Relationship between harmonics and symmetrical components

G. Atkinson-Hope Department of Electrical Engineering, Cape Technikon, South Africa

Abstract New terminology is introduced to make clear the relationship between harmonics and symmetrical components. Three-phase sets are classied in terms of symmetrical sets and asymmetrical sets. Subclasses are introduced with the names symmetrical balanced sets, symmetrical unbalanced sets, asymmetrical balanced sets and asymmetrical unbalanced sets to show that a threephase set can resolve to either one, two or three symmetrical component sets. The results from four case studies show that these subclasses and their resolution to symmetrical component sets improve understanding of harmonic analysis of systems having balanced and unbalanced harmonic sources and loads. Keywords asymmetrical sets; harmonic ows; harmonic sources; symmetrical component sets; symmetrical sets

Any periodic wave shape can be broken down into or analysed as a fundamental wave and a series of harmonics.1 Three-phase harmonic analysis requires a clear understanding of the relationship between symmetrical component injections from harmonic sources (e.g. adjustable speed drives, ASDs) and their relationship to harmonic ows (symmetrical components) arising from the application of a harmonic source to a linear system. A limited number of references contain brief information concerning harmonics and symmetrical components.15 Reference 1, provides a paragraph on this topic and uses the heading Relationship between Harmonics and Symmetrical Components. It includes a table that is supported by a brief explanatory paragraph. The table expresses harmonics in terms of positive, negative and zero sequences. It states that these sequences are for harmonics in balanced three-phase systems. The heading refers to symmetrical components while the content refers to balanced three-phase systems. Herein lies the anomaly. Classically, symmetrical components (especially zero sequence) are only applied in unbalanced systems. The following questions arose after reading the Ref. 1 paragraph. (a) Do symmetrical components (especially zero sequence), in the classical sense, apply in balanced as well as unbalanced non-sinusoidal systems and is this a break from tradition? (b) What do the terms, symmetrical, asymmetrical, balanced, unbalanced and symmetrical components mean? (c) What are the conditions under which a system must operate so that harmonics resolve to positive, negative and zero sequences and is the table given in Ref. 1 correct? The terminology used is found inadequate for describing non-sinusoidal systems. There is thus a need to introduce a three-phase terminology that will show the
International Journal of Electrical Engineering Education 41/2

94

G. Atkinson-Hope

relationship and make the comparison between injections (currents) and harmonic ows (voltages and currents) meaningful. References 3 provides the basis for the solution by providing denitions for threephase sets, symmetrical sets and symmetrical component sets. The purpose of this paper is to introduce an approach to harmonic analysis based on the classication of three-phase sets and to make to comparison between injections from harmonic sources and corresponding harmonic ows quantiable by expressing the results in terms of the number of symmetrical component sets found. Harmonic ows and their resolution to symmetrical components depends upon the magnitudes and phase sequences of the injections from a harmonic source, on the systems sequence impedances, on three- and four-wire connections and on whether the customers linear load on the system is balanced or unbalanced. Therefore, what is injected in terms of symmetrical component sets by a harmonic source is not necessarily received by the system, i.e. the harmonic ows may resolve to one, two or three symmetrical component sets and this depends upon the type of three-phase set found. Therefore, any three-phase harmonic may be partially made up of any of the symmetrical component sets.4,6 Four case studies are reported and they show a novel method for teaching the ow of power system harmonics. It is important to use case studies as part of ones teaching as they link learning to concepts and improve understanding. They show how the method of symmetrical components can be extended to a systems response to harmonic ows. When taught as a group, the four case studies improve cognitive skills by showing that the symmetrical component responses under unbalanced situations are different to the balanced state. Classication of three-phase sets At a point/node within a sinusoidal system, a set of currents (or voltages) is found, for example: ia = 2 IA cos (wt + a a1 ) ib = 2 IB cos (wt + a b1 ) ic = 2 IC cos (wt + a c1 ) (1) (2) (3)

Such a set is called a three-phase set (TPS) and is dened as a group of three interrelated currents (or voltages) that have the same period.3 Therefore, a system is comprised of a number of TPSs. The type of TPS found at the various points/nodes in a given system depends upon the system conguration. Symmetrical and asymmetrical sets The magnitudes and phase displacement of the three phasors of a TPS are variables. A TPS can therefore be either a symmetrical set (SS) or an asymmetrical set (AS). If each phasor of a TPS is equal in magnitude and has either a 120 or 0 phase displacement, the set is a SS.3,7 A TPS which does not have the characteristics of a
International Journal of Electrical Engineering Education 41/2

Relationship between harmonics and symmetrical components

95

SS in an AS. A TPS can have either an abc, acb or an in-phase order of rotation for its phasors. Figure 1 shows the classication of a TPS in the form of a diagram. There are thus six classes of TPS. Meaning of the terms balanced and unbalanced If the three phasors (IA, IB and IC) of a TPS at a given point in a system have the same magnitude and are displaced from each other by 120, the set is balanced. In this case, the phasors form an equilateral triangle and their sum is zero. With reference to a TPS at a given point in a system, the term balanced means the phasors of the set sum to zero (IA + IB + IC = 0). Therefore, if the three phasors of a set at a given point in a system do not sum to zero (IA + IB + IC 0), the TPS is unbalanced.810 Symmetrical and asymmetrical balanced and unbalanced sets If the TPS is a SS with 120 phase displacement, the phasors sum to zero, therefore the TPS is a symmetrical balanced set (SBS). There are thus two types of SBS, one with an abc order and one with and acb order. Whereas, if the TPS is a SS with a 0 phase displacement, the phasors do not sum to zero and the TPS is a symmetrical unbalanced set (SUS). If the phasors of an AS sum to zero, the TPS is an asymmetrical balanced set (ABS). Whereas, if the phasor sum of an AS is not zero, the TPS is an asymmetrical unbalanced set (AUS).9 Figure 2 shows the subclasses of TPSs. Symmetrical component sets resolved from symmetrical and asymmetrical sets Understanding the relationship between symmetrical components and three-phase harmonics provides a rm foundation for power system analysis.2 The method of symmetrical components as an analytical tool is traditionally used to examine sinusoidal systems under balanced and unbalanced customer loading conditions.11 As harmonics are sinusoidal quantities, the method can be extended to steady-state harmonic analysis.4

Fig. 1

Classication of a three-phase set (TPS).

International Journal of Electrical Engineering Education 41/2

96

G. Atkinson-Hope

Fig. 2

Subclasses of three-phase sets.

The symmetrical components of any TPS (IA + IB + IC) are: I 0 = 1 3 ( I A + I B + IC ) I1 = 1 3 ( IA + aIB + a 2 IC ) I2 = 1 3 ( IA + a IB + aIC )
2

(4) (5) (6)

The symmetrical component transformation is a general mathematical technique developed by Fortescue and provides that any TPS may be resolved into symmetrical components.11 Therefore, if the TPS in equation (4) is: (i) SBS or SUS The TPS resolves to only one symmetrical component set (SCS). The other two SCSs are absent. If the order of the SBS is abc, it resolves to a positive sequence symmetrical component set (SCS+), whereas if it is acb, it resolves to a negative sequence symmetrical component set (SCS-). If the set is SUS, it resolves to a zero sequence symmetrical component set (SCSz).6,12 (ii) ABS or AUS The TPS resolves to more than one SCS. If the set is ABS, it resolves to two SCSs; one SCS+ and one SCS-. If the set is AUS, it resolves to three SCSs; a SCS+, SCS- and SCSZ.9 Figure 3 and Table 1 show the SCSs resolved from the subclasses of TPSs. Modelling three-phase harmonic current sources in terms of symmetrical component sets Harmonic sources are considered to be injection sources and for most studies are treated as simple sources of harmonic currents.4
International Journal of Electrical Engineering Education 41/2

Relationship between harmonics and symmetrical components

97

Fig. 3 TABLE 1
Three-phase sets (TPS) Type SBS SBS SUS ABS AUS Order abc acb in-phase abc/acb abc/acb in-phase

SCSs resolved from TPSs.

SCSs resolved from subclasses of TPSs

Symmetrical component sets (SCS) Positive SCS+ Absent Absent SCS+ SCS+ SCS+ Negative Absent SCSAbsent SCSSCSSCSzero Absent Absent SCSZ Absent SCSZ SCSZ

Balanced source The main contributor to system harmonic distortion is the six-pulse convertor. Under perfect conditions it is an ideal convertor and its wave shapes are periodic. The Fourier series for the current in its A phase is:11 iA = (2 p ) 3 Id [(cos wt ) - (1 5)(cos 5wt ) + (1 7)(cos 7wt ) - (1 11)(cos 11wt ) + (1 13)(cos 13wt ) - (1 17)(cos 17wt ) + (1 19)(cos 19wt ) - ] (7)

International Journal of Electrical Engineering Education 41/2

98

G. Atkinson-Hope

where (2/p) 3 Id is the maximum value of the fundamental frequency component (I1m). The RMS magnitude of the fundamental frequency and each harmonic component are I1 = [(2/p) 3 Id] 2 and Ih = I1/h, respectively, where h is the harmonic number. Triplen harmonic components are absent. Only 6k 1 harmonics are present, for integer values k = 1, 2, 3, . . . , n. They are called characteristic harmonics.13 The general expression for the currents in the A, B and C phases are therefore: iA = I1m cos wt - I5 m cos 5wt + I7 m cos 7wt - I11m cos 11wt + I13m cos 13wt - I17 m cos 17wt + I19 m cos 19wt iB = I1m cos(wt - 120) - I5 m cos(5wt - 240) + I7 m cos(7wt - 120) - I11m cos(11wt - 240) + I13m cos(13wt - 120) - I17 m cos(17wt - 240) + I19 m cos(19wt - 120) iC = I1m cos(wt + 120) - I5 m cos(5wt + 240) + I7 m cos(7wt + 120) - I11m cos(11wt + 240) + I13m cos(13wt + 120) - I17 m cos(17wt + 240) + I19 m cos(19wt + 120) Changing - to + polarities, the equations are:14,15 iA = I1m cos wt + I5 m cos(5wt + 180) + I7 m cos 7wt + I11m cos(11wt + 180) + I13m cos 13wt + I17 m cos(17wt + 180) + I19 m cos 19wt iB = I1m cos(wt - 120) + I5 m cos(5wt - 60) + I7 m cos(7wt - 120) + I11m cos(11wt - 60) + I13m cos(13wt - 120) + I17 m cos(17wt - 60) + I19 m cos(19wt - 120) iC = I1m cos(wt + 120) + I5 m cos(5wt + 60) + I7 m cos(7wt + 120) + I11m cos(11wt + 60) + I13m cos(13wt + 120) + I17 m cos(17wt + 60) + I19 m cos(19wt + 120) (13) (12) (11) (10) (9)

(8)

At each frequency a SBS is obtained. By denition, harmonics result from periodic steady-state operating conditions and therefore their prediction should be formulated in terms of harmonic phasors, i.e. in the frequency domain,11 namely: wt _ 0 IA 5 _ - 120 IB5 _ + 120 IC 5 5wt _ + 180 IA 7 _ - 60 IB7 _ + 60 IC 7 7wt _ 0 _ - 120 _ 120

I A1 IB1 IC1

International Journal of Electrical Engineering Education 41/2

Relationship between harmonics and symmetrical components

99

Fig. 4

SCSs used to model an ideal six-pulse convertor.

IA11 IB11 IC11

11wt _ + 180 IA13 _ - 60 IB13 _ + 60 IC13

13wt 17wt 19wt _ 0 IA17 _ + 180 IA19 _ 0 _ - 120 IB17 _ - 60 IB19 _ - 120) _ + 120 IC17 _ + 60 IC19 _ + 120)

The SBSs at 1st, 7th, 13th and 19th harmonic frequencies have an abc order of rotation, equal magnitudes and are 120 apart. They are thus SCS+s. At the 15th, 11th and 17th harmonic frequencies the SCSs have an acb order of rotation, equal magnitudes and are 120 apart and are thus SCS-s. SCSs resolved from SBSs can thus be used to model an ideal six-pulse convertor, as shown in Fig. 4. Each injection is only one SCS, they, however, have different frequencies. As all the injections are SBSs the convertor can be said to operate as a balanced harmonic current source.16 Unbalanced source Normally the convertor injects only characteristic harmonics. If not red symmetrically the convertor injects non-characteristic harmonics (triplens and even harmonics) besides characteristic harmonics. Since convertors typically inject currents with half-wave symmetry, even harmonics are ignored. Under such non-ideal conditions, the actual magnitudes of the TPSs of a six-pulse convertor are not inversely proportional to the harmonic number h, therefore I h I1 h (14)

Magnitudes are based on measurements or published data. TPSs at triplen harmonic frequencies typically have an in-phase order of rotation and are SUSs. Their injections are therefore treated as SCSZs. Under non-ideal conditions the convertor spectrum would be1 as shown in Fig. 5. As the source injects both SUSs and SBSs it can be said to operate as an unbalanced harmonic source16 (Table 2).
International Journal of Electrical Engineering Education 41/2

100

G. Atkinson-Hope

Fig. 5

Convertor spectrum under non-ideal conditions. Source operating as an unbalanced harmonic source
Unbalanced harmonic source

TABLE 2

NR h 1 3 5 7 9 Sequence Positive* Zero Negative* Positive* Zero

Symmetrical components h 11 13 15 17 etc Sequence Negative* Positive* Zero Negative*

* Indicates a balanced harmonic source.

Case studies A three-phase four-wire radial distribution system is used for the four case studies conducted. The system is comprised of a three-phase sinusoidal voltage source (star point connected to neutral), a distribution network (HV and LV sections including a star/star transformer, both sides have neutral returns providing four-wire distribution) and a load (customer). The customer load is comprised of a motor load (three-wire), a linear star load with a neutral return, a single-phase linear load with a neutral return and a harmonic current source (six-pulse convertor). The results of the case studies are given in Table 3. The SUPERHARM software program is utilized as it is able to analyse both balanced and unbalanced non-sinusoidal systems.15 The injections from the harmonic source are compared to corresponding harmonic ows in the various sections of the system in terms of SSs, ASs and SCSs. The number of SCSs is shown in brackets in Table 3.
International Journal of Electrical Engineering Education 41/2

TABLE 3
Injections from a harmonic source compared to harmonic ows Distribution network HV and LV sections LN(V) LL(V) (A) LL(V) (A) Induction motor load

Results of case studies

Harmonic current source (injections)

Case

Harmonic frequency

Type of set (number of symmetrical component sets in brackets)

Customer linear load Balanced Unbalanced Balanced Unbalanced Balanced Unbalanced SBS(1) SBS(1) ABS(2) ABS(2)

Relationship between harmonics and symmetrical components

1 2 3 4 SBS(1) SBS(1) AUS(3) AUS(3) N/A SUS(1) N/A AUS(3) N/A None N/A ABS(2) SBS(1) SBS(1) ABS(2) ABS(2) SBS(1) SBS(1) AUS(3) AUS(3) N/A SUS(1) N/A AUS(3)

Fundamental

SBS(1) SBS(1) SBS(1) SBS(1)

SBS(1) SBS(1) AUS(3) AUS(3)

SBS(1) SBS(1) ABS(2) ABS(2)

SBS(1) SBS(1) (AUS)(3) AUS(3)

SBS(1) SBS(1) ABS(2) ABS(2) SBS(1) SBS(1) ABS(2) ABS(2) N/A None N/A ABS(2)

1 2 3 4

6k 1

SBS(1) SBS(1) SBS(1) SBS(1)

SBS(1) SBS(1) ABS(2) ABS(2) N/A None N/A ABS(2)

1 2 3 4

Triplen

None SUS(1) None SUS(1)

101

International Journal of Electrical Engineering Education 41/2

102

G. Atkinson-Hope

The harmonic current source for cases 1 and 3 is modelled as a balanced source, whereas for cases 2 and 4 an unbalanced source is used. The same spectrum for the harmonic source is used for all cases, except triplen harmonics are added for the unbalanced source investigations. For cases 1 and 2 the customer load on the system is balanced (the single-phase linear load is ignored) whereas for cases 3 and 4 the load is unbalanced. In case 1, the fundamental and 6k 1 injections and corresponding harmonic ows are all single SCSs. Therefore, the SCS+s and SCS-s that are injected, propagate through the three- and four-wire sections of the system. This is expected as the load is balanced. For case 2, the fundamental and 6k 1 injections and ows are the same as case 1. The triplen harmonics injected and the corresponding ows are SUSs and are SCSZs. No SCSZs appear in the LL voltages nor in the current ow in the threewire motor section of the system. In case 3, the injections are SBSs and are therefore single SCSs. Within the sections of the system the ows are found either to be ABSs or AUSs and they resolve to two and three SCSs, respectively. The ABSs are found in the three-wire sections whereas the AUSs appear in the four-wire sections. Therefore, the corresponding SCS ows are different to that which is injected. The case 4 results for the fundamental and 6k 1 harmonics are the same as case 3. The triplen harmonic SUSs give rise to ABSs and AUSs within the system and appear in the same sections as case 3. When compared to case 2, the results show that triplen harmonics are not purely SCSZs. In systems having an unbalanced harmonic source and an unbalanced load, the triplen harmonic ows can resolve to either two SCSs (SCS+ and SCS-) or to three SCSs (SCS+, SCS- and SCSZ). Conclusions The table and paragraph given in Ref. 1 does not adequately describe the relationship between harmonics and symmetrical components, especially harmonic sources and does not cover unbalanced loads. In general authors of power system literature ignore the symmetrical set concept as opposed to the asymmetrical set. The role that the symmetrical set plays in harmonic analysis and in the modelling of harmonic sources should not be ignored. Classically, positive, negative and zero sequence components have the same frequency. The positive, negative and zero sequence symmetrical components injected by an unbalanced harmonic source have different frequencies. The new terminology/subclasses of three-phase sets introduced distinguish balanced and unbalanced harmonic sources from each other and are found suitable for comparing injections to harmonic ows in systems in terms of symmetrical components. The results show that what is injected in terms of symmetrical component sets by a harmonic source is not necessarily received by the system, i.e. the harmonic ows may resolve to one, two or three symmetrical component sets and this depends upon the type of three-phase set found at a given point/node in a system.
International Journal of Electrical Engineering Education 41/2

Relationship between harmonics and symmetrical components

103

The results show that if the harmonic source and the load is balanced, the line voltages and 6k 1 harmonic currents in the motor section are symmetrical sets. Only one SCS per harmonic frequency is applied to the motor. If the harmonic source is unbalanced and the load remains balanced the line voltages and currents in the motor section are unaffected by the injected triplen harmonics. If the source is balanced and the load unbalanced, the 6k 1 harmonics in the line voltages and currents are asymmetrical balanced sets, therefore two SCSs per harmonic frequency (a positive and a negative sequence symmetrical component set) is applied to the motor. If the source and load are both unbalanced, the heating effect on the motor is further increased as triplen harmonics as asymmetrical balanced sets, are added to the 6k 1 harmonic asymmetrical balanced sets resulting in the motor being subjected to two SCSs per harmonic frequency (a positive and a negative sequence symmetrical component set) at triplen and 6k 1 harmonic frequencies adding to the decline in the life expectancy of the motor. The results show that triplen harmonics as ows act like 6k 1 harmonics and are not purely zero sequence, as is commonly believed, but resolve to two SCSs per harmonic frequency (a positive and a negative sequence symmetrical component set) in the three-wire sections and to three SCSs per harmonic frequency (a positive, a negative and a zero sequence symmetrical component set) in the four-wire sections when the source and load are unbalanced. The results show that both symmetrical sets and asymmetrical sets and their resolution to symmetrical component sets are important analytical tools for harmonic analysis of systems having balanced and unbalanced harmonic sources and loads. References
1 2 3 4 5 6 7 8 9 10 IEEE Recommended Practice for Electric Power Distribution for Industrial Plants, IEEE Standard 141, (1993), 447448, G. W. Massey, Power distribution system design for operation under non-sinusoidal load conditions. IEEE Trans. Ind. Applic. 31 (1995), 513514. IEEE Standard Dictionary of Electrical and Electronic Terms, IEEE Standard 100, (1996), pp. 73, 74, 681, 768, 796, 976, 1073, 1215. R. C. Dugan, M. F. McGranaghan and H. W. Beaty, Electrical Power Systems Quality (McGrawHill, New York, 1996), pp. 172174. J. L. Blackburn, Symmetrical Components for Power System Engineering (Marcel Dekker, New York, 1993), pp. 3738. E. W. Kimbark, Power System Stability, Power System Series, vol. 1 (IEEE Press, New York, 1995), pp. 193196. R. Roeper, Short Circuit Currents in Three-Phase Networks (Siemens & Pitman Publishing, London, 1972), p. 8. G. W. Stubbings, Elements of Symmetrical Component Theory (Isaac Pitman & Sons, London, 1937), p. 25. V. Del Toro, Electric Power Systems (Prentice Hall, Englewood Cliffs, NJ 1992), p. 372. W. V. Lyon, Applications of the Method of Symmetrical Components (McGraw-Hill, London, 1937), p. 18.
International Journal of Electrical Engineering Education 41/2

104 11 12 13 14 15 16 17 18

G. Atkinson-Hope J. Arrillaga, B. C. Smith, N. R. Watson and A. R. Wood, Power System Harmonic Analysis, John Wiley & Sons, England, (1997), pp. 2, 35, 136. P. Kemp, Alternating Current Electrical Engineering (McMillan, London, 1958), pp. 170171. IEEE Tutorial Course, Power System Harmonics, IEEE Service Centre, Piscataway, (1984), pp. 10, 7379. A. B. Carlson, Circuits (John Wiley & Sons, New York, 1996), p. 243. Electrotek Concepts, Superharm Users Guide, Version 4.0, (1999), pp. 314, 316, 329. DigSilent Power Factory software Manual, Verion 11.1, 1999, pp. 1216. IEEE Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems, IEEE Standard 519, (1992), pp. 27, 35. University of Manchester Institute for Science and Technology and Electrotek Concepts Power Quality Workshop Notes, Manchester Centre for Electric Energy, Harmonic Sources, Session 2C, Table 11, May 1999.

International Journal of Electrical Engineering Education 41/2