A few benchmarks

[rethore]

The repo should have a few test cases to showcase and test the models:

• Horns Rev I
• Nysted
• Middelgrunden
• Lillgrund

So that should include the input files, and some normalised validation datasets already available in public reports or articles. The demonstration should be done through Jupyter notebooks and the test should be done in a new directory call benchmarks, that should be run with the other unit tests.