A

itmn

oast©

K

1

actcthwswmir

0d

Electric Power Systems Research 77 (2007) 821–830

Comparative evaluation of two models of UPQC for suitableinterface to enhance power quality

Malabika Basu a,∗, Shyama P. Das b, Gopal K. Dubey b,�a Department of Electrical Engineering, Dublin Institute of Technology, Kevin Street, Dublin 8, Ireland

b Department of Electrical Engineering, Indian Institute of Technology, Kanpur, India

Received 5 October 2005; received in revised form 27 June 2006; accepted 17 July 2006Available online 18 September 2006

bstract

Majority of the dispersed generations from renewable energy sources are connected to the grid through power electronic interface, whichntroduce additional harmonics in the distribution systems. Research is being carried out to integrate active filtering with specific interface suchhat a common power quality (PQ) platform could be achieved. For generalized solution, a unified power quality conditioner (UPQC) could be the

ost comprehensive PQ protecting device for sensitive non-linear loads, which require quality input supply. Also, load current harmonic isolationeeds to be ensured for maintaining the quality of the supply current.

The present paper describes two control scheme models for UPQC, for enhancing PQ of sensitive non-linear loads. Based on two different kindsf voltage compensation strategy, two control schemes have been designed, which are termed as UPQC-Q and UPQC-P. A comparative loading

nalysis has developed useful insight in finding the typical application of the two different control schemes. The effectiveness of the two controlchemes is verified through extensive simulation using the software SABER. As the power circuit configuration of UPQC remains same for bothhe model, with modification of control scheme only, the utility of UPQC can be optimized depending upon the application requirement.

2006 Elsevier B.V. All rights reserved.

ffmiitantw

o

eywords: Distributed generation; Power quality; VA rating analysis; UPQC

. Introduction

Distributed generation (DG) systems have both advantagesnd disadvantages in relation to grid power quality (PQ). Theyan increase the efficiency of systems by local power genera-ion. More reliable and uninterrupted power can be provided toustomers, with energy cost savings [1]. World wide DG pene-ration in the grid is on the rising. For example, Denmark has aigh penetration of wind energy in the country with 14% of thehole electrical energy consumption supplied from wind [2]. A

tudy by EPRI indicates that by 2010, 25% of new generationill be DG and at least it will be 20% of the total electrical utility

arket, worth of USD 72 billion [3]. Deregulation of electric-

ty market may contribute to rising penetration level of DG fromenewable energy sources (wind, solar, biomass, etc.) in the near

∗ Corresponding author. Tel.: +353 1 4024996; fax: +353 1 4024992.E-mail address: mbasu@dit.ie (M. Basu).

� Deceased.

ssadpnta

378-7796/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.epsr.2006.07.008

uture [4]. From the perspective of environmental protection, DGrom renewable energy sources is of great importance, as theyinimize harmful emissions. As most of the DG systems are

nterfaced to the grid through power electronic interface, hencenjection of additional higher frequency harmonics in the sys-em is obvious. Therefore, additional grid integration problemsre equally worrying from electrical pollution point of view ifot attended properly. Furthermore, variable wind speed, varia-ion in solar and tidal power, etc., are uncontrollable parametershich are bound to affect the generated power quality.Research is being carried out to integrate active filtering

ptions into the integrating power electronic converters them-elves [2,5], but they need to be case specific. From the per-pective of sensitive non-linear loads in the distribution system,common platform of PQ needs to be ensured; as PQ varies

ue to various types of sources of generation. Hence, suitable

ower conditioning interfaces are recommended for sensitiveon-linear loads. These type of loads primarily include produc-ion industries (like automotive plants, paper mills, chemicalnd pharmaceutical industries, semiconductor manufacturing

8 ystem

palrelotcr

famctpadwcasac

U3dlcs

2

s

1uiitvlicitdButeot

•

•

22 M. Basu et al. / Electric Power S

lants, etc.), and critical service providers like medical centres,irports, broadcasting centres, etc. Typical grid integration prob-ems associate with voltage and frequency compatibility andequirement of active and reactive power. A power conditioningquipment can act as an interface between the grid and sensitiveoads, so that the load can remain insensitive to the variationf power quality from the utility. Unified power quality condi-ioner (UPQC) happens to be the most comprehensive poweronditioning equipment that can mitigate both voltage and cur-ent quality problems [6,7].

In this paper, two models of UPQC are discussed and analysedrom the perspective of VA loading and applications. Function-lly UPQC is a combination of series and shunt active filter, foraintaining desired quality of both the incoming voltage and

urrent. But its coordinated control gives it unique feature inerms of shared responsibility and reduced VA rating as com-ared to individual dynamic voltage restorer (DVR) [8,11] orctive power filter (APF) [6,9,11]. The two control schemesescribed in this paper have common current control strategy,hich is based on hysteresis current control. The series voltage

ompensation can be performed in a number of ways, whichre non unique. Based on the two extreme options, two controlchemes have been designed and their performance and ratingre analysed. The insight gained could be useful for design ofontrol strategy of UPQC for various applications.

The organization of the paper is as follows. In Section 2PQC topology and power flow strategy is described. Sectiondescribes the two proposed control strategies for UPQC for

ifferent applications. Section 4 discusses analytically the VAoading and rating issues. Simulation results in support of theontrol strategy are provided in Section 5. Finally, the conclu-ions are presented in Section 6.

. UPQC topology and power flow strategy

A three-phase UPQC consists of two three-phase voltageource inverters connected in cascade as shown in Fig. 1. Inverter

Fig. 1. Power circuit diagram

s Research 77 (2007) 821–830

(Series Inverter (SEI)) is connected in series with the incomingtility supply through a low pass filter and a voltage inject-ng transformer. Inverter 2 (Shunt Inverter (SHI)) is connectedn parallel with the sensitive load, whose power quality needso be strictly maintained. The main purpose of SHI is to pro-ide required VAR support to the load, and to suppress theoad current harmonics from flowing towards the utility and its operated in current controlled mode. SEI is responsible forompensating the deficiency in voltage quality of the incom-ng supply, such that the load end voltage remains insensitiveo the variation of utility supply. The two models of UPQCiscussed in this paper have same power circuit configuration.ut as the control strategies are different in SEI, the individ-al loading of SHI and SEI varies and the overall rating ofhe UPQC differs, which is the thrust of this paper and isxplained in the subsequent sections. The UPQC also has a fewther important components that are essential for interfacing ofhe same.

The SHI is connected through a boost inductor LSHI, whichcan boost up the common dc link voltage to the desired valuethrough appropriate control. The size of the inductor has tobe chosen carefully, as increase in size would cause slowerresponse to current control.The dc link capacitor C provides the common dc link voltageto both SEI and SHI. Ideally once charged, the dc link voltageshould not fall off its charge, but due to finite switching lossesof the inverters, inductor and capacitor, some active power isconsumed and the charge of the dc link voltage needs to bemaintained in a closed loop control, through the SHI. Thechoice of the reference dc link voltage depends upon the per-centage of voltage sag to be mitigated and amount of VARto be shared. The higher of the two values is to be chosen

to comply with all needs. It is to be noted that as the C ischarged continuously through SHI, it does not require addi-tional source of voltage support. The online charging alsohelps UPQC in mitigating voltage unbalance or under-voltage

of three phase UPQC.

ystem

•

upatc

3

Ate

3

oslisnS

dfacspsrsatae

3f

qnsd

Soa

uvVasltfoanbtcipfss“2” to the parameters. Consequently the load current changesto IL2. The SHI injects IC2 in such a way that the active powerrequirement of the load is only drawn from the utility. Therefore,from the utility side the load power factor is always unity. It can

M. Basu et al. / Electric Power S

situations for longer durations, as it is not limited by the stor-age capacity of a separate voltage source.The SEI needs to be connected to the supply side througha series injection transformer and a low pass filter (LPF), toeliminate the high switching frequency ripple of the inverter.The filter may inject some phase shift, which could be loaddependent, but suitable feedback control is to be designed todynamically adjust the shift, which is described in the controlsection.

The active power flow through the UPQC originates from thetility, as it is the only source of active power. But the reactiveower and load harmonic currents are shared between the SHInd loads primarily. Therefore, SHI provides harmonic isolationo the utility. SEI may also share some VAR depending upon theontrol, described further in the subsequent section.

. Control strategies

In this section two control schemes of UPQC are discussed.s the SHI control scheme remains same in both the scheme,

hat is discussed first and then the two different SEI schemes arexplained.

.1. Control scheme for SHI

To ensure fast elimination of higher order current harmonicsf the load, hysteresis controller is designed for controlling thewitching of the SHI. Based on the active power demand of theoad, a suitable sinusoidal reference is selected for the incom-ng utility current and in addition appropriate hysteresis band iselected. Narrower hysteresis band ensures higher THD elimi-ation, at the cost of higher switching frequency of the inverter.uitable trade off in design is required to optimize all criteria.

As discussed earlier, the dc link voltage ideally should notecay, unless some active power loss occurs in the UPQC. There-ore, the deviation of the dc link voltage acts as a measure ofctive power requirement from utility supply. The error is pro-essed through a PI controller and a suitable sinusoidal referenceignal in phase with the supply voltage is multiplied with the out-ut of the PI controller, to generate the reference current for theupply. Hysteresis band is imposed on top and bottom of thiseference current. The width of the hysteresis band is adjusteduch that the supply current THD remains within internationalgencies specified limit. As the supply current hits the upper orhe lower band, appropriate switching of the SHI takes place sos to compel the supply current to remain within the band, byither aiding its dc link voltage to utility supply or by opposing.

.2. Control scheme A for SEI (quadrature compensationor SEI) [7]

In this scheme the injected voltage from SEI maintains a

uadrature advance relationship with the supply current, so thato real power is consumed by SEI in the steady state. This is aignificant advantage when UPQC mitigates under-voltage con-itions. The SEI also shares the VAR of the load along with the

F(at

s Research 77 (2007) 821–830 823

HI, so the VA loading of SHI reduces. To highlight this aspectf quadrature voltage injection, this scheme will be henceforthddressed as UPQC-Q.

Fig. 2 shows the current and voltage required from UPQCnder a typical load power factor condition for a typicaloltage sag. When the supply voltage has no deficiency;S = VL1 = VS1 = V0 (a constant), and the series injected volt-ge Vinj requirement is zero. This state is represented by addinguffix “1” to all the voltage and current quantities of interest. Theoad current is IL1 (IL1 = IL) and the SHI compensates the reac-ive component IC1 of the load, resulting in unity input poweractor. Thus, the current drawn by the SHI is −IC1, which ispposite to the load reactive current IC1. As a result, the loadlways draws the in-phase component IS1 from the supply. Foron-linear loads, the SHI not only supplies the reactive current,ut also the harmonic currents required for the load. Thus, afterhe compensation action of the SHI, only the fundamental activeomponent of the current is required to be supplied from the util-ty. As soon as the load voltage VL sags, due to utility voltageroblems, the UPQC is required to take action to compensateor the sag, so that VL is restored to its desired magnitude. Aseen from Fig. 2, the restoration of VL is achieved by specificallyelecting γ = 90◦. This condition is represented by adding suffix

ig. 2. Phasor diagram of UPQC for fundamental power frequency, when θ < Φ.a) Describes STATE 1, when the supply voltage equals the desired load voltagend (b) describes STATE 2, when the supply voltage sags and UPQC injects Vinj

o maintain the load voltage at its desired level.

8 ystem

bI

V

w

I

uoti

Uattfnd

fittcit

dtrtitncl

3

24 M. Basu et al. / Electric Power S

e observed from the phasor diagram that the utility current isS2, and is in phase with VS2.

If the active power demand is constant,

S1IS1 = VS2IS2 (1)

hich can be written as

S2 = VS1IS1

VS2(2)

Between the control loops, the hysteresis current control loopsed with the SHI is much faster than the voltage control loopf the SEI. The two loop speeds are chosen such that in no casehese two controllers can interfere with each other and causenstability.

Fig. 3 gives an overview of the control schematic for thePQC-Q. The supply voltage peak detector would indicate of

ny voltage sag, which would require to be compensated byhe SEI of UPQC-Q. One fast feed-forward path is designed

o determine the initial modulating index of the SEI. A slowereedback path through another PI controller is implemented toullify the injected phase angle error, which may occur due toynamic load change and the presence of LPF.

wm

Fig. 3. Control block dia

s Research 77 (2007) 821–830

Because of quadrature voltage injection by SEI, the low passlter (LPF), and the load current will appear to be inductive to

he SEI and significant variation in the load current would alterhe phase angle that cannot be predetermined. But a feedbackontroller that compares the actual injected voltage (Vinj) to thedeal injected voltage (V ∗

inj) can eliminate this error caused dueo dynamic load change.

Apart from these, there are transformer leakage reactancerop, resistance drop and the voltage drop due to the low pass fil-er (LPF) connected at the output of the SEI to filter the switchingipples of the SEI. The load active power demand increase leadso increase in the source current. This current is also reflectedn the primary side of the series transformer. Thus, the drop inhe above mentioned elements change the injected voltage mag-itude and phase which need to be corrected by a closed-loopontrol with good dynamics. It should be noted that only open-oop control for the SEI is inadequate.

.3. Control scheme B for SEI [10]

In this scheme, in general the injected voltage is in phaseith the supply voltage when the supply is balanced. Thereforeostly the SEI would consume active power. To distinguish from

gram of UPQC-Q.

M. Basu et al. / Electric Power System

twimosspttUtivdpTpc

i

c2qstcbdwbtdt

4

4

dooFc

V

Fig. 4. Phasor diagram of UPQC-P.

he earlier SEI control scheme this type of UPQC control schemeill be henceforth addressed as UPQC-P. By virtue of in phase

njection, UPQC-P will mitigate voltage sag conditions by mini-um injected voltage. The phasor diagram of Fig. 4 explains the

peration of UPQC-P for the fundamental frequency. When theystem voltage and current are in phase due to the action of thehunt compensator, the series converter handles purely activeower. As seen from Fig. 4, the SHI current increases whenhere is a supply voltage sag, as the SEI consumes active powerhrough the SHI. When the supply sag is created, the SEI of thePQC-P should compensate for the fall in voltage to maintain

he load voltage to its specified value. The injected voltage beingn-phase with the supply voltage, the supply current and injectedoltages are also in-phase with each other. Hence, the SEI han-les only active power. The SEI delivers this additional activeower by drawing the same from the dc link of the UPQC-P.herefore, it acts as an active load to the SHI. As seen from the

hasor diagram, IC2 has an additional active and same reactiveomponent as IC1.

The control scheme is based on abc–dqo analysis of thencoming voltage. In steady state and balanced supply voltage

i

I

Fig. 5. Control block for voltage unba

s Research 77 (2007) 821–830 825

ondition, the d-component of voltage will be 396.7 V dc for30 V rms phase voltage, which is considered as reference. Theand o-component of voltage will be zero. If there is a balanced

upply voltage sag, the d component of voltage will deviate fromhe reference voltage, but q and o-component will remain zero. Inase of unbalanced supply voltage sag q and o-components wille ac quantities and d-component of voltage will contain bothc and ac quantities. The ideal reference being known the SEIould operate in such a manner so that the difference in voltageetween the reference d–q–o quantities and the actual quanti-ies are supplied by SEI. A closed loop feedback can ensure theynamical changes are taken care of. Fig. 5 shows the detail ofhe SEI control scheme for UPQC-P.

. Comparative VA rating calculation and analysis

.1. UPQC-Q

The overall VA handled by the UPQC is an important factoreciding its size. The power loss is also related to the VA loadingf the UPQC. Here, the loading calculation has been carried outn the basis of linear load for fundamental frequency [10]. Fromig. 2 it can be found out that the load voltage is to be keptonstant at Vo p.u. irrespective of the supply voltage variation.

S = VL1 = VL2 = VS1 = Vo p.u. (3)

The load current is assumed to be constant at the rated value,.e.,

L = IL1 = IL2 = Io p.u. (4)

lance compensation in UPQC-P.

8 ystem

wli

V

V

c

I

wθ

V

∴

∴

t

I

V

w(

4

t

p

V

I

wr

i

V

V

W

I

∴

a

I

∴

Af

4

osrtrurTmt

iUimtFoUT

26 M. Basu et al. / Electric Power S

ith fundamental p.f. = cos φ. Assuming the UPQC-Q to be loss-ess, the active power demand in the load remains constant ands drawn from the source, i.e.,

SIS = VLIL cos φ (5)

In case of a sag when VS2 < VS1, where x denotes the p.u. sag,

S2 = (1 − x)VS1 = Vo(1 − x) p.u. (6)

Now, to maintain constant active power under the voltage sagondition (as explained in (1)),

S2 = VS1IL cos φ

VS1(1 − x)= Io cos φ

1 − xp.u. (7)

As the voltage injected (Vinj) by the SEI is in quadratureith the supply, the resultant load voltage VL2 makes an angle(Fig. 2) with the supply VS2, which implies

inj =√

V 2S1 − V 2

S2

Vinj

VS2= tan θ, Vinj = VS2 tan θ, Vinj = Vo(1 − x) tan θ p.u. (8)

SEI VA Rating = VinjIS2 = VoIo cos φ tan θ p.u. (9)

The SHI current can be calculated from the trigonometry ofhe vector diagram (Fig. 2)

C2 =√

I2L2 + I2

S2 − 2IL2IS2 cos(φ − θ)

= Io

√(1 − x)2 + cos2 φ − 2 cos φ cos(φ − θ)(1 − x)

1 − xp.u.

(10)

It follows that the rating of the SHI is

L2IC2 = VoIo

√(1 − x)2 + cos2 φ − 2 cos φ cos(φ − θ)(1 − x)

1 − x

+ I2o

(1 − x)2 + cos2 φ − 2 cos φ cos(φ − θ)(1 − x)

(1 − x)2 ZSHI p.u.

(11)

here ZSHI is the shunt inductance impedance. Adding (9) and11), the total VA rating of the UPQC-Q can be evaluated.

.2. UPQC-P

The loading calculation of UPQC-P has been carried out onhe basis of linear load.

From phasor diagram of Fig. 4, it can be found that for eachhase

L1 = VL2 = VS1 = Vo p.u. (12)

If the load current is assumed to be

L = IL1 = IL2 = Io p.u. (13)

ith fundamental p.f. = cos φ, active power demand in the loademains the same,

.e., VSIS = VLIL cos φ (14)

iU

t

s Research 77 (2007) 821–830

In case of sag when VS2 < VS1, where x denotes the p.u. sag,

S2 = (1 − x)VS1 = Vo(1 − x) p.u. (15)

Now, to maintain constant active power

S1IS1 = VS2IS2 (16)

hich leads to,

S2 = VS1IL cos φ

VS1(1 − x)= Io cos φ

1 − xp.u. (17)

SEI VA Rating = VinjIS2 = VoIo(x cos φ)

1 − xp.u. (18)

nd

C2 =√

I2L1 + I2

S2 − 2IL1IS2 cos φ

= Io

√(1 − x)2 + cos2 φ{1 − 2(1 − x)}

1 − xp.u. (19)

SHI VA Rating

= VoIo1

1 − x

√(1 − x)2 + cos2 φ{1 − 2(1 − x)}

+ I2o

(1 − x)2 + cos2 φ{1 − 2(1 − x)}(1 − x)2 ZSHI p.u. (20)

dding (18) and (20), the total VA rating of the UPQC-P isound.

.3. Comparative analysis of VA loading

Figs. 6–8 show the comparison of SEI, SHI and total loadingf UPQC, respectively. The 10 points in each set are for p.u.upply voltage sag from 5% to 50%, which are typical. Thisange has been chosen as the most practical cases are observedo be in this range as available from PQ survey reports. A wideange of load power factor has been chosen from 0.6 lagging tonity power factor (u.p.f.), with ZSHI = 1 p.u. in all cases. Theating of the equipment has been estimated from (9) and (11).he maximum loading within the opearting zone would deter-ine the rating of the individual inverter, and the summation of

he two would yield the total rating of the UPQC.As observed from Fig. 6, it is seen that loading on the SEI

ncreases as % sag increases. The SEI maximum loading underPQC-P control rating will be 1 p.u. (based on maximum load-

ng at 50% sag at u.p.f. load p.f.) to successfully cater theentioned region of voltage sag under the specified power fac-

or variation. Corresponding UPQC-Q SEI loading is 1. 73 p.u.rom Fig. 7 it is observed that the maximum loading conditionccurs at similar condition mentioned above. Maximum SHI forPQC-P is 2 p.u., whereas for UPQC-Q it is as high as 4.73 p.u.he total VA loading is the sum of the two individual load-

ng, and thus the maximum UPQC-P rating would be 3 p.u. andPQC-Q would be 6.46 p.u.It is interesting to note that UPQC-Q does not seem to be

he natural choice considering the double rating as compared to

M. Basu et al. / Electric Power Systems Research 77 (2007) 821–830 827

oadin

Uctit

V

Fig. 6. SEI l

PQC-P. However, considering all the detail of the SHI loading

urves in Fig. 7, it can be observed that the loading of SHI ofhe UPQC-Q is considerably lower than that of the UPQC-Pn the low power factor load region. Hence, depending uponhe load requirement, UPQC-Q could be a better choice, where

fasa

Fig. 7. SHI loadin

g of UPQC.

AR demand of the load may be high, and typically the need

or VAR compensation would be essential. Thus, considering thepplication area of operation upto 0.8 lagging p.f., and upto 50%upply voltage sag, typically with the same rating of UPQC-Qnd UPQC-P around 2.8 p.u. (Fig. 8), the overall VA loading

g of UPQC.

828 M. Basu et al. / Electric Power Systems Research 77 (2007) 821–830

load

oas

ttcSfi(p

x

t

bo

5

scwFct

Fig. 8. Total VA

f UPQC-Q would be much smaller than UPQC-P. Hence, thessociated losses of the equipment would be less and overallystem efficiency would be higher.

There is another important observation in Fig. 7 which dis-inctly shows that the loading on the SHI is mutually related tohe load power factor and % voltage sag. For each power factor,ertain percentage of sag creates zero loading condition of theHI. From Fig. 2 it can be observed that for a typical load poweractor condition and supply voltage sag, IC2 can reach zero valuef θ = φ. Following this condition to minimize IC2 w.r.t. x using10), we get the relationship between voltage sag and the loadower factor condition, which is given by

+ cos φ = 1 (21)

If θ < φ, SHI and SEI share the VAR of the load. But if θ > φ,hen SHI current has to increase with the opposite sign to bring

ifbU

Fig. 9. Three phase load a

ing of UPQC.

ack leading power factor to unity, and this increases the loadingf the SHI additionally.

. Simulation results

The analysis of UPQC control schemes has been extensivelyimulated in SABER software, which can implement extensiveontrol schemes. A 400 V (L–L) three-phase three-wire systemith non-linear diode bridge rectifier load has been considered.ig. 9 shows the typical three phase load currents and supplyurrents. It is seen clearly that the quasi square wave shapes ofhe load current, with high THD of 24% do not reflect in the

ncoming supply current. The hysteresis controller of SHI hasorced the input current to be sinusoidal and the THD has beenrought down within 5%. This control is equally effective inPQC-P and UPQC-Q.

nd supply currents.

M. Basu et al. / Electric Power Systems Research 77 (2007) 821–830 829

Fig. 10. Load, supply and injected voltages of phase A, under normal and 20% supply voltage sag condition.

le un

aioisf

aia

ct

gpCaps3pva

Fig. 11. Load voltage and supply voltage profi

Fig. 10 explains the operation of UPQC-Q under 20% bal-nced supply voltage sag. The load voltage, supply voltage andnjected voltage of phase A are plotted. The harmonic spectraf load voltage remain satisfactory (THD within 3.3%). Thenjected voltage maintains a quadrature relationship with theupply voltage as per the control scheme and can be verifiedrom Fig. 10.

Fig. 11 explains the operation of UPQC-P under 20% bal-nced supply voltage sag for a duration of 0.15 s. There is annstantanesous undershoot at the instant of occurrence of sag

round 14%, which cannot be avoided.

The additional advantage of UPQC-P type control is that itan mitigate unbalanced voltage sag. Figs. 12 and 13 presenthe performance of UPQC-P for unbalanced voltage sag miti-

fv5m

Fig. 12. Load voltage and supply voltage profile u

der normal and 20% balanced sag condition.

ation. In Fig. 12 it is found that at t = 0.1 s, the peak of threehase voltages become 300, 275 and 250 V in phases A, B and, respectively. But the lower trace of load voltages are balancednd are maintained to the desired value of 230 V (rms) (325 Veak). From Fig. 13, it is found that when a supply voltageag occurs, q-component voltage becomes ac peak to peak of5 V with frequency 100 Hz, o-component voltage becomes aceak to peak of 50 V with frequency 100 Hz. The d-componentoltage is seen to be reduced by 60 V with a superimposed volt-ge ripple of 35 V (peak to peak, with frequency 100 Hz). It is

ound that after series injection of voltage by the UPQC, the loadoltage harmonic spectra remain within IEEE specified limit of% THD. Thus, the simulation results show satisfactory perfor-ance of UPQC-P.

nder normal and unbalanced sag condition.

830 M. Basu et al. / Electric Power Systems Research 77 (2007) 821–830

oltage

6

aqdhsvstsatvptaias

cumdbpnS

is

vAita

R

Fig. 13. d–q–o component of v

. Conclusion

The present paper investigates the performance of UPQCs a suitable interfacing equipment for enhancement of poweruality. Two control schemes have been analysed based on theifferent voltage compensation schemes of the SEI. UPQC-Qas the advantage of VAR sharing between the two compen-ators. The SEI, while injecting voltage to mitigate the supplyoltage sag, shares a part of VAR of the load and does not con-ume any active power. But at higher load power factor (>0.9),he loading requirement of UPQC-Q is quite high due to exces-ive high SHI rating. The SHI rating of the UPQC-Q increases athigher rate to compensate the effective leading input power fac-

or created by quadrature voltage injection. This shifts the loadoltage angle, seen from the utility side. Therefore, for higherower factor loads UPQC-P rating would be substantially lowerhan that of UPQC-Q. Also UPQC-Q cannot compensate unbal-nced voltage sag. But for applications where VAR demands very high UPQC-Q could be a potential control scheme forction, as it can effectively reduce the input power factor angleeen from the utility side.

The SEI control scheme of UPQC-P is based on d–q–oomponent analysis. UPQC-P can mitigate the supply voltage-nbalance problem also besides voltage sag as the individualodulating signals can vary in phase because they would be

irectly derived from the d–q–o component analysis. Underalanced voltage sag condition, the load voltage angle after com-ensation is not altered. Therefore, the SHI of the UPQC-P doesot require to compensate any additional VAR created due to

EI control action.

Comparative loading analysis has brought useful insightn finding the typical application of the two different controlchemes. The effectiveness of the two control schemes is

[

[

under balanced sag condition.

erified through extensive simulation in the software SABER.s the power circuit configuration of UPQC remains same

n both models, with modification of control scheme only,he utility of UPQC can be optimized depending upon thepplication requirement.

eferences

[1] K. Kowalenko, Distributed Power Offers an Alternative to Electric Utilities,vol. 25, IEEE Press, Piscataway, NJ, 2001.

[2] F. Blaabjerg, Z. Chen, S.B. Kjaer, Power electronics as efficient interfacein dispersed power generation systems, IEEE Trans. Power Electron. 19(September (5)) (2004) 1184–1194.

[3] F.Z. Peng, Editorial: Special issue on distributed power genera-tion, IEEE Trans. Power Electron. 19 (September (5)) (2004) 1157–1158.

[4] T.S. Perry, Deregulation may give a boost to renewable resources, IEEESpectr. (January) (2001) 87.

[5] K.J.P. Macken, et al., Distributed control of renewable generation unitswith integrated active filter, IEEE Trans. Power Electron. 19 (September(5)) (2004) 1353–1360.

[6] H. Fujita, H. Akagi, The unified power quality conditioner: the integrationof series and shunt active filters, IEEE Trans. Power Electron. 13 (March(2)) (1998) 315–322.

[7] M. Basu, S.P. Das, G.K. Dubey, Performance study of UPQC-Q for loadcompensation and voltage sag mitigation, in: IEEE IECON 02, Sevilla,Spain, 5–8 November, 2002, pp. 698–703.

[8] S.S. Choi, B.H. Li, D.M. Vilathgamuwa, Dynamic voltage restoration withminimum energy injection, IEEE Trans. Power Syst. 15 (February (1))(2000) 51–57.

[9] K. Chatterjee, B.G. Fernandes, G.K. Dubey, An instantaneous reactive volt-ampere compensator and harmonic suppressor system, IEEE Trans. PowerElectron. 14 (March (2)) (1999) 381–392.

10] M. Basu, Some Investigations on Unified Power Quality Conditioner, Ph.D.Dissertation, Department of Electrical Engineering, IIT-Kanpur, India,2003.

11] N.G. Hingorani, L. Gyugyi, Understanding FACTS: Concepts and Tech-nology of Flexible AC Transmission Systems, IEEE Press, 2001.