Volume 3, Issue 3 AUG 2015
A POWER QUALITY COMPENSATOR WITH RENEWABLE POWER GENERATION
SYSTEMS USING ACTIVE POWER FILTER
P.MOUNIKA1, SUBHASH RATHOD2
1
2
PG Scholar , VIJAY RURAL ENGINEERING COLLEGE , Telangana, India
Associate Professor , VIJAY RURAL ENGINEERING COLLEGE, Telangana, India
Abstract- In this paper we proposed a propelled
voltage loops. PI controllers must be designed based
controller called as fuzzy logic controller with dynamic
on the equivalent linear model, while predictive
force channel actualized as four-leg voltage-source
controllers use the nonlinear model, which is closer
inverter utilizing a prescient control plan is exhibited. The
utilization of four-leg voltage-source inverter with fuzzy
controller
permits
the
remuneration
of
current
symphonious segments, and additionally uneven current
to real operating conditions. An accurate model
obtained using predictive controllers improves the
performance of the active power filter, especially
created by nonlinear burdens. The MATLAB/SIMULINK
during transient operating conditions, because it can
model is given results.
Quickly follow the current-reference signal while
maintaining a constant dc-voltage. But in this paper
we proposed a fuzzy logic controller instead of pi
INTRODUCTION
controller at dc link.
RENEWABLE generation affects power quality
due to its nonlinearity, since solar generation plants
and wind power generators must be connected to the
grid through high-power static PWM converters. The
non uniform nature of power generation directly
affects voltage regulation and creates voltage
So far, implementations of predictive control in
power converters have been used mainly in induction
motor drives. In the case of motor drive applications,
predictive control represents a very intuitive control
scheme that handles multivariable characteristics,
simplifies the treatment of dead-time compensations,
distortion in power systems. This new scenario in
and permits pulse-width modulator replacement.
power
However, these kinds of applications present
distribution
systems will require more
disadvantages related to oscillations and instability
sophisticated compensation techniques.
Although active power filters implemented with
three-phase four-leg voltage-source inverters (4LVSI) have already been presented in the technical
literature, the primary contribution of this paper is a
predictive
control
algorithm
implemented
specifically
Traditionally,
active
for
designed
this
and
application.
created from
unknown
load
parameters. One
advantage of the proposed algorithm is that it fits
well in active power filter applications, since the
power converter output parameters are well known.
These output parameters are obtained from the
converter output ripple filter and the power system
been
equivalent impedance. The converter output ripple
controlled using pretended controllers, such as PI-
filter is part of the active power filter design and the
power
filters have
type or adaptive, for the current as well as for the dc-
IJOEET
power system impedance is obtained from well-
149
Volume 3, Issue 3 AUG 2015
known standard procedures. In the case of unknown
system impedance parameters, an estimation method
can be used to derive an accurate R–L equivalent
impedance model of the system .This paper presents
the mathematical model of the 4L-VSI and the
principles of operation of the proposed predictive
control scheme, including the design procedure. The
complete description of the selected current reference
Fig.2. Three-phase equivalent circuit of the proposed
shunt active power filter
generator implemented in the active power filter is
also presented. Finally, the proposed active power
filter and the effectiveness of the associated control
scheme compensation are demonstrated through
simulation and validated with experimental results
obtained in a 2 kVA laboratory prototype.
Extract the maximum energy possible from wind and
sun. The electrical energy consumption behavior is
random and unpredictable, and therefore, it may be
single- or three-phase, balanced or unbalanced, and
linear or nonlinear. An active power filter is
connected in parallel at the point of common
FOUR-LEG CONVERTER MODEL
coupling to compensate current harmonics, current
unbalance, and reactive power. It is composed by an
Fig. 1 shows the configuration of a typical power
distribution system with renewable power generation.
It consists of various types of power generation units
and different types of loads. Renewable sources, such
as wind and sunlight, are typically used to generate
electricity for residential users and small industries.
Both types of power generation use ac/ac and dc/ac
electrolytic capacitor, a four-leg PWM converter, and
a first-order output ripple filter, as shown in Fig. 2.
This circuit considers the power system equivalent
impedance Zs, the converter output ripple filter
impedance Zf, and the load impedance ZL.
The four-leg PWM converter topology is shown in
Fig. 3.
static PWM converters for voltage conversion and
battery banks for long term energy storage. These
converters perform maximum power point tracking to
This converter topology is similar to the conventional
three-phase converter with the fourth leg connected
to the neutral bus of the system. The fourth leg
increases switching states from 8 (23) to 16 (24),
improving control flexibility and output voltage
quality [21], and is suitable for current unbalanced
compensation. The voltage in any leg x of the
converter, measured from the neutral point (n), can
Fig.1. Stand-alone hybrid power generation system
be expressed in terms of switching states, as follows:
with a shunt active power filter
IJOEET
150
=
−
,
= , , , .
(1)
Volume 3, Issue 3 AUG 2015
applied to the power converter, according to
predefined optimization criteria. The predictive
The mathematical model of the filter derived from
control algorithm is easy to implement and to
understand, and it can be implemented with three
the equivalent circuit shown in Fig. 2 is
main blocks, as shown in Fig.3.
=
−
−
(2)
Where Req and Leq are the 4L-VSI output
parameters expressed as Thevenin impedances at the
converter output terminals Zeq .Therefore, the
Thevenin equivalent impedance is determined by a
series connection of the ripple filter impedance Zf
and a parallel arrangement between the system
equivalent impedance Zs and the load impedance ZL
Fig.3. Proposed predictive digital current control
block diagram
=
+
≈
+
(3)
1) Current Reference Generator: This unit is
designed to generate the required current reference
For this model, it is assumed that ZL _ Zs , that the
that is used to compensate the undesirable load
resistive part of the system’s equivalent impedance is
current components. In this case, the system voltages,
neglected, and that the series reactance is in the range
the load currents, and the dc-voltage converter are
of 3–7% p.u., which is an acceptable approximation
measured, while the neutral output current and
of the real system. Finally, in (2) Req = Rf and Leq = Ls
neutral load current are generated directly from these
+ Lf .
signals (IV).
DIGITAL PREDICTIVE CURRENT
2) Prediction Model: The converter model is used to
CONTROL
predict the output converter current. Since the
The block diagram of the proposed digital
controller operates in discrete time, both the
predictive current control scheme is shown in Fig. 4.
controller and the system model must be represented
This control scheme is basically an optimization
in a discrete time domain [22]. The discrete time
algorithm and, therefore, it has to be implemented in
model consists of a recursive matrix equation that
a microprocessor. Consequently, the analysis has to
represents this prediction system. This means that for
be developed using discrete mathematics in order to
a given sampling time Ts, knowing the converter
consider additional restrictions such as time delays
switching states and control variables at instant kTs, it
and approximations. The main characteristic of
is possible to predict the next states at any instant [k +
predictive control is the use of the system model to
1]Ts .Due to the first-order nature of the state
predict the future behavior of the variables to be
equations that describe the model in (1)–(2), a
controlled. The controller uses this information to
sufficiently accurate first-order approximation of the
select the optimum switching state that will be
derivative is considered in this paper.
IJOEET
151
Volume 3, Issue 3 AUG 2015
produces this minimal value and applies it to the
≈
[
]
[ ]
(4)
converter during the k + 1 state.
The 16 possible output current predicted values can
CURRENT REFERENCE GENERATION
be obtained from (2) and (4) as
A dq-based current reference generator scheme is
used to obtain the active power filter current
[ + 1] =
(
[ ]−
+ 1−
[ ])
reference signals. This scheme presents a fast and
[ ]
(5)
accurate signal tracking capability. This characteristic
avoids voltage fluctuations that deteriorate the
current reference signal affecting compensation
As shown in (5), in order to predict the output
performance [28]. The current reference signals are
current io at the instant (k + 1), the input voltage
obtained from the corresponding load currents as
value vo and the converter output voltage vxN, are
shown in Fig.4. This module calculates the reference
required. The algorithm calculates all 16 values
signal currents required by the converter to
associated with the possible combinations that the
compensate reactive power, current harmonic and
state variables can achieve.
current imbalance. The displacement power factor
3) Cost Function Optimization: In order to select the
(sinφ(L) ) and the maximum total harmonic distortion
optimal switching state that must be applied to the
of the load (THD(L) ) defines the relationships
power converter, the 16 predicted values obtained for
between the apparent power required by the active
io[k + 1] are compared with the reference using a cost
power filter, with respect to the load, as shown
function g, as follows:
∅( )
( )
=
g[k + 1] = (i∗ [k + 1] − i [k + 1]) +
(7)
( )
(i∗ [k + 1] − i [k + 1]) +
(i∗ [k + 1] − i
[k + 1]) +
(i∗ [k + 1] − i [k + 1])
(6)
Where the value of THD(L) includes the maximum
compensable harmonic current, defined as double the
sampling frequency fs. The frequency of the
The output current (io) is equal to the reference
(i*o) when g = 0. Therefore, the optimization goal of
the cost function is to achieve a g value close to zero.
The voltage vector vxN that minimizes the cost
function is chosen and then applied at the next
sampling state. During each sampling state, the
maximum current harmonic component that can be
compensated is equal to one half of the converter
switching frequency.
The dq-based scheme operates in a rotating
reference frame; therefore, the measured currents
must be multiplied by the sin (wt) and cos (wt)
switching state that generates the minimum value of
signals. By using dq-transformation, the d current
g is selected from the 16 possible function values.
component is synchronized with the corresponding
The algorithm selects the switching state that
phase-to-neutral system voltage, and the q current
component is phase-shifted by 90◦. The sin (wt) and
IJOEET
152
cos (wt) synchronized reference signals are obtained
Volume 3, Issue 3 AUG 2015
The resulting signals id and iq are transformed back to
from a synchronous reference frame (SRF) PLL [29].
a three-phase system by applying the inverse Park
The SRF-PLL generates a pure sinusoidal waveform
and Clark transformation, as shown in (9). The cutoff
even when the system voltage is severely distorted.
frequency of the LPF used in this paper is 20 Hz
Tracking errors are eliminated, since SRF-PLLs are
i∗
i∗
i∗
designed to avoid phase voltage unbalancing,
harmonics (i.e., less than 5% and 3% in fifth and
seventh, respectively), and offset caused by the
nonlinear load conditions and measurement errors
=
[30]. Equation (8) shows the relationship between the
real currents iLx(t) (x = u, v,w) and the associated dq
components (id and iq )
1
⎡
1
0 ⎤
⎢√2
⎥
1 √3 ⎥
2⎢ 1
−
2
2 ⎥
3 ⎢√2
⎢
⎥
⎢ 1 − 1 − √3⎥
⎣√2
2
2⎦
i
1
0
0
× 0 sin ωt − cos ωt i∗
0 cos ωt sin ωt i∗
(9)
The current that flows through the neutral of the
load
is
compensated
by
injecting
the
same
instantaneous value obtained from the phase-currents,
phase-shifted by 180◦, as shown next
Fig.4. dq-based current reference generator block
i∗ = −(i
+i
+i
)
(10)
diagram
One of the major advantages of the dq-based
current reference generator scheme is that it allows
the implementation of a linear controller in the dc-
(8)
voltage control loop. However, one important
A low-pass filter (LFP) extracts the dc component
disadvantage of the dq-based current reference frame
of the phase currents id to generate the harmonic
algorithm used to generate the current reference is
reference components −id. The reactive reference
that a second order harmonic component is generated
components of the phase-currents are obtained by
in id and iq under unbalanced operating conditions.
phase-shifting
dc
The amplitude of this harmonic depends on the
components of iq by 180◦. In order to keep the dc-
percent of unbalanced load current (expressed as the
voltage constant, the amplitude of the converter
relationship between the negative sequence current iL,
reference current must be modified by adding an
2 and the positive sequence current iL, 1). The
active power reference signal ie with the d-
second-order harmonic cannot be removed from id
component, as will be explained in Section IV-A.
and iq, and therefore generates a third harmonic in the
IJOEET
the
corresponding
ac
and
153
reference current when it is converted back to abc
Volume 3, Issue 3 AUG 2015
represents a simple and effective alternative for the
frame [31]. Fig. shows the percent of system current
dc-voltage control.
imbalance and the percent of third harmonic system
The dc-voltage remains constant (with a minimum
current, in function of the percent of load current
value of 6 vs(rms)) until the active power absorbed
imbalance. Since the load current does not have a
by the converter decreases to a level where it is
third harmonic, the one generated by the active power
unable to compensate for its losses. The active power
filter flows to the power system
absorbed by the converter is controlled by adjusting
the amplitude of the active power reference signal ie ,
which is in phase with each phase voltage. In the
block diagram shown in Fig. 5, the dc-voltage vdc is
measured and then compared with a constant
reference value v* dc. The error (e) is processed by a
PI controller, with two gains, Kp and Ti . Both gains
are calculated according to the dynamic response
Fig.6. Relationship between permissible unbalance
requirement. Fig. 7 shows that the output of the PI
load currents, the corresponding third-order harmonic
controller is fed to the dc-voltage transfer function
content, and system current imbalance (with respect
Gs, which is represented by a first-order system (11)
to positive sequence of the system current, is,1 ).
G(s) =
=
√
∗
(11)
In this paper control is done by the fuzzy logic
controller. The control scheme consists of a Fuzzy
controller, a limiter, and a three phase sine wave
generator for the generation of the internal structure
Fig.5. DC-voltage control block diagram
A. DC-Voltage Control
of the control circuit. The control scheme consists of
a Fuzzy controller, a limiter, and a three phase sine
The dc-voltage converter is controlled with the
wave generator for the generation of reference
fuzzy logic controller (FLC). This is an important
currents and switching signals. The peak value of the
issue in the evaluation, since the cost function (6) is
reference current is estimated by regulating the DC
designed using only current references, in order to
link voltage. The actual capacitor voltage is
avoid the use of weighting factors. Generally, these
compared with a set reference value. The error signal
weighting factors are obtained experimentally, and
is then processed through a Fuzzy controller, which
they are not well defined when different operating
contributes to the zero steady error in tracking the
conditions are required. Additionally, the slow
reference current signal. A fuzzy controller converts
dynamic response of the voltage across the
a linguistic control strategy into an automatic control
electrolytic capacitor does not affect the current
strategy, and fuzzy rules are constructed either by
transient response. For this reason, the PI controller
expert experience or with a knowledge database.
IJOEET
154
Firstly, the input Error ‘E’ and the change in Error
Volume 3, Issue 3 AUG 2015
A fuzzy inference system (or fuzzy system) basically
‘4E’ have been placed with the angular velocity to be
consists of a formulation of the mapping from a given
used as the input variables of the fuzzy logic
input set to an output set using fuzzy logic. This
controller. Then the output variable of the fuzzy logic
mapping process provides the basis from which the
controller is presented by the control Current Imax.
inference or conclusion can be made. A fuzzy
To convert these numerical variables into linguistic
inference process consists of the following steps:
variables, the following seven fuzzy levels or sets are
chosen: NB (negative big), NM (negative medium),
NS (negative small), ZE (zero), PS (positive small),
PM (positive medium), and PB (positive big) as
shown in below fig
Step 1: Fuzzification of input variables.
Step 2: Application of fuzzy operator (AND, OR,
NOT) in the IF (antecedent) part of the rule.
Step 3: Implication from the antecedent to the
consequent (THEN part of the rules).
Step 4: Aggregation of the consequents across the
rules.
Step 5: Defuzzification.
Rule Base: The elements of this rule base table are
The crisp inputs are converted to linguistic
determined based on the theory that in the transient
variables in fuzzification based on membership
state, large errors need coarse control, which requires
function (MF). An MF is a curve that defines how the
coarse input/output variables, while in the steady
values of a fuzzy variable in a certain domain are
state, small errors need fine control, which requires
mapped to a membership value μ (or degree of
fine input/output variables. Based on this, the
membership) between 0 and 1. A membership
elements of the rule table are obtained as shown in
function can have different shapes. The simplest and
below fig
most commonly used MF is the triangular-type,
which can be symmetrical or asymmetrical in shape.
A trapezoidal MF has the shape of a truncated
triangle. Two MFs are built on the Gaussian
distribution curve: a simple Gaussian curve and a two
sided
composite
of
two
different
Gaussian
distribution curves. The bell MF with a flat top is
FUZZY LOGIC CONTROLLER
somewhat different from a Gaussian function. Both
Gaussian and bell MFs are smooth and non-zero at all
points. The implication step helps to evaluate the
consequent part of a rule. There are a number of
implication methods in the literature, out of which
Mamdani and TS types are frequently used. Mamdani
proposed this method which is the most commonly
used implication method. In this, the output is
IJOEET
155
truncated at the value based on degree of membership
to give the fuzzy output.
Takagai-Surgeon-Kang method of implication is
Volume 3, Issue 3 AUG 2015
A simulation model for the three-phase four-leg
PWM converter with the parameters shown in Table I
has been developed using MATLAB-Simulink. The
different from Mamdani in a way that, the output
objective
is
to
verify
MFs is only constants or have linear relations with
compensation effectiveness of the proposed control
the inputs. The result of the implication and
scheme under different operating conditions. A six-
aggregation steps is the fuzzy output which is the
pulse rectifier was used as a nonlinear load. The
union of all the outputs of individual rules that are
proposed
validated or “fired”. Conversion of this fuzzy output
programmed using an S-function block that allows
to crisp output is defines as defuzzification.
simulation of a discrete model that can be easily
There are many methods of defuzzification out of
implemented in a real-time interface (RTI) on the
which Center of Area (COA) and Height method are
dSPACE DS1103 R&D control board.
predictive
the
control
current
harmonic
algorithm
was
frequently used. In the COA method (often called the
center of gravity method) of defuzzification, the crisp
output of particular variable Z is taken to be the
Simulations were performed considering a 20 [μs] of
geometric center of the output fuzzy value μout(Z)
sample time. In the simulated results shown in Fig.
area, where this area is formed by taking the union of
8, the active filter starts to compensate at t = t1. At
all contributions of rules whose degree of fulfillment
this time, the active power filter injects an output
is greater than zero. Here in this scheme, the error e
current
and change of error ce are used as numerical
components, current unbalanced, and neutral current
variables from the real system. To convert these
simultaneously. During compensation, the system
numerical variables into linguistic variables, the
currents is show sinusoidal waveform, with low total
following seven fuzzy levels or sets are chosen as:
harmonic distortion (THD = 3.93%).At t = t2, a three-
NB (negative big), NM (negative medium), NS
phase balanced load step change is generated from
(negative small), ZE (zero), PS (positive small), PM
0.6 to 1.0 p.u. The compensated system currents
(positive medium), and PB (positive big)
remain sinusoidal despite the change in the load
iou
to
compensate
current
harmonic
current magnitude. Finally, at t = t3, a single-phase
The fuzzy controller is characterized as follows:
load step change is introduced in phase u from 1.0 to

Seven fuzzy sets for each input and output.
1.3 p.u., which is equivalent to an 11% current

Triangular membership functions for
imbalance. As expected on the load side, a neutral
simplicity
current flows through the neutral conductor (iLn ), but
Fuzzification using continuous universe of
on the source side, no neutral current is observed (isn
discourse.
). Simulated results show that the proposed control


Implication using Mamdani's 'min' operator.

Defuzzification using the 'height' method.
scheme effectively eliminates unbalanced currents.
Additionally, Fig.6 shows that the dc-voltage remains
stable throughout the whole active power filter
SIMULATED RESULTS
operation.
IJOEET
156
Volume 3, Issue 3 AUG 2015
TABLE I
SPECIFICATION PARAMETERS
VARIABLE
DESCRIPTION
VALUE
v
Source voltage
55[v]
f
System frequency
50[Hz]
v
dc-voltage
162[v]
C
dc capacitor
2200[µF](2.0pu)
L
Filter inductor
5.0[mH](0.5pu)
R
Internal resistance
0.6[Ω]
within L
T
Sampling time
20[µs]
T
Execution time
16[µs]
Fig.6. Simulated waveforms of the proposed control
scheme (d) Load neutral current. (e) System neutral
current. (f) System currents
Fig.6. Simulated waveforms of the proposed control
scheme (g) DC voltage converter
Fig.6. Simulated waveforms of the proposed control
CONCLUSION
scheme (a) Phase to neutral source voltage (b) Load
Current (c) Active power filter output current
In this paper we enhanced the dynamic
current sounds and a receptive force remuneration
plan for force circulation frameworks with era from
renewable sources has been proposed to enhance the
present nature of the appropriation framework. Focal
points of the proposed plan are identified with its
straightforwardness, demonstrating, and execution.
The utilization of a prescient control calculation for
the converter current circle turned out to be a viable
answer for dynamic force channel applications,
enhancing current following capacity, and transient
IJOEET
157
reaction. Reproduced and exploratory results have
Volume 3, Issue 3 AUG 2015
implementation,” IEEE Trans. Power Electron., vol.
demonstrated that the proposed prescient control
26, no. 12, pp. 3501– 3513, Dec. 2011.
calculation is a decent distinct option for established
[7] X.Wei, “Study on digital pi control of current
straight control routines. The prescient current
loop in active power filter,” in Proc. 2010 Int. Conf.
control calculation is a steady and powerful
Electr. Control Eng., Jun. 2010, pp. 4287–4290.
arrangement. Reenacted and exploratory results have
[8] R. de Araujo Ribeiro, C. de Azevedo, and R. de
demonstrated the remuneration viability of the
Sousa, “A robust adaptive control strategy of active
proposed dynamic force channel.
power filters for power-factor correction, harmonic
compensation, and balancing of nonlinear loads,”
IEEE Trans. Power Electron., vol. 27, no. 2, pp. 718–
REFERENCES
[1] J. Rocabert, A. Luna, F. Blaabjerg, and P.
Rodriguez, “Control of power converters in AC
microgrids,” IEEE Trans. Power Electron., vol. 27,
730, Feb. 2012.
[9] J. Rodriguez, J. Pontt, C. Silva, P. Correa, P.
Lezana, P. Cortes, and U. Ammann, “Predictive
current control of a voltage source inverter,” IEEE
no. 11, pp. 4734–4749, Nov. 2012.
[2] M. Aredes, J. Hafner, and K. Heumann, “Threephase four-wire shunt active filter control strategies,”
Trans. Ind. Electron., vol. 54, no. 1, pp. 495–503,
Feb. 2007
IEEE Trans. Power Electron., vol. 12, no. 2, pp. 311–
318, Mar. 1997.
[3] S. Naidu and D. Fernandes, “Dynamic voltage
restorer based on a fourleg voltage source converter,”
P.MOUNIKA Completed B.Tech.
Gener. Transm. Distrib., IET, vol. 3, no. 5, pp. 437–
447, May 2009.
in Electrical &Electronics Engineering in 2012 from
[4] N. Prabhakar and M. Mishra, “Dynamic
VIDYA VIKAS INSTITUTE OF TECHNOLOGY,
hysteresis current control to minimize switching for
Chevella,
three-phase four-leg VSI topology to compensate
Hyderabad and M.Tech in Power Electronics in
nonlinear load,” IEEE Trans. Power Electron., vol.
2015(pursuing)
25, no. 8, pp. 1935– 1942, Aug. 2010.
ENGINEERING COLLEGE Affiliated to JNTUH,
[5] V. Khadkikar, A. Chandra, and B. Singh, “Digital
Nizamabad,
signal processor implementation and performance
smile.kushi111@gmail.com
Rangareddy,
Affiliated
from
Telangana,
VIJAY
India.
to
JNTUH,
RURAL
E-mail
id:
evaluation of split capacitor, four-leg and three hbridge-based three-phase four-wire shunt active
filters,” Power Electron., IET, vol. 4, no. 4, pp. 463–
470, Apr. 2011.
[6] F. Wang, J. Duarte, and M. Hendrix, “Gridinterfacing converter systems with enhanced voltage
quality
for
microgrid
application;concept
and
SUBHASH RATHOD Completed
B.Tech in Electrical &Electronics Engineering in
2004 from KITS(S), Huzurabad , Karimnagar,
IJOEET
158
Volume 3, Issue 3 AUG 2015
Affiliated to JNTUH, Hyderabad and M.Tech in
Power Electronics Drives in 2008 from NIT
Warangal. Working as Associate Professor at VIJAY
RURAL ENGINEERING COLLEGE ,Nizamabad,
Telangana, India. Area of interest includes research
in
Power
Electronics
&
Drives.
E-mail
id:
rathodmt@gmail.com
IJOEET
159