Fix random state for kmean clustering

[koen-vg]

When the kmeans algorithm is used to cluster networks, this is not deterministic by default. The result is that repeated runs of the simplify_network and cluster_network rules can and usually do produce different results that vary randomly. This makes results less reproducible when given only a pypsa-eur configuration file.

Changes proposed in this Pull Request

The fix is to supply a fixed random state to the k-means algorithm.

It might be considered if this should be done PyPSA itself, in the busmap_by_kmeans functions in networkclustering.py. That's an equally valid option, but maybe a bit opinionated.

I have not checked in a very rigorous way if there are any other sources of randomness in the network building process, but I don't think so.

Checklist

• I tested my contribution locally and it seems to work fine.
• N/A Code and workflow changes are sufficiently documented.
• N/A Newly introduced dependencies are added to envs/environment.yaml and envs/environment.docs.yaml.
• N/A Changes in configuration options are added in all of config.default.yaml, config.tutorial.yaml, and test/config.test1.yaml.
• N/A Changes in configuration options are also documented in doc/configtables/*.csv and line references are adjusted in doc/configuration.rst and doc/tutorial.rst.
• A note for the release notes doc/release_notes.rst is amended in the format of previous release notes.

[fneum]

That's a good idea!