FluxTubes: Alternative approximation

[AHaumer]

When working with FluxTubes, I found it nearly impossible to implement new materials using the approximation formula of Roschke. Additionally, when the model drives the material very high into saturation ("overfluxed region"), the behaviour is not always as expected (e.g. polarization J doesnt show a horizontal asymptote but decreases with increasing field strength). Unfortunately the original authors are no longer available.
So I decided to implement an alternative approximation: Smoothing Splines with Exponential Extrapolation.
Everything is explained in the documentation, and there will be a contribution to the Modelica 2027 conference.
Additionally, VoestAlpine allowed me to use their measured data of fully processed electric sheet (as used in rotating electric machines). So I was able to add 10 material records. The new parameter records contain the used raw data, so the approximation is traceable. It should be easier for users to implement their own material.
Additonally, 4 new examples have been implemented to demonstrate and test the new approximation thoroughly.
You might call it an enhancement, but it solves the users' problem of implementing new material records.
The solution is 100% backwards-compatible, and has been tested with (in alphabetical order) Dymola, OpenModelica and WolframSystemModeler.

[AHaumer]

One of the original authors, Thomas Boedrich, wrote to me:
Unfortunately we (he and Johannes Ziske) do not have the ressources to review that PR, but we fully agree with you that this is a valuable and important enhancement.

[AHaumer]

I have a set of functions to determine the approximation parameters from raw data.
These functions run both in Octave and Matlab.
Should I provide the code (and the raw data files)? In Ressources/xxx?