IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO.

2, APRIL 2008 937

Optimum Control of Selective and Total Harmonic

Distortion in Current and Voltage Under
Nonsinusoidal Conditions
Sincy George, Member, IEEE, and Vivek Agarwal, Senior Member, IEEE

AbstractActive-filter-based power-quality improvement

schemes, in general, focus on the reduction of total harmonic dis-
tortion (THD) and improvements in power factor (PF). This may
not be adequate because standards, such as IEEE-519/IEC 61000,
require that apart from PF and THD, selective (specific) harmonic
distortion (SHD) too, should be controlled. The issue is further
complicated due to the presence of nonsinusoidal current and
voltage waveforms in the power system. Under such conditions,
any attempt to achieve unity PF does not result in harmonic-free
current (or voltage). Similarly, compensation for current (or
voltage harmonics) does not yield unity PF. The best solution to
this tradeoff is the optimization of PF, SHD, and THD. This paper
presents an algorithm which restricts the THD and SHD within the
specified limits, while optimizing the PF. The algorithm uses the
Lagrange optimization technique to minimize the total apparent
power. The average active power demanded by the load dictates
the equality constraint, while the maximum limits on THD and Fig. 1. Generation of nonsinusoidal voltage at the PCC.
SHD lead to inequality constraints for the optimization problem.
All of the analytical, simulation, and experimental results of this
work are presented.
core saturation in transformers due to the presence of the dc
Index TermsNonsinusoidal voltage, optimization, power factor
(PF), selective harmonic distortion (SHD), total harmonic distor- component. The triplen harmonic currents, on the other hand,
tion (THD). add arithmetically, instead of canceling, at the neutral point,
and can go up to 1.73 times the phase current [4]. This results
in the overloading of neutral conductors and requires special
I. INTRODUCTION consideration, as neutral conductors are not provided with
overcurrent protection.

I N recent years, there has been an increased use of

power-electronic systems among the power-supply utilities
and consumers. These nonlinear loads draw nonsinusoidal
current from the source, causing nonsinusoidal voltage drops
across the transmission line and transformer impedances [1].
Additionally, voltage source-type harmonic producing loads
[2], such as diode rectifiers with large filter capacitors, inject
voltage harmonics into the system. This results in nonsinusoidal
supply voltage waveforms at the point of common coupling
(PCC) as shown in Fig. 1. The behavior of each voltage and
current harmonic depends on its order. Harmonics adversely
affect virtually every component in the power system [3].
Though every harmonic component is detrimental, certain
harmonic components are particularly harmful for equipment
or applications. Even harmonics, especially the second order,
cause waveform distortion characterized by asymmetries be-
tween the positive and the negative half cycles. This results in
Manuscript received August 15, 2005; revised June 21, 2007. Paper no.
TPWRD-00477-2005.
The authors are with the Department of Electrical Engineering, Indian In-
stitute of Technology-Bombay, Powai, Mumbai 400 076, India (e-mail: sin-
cyeapen@yahoo.com; agarwal@ee.iitb.ac.in).
Digital Object Identifier 10.1109/TPWRD.2007.916011
0885-8977/$25.00 2008 IEEE
In induction machines, harmonic pairs, such as fifth and sev-
enth, can produce intermediate harmonic current in the rotor. It
also induces mechanical oscillations in a turbinegenerator or
the motor-load system [5]. Under nonsinusoidal voltage condi-
tions, unoriented core material used in rotating machines pro-
duces higher magnetic loss [6].
Realizing the significant impact of these harmonics on the life
and performance of equipment and components connected to
the power system, regulatory agencies have specified limits for
selective harmonic distortion (SHD) in addition to the total har-
monic distortion (THD). Accordingly, standards, such as IEC
61000 [7] and IEEE 519 [8], specify the limit for both individual
and THD in voltage and current. The current THD and SHD
limits specified by IEC 61000 34 (1998) for a balanced 3- ,
low-voltage equipment for a selected range, is given in Table I.
Various techniques have been reported to compensate the har-
monic distortion in current using shunt active filters. Most of
these techniques [9][11], assume a sinusoidal supply voltage
and an attempt is made to achieve a sinusoidal source current.
A few compensation techniques [12], [13] have also consid-
ered the harmonics present in the supply voltage. But there is
not enough literature available for selective harmonic control,
though its need has been highlighted [14].

Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
938 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO. 2, APRIL 2008

TABLE I where is the order of harmonics present in , is the

CURRENT DISTORTION LIMIT FOR EQUIPMENT ( > 16 A PER PHASE) [7] rms value of the th order harmonic component of and can
be determined using the discrete Fourier transform. is given
by

(2)

where is the order of harmonic present in , is the rms

value of the th-order harmonic component of and is the
corresponding phase angle between and .
In order to shape the same as , without any phase
When the supply voltage is nonsinusoidal, any attempt to shift between them, the following conditions need to be
make the current a perfect sine wave, leads to a poor power satisfied.
factor (PF). On the other hand, achieving a unity power factor 1) Similar order of harmonics should be present in both the
by shaping the current waveform similar to voltage, results in waveforms and .
distorted current [12], [13]. Similarly, making the load voltage 2) Ratio of the rms value of each harmonic component of
waveform sinusoidal by using a series active filter, may reduce one waveform to the corresponding component in the
the PF since the current may be distorted. However, making the other waveform should be equal.
load voltage in phase and of the same shape as current may im- 3) Phase angles corresponding to similar harmonic compo-
prove the PF, but the voltage distortion will be greater. Thus, nents in the two waveforms should be zero.
there is a tradeoff between the requirements of high PF and low Accordingly, should be modified to as follows:
THD. The best solution to this tradeoff is the optimization of
PF and THD (i.e., shape the source current or the load voltage
(3)
in such a way that the PF is optimum while keeping THD within
the limit). However, during this process, some of the individual
harmonics may exceed their limit. where is the rms value of the modified and is given by
This paper presents a flexible algorithm applicable to shunt
or series active filters under nonsinusoidal supply voltage and (4)
current conditions to achieve the following objectives.
where is the harmonic ratio of the rms value of the harmonic
1) Restricting the individual harmonic components of
component of to the corresponding component of , and
source current or load voltage according to the require-
varies from 1 to i.e.,
ments.
2) Limiting the THD in current or voltage waveforms ac-
(5)
cording to the requirements.
3) Optimization of PF in conjunction with 1) and 2).
If the ratio is the same for all harmonics, the shape of the
The proposed algorithm fully complies with the requirements of
waveforms (1) and (3) will be the same. Similarly, it is possible
IEEE-519 or IEC-61000 or any other standard for that matter.
to shape the same as . Also, by controlling the values
of , the shape can be controlled. This principle is used in
II. PROPOSED ALGORITHM FOR TOTAL
this paper to shape the current or voltage to the desired one. The
AND SELECTIVE HARMONIC CONTROL
most appropriate values for are obtained using the Lagrange
An algorithm is proposed for shunt and series active filter for optimization technique.
the compensation of selective and THD in current and voltage,
respectively, using the Lagrange multiplier optimization tech- B. Application of the Technique to Shape Voltage or Current
nique [15]. The algorithm also optimizes the PF while ensuring When a nonlinear load is connected to a nonsinusoidal supply,
THD within the specified limit in addition to selective harmonic due to the characteristics of supply and load, similar or dissim-
control. A compensation algorithm applicable to both shunt and ilar order of harmonics may be present in the voltage and cur-
series active filters is described. rent. With this assumption, let the supply voltage contain
harmonics of order and and the current contain har-
A. Generalized Algorithm for Shaping the Waveforms
monics of order and . The instantaneous voltage and cur-
Let and represent two nonsinusoidal waveforms of rent, containing these harmonics, are expressed by
different shapes with a certain phase angle between each har-
monic component that is present in them. This can be repre-
sented as

(1) (6)

Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
GEORGE AND AGARWAL: OPTIMUM CONTROL OF SHD AND THD 939

where is the objective function, represents the equality con-

straint, and , , and , etc., represent the inequality con-
straints. , , , , and represent unknown quantities. More
constraints can be added to this equation. However, it will in-
(7)
crease the number of unknowns and, therefore, the computa-
tional time will increase.
where and are the rms values of the th components For a given active power, the PF can be improved by min-
of voltage and current, respectively, is the arbitrary angle of imizing the total apparent input power [16]. The definition
the th harmonic component of voltage, is the phase angle of apparent power under nonsinusoidal conditions is a matter
of the th harmonic component of current, and of debate. However, in this work, we have considered
and the summation do not proceed with integer steps in any set where and .
of harmonic components. Further, the total PF of the system is considered which is due
A shunt active filter is used for limiting the SHD and THD in to distortion as well as the displacement PF. is taken as the
current. The desired source current is obtained from (6) objective function in the case of both shunt and seriesAFs. For
and (7) by setting and retaining the order of harmonics a shunt AF, the objective function is constructed using (6) and
to be the same as that of the supply voltage (8) as

(11)

(8) where harmonics and have been combined and denoted

by .
Substituting , the objective function
where is the rms value of the th harmonic component of for a shunt AF is given by
the desired source current. It is given by ,
where is the control variable defined as the admittance
of the th-order harmonic component which is the ratio of the (12)
th-order harmonic component of the source current after com-
pensation to the corresponding voltage component.
Similarly, a series AF can be used for the compensation of Similarly, the objective function for the series AF can
voltage harmonics. Since reactive power cannot be compensated be formed by using and as
by a series AF, assume that the current (7) is in phase with the
supply voltage (6) for simplicity [i.e., make in (7)]. The
desired load voltage is computed in a similar way as (8) (13)
and is given by
2) Equality Constraint: The source current is calculated in
such a way that it should supply only the mean value of the in-
stantaneous real power demanded by the load after compensa-
tion. The compensating circuit supplies the remaining power.
Thus, the equality constraint is formed by applying the con-
(9)
dition that the mean value of instantaneous power demand be-
fore and after compensation should be the same.
where As the displacement angle between the supply voltage and
where is the control variable defined as the current is zero after compensation, the mean value of the instan-
impedance of the th-order harmonic component which is taneous real power in the case of the shunt active filter is given
the ratio of th-order harmonic component of load voltage after by
compensation to the corresponding load current component.
Since optimization technique computes the most suitable
value of for a shunt or series AF, the Lagrangian-multiplier (14)
technique is used that optimizes the PF while limiting the SHD
and THD in current or voltage.
Hence, for a shunt active filter, the equality constraint
1) Objective Function : According to the Lagrange op-
is given by
timization technique, an augmented Lagrange function can be
written as
(15)
(10)

Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
940
Similarly, for a series active filter, the equality constraint
is written as
3) Inequality Constraints: There are two inequality con-
straints to be met. The SHD and THD in current or voltage
should be within the specified limits.
1) Total harmonic distortion.
Let the THD limit in current be and the SHD limit
be . Then, the inequality constraint for the shunt
active filter can be determined as below
Similar expression can be obtained for a series active
filter as follows:
2) Selective harmonic elimination.
Let the th-order current harmonic component be lim-
ited and the corresponding harmonic distortion limit be
. The individual harmonic distortion factor is the
ratio of the magnitude of the corresponding harmonic
component to the fundamental component of the current.
For a shunt active filter, inequality constraint
corresponding to the selective harmonic distortion can
be formed as
A similar expression can be written for a series active filter as
4) Lagrange Function: The objective is to minimize , given
the equality constraint and the inequality constraints
Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO. 2, APRIL 2008
, and . In general, in order to shape to a desired
value, using (1) and (3), the augmented function is written as
(16)
(25)
where , , and are the unknown quantities and
and are the respective limits specified by the standards,
which the THD and SHD cannot exceed.
By applying the necessary conditions for the constrained
local minima of [15], the unknowns can be computed.
(17)
After computing , the rms value of the harmonic component
of can be obtained by using (4). Using the inverse DFT,
can be computed.
(18) Similarly, the desired source current and desired load
voltage can be determined and the compensation current
and compensation voltage for shunt and series filter, respec-
tively, can be computed as
(19) (26)
(27)
(20) III. SIMULATION RESULTS
To verify the algorithm for selective and total harmonic
control using shunt or series active filters presented in the
preceding sections, a balanced, 3- , 4-wire, 415-V, 50-Hz,
trapezoidal voltage supply, having 21.02% THD and 20.5%
third harmonic distortion is considered. To verify the perfor-
mance of the algorithm, simulation studies have been carried
out for the following three cases.
, Case 1) Verification of the algorithm to limit THD and SHD
in current per IEC-61000 34 by using a shunt active
filter.
Case 2) Capability of the algorithm to limit even harmonic
distortion in current, using a shunt active filter.
Case 3) Verification of the algorithm to limit THD and SHD
(21) in voltage per IEC 61000 2-2, using a series active
filter.
1) Case 1: Fifth harmonic distortion limited to 14%
(22) and THD limited to 16% using a shunt active filter.
A combination of 3- diode rectifiers and 3- resistive net-
(23) works act as the load which consumes around 55-kW power.
The PF of the circuit without compensation is 0.95, while the
current THD is 25.67%, which is greater than
(24) . MATLAB-SIMULINK software is used for the verifica-
tion of the algorithm. Waveforms of the nonsinusoidal supply
voltage and load current for phase-a of the sample power
system considered are shown in Fig. 2.
The harmonic spectra of the supply voltage and source cur-
rent (the same as load current) before compensation are shown
in Fig. 3. The voltage spectrum in Fig. 3(a) shows that the third
harmonic component is predominant and is about 20% of the
GEORGE AND AGARWAL: OPTIMUM CONTROL OF SHD AND THD
Fig. 2. Phase-a waveforms of supply voltage (v ) and load current (i ) be-
fore compensation for Case 1) (Current THD = 25:67%, fifth harmonic com-
ponent of current = 16:83%, and PF = 0:953).
Fig. 3. Harmonic spectra of (a) supply voltage and (b) source current before
compensation for Case 1) (the inset of figure (b), indicates the current har-
monic spectrum excluding the fundamental).
fundamental voltage. However, 5th and 7th harmonic compo-
nents are predominant in the load current. For clarity, the mag-
nitude of the higher order harmonics is plotted separately and
shown as the inset of Fig. 3(b). It can be seen that the fifth
harmonic component of the current is around 16.83% and the
7th harmonic component is 11.05% of the fundamental with a
THD of 25.67%. Therefore, according to the standard (Table I),
the THD as well as the individual harmonic distortion in current
exceed their limits.
Fig. 4 shows the harmonic spectrum of the desired source cur-
rent. It shows that the THD and the fifth harmonic distortion are
within the set limits. However, the third harmonic component of
current is increased after control. The specialty of the algorithm
is that in case the third harmonic component exceeds the speci-
fied limit, by adding one more constant inequality, this compo-
nent can also be limited.
2) Case 2: The second harmonic distortion in the
source current is limited to 3% and THD is limited to
5% using the shunt active filter.
Considering the severe impact of even harmonics on the
system, IEEE 519 requires that even harmonics be limited
to 25% of the odd harmonic limit. To show the effectiveness
Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
941
Fig. 4. Harmonic spectrum of the desired source current with I = 14%
and I = 16% for Case 1).
Fig. 5. Waveforms of load current (i ), desired source current after compen-
sation (i ), and the reference compensation current (i ) with I = 3%
and I = 5% [Case 2)].
of the proposed algorithm in restricting the second harmonic
component, a half-wave rectifier load of 10 kW is connected to
the same supply as considered in Case 1).
Fig. 5 shows the waveform of load current, desired source cur-
rent after compensation, and the compensation current, which
limits the current THD to 5% and the second current harmonic
distortion to 3%.
Before compensation, the source current (which is the same
as the load current) flows only during the positive half cycle of
the supply due to which even harmonics are injected into the
system. However, after compensation, the current is uniformly
distributed during a positive and negative half cycle with less
distortion, thus achieving an optimum PF. Harmonic spectra of
the source current before compensation and the desired source
current are shown in Fig. 6. Harmonic spectrum of the load cur-
rent in Fig. 6(a) indicates that in addition to odd harmonics,
even harmonics are also present in the system. It shows that the
2nd and 4th harmonic components are around 50% and 20%,
respectively, of the fundamental component. Fig. 6(b) clearly
proves that using the proposed algorithm, it is possible to limit
the THD to any specified value and, at the same time, individual
harmonics (second harmonics in this case) can be limited to the
specified value.
Using the data of Figs. 5 and 6, it is observed that the second
harmonic component is reduced from 50% to 3% of its fun-
damental component after compensation. The current THD is
found to be less than 5%. It can be verified that the algorithm is
942 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO. 2, APRIL 2008

Fig. 8. Waveforms of supply voltage v ( ) before compensation and the de-

( )
sired load voltage v in phase-a with V = 5% and V = 8% [Case
3)].

Fig. 6. Harmonic spectra of: (a) load current and (b) source current after com-
pensation with I = 3% and I = 5%
for Case 2).

Fig. 9. Harmonic spectrum of the desired load voltage with V = 5% and

V = 8%[Case 3)].

TABLE II
COMPUTED VALUES OF HARMONIC DISTORTION AND HARMONIC COMPONENT
OF LOAD VOLTAGE BEFORE AND AFTER COMPENSATION FOR CASE 3

Fig. 7. Phase-a waveforms of the desired and the resulting source current after
compensation for I = 3%and I = 5%
in current (Case 2).

capable of limiting the current THD and the second-order cur-

rent harmonic distortion to within the specified values of 5%
and 3%, respectively. At the same time, the PF too is improved
to 0.986 from 0.808.
After computing the reference compensating current, the
pulses for the inverter circuit are obtained using hysteresis cur-
rent control. A 3- , three-leg inverter is used for injecting the
compensation current into the system Fig. 7 gives the phase-a
waveforms of the desired source current and resulting source
current ( -resultant) after compensation for Case 2). It can be
seen that the actual current tracks the desired source current. puted values of the individual harmonic distortion and magni-
This is also applicable to Case 1). tude of harmonic components of the supply voltage before and
3) Case 3: Limiting third harmonic distortion to 5% after compensation.
with in load voltage using the series active filter. It indicates that the third harmonic distortion in the load
The proposed algorithm is also verified for compliance as per voltage is reduced from 20.5%, to within the set value of 5%.
the IEC 61000 2-2 limit of 8% on and 5% on . Even the ninth harmonic distortion is reduced from 2.19%
Fig. 8 shows the waveforms of the phase-a supply voltage to 0.29%. Though 5th and 7th harmonic distortion is slightly
(the same as the load voltage before control) and the desired load increased, it is within the specified limit as per the standard.
voltage , obtained using the proposed algorithm. The refer- Table III lists the PF, voltage THD, and third ordervoltage
ence compensation voltage is computed by subtracting the harmonic distortion before and after compensation.
desired load voltage from the supply voltage. Fig. 9 shows the
harmonic spectrum of the desired load voltage. A comparison IV. HARDWARE IMPLEMENTATION OF THE ALGORITHM
of Figs. 3(a) and 9 shows that the third harmonic distortion is The experimental verification of the proposed algo-
reduced from 20.5% to less than 5%. Table II shows the com- rithm is carried out on an experimental prototype with a

Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
GEORGE AND AGARWAL: OPTIMUM CONTROL OF SHD AND THD 943

TABLE III
COMPARISON OF POWER FACTOR AND VOLTAGE THD BEFORE AND AFTER
COMPENSATION [CASE 3)]

Fig. 11. Experimental waveforms of load current (i ), source current after

compensation (i ), and the compensation current (i ) for I = 10% and
I 0
= 5% [x axis: time 5 ms=div ; y axis: i 0
2 A=div , i 0
2 A=div ,
i 0 1 A=div ].

Fig. 10. Experimental waveforms of supply voltage (v ) and load current

(i ) without any compensation [x axis: 5 ms/div, y axis: v 20 V=div ; 0
i 0 2 A=div ].

TMS320LF2407A DSP kit used for computation and control

purposes. The DSP performs optimization according to the
algorithm presented in Section II and generates an appropriate
reference compensation current (or voltage) required by the
shunt (or series) active filter for the desired compensation.

A. Selective and Total Current Harmonic Control Using Shunt

Active Filter (With and )
In order to validate the proposed control algorithm, a non-
sinusoidal, 3- , 50-Hz, 45-V power supply was fabricated. A
six-pulse diode rectifier with a 40- resistor forms the nonlinear
load. The waveforms of the supply voltage and load current in
phase-a are shown in Fig. 10.
It is found that the source current THD is around 21.76% and Fig. 12. Harmonic spectra of the experimental waveforms of source current. (a)
Before compensation. (b) After compensation for I = 10% and I =
the fifth harmonic distortion in the source current is 20.7%. The 5%.
PF is 0.967. The proposed algorithm is tested for limiting the
current THD to 10% and 5th harmonic distortion to within 5%.
The corresponding reference compensation current is obtained
using the DSP, and pulses for the inverter devices are gener-
ated using hysteresis current control. The source current before
compensation (the same as the load current), the compensation
current, and the resulting source current for the aforementioned
set limit is as shown in Fig. 11.
It is found that the THD in source current after compensa-
tion is 10.8% and the fifth harmonic distortion is 4.6%. The PF
is optimized to 0.977. The waveform data are loaded into the
MATLAB workspace and the harmonic spectrum is plotted as
shown in Fig. 12. It is evident from the spectrum that the fifth
Fig. 13. Experimental waveforms of the supply voltage (v ), load current
harmonic distortion as well as THD in source current is within (i ), and the load voltage after compensation (v ) in phase-a for V = 3%
the limit after control. and V = 1%.

B. Selective Voltage Harmonic Control Using Series Active

Filter (with and ) voltage after compensation for the 3% voltage THD is shown in
The experimental verification is performed on the same Fig. 13. It is found that the voltage THD improves from 7.7 to
power system which is considered for the shunt active filter. 2.88% and the third harmonic distortion is reduced from 3.36
The waveform of the supply voltage, load current, and the load to 1.03% after compensation. The PF is optimized to 0.9607.

Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
944
Fig. 14. Waveforms of desired load voltage and desired source current for
V = 3% and I = 10%
using a series-shunt active filter.
V. CONCLUSION
A versatile algorithm has been proposed to limit the total and
individual harmonic distortion in current or voltage under non-
sinusoidal supply voltage and current conditions, using a shunt
or series active filter, respectively. In addition to controlling the
harmonics to meet any specified standard requirements, the pro-
posed algorithm also optimizes the total PF of the system which
includes distortion as well as displacement PF.
The specialty of the algorithm is its flexibility to meet any
requirement. Even though the control of only one selective har-
monic component has been considered in this paper, it is pos-
sible to control more components by adding more inequality
constraints. It is also possible to change the objective function or
equality constraints as per the requirement. The algorithm takes
into account the harmonics present in both the voltage and the
current, which is a more practical situation. The Lagrange opti-
mization technique is used to solve the tradeoff between the PF
and THD. The technique optimizes the PF while satisfying the
constraints such as THD and the selective harmonic distortion
limit. The algorithm is applicable to 1 and 3 under sinu-
soidal and nonsinusoidal conditions.
An important feature of this algorithm is that it is also ap-
plicable in the case of a combination of series and shunt active
filters (series-shunt active filter) to generate reference compen-
sation voltage and current. It is possible to completely eliminate
the harmonics present in the voltage and current using a series-
shunt active filter and, as such, there is no need for any optimiza-
tion. However, optimization will be needed, if it is sufficient to
limit the voltage and current harmonics to within a nonzero limit
imposed by standards, instead of their complete elimination. To
show the applicability of the algorithm, simulation is carried out
on the same supply as in Case 1) (Section III). The desired load
voltage and source current, obtained with voltage THD limited
Authorized licensed use limited to: University of Waterloo. Downloaded on October 17, 2008 at 18:51 from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 23, NO. 2, APRIL 2008
to 3% and current THD limited to 10%, are shown in Fig. 14.
More details of this aspect will be presented in a future paper.
REFERENCES
[1] H. Akagi and H. Fujita, A new line conditioner for harmonic compen-
sation in power systems, IEEE Trans. Power Del., vol. 10, no. 3, pp.
15701575, Jul. 1995.
[2] F. Z. Peng, Application issues of active power filters, IEEE Ind. Appl.
Mag., vol. 4, no. 5, pp. 2130, Sep./Oct. 1998.
[3] V. E. Wagner, Effect of harmonics on equipment, IEEE Trans. Power
Del., vol. 8, no. 2, pp. 672680, Apr. 1993.
[4] IEEE Recommended Practice for Powering and Grounding Sensitive
Electronic Equipments, IEEE Std. 1100-1992.
[5] IEEE Recommended Practice for Electric Power Distribution for In-
dustrial Plants, IEEE Std. 141-1993.
[6] M. S. Lancarotte and A. de A Penteado, Jr., Estimation of core losses
under sinusoidal or non-sinusoidal induction by analysis of magnetiza-
tion rate, IEEE Trans. Energy Convers., vol. 16, no. 2, pp. 174179,
Jun. 2001.
[7] Electromagnetic Compatibility (EMC)Part 34 LimitsLimitation
of Emission of Harmonic Current in Low-Voltage Power Supply Sys-
tems for Equipment With Rated Current Greater Than 16 A, IEC Std.
61000-3-4-1998.
[8] IEEE Recommended Practice and Requirements for Harmonic Control
in Electrical Power Systems, IEEE Std. 519-1992.
[9] A. Cavallini and G. C. Montanari, Compensation strategies for shunt
active-filter control, IEEE Trans. Power Electron., vol. 9, no. 6, pp.
587593, Nov. 1994.
[10] A. Chandra, B. Singh, B. N. Singh, and K. Al-Haddad, An improved
control algorithm of shunt active filter for voltage regulation, harmonic
elimination, power-factor correction, and balancing of non-linear
loads, IEEE Trans. Power Electron., vol. 15, no. 3, pp. 495506, May
2000.
[11] Z. Li, H. Jin, and G. Joos, Control of active filters using digital signal
processors, in Proc. IECON, 1995, pp. 651655.
[12] S. M.-R. Rafiei, H. A. Toliyat, R. Ghani, and T. Gopalarathnam, An
optimal and flexible control strategy for active filtering and power
factor correction under non-sinusoidal line voltages, IEEE Trans.
Power Del., vol. 16, no. 2, pp. 297305, Apr. 2001.
[13] S. George and V. Agarwal, A novel technique for optimising the har-
monics and reactive power under non-sinusoidal voltage conditions,
in Proc. 28th Annu. Conf. IEEE Industrial Electronic Society, 2002, pp.
858863.
[14] J. A. Orr and A. E. Emanuel, On the need for strict second harmonic
limits, IEEE Trans. Power Del., vol. 15, no. 3, pp. 967971, Jul. 2000.
[15] A. Ravindran, T. P. Don, and J. S. James, Operations Research Prin-
ciples and Practice, 2nd ed. New York: Wiley, 2003, pp. 524542.
[16] W. Shepherd and P. Zand, Energy Flow and Power Factor in Non-
sinusoidal Circuits. Cambridge, U.K.: Cambridge University Press,
1979.
Sincy George (M95), photograph and biography not available at the time of
publication.
Vivek Agarwal (SM01), photograph and biography not available at the time
of publication.