Ambiguity and mandatory requirement of 'max_loading_percent' in N-1 contingency analysis #2684
Description
Problem Description
When using run_contingency, I encountered two issues related to the handling of the max_loading_percent attribute:
1. Mandatory presence of max_loading_percent
If the elements in the network (e.g. net.line) do not have a max_loading_percent column, the N-1 analysis fails with a KeyError.
- In practice, this attribute should not always be mandatory: for N-0 runs, the calculation works fine without it, but for N-1 it breaks because _update_contingency_results tries to access it unconditionally.
- This makes it impossible to run contingency analysis on networks that do not define this attribute.
2. Ambiguity of the attribute’s meaning
The name max_loading_percent is used in two different contexts:
- as a threshold: the maximum allowed loading beyond which an element is considered overloaded,
- as a result: the highest observed loading of an element during N-1 scenarios.
This dual use can easily cause confusion, since the same name refers to both an input parameter and a computed result.