CL solver improvements

[anilyil]

Purpose

This PR overhauls the CL solver in ADflow. Several features are added with the goal of increasing solver robustness and performance. The solver used to take about 3 times the cost of a standalone analysis for a CL solve with 1e-6 tolerance. Now it takes about 1.1-1.5 times a standalone analysis cost. We also print more information during iterations.

The features added to the solver is relatively easy to understand, so I won't explain them here. The biggest change is the ability to provide a relative convergence target, or a list of relative convergence targets that are used at each iteration. Besides checking for CL convergence, solver also checks if CFD L2 convergence is achieved. If not, we keep running more iterations. This is responsible from bulk of the performance improvements.

One major change separate from the CL solver is the modifications to the ANK solver to handle restarts better. Originally, we were setting ANK CFL to the initial ANKCFL0 value at each restart. The default is 5.0. Based on nonlinear convergence, ANK CFL would usually end up slightly higher than this and this has been working well enough for most optimization cases.

However, when we do a CL solve type iteration, there is no point to dump the last ANK CFL. The results now jump only due to the change in AoA, and the flowfield correction. As a result, we should be able to just keep whatever ANK CFL we had in the last iteration.

To do this, I am now saving the ANK_CFL in aero problems' adflowData attribute, which is a small object to keep some ADflow data. If the ADflow option ANKCFLReset is set to False, then we replace the ANK CFL from what it was in the last iteration. The default value of ANKCFLReset is True, and this recovers the original ANK behavior where we initialize the ANK CFL every time the solver restarts.

Expected time until merged

After #231 is merged, I need to update this PR. Currently it looks like there are tons of changes but in reality this is a standalone feature. I also need to add tests.

Once I do these, it should be ready to be merged within a week after reviews.

Type of change

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (non-backwards-compatible fix or feature)
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Documentation update
• Maintenance update
• Other (please describe)

Testing

Checklist

• I have run flake8 and black to make sure the Python code adheres to PEP-8 and is consistently formatted
• I have formatted the Fortran code with fprettify or C/C++ code with clang-format as applicable
• I have run unit and regression tests which pass locally with my changes
• I have added new tests that prove my fix is effective or that my feature works
• I have added necessary documentation

[anilyil]

Closing this PR to create a separate one with a clean history for the cl solver