Add a virtual PhaseFunctionBase class

[lhy11009]

Pull Request Checklist. Please read and check each box with an X. Delete any part not applicable. Ask on the forum if you need help with any step.

Introduction to the PR

In this pull request, I introduce a new class, PhaseFunctionBase, which serves as a pure virtual base class for different implementations of the PhaseFunction class. My motivation for this design stems from the need to create a PhaseFunction that returns discrete values (e.g., 0 or 1) based on a table lookup.

To provide more context, I aim to reuse much of the existing structure from the PhaseFunction class originally implemented by @gassmoeller . For instance, the vector structure for storing phase function values will remain intact, and we will continue parsing the relevant data from the parameter file.

The main changes involve overriding several key functions, specifically:

• compute_value: This will now perform a table lookup instead of using continuous functions.
• compute_derivative: Since an analytical solution is not applicable in this case, we may adopt a perturbation-based approach, similar to how viscosity derivatives are computed in the material model.
• declare_parameters and parse_parameters: These will be reloaded to accommodate the changes in functionality.

As you can see in the current format of the PR, the new class is purely virtual, which provides flexibility for future phase function implementations. However, I am unsure if making the base class purely virtual is the best approach for this problem. I would appreciate feedback on whether this is the right path forward.

Before your first pull request:

• I have read the guidelines in our CONTRIBUTING.md document.

For all pull requests:

• I have followed the instructions for indenting my code.

For new features/models or changes of existing features:

• I have tested my new feature locally to ensure it is correct.
• [] I have added a changelog entry in the doc/modules/changes directory that will inform other users of my change.

[lhy11009]

Another thread: Question on PhaseFunction<dim>::compute_derivative Implementation

I have a question regarding the existing implementation of PhaseFunction<dim>::compute_derivative. While reviewing the code, I noticed that if the transition width is set to 0.0, the function returns 0.0. Conceptually, I would expect that a transition width of 0.0 represents a very sharp transition, so I'm wondering why the value is set to 0.0 in this case.

Could you provide some clarification on the reasoning behind this behavior?