3 phase transformer losses division by 3

[elodin]

As based on #1073, the overall trafo losses have to be divided by 3 phases for 3 phase calculations.
This request does that only for 3 phase calculations (using runpp_3ph) and shows more consistent results for the bus P_mw values for a variety of tested grids. The comparison is made to the results of the same (but balanced) grid, calculated using runpp.
I.e. for a unbalanced version of the SimBench LV-rural2 grid, it shows the following results on buses:

Before change - Max. Difference of p_mw between 1 and 3 phase:
0.0019667287444448955
In percent, this means:
3.3230423588091575

After change - Max. Difference of p_mw between 1 and 3 phase:
0.00015141917685417006
In percent, this means:
0.2558422660184597

[codecov]

Codecov Report

Merging #1201 (9a04719) into develop (9f893e3) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff            @@
##           develop    #1201   +/-   ##
========================================
  Coverage    87.61%   87.61%           
========================================
  Files          162      162           
  Lines        15773    15773           
========================================
  Hits         13819    13819           
  Misses        1954     1954           

Impacted Files	Coverage Δ
pandapower/build_branch.py	97.40% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9f893e3...9a04719. Read the comment docs.