Parallel Operation of Inverters - Control of Power Inverters in Renewable Energy and Smart Grid Integration - page 285

A natural problem for parallel-operated inverters is how to share the load among them. A
key technique is to use droop control (Barklund et al. 2008; Brabandere et al. 2007; Guerrero
et al. 2005, 2007, 2011; Majumder et al. 2010; Mohamed and El-Saadany 2008a; Tuladhar
et al. 1997; Zhong and Weiss 2011), which is widely used in conventional power generation
systems (Diaz et al. 2010). The advantage is that no external communication mechanism is
needed among the inverters to achieve good sharing for linear and/or non-linear loads (Borup
et al. 2001; Chandorkar et al. 1993; Coelho et al. 2002; Guerrero et al. 2004, 2006a; Tuladhar
et al. 1997, 2000).
The equal sharing of linear and non-linear loads has been intensively investigated (Borup
et al. 2001; Guerrero et al. 2004, 2005, 2006a, 2007). A voltage bandwidth droop control was
used to share non-linear loads in (Tuladhar et al. 1997) and a small signal injection method
was proposed to improve the reactive power sharing accuracy in (Tuladhar et al. 2000), which
can also be extended to harmonic current sharing. Injecting a harmonic voltage according to
the output harmonic current can be used to improve the total harmonic distortion (THD) of the
voltage (Borup et al. 2001). In (Guerrero et al. 2004, 2005), it was pointed out that the output
impedance of inverters could play an important role in power sharing and a droop controller
for inverters with resistive output impedances was proposed for sharing linear and non-linear
loads (Guerrero et al. 2006a, 2007).
Although significant progress has been made for the equal sharing of linear and non-
linear loads, it was reported that the accuracy of reactive power sharing (for the Q − E

and P droop) is not high (Guerrero et al. 2006b; Lee et al. 2010; Li and Kao 2009).

Moreover, some approaches developed for equal sharing, e.g. the one proposed in (Marwali
et al. 2004), cannot be directly applied to proportional sharing. Another issue is that the output
voltage drops due to the increase of the load and also due to the droop control (Sao and
Lehn 2005).
Adding an integral action to the droop controller is able to improve the accuracy of load
sharing for grid-connected inverters (Dai et al. 2008a; Li et al. 2004; Marwali et al. 2004)
but it does not work for inverters operated in the stand-alone mode and also there is an issue
associated with the change of the operation mode. Adding a virtual inductor and estimating the
effect of the line impedance is able to improve the situation by changing the droop coefficients
(Li and Kao 2009). A Q-V dot droop control method was proposed in (Lee et al. 2010) to
improve the accuracy of reactive power sharing following the idea of changing the droop
coefficients in (Li and Kao 2009) but a mechanism to avoid the output voltage variation was
necessary, which reduces the accuracy of the reactive power sharing, as can be seen from the
results given there. These strategies are sensitive to numerical computational errors, parameter
drifts and component mismatches.
Other trends to solve the power sharing problems for parallel-operated inverters are to make
the output impedance as accurate as possible over a wide range of frequencies (see (Guerrero
et al. 2008; Yao et al. 2011) and the references therein) and to introduce the secondary
control to bring the deviated voltage and frequency back to the rated values (Guerrero et al.
2009, 2011). However, this inevitably needs communication among the inverters (even of low
bandwidth) and the advantage of droop control is lessened. The secondary control also leads
to slow responses and/or instability because of the delay introduced in the measurement and
communication channel. The complexity of the system is also increased, as evidenced by the
number of control loops/levels involved in such systems. This, again, increases the chance of
instability.

− ω