PowerGrid metrics are label dependant

[jeromedejaegher]

Hi,

On a Rte project, we think about using all or part of LIPS framework. However, PowerGrid metrics are label dependant.
For example, "verify_kirchhoff_law" uses "YBus", "p_or", "p_ex", q_or", "q_ex". This means that we cannot verify kirchoff law on DC approximation, or if we label variables as var_in and var_out.

Moreover, it is not clear how to build the YBus var from a given topology.
On "verify_joule_law", it is not clear why the resistance value is equals to the ohm_per_km value...
Default values and/or constructor and/or documentation would help using this project on real networks data.

Thank you,

Jérome.

[Mleyliabadi]

Hi,

I have added a new functionality in LIPS to be able to change the name of the variables dynamically whether using a specific configuration file or using directly a Data Class concept. These updates are available on a fork repository and a specific branch: 18-powergrid-metrics-are-label-dependant. I will merge them once their stability are proved.

A Jupyter Notebook containing an example of how to use this new functionality is provided.

The construction of the YBus var from a given topology and resistance values are out-of-scope of the LIPS framework.

Don't hesitate to use and test this new functionality and share with us your feedbacks.

Thanks,

Milad