More rigorously test power flow in the loop

[GabrielKS]

As of 2025-01-23, the only real data test performed in #1040 is that for the first and last timestep in a sample simulation, the active powers in the PSI results match the active powers in the PSS/E export for all thermal generators. We should be testing both more granularly and more completely. In particular, it would be great if these things were tested:

• That input_key_map is constructed (_make_pf_input_map!) correctly for a wide variety of potential configurations
• That data is correctly written (update_pf_data!) from the OptimizationContainer to the PowerFlowData (currently unsure what data source to use for reactive power here, but we can at least granularly test active power)
• That data is correctly written back (calculate_aux_variable_value!) from the PowerFlowData to the OptimizationContainer
• That final results are consistent beyond just active power (good starting point)
• What happens when the AC power flow fails to converge