Performance improvement of parallel active power filters using droop control method

Performance Improvement of Parallel Active Power
Filters Using Droop Control Method
Ghazal Falahi
School of Electrical engineering
Sharif University of technology
Tehran, Iran 11365-9363
Email: Falahi@ee.sharif.edu
Hossein Mokhtari , Member IEEE
School of Electrical engineering
Sharif University of technology
Tehran, Iran 11365-9363
Email: Mokhtari@sharif.edu
Abstract-In this paper, a new method based on droop control
scheme is proposed for controlling parallel operation of active
filters. The harmonic components of the load current are
extracted by an enhanced phase-locked loop (EPLL). In the
parallel group, each filter operates as a conductance and the
harmonic workload is shared among them. A droop relationship
between the conductance and non-fundamental apparent power
controls the operation of each unit. The non-fundamental
apparent power has been calculated based on IEEE Std 1459.
Principles of operation are explained in this paper and simulation
results which are presented approve the effectiveness of this
method. The results indicate a significant reduction in Total
Harmonic Distortion (THD) in a rectifier application.
Keywords-Power quality; Parallel active filters; Power system
harmonics; Droop
. INTRODUCTION
The increasing use of power electronic devices has resulted
in harmonic pollution in power networks. The main problem
stem from the flow of non-active energy caused by harmonic
currents and voltages. Flowing non-sinusoidal current into the
network has the drawback of deteriorating the harmonic
pollution and degrading power quality [1].
Mitigation equipments such as Active Power Filters (APFs)
are designed and used to improve power quality. APFs are one
of the emerging solutions to surpass power system harmonics
and enhance power quality. This is mainly due to
technological progress in the power switching devices, Digital
Signal Processors (DSPs) and new control algorithms [2].
In many cases, it is favorable to connect APFs in parallel
instead of using filters with increased capacities. This results
in higher reliability and minimizes the installed converters
rating and cost. One of the most effective and reliable methods
for controlling the operation of parallel inverters is droop
method. The droop method is usually used to achieve good
sharing among units when communication between the
inverters is difficult due to their physical location or when a
more reliable and flexible system is required. This method
avoids any control wire interconnection among different
modules and that’s why it is often named as wireless or
independent parallel control [3]-[5].
This paper proposes a control method for parallel operation of
Active Filter Units (AFUs) based on droop strategy. In this
case, each filter operates as a conductance and several units
can share the non-fundamental load power without any
interconnection. The droop coefficient of each unit is
determined by its capacity and the harmonic filtering
workload is shared among AFUs in proportion to their
capacity. With increasing or decreasing of nonlinear loads in
the system the droop control method will help AFUs to
dynamically adjust their non-fundamental filtering capacity to
acquire satisfactory compensation. The non-fundamental
power is calculated based on IEEE Std 1459. Also a new
approach for harmonic extraction based on EPLL has been
introduced. Fig. 1 shows the power circuit of parallel
connected AFUS and Fig .2 shows the proposed load sharing
and control method.
As compared to existing harmonic extracting methods,
EPLL-based method provides higher degree of immunity and
insensitivity to noise, harmonic and other types of pollutions.
Simple Structure of EPLL-based method simplifies its
implementation in digital software and/or hardware
environments as an integral part of digital control platform for
power electronic converters.
Fig. 1. Power circuit of parallel connected AFUs
978-1-4244-2487-0/09/$25.00 ©2009 IEEE

Fig. 2 Load sharing and control method Fig. 3. EPLL structure
The dominant feature of the proposed method over
conventional methods is the frequency adaptivity which
permits desired operation when the center frequency of the
base signal varies. This system is also capable of coping with
the unbalanced system conditions [6].
II. Operation principals of parallel APF
This section explains the principals of operation of the
proposed parallel APF system.
A. Harmonic detection method
An estimation of fundamental component is obtained by
means of an adaptive nonlinear notch filter, i.e. EPLL. The
overall structure of the EPLL is in accordance with a
conventional PLL. The basic structure has three independent
internal parameters K, KPKv and KiKv. Parameter K
dominantly controls the speed of convergence of amplitude
and parameters KPKv and KiKv control the rate of convergence
of phase and frequency respectively. As compared with the
conventional PLL, the EPLL method generates a more
accurate angle and estimation of fundamental component in a
polluted environment. An implementation of the EPLL is
shown in Fig. 3. The input signal is compared with its
extracted smooth version to generate an error signal which is
used by low-pass filter (LF) to generate a driving signal for
VCO.
The EPLL is actually a band-pass filter and the continuous
time differential equations governing it’s dynamic are derived
from the block diagram of Fig. 3 as [6]:
(1)
(2)
Where y(t) is the fundamental component and e(t) is the error
signal. Fig. 4 compares the phase angle extracted by an EPLL
and a conventional PLL. Also an estimation of the
fundamental component of input signal is presented.
B. Current controller
After detecting the harmonic component of phase current, a
current regulator is used for accurate tracking of the reference
current by the AFUs and the voltage commands are calculated
as follows:
(3)
Where the Lx is the output inductor of the AFUx and T is
the sampling period. The voltage commands are used as the
reference voltages for a Pulse Width Modulator (PWM) and
the gating signals are generated to provide an effective
tracking of current commands [8].
EPLL(Rad) Ia1(A) PLL(Rad)
Fig. 4. Phase angle extracted by an EPLL and conventional PLL(Rad/sec) and
an estimation of fundamental component

C. Control algorithm for parallel operation
A proper method is needed to contro
nonlinear loads among AFUs. The conventi
parallel power converters requires interco
converters to achieve balanced load sharin
conventional methods uses a voltage contro
and several “slave” units. However a confi
master/slave strategy is not redundant due
master unit. To achieve true redundancy, a
able to operate independently [9].
To achieve true redundancy, each AFU can
that it behaves like a harmonic conductan
output current of each AFU is related to t
node to which the AFU is installed, i.e.:
IAFx=Gx.EAFx
Where the IAFx is the output current of A
harmonic conductance and EAFx is the phase v
The proposed control method is a load-
that will share the non-fundamental appar
AFUs. Droop control method has been ex
uninterruptible power supply (UPS) syst
various units to share loads without any
Therefore the reliability of the system is enh
operation of AFUs, the droop control met
This control technique can be defined as
between the conductance and the non-fund
power. Therefore, the harmonic workload
among the AFUs. For the proposed pow
equations can be given as:
G1=G0+d1(SN1-SN10)
G2=G0+d2(SN2-SN20)
:
Gx=G0+dx(SNx-SNx0)
In the above equations, Gx is the condu
G0 is the rated conductance, di is the slope of
is the non-fundamental power of AFUx an
non-fundamental apparent power. The droo
shown in Fig. 5. The base of compensation
1459 and the Non-fundamental apparent po
according to:
Where S1x is fundamental apparent power o
the apparent power of AFUx. Vabcx, Iabcx are
and currents of the AFUx and Vabcx1, Iabcx1 ar
phase voltages and currents of AFUx respect
ol the sharing of
ional approach for
nnection between
ng. One of these
oller as a “master”
guration based on
to dependency of
all units should be
n be operated such
nce. Therefore the
the voltage of the
(4)
AFUx , Gx is the
voltages of AFUx.
-sharing technique
ent power among
xtensively used in
tems and allows
y communication.
hanced. In parallel
thod can be used.
s the relationship
damental apparent
d can be shared
wer system, droop
)
) (5)
uctance command,
f the equation, SNx
nd SN0 is the rated
op characteristic is
is IEEE Standard
ower is calculated
(6)
of AFUx and Sx is
the phase voltages
re the fundamental
tively.
Fig. 5. Droop
III. Simula
Simulations have been
environment to investigate the
in different conditions. The par
in Table 1. Since the nonlinear
and 7th
harmonics are dominan
APFs are installed in parallel
line. The THD of the load cu
simulation results. The THD o
the operation of APFs, is appro
and its fundamental componen
compensated source current i
inverter reference and output c
in Fig. 8 (a) and (b), it is sho
condition the filters currents a
coefficient. The simulation re
APFs can effectively suppre
unbalanced and distorted volta
10 show the three phase syste
unbalanced conditions and
respectively and Fig. 11 show
instantly in response to the load
The simulation results verify
leads to a better harmonic extra
improved compensation is ac
compensated source curren
compensation signal generatio
EPLL harmonic extraction me
traditional synchronous refe
unbalanced and distorted condi
Table1. Simula
Source voltage 220v (line-
Transmission line
parameters
R=0.05, L=
Active filter 2 AFUs, L=
G0=0 , d1=8
PWM A sine/trian
Nonlinear load A diode rec
Fig. 6 Load current and it’s fundame
10m
p characteristic
ation results
carried out in the PSCAD
proposed droop control method
rameters of simulation are given
r load is a 6-pulse bridge, the 5th
nt harmonics in the load current.
on the same position along the
urrent is 27% as shown in the
of the source current, thanks to
oximately 3%. The load current
nt are shown in Fig. 6 and the
is shown in Fig. 7. Also, the
current of two AFUs are shown
own that under different droop
are proportional to their droop
esults indicate that the parallel
ss the harmonics even under
age conditions. Fig. 9 and Fig.
em voltages under distorted and
increase of nonlinear load
ws that the source current rises
d increase.
the fact that the EPLL method
action, and as a consequence, an
chieved. Fig. 12 compares the
nt THD for two different
on methods. It is clear that the
ethod is more accurate than the
erence frame method under
itions.
ation parameters
-line), 60Hz
=4mH
=6mH, SN10= 800VA, SN20=800VA,
8x10-4
, d2=4x10-4
ngle PWM, fpwm=10 KHZ
ctifier (6-pulse bridge), RL load.
ental component y-axis:2A/div, x-axis:
ms/div

Fig. 7 System current after compensation y-axis:2.5A/div, x-axis: 10ms/div
(a)
(b)
Fig. 8 (a) Reference and output current of AFU1 , y-axis:0.5A/div, x-axis:
5ms/div (b) Reference and output current of AFU2 , y-axis:0.5A/div, x-axis:
5ms/div
Fig. 9 system phase voltages y-axis:100v/div, x-axis: 10ms/div
Fig. 10 Load current and it’s fundamental component in response to load
increase y-axis:2.5A/div, x-axis: 20ms/div
VI. Conclusion
In this paper, a new system is proposed for better operation
of paralleled APFs based on droop method and an EPLL. A
droop relationship between conductance command and non-
fundamental apparent power (G-SN) controls the sharing of
nonlinear workload among various AFUs. This definition is
based on IEEE1459 Std. A nonlinear load, i.e. a 6-pulse
bridge, is considered to verify the performance of the
proposed technique. Computer simulation shows the
effectiveness of the proposed control technique for harmonic
suppression. The droop characteristic adjusts the filters
Fig. 11 Transition of system current in response to load increase y-axis:
5A/div, x-axis: 20ms/div
(a)
(b)
Fig. 12. (a) THD in abc/dq detection method y-axis:1percent/div, x-axis:
50ms/div (b) THD in EPLL detection method y-axis:1percent/div, x-axis:
50ms/div
capacity based on non-fundamental power while keeping the
source THD within desired limits.
References
[1] IEEE Trial-Use Standard Definitions for the Measurement of Electric
power Quantities Under Sinusiodal , Non-Sinusiodal, Balanced or
Unbalanced Conditions, IEEE Std 1459.
[2] LIU Jinjun, WANG Xiaoyu, YUAn Chang ,WANG Zhaoan, “On the
Control of Active Power Filters, IEEE,The 7th
International Conference
on Power Electronics,Oct. 2007.
[3] J.M.Guerrero, L.Garcia de Vicuna, J.Matas, J.Miret, and M.Castilla, “A
Wireless Load Sharing Controller to Improve Dynamic Performance of
Parallel-connected UPS Inverters,” in Proc. IEEE-PESC’03 Conf. ,
pp.1408-1413.
[4] S.J.Chiang, W.J.Ai, ”Parallel Operation of Three Phase Four Wire
Active Power Filters Without Control Interconnection,” IEEE PESC
2002 Vol. 3, pp. 1202-1207.
[5] Josep M. Guerrero, Nestro Berbel, Luis Garcia de Vicuna, Jose Matas,
“Droop Control Method for the Parallel Operation of Online
Uninterruptable Power Systems Using Resistive Output
Impedance,”IEEE APEC 2006, pp.1716-1722
[6] .M.Karimi-Ghartemani, M.R.Irvani, “A Method for Synchronization of
Power Electronic Converters in Polluted and Variable-Frequency
Environments,” IEEE Trans. Power Syst., vol. 19, pp.1263-1270,
Aug.2004.
[7] Lusian Asiminoaei, Lascu Cristian, Frede Blaabjerg, “Harmonic
Mitigation Improvement With a New Parallel Topology For Shunt
Active Power Filters,”IEEE PESC 2006,pp.1-7.
[8] Po-Tai Cheng, Tzung-Lin Lee ,”Distributed active filter systems
(DAFs):A new Approach to Power System Harmonics,” IEEE
Transaction on Industry Application, Vol. 42, No. 5, pp1301-1309
Sept./Oct. 2006,.
[9] Josep M. Guerrero, Luis Garcia de Vicuna,”A Wireless Controller to
Enhance Dynamic Performance of Parallel Inverters in Distributed
Generation System,”IEEE Transaction on Power Electronics , Vol.19,
No.5, pp. 1205-1213, September 2004.

Performance improvement of parallel active power filters using droop control method

More Related Content

What's hot (19)

Viewers also liked (19)

Similar to Performance improvement of parallel active power filters using droop control method (20)

Recently uploaded (20)

Performance improvement of parallel active power filters using droop control method