I am looking at the L2 value of -80 Watts with a pf of 0.17. That doesn’t look right, but now as I look closely at your diagram, I see those CTs are on the SolarEdge (SE) cables. I don’t know much about that system with battery, so I’ll reserve judgement until you put CTs on the main line somewhere as suggested and I see the results.
I’m assuming theSE CTs are on the main incoming cables upstream from where the inverter connects, and that it monitors grid power usage so that it will divert PV power to the battery rather than export power to the grid. If I have that wrong, please advise.
You want to understand what the SE is doing. To determine how well it is managing energy using the IoTaWatt, I would suggest you reconfigure to the first option, at least initially, so you can see and plot exactly what is happening.
Just saying, you can’t get consistently accurate results with two different phases in one CT, regardless of the voltage/phase that you use. Prese_PTP1 and Luci_PTP1 cannot be right with a reference of phase B-C. With derived reference and a phase-to-neutral load, the phase can only be A, B or C. While you may have found complementary errors with the phase shift and power factor of your load that approximates what you feel is correct, it will not yield consistently correct results.