Parallel Operation of Inverters - Control of Power Inverters in Renewable Energy and Smart Grid Integration - page 302

the voltages were different). Inverter 1 was able to pick up the load, gradually in the case of
the robust droop controller and very quickly in the case of the conventional droop controller.
The robust droop controller could share the real power very accurately but the conventional
one could not. The robust droop controller has considerably relaxed the trade-off between the
sharing accuracy and the voltage drop. The voltage from the inverters equipped with the robust
droop controller is very close to the rated voltage but the voltage from the inverters equipped
with the conventional controller is only 3 of the rated voltage. The voltage set-point of the
conventional droop controller has to be lower than the rated voltage due to the droop effect.
The bigger the voltage drop ratio, the lower the voltage set-point. It can also be clearly seen
that E 1 = E 2 because the per-unit output impedance is different and also there are numerical

errors and component mismatches etc. Because of the reduced deviation in the voltage, the
reactive power becomes bigger. This leads to a slightly bigger deviation in the frequency but it
is expected because of the Q

− ω

droop. The current sharing reflects the power sharing well.
It is worth noting that there was no need to change the operation mode of Inverter 2 when
connecting or disconnecting Inverter 1.

19.6.5.2 Inverters having the Same per-unit Output Impedance with a Linear Load

The current feedback gains were chosen as K i 1 = 4 so that the output impedance

is consistent with the power sharing ratio 2 : 1. The results from the robust droop controller
with K e =

2 and K i 2 =

10 are shown in the left column of Figure 19.9 and the results from the conventional
droop controller are shown in the right column of Figure 19.9. The robust droop controller
was able to share the load according to the sharing ratio and considerably outperformed the
conventional droop controller in terms of sharing accuracy and voltage drop. The difference
between the voltage set-points can be clearly seen. This indicates the effect of numerical errors,
parameter drifts and component mismatches etc. because the voltage set-points were supposed
to be identical without these uncertain factors. Comparing the left columns of Figures 19.8
and 19.9, there were no noticeable changes in the performance for the robust droop controller
but the difference in the voltage set-points was decreased. Comparing the right columns of
Figures 19.8 and 19.9, the sharing accuracy and the voltage drop were improved slightly and
the voltage set-points became closer to each other when the per-unit output impedances were
the same.
Another experiment was carried out with K e = 1 to demonstrate the role of K e and the

results are shown in the right column of Figure 19.10. The results with K e = 10 shown in the

left column of Figure 19.9 are shown in the left column of Figure 19.10 again for comparison. It
can be seen that a large K e helps speed up the response and reduce the voltage drop. However,
a large K e causes large ripples in the current.

19.6.5.3 With a Non-linear Load

A non-linear load, consisting of a rectifier loaded with an LC filter and the same rheostat
used in the previous experiments, was connected to Inverter 2 initially. A similar procedure
to connect/disconnect Inverter 1 was followed in the experiment. The relevant curves of the
experiment are shown in the left column of Figure 19.11 for the case with the robust droop
controller and in the right column of Figure 19.11 for the case with the conventional