Sie sind auf Seite 1von 8

E TEP

Instantaneous SinwoidalandHarmonicActiveand
Deactive Currents in Three-phahasePower Systems
J. L. Willems

Abstract
In this contribution a new decomposition of sinusoidal three-phase currents is proposed, The main feature is
that it leads to the compensation of non-active currents partly by elements without energy storage and partly
by means of linear elements with energy storage (inductors and capacitors). The proposed decomposition is
compared with existing procedures, such as those developed by Akagi-Nabae and by Czarnecki. The possible
generalization of this concept to distorted voltages and currents is discussed.

1 Introduction reactive power compensation are briefly recalled first.


Consider a three-phase power system with positive-se-
In a power system delivering electrical energy to a quence voltages. The voltage phasor vector is
load the currentsconsist of a part necessary for the trans-
fer of power and another part that does not contribute to
the power transfer. The latter part causes additional loss
and is responsible for a decrease of the power factor. For
the analysis of the performance and the efficiency of the
power transmission it is therefore important to be able to
distinguishthe active and the non-active ordeactive cur-
rr = E.,I [ =
with a = &2d3. The instantaneousvoltages are given by
(1)

rent components and to design procedures for compen- (I


I-L1 145 cos(wt + q,,)
sation of the non-active current components.
To analyse the performance of the power uansmis- u (t) = lcLll-b
cos(wt + (pLI- 2x13)
sion for sinusoidal as well as distorted three-phase cur- LL1 U I
d cos(wt + (pLI + 2x13)
rents, various decompositionshave been proposed. Two
fundamentalapproachesare due to Czarnecki [4-61 and where
to Akagi et al. [l]; both decomposition techniques are
useful with respect to the analysis of overloadingdue to -u
V L I -I-LI le’9~l.
non-active current components as well as the compen- A load is connected to the power system. The load
sation of these components. can be reactive and unsymmetrical. Let the load current
An elegant method for analysing both decomposi- phasor vector be
tion procedures has been developed by F e m m and co-
authors [ 10, 121. In the present contribution the merits
and drawbacks of both decomposition techniques are
briefly discussed. On the basis of that discussion a new
decompositionprocedure is proposed for the case of si-
nusoidal three-phase currents and voltages. This new
kL3 1
technique combines interesting features from the Czar- The active power delivered to the load is given by
necki and the Akagi-Nabae decompositionprocedures;
compensation can be achieved partially by elements p, = Re(U,,lt, +U,lt* +UL,LL} (3)
without energy storage and partially by linear inductors
and capacitors.Moreover, the possible generalization of where “Re” denotes the real value and * the complex con-
the concept to the case of distorted currents and voltag- jugate of a complex number. The line losses are propor-
es is discussed. tional to the square of the rms line current:
lllIl’= ILLII?+ &’I? +llL#
where 11~Il
denotes the length of a vectorx defined by
2 Ciassical Approach to Compensation in 11x_112= x_Tx_*.
Sinusoidal Steady State
The efficiency of the power transmission may be
To be able to compare the proposed decomposition characterized by the ratio of the line losses to the active
technique to existing procedures, the classical ideas of power, or by the power factor which is equal to the ratio

ETEP Vol. 4,No. 5, September/October 1994 335


ETEP
of the active power Pa to the apparent power S. The ap- where Q denotes the reactive power delivered to
parent power is defined by the load
s= ll~II.ll~ll (4) - the unsymmetrical current which could be split up
with further into the negative-sequenceand the zero-se-
quence currents.
Note that Czarnecki considers more general situa-
It is the purpose of compensation to increase the tions than the cases analyzed in the present discus-
power factor or equivalently to reduce the transmission sion, namely problems with distorted voltages and
loss, without altering the active power delivered to the currents.
load.
To analyse the efficiencyof the power transmission, The active power considered above corresponds to
in classical power system theory the current is decom- the average real power. The active current 1, is hence a
posed into an active current component 1, and a deac- current which yields the same average power as the load
tive current component Id.The active current is defined current. It is the current with the smallest rms value hav-
as : ing this property, and hence it realizes the smallest line
losses and the largest power factor. Note that 1, and 1,
correspond to the projection of the current vector 1
(5) along U and j P , respectively. This interpretation is
Lr,, J interesting for the comparison with results discussed in
later sections.
with G = f',/ll~ll', and
I d = 1- 1,. (6) 3 Akagi-Nabae's Current Decomposition
The following properties are readily derived:
The classical approach to deactive current compen-
The active power corresponding to the active current
sation, as discussed in the previous section, requires
1, is equivalent to the active power corresponding to
the load current 1: compensating elements with energy storage, since the
instantaneous power correspondingto the active current
Re{U71*)=Re (U'11). vector 1, is different from the instantaneous power cor-
The deactive current Idcorresponds to zero active responding to the load current vectorl. Although the de-
active current corresponds to zero active (or average)
power:
power, the associated instantaneouspower does not van-
Re { U r [ i )= 0. ish. Parallel capacitors and/or inductors, used as com-
The line losses are proportional to pensating elements, realize the storage of electrical en-
ergy; the average power delivered is zero.
11111=
' 111_,11'+ llldll'. One may wonder if (partial) compensation, that is
reduction of the rms deactive line currents and corre-
Hence the line losses can be split up into a part due to
the active current and a part due to the deactive cur- sponding line losses, is possible without energy storage.
This may be desirable for economical and/or technical
rent. Reducing the latter without altering the former
reasons [3,9]. To analyse this question, Akagi et al. [ 11
is the aim of compensation.
introduced new concepts, the instantaneous active
If an element taking the current -Id, or equivalently power and the instantaneous reactive power, and corre-
delivering the current Id,is connected parallel to the sponding currents, to see which current components can
load, then the power system has to deliver only the ac- be compensated without affecting the instantaneous
tive current I,. The active power delivered by the power. Such compensation (corresponding to current
power system is unchanged. The parallel element, and loss reduction) can be achieved by means of (power
called the compensating element, does not take nor electronic) elements without energy storage.
generate active power. However, the instantaneous In particular for three-phasenetworks, the current is
power delivered by the power system is changed by split up into the instantaneousactive currentip(t)and the
the compensation. instantaneous deactive Current i,(t), such that [ 1,2, 101:
It can be shown that the compensatingelement deliv- +
i ( t ) = ip(t) i,(t),
ering the current Idto the load can always be realized
u (t)Ti(t)= u (t)TiP(t),
by means of a linear reactive element (consisting of
inductors and capacitors) [ 1.51. u ( t f i & ) = 0.
Czarnecki [4-6] analyses the causes of the overload- Moreover the decomposition should be such that
ing of a power line in detail; he shows that for the sit- the current component i , ( t ) is the smallest current hav-
uation considered here the overloading current or the ing these properties. Actually Akagi and coauthors use
deactive current corresponds to the sum of the terminology "instantaneous reactive current". Here
the terminology has slightly been changed to avoid con-
- the reactive current L,=jBCJ (7) fusion with the concept of reactive current in the sinu-
with B=-Q/IIUII? soidal situation.

336 ETEP Vol. 4.No. 5, September/October 1994


ETEP
The analysis of the problem yields the following re- For this example one obtains
sults [7,8,13. 14, 161:
- i , ( t ) is the orthogonal projection of the vectori(t) on
the vector u ( t ) ,
- i,(r) is the component of the vectori(f)orthogonal to
the vector u(r).
Explicitly:

i
Note that

!
I191u‘7[1-cos(2ot)] cos(ot)
IIi(t)II’= IIi,(r)tI’ +IIi,(t)f. (9) i, (f)= p’(dqcos(wt+2x/3)-cos(2wt) cos(or-2x/3)] .
The current component i, can be annihilated by a lj*1&[cos (wt-2x~)-cos(2wr)cos ( o r + 2 x ~ ) ]
parallel compensating element which not only does not
require energy but also no energy storage. The compen- One readily checks
sating element should deliver the current i,. It reduces
u (t)Ti,(r>= 0.
the line current and the line losses as much as it is pos-
sible by means of non-active elements without energy The currents i , ( t ) and i , ( r ) are clearly non-sinu-
storage. soidal.
A further decomposition is possible The above discussion and the example clearly show
that on the one hand the compensation of the instantane-
i , ( f ) = i,(r) + ipl(r) (10) ous deactive current has the advantage of only requiring
where i , ( t ) is the active current vector corresponding to elements without energy storage. However, it has the
the phasor 1, discussed in the previous section. The cur- drawback of introducing harmonics; moreover, the ad-
rent i*(r) is the component in the instantaneous active ditional compensation of the deactive current contained
current which corresponds to non-zero instantaneous ac- in the instantaneous active current is not possible by
tive power but to zero average active power. One readi- means of (linear) inductors and capacitors. In the next
ly checks section it is therefore discussed which sinusoiduf deac-
tive current can be compensated using elements without
energy storage.

where T denotes the period. Hence the rms current and 4 New Decomposition Technique
the average line loss can further be reduced by compen-
sating i @ ( t ) :this obviously requires elements with ener- From the discussion of the previous section the
gy storage. question arises which is the largest sinusoidal current
An important feature which however has not been (part of the instantaneous deactive current) that can be
emphasized in the literature is that sinusoidal u ( t ) and compensated by elements without energy storage. Then
i(t) do not necessarily lead to sinusoidal i9(r)and i , ( t ) , the remaining part of the deactive current is sinusoidal
ascanbeclearly seenfromeq. (8).Thus,ifr,(t)andi,(t) and can be compensated by means of linear inductive
are non-sinusoidal, the compensation of i, (f) necessar- and capacitive elements [ 151 which obviously require
ily requires (power electronic) non-linear elements, and energy storage. This leads to the following problem
introduces harmonics in the line current. Also, and this statement [ 181: find a decomposition of the current
is probably more important, the compensation of the
current i,,,,(f), requiring elements with energy storage, i ( t ) = i,(t) + iqs(t) (11)
cannot be realized by means of linear inductors and ca- with the following properties:
pacitors.
This observation can readily be illustrated by means - The currents i , ( f ) and i,(t) are sinusoidal. Let the
of a simple example. Let the voltage and the load current corresponding phasors be denoted as Ipsand I,.
be given by - The instantaneous power corresponding to i, (t) must
be zero for all t:
u (t)riqs(r) = 0.
- From all decompositions satisfying the two conditions
above, that one should be selected for which the current
ip,(r) is as small as possible. A measure for this is the
rms value of i, ( r ) or equivalently the magnitude of 1,.
In other words. one wants to determine the maximal
current or line loss reduction which can be realized with-

ETEP Vol. 4. No. 5 , SrptembedOctober 1994 337


E TEP
out energy storage. but also without destroying the sinu- The currents I, and are orthogonal in the mean
soidal character of the line current. square sense. The expression above shows how much re-
The instantaneous power corresponding to sinu- duction of the line current and the losses is possible by
soidal voltage andcurrent vectorsu (t) andi(t), with cor- compensation of IqSand of I , - -Iqs.
responding phasorsJ L and 1,is given by
p ( t )= u(t)'i(t) = A + Bcos(2wr + p).
5 Balanced Power System Voltages
It consists of a constant component
If the three-phasevoltages are balanced, as in eq. ( l),
A = Re (U'r) the computation of the current Ipsfrom the geometrical
equal to the active power, and an oscillating component interpretation in Lemma 1 is simplified, because the
[ 171: three phasors U ,If and j u are mutually orthogonal.
The current Ips is given by
~ c o( s2 0 1+ #3) = I U ' ~I cos[ 2ur + arg ( Q'L)I.
Hence the instantaneous powers corresponding to
the voltage u (t) and the load current i (t) on the one hand,
and to the same voltage and the current i'(t), with pha-
sot I', on the other hand, are equal if and only if

where "Im" denotes the imaginary pan. The second ex- It is readily checked that 1, corresponds to the same
pression can also be written as instantaneous power as the current I or the current I,.
Re { u'f ) = Re [ U'l'*). Only the first component is necessary for the average
power; it equals the active current I,. The second com-
The third expression is equivalent with ponent ofl, corresponds to zero average power, but it
Re (-jU'l) =Re {-j&!'l') is necessary for realizing the oscillating term in the
power between source and load. It can readily be
or checked that the instantaneouspower corresponding to
1q\ =I - l p
Let the phasor vectors be considered as elements of is zero at all time.
a vector space with three complex entries where the For the case considered in the present section, the
internal product of two vectors x and y is defined as current components -Ip and lqs can easily be expressed
Re { x r y * )The
. above relations express that the currents using the positive-sequence, negative-sequence, and
landf'correspond to the same instantaneouspowerwith zero-sequence components of the load current. Let this
the voltage U if the internal products of 1 with U , andu decomposition for the load current phasor be:
jLf are respectively the same as the internal products of
l'with 111, L f andj L f .This leads to the geometrical tech- I=-II+ 1 2 + 1 0 .
nique expressed in the lemmabelow, which is used in the Moreover let 1 be decomposedinto an active and a
next two sections to derive the current 1,. reactive component, or otherwise stated an in-phase and
a quadrature component with respect to the voltage U :
Lemma 1: Thephasorl, of the smallestsinusoidal cur-
rent which yields the same instantaneouspower as 1for I1 = I l a + 1 l r .
a voltage 11 is given by the projection of I on the sub- Then it is clear that
space spanned by the vectors p,_V* and jug.
11/112=11~,,112+11~~,112+ 1112112+ Il~o112.

The sinusoidal currentsi,(r) andi,(t) with the pha- The powers corresponding to these current compo-
sorsI,andI,,=I-&,,arerespectivelycalledtheinsran- nents are as follows:
taneous sinusoidal active current and the instantaneous
sinusoidal deactive current, in contrast with the instan-
- Ilacorresponds to a constant instantaneous power. It
is the same as the current 1, discussed previously.
taneous active current and the instantaneous deactive
current, as defined by Akagi-Nabae [ 11. - I 2 corresponds to an oscillating instantaneous
With these concepts compensation can be realized power component with zero average value. It is
as follows. The current iqs(t) is compensated by an ele- equal to I ps - I,.
ment without energy storage; the current i p ( t )is funher
reduced to i.0) by means of linear elements with ener- - flrand l o correspond to zero instantaneous active
power. Their sum equals I,+.
gy storage (capacitors andor inductors). The currents
satisfy Hence the compensator without energy storage,
which minimizes the line loss while taking a sinusoidal
current, is the compensator which delivers the positive-

338 ETEP Vol. 4, No. 5, Septernber/October 1994


ETEP
sequence reactive load current and the zero-sequence system with balanced positive-sequence voltages [3.91.
load current. Summarizing one concludes that. if com- In this case the instantaneous sinusoidaldeactive current
pensation is restricted to sinusoidal currents, the load coincides with the instantaneous deactive current, as de-
current should be split up as follows: fined by Akagi-Nabae [I]. Indeed the instantaneous de-
The urtive current 1, or I lu: This is the minimal cur- active current is a positive-sequence reactive current. In
rent corresponding to the active power delivered to this case all deactive current compensation is realizable
the load. by elements without energy storage.

The negative-sequence current i2:This current does


not contribute to average active power transfer; it 6 Unbalanced Power System Voltages
does however contribute to instantaneous active
power transfer. This current component can be com- If the three-phase voltages are not balanced, the
pensated by elements with energy storage, in partic- computation of the current 1, from the geometrical
ular by means of linear inductors and capacitors. interpretation of Lemma 1is more complicated;the rea-
son is that Cr is not orthogonal to U*orj U*,
although the
The reactive positive-sequence current I,, and the
zero-sequence current I,: These current compo-
latter two vectors are still mutually orthogonal. The re-
sulting expression of the current Ips:
nents do not contribute to instantaneous active
power transfer. These currents can be compensated I, = aL!-+ Pcr (18)
by inductors and capacitors, but also by means of
with
elements without energy storage.
The sum of the currents a)and B) is the instantane-
ous sinusoidal active current Ip. The sum of the current
components y) is the instantaneous sinusoidal reactive
current l,,,. and
The relation of the new components with Akagi-
Nubae’s components [ 1 ] is as follows. Component a)is
part of Akagi-Nnbue’s instantaneous active current
while the components y) are part of Akagi-Nabae’s in-
stantaneous deactive current. The distinction lies in The cases where the denominator of a vanishes
m,
component which in Akugi-Nabae’s theory is divid- should be excluded. These cases do not occur in practi-
ed up into an active and a deactive part. Because of the cal situations. Mathematically they correspond to cases
requirement of sinusoidal currents this component p) is where the three vectors U , and jg are not linearly
completely included in the instantaneous sinusoidal ac- independent. The expression of Ipsreduces to eq. ( 16)if
tive current. the voltages are balanced. The analytical expressions of
It is instructive to compare also the new decompo- the instantaneous sinusoidal active and reactive currents
sition with the decomposition procedure proposed by can be much simplified in the practically important par-
Cxrnecki [4, 51. This author distinguishes the follow- ticular case that the voltage and current vectors do not
ing components for three-phase systems in sinusoidal contain zero-sequence components:
steady state: -
U=iJ++U-
I. the active current, I = 1’ -+ 1-
11. the reactive current, with
111. the unsymmetrical current. u+=[I a’ a ] T U , ,
To see the relationship with the current compo- u-=[ 1 a a2]Tfi*,
nents defined in the present paper, the unsymmetrical
current should be further split up into a negative-se- I+= [ 1 a2 a ] T ‘ l , ,
quence and a zero-sequence component. Component 1. I-=[ I a a’]Q2.
corresponds to a).The negative-sequence part of III.
corresponds to /3). The zero-sequence part of 111. to- It can readily be shown that in this case the instanta-
gether with 11. correspond to y). neous sinusoidal deactive current
To illustrate the results consider the compensation
Iqs=I-lps
of a single-phase resistive load connected between two
phases of a three-phase power system with balanced is the projection of the current vector I on the vector
positive-sequence voltages [ 151.In thiscase thedeactive j(u+- U-). This yields a particularly simple expression
current is purely negative sequence. The instantaneous of this current:
sinusoidal deactive current is zero for this case. If only
compensation by means of sinusoidal currents is consid-
ered. then partial compensation by means of elements
without energy storage is not possible.
As a second illustrative example consider a symmet-
rical reactive load connected to a three-phase power or

ETEP Vol. 4. No. 5 . SeptembedOctober 1994 339


ETEP
could be Compensated by elements without energy
storage. Although Akagi-Nabae's deactive current
corresponds to zero instantaneous power, in general.
even for sinusoidal voltages, the fundamental fre-
quency component of Akagi-Nabae's deactive cur-
This expression was derived independently by Prof. rent does not correspondto zero instantaneouspower:
Depenbrock (private communication).Hence hence it cannot be compensated by elements without
A+ A- energy storage. This is readily seen from the example
-Iqs- - v -Ic -(u+-u-),
2 J - considered in Section 3. The fundamental frequency
3(1ai' +k21) component of the instantaneousdeactive current de-
rived for that example is:
with Q+and @denoting the reactive powers correspond-

i 1
ing respectively to the vectors of the positive-sequence ' 45 [cos(wt)]/2
111
voltages and currents, and to the vectors of the negative- iqf(t)= IJ'I&[cos(ut + 2 ~ / 3 ) ] / 2.
sequence voltages and currents. For the case of balanced
voltages this reduces to the current 1I r ;this corresponds /L'l f i [cosrwt - 2n/3)]/2
to the result obtained in the previous section. Note that It is readily checked that, although i , ( t ) is orthogonal
the instantaneous sinusoidal deactive current contains to u ( t ) , the same is not true for the current iqf(t).
positive- and negative-sequence components;they have
the same ratio as the correspondingvoltage components. In relation with the problem considered in this paper,
The current components are in quadrature with the cor- the opinion of experts in power electronics would be
responding voltage Component; the phase differences of welcome on the following questions:
90" for both components have opposite signs. - Is it interesting from a practical point of view to
It would be interesting to analyse the relationship compensate separately deactive currents which
between the current decomposition technique obtained correspond to zero instantaneous power?
above and the decomposition procedures proposed by
Czarnecki [4,5] and by Ferrero and co-authors [ lo]. The Is it advantageous to require these currents to be si-
expression of -Iq, shows that the compensation of the in- nusoidal, such that the compensationof the remain-
stantaneous sinusoidal deactive current does not aim at ing part of the deactive currents can be realized by
annihilatingor reducing the reactive power Q++ but c, means of linear inductive or capacitive elements?
rather at annihilatingthe difference Q+- @; this supports If the answers to these questions are affirmative, it is
the reservation against the generalization of the concept interesting to try to generalize the decomposition
of reactive power to non-standard situations (asymmet- technique of the present paper to the case of distorted
rical currents and/or voltages, distorted voltages and cur- voltages and/or currents. Some ideas are discussed in
rents). The difference between the reactive powers cor- the next section.
responding to the positive- and the negative-sequence
componentsalso shows up in Ferrero's papers [I I] in the
expression of the average value of the imaginq Park
power; the above analysis leads hence to a physical inter- Distorted Voltages and Currents
pretation of that average value. If the current contains
In the present section it is tried to generalizethe con-
zero-sequence components, but the voltage does not,
cepts of instantaneous sinusoidal active and deactive
then the zero-sequence current components are included
currents to the case of distorted voltages and currents.
in the sinusoidal instantaneous deactive current.
Assume that the current waveform does not contain any
harmonic order which is not present in the voltage wave-
form; this is a reasonable assumption since the analysis
7 Remarks is aimed at the linearity of the cornpensation technique.
A possible concept generalizing the instantaneous sinu-
- For balanced voltages the sinusoidal instantaneous soidal active current could be the instantaneous har-
deactive current can be computed from the instanta- monic active current iph(t)defined as follows:
neous reactive power defined by A h g i [ 1,2] in the
The currentiph(t)correspondsto the same instantane-
same way as the instantaneous reactive current, by
ous power as the load current i ( t ) :
only considering the constant term in the instantane-
ous reactive power. This is however not true for un- u (t)Ti*(t)= u ( t > 5 ( r ) .
balanced power system voltages.
Hence the instantaneous harmonic deactive current
- To derive the instantaneous sinusoidal deactive cur- iqh= i ( r )- i p h ( t corresponds
) to zero instantaneous
rent one could conjecture the following algorithm: power.
compute the instantaneous deactive current accord-
The current waveform iph(t) does not contain har-
ing to Akagi-Nubue's procedure. and take the funda-
monic orders which are not present in the voltage
mental frequency component of this current as the in- waveform.
stantaneous sinusoidal deactive current. It is impor-
tant to realize that this is not a correct way of getting Among all currents satisfying the above two condi-
the instantaneous sinusoidal deactive current which tions i p l , ( f )is the one with the smallest rms value.

340 ETEP Vol. 4.No. 5 , SepternbdOctober 1994


ETEP
It is clear that for the sinusoidal case this decompo-
sition corresponds to the decomposition considered in
Section 4. lrsJ
The concreteexpression of these conditionsin terms on a subspace in the space of the vectors with six com-
of the data corresponding to the waveforms of the volt- plex elements. This example shows that this approach
age and the current is rather complex. One of the reasons leads to a complex procedure if the number of harmon-
is that the instantaneous power contains sinusoidal fluc- ic orders is relatively high.
tuations of various frequencies; each frequency can cor- In the particular, but practically most important
respond to different harmonic orders of the voltage and case that voltages and currents do not contain zero-se-
the current. Therefore conditions are required between quence components, the solution can be expressed very
the voltage and currentcomponentsof different harmon- similarly to the results obtained in Section 6. Let the
ic orders to have zero instantaneouspower to correspond p h a s o r v e c t o r s ~ I , U s andL5beexpressedas
,~I thesum
to a current waveform and a voltage waveform. As an of positive-sequence, and negative-sequence compo-
example consider the voltage vector nents:
u 0 )= U I (0 + W ) UI=U;+U;, us = u;+ us,
containing first and fifth harmonics: LI =[;+IT, Is = I ; +I:.
Then the phasors of the first and fifth harmonic com-
ponents of the instantaneousharmonic deactive current
A cos(or + 2x13) ',i are obtained as the projection of the vector

B cos (5wr)

B C O S ( ~-U
2x13)
~ on the vector

Then a current containing first and fifth harmonics


C sin (ax)
and hence
Csin(wr +2x/3)

D sin (5wr)

Dsin(5wt - 2x13)
corresponds to zero instantaneous power provided
AD + B C = 0. This result can readily be extended to the general
To see how the currents iphand iqhcould be derived case with an arbitrary number of harmonic compo-
in general, an example is considered where the wave- nents. The numerator of the multiplier is the differ-
forms of the voltage and the current contain fundamen- ence between the positive-sequence reactive powers
tal frequency components as well as fifth harmonics. Let and the negative-sequence reactive powers. The de-
the correspondingphasor vectors be denoted by U I, L I, nominator equals the square of the rms value of the
u5,15,and the phasor vectors of the first and fifth har- voltage. As in Section 6 the question arises whether
monic components of i p h ( f ) by d, and J 5 .Then the vec-
these components i p h ( t and) iqh(r) have a meaning in
tors J I and J 5 should be chosen such that the current decompositionsproposed by Czarnecki [ 5 ]
or by Ferrero [ 6 ] .
llJl 112 + llJ# Here also remarks on the relation to the definitionof
is as small as possible, with the following constraints reactive power and in particular to Eudeunu's definition
are relevant as in Section 6. The aim of the compensa-
Re I Y :I; + U TI; I = Re { K TJ; + U,',f;I. (19) tion of the instantaneous harmonic deactive current
IU:~JI= lU,'s5I7 (20) i q h ( fis) hence to annihilate the difference between the
reactive powers for the positive-sequence and the nega-
Iv :r v
I I= I fJI I, (21) tive-sequence components; this quantity also shows up
in the papers by Ferrero et al. [ 1I ] in the expression of
IUT1; + U;rLI I = lUT& + U?-_JI I. (22) the average value of the imaginary Park power; eq. (24)
1 V f ~ +s U:LI I = IY :J j + UTJ I 1. (23) hence leads to a physical interpretation of that average
value. Note finally that if there are no zero-sequence
These conditions lead to the conclusion, similar to voltage components, the zero-sequence current compo-
the result of Lemma 1. thatdl andJ5 can be obtained as nents should be included completely in the instantane-
the orthogonal projection of the vector ous harmonic deactive current.

ETEP Vol. 4. No. 5, SeptembdOctober 1994 34 I


ETEP
9 Conclusion [J] Czrnecki, L S.: Orthogonal decomposition of the cur-
rents in a three-phase nonlinear asymmetrical circuit
A new decomposition of sinusoidal currents in with a non-sinusoidal voltage source. IEEE Trans. on In-
strumentat. a. Meas. IM-37 ( 1988) pp. 30-34
three-phase power systems has been proposed. It has the [ 5 ] Czarnecki. L S.: Scattered and reactive current, voltage
interesting feature to combine the use of compensating and power in circuits with non-sinusoidal waveforms
elements without energy storage with the feature that the and their compensation. IEEE Trans. on Instrumentat. a.
current components are sinusoidal and compensation is Meas. IM-40 (1991) pp. 563-567
possible with linear components.The relationshipof the [6] Czarnecki, L S.: Current and Power Equations at Bidi-
new components with known decompositionshas been rectional Flow of Harmonic Active Power in Circuits
discussed in detail. In a stnightforward way the analy- with Rotating Machines. ETEP Eur. Trans. on Electr.
Power Engng. 3 ( 1993) no. 1, pp. 45 -52
sis can be generalized-at the cost of increased complex- [7] Depenbmck, M.: The FDB-method, a generally appli-
ity - to polyphase power systems. A possiblegeneraliza- cable tool for analysing power relations. ICHPS V, At-
tion to distorted voltages and currents has been dis- lanta, GeorgiaNSA 1992, Reprints pp. 135- 141
cussed in this contribution. It depends on practical con- [8] Depenbrock. M.: Marshall. D. A.; Van wk, J. D.: For-
mulating requirements for a universally applicable
siderations whether or not the proposed decomposition
technique is relevant for the purpose of the compensa- power theory as control algorithm in power compensa-
tors. 2nd Int. Workshop on Power Definitions and Meas.
tion of deactive currents. under Non-Sinusoidal Cond., Stresdtaly 1993, Pre-
prints pp. 1-25
[9] Epstein. E.: Yair. A.: Alexandrovit:, A.: Three-phase
10 List of Symbols, Sub- and Superscripts static Var compensator with reduced energy storage.
etzArchiv 8 (1986) no. 6, pp. 213-219
10.1 Symbols [lo] Ferrero, A.: Superti Furga, G.: A new approach to the
definition of power components in three-phase systems
under non-sinusoidal conditions. IEEE Trans. on Instru-
instantaneouscurrent vector ment. a. Meas. IM-40 (1991) pp. 568-577
instantaneous voltage vector [ I I 1 Ferrero, A.; Morando. A. F: Ortoboni, R.; Supeni Furga,
current phasor G.: On the Meaning of the Park Power Components in
L! voltage phasor Three-Phase Systems under Non-Sinusoidal Conditions.
P, Q,S active, reactive, apparent power ETEP Eur. Trans. on Electr. Power Engng. 3 (1993)
G conductance no. I , pp. 33-43
B susceptance [ 121 Cristafdi.L ; Ferrero. A.; Supeni Furga, G.: Power and
current decompositions into time and frequency domain
j imaginary unit components: analysis and comparison. 2nd Int. Work-
cp phase lag between voltage and current shop on Power Definitions and Meas. under Non-Sinu-
ej2x13
a soidal Cond., StresdItaly 1993, Preprints pp. 83-92
1...I absolute value [ I3 1 Murshall. D. A.; VenteK F: i?; Van L@k, J. D.: An evafu-
11.. .II norm ation of the instantaneous calculation of load current
components. ETEP Eur. Trans. on Electrical Power
10.2 Subscripts Engng. 3 ( 1993) no. 1, pp. 53-59
[ 141 Rosetfo.L ; Tenti. P.: Evaluation of instantaneous power
terms in multi-phasesystems: techniques and application
1,2,0 positive sequence, negative sequence, to power conditioning equipment. 2nd Int. Workshop on
zero sequence Power Definitions and Meas. under Non-Sinusoidal
L 1, L2, L3 phases of a three-phase system Cond., StresdItaly 1993, Preprints pp. 35-44
a, r, d active, reactive and deactive quantities [IS] wiLkms, 3. L: Current compensation in three-phase
instantaneous active and deactive power systems. ETEP Eur. Trans. on Electr. Power
P7 q Engng. 3 (1993) no. I, pp. 53-59
PS. qs active and deactive sinusoidal quantities [ 161 Wllems. J. L.: A new interpretation of the Akagi-Nabae
Ph, qh active and deactive harmonic quantities power components for non-sinusoidal three-phase situa-
tions. IEEE Trans. on Instrumentat. a. Meas. IM-41
10.3 Superscripts (1992) pp. 523-527
[ 171 willem, J. L:The computation of the alternatingtoque
+, - positive sequence, negative sequence of an a.c. machine. Int. J. of Electr. Engng. Educ. 19
* complex conjugate (1982) pp. 79-85
[ 181 Wllems. J. L;Aeyels, D.: New decomposition for 3-
T matrix transpose phase currents in power systems. IEEb. C 140( 1993)
pp. 307 - 3 10
References This paper was presented at the “2nd InternationalWorkshop
[I] Akagi. H.: Kanazawa.:!l Nabae, A.: Instantaneous reac- on Power Definitions and Measurements under Non-Sinu-
tive power compensators comprising switching devices soidal Conditions” in Stresdtaly (September 8- 10, 1993)
without energy storage components. IEEE Trans. on Ind.
Appl. IA-20 (1984) pp. 625-63 I
[2] Akagi, H.; Nabue. A.: The p-q Theory in Three-Phase Acknowledgment
Systems under Non-Sinusoidal Conditions. ETEP Eur.
Trans. on Electr. Power Engng. 3 ( I 993) no. 1, pp. 27 -3 I This paper presents research results of the Belgian Programme
[3] Ale.randrovir:. A.; kir: A.: Epstein. E.: Analysis of static on Interuniversity Poles of Attraction, initiated by the Belgian
VAR compensator with optimal energy storage element. State, Prime Minister’s Office for Science, Technology and
IEEE Trans. on Ind. Electron. 1E-3 I (1984) pp. 28-33 Culture. The scientific responsibility is assumed by the author.

342. ETEP Vol. 4. No. 5. September/October I994

Das könnte Ihnen auch gefallen