LITERATURE REVIEW 2.1 Introduction Power quality issues and remedies are relevant research topics and a lot of advanced researches are being carried out in this area. These issues are mainly due to increased use of power electronic devices, nonlinear loads and unbalance in power systems. Dynamic loads cause power quality problems usually by voltage or current variations such as voltage dips, fluctuations, momentary interruptions, oscillatory transients, harmonics, harmonic resonance etc.. Various publications define power quality in different aspects. According to IEEE Recommended Practice for Monitoring Power Quality (IEEE Std 11591995), Power quality is defined as “concept of powering and grounding sensitive equipment in a manner that is suitable for operation of that equipment.”
2.2 Power quality issues- Definitions Definitions for power quality issues in power systems with non sinusoidal waveforms and unbalanced loads are detailed in [5-9]. The definitions and terminology used in conjunction with power quality are as follows: Voltage quality can be interpreted as the quality of voltage delivered by the utility to the consumers and is concerned with the deviations of the voltage from the ideal one. The ideal voltage is a single frequency sine wave of constant frequency and constant magnitude. Current quality deals with the deviations of the current from the ideal one which should be sinusoidal wave current of constant frequency and required magnitude and should also be in phase with the supply voltage. Voltage quality deals with what the utility delivers to the customer and current quality deals with what the customers take from the utility and are mutually dependent. Power quality is the combination of voltage quality and current quality. Power quality is concerned with deviations of voltage and/or current from the ideal.
Voltage magnitude variation is the increase or decrease in voltage magnitude due to load variations, transformer tap–changing, switching of capacitor banks or reactors etc. Voltage frequency variation is the variation in frequency of supply voltage due to the imbalance between load and generation units. Current magnitude variation is the variation of the load current magnitude which also results in voltage magnitude variations. Current phase variation – Ideally, the voltage and current waveforms should be in phase so that the power factor perceived by the source is unity. Deviation from this situation is termed as current phase variation. Voltage and current imbalances – Voltage imbalance in three phase systems where the rms values of the voltages in each phase or the phase angle differences between consecutive phases are not equal, can affect the ratio of negative sequence and positive sequence voltage components. This can result in large differences between the highest and lowest values of voltage magnitude and phase difference. The voltage imbalance leads to large load current imbalances. Voltage fluctuation –The fast variation in voltage magnitude is called voltage fluctuation or ‘voltage flicker’ and can affect the performance of the equipment. Harmonic voltage distortion – The ideal voltage waveform is a sinusoidal wave of constant frequency. But, when there is voltage distortion, it may be a sum of sine waves with frequencies which are multiples of fundamental frequency. These non-fundamental components contribute to harmonic distortion. The harmonic current components result in harmonic voltage components and hence a non-sinusoidal voltage in the system. Harmonic current distortion – Harmonic current distortion is the complementary phenomenon of harmonic voltage distortion. They are mutually dependent as harmonic voltage distortion is mainly due to non-sinusoidal load currents. Inter-harmonic voltage and current components are generated by equipment such as cyclo-converters, heating controllers and arc furnaces, which generate current components at such frequencies which are not integral multiples of fundamental frequency. In fact, there may be sub-harmonic frequency currents as well. These inter-harmonic components can
cause resonance between the line inductances and capacitor banks. The sub-harmonic currents can lead to saturation of transformers and in turn to damage of synchronous generators and turbines. Voltage notching - In three phase converters during commutation from one device to another, short circuits for short durations can cause voltage reduction or notching. Voltage notching leads to higher order harmonics. Interruptions – Supply interruption is a condition in which the voltage at the supply terminals is close to zero or less than 10% according to IEEE Standard 1159 -1995. Faults or protection equipment mal-tripping can cause interruptions. Under voltages –Short duration under voltages are known as voltage sags and longer duration under voltages are called under voltages. Voltage sag is a reduction in the supply voltage magnitude followed by a voltage recovery after a short period of time. Voltage sags are mostly caused by short circuit faults in the system and by starting of large motors. Over voltages- Over voltages of very short duration and high magnitude are called transient over voltages/voltage spikes/voltage surges. Over voltages with duration between one cycle and one minute are called voltage swells or temporary power frequency over voltages. Longer duration over voltages are called over voltages. Over voltages are caused by lightning strokes, switching operations, sudden load reduction, single phase short circuits and nonlinearities. Electromagnetic compatibility (EMC) – EMC is defined by IEC (International Electrotechnical Commission) as the ability of a device, equipment or system to function satisfactorily in its electromagnetic environment without introducing intolerable electromagnetic disturbances to anything in that environment.
2.3 Sources of power quality issues The increasing cost of energy led to the introduction of efficient adjustable speed drives using static power converters in 1970’s. This brought about a wide change in application of utilisation equipment in industrial power systems. To minimise the electrical energy costs, which are made up of kVA demand and kWh charges, users began to apply capacitors in their system to lower the demand charges. Wide usage of capacitor banks with static power
converters introduced harmonic resonance problems . The causes of these power quality problems are generally complex and difficult to detect. In earlier days, the main sources of waveform distortion were electric arc furnaces, fluorescent lamps, electrical machines and transformers. (i)
Arc-furnace: In Arc furnace, the voltage-current characteristics of electric arcs are highly nonlinear. Following arc ignition, the voltage decreases due to the short-circuit current, which is limited only by the power system impedance.
Fluorescent lamp: In a fluorescent lamp, the voltage builds up in each half cycle till it emits light. Its current is limited by the non-linear reactive ballast and hence distorted.
(iii) Rotating machines: They also generate harmonics because the windings are embedded in slots which are not exactly sinusoidally distributed and mmf becomes distorted. Generally, harmonics produced by rotating machines are considered negligible compared to those produced by other sources. Also, large generators are usually connected to power grid through delta-connected transformers thus blocking the flow of third harmonic current. (iv) Power transformers: They use magnetic materials that are operated often in the nonlinear region for economic purposes resulting in the distorted (mainly third harmonics) transformer magnetizing current even if the applied voltage is sinusoidal. Large numbers of power electronic loads installed in power systems, also generate harmonics. Major sources are identified as Desktop computers, TVs, Fax Machines, Copiers, Microwave ovens, Electric vehicle battery chargers, Thyristor converters, UPS, ASDs, Welding machines, Static var compensators, Inverters, SMPS, Fluorescent lighting etc. The switching or commutation of power semiconductor devices generates voltage or current transients that are characterized by a spectrum of frequencies. Static VAR compensators are balanced three-phase devices that use thyristors to control the conduction time of shunt capacitors or inductors during each half cycle in order to maintain a desired terminal voltage. It generates non sinusoidal currents .
Harmonics is considered to be a major power quality issue. The generation of harmonic currents by typical harmonic loads are illustrated in the following subsection. 2.3.1 Typical harmonic generating loads Of the several harmonic producing loads such as transformers, arc furnaces, welding equipments etc., for the purpose of this work, it is decided to consider four types of widely used loads, both steady state and dynamic, which will introduce non sinusoidal currents in three phase AC network. The major harmonic loads considered in this work are: (a) Three phase diode bridge rectifier (b) Three phase thyristor converter (c) DC motor drive and (d) Induction motor drive 220.127.116.11 Three phase diode bridge rectifier It is the most extensively used rectifier topology and is shown in Figure 2.1.This is used for feeding R-L load, DC Motor or charging battery. The load side inductance is assumed to be large enough to keep continuous load current [ 15, 16 ]. One diode from the upper arm (D , 1
D and D ) and one diode from the lower arm (D , D and D ) conduct always. Thus the 3
converter has six different conduction modes and in each mode one of the diode pairs of D D , D D , D D , D D , D D and D D conducts. Each conduction mode lasts for π/3 1
radians and each diode conducts for 2π/3 radians in each cycle.
Figure 2.1 Three phase diode bridge rectifier.
Figure 2.2 shows source voltage and source current waveforms for the circuit in Figure 2.1, when E = 0, where E = back emf [ 15, 16].The source current is highly non-sinusoidal, but in phase with voltage waveform.
Figure 2.2 Source voltage and Source current waveforms in phase a - with three phase diode bridge rectifier feeding R – L load 18.104.22.168 Three phase thyristor converter Three phase controlled converter circuits are classified into various categories such as: 3 pulse, 6 pulse, 12 pulse, etc. In this work, six pulse converters are used as the non-linear load. The circuit of three phase fully controlled converter with R – L load is shown in Figure 2.3. One of the thyristors in upper arm, triggered at highest instantaneous positive value and one thyristor in the lower arm triggered at the lowest instantaneous negative value are conducting, when overlapping is ignored. For six pulse operation, each thyristor is conducting for one-third of a period, 2/3 radians[15, 16].
Figure 2.3 Three phase thyristor converter 11
The average output DC voltage is Edc=
3 3 E m cos , where E m is the peak value of line to
neutral voltage and is the triggering angle of thyristor. As the firing angle α changes, from 0º to 90º, the voltage also changes from maximum to zero and converter is said to be in rectification mode. For the angles in the range 90º to 180º, the voltage varies from 0 to negative maximum and the converter is in the inversion mode. The source voltage and source current waveforms with three phase thyristor converter load at firing angle α of 60º are shown in Figure 2.4. In this case also source current is non-sinusoidal and displaced from the voltage waveform.
Figure 2.4 Source voltage and Source current waveforms in phase a - with three phase thyristor converter (α = 60º) feeding R-L load 22.214.171.124 DC motor drive DC Motor drives are used in a very wide power range from a few watts to many thousands of kW in applications ranging from very precise, high performance position controlled drives in robotics to variable speed drives for adjusting flow rates in locomotives. The thyristor converter providing variable armature voltage for the drive motor is shown in Figure 2.5[15,16].
Figure 2.5 Three phase thyristor converter circuit for speed control of DC motor
It is a two quadrant drive. The field is separately excited and field current held constant. The thyristors are fired at an interval of 60 and ripple in the motor terminal voltage is 6 pulses/cycle. If the firing angle is delayed more than 90, motor terminal voltage becomes negative and it is the inversion mode of operation of converter. Practically, the motor is operated at a desired speed meeting the load torque which depends on the armature current. The source current waveform of 400V, 5HP DC motor drive under dynamic conditions is given in Figure 2.6.
Figure 2.6 Source current waveform of DC drive under dynamic conditions A speed feedback with an inner current loop is used to provide faster response to any dynamic disturbance in the speed, load torque and supply voltage. The output of the speed controller, which sets the current reference for the current loop, is compared with armature current. The error is processed through a current controller whose output adjusts the firing angle of the converter and brings the motor speed to the desired value. The speed controller also limits the armature current. 126.96.36.199 Induction motor drive Many induction motor applications require multiple speeds or adjustable speed ranges. Variable frequency control method is preferred for speed control of squirrel cage induction motor. In this method, by changing stator voltage frequency, motor speed can be controlled. The variable frequency control below the rated frequency is carried out by varying the machine phase voltage along with frequency such that flux is maintained constant. It allows good running and dynamic performance with squirrel cage induction motor[15, 16]. Due to the availability of power electronic devices with improved ratings and characteristics, low cost and reliable variable frequency drives are developed. The variable frequency supply to an induction motor, for speed control, can be obtained by using voltage source inverter, current source inverter or cycloconverter. The schematic diagram of the PWM based voltage source inverter method for speed control of induction motor is shown in Figure 2.7.
Figure 2.7 Three phase voltage source inverter for speed control of induction motor The choice of the switching frequency with PWM inverter depends on number of factors – switching losses, harmonic content in waveforms, torque pulsation, motor losses etc. High switching frequency within the inverter will increase the inverter switching losses but will reduce the harmonic content of the current waveforms and hence a smoother torque [15, 16]. Harmonic analysis of some typical loads are carried out in laboratory and explained in section 2.3.2. 2.3.2 Study of typical nonlinear loads In this section, a few cases of typical non linear loads mentioned in section 2.3.1 are reviewed in detail to obtain the waveforms of distorted source currents. 188.8.131.52 Rectifiers Rectifiers convert the AC supply into DC voltage source, which is directly connected to loads such as heater coils, furnaces, DC motors, etc., or further converted to AC as in the case of UPS systems, variable frequency AC drives (VFD), switched mode power supplies (SMPSs), induction heating inverters, etc. Basically there are two types of rectifiers called uncontrolled rectifiers and controlled rectifiers . They are discussed in the following paragraphs. 184.108.40.206.1 Three phase diode bridge rectifier Uncontrolled rectifiers are used as front-end converters in SMPS, VFD, DC power supply, and UPS. Generally uncontrolled rectifiers are connected directly to a DC smoothing capacitor. Distorted source voltage (line) and source current waveforms with three phase diode bridge rectifier fed resistive load of 1kW are shown in Figure 2.8. The predominant harmonic component in the current waveform is 5th. The fundamental component of the input current is almost in phase with the respective phase voltage. Thus the apparent input
power factor is close to unity. Voltage distortion occurs in uncontrolled rectifiers due to voltage clamping. When the diodes conduct, the input line voltage gets clamped to the DC voltage across the capacitor. This can also affect other equipment connected to the same supply. These can be overcome if proper snubber, that can absorb the difference in voltages, is used for each rectifier . Radio frequency interference and electromagnetic interference are problems created by rectifiers and other power electronic equipment due to fast switching of voltage and current.
220.127.116.11.2 Single phase half controlled converter Harmonic analysis of controlled rectifier fed 1kW resistive load is carried out using Power Analyzer (Fluke 434B). Source voltage and distorted source current waveforms under varying load conditions are shown in Figure 2.9.
Figure 2.8 Experimental results - Source voltage and current waveforms – 3 diode bridge rectifier feeding resistive load
Figure 2.9 Experimental results - Source voltage and current waveforms – 3 thyristor converter feeding resistive load (a) ( =120º) and (b) ( = 60º)
18.104.22.168 Single phase voltage regulator using TRIAC-DIAC AC phase controllers are used to convert fixed AC to variable AC. Main applications are heating, lighting control, and speed control of 1 and 3 ac drives. For the phase control applications, the triac-diac circuit is used to generate a sharp triggering pulse [15,16]. The circuit block diagram is shown in Figure 2.10. It is simple and compact, but introduces harmonics in supply current. A single phase voltage regulator used for current control of resistive load, lighting load and speed control of fan. Source voltage and distorted source current waveforms under various load conditions are shown in Figure 2.11.
Figure 2.10 Block diagram of single phase AC voltage regulator
Figure 2.11 Experimental results - Source voltage and current waveforms – AC Voltage regulator for (a) speed control of fan (b) resistive load (c) incandescent lamp 22.214.171.124 Lighting loads One of the best methods of energy saving in lighting sector is the implementation of energy efficient lamps. The fluorescent tube lights (FTL) and compact fluorescent lamps (CFL) are well known for their energy efficacy as compared to incandescent bulbs. But the wide spread usage of CFL will cause harmonic pollution and more reactive power demand (power factor being low), if not properly designed and selected . Source voltage and distorted source current waveforms with 5W CFL are shown in Figure 2.12. Fifth harmonic is the predominant harmonic and magnitude of harmonics decreases as order increases.
Figure 2.12 Experimental results - Source voltage and current waveforms – CFL 126.96.36.199 Chopper fed DC drive Controlled rectifiers are used in variable speed DC drives, DC power plants, induction heating, welding furnace control, etc. Source voltage (line) and current waveforms of a DC motor drive (230V,1500RPM, 4.5A) under loaded condition are shown in Figure 2.13.
Figure 2.13 Experimental results - Source voltage and current waveforms – MOSFET chopper fed DC motor drive 188.8.131.52 UPS In applications, such as medical intensive care systems, chemical plant process control, safety monitors, major computer installation, where even a temporary loss of supply causes severe consequence, there is need to provide an uninterruptible power supply system (UPS). Here the load can be permanently fed from the inverter. When the supply is energised, the rectifier keeps the battery charged to its prime condition. When the mains supply fails, the battery will supply the required power to the load, thus avoiding any interruption to the load [15, 16]. Source voltage and distorted source current waveforms with computer loads are shown in Figure 2.14. The harmonic content of the waveforms is dominated by odd harmonics. The contributions of all the even harmonic components of the phase current are negligible. The dominant phase current harmonic is the third, followed by the fifth, seventh and ninth harmonic. The total THD averaged around 130%. The THD of the source current increases during light load conditions and decreases during heavy load conditions.
Figure 2.14 Experimental results - Source voltage and current waveforms - Load A (Three PCs)
184.108.40.206 Unbalanced Load Combination of the loads mentioned in sections 220.127.116.11 to 18.104.22.168 are connected in each phase and three phase source current waveforms with unbalanced load conditions are shown in Figure 2.15.
Figure 2.15 Experimental results - Three phase source current waveforms under unbalanced conditions 22.214.171.124 Variable frequency drive The basic function of variable frequency drive is to act as a variable frequency generator in order to vary speed of a motor as per the user setting. It is used for speed control of 415V, 50Hz, 3kW, 4.5A, 1440 RPM induction motor and source voltage (line to line voltage) and current waveforms under full load conditions are shown in Figure 2.16. The spiky current is due to DC link capacitor, which will happen with any VSI after the DC link.
Figure 2.16 Experimental results - Source voltage and current waveforms – variable frequency drive under full load
126.96.36.199 Wind Farms Induction generators are commonly used as wind generators and when wind electric induction generators draw lagging reactive power, voltage drop in the line highly affects voltage regulation and power delivering capability of generators. Harmonic distortions are generated in the line currents by AC/DC or DC/AC converters and these harmonic distortions in line currents cause harmonic distortions in line voltages [11-14]. For the purpose of this work, the generation of harmonics by various loads and cancellation of harmonics are taken up. The most common harmonic index used to indicate harmonic content of distorted waveform is termed as Total Harmonic Distortion (THD). THD is the ratio of harmonic current to the fundamental component of load current and is expressed in %.THD is valid only for periodical waves which possess a fourier series . For non periodic waveforms, power quality is analysed by defining Total Demand Distortion (TDD) . It is the ratio of harmonic current to the fundamental component of rated load current or maximum demand current and is expressed in %.
2.4 Effects of harmonics on apparatus Harmonics increases copper loss, iron loss, dielectric loss and thermal stress in cables, transformers and rotating machines. 2 R I 12rms R(1 THD I ) 2  (i) Copper loss: The copper loss PR I rms
PR ( p.u .)
PR 1 THDI2 PR1
PR = Copper loss due to all frequency components of current PR1 = Copper loss due to fundamental component of current Irms= Total distorted current(rms) I1rms=Fundamental component of current (rms) THDI = Source current THD (%) (ii) Core loss: The core loss consists of hysteresis and eddy current losses. Hysteresis loss is loss in iron core when it is magnetized by an applied excitation or is rotating in magnetic field. Eddy current loss is associated with the flow of eddy currents induced in the armature core of rotating machine. Both these losses vary with frequency of current.
The hysteresis losses with harmonics  is Ph ( p.u .)
Ph v I hp .u . Ph1 h
Ph = Total hysteresis loss Ph1= hysteresis loss due to fundamental component of current Ih p.u. = hth harmonic current in p.u. of rated current =
= exponent depending on core material 1.6 The eddy current losses  with harmonics is
Pe ( p.u.)
Pe v h 2 I hp .u . Pe1 h
Equipment derating is a preventive requirement to keep the system within thermal limits. (iii) Dielectric loss: Peak voltage is increased by harmonics, which results in insulation stress or cable insulation breakdown. Dielectric loss  in capacitor or insulation loss in cable is
Pdh ( p.u .) hVhp2 .u .
Pdh = hth harmonic dielectric loss Vh p.u. = hth harmonic peak voltage in p.u. The effects of harmonics on major power system equipment are mentioned in the following paragraphs. Capacitor banks are overloaded by harmonic currents causing the fuses to blow. It also forms parallel resonant circuit with source inductance resulting in amplification of harmonics close to resonant frequency. In rotating machines, pulsating torques are produced due to interaction of either harmonics generated magnetic field and fundamental current or fundamental magnetic field and harmonic current. When any of these additional generated harmonic frequencies is close to natural frequency of oscillation of generator, super or sub synchronous resonance occurs. Harmonics also leads to (i) interference with power line carrier systems, (ii) mis-operation of systems which are used in remote switching, load control and metering, (iii) over voltages and excessive currents on the system due to resonance, (iv) dielectric breakdown of insulated cables resulting from harmonic over voltages, (v) unstable operation of firing circuits based on zero crossing detection or latching, (vi) excessive heating of transformers
due to frequency dependent core, and (vii) effects on computer and computerised automation production [12, 14].The above discussions show that harmonics causes various detrimental effects in power system equipment. To set an allowable limit to the harmonic distortions and individual harmonics, power quality standards are developed.
2.5 Power quality standards These are formal agreements between industry, user and government to generate, test, measure, manufacture and consume electric power. Mostly followed power quality standards are IEEE, ANSI, European Norms (EN), International Electro technical Commission (IEC) , Information Technology Industry Council (ITIC), U.S. National Electric Manufacturers Association (NEMA) etc. Out of these standards, IEEE standard 519-1992  is more practical and provides theoretical background. It contains 13 sections (i)
application of standards (section 1)
definition and letter symbols (section 2)
(iii) standard references (section 3) (iv) converter theory and harmonic generation (section 4) (v)
system response characteristics (section 5)
(vi) effects of harmonics (section 6) (vii) reactive power compensation and harmonic control (section 7) (viii) calculation methods (section 8) (ix) measurements (section 9) (x)
recommended practices for individual consumers (section 10)
(xi) recommended harmonic limits on system (section 11) (xii) recommended methodology for evaluation of new harmonic sources(section 12) (xiii) bibliography (section 13). IEEE 519 - 1992 standard is used as the reference in this work. It sets limits on voltage and current harmonic distortion at PCC (usually secondary side of the supply transformer).THD at PCC depends on percentage of harmonic distortion from each nonlinear device, total capacity of transformer and the rating of load. Two criteria that are used in IEEE 519 to evaluate harmonic distortion are (i)
Limitation on harmonic current that a user can transmit/inject into the utility system
Limitation on voltage distortion that utility must furnish the user
The standard limits for the above quantities are shown in Table 2.1 and Table 2.2.
Table 2.1 IEEE – 519 Maximum odd harmonic current THD limits (% of fundamental) for nonlinear loads at PCC at voltages of 2.4kV – 69kV h<11 h≥35 THD I sc 11h17 17h23 23h35 IL <20 4.0 2.0 1.5 0.6 0.3 5 20-50 7.0 3.5 2.5 1.0 0.5 8 50-100 10.0 4.5 4.0 1.5 0.7 12 100-1000 12.0 5.5 5.0 2.0 1.0 15 >1000 15.0 7.0 6.0 2.5 1.4 20 where, Isc = Short circuit current; IL =Nominal fundamental load current Even order harmonics are limited to 25% of above harmonic limits. Table 2.2 IEEE 519 Harmonic voltage THD limits (% of fundamental) at PCC for utilities
Max. individual harmonics THDv (%)
2.3kV – 69kV 3.0 5.0
69kV – 138kV 1.5 2.5
>138kV 1.0 1.5
Hence to keep up with the recommended standards, mitigation techniques are suggested.
2.6 Harmonic modeling The harmonics generated by the nonlinear loads, propagate throughout the network and affect all apparatus in the system. Hence, they are to be suppressed at the source of harmonic generation itself, by using suitable mitigation techniques. Therefore, the harmonic source detection and analysis are to be conducted before installing power quality improvement techniques. They are aimed at computing bus harmonic voltages, branch harmonic currents, voltage and current total harmonic distortions as well as detecting resonance conditions [8,10,25]. When conducting harmonic studies, system components are to be correctly modeled to ensure accurate and reliable harmonic distortion results. Harmonic analysis based on positive sequence representations was generally sufficient, when devices which were well balanced among the three phases (such as HVDC, SVC etc.) were used. But in utility secondary distribution systems, etc. for unbalanced harmonic analysis, the propagation of harmonics in each phase of a power system is to be assessed and hence needs full phase representation of network. Propagation of harmonic components in power systems depends on characteristics of harmonic sources as well as frequency responses of linear network components. Detailed analysis of influence of load characteristics on propagation of disturbances, cycloconverter harmonics, and six pulse converter harmonics are given in [20-22].Concepts and characteristics of power system 22
harmonics, modeling of harmonic sources and network components and techniques of network wide harmonic analysis are discussed in . General passive loads are represented by CIGRE model . Distributed single phase power electronic loads are modelled as fixed harmonic current injectors in distribution system studies. Their fundamental current component is varied in proportion with load power. Harmonic modeling of rotating machines, transformers, transmission cables, underground cables, arc furnaces, energy efficient lightings and some household electronic appliances are discussed in [25,26]. Éloi Ngandui, et al  frames the probability density functions of amplitudes and phase angles of various harmonic loads and results are validated by Monte-carlo simulations.
2.7 Power quality improvement techniques After harmonic modeling and analysis of the test system, the next step is the selection of suitable technique for harmonic cancellation or power quality improvement. The harmonic mitigation techniques can be classified into precautionary (preventive) solutions and corrective (remedial) solutions . Phase cancellation or harmonic control in power converters, and usage of low distortion loads are preventive solutions. Usage of harmonic filters for compensation of harmonics, reactive power and unbalance are corrective solutions. One has to consider (i) type of the filter (ii) control strategy and (iii) controllers used to execute control action while implementing these techniques. The following are types of the filters generally used. (a) Fixed element passive filters (b) Active filters (c) Hybrid filters 2.7.1 Fixed element passive filters Since the first installation of a passive tuned filter in the mid 1940’s, development of filter technology has advanced in steps. Conventionally tuned passive LC filters have been used to compensate a portion of reactive power and provide low impedance path to the harmonic currents in power system. A number of configurations are suggested such as single tuned, double tuned, triple tuned, quadruple tuned, damped, automatically tuned etc. Series passive filter is required to prevent a particular component from entering selected plant components or parts of a power system which offers large impedance to the relevant
frequency component. Shunt passive filter offers very low impedance to the tuned harmonic frequency and prevents from entering rest of the system as shown in Figure 2.17.
Figure 2.17 Fifth order harmonic filter provides low impedance path to 5th harmonics generated by the load The sharpness of tuning of filter is denoted by Q-factor. For tuned passive filters, Q-factor is recommended in the range 30 – 60 and for damped filters it is 0.5 - 5. But variations in the filter capacitance and inductance due to ageing and temperature may cause detuning from nominal tuned frequency. Double tuned filter has the advantage of reducing the power loss at fundamental frequency and is recommended for high voltage applications, because of the reduction in the number of inductors subjected to full line voltages.
Figure 2.18 Transformation from (a) two single tuned filters to (b) equivalent double tuned filters (c) impedance versus frequency of filter double tuned for 5th and 7th  Two single tuned filters, its equivalent double tuned filter configuration and its impedance characteristics are shown in Figure 2.18.The relationship between single tuned and double tuned filter parameters are shown in equations (2.6) – (2.10).
C1 C a C b
C a C b (C a C b )( La Lb ) 2 ( La C a Lb C b ) 2
La Lb La Lb
( La C a Lb C b ) 2 (C a C b ) 2 ( La Lb )
a 2 (1 x 2 ) (1 x 2 )(1 ax 2 ) 1 x2 R2 Ra R R b 1 2 2 2 2 2 2 (1 ax ) (1 ax )(1 x ) (1 x )(1 ax ) Where, a
Ca and x Cb
Lb C b .R1 includes resistance of inductor L1 . La C a
Triple and quadruple tuned filters are rarely used because of the difficulty of adjustment. Automatically tuned filters use a control system to measure reactive power and hence control the value of inductance and capacitance based on sign and magnitude of reactive power. It has advantages such as low inductor and capacitor ratings. For filtering a range of harmonic frequencies, damped filters are recommended, as shown in Figure 2.19. It provides low impedance for a wide spectrum of harmonics, without many parallel branches and reduces switching and maintenance problems. But it has disadvantages such as higher VA ratings, high losses in reactor and resistor. They are less sensitive to temperature variation, frequency deviation and component manufacturing tolerances .
Figure.2.19 High pass damped filters: (a) first order (b) second order (c) third order (d) C-type 
The drawback of passive filter is that it draws large reactive current of fundamental frequency causing increase in the source current. Also when load current changes, this type of filter is unable to adjust the amount of compensation. In automatically tuned passive filter proposed by Arillaga , the passive filter component values can be controlled as the load varies. The author feels a better term to describe such filter is adaptive filter. According to Wikipedia, “An adaptive filter is a filter that self-adjusts its transfer function 25
according to an optimization algorithm driven by an error signal”. Normally, adaptive filters are used in digital signal processing. However, the author proposes selection of filter parameters, namely, capacitance and inductance with the help of properly tuned knowledge base using Artificial Neural Network (ANN) to provide fast compensation. Hence digital controller is necessary. Practically passive filters have drawbacks like dependence of filtering characteristics on source impedance, detuning, parallel/series resonance between power system components, high no load losses, bulky size and fixed compensation. It cannot solve random variations in the load current waveform . To overcome the difficulties explained with the passive filters, active filters have been developed, which provide dynamic and adjustable solutions to power quality issues. 2.7.2 Active filters Principle of active filters was originally presented by H.Sasaki and T.Machida in 1971 .In 1976, L.Guguyi, et al presented a family of active filters consisting of PWM converters using power transistors and established the concept of active filter. These active filter circuitries have been developed employing modern fast switching power devices with turn off capability like insulated gate bipolar transistors (IGBTs).The Active filters (AFs), which have been used over the past few years, are the dynamic solutions to the power quality issues. They are pulse width modulated (PWM) voltage source or current source inverters (VSI/CSI) which help in power factor correction, reactive power compensation, harmonic cancellation and unbalance elimination with DC Capacitor back up based on selected control strategy. It injects harmonic component of voltage or current equal in magnitude, but opposite in phase, at the point of common coupling (PCC), thereby improving the power quality of the connected power system .
Figure 2.20 Basic active filter – Concept of harmonic current cancellation
Figure 2.20 illustrates the concept of the harmonic current cancellation so that the current supplied from the source is sinusoidal. The symbols IL, IF and IS represent
current, injected filter current and source current respectively. Hence, with the active filter, the source supplies only the real part of the fundamental component of current required by the load and the rest (harmonic and reactive currents) required by the load are supplied by the active filter. Bhim singh et al  has presented an extensive review of active filters to compensate harmonic current, reactive power, neutral current, unbalance current and harmonics. The published work on active filters are categorised based on criteria such as converter type, topology, supply system, control strategies, compensating variables, power rating and speed of response. The inverters commonly used in active filters are current fed PWM inverter and voltage fed PWM inverter, and they are shown in Figure 2.21. Voltage fed PWM inverter is more dominant, since it is lighter, cheaper and expandable to multilevel and multistep versions, to enhance the performance with lower switching frequencies .
Figure 2.21(a) Voltage fed PWM inverter
(b)Current fed PWM inverter 
Figure 2.22 Active filters -Topology 
The topologies used in active filters are series filters, shunt filters, unified power quality conditioners and hybrid filters, as shown in Figure 2.22. An active series filter is connected in series with the mains using a matching transformer to eliminate voltage harmonics and to balance and regulate terminal voltage of the load or line. It is shown in Figure 2.23(a). It also helps to damp out harmonic propagation caused by resonance with line impedance and passive shunt compensators. Series active filters are less common in industries than parallel active filters. It is because of the drawback of series circuits, that they have to handle high load currents, which increase their current rating considerably. This category of filters are mainly used to improve the quality of system voltage, which is important for voltage sensitive devices such as superconducting magnetic energy storage and power system protection devices . Shunt active filters have the advantage of carrying only the compensation current plus a small amount of active fundamental current supplied to compensate filter losses. It is shown in Figure 2.23 (b). They can be made suitable for a wide range of power ratings, by connecting several filters in parallel to supply higher currents. Shunt active filters are again subdivided into standard inverter, switched capacitor, lattice structured and voltage regulator filters . Unified power quality conditioner is a combination of active shunt and active series filters. Here, as shown in Figure 2.24, the DC link storage capacitor or inductor is shared between two voltage source or current source bridges operating as active series and active shunt compensators. It eliminates voltage and current harmonics. Its main drawbacks are large cost and control complexity . The active filters are again classified, based on supply and/ or the load, as single phase (two wire) and three phase (three wire or four wire).Two wire active filters are used in the case of nonlinear loads such as domestic appliances, as shown in Figure 2.25.
Figure 2.23 (a) Series active filter 
(b) Shunt active filter 
Fig.2.24 Unified power quality conditioner 
Figure 2.25 (a) Two wire series active filter 
(b) Two wire shunt active filter 
Three wire active filters are used for three phase nonlinear loads without neutral as in adjustable speed drives. Four wire active filters are used when a large number of single phase loads are supplied from three phase mains with neutral conductor. These loads may cause excessive neutral current, harmonic, and reactive power burden. Four wire shunt active filters are in different configurations such as capacitor midpoint type, four pole and three bridge structures, as shown in Figure 2.26. Capacitor midpoint type is used in smaller ratings and DC bus capacitor carries entire neutral current. In four pole switch type, fourth pole is used to stabilize the neutral of active filter. The three bridge configuration allows proper voltage matching for solid state devices and improves reliability of system. The configurations such as active series, active shunt, and the combination of both (as in unified line conditioners) are used in all the above cases .
Figure 2.26 (a) Capacitor midpoint four wire shunt active filter 
Figure 2.26 (b) Four pole four wire shunt active filter 
Figure 2.26(c) Three bridge four wire shunt active filter 
Figure 2.27 Active filter configurations  Active filter configurations are again classified according to the compensating variable as shown in Figure 2.27 .
The variables can be (i) Reactive power compensation: Active filters are rarely used for reactive power compensation, since other quasi dynamic, cheaper and slow response reactive power compensators are available in the market. It is suitable for low power applications only. (ii) Current harmonic compensation: Active filters are used for compensation of current harmonics in low and medium power applications. Current harmonic compensation reduces voltage harmonics at PCC. (iii) Voltage harmonic compensation: Voltage harmonic compensation is not widely addressed because power supplies usually have low impedance. But it is very significant for harmonic voltage sensitive devices such as power system protection devices and super conducting magnetic energy storage. The reduction of voltage harmonics at the point of common coupling reduces a great amount of current harmonics. (iv) Balancing of mains voltage: Three phase system voltage unbalance is mainly due to significant amount of supply impedance especially in low and medium power systems. Active power filters add corresponding amount of instantaneous voltage to each phase to follow the reference sinusoidal waveform. (v) Balancing of mains current: Active power filters are used for balancing of mains current in low power applications. (vi) Multiple variable compensation: Different combinations of above mentioned compensating variables are used to improve the effectiveness of filters. (a) Harmonic currents with reactive power compensation: Active power filters are most commonly used for compensation of harmonic currents and reactive power compensation. However, because of the limits imposed by the ratings of power switches, it is recommended for low power applications. (b) Harmonic voltages with reactive power compensation: This system is only suitable for low power applications. (c) Harmonic currents and voltages: Series or parallel combination of active filters is used for both harmonic current and voltage compensation. However, these complex configurations are used only for very sensitive devices such as power system protection equipment and superconducting magnetic energy storage system.
(d) Harmonic currents and voltages with reactive power compensation: This technique requires the use of parallel/series active filter combination. It is not employed very often because its control is rather difficult .
Active power filters are categorised into three according to power rating and speed of response required in the compensated system. The power rating of the compensated system and its speed of response are significant while selecting active filter. Cost of system increases proportional to the required speed of response. The categories are (i)low power (ii) medium power and (iii) high power applications, as shown in Figure 2.28 .
Figure 2.28 Classification of active filters according to rating  (i) Low power applications: It is concerned with systems of power ratings below 100kVA, mainly in residential areas, commercial buildings and hospitals. These dynamic active filters are high pulse number PWM VSI or CSI. Their response time is relatively much faster ranging from tens of microseconds to milliseconds and can be used for single phase and three phase systems. (ii) Medium power applications are used in three phase systems ranging from 100kVA to 10MVA (for current harmonic compensation). Speed of response is in the range 100ms-1 sec. (iii) High power applications: Implementation of very high power dynamic filter is extremely ineffective because of the lack of high switching frequency power devices that can control the current flow at high power ratings. This is a major limitation for such systems. Here, response time is in the range of tens of seconds. 32
However, the rating of active filters is very close to load rating (up to 80%) and hence cost of shunt active filters is high. They are difficult to be implemented in large scale. Additionally, the effectiveness of active filter is determined by its control algorithm. The above reasons led to different solutions to improve the practical utilisation of active filters. One of them is to use a combined system of passive filters and active filters as hybrid filters.
2.7.3 Hybrid filters Section 2. 7. 1 deals with passive filter and section 2. 7. 2 deals with active filters. Hybrid filters are combinations of more than one active filter or passive filter. The combination of filters operating in parallel is called shunt hybrid filter. It can improve power quality, at the same time, reduce filter rating. Generally, passive filters are used to eliminate lower order harmonics. The higher order harmonics, which are much less compared to lower order harmonics, are eliminated by active filter. Hybrid filters are quite popular because the solid state devices used in the active part can be of reduced size (about 5% of load size). According to research publications , hybrid filters are classified based on the number of elements in topology, nature of supply system and types of converters used (Figure 2.29). Main classification of hybrid filters is made on the basis of supply system, with topology of filters as sub-classification. The supply system can be single phase 2 wire, three phase 3 wire or three phase 4 wire. The number of elements in the topology can be two, three or more, and it can be active filter or passive filter. Each category of the above mentioned hybrid configurations are classified into: (i)hybrid of two passive elements, (ii) hybrid of three passive elements, (iii) hybrid of one active and one passive filter, (iv) hybrid of three elements- two passive with one active, (v) one passive with two active filter elements, (vi) hybrid of two active elements and (vii) hybrid of three active filter elements .The hybrid filters of more than three elements are rarely used because of complexity considerations.
(i) Hybrid of two passive elements: Two circuit configurations commonly used are: (a) combination of passive series and passive shunt, and (b) combination of passive shunt and passive series filters, as shown in Figure 2.30.
(ii) Hybrid of three passive elements: Two circuit configurations used are: (a) combination of passive series, passive shunt and passive series filters (b) combination of passive shunt, passive series and passive shunt filters (Figure 2.31).
Figure 2.29 Hybrid filters- Classification 
Figure 2.30 Hybrid of two passive elements (a) combination of passive series and passive shunt filters (b) combination of passive shunt and passive series filters 
Figure 2.31 Hybrid of three passive elements (a) combination of passive series passive shunt and passive series filters (b) combination of passive shunt, passive series and passive shunt filters (iii) Hybrid of two active elements: Two configurations are shown in Figure 2.32.
Figure 2.32 Hybrid of two active elements: (a) active series and active shunt filters, and (b) active shunt and active series filters 
(iv) Hybrid of one active and one passive element: Various configurations are shown in Figure 2.33.
Figure 2.33 Hybrid of one active and one passive elements (a) series connected passive series and active series filters, (b) parallel connected passive series and active series filters, (c) passive series and active shunt filters, (d)active shunt and passive series filters, (e)active shunt and passive shunt filters, and(f)series connected passive shunt and active shunt filters  (iv) Hybrid of three active elements: Two configurations are shown in Figure 2.34.
Figure 2.34 Hybrid of three active elements(a)combination of active series, active shunt in parallel with series combination of active series filter and nonlinear load, and (b) combination of active shunt, active series, and active shunt filters across the load 
(v) Hybrid of one active and two passive elements: Different combinations are shown in Figure 2.35.
Figure 2.35 Hybrid of one active and two passive elements: (a) passive shunt, passive series and active series, (b)passive series, passive shunt and active series, (c)passive series in series with parallel connected active series and passive series filters, (d) passive shunt and parallel connected active series and passive series filters, (e) passive series, active shunt and passive series filters, (f) parallel connected passive shunt, active shunt and passive series filters, (g) active series, passive shunt and passive series filters, (h) series connected passive shunt with active series and passive series filters, (i) series connected passive series with active series in parallel with passive series filters, (j) passive series, parallel connected passive shunt and active shunt filters, (k) passive shunt, passive series and active shunt filters, (l) series connected passive series ,active series and passive shunt filters,(m)passive shunt, active series and passive shunt filters, (n)passive shunt, series connected passive shunt with active shunt filters, (o) passive shunt and series connected active series with passive shunt filters, (p) active shunt, passive series and passive shunt filters,(q) parallel connected active series with passive series and passive shunt filters, and (r) passive shunt and parallel connected passive shunt with active series filters 36
(vi) Hybrid of two active and one passive element: Various configurations are shown in Figure 2.36.
Figure 2.36 Hybrid of two active and one passive elements (a) active shunt, passive series and active series filters, (b) active series, active shunt and passive series filters, (c) active series, parallel combination of active series and passive series filters, (d) active series, passive shunt and active series filters, (e) active series, passive shunt and active series filters, (f) active shunt, passive shunt, active series filters, (g) passive series, active shunt, active series filters, (h) series connected active shunt with passive shunt and active series filters, (i) active series, passive series ,active series filters, (j) active series, active shunt and passive shunt filters, (k) active shunt, active series and passive shunt filters, (l) active series, passive series and active shunt filters, (m) active shunt, passive series and active shunt, (n) series connected active series, series connected active shunt and passive shunt, (o) series connected active shunt, series connected active shunt and passive shunt filters, (p) passive shunt, active series ,active shunt filters, (q) parallel connected passive series with active series and active shunt filters, and (r) active shunt, in series with parallel connected active shunt and passive shunt filters
Hybrid filters are developed and tested successfully in applications such as high power nonlinear loads , three phase three wire railway traction loads , transformer less shunt hybrid power filter , three phase anti resonance hybrid filter .Design criteria, and implementation of hybrid filter to obtain flexible and robust operation are also discussed [35-45].The complex control circuits constitute a major portion of the total cost of hybrid filters. The different types of filters used to compensate harmonics are reviewed in section 2.7.The passive filter will remove harmonics for which it is tuned without any additional control strategy or controllers. However, the remaining types of filters have to be controlled for efficient harmonic suppression. The review of control strategies is given in the following section.
2.8 Control strategies Control strategy of active or hybrid filter is implemented in three stages [29,30]. In the first stage, essential voltage and current signals are sensed using Hall effect sensors and isolation amplifiers to acquire system conditions. Compensating signals are derived based on suitable control techniques in the second stage. For implementing control algorithm, several instantaneous voltage and current signals such as supply terminal voltages, DC bus voltage across capacitor, load currents, source currents, compensation currents, and DC link current are to be monitored. The various control techniques, classified in Figure 2.37, are used to derive compensating signals. These techniques are classified as: (i) time domain control, and (ii) frequency domain control.
Figure 2.37 Control techniques of active filters 
Continuous time domain control uses analog filter to determine supply current harmonics. It is simple to implement. Two categories of continuous time domain control are (a) high pass filter method, and (b) low pass filter method. High pass filter removes low order frequency signals in the load current and resulting high frequency components constitute the desired reference signal. It is vulnerable to noise. Low pass filter method filters the fundamental component and desired reference is obtained by subtracting it from the total load current. Low pass filter method is more commonly used. The significant time domain algorithms, which are used for single phase or three phase systems are: (i) Instantaneous reactive power algorithm , (ii) Synchronous detection algorithm , (iii) DC bus voltage algorithm , (iv) Icos algorithm  etc., If any one of the existing control algorithms of active filter is implemented using intelligent techniques such as artificial neural networks, fast reactive power and harmonic compensations can be provided. The intelligent controllers make use of a knowledge base for the generation of compensation signals and switching pulses for the active filter [58, 60-62]. Frequency domain approaches are suitable for both single and three phase systems. Basically they are derived from conventional Fourier analysis. (i) Conventional Fourier and FFT algorithms: Load current is sensed and transformed into frequency domain. The fundamental component is removed from the transformed current signal and inverse transform is applied to obtain a time domain compensating signal. This technique needs to take samples of one complete cycle to generate Fourier coefficients and therefore are suitable for slowly varying load conditions. (ii) Modified Fourier series technique: Only the fundamental component of current is calculated and this is used to separate the total harmonic signal from the sampled current waveform. The computation time is much less than other techniques used for single and three phase applications [29,30]. In the third stage of control, the gating signals for switching devices are generated using PWM, hysteresis, sliding mode or intelligent technique based controllers [29, 30]. Realisation of control algorithm can be done through analog or digital control circuits. The controllers are discussed in section 2.9.
2.9 Analog and digital controllers Analog controllers use circuit components such as operational amplifiers, diodes, transistors, pulse transformers, multivibrators and zero order hold circuits. It consists of large number of components and any change in design requires rearranging the whole circuit and hence time consuming. The major drawbacks are ageing of components, deviation of component values with respect to the environmental conditions and difficulty to produce the controller circuits in bulk quantities. To avoid the limitations of analog circuit controllers, digital controllers are developed. Digital controllers are more flexible, cheap, adaptable and can be easily manufactured in large quantities. Major component of a digital controller is digital signal processor, which receives the input signals, computes the algorithm and generates optimised PWM signals. An improved low cost sensor technology, compact isolation amplifiers, low cost, highly accurate and fast digital signal processors, microcontrollers and application specific integrated circuits (ASICs) have made possible the implementation of complex control algorithms for real time control at an acceptable price [46,47]. FPGA based implementation  was also developed. Anis Ibrahim et al  reviews applications of advanced mathematical tools such as fuzzy logic [50, 51], neural network  and wavelet transforms  in power quality. Scope for improvement still remains in the configuration and/or control aspects of the active and hybrid filters. A digital controller for an ANN based adaptive shunt hybrid filter, i.e., combination of ANN controlled shunt active filter and adaptive shunt passive filter, is proposed by the author for harmonic and reactive power compensation for nonlinear loads. The load in the system will be varying from instant to instant and computations of reference current are to be carried out repeatedly. If these computed values of reference currents corresponding to different values of load current are used as knowledge base for training ANN network, compensation can be provided at a faster rate. An ANN based digital processor is proposed by the author, which generates PWM signals to switching devices of shunt active filter, and gating signals for thyristors in the adaptive shunt passive filter. It helps the active filter to generate suitable compensation signals and adjusts passive filter component values, based on the harmonic and reactive power compensation requirement of the non-linear load.
2.10 Conclusion After reviewing issues related to power quality, this chapter discusses different types of filters used to mitigate distortion and unbalance in power supply. The earliest type of filters were shunt /series passive filters connected at PCC which provide a low/high impedance path to the selected harmonics, preventing most of the selected harmonics from appearing at the source. But they have drawbacks such as resonance, fixed compensation, high no load losses, bulky size etc. As better options for complete compensation of distortions, active power filters have been researched and developed. Many mature control algorithms are available in literature for the control of three-phase active filters. Most of them use tedious computations and complex circuits and hence are highly expensive and slow in response. Hence an effective option, combination of passive and active filters, named hybrid filter, is implemented. Hybrid filters allow designing active filters for only a fraction of total load power. This chapter surveys various configurations and control algorithms of active filter and hybrid filter, proposed and tested by different authors. Their merits and demerits are briefly touched upon. The author feels there is scope for improvement of hybrid filter performance, especially under dynamic conditions of the load and an ANN based digital controller is proposed for adaptive shunt hybrid filter. This chapter explains the significant points to be taken into account while considering power quality improvement schemes such as: (i) type of the filter, (ii) control algorithm – their suitability under static/dynamic conditions, and (iii) type of controller –analog/digital.
The expected outcome of this research work is to develop an effective and lower rated filter configuration for compensation of reactive power, harmonics and imbalance in source current under steady state and dynamic conditions.