[IMPROVEMENT] Round-trip PF -> SE to test SE convergence

[mgovers]

Problem faced

We currently generate some state estimation validation cases based on power flow validation cases. However, this is not always exhaustive and there may, e.g., be some edge cases/topologies/... tested in PF but not in SE.

In addition, state estimation should often be able to converge within one iteration if the input is "perfect", but it is not tested whether that is actually the case (note that the flat start may still prevent such "perfect" convergence).

Improvement proposal

Since PF output can also be used as input for SE, we can actually solve both problems at once by doing a round-trip:

• We take the PF validation cases
• We run a PF
• We use the PF results for input for the SE (each appliance and branch shall be enriched with a power sensor)
• we can run this in Nightly, no need to run this for every commit.

Community Meeting

• I wish to discuss this improvement in the bi-monthly PGM community meeting. Please join the PGM community meeting.