2010 International Conference on Power System Technology

Voltage and Frequency Control of Inverters

Connected in Parallel Forming a Micro-Grid
Chunsheng Wu, Hua Liao, Zilong Yang, Yibo Wang, Honghua Xu

Abstractthe communication mechanism to realize the

power share of grid-forming inverters is analyzed and a droop
control method through active and reactive current is proposed
in this paper. The control strategies of inverters connected in
parallel forming a micro-grid are described in detail. The droop
controller is divided into three units that are current (power)
decoupling, the first-order inertia and droop control. From the
point of synchronous generator, the design principles of firstorder inertia constant and the droop coefficients are analyzed.
Finally, simulation models are established in Simulink/Matlab
and two prototypes with the capacity of 10kVA are established in
laboratory. The simulation and experiment results verify the
feasibility of the proposed method.
Index Terms-- Droop Control, Inverter, Micro-grid, Parallel,
Voltage and Frequency Control.

regulation. Namely, frequency has a strong coupling feature

against active power, while voltage has a strong coupling
feature against reactive power. However, compared to the
parallel operation of synchronous generators, there is essential
deference for parallel operation of inverters. Analyzing
operation characteristics of droop control inverters, voltage
and frequency control strategies of inverters connected in
parallel forming a micro-grid is proposed from the view of
modular design, which makes inverters has the prosperity of
synchronous generators.
In this paper, control strategies of droop controller are put
forward. Analysis of droop coefficients influence on stability
was done. The design method of first-order inertia link is
given in detail. Finally, simulation models are established in
Simulink/Matlab and two 10kVA prototypes are established in
laboratory to verify the feasibility.

I. INTRODUCTION

T is not only an important technical means to increase the

reliability and the capacity of the power supply for parallel
technology of inverters, but also a key technology to
develop from a traditional centralized mode to a distributed
mode for AC power supply system. Because the renewable
energy sources such as solar power generation are influenced
by local conditions and geographical area limitations, a
relative large number of inverters with unequal capacity are
connected in parallel to form a micro-grid, which can reduce
the cost of design and production, improve the reliability and
flexibility [1].
Parallel operation of inverters technology is first used in the
parallel control of UPS, but the mature technologies adopted
in UPS are still centralized control technologies with
communication lines, which not meet the need of micro-grid
application. At present, droop control is widely adopted for
parallel inverters forming a micro-grid [2-7].
Droop control theory is based on the external characteristic
of synchronous generator under the condition of power

II.

DROOP CONTROL THEORY

A. Transmission Line Power Transfer Theory[8-10]

Transmission of active power and reactive power in
transmission line depends on the voltage amplitude and phase
angle at sending side and receiving side respectively, which is
the theoretical basis of synchronous generators operating in
parallel. A simple model of power transmission in high
voltage transmission line is shown in figure 1.
us

us

jX

I&s S&s

ur 0

S&r = Pr + jQr

S&s = Ps + jQs
X : line inductance
: phase angle
us : voltage at sending side
ur : voltage at receiving side

I&s

jX

ur

: power angle
I& : current at sending side
S&s : apparent power at sending side
S&r : apparent power at receiving side

Fig1 Simple model of power transmission

This work was supported in part by Important direction of Chinese

Academy of Sciences Knowledge Innovation Program.
Chunsheng Wu is with the Department of Renewable Energy Generation,
IEE CAS, Peking, China. (e-mail: [email protected]).
Hua Liao is with the Department of Renewable Energy Generation, IEE
CAS, Peking, China. (e-mail: [email protected])
Zilong Yang is with the Department of Renewable Energy Generation,
IEE CAS, Peking, China. (e-mail: [email protected])
Yibo Wang is with the Department of Renewable Energy Generation, IEE
CAS, Peking, China. (e-mail: [email protected])
Honghua Xu is with the Department of Renewable Energy Generation,
IEE CAS, Peking, China. (e-mail: [email protected])

978-1-4244-5939-1/10/$26.002010 IEEE

As shown in figure 1, active power and reactive power

flowing into the line at sending side can be described with
following equations:
*

u u e j
S&s = Ps + jQs =u&s I&* =u&s s r

jX

uu
Ps = s r sin *
X

(1)

( 2)

Qs =

us (us ur cos )
X

(3)

u s ur

X
u (u u )
Qs = s s r
X

( 4)

In general transmission power level, the power angle is very

small, so it can be assumed that sin = and cos = 1 .
In this case, equations (1) and (2) are simplified as
followings:

Ps =

( 5)

From the equations (4) and (5), we can draw two

conclusions:
(1) The active power has a strong coupling to the power
angle (corresponding to the frequency of power grid).
(2) The reactive power has a strong coupling to the voltage
at sending side.
According to the two conclusions above, the frequency and
voltage droop regulation through respectively active and
reactive power:

f f 0 = k p ( P P0 )

u u0 = kq (Q Q0 )

( 6)
(7)

f and u are grid frequency and voltage, f 0 and u0 are

nominal frequency and nominal grid voltage respectively,
P and Q are the output of active power and reactive power,

P0 and Q0 are the momentary set points for active power and
reactive power of the inverter, k p and kq are the frequency
coefficient and voltage coefficient respectively.
B. Voltage and Frequency Control[11]
Equations (6) and (7) are the basis of voltage and
frequency control of inverters connected in parallel forming a
micro-grid.
The power share of two inverters based on droop control is
described in figure 2. Assuming frequency droop coefficients
of the two inverters as k p1 and k p 2 , the output active power of

Fig. 2. Power share of two inverters connected in parallel

Comparing with equations (8) and (9),

P1 k p 2
=
P2 k p1

(10)

Promoting equation (10) to multiple inverters, the

communication mechanism of active power share based
droop control is:

P1 k p1 =P2 k p 2 = ... = Pn k pn

(11)

Similarly, the communication mechanism of reactive

power share based droop control is:

Q1 kq1 =Q2 kq 2 = ... = Qn kqn

(12)

In equations (4) and (5), the voltage at sending side us can

be regarded as a constant, thus:

Ps ur
=
us X
Q u u
I re = s = s r
us
X

Ia =

(13)
(14)

the two inverters as P10 and P20 at the nominal frequency of

I a and I re are active current and reactive current

the grid f 0 , when the loads increasing, two inverters will

respectively, so the frequency and voltage droop control

characteristics then become:

change its output respectively and make the grid frequency

operation at f1 .
Thus we can get:

1
f1 f 0
k p1

(8)

1
f1 f 0
k p2

(9)

P1 = P11 P10 =
P2 = P21 P20 =

f f 0 = ka ( I a I a 0 )

u u0 = kre ( I re I re 0 )

(15)

(16)

I a 0 and I re 0 are the momentary set points for active

current and reactive current of the inverter, ka and kre are the
frequency coefficient and voltage coefficient respectively.
This approach has some benefits:
(1) When the voltage u1 change significantly from nominal
value u0 (e.g. A short-circuit occurs at u1 ), the result of
infinite currents will be avoided;

(2) The computing process of digital controller is simplified

and improves the algorithm accuracy.

as example. Assuming
can be written:

III. CONTROL STRATEGY

A. Control Scheme
The aim of the control strategy for three-phase voltage
source is to control frequency and voltage as constant value.
So in this way, the inverters can not operation in parallel,
otherwise the small differences between voltage and phase
will cause large system circulation. But the aim of the droop

control is to obtain the frequency reference f

and voltage

reference u . Adjusting frequency and voltage of grid

dynamically to realize power share of the grid-forming
inverters, the voltage and frequency stability of the grid are
maintained Figure 3 shows the control scheme of three-phase
forming-grid inverter in dq rotation coordinate system.
Vdc

+
-

u d

u q

ia
ib
ic

uabCf
ubcCf
ucaCf

ud
uq

Take the frequency versus active current droop regulation

uabC
ubcC
ucaC
Z

uabC C
ubcC
ucaC

f f0 =

M=

ka ,

D= 1

1
(id id 0 )
Ms + D

ka , equation (18)
(20)

Equation (20) can be described in figure 4.

I a

id 0

1
Ms + D

f = f f 0

id
Fig. 4. Control diagram of first-order inertia unit

From the figure 4, we can find that the droop control of

frequency versus active current is similar to that of
synchronous generator in absence of governor after
introducing first-order inertia unit. M stands for the first-order
inertia constant of the generator system, and D stands for the
damping coefficient of the generator system respectively,
which together determine the respond speed to loads changing.
Taking M = 1s, D = 2, the bode diagram of first-order
inertia in the frequency domain is shown in figure 5.

idf
iqf
ud

id
iq

iaf
ibf
icf

ikf
iq

1
cs +1

id f

ia
ib
ic

id
iqf

Fig. 3. Control scheme of three-phase forming-grid inverter

As shown in figure 3, different regular three-phase voltage

source inverter, droop controller is introduced, which can be
divided into three units that are current (power) decoupling,
the first-order inertia and droop control.
In a three-phase symmetrical system, the coordinate
transformation from three-phase stationary coordinate system
to rotation coordinate system for currents can be realized:

iaf
id

i = T3s 2 r ibf
q
icf

(17)

B. Implementation of First Order Inertia

After introducing the first-order inertia, equations (15) and
(16) can be written:

ka
(i i )
cs +1 d d0
k
u u0 = re (iq iq 0 )
cs +1
f f0 =

(18)
(19)

Fig. 4. Bode diagram of the first-order inertia unit

For the amplitude-frequency characteristics shown in figure 4,

the first-order inertia unit has a good performance of low filter

= 1/

c , otherwise
(LF) in less than turn-point frequency c
attenuate at -20dB more than turn-point frequency.
Equation (20) describes the first-order inertia unit in
continuous domain. But in digital control system, it must be
discredited. Through bilinear transform of equation (20), the
expression can beget in the z domain:

f ( z) f0 =

z +1
(id ( z ) id 0 ) (21)
(2 Mf s + D) z + ( D 2 Mf s )

f s is the sampling frequency, f ( z ) is output frequency and

id ( z ) is the output current of discredited system respectively.
Taking f s = 600 Hz and f s = 6kHz , the bode diagram of
the first-order inertia in z domain is shown in figure5.

The bode diagram in z domain shows that the first-order

inertia unit still has a good performance of low-pass if the
sampling frequency is relative high compare to the turin-point
frequency of the control system.

kre =

(23)

I re max

When the reactive current I re is greater than zero, the gridforming inverter will work at capacitive region, otherwise
active current I re is less than zero, and it will work at
inductive region. But no matter what kind of modes the gridforming inverters work, the reactive power share mechanism
of every grid-forming inverter is following as equation (12).
IV. SIMULATION AND EXPERIMENT RESULTS
A. System Description
In order to verify the feasibility of the control scheme
proposed above, a system formed by two inverters with the
same parameters is designed. The system parameters
description is shown in table1.
TABLE 1. SYSTEM PARAMETERS DESCRIPTION

Parameters Symbol

Value

Physical Meanings

Fig. 4. Bode diagram of the first-order inertia unit

Sinv1=Sinv2

10kVA

Apparent power

C. Implementation of Droop Control

As analyzed above, the droop coefficient of the gridforming inverter is corresponding to the load damping
constant of synchronous generator, which reveals the
percentage of load changing under the 1% changing of the
frequency or voltage.
Only thorough understanding of the droop coefficients
physical meanings, optimal value of droop coefficients can be
designed. Figure 4 shows the droop control characteristics.

kp1=kp2

0.033(Hz/A)*

Frequency droop coefficient

kq1=kq2

0.735(V/A)*

Voltage droop coefficient

f01= f02

50(Hz)

Nominal frequency

u01= u02

220V

Nominal voltage

Ia01= Ia02

Active current at nominal

frequency point

Ire01= Ire02

reactive current at nominal

voltage point

Lf1= Lf2

1.5mH

Filter inductance

Cf1= Cf2

22uF

Filter capacitance

r/X

0.1

Line impedance ration

f ( Hz )

u (V )

f max

umax

f0
Rectifier
Region

I a min

f
f
f min

u0
Inverter
Region

I a ( A)

I a max

(a) active current versus frequency

u
u
umin

Inductive
Region

I re min

Capacitive
Region

I re ( A)

I re max

(a) reactive current versus voltage

Fig. 4. Diagram of droop control characteristics

Figure 4 (a) describes the characteristics of active current

versus frequency droop regulation. There is a physical limit,
determined by the ratings of the inverter, that is I a max . The
frequency of the grid is also limited to the range ( f min , f max ) .
So the droop coefficient of active current versus frequency is
defined as:

ka =

I a max

(22)

Note*: The frequency droop coefficient is designed at the

maximum percentage change of 1% and voltage droop
coefficient is designed at the maximum percentage change of
5%.
B. Simulation Results
In matlab/simulink simulation platform, the system model is
built.
Simulation condition 1: Take the first-order inertia constant
c =0.1s; the total loads demand of active current and reactive
current is 20A and 15.6A respectively at t=0s, and increasing
to 39A and 31.2A respectively at t=5s; .

When the active current I a is greater than zero, the gridforming inverter will work at inverter mode, otherwise active
current I a is less than zero, and it will work at rectifier mode.
But no matter what kind of modes the grid-forming inverters
work, the active power share mechanism of every gridforming inverter is following as equation (11).
Similarly the droop coefficient of reactive current versus
voltage is defined as:

Fig. 5. Output active current of two grid-forming inverter

The output active current of the two grid-forming inverter is

shown in figure 5. Before t=5s, id 1 = id 2 =10A, at t=5s,the
loads demand increasing, id 1 = id 2 =20A
The output active current of the two grid-forming inverter is
shown in figure 6. Before t=5s, iq1 = iq 2 =7.8A, at t=5s,the
loads demand increasing,

iq1 = iq 2 =15.6A.

Fig. 10. Voltage of the grid

=0.8s

The frequency and the voltage of the grid are shown in

figure 9 and 10 respectively.
Compare to figure 7 and 8, we can draw some conclusions:
(1) The more first-order inertia constant, the more stable for
the frequency and the voltage of the grid;
(2) The more first-order inertia constant, the lower
responding speeds when loads changing.

Fig. 6. Output reactive current of two grid-forming inverter

The frequency and the voltage of the grid are shown in figure
7 and 8 respectively. Before t=5s, the frequency and the
voltage of the grid is 49.947Hz and 216V, at t=5s,the loads
demand increasing, the frequency and the voltage of the grid
is decrease to 49.897Hz and 212V.

C. Experiment Results
Two prototypes with the capacity of 10kVA are established
in laboratory, which have the same parameters as those of the
simulation model. Take c =0.4s, sampling frequency

f s =600Hz, loads R=30. The output current of the A-phase

working at steady-state is shown in figure 11.

Fig. 7. Frequency of the grid

Fig. 11. The output current of the A phase working at steady-state

Fig. 8. Voltage of the grid

From the results of the simulation, we can find that droop

control is a kind of difference adjusting, power share of every
grid-forming inverter is at the expense of the deviation of the
grid frequency and the voltage, so it is necessary to limit the
range of the frequency and the voltage in the grid to ensure
safety.
Simulation condition 2: the same conditions as the
simulation 1, change the first-order inertia constant as
c =0.8s.

Fig. 9. Frequency of the grid with

In figure 11, channel 1 is output current waveform of gridforming inverter 1, channel 2 is that of grid-forming inverter 2,
and channel 3 is that of loads. From the results, the current
value of the loads is 7.3A, and that of grid-forming inverter1
and grid-forming inverter2 is 3.7A and 3.6A respectively.
Changing the loads demand from R=30 to R=18, and
from R=18 to R=30, the output current waveforms of the
grid-forming inverters and the loads is shown in figure 12 and
figure 13.

=0.8s
Fig. 12. The output current with increasing loads

[10] Engler A, Soultanis N. "Droop control in Lv-Grids", International

Conference on Future Power Systems, 2005.
[11] Piagi P. "Microgrid Control", Ph.D Thesis, University of Wisconsinmadison, 2005

VII. BIOGRAPHIES

Fig. 13 The output current with decreasing loads

V. CONCLUSIONS
After analysis and experimental verification, conclusions
were obtained.
(1) Inverters do not have the same droop characteristics as
the synchronous generators, and has rapid voltage and
frequency response with the loads change. After introducing
the first-order inertia, voltage and frequency regulation
characteristics of parallel inverters forming a micro-grid are
similar to those of synchronous generators.
(2) Appropriate droop coefficients can guarantee every
inverter to share load respectively and maintain the voltage
and frequency of micro-grid stability.
(3) The more first-order inertia time constant can improve
the voltage and frequency of micro-grid stability, but decrease
the dynamic response speed. So it is very important to set a
right first-order inertia time constant for micro-grid stability.
VI. REFERENCES
[1]

[2]

[3]

[4]
[5]

[6]

[7]

[8]
[9]

Guerrero J M, Berbel N, Matas J, et al. "Decentralized control for

parallel operation of distributed generation inverters in microgrids using
resistive output impedance", Proceedings of the 32nd Annual
Conference of the IEEE-Industrial-Electronics-Society, Paris, FRANCE,
F Nov 07-10, 2006 .
De Brabandere K, Bolsens B, Van Den Keybus J, Et Al. "A voltage and
frequency droop control method for parallel inverters", Proceedings of
the 35th Annual IEEE Power Electronics Specialists Conference (PESC
04), Aachen, GERMANY, F Jun 20-25, 2004 .
Guerrero J M, De Vicuna L G, Matas J, Et Al. "A wireless controller to
enhance dynamic performance of parallel inverters in distributed
generation systems", Proceedings of the 34th Annual IEEE Power
Electronics Specialists Conference, Acapulco, Mexico, F Jun 15-19,
2003.
Brabandere K D. "Voltage and frequency droop control in low voltage
grids by distributed generators with inverter front-End ", Ph.D.
dissertation, Katholieke University, 2006.
Ju H X, Ding M, Su J H, et al. "Communicationless parallel inverters
based on inductor current feedback control", Proceedings of the 22nd
Annual IEEE Applied Power Electronics Conference (APEC 2007),
Anaheim, Feb 25-Mar 01, 2007
Kan J R, Xie S J, Ieee. "Research on the power sharing of the parallel
inverters without control interconnection basing on droop characteristic",
Proceedings of the CES/IEEE 5th International Power Electronics and
Motion Control Conference, Shanghai, Peoples R China, Aug 14-16,
2006 .
[55] Loix T, De Brabandere K, Driesen J, Et Al. "A three-phase
voltage and frequency droop control scheme for parallel inverters ",
Proceedings of the 33rd Annual Conference of the IEEE-IndustrialElectronics-Society, Taipei, TAIWAN, F Nov 05-08, 2007
TAYLOR, Translated by Weisheng Wang, "Stability analysis of power
system", Chinese power press, 2002.
Guerrero J M, Berbel N, Matas J, et al. "Control of line-interactive UPS
connected in parallel forming a microgrid", Proceedings of the IEEE
International Symposium on Industrial Electronics, Vigo, SPAIN, F Jun
04-07, 2007

Chunsheng Wu was born in Anhui Province,

China, in 1980. He received the Msc.E.E. degree
from Inner Mongolia University of Technology
(IMUT) in Inner Mongolia, and the Ph.D. degree
from the Institute of electrical engineering, Chinese
Academy of Science, Beijing, China, in 2003 and
2009, respectively.
His employment experience included Electronic
Engineer in Guangdong Kelon Electrical and
Mechanical Co., Ltd, from 2003 and 2005.
Since 2009, he has been an Assistant Professor in the Renewable Energy
Generation Research and Development Center, Institute of electrical
engineering, Chinese Academy of Science. His research interests are microgrid control, converters control, PV generating system, solar tracker control
etc.
Hua Liao was born in Hubei Province, China, in
1976. He received the Msc.E.E. degree and Ph.D.
degree Institute of electrical engineering, Chinese
Academy of Science, Beijing, China, in 2004 and
2008, respectively.
His employment experience included Electrical
Engineer in Shougang Group in 2001.
Since 2008, he has been an Assistant Professor
in the Renewable Energy Generation Research and
Development Center, Institute of electrical
engineering, Chinese Academy of Science. His research interests are control
technology of DC/DC converter, energy storage system.
Zilong Yang was born in Hebei Province, China,
in 1980. He received the Msc.E.E. degree from
Yanshan University in Hebei Province, and the Ph.D.
degree from the Institute of electrical engineering,
Chinese Academy of Science, Beijing, China, in
2003 and 2010, respectively.
He is presently an Assistant Professor in the
Renewable Energy Generation Research and
Development Center, Institute of electrical
engineering, Chinese Academy of Science. His
research interests are in the areas of renewable energy generating techniques,
energy management system of PV microgrid.
.
Yibo Wang was born inXinjiang Province,
China, in 1977. He received Ph.D from the Institute
of electrical engineering, Chinese Academy of
Science, Beijing, China, in 2009, respectively.
In 1998, he joined School of Electrical
Engineering Xinjiang University. Since 2009, he has
been an Associate Professor in the Renewable
Energy Generation Research and Development
Center, Institute of electrical engineering, Chinese
Academy of Science. His research interests are PV
generating system, power system control.
Xu Hong-hua was born in 1967. He is a
professor in Institute of Electrical Engineering, CAS,
China. He is the chair of renewable energy
generating group in IEE, CAS. His special interest
includes wind and PV power planning and
renewable energy generating techniques.