]

I apologize for the long review turnaround. We have a lot of open PRs at the moment.

The results look very promising. Could you please also add the analytical solution in the plots to be able to get a better feeling of how accurate the results are?

[sherin-joseph]

Thanks for the detailed review. I added the analytical profiles to the plots. These figures show the transient relaxation from the Newtonian initial profile toward the Carreau-Yasuda analytical profiles. The n = 0.25 case is much slower to converge, so I am rerunning the remaining cases for a longer final time and will update the figures once those runs finish.

I apologize for the long review turnaround. We have a lot of open PRs at the moment.

The results look very promising. Could you please also add the analytical solution in the plots to be able to get a better feeling of how accurate the results are?