Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Review on active filters and its performance with grid connected fixed
and variable speed wind turbine generator
Sajid Hussain Qazi a,b,n, Mohd Wazir Mustafa a
a
Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Malaysia
b
Department of Electrical Engineering, Mehran UET SZAB Campus, Khairpur Mir's, Pakistan

art ic l e i nf o a b s t r a c t

Article history: In this paper a thorough review on the configuration of active filters have been carried out along with the
Received 30 June 2015 comparison of its operation by mitigating load harmonics when attached with grid connected fixed and
Received in revised form variable speed wind turbine generator. For extraction of reference signal, Synchronous Reference Frame
25 November 2015
(SRF) technique has been used and for generating driving pulses Hysteresis Band Current Control (HBCC)
Accepted 1 December 2015
method has been used for active filter. MATLAB/SIMULINK software utilized to verify effectiveness of
proposed system. The simulation results represent that proposed system is operating under the har-
Keywords: monic limit specified in IEEE-519 standard with variable speed of wind.
Harmonic & 2015 Elsevier Ltd. All rights reserved.
Active Filter
WTG
Wind Speed

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
2. Wind turbine system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
2.1. Fixed speed wind turbine system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
2.2. Variable-speed wind turbine system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
3. Design of wind turbines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
4. Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
4.1. Effect of harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
4.2. Effect of harmonics on the wind turbine generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
4.3. International standard for harmonics control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
5. Methods for mitigating harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
5.1. Passive filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
5.2. Active filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
6. Classification of active power filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
6.1. Classification based on power rating and speed of response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
6.1.1. Low power applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
6.1.2. Medium power applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
6.1.3. High power application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
6.2. Classification based on power circuit connection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
6.2.1. Shunt active power filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
6.2.2. Series active power filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
6.3. Classification based on compensated variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
6.3.1. Reactive power compensation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
6.3.2. Compensation of harmonic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
6.3.3. Balancing 3-Ø systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
6.4. Classification based on control technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430

n
Corresponding author at: Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Malaysia. Tel.: þ 60 137190744.
E-mail address: [email protected] (S.H. Qazi).

http://dx.doi.org/10.1016/j.rser.2015.12.049
1364-0321/& 2015 Elsevier Ltd. All rights reserved.
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 421

6.4.1. Open loop control system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430

6.4.2. Close loop control system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
6.5. Classification based on reference signal estimation technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
6.5.1. Current/voltage reference synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
6.5.2. Current/voltage reference calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
6.5.3. Time domain technique. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
6.5.4. Frequency domain technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
7. Application of active filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
7.1. Inverter (power unit). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
7.2. Reference signal extraction technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
7.2.1. Synchronous reference frame method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
7.3. Gate driver signal generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
8. Proposed system and its parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
9. Simulation results at fixed speed WTG without active filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
10. Simulation results at variable speed WTG without active filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
11. Simulation result at fixed speed WTG with active filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
12. Simulation result at variable speed WTG with active filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
13. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436

1. Introduction generator considering profile of total harmonic distortion (THD)

level is the ultimate interest of this research. At the end, the
The problems of power quality now-a-days are of great concern effective solicitation of the proposed methodology will eventually
as the perturbations due to PQ problems usually create compli- function under international standard of harmonic control i.e.
cated issues. These issues are normally due to increased penetra- IEEE-519 with improved efficiency.
tion of power electronics devices having non-linear characteristics,
which changes sinusoidal nature of supply by adding harmonic
components [1]. If these loads are connected in a grid-connected 2. Wind turbine system
wind energy conversion system (WECS), it will definitely affect the
waveform of current and voltage at the point of common coupling As the requirement for energy is consistently expanding, spe-
(PCC) and the wind-turbine generator (WTG), which in return cialists have been showing up for novel types of assets to cut
produces more heat and ultimately decreases efficiency and life reliance on fossil energy. Contrasted with solar based and other
span of WTG [2]. A compensation technique should be utilized to renewable energy, wind energy is less expensive and more legit-
mitigate the effect of these harmonics from PCC and WECS and to imate. Throughout the years wind power turned into the most
improve supply waveform. For that an appropriate harmonic quickly developing renewable energy into power era [7]. Espe-
extraction technique has to be applied for proper performance of cially wind energy is a standout amongst the most hopeful
the scheme. renewable energy (RE) sources, which is free from Green House
Passive and Active filters have been implemented to reduce Gasses (GHG) emanations, and it can possibly fill the energy
harmonic distortion. Prasad et al. [3] utilized passive filter to demand because of its accessibility, which energizes interest
eliminate harmonics from the system, however passive filter have around the world. It is one of the fastest developing and cost-
drawback that it only indemnify the system for under and over effective assets among the distinctive RE sources that have begun
compensation. Active Power Filter offers a liberal solution to the to be rehearsed worldwide for sustainable atmosphere friendly
issues of power quality, which allow the compensation of current power network [8].
harmonics and together with reactive power compensation. Looking at the most recent advanced technologies, wind energy
Dos Reis et al. [4] applied APF to mitigate harmonics from generators amazingly made strides. These advances brought on a
permanent magnet synchronous generator (PMSG) based wind perceptible ascent in wind turbines presence among generation
turbine generator in its islanding mode with fixed wind speed. Tsai technologies in recent years [9]. At that, places were a few types of
et al. [2] also proposed APF with its variable frequency operation generators, which have been used as wind energy converters. The
under islanding mode of wind turbine system (WTS). Gaillard et al. most recognized type of generator in the days of yore was syn-
[5] employed APF with doubly fed induction generator (DFIG) chronous generators while nowadays by assembling innovation
based fixed speed wind turbine to analyse the operational features enhancements, impelling generators of various types turn out to
of proposed system under the same islanding mode. Hoseinpour be more acknowledged inside of wind energy conversion field
et al. [6] put forward modified modulation technique for APF with [10,11]. Induction generators, especially doubly fed induction
PMSG wind turbine generator under islanding mode with fixed generators are turning out to be more well known in renewable
speed. All of them used different reference signal extraction energy work [12,13]. To the extend, commercial utilization of wind
techniques and modulation methods. turbines are concerned, the quickest advancement of the direct-
In this paper, a thorough review on active power filter config- driven wind turbine (WT) with a permanent magnet synchronous
uration is carried out and based on review active filter with better generator (PMSG) can be attributed to its bare structure, low cost
extraction technique and improved driving signal generation sys- of repairs, elevated conversion efficiency and high reliability
tem will be used with grid connected fixed and variable speed [14,15]. At present, two primary types of wind turbines available
wind turbine system separately to analyse the effectiveness of as:
proposed system. Under this study, effect of wind turbine at PCC
under fixed and varying nature of wind speed will be investigated. (i) Fixed Speed Wind Turbines,
Furthermore, the effect of nonlinear load at PCC and on the wind (ii) Variable Speed Wind Turbine
422 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Nomenclature usq Stator quadrature axis voltage

SRF Synchronous Reference Frame
PCC Point of Common Coupling P Pole pair
HBCC Hysteresis Band Current Control d-q Direct-Quadrature
THD Total Harmonic Distortion ωr Rotor Speed
PLL Phase locked loop WECS Wind Energy Conversion System
APF Active Power Filter ψm Amplitude of flux induced
id Direct axis current RE Renewable Energy
PWM Pulse Width Modulation vLa, vLb, vLc Three phase source voltages
iq Quadrature axis current GHG Green House Gases
PMSG Permanent Magnet Synchronous Generator iLa, iLb, iLc Three phase load currents
Ld Direct axis inductance WT Wind Turbine
WTG Wind Turbine Generator isa, isb, isc Extracted reference currents (SRF Technique)
Lq Quadrature axis inductance SCIG Squirrel Cage Induction Generator
WTS Wind Turbine System id ; iq Fundamental active and reactive current component
Rs Stator Resistance WRIG Wound Rotor Induction Generator
DFIG Doubly Feed Induction Generator ι~d ; ι~q Harmonic active and reactive current component
usd Stator direct axis voltage HVDC High Voltage DC
PI Proportional Integral DSP Digital Signal Processing

2.1. Fixed speed wind turbine system

Fixed-speed WECS are electrically straightforward devices,

comprising of an aerodynamic rotor connected through gearbox
and shaft and driving a squirrel cage induction generator (SCIG) or
a wound rotor induction generator (WRIG). The slip, and subse-
quently the speed of rotor of the generator fluctuates with the
amount of power generated. Variation in rotor speed is ordinarily
very low (roughly 1 to 2 percent); therefore, this WECS is known
as constant or fixed speed WECS [16]. From the design and man-
ufacturing point of view, the capital cost of this type of WT is low Fig. 1. Fixed Speed Squirrel Cage Induction Generator.
because no power electronic devices are used [17]. These are
intended to accomplish maximum efficiency at any one specific
wind speed. At wind speeds above and below the specified wind
speed, the energy catch does not achieve the maximum quality.
Fixed-speed WECS is mechanically basic, stable, reliable and
well-demonstrated. They require less cost for maintenance and
electrical parts. Alternately, these experience the ill effects of the
inconveniences of mechanical stress, limited power quality con-
trol, and poor wind energy conversion efficiency. Fig. 1 illustrating
fixed speed wind turbine with squirrel cage induction generator
[9].
In United States fixed wind speed turbines were used on the
coast of Long Island, New York. They designed, fabricated, and
tested 13 different small wind turbine designs (ranging from 1 kW
to 40 kW), five large (100 kW to 3.2 MW) horizontal-axis wind
turbine (HAWT) designs, and several vertical axis wind turbine
(VAWT) designs ranging from 5 to over 500 kW. Fig. 2 showing the
wind turbine used in that project.

2.2. Variable-speed wind turbine system

As the extent of the WECS is getting to be bigger and the

entrance of wind power inside of a power network is expanding, Fig. 2. Wind turbines on the coast of Long Island, NY.
the inalienable issues of fixed speed WECS turned out to be more
prominent, particularly in ranges with moderately weak grid of interest of variable-speed WECS are enhanced power capture,
configuration. To surmount these issues and to take after with the improved system efficiency, better power quality with less flicker,
grid codes connection necessities, the propensity in present day shrunk mechanical stress, decreased fatigue, and lessened acoustic
WECS innovation is to apply variable-speed ideas. noise. Moreover, the existence of power converters in wind tur-
With the increment in power electronic converters, which are bines likewise gives high-potential control capacities for both vast
connected to attach wind turbines to the power system, variable- current wind turbines and wind farms to fulfill the high specia-
speed wind energy ideas are getting normal. The essential points lized codes forced by the power grid administrators. The essential
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 423

Fig. 4. Variable Speed Wind Turbine System (PMSG).

Fig. 3. Variable Speed Wind Turbine System (DFIG).

qualities of variable-speed WECS are the ability to control active

and reactive power (frequency and voltage control), prompt
response under the situation of transient and dynamic in power
system, impact on network stability and improved power quality.
As the wind speed is constantly variable, the electrical frequency
of the generator likewise differs and must hence be decoupled
from the grid frequency, and this could be accomplished by
applying power electronic converters in the middle of generator
and power grid. The power converters decouple grid system fre-
quency from the mechanical frequency of rotor to empower vari-
able operation of wind turbine.
Classical synchronous generator and concept of doubly feed
induction generator (DFIG) can be used for variable-speed WECS
[17]. The block diagram of DFIG based wind turbine is shown in
Fig. 3. The development of DFIG is more promising for variable Fig. 5. Zorlu Energy Jhimpir, Pakistan.

wind speed performance because of its smaller size and cost of

power converters, less contents of harmonic in power system 3. Design of wind turbines
produced by the power converter, better controllability, require
In the wind turbine industry there are two basic types of tur-
small space, a better stability of the power system and the possi-
bines to choose, vertical axis wind turbines and horizontal axis
bility to generate reactive power.
wind turbines. They both have their advantages and disadvantages
Other than induction generator a synchronous generator with
[20]. Some of the features are summarized in Table 1.
slow rotating speeds (5–30 rpm) can be utilized. Synchronous
generator with the high number of poles (multi pole synchronous
generator) with established excitation or permanent magnets can
4. Harmonics
be joined specifically to the turbine without a gearbox (Direct
Drive Wind Turbine) [18] as shown in Fig. 4. Furthermore, the
A harmonic can be defined as “A sinusoidal component of a
power converters used to connect WTG with grid have rating
periodic wave or quantity having a frequency that is an integral
equivalent to the power rating of the generator. Because of the multiple of the fundamental frequency” [21]. Harmonics are
huge number of poles and low turning speed the generator must basically the consequence of power electronics (non-linear) based
propagate extensive torque. Thus, it has generally heavy in weight loads, which are increasing enormously in past decades and
and measurements, which unavoidably influences the develop- indicating that it will be increased even more in future. These
ments and measurements of the nacelle. This sort of power con- loads draw a non-sinusoidal current from sinusoidal voltage. The
verters conveys full power of the generator [18]. The depletion in types of devices are designed on basis of power electronics
price of permanent magnets and their accessibility in the market incorporating adjustable speed drives, power electronic supplies,
have altogether impacted the improvement and use of permanent DC engine drives, battery chargers, electronic ballasts are
magnet generators for direct-driven systems. responsible for the increment in harmonic distortion. These loads
Nowadays, variable speed wind turbines are extensively used are the essential sources of harmonics in the distribution system,
by technologists and power supply companies. One of the example harmonic currents generated by these loads are infused back to
of variable speed wind turbine is Zorlu Wind Energy Jhimpir, the grid through PCC bringing about distortion in voltage supply
Pakistan, using Turkish made VENSYS 62 wind turbines of 1.2 MW [22].
capacity and 1.8 MW capacity WT, supplied by Vestas, Denmark.
The site of Zorlu Energy is shown in Fig. 5. 4.1. Effect of harmonics
The important differences between variable speed operation, as
employed in modern large wind turbines and the older conven- Power system issues identified with harmonics are uncommon,
tional fixed speed operation are: yet it is possible for a number of undesirable impacts to happen.
Abnormal amount of harmonic distortion can bring about such
(i) Fixed speed operation have low reactive power control and impacts as expanded transformer, capacitor, motor or generator
the grid voltage level cannot be controlled [19]. heating, disoperation of electronic apparatus (which depends on
(ii) Variable speed operation below rated power can enable voltage zero crossing or is delicate to wave shape) [23]. The issue
increased energy capture; and of resonance between transformer inductance and line capaci-
(iii) Variable speed capability above rated power (even over quite a tance would occur due to increase in eddy current losses and
small speed range) can substantially relieve loads, ease pitch hysteresis losses in the reactor of transformers. As increment in
system duty and much reduce output power variability. reactive power in capacitors thus, it will increase dielectric losses,
424 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Table 1
Summary on design features of wind turbines.

Feature Horizontal axis wind turbine Vertical axis wind turbine

Axis of rotation Parallel to ground Perpendicular to ground

Number of blades working at a time (If 03 All work at a time Only one at a time
blades)
Working wind direction It only works in a specific wind direction, when wind direction change it It works in all wind directions
have to move itself in that direction
Efficiency at a given wind speed High efficiency Low efficiency
Area needed for installation More area needed for HAWT, height is also the concern Less area needed as compare to HAWT
Location of installation HAWT cannot be installed near population VAWT is designed for installation in urban
areas

which may lead to failure of the capacitor. Because of skin and Table 2
proximity effect, an increase in heat and hence breakdown of Voltage Distortion Limits.

dielectric and failure of cable insulation may occur. Changing the

Bus voltage at PCC Individual harmonic Total harmonic distortion THD
rate of rise of transient recovery voltage and operation of blowout (%) (%)
coil may affect the switchgear because of these harmonics further,
V r 1.0 kV 5.0 8.0
current-carrying parts will be affected by skin and proximity
1.0 kV o V r69 kV 3.0 5.0
effects. Because of interference of harmonics it may affect time- 69 kV o V r161 kV 1.5 2.5
delay characteristics of relay, especially microprocessor based 161 kV o V 1.0 1.5
relays. Likewise, in cables false tripping may occur. Due to increase
in heat, torque ripple may be produced in rotor of generators,
which adds to maintenance requirements and may finally con- Table 3
tribute to bankruptcy. Due to harmonics, it upturns the eddy Current distortion limits (120 V–69 kV).
current and skin effects losses in motors. Depending on the design
ISC/IL h o 11 11r h o17 17r h o 23 23r h o 35 35Z h
features and conventions of operation of protective devices, har-
monics may affect their functioning. It may stay in either series or o 20 4 2 1.5 0.6 0.3
parallel resonances in the power system network. In improver, it 20–50 7 3.5 2.5 1 0.5
50–100 10 4.5 4 1.5 0.7
may cause disoperation of electronic devices (which depends on 100–1000 12 5.5 5 2 1
voltage zero crossing detection or sensitivity to the voltage 41000 15 7 6 2.5 1.4
waveform) [1] and may cause interference with telephones.

4.2. Effect of harmonics on the wind turbine generators Table 4

Current distortion limits (69 kV–161 kV).
Variable-speed WT can be used in stand-alone mode or can be
ISC/IL h o 11 11r h o17 17r h o 23 23r h o 35 35Z h
connected to the grid. PMSG and DFIG are widely used because of
their high performance and power quality features when com- o 20 2 1 0.75 0.3 0.15
pared with fixed speed WT. The majority of research interests 20–50 3.5 1.75 1.25 0.5 0.25
50–100 5 2.25 2 0.75 0.35
related to grid-connected WECS due to the increasing ratio of 100–1000 6 2.75 2.5 1 0.5
installing WT with the grid in the last decade [24]. In this aspect, 41000 7.5 3.5 3 1.25 0.7
the control and operation of WT depend on active and reactive
power control, fault ride through probability and compensation for
non-linear and unbalanced loads. The typical characteristic of
Table 5
unbalanced load causing a non-linear current with high THD Current harmonic distortion limits ( 4 161 kV).
value, due to this non-linear current, stator output voltage at PCC
becomes non-sinusoidal with odd harmonic's (6n þ1), multiples of ISC/IL h o 11 11 rh o 17 17r h o 23 23r h o35 35Z h

fundamental frequency. ωs which deteriorates the performance of o 20 2 1 0.75 0.3 0.15

other connected loads with generator [25]. Therefore, it is neces- 20–50 3 1.5 1.15 0.45 0.22
sary to improve the quality of power by mitigating these current
harmonics generated due to non-linear loads, so to get pure
sinusoidal output at PCC.
5. Methods for mitigating harmonics

4.3. International standard for harmonics control To overcome the harmonic's problem, several methods have
been proposed in the past several years to mitigate harmonic
In 1981, the IEEE presented the IEEE standard 519 "IEEE problems in the grid-connected variable-speed wind energy con-
Recommended Practices and Requirements for Harmonic Control version system in order to improve the power quality of wind
in Electrical Power Systems" to offer rules for managing harmonics energy [27]. Below are the methods used to mitigate harmonic
introduced by nonlinear loads so that power quality issues could contents from WT;
be whitewashed. According to modification of IEEE-519-1992,
breaking points for voltage and current harmonics are demon- (i) Passive Power Filter
strated in Tables 2–5 [26]. (ii) Active Power Filter
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 425

5.1. Passive filter width modulation (PWM) inverter is utilized as APF and reactive
power compensator. A capacitor on the DC side with regulated DC
The utilization of an extra system components for the mitiga- voltage is connected with voltage source APF, whereas an inductor
tion of harmonic, this methodology is called as power filters. In the with regulated DC current is connected with current source APF. In
power filter group Passive filters were introduced first for har- terms of filter losses and its capacity to diminish PWM carrier
monic compensation. The inductors, capacitors and damping harmonics; voltage source APF is considered to be better, whereas
resistors can be joined either in arrangement of series or in parallel the current source type APF is superior in terms of compensating
to design a Passive filters [1]. A few positions of passive filter with current dynamics with reliability and protection. However, there are
distinctive filtering qualities and for diverse application demands few disadvantages of the APF:
are accessible. A series or either parallel connection of passive
filters can be made with power network. The operation of shunt (i) Initial and running cost of APF is high, moreover increased
passive filter is tuned at the prerequisite harmonic order to be losses and complex control system.
isolated out. The request can be of first, second, or third order, (ii) The exchange-off in the design of APFs with high-power rating
tuned on the filtering necessities. The operation of the shunt and fast dynamic current response.
passive filter is fundamentally influenced by the source impe-
dance. In the series arrangement, passive filter is joined in Contingent upon topology of connection, APF can be delegated
arrangement with the power system to counteract flow of har- as shunt APF, series APF or combined shunt series APF, as shown in
monic current by mean of a high impedance way to it. Customa- Fig. 6 [1].
rily, the passive filter with thyristor-controlled reactor (TCR) is
utilized to harmonic filter, in addition to compensation to reactive
current components [28]. Advantages of passive filters are simple 6. Classification of active power filters
maintenance, lower expenses and less complexity. However, in
practice, these filters have numerous disadvantages [25–28], The typical active power filter components and their inter-
which can be itemized as follows: connections are shown in Fig. 7. The data is supplied to the
reference current/voltage estimator, generated by nonlinear loads.
(i) The passive filter consider only one component; harmonic or The overall system control depends on reference signal generated
fundamental current component. It should consider both from the current/voltage estimator [41]. Consequently, generating
when they flow through the filter. switching signals to control the PWM generator. Furthermore,
(ii) With increase in harmonic current, the filter may be output of the PWM generator controls the active power filter. The
overloaded. APF can be connected in series or shunt depending on the com-
(iii) Existence of parallel resonance causes over voltage in the pensation required.
system. On the basis of the above categories, an APF can be classified
(iv) Performance depend upon source impedance, which is not according to subsequent measures [42].
known and may vary with network configuration.
(v) It may cause series resonance with AC source, leading to flow (i) The speed of response and power ratings,
of high currents within the filter. (ii) The criteria for the configuration of power-circuit configura-
(vi) The operating frequency should be considered in design of tion and its connections;
filter, as in the AC system frequency varies around its minimal (iii) The compensation of system parameters (e.g. current harmo-
value as per the varying load conditions. nics compensation, power factor improvement, unbalanced
(vii) Passive filter only eliminate frequencies to whom they are three-phase system etc.);
tuned to. (iv) The techniques employed for control of APF; and
(v) The reference signal estimation technique utilized.
5.2. Active filter
6.1. Classification based on power rating and speed of response
To overcome the problems of passive filters, active filters have
been introduced and researched. Over the course of the years, a The control technique to implement the required filter for
numerous setups of active power filter with their specific control compensation, depends on the power rating and speed of
methodologies have been proposed [29,30], and they have been response. These factors, power rating and speed of response follow
perceived progressively as an advantageous answer for the issues a reciprocal relationship [43–45]. Fig. 8 below showing classifica-
that have been made by high-power nonlinear loads [31–33]. tion based on these two factors. As a rule, the cost of any specific
Active power filter (APF) can be distributed into two categories; system is corresponding to the required response speed.
DC and AC power filters [34]. The configuration of DC APFs is to
make up for current and/or voltage harmonics from the high-power 6.1.1. Low power applications
utilizations of thyristor, for example, in high-voltage DC (HVDC) This application of this type is mainly concerned with power
system and high-power drives [35–38]. AC APFs are usually dis- system of rating below 100 kV A. Essentially, it is associated with
tinguished as Active Power Filters, and are likewise called as active domestic areas, hospitals and for the factories ranging from small
power line conditioners, active filters and active power quality to medium-sized loads. They have fast response time relatively,
conditioners. APFs can be employed with a medium or low voltage ranging from several microseconds to milliseconds. This type of
distribution system for compensating harmonics, supplying reactive active filter includes two categories, Single Phase APF and Three
power, and/or neutral current. It can likewise be utilized for eradi- Phase APF.
cation of voltage harmonics, control over terminal voltage, sup-
pressing voltage flicker, and in three-phase system it improves (i) Single phase system active filter
voltage balance [39,40]. When contrasted with passive filters, active Single phase APFs are normally available in low power ratings
filters have superior characteristics on compensating harmonic with [46–53]. They are appropriate for commercial or educational
varying line impedance in AC power systems and the variation in buildings with computers load [52], where the current har-
frequency of harmonic currents. Usually, a voltage source pulse- monics are dealt with at point of common coupling.
426 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Fig. 6. (a) Shunt APF, (b) Series APF, (c) Combined Series Shunt APF.

Fig. 7. Generalized block diagram of APF.

Appropriately, on a particular distribution system, several harmonic in the neutral line. This type of active filter operates
lower-power filters can be joined rather than using one filter more efficiently under varying operating conditions. The fore-
with large value on the incoming supply. As the consequences, most benefit of single-phase filters is their operation at higher
within one building large number of the single-phase loads frequencies which leads to improved performance as they
installed, it concern about the presence of large amounts of have to deal with only low power applications.
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 427

Fig. 8. Classification of APF according to power rating and speed of response.

(ii) Three phase system active filter then the main problem and is commonly compensated for by
The choice of filter configuration for three-phase applications utilizing customary static power conditioners filters [34], or by
depends on loads, either balanced or not. For the low power using parallel connected [44] sets of synchronous condensers and
loads, relatively 100 kV A, three single phases or one three employing cascaded multilevel-inverter VAR compensators
phase filter/compensator can be used with three phase sys- [60,62–65]. The response time for this application is in the range
tem. For a system with balanced loads, a single three-phase- of several seconds.
inverter configuration is employed [34,43,54–56]. This is One of the application of high power active filters is in Japanese
acceptable if the aim is simply mitigate current harmonics bullet train which uses several active filters in parallel combina-
but not the constraint for balancing voltages or currents in tion. The control techniques and requirements for co-ordination of
each phase, assuming same magnitudes and phase angles in these filters are however complicated [65,66].
every phase.
6.2. Classification based on power circuit connection
6.1.2. Medium power applications
The application of this type is mainly concerned with power The selection of application of active filters depends on Power-
system of rating from 100 kV A to 10 MV A [34,57]. The medium- circuit configurations. Some circuit configurations are suitable for
to high voltage distribution system [58], and drive systems with certain features of power ranges and control, discussed in this
high-power high-voltage [59], which unbalance the phases more, section and explained in Fig. 9.
fall under the classification of this category. The foremost purpose
of this type of filter is to eradicate or reduce the current harmo- 6.2.1. Shunt active power filter
nics. At high voltage distribution systems the application of active
filters is considered not feasible due to economic considerations The shunt active power filter, the most important and most
and reactive-power compensation [44], due to the high voltage widely used in industrial processes [34,54,55,67,68]. It is con-
and its associated problems of isolation and series/parallel con- nected to power system in parallel as shown in Fig. 10. The
nections of switches. Alternative approaches, including capacitive advantage of shunt APFs that it only carries the compensation
and inductive static compensators (such as LC circuits controlled current for compensating losses in the system and to maintain
by relay), tune able harmonic filters, line-commutated thyristor constant DC voltage at the terminals of capacitor with the addition
converters, synchronous condensers and cascaded multilevel- of a small active fundamental current component [42]. Even so,
inverter VAR compensators [60,61], are considered more appro- shunt APF can be utilized for compensating reactive power at
priate. The response time for the application of this filter ranges on high-voltage-distribution as well [42]. This configuration of filter is
the order of tens of milliseconds. commonly used to cancel current harmonics fed to the voltage
sources. Connecting several active filters in parallel will provide
6.1.3. High power application high currents, suitable for loads with high power ratings. However,
There are some limitations over the application of very high- the impedance of the source side is higher than the impedance of
power dynamic filters, as they are greatly cost inadequate, the diode rectifier. Thus, the injected compensating current from the
dearth of fast switching-frequency power strategies that can shunt active filter flows into both the source and diode rectifier. As
control the flow of current at such high power ratings [34,44]. The a result of this, shunt active filters may not completely cancel the
modern state-of-art semiconductor devices also cannot withstand harmonics, but can increase the DC ripple and the AC peak current
extra high voltage of hundred kilovolts, it can tolerate a few of the diode rectifier. A series inductance can be used with shunt
kilovolts. The combination of these switches in series-parallel is APF to avoid these problems.
conceivable, but challenging to implement and cost-ineffectual.
Fortunately, the system above 10 MVA has less harmonic pollution 6.2.2. Series active power filter
and is not of such a noteworthy issue as in systems with low The series APF does not compensate for load current harmo-
power ratings. In those systems the effect of harmonics would be nics, but act as an isolator to harmonics current from the source
minimized at the low-power, by the fitting of some medium- and side [69]. Series APFs work as a controllable voltage source [70]
low-power active filters. The compensation for reactive power is and they are industrially less common than the shunt APFs.
428 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Fig. 9. Division of active filter according to its power connection.

A Conn1

N B Conn2

C Conn3

Supply Nonlinear
Load
Conn1

Conn2

Conn3

Parallel APF

Fig. 10. Parallel APF.

Because of this reason that they cause to wield the full load quasi dynamic reactive power compensators with slower response
current, therefore, increase in the current rating while comparing and cheaper devices are accessible in the market. However, active
with the Shunt APF. Thus, it increases I2R losses and filter size [1]. filter used to compensation of reactive power is best suited for low
The core advantage of series APF over shunt APF is that it elim- power applications; subsequently it requires the same magnitude of
inates harmonics from voltage waveform and maintains a pure current as of rated load current for reactive-power compensation.
sinusoidal three phase voltages across the load [34,42,71]. This, Therefore, using active filter without the use of other power factor-
indeed, implies that this category of filter benefit load as this is correction devices, such as thyristor-controlled reactors and capaci-
used for improvement in the voltage quality of the system. It tors; particularly in single-phase systems [42], where the require-
provides a perfect sinusoidal waveform to the voltage sensitive ment is for accurate compensation without harmonics generation,
devices (such as power system protection devices and super- would be waste of sophisticated apparatus.
conductive magnetic-energy storage, etc.) Fig. 11, presenting active
filter connected in series with power system. Table 6, surveys the 6.3.2. Compensation of harmonic
elements of both shunt APF and series APF [1,42]. The compensation of harmonic is the most important para-
meter of system which is required in power networks and it is
6.3. Classification based on compensated variable sectioned into two categories, voltage harmonics compensation
and current harmonic compensation as follows.
The purpose of Active filters is to improve some of the power
system characteristics under interrogation. These signified para- (i) Current harmonic compensation
meters of power system characterized according to their control, In the power system with application of low and medium
leading to the subsections existing in Fig. 12. power ratings, compensation of current harmonics is of high
importance [34,43,54,61,68,70]. If, current harmonics reduces,
6.3.1. Reactive power compensation it decreases the distortion level in voltage waveform at point
Reactive power compensation is not only the factor which is of common coupling. The design criteria of power system
requiring active filters [56,72–74]. Conversely, concurrence of power depend on magnitude and waveform of current. Furthermore,
factor compensation with current harmonics is legitimately addres- it is always recommended to decrease the losses in cables and
sed by many publications as function of active filters. In contrast, feeders losses by reducing amount of RMS value of the total
owing to the fact, active-filter configurations rarely treat the problem current, which indicate the decrease in current harmonics.
of power-factor correction separately, because of the way that other This is because the total value of RMS current for nonlinear
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 429

A A A Conn1

N B B B Conn2

C C C Conn3

Supply Supply Nonlinear

Inductance Load

C
Three-Phase
Transformer

c
Conn1

Conn2

Conn3
Series APF

Fig. 11. Series APF.

Table 6
Evaluation of shunt and series APF.

Category Shunt APF Series APF

Connection with system – Parallel with distribution system – Connected in series with distribution system
Act as Current source – Voltage source
Filter rating – Voltage rated at full load rating, – Current rated at full load rating,
– Current rating comprise partially harmonics and partly – Voltage rating is partially compensated voltage component
reactive current component
Functioning – Harmonic load Current filtering, – Mitigation of Voltage harmonic, sag, swells,
– Compensation for Reactive current, – Mitigating Current harmonic,
– Mitigate Current unbalance. – Compensating Reactive current,
– Mitigate Current unbalance.
Characteristics of – Source impedance have no effect on compensation for – Source and load impedance have no effect on compensation for vol-
Compensation current source loads. tage source loads.
Application – Injected current may cause overcurrent when applied to a – When working with current source load a low-impedance parallel
voltage source load. branch (for improvement of power factor)
Load considered – Non-linear/inductive current source loads or harmonic – Non-linear/capacitive voltage source loads or harmonic containing
containing current source loads. voltage source loads

Fig. 12. Subsections of active filter according to compensated variables.

load is equivalent to the aggregate of the squares of the RMS [34,50]. At the point of common coupling of consumers, utility
values of each of the individual harmonics, as in Eq. (1) [2]. companies keep the terminal voltage within the normal limits
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi for voltage sag and total harmonic distortion and does not
I 23 þ I 25 þ I 27 þ … þ I 2n generally change much when load changes. For the harmonic-
THD ¼ ð1Þ
I1 voltage sensitive equipment, which require pure sinusoidal
voltage such as, protection devices and super-conducting
(ii) Voltage harmonic compensation energy storage [75]. The compensation for voltage and current
The issue of voltage harmonics compensation voltage is not harmonics are correlated; if at PCC the decrease in voltage
been addressed due to low impedance of power supplies harmonics ultimately reduce level of current harmonics.
430 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Though, the voltage harmonic compensation at the point of reactive power by simply injecting fixed amount of power in term of
common coupling does not eradicate the necessity for com- reactive current into the system. These types of systems do not check
pensating current-harmonic due to loads with nonlinear how much the compensation is effective. Likewise, in case of the
characteristics. customary techniques of filtering, such as switched passive-filter
banks, with addition to the following techniques:
6.3.3. Balancing 3-Ø systems
The problem of unbalanced currents and voltages in 3-Ø low (i) Eradication of harmonics by injecting third harmonic,
and medium power systems, consequently the currents and the (ii) Devices for cancellation of harmonic,
voltages are also not equally spaced at 120°. Active filter utilized to (iii) Systems with known constant load-harmonic pattern [45,60].
maintain and balance 3-Ø current and voltage.
6.4.2. Close loop control system
(i) Balancing 3-Ø system voltages Unlike, open-loop systems, the closed-loop technique add in with
The system voltage imbalance relies on upon the level of a feedback system, which then sense the under considered variables.
current imbalance and the supply impedance magnitude. These systems are more accurate in cancellation of harmonic and
Ultimately, causing unequal and unevenly spaced three-phase reduction in reactive-power. Practically, this type of control system
voltages. The significant value of supply impedance is the comprises of all new techniques for controlling harmonically idle
main cause of this problem. The solution to the problem like systems. These controllers are further divided into constant capacitor
this is to supplement the corresponding amount of instanta- voltage, constant inductor current, optimization techniques, linear
neous voltage to each phase of system, forcing it to follow the voltage control and some other techniques.
reference sinusoidal waveform. This type of problem occur
normally in low power rating system, because the supply (i) Constant capacitor voltage technique
impedance has no effect in medium- and high-power rating This method, is suitable for a DC link capacitor connected to
systems [68,76]. single-and three-phase inverter configurations, depend to the fact
(ii) Balancing 3-Ø system currents that this capacitor voltage will control the waveform of current
Likewise voltage balancing, current compensation is also simply by linking the capacitor with smoothing inductor with the
concerned with lower power applications of three-phase power system. The conventional technique (PWM) is then used to
system. The reason over this concern is that the magnitudes regulate resulting current. Furthermore, the voltage across term-
of supplied currents to the grid be subject to the level of inals of capacitor fluctuates, as energy is supplied to or taken from
imbalance in system, which mostly occurs in residential loads. the capacitor. To maintain constant voltage, a set reference voltage
The considered compensator under [68,76–78], have some is chosen. The error difference among the measured voltage of
limitations on their power capabilities because sometimes capacitor and its set value for reference governs the systems active
they are enforced to supply the rated value of current. The component of power needed to compensate for filter losses. This
three single phase H-bridge inverters are mainly used to error signal is then given to the current-controller error signal to
develop power circuit of this system devising the similar calculate the complete system error to be fixed by the current
energy-storage element. controller [34,45,47,50,54,55,67,76].
(ii) Constant inductor current technique
6.4. Classification based on control technique Unlike, capacitor is connected in DC link in capacitor voltage
control system; likewise, inductor in the DC link is con-
The classified control techniques presently available are shown nected. The operation of this technique is same as the
in Fig. 13. These techniques categorize into two parts, close loop operation of the constant-capacitor-voltage control system
and open loop controls. just the voltage of capacitor is interchanged with the current
of inductor. The two central approaches are then used to
6.4.1. Open loop control system exercise this technique:
The load current and the harmonics contents are sensed by open  Current pulse-width modulation: Likewise, the case of
loop control system. This system compensate for harmonics and/or constant capacitor-voltage control system, the value of

Fig. 13. Control techniques for active filter.

 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 431

Fig. 14. Reference estimation techniques for active filter.

current signal in a specific interval of time depends on PWM 6.5. Classification based on reference signal estimation technique
technique to represent average value of current
[43,44,54,61,72]. The Fig. 12, illustrating reference estimation techniques for
 Current pulse-amplitude modulation: On the basis of this active filters, as described in Fig. 5, estimated reference signal
control technique, required current waveform provided by for current/voltage is being processed by the feedback loops
active filter is controlled by amplitude modulation. However, constitutes a noteworthy and crucial measure for active-
the idea is established, yet it is impractical to use it with filtering techniques [22,82]. Estimation techniques elaborated
present technology of power electronics. in Fig. 14 by giving it the obliged reference for further pro-
(iii) Optimization techniques cessing; these estimation techniques can be categorized as
The control technique for optimization of switched-capacitor follows.
and lattice-filter circuits is the identical [46,51,52]. The rising
rate of current and its amplitude depends on size of capacitor 6.5.1. Current/voltage reference synthesis
and initial voltage. These two aspects patterns the switching To regulate harmonics in the system current/voltage, ana-
signals also, they extensive flexibility in shaping the current logue signal filter to be used in this technique. The imple-
waveform drawn by the filter [42]. The reduction in a mentation of this technique is simple because of analogue
predetermined numeral of individual harmonics of load– devices, in the time domain. However, one serious drawback,
current is the foremost task of controller, with this, also an error signal is introduced by signal active filter in phase
decreasing magnitude of THD level or the filter current angle and magnitude of system. Two categories emerged in this
fundamental components. Though, this could not be achieved
technique,
promptly, a time delay occurs during the detection in the
variance of harmonic current and the applying the appropriate
(i) High-Pass Filter
set of switching signals acquired from the procedure of
(ii) Low-Pass Filter
optimization. Thus, this system is primarily suitable for con-
stant or slowly varying loads.
6.5.2. Current/voltage reference calculation
(iv) Linear Voltage Control Technique
The harmonic calculation is usually implemented on account of
This technique is only appropriate for the active filter with
the major downside of the system beforehand utilized, which
capability of voltage-regulation [42,45]. Through continuous
integrates phase-angle and magnitude errors. Most ordinary
charging and discharging of capacitor, the voltage across the
techniques for estimation can be named as;
output capacitor can be linearly controlled. From the har-
monic reference value, capacitor-voltage reference is calcu-
(i) Time Domain,
lated which consist of the changing rate of the load–current
(ii) Frequency Domain, and
harmonics [42]. In this technique, sudden changes of inver-
(iii) Other modern Techniques
ter voltage waveform could be controlled through contin-
uous and smooth variation in the capacitor voltage, which
6.5.3. Time domain technique
ensures that variation in the supply current is controlled
Time-domain technique is the basis of instantaneous approx-
and therefore, the reduced switching frequency. This, actu-
imation of reference signal (current/voltage) from biased and
ally, decreases the magnitude of injected high-frequency
harmonically contaminated voltage and current signals. These
harmonics into the supply due to the existence of the PWM
methodologies are appropriate for both single-phase and three-
inverter.
(v) Other Techniques phase systems excepting for the synchronous-detection theorem
The additional techniques to control are addressed in [79–81], [83,84] and synchronous-reference-frame theorem [81,85–90],
they just simply offer some changes in the abovementioned which are suitable for three phase system. The following seven
procedures, giving basically more current or better execution categories come under the shadow of time domain techniques,
over their antecedents. These systems may incorporate the
utilization of state-of-the-art adaptive, predictive and sliding- (i) Instantaneous Reactive Power Technique,
mode controllers, which are typically hard to actualize with- (ii) Synchronous Detection Technique,
out the utilization of DSPs [81]. These strategies can be (iii) Constant Active Power Technique,
implemented in either the time domain or in the frequency (iv) Constant Power Factor Technique,
domain. (v) Fictitious Power Compensation Technique,
 432 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

(vi) Synchronous Reference Frame Technique, system to synchronously rotating components [85–88]. The
(vii) Synchronous Flux Detection Technique. transformed quantities can be easily analysed, as the compo-
nents of fundamental- frequency are transformed into DC
(i) Instantaneous Reactive Power Technique quantities [89]. In this technique active and reactive compo-
This technique is appropriate only for three-phase systems; in nents of the system are represented by direct and quadrature
which instantaneous power of the load is calculated. The components, respectively. The system being modulated at
calculated power consists of a DC component and an oscillat- different frequencies as harmonics of high-order are still
ing component. More than a certain interim of time these remain in the signal. These are the undesired components
oscillating components are being separated from power. Then, represent reference harmonic current which need to be
calculation of reference signal by correspondingly distributing eliminated from the system. The technique is quite stable
the total current to each of the three phases. This, whole since its controller mainly deals with DC quantities of the
operation carried out under the supposition that the voltage system. Three phase systems are pertinent to use this
waveforms are sinusoidal and the three-phase system is technique.
balanced [67,72,91]. Certainly, the performance of this techni- (vii) Synchronous Flux Detection Technique
que will be deprived, if applied to polluted supply system [67]. Likely, Synchronous reference frame technique, this technique
(ii) Synchronous Detection Technique too applies Park transformation to transfer the systems voltage
This technique, merely depends on the fact of balanced three and current into synchronously rotating frame of reference.
phase current, and its application is similar to instantaneous Conversely, the transformation technique is applied on the
reactive power technique [67]. The calculated average power filter inductance flux linkage, and the output voltages and
is divided equally between the three phases. The reference currents control in separate integral loops [94]. A time delay
signal is synchronized in relation with the supply voltage for occurs due to the presence of these integral loops, and is
each phase. However, this procedure is easy to implement, but dependent on the frequency response of the special feed
on the other side it has some limitations that is it significantly forward and feedback integrators [42].
depends on the voltage signal harmonics.
(iii) Constant Active Power Technique
In this technique the systems active power is kept under control 6.5.4. Frequency domain technique
to maintain instantaneous power constant, however keeping These techniques are appropriate for both systems; single- and
zero reactive power. Meanwhile, it calculates average and three phase systems. These techniques are the consequent from
instantaneous powers of the load. This methodology performs the conventional Fourier analysis and comprise the following
legitimately well under normal conditions. However, if the three subdivisions.
supply is contaminated its performance is deteriorating [55].
(iv) Constant Power Factor Technique (i) Conventional Fourier and FFT Technique,
This technique is very similar to that constant active power (ii) Sine-multiplication Technique, and
method, except to the fact that it compels current signal to (iii) Modified-Fourier-series Technique.
take after the voltage-reference waveform. This infers that
power factor would be fixed to unity and it would work (i) Conventional Fourier and FFT Technique
suitably under associated system of VAR and current- Using fast Fourier transforms, is time consuming process. To
harmonic compensation [55,67]. eliminate fundamental components from the transformed
(v) Fictitious Power Compensation Technique harmonic current signal the inverse transform is then applied
This principle of fictitious power compensation technique is to obtain a time-domain signal [54,56]. This technique is best
developed in [43–45]. This technique is proven to be operating suited for loads with slowly varying state, as this method owe
satisfactorily, despite of many oppositions to the concept by generate Fourier coefficient for that it needs to take sample of
[92,93]. The controller of this system is designed as such to one complete cycle (or an integral number of cycles).
decrease the undesired constituent of power. This technique (ii) Sine Multiplication Technique
can be applied with both single- and three-phase systems, but In this process, the method for multiplication of current signal
the large amount of computation will be required. with sine wave of the fundamental frequency to obtain
(vi) Synchronous Reference Frame Technique assimilated outcome. In this process using only a low pass
The Park transformations is the basis of this technique for filter, can reduces in all the high-order harmonics [47,53]. The
transforming the three phase stationary reference frame operation of this methodology is similar to FFT technique but
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 433

differently executed. Similar to FFT its processing time is also

slow (take more than one complete mains cycle).
(iii) Modified Fourier series Technique
One of the technique addressed in [68]. The norm behind is
that to separate harmonic signal from measured load current
only the fundamental component of current being calculated
and used. The utilization of this technique relies on upon
adjustment in equations of main Fourier series in order to
produce a recursive formula with a sliding window. In the
procedure of this method, computed sine and cosine compo-
nents obtained from every cycle are stored by using two
different circular arrays. The sum of cosine and sine are being
continuously updating after every newly computed value of
the chosen coefficient. The processing time is much lower
than that of other techniques employed for the applications of
single phase systems. Nevertheless, this technique is fairly
suitable for both systems; single and three-phase.
Alternative improved Fourier-series technique was initiated in
[95]. It decomposes the three phase signals into direct (d) and
Fig. 16. Components of shunt APF.
quadrature (q) axes. All the non-active components of load
current have been compensated in this method. The dq-
reference frame generated from non-active current definition
is used to supply the desired supply currents. This method also
uses Sliding-window computation techniques to estimate the
filter current reference value. This technique is appropriate
only for system of three-phase arrangements. The time
domain filtering has advantage over frequency domain that
the measured data can be filtered without waiting for the
whole data [96].

7. Application of active filter Fig. 17. 3-Arm, IGBT-diode bridge.

As discussed, proposed grid connected wind turbine system is devices for smoothing, and the DC link capacitor intend to main-
feeding a nonlinear load has been chosen to study the perfor- tain and regulate the DC voltage. Universal bridge with 3-arm
mance of the APF system under fixed and variable speed of wind. IGBT-Diode Bridge displayed in Fig. 17, from Sim Power System
It has been experimented that due to the variant nature of non-
toolbox of MATLAB/SIMULINK has been selected for this purpose.
linear loads the THD level of current and terminal voltage fall
underneath the limits specified in IEEE-519 standard and in actual
AF system is utilized to inject a current equal in magnitude but in 7.2. Reference signal extraction technique
phase opposition to harmonic current to obtain a purely sinusoidal
current wave in phase with the supply voltage. Fig. 15 displays the As perceived from literature review, the reference signal
schematic diagram of a grid connected WTG with APF system. The extraction technique is the crucial component that confirms the
operation of shunt APF depend on three particulars; the inverter accurate operation of APF. The extraction process is began through
(power unit), reference signal extraction technique and gate driver the detection of obligatory current information of system vari-
signal generator as illustrated in Fig. 16. ables. Customary variables to be extracted are load current, AC
source current, compensation current and DC-link current of the
7.1. Inverter (power unit) APF. Depending on the feedback of system variables, reference
signals estimation in terms of current levels are estimated in
The inverter comprises of six numbers of IGBTs semiconductor frequency-domain or time-domain. Reference extraction techni-
switching device in PWM VSI configuration mode. The ripples que cannot be considered as control system, this is a self-
produced by filter, inductor connected in series with the switching determining task which provide essential reference signal to the
controller for further processing.

7.2.1. Synchronous reference frame method

In SRF technique only three phase current being converted
from stationary rotating frame to synchronous rotating frame
(direct, quadrature and zero sequence, dq0) components [81,97].
For conversion from stationary to synchronous frame Park0 s
transformation have been used. As shown in Fig. 18, idea about
conversion and extraction of reference signal using SRF technique
is given.
Active and reactive components of load voltage and current are
decomposed to direct and quadrature components respectively
Fig. 15. Schematic diagram of APF connected system. [22]. The decomposition from a-b-c to d-q frame uses Park's
434 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

transformation under Eq. (2) as, Eqs. (5)–(7).

2 3
" # "    # ia isa ¼ id sin ðωtÞ þ cos ðωtÞ ð5Þ
id sin θ sin θ  32π
sin θ þ 2π
3 6i 7
¼     4 b5 ð2Þ
iq cos θ cos θ  2π cos θ þ 2π   

3 3
ic 2π 2π
isb ¼ id sin ωt  þ cos ωt  ð6Þ
3 3
where
   
id and iq ¼ d and q axis current 2π 2π
isc ¼ id sin ωt þ þ cos ωt þ ð7Þ
ia, ib and ic ¼ Three phase load current 3 3
The d and q components of current contain active and reactive
power components, as in Eqs. (3) and (4)). where,
i*sa, i*sb and i*sc ¼Extracted reference current signal.
id ¼ ιd þ ι~d ð3Þ This theorem is appropriate only to three-phase system. The
SRF technique is quite stable as its controller only deals with DC
iq ¼ ιq þ ι~q ð4Þ
quantities. Besides, the instantaneous computation of quantities
ιd and ~ιq ¼ fundamental active and reactive current this technique owe time delays in filtering the DC quantities [101–
components 103]. The extracted reference signal is further transferred to gate
ι~d and ι~q ¼ harmonics active and reactive current components driver controller which will then generate appropriate gate pulses
For harmonic compensation strategy [97–99], the source only for inverter to inject desired compensation current in the system.
supply the fundamental component of d-axis load current and to
filter dc component of d-axis load current low pass filter have been 7.3. Gate driver signal generator
used. The function of PI controller to eliminate steady state error of
DC component of inverter and maintain constant dc voltage across As comprehend in the previous chapter, the objective of gate
capacitor. The dc capacitor voltage is sensed and compared with driver controller is to generate suitable gating signals for switching
reference voltage to calculate error voltage. Furthermore, the APF based on the extracted reference signals. The performance of
output of PI controller is subtracted from d-axis current to eradi- an APF is merely depend on the selection of gate driver controller
cate steady state error and to regulate capacitance in dynamic techniques. Thereupon, the optimal choice and strategic imple-
conditions. mentation of the gate driver technique is very important for the
The transformed d-q output signals depend on the load cur- achievement of a satisfactory APF performance.
rents and the phase locked loop (PLL) performance [100], as sinθ Hysteresis band current controller (HBCC) has been selected for
and cosθ is provided by PLL circuit for synchronization. The rota- driving gate pulses for APF in this research study. It imposes a
tion speed of PLL circuit of the rotating reference frame ωt set as bang–bang type instantaneous control that forces the APF com-
fundamental frequency component. Where ω is power supply pensation current (if) signal to come with its estimated reference
frequency. The technique is then established to extract reference signal (if,ref) within a certain tolerance band. This type of control
signal in d-q rotating frame which is converted back to a-b-c system is on view in a block diagram in Fig. 19 [22]. Following this
stationery frame using inverse Park's transformation as in control system, a signal deviation (H) is designed and imposed on
if,ref to form the upper and lower edges of a hysteresis band. The if

Fig. 20. Gate driver signal generated by hysteresis control.

Fig. 18. Block diagram for SRF technique.

Fig. 19. Hysteresis current control method.

 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 435

A
T1 TL T2

A A A aA A a aA A A A a aA

N B B B bB B b bB B B B b bB B

C C C cC C c cC C C C c cC
120 kV 2500 MVA B120 B25 30 km line B575
120 kV/25 kV 25 kV/ 575 V
X0/X1=3 (120 kV) (25 kV) (575 V) C
47 MVA 2.5 MVA

A Conn1
PMSG Wind Turbine
N B Conn2
C
Conn3
3.3ohms
Grounding
Transformer Load

Fig. 21. Simulink model of proposed system.

is then measured and compared with if,ref; the resulting error is

subjected to a hysteresis controller to determine the gating signals Table 7
Parameters of PMSG generator.
when exceeds the upper or lower limits set by estimated reference
signal þH/2 or  H/2. Rated power 1.5 MV A
Providing that, when error is within the hysteresis band, no
switching action will be taken. Switching occurs whenever the Stator voltage 575 V
DC link voltage 1150 V
error hits the upper or lower limit hysteresis band. The APF is
No. of poles 48
therefore switched in such a way that the peak-to-peak compen- Rotor speed 15–20 rpm
sation current/voltage signal is limited to a specified band deter-
mined by H as illustrated by Fig. 20.
Table 8
The remarkable dynamic performance and controllability of
Parameters of grid.
current ripple within a specified band are the foremost benefits
of using the hysteresis current controller. Additionally, the Element Parameter
implementation of this control scheme is simple and robust.
Power grid 120 kV
However, this control scheme exhibits some disappointing fea-
Transformer (T1) 120 kV/25 kV, 47 MV A
tures. The main drawback is producing uneven switching fre- Transformer (T2) 25 kV/575 V, 2.5 MV A
quency. Furthermore, there is probably generation of unwanted Transmission Line (TL) R1 ¼ 0.153 Ω
resonances on the power distribution system [104,105]. Besides, R0 ¼ 0.43 Ω
L1 ¼ 1.04 mH
the uneven switching also affects the APF efficiency and relia-
L0 ¼ 3.3 mH
bility [106]. C1 ¼11.3e  009 F
C0 ¼5.0e  009 F
Length¼ 30 km
8. Proposed system and its parameters
Table 9
The Simulink model of proposed system is given in Fig. 21 and Load rating.
the rating of PMSG wind turbine generator, power grid and non-
linear load are given in Tables 7–9 respectively. Type Combination Rating

3-Phase Rectifier Series R and L R ¼30 Ω

L ¼ 0.1 H
9. Simulation results at fixed speed WTG without active filter Single Phase Rectifier (01) Series R and L R ¼1e50 Ω
L ¼ 0.01 e  3 H
Single Phase Rectifier (02) Parallel R and L R ¼50 Ω
To check the effect of harmonics in the form of total harmonic
L ¼ 0.01 e  3H
distortion on the proposed system, initially system is operated at Single Phase Rectifier (03) R only 100 Ω
fixed wind speed (8 m/s) without active filter. The simulation
results shown increased THD level of current at terminals of wind
turbine generator and at PCC. Table 10 showing the THD level of Table 10
WTG and PCC current with voltage at PCC. THD Levels without active filter and fixed speed.

Phase WTG current (%) PCC current (%) PCC voltage (%)

10. Simulation results at variable speed WTG without active A 38.25 92.18 9.63
filter B 53.45 71.38 9.85
C 55.87 65.38 6.90

Further, the proposed system operated at variable speed of

wind (8–12 m/s using random block in MATLAB/SIMULINK)
without active filter. The simulation results have shown decrease standard specified in IEEE-519. Table 11 showing the simulation
in THD level of WTG and PCC current as compared to fixed speed results under variable speed operation of wind without active
operation but still the system is not operating under the harmonic filter.
436 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

Table 11 References
THD levels without active filter and variable speed.

[1] Massoud AM, Ahmed S, Abdel‐Khalik AS. Active power filter. Power elec-
Phase WTG current (%) PCC current (%) PCC voltage (%)
tronics for renewable energy systems, transportation and industrial appli-
cations; 2014. p. 534–72.
A 35.95 62.85 8.30 [2] Tsai J, Tan K. H APF harmonic mitigation technique for PMSG wind energy
B 33.45 53.36 7.39 conversion system. In: Power engineering conference. AUPEC 2007. Aus-
C 44.06 32.86 6.79 tralasian Universities; 2007. p. 1–6.
[3] Prasad A, Ziogas PD, Manias S. An active power factor correction technique
for three-phase diode rectifiers. IEEE Trans Power Electron 1991;6:83–92.
Table 12 [4] dos Reis FS, Ale J, Adegas F, Tonkoski R, Slan S, and Tan K. Active shunt filter
THD levels with active filter and variable speed. for harmonic mitigation in wind turbines generators. In: Power electronics
specialists conference. PESC'06. 37th IEEE; 2006. p. 1–6.
[5] Gaillard A, Poure P, Saadate S, Machmoum M. Variable speed DFIG wind
Phase WTG current (%) PCC current (%) PCC voltage (%)
energy system for power generation and harmonic current mitigation.
Renew Energy 2009;34:1545–53.
A 4.39 1.06 0.95 [6] Hoseinpour A, Masoud Barakati S, Ghazi R. Harmonic reduction in wind
B 2.92 1.08 1.16 turbine generators using a Shunt Active Filter based on the proposed mod-
C 6.96 1.71 0.84 ulation technique. Int J Electr Power Energy Syst 2012;43:1401–12.
[7] Slootweg J and Kling W, Is the Answer Blowing in the Wind? IEEE power &
energy magazine, November, ed: December; 2003.
Table 13 [8] Shafiullah G, Oo AMT, Shawkat Ali A, Wolfs P. Potential challenges of inte-
THD levels with active filter and variable speed. grating large-scale wind energy into the power grid – a review. Renew
Sustain Energy Rev 2013;20:306–21.
Phase WTG current (%) PCC current (%) PCC voltage (%) [9] Mohd Zin AAB, Pesaran HA M, Khairuddin AB, Jahanshaloo L, Shariati O. An
overview on doubly fed induction generators0 controls and contributions to
wind based electricity generation. Renew Sustain Energy Rev 2013;27:692–708.
A 1.11 0.99 0.91
[10] Lang Y, Zargari N, Kouro S. Power conversion and control of wind energy
B 2.59 1.59 1.07 systems, vol. 74. USA: John Wiley & Sons; 2011.
C 1.15 1.86 0.85 [11] Bansal R, Bhatti T, Kothari D. Bibliography on the application of induction
generators in nonconventional energy systems. IEEE Trans Energy Convers
2003;18:433–9.
11. Simulation result at fixed speed WTG with active filter [12] Shadhu Khan, P JCK. Three-phase induction generators: a discussion on
performance. Electr Mach Power Syst 1999;27:813–32.
[13] Bansal R. Three-phase self-excited induction generators: an overview. IEEE
Moreover, the established system is now operated with active Trans Energy Convers 2005;20:292–9.
filter and fixed wind speed. Simulation results shown in Table 12 [14] Polinder H, Van der Pijl FF, De Vilder G-J, Tavner PJ. Comparison of direct-
drive and geared generator concepts for wind turbines. IEEE Trans Energy
giving a substantial decrement in the THD level of WTG and PCC Convers 2006;21:725–33.
current along with voltage at PCC. However, the effect of harmo- [15] Wu Z, Dou X, Chu J, Hu M. Operation and control of a direct-driven PMSG-
nics on WTG current is still above the set limit under IEEE-519 as based wind turbine system with an auxiliary parallel grid-side converter.
Energies 2013;6:3405–21.
mentioned in Table 4. To get the better results system is again [16] Nouh A and Mohamed F. Wind energy conversion systems: classifications
operated with active filter but at variable speed of wind. and trends in application. In: 5th international renewable energy congress
(IREC); 2014. p. 1–6.
[17] Zhang Y and Ula S, Comparison and evaluation of three main types of wind
turbines. In: Transmission and distribution conference and exposition. T&
x00026; D. IEEE/PES; 2008. p. 1–6.
12. Simulation result at variable speed WTG with active filter [18] Rolan A, Luna A, Vazquez G, Aguilar D, and G. Azevedo. Modeling of a vari-
able speed wind turbine with a permanent magnet synchronous generator.
In IEEE international symposium on industrial electronics. ISIE 2009; 2009.
The simulation results from the system when operated at p. 734–39.
variable wind speed with active filter are shown in Table 13. The [19] Shewale AJ, Gagangras AR, and Lokhande NM. Comparison of various wind
proposed system under this condition satisfying the IEEE-519 turbine generators.
[20] Norman R. Horizontal and vertical axis wind turbines. In Von Karman Inst.
standard and reduced the effect of harmonics from the system. for Fluid Dyn. Wind Energy Conversion Devices 20 p (SEE N83-12526 03-
44); 1981.
[21] Jay F and Goetz J. IEEE standard dictionary of electrical and electronics terms;
1988.
[22] Salam Z, Tan PC, Jusoh A. Harmonics mitigation using active power filter: a
13. Conclusion
technological review. Elektrika 2006;8:17–26.
[23] Ellis RG, Eng P. Power system harmonics – a reference guide to causes, effects
Review and application of active filter with grid connected and corrective measures. An Allen-Brandley Series of Issues and Answers-
Rockwell Automation; 2001. p. 3.
wind turbine generator for mitigating load harmonics from WTG
[24] Oğuz Y, Güney İ, Çalık H. Power quality control and design of power con-
and PCC have been carried out in this paper. For extraction of verter for variable-speed wind energy conversion system with permanent-
harmonic reference signal, synchronous reference frame technique magnet synchronous generator. Sci World J 2013;2013:1–14.
[25] Phan V-T, Lee H-H. Control strategy for harmonic elimination in stand-alone
has been utilized and for generating active filter driving pulses
DFIG applications with nonlinear loads. IEEE Trans Power Electron
hysteresis band current control method has been used in this 2011;26:2662–75.
research. The performance of active filter with grid connected [26] I. F II. IEEE recommended practices and requirements for harmonic control in
electrical power systems; 1993.
variable speed WTG is better as compared to fixed speed wind [27] Das S, Gupta S, and Bhargava S. Harmonic reduction methods for wind
turbine. Additionally, simulation results also showing that active energy conversion systems: a review. In 2014 IEEE students' conference on
filter with variable speed WTG satisfying harmonic limits specified electrical, electronics and computer science (SCEECS); 2014. p. 1–8.
[28] Rahmani S, Hamadi A, Al-Haddad K. A new combination of shunt hybrid
in IEEE-519 standard. With addition to current harmonic mitiga- power filter and thyristor controlled reactor for harmonics and reactive
tion proposed configuration also decreases the voltage THD level. power compensation. Electrical power energy conference (EPEC). IEEE; 2009.
p. 1–6.
By eliminating the effect of harmonics from WTG and PCC, supply [29] Liu Q, Peng L, Kang Y, Tang S, Wu D, Qi Y. A novel design and optimization
voltage waveform will be sinusoidal in nature it will simulta- method of an filter for a shunt active power filter. IEEE Trans Ind Electron
neously increase the performance of generator and reduces the 2014;61:4000–10.
[30] Qasim M, Kanjiya P, Khadkikar V. Artificial-neural-network-based phase-
losses caused by these harmonics from the power system and locking scheme for active power filters. IEEE Trans Industrial Electron
customers devices. 2014;61:3857–66.
 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438 437

[31] Kawahira H, Nakamura T, Nakazawa S, Nomura M. Active power filters. [64] Scheidecker D and Tennakoon S. Optimisation of the number of levels in a
Proceedings of the 1983 international power electronics conference; 1983. p. multi-level advanced static VAR compensator. In: European conference on
981–92. power electronics and applications; 1995. p. 2.493–98.
[32] Akagi H, Kanazawa Y, Nabae A. Instantaneous reactive power compensators [65] Peng FZ and Lai JS. Dynamic performance and control of a static var gen-
comprising switching devices without energy storage components. IEEE erator using cascade multilevel inverters. In: Industry applications con-
Trans Ind Appl 1984:625–30. ference record of the thirty-first IAS annual meeting, IAS'96, IEEE; 1996. p.
[33] Akagi H, Nabae A, Atoh S. Control strategy of active power filters using 1009–15.
multiple voltage-source PWM converters. IEEE Trans Ind Appl 1986:460–5. [66] Inarida S, Miyake W, Mizuguchi N, Hariyama T, and Hoshi K. Development of
[34] Akagi H. New trends in active filters for power conditioning. IEEE Trans Ind three-level power converter system using IGBTs for Shinkansen trains. In
Appl 1996;32:1312–22. European conference on power electronics and applications; 1997. p. 1.216–
[35] Xiao G, Liu J, and Yang J. The experimental studies on active power filter for 20.
HVDC system. In Proceedings of the third international power electronics [67] Jou HL. Performance comparison of the three-phase active-power-filter
and motion control conference, 2000. IPEMC; 2000, p. 1376–79. algorithms. IEE Proceedings-generation, transmission and distribution, vol.
[36] Verdelho P and Marques G. Multiple applications control system for the 142; 1995. p. 646–52.
PWM voltage converter connected to the AC mains. In: Fifth European [68] Akagi H. New trends in active filters. In European conference on power
conference on power electronics and applications; 1993. p. 42–6. electronics and applications; 1995. p. 0.017–26.
[37] Pereira M, Wild G, Huang H, and Sadek K. Active filters in HVDC systems: [69] Dixon JW, Venegas G, Moran LA. A series active power filter based on a
actual concepts and application experience. In: Proceedings of the interna- sinusoidal current-controlled voltage-source inverter. IEEE Trans Ind Elec-
tional conference on power system technology. PowerCon; 2002. p. 989–93. tron 1997;44:612–20.
[38] Gole A, Meisingset M. An AC active filter for use at capacitor commutated [70] Peng FZ, Akagi H, Nabae A. Compensation characteristics of the combined
HVDC converters. IEEE Trans Power Deliv 2001;16:335–41. system of shunt passive and series active filters. IEEE Trans Ind Appl
[39] Singh B, Al-Haddad K, Chandra A. A review of active filters for power quality 1993;29:144–52.
improvement. IEEE Trans Ind Electron 1999;46:960–71. [71] Fujita H, Akagi H, Kohata M. Combined rectifying system including double‐
[40] Qasim M, Khadkikar V. Application of artificial neural networks for shunt series capacitor‐smoothed diode rectifier and series active filter. Electr Eng
active power filter control. IEEE Trans Ind Inf 2014;10:1765–74. Jpn 1997;120:85–94.
[41] Y. Fang, S. Hou, and J. Fei, Harmonic Suppression of three-phase active power [72] Ooi B-T, Dai S-Z. Series-type solid-state static VAR compensator. IEEE Trans
filter using backstopping approach. Power Electron 1993;8:164–9.
[42] El-Habrouk M, Darwish M, Mehta P. Active power filters: a review. IEE [73] Harley R, Rigby B, Jennings G. Design of a controlled converter which emu-
Proceedings electric power applications vol. 147; 2000. p. 403–13. lates a series capacitive compensator for long power lines. Proc. int. conf.
[43] Enslin JH, Van Wyk JD. Measurement and compensation of fictitious power power electronics and motion control (PEMC’94); 1994. p. 213–8.
under nonsinusoidal voltage and current conditions. IEEE Trans Instrum [74] Joos G, Moran L, Ziogas P. Performance analysis of a PWM inverter VAR
Meas 1988;37:403–8. compensator. IEEE Trans Power Electron 1991;6:380–91.
[44] Enslin JH and Van Wyk JD. A new control philosophy for power electronic [75] Shirahama H, Sakurai Y, Matsuda Y, Ishigaki Y, Murai K. Instantaneous
converters as fictitious power compensators. In: 19th annual IEEE power control method with a GTO converter for active and reactive powers in
electronics specialists conference. PESC'88 Record. 1988, p. 1188–96. superconducting magnetic energy storage. IEEE Trans Power Electron
[45] van Harmelen GL, Enslin JH. Real-time dynamic control of dynamic power 1994;9:1–6.
filters in supplies with high contamination. IEEE Trans Power Electron [76] Rastogi M, Mohan N, Edris A. Hybrid-active filtering of harmonic currents in
1993;8:301–8. power systems. IEEE Trans Power Deliv 1995;10:1994–2000.
[46] Welsh M, Mehta P, Darwish M. Genetic algorithm and extended analysis [77] Bhavaraju V, Enjeti PN. Analysis and design of an active power filter for
optimisation techniques for switched capacitor active filters-comparative balancing unbalanced loads. IEEE Trans Power Electron 1993;8:640–7.
study. IEE Proceedings electric power applications; 2000. p. 21–6. [78] Verdelho P, Marques G. An active power filter and unbalanced current
[47] Hsu CY, Wu HY. A new single-phase active power filter with reduced energy- compensator. IEEE Trans Ind Electron 1997;44:321–8.
storage capacity. IEE Proceedings electric power applications vol. 143; 1996. [79] Williams SM, Hoft RG. Adaptive frequency domain control of PWM switched
p. 25–30. power line conditioner. IEEE Trans Power Electronics 1991;6:665–70.
[48] Torrey DA, Al-Zamel AM. Single-phase active power filters for multiple [80] Saetieo S, Devaraj R, Torrey DA. The design and implementation of a three-
nonlinear loads. IEEE Trans Power Electron 1995;10:263–72. phase active power filter based on sliding mode control. IEEE Trans Ind Appl
[49] Nastran J, Cajhen R, Seliger M, Jereb P. Active power filter for nonlinear AC 1995;31:993–1000.
loads. IEEE Trans Power Electron 1994;9:92–6. [81] Jeong S-G, Woo M-H. DSP-based active power filter with predictive current
[50] Bhavaraju V, Enjeti P. A fast active power filter to correct line voltage sags. control. IEEE Trans Ind Electron 1997;44:329–36.
IEEE Trans Ind Electron 1994;41:333–8. [82] Xu J, Yang J, Shen A, Chen J. A novel ANN-based harmonic extraction method
[51] Koozehkanani Z, Mehta P, and Darwish M. Active symmetrical lattice filter tested with ESN, RNN and MLP in shunt active power filters. Int J Wirel Mob
for harmonic current reduction. In European conference on power electro- Comput 2014;7:123–31.
nics and applications; 1995. p. 1.869–73. [83] Wu TF, Shen CL, Chiu JY, and Chen CC. An APF with MAPPT scheme to improve
[52] Koozehkanani ZD, Mehta P, Darwish M. An active filter for retrofit applica- power quality. In TENCON 2001. Proceedings of IEEE region 10 international
tions; 1996. conference on electrical and electronic technology; 2001. p. 620–6.
[53] Koczara W, Greetham S, Haydock L. An investigation of the reduction of [84] Buso S, Malesani L, Mattavelli P. Comparison of current control techniques
harmonics andreactive power in a single phase system using an adaptive for active filter applications. IEEE Trans Ind Electron 1998;45:722–9.
filter. In: Proceedings of power electronics and motion control conference, [85] Jintakosonwit P, Fujita H, Akagi H. Control and performance of a fully-digital-
PEMC94; 1994. p. 196–201. controlled shunt active filter for installation on a power distribution system.
[54] Grady WM, Samotyj MJ, Noyola AH. Survey of active power line conditioning IEEE Trans Power Electron 2002;17:132–40.
methodologies. IEEE Trans Power Deliv 1990;5:1536–42. [86] Khositkasame S and Sangwongwanich S. Design of harmonic current
[55] Cavallini A, Montanari GC. Compensation strategies for shunt active-filter detector and stability analysis of a hybrid parallel active filter. In: Proceed-
control. IEEE Trans Power Electron 1994;9:587–93. ings of the power conversion conference. Nagaoka; 1997. p. 181–6.
[56] Viriya P, Matsuse K. Low-harmonic GTO converter for fundamental power [87] Routimo M, Salo M, and Tuusa H. A novel control method for wideband
factor compensation. IEEE Trans Power Electron 1991;6:371–9. harmonic compensation. In: The fifth international conference on power
[57] Ledwich G, Doulai P. Multiple converter performance and active filtering. electronics and drive systems. PEDS; 2003. p. 799–804.
IEEE Trans Power Electron 1995;10:273–9. [88] Li R, Johns A, Elkateb M, and Robinson F. Comparative study of parallel
[58] Xu W, Mansour Y, Siggers C, Hughes M. Developing utility harmonic reg- hybrid filters in resonance damping. In International conference on electric
ulations based on IEEE STD 519-BC, Hydro's approach. IEEE Trans Power power engineering. PowerTech Budapest 99; 1999. p. 230.
Deliv 1995;10:1423–31. [89] Bhattacharya S and Divan D. Synchronous frame based controller imple-
[59] Moran LA, Dixon JW, Wallace RR. A three-phase active power filter operating mentation for a hybrid series active filter system. In: Conference record of
with fixed switching frequency for reactive power and current harmonic the industry applications conference thirtieth IAS annual meeting, IAS'95.
compensation. IEEE Trans Ind Electron 1995;42:402–8. IEEE; 1995. p. 2531–40.
[60] Peng FZ, Lai J-S, McKeever JW, VanCoevering J. A multilevel voltage-source [90] Peng FZ, Akagi H, Nabae A. A new approach to harmonic compensation in
inverter with separate DC sources for static var generation. IEEE Trans Ind power systems-a combined system of shunt passive and series active filters.
Appl 1996;32:1130–8. IEEE Trans Ind Appl 1990;26:983–90.
[61] Fukuda S, Endoh T. Control method for a combined active filter system [91] Cho GC, Jung GH, Choi NS, Cho GH. Analysis and controller design of static
employing a current source converter and a high pass filter. IEEE Trans Ind var compensator using three-level GTO inverter. IEEE Trans Power Electron
Appl 1995;31:590–7. 1996;11:57–65.
[62] Fujita H, Tominaga S, Akagi H. Analysis and design of a dc voltage-controlled [92] Czarnecki LS. Comments on Measurement and compensation of fictitious
static var compensator using quad-series voltage-source inverters. IEEE power under nonsinusoidal voltage and current conditions by JHR Enslin and
Trans Ind Appl 1996;32:970–8. JD Van Wyk. IEEE Trans Instrum Meas 1989;38:839–41.
[63] Joos G, Huang X, and Ooi BT. Direct-coupled multilevel cascaded series VAR [93] Czarnecki L. Comments, with reply, on A new control philosophy for power
compensators. In: Industry applications conference record of the thirty- electronic converters as fictitious power compensators by JHR Enslin and JD
second IAS annual meeting, IAS'97, IEEE; 1997. p. 1608–15. Van Wyk. IEEE Trans Power Electron 1990;5:503–4.
438 S.H. Qazi, M.W. Mustafa / Renewable and Sustainable Energy Reviews 57 (2016) 420–438

[94] Bhattacharya S, Veltman A, Divan DM, Lorenz RD. Flux-based active filter [101] Shi P, Zhang Y, Chadli M, Agarwal RK. Mixed H-infinity and passive filtering
controller. IEEE Trans Ind Appl 1996;32:491–502. for discrete fuzzy neural networks with stochastic jumps and time delays;
[95] Blajszczak G. Non‐active power compensation using time‐window method. 2015.
Eur Trans Electr Power 1992;2:285–90. [102] Shi P. Filtering on sampled-data systems with parametric uncertainty. IEEE
[96] Al-Odienat AI, Al-Mbaideen AA. Optimal length determination of the moving Trans Autom Control 1998;43:1022–7.
average filter for power system applications. Biomedicine 2015;10. [103] Yin Y, Shi P, Liu F, Teo KL. Filtering for discrete-time nonhomogeneous
[97] Soares V, Verdelho P, Marques GD. An instantaneous active and reactive Markov jump systems with uncertainties. Inf Sci 2014;259:118–27.
current component method for active filters. IEEE Trans Power Electron [104] Chen D and Xie S. Review of the control strategies applied to active power
2000;15:660–9. filters. In: Proceedings of the 2004 IEEE International conference on electric
[98] Mikkili S, Panda A. Real-time implementation of PI and fuzzy logic con- utility deregulation, restructuring and power technologies (DRPT 2004);
trollers based shunt active filter control strategies for power quality 2004. p. 666–70.
improvement. Int J Electr Power Energy Syst 2012;43:1114–26. [105] Malesani L, Mattavelli P, and Buso S. Dead-beat current control for active
[99] Mikkili S and Panda A. Type-1 and Type-2 Fuzzy logic controller based Shunt filters. In Proceedings of the 24th annual conference of the IEEE industrial
active filter I d-I q control strategy for mitigation of harmonics with Trian- electronics society, 1998. IECON'98; 1998. p. 1859–64.
gular membership function. In: 2012 IEEE international conference on power [106] El-Habrouk M, Darwish M. Design and implementation of a modified Fourier
electronics, drives and energy systems (PEDES); 2012. p. 1–6. analysis harmonic current computation technique for power active filters
[100] Sharma A, Upadhyay A. Harmonic mitigation using inverter based hybrid using DSPs. In: IEE Proceedings electric power applications; 2001. p. 21–8.
shunt active power filter. Int J Electron Electr Eng 2014;7:787–98.