running timeseries with runpp_3ph while also passing the parameter trafo_loading="power" in runpp_3ph

[recklessAxolotl]

Hello,

I've had some trouble with the results of an asymmetric powerflow in a timeseries I run.
I passed in the runpp_3ph function to the timeseries, in order to run the asymmetric load flow instead of the balanced one:
run_timeseries(net, time_steps=range(0, 96), run=pp.runpp_3ph)

But if I do it like this I get infeasable results.

Is there a way to also pass a parameter to this pp.runpp_3ph function? It does not work like this:
run_timeseries(net, time_steps=range(0, 96), run=pp.runpp_3ph(net, trafo_loading="power"))

Or if not, could I get to the results of trafo loading_percent calculated like when trafo_loading="power" is set some other way?

I tried running without timeseries, so just a single time step just with the usual runpp_3ph(net, trafo_loading=power) function. There my results are feasable. But I can't do it like this using loops, because my model depends on pandapower controllers that wouldn't work without a timeseries. The test I made was just a simplified version of the model.

[recklessAxolotl]

(Background info:)
(Trafo "loading_percent" from "res_trafo_3ph" is very high and not really feasable. (I'm comparing trafo loading_percent with the summed consumed power of all loads divided by the trafo's rated apparent power - to check for feasability.)
I've tried running one time step without the timeseries function but just with the usual runpp_3ph function and its parameter "trafo_loading" set to "power" (instead of the default which is "current"). This way I got good results. But my problem is that I have to use the timeseries function, as I am depending on pandapower controllers that work in combination with it.)

(Findings with trafo_loading=current:)
(I get very different results, depending on the trafos vector group (Dyn or YNyn). Actually the trafo's vector group should be Dyn though.
I also tried with unbalanced or balanced load data (P and Q values of the asymmetric loads).
(I created balanced data from my unbalanced data by just averaging P and Q over the phases).
Findings (trafo_loading=current):
Dyn (unbalanced or balanced data) -> infeasable results
YNyn (unbalanced or balanced data) -> feasable results)