You are on page 1of 10

Proc. Natl. Sci. Counc. ROC(A) Vol. 25, No. 6, 2001. pp.

329-338

(Invited Review Paper)

A Review of Methods for Calculation of Frequency-dependent Impedance of Overhead Power Transmission Lines
YAW-JUEN WANG AND SHI-JIE LIU
Department of Electrical Engineering National Yunlin University of Science and Technology Touliu, Yunlin, Taiwan, R.O.C. (Received October 23, 2000; Accepted May 18, 2001) ABSTRACT This paper reviews the three most commonly used methods for calculation of frequency-dependent impedance of overhead transmission lines: Carson’s method, the complex depth of earth return method and the finite element method. The historical path along which these methods were developed is also reviewed. This is followed by an illustrative example that compares the frequency-dependent impedances of a 500 kV transmission line calculated using the tree methods. It is concluded that the complex depth of earth return method can, in most cases, replace Carson’s method without impairing the calculation accuracy, and that the finite element method, though more complicated and computationally more cumbersome, is capable of treating problems resulting from practical geometric allocation of transmission conductors and is a promising method for use in the near future. Key Words: Carson’s method, complex depth of earth return, finite element method, frequency-dependent impedance

I. Introduction
The history of the electrical power industry can be traced back to Edison’s Pearl Street Power Station in New York City in 1882 that supplied 110-V DC power. Then in 1889, the first single-phase AC line between Oregon and Portland (US) began to operate. The first three-phase AC transmission line became operational in Germany in 1891 (Glover and Sarma, 1989). The transmission of electric power from DC to polyphase AC implies that the working frequency of the power system has increased from zero Hz to several different values. For instance, the early Central European and American electric railways were operated at 162/3 Hz and 25 Hz, respectively (Hill, 1994). Modern electric power systems work at 50 or 60 Hz, and some special power systems, such as those used in aircraft, submarines and military equipment, are designed to operate at 400 Hz. The change of power transmission technology from DC to AC allowed the voltage level to be shifted easily by transformers, and the increase in the operational frequency also allowed the volumes and weights of power system components to be reduced. However, several types of disturbances arising from the increase in frequency, such as harmonics, transients and electromagnetic interference, have become new challenges to electrical engineers, and the frequency-dependent impedance of transmission lines has also become one of the

most important power system parameters. For instance, an electrical engineer needs to know the power line impedance from the fundamental frequency (i.e., 50/60 Hz) to about the 25th harmonic frequency (i.e., 1,250/1,500 Hz) to conduct harmonic analysis, and that from DC to about 1 MHz to conduct transient analysis. It is noteworthy that calculation of frequency-dependent impedance is important not only in power engineering, but also in other areas involving high frequency applications like communications and semiconductor microchip design. In fact, as the frequency increases to several GHz (109 Hz), a small segment of the conductor in a microchip behaves just like a 100-km-long power transmission line working at 60 Hz. Consequently, the methods for calculating power line frequency-dependent impedance are also applicable to the analysis of electromagnetic interference in printed circuit boards and integrated circuits. This paper discusses three most commonly used methods for calculating the frequency-dependent impedance of overhead power transmission lines: Carson’s method, the complex depth of earth return method and the finite element method. An example of a three-phase 500 kV transmission line is then used to compare the three methods.

II. Literature Review
The term “frequency-dependent impedance” in this paper

− 329 −

Zij is equal to Zji. Because of the skin and proximity effects. 1989... 2000) has been one of the most rapidly developing topics in electrical engineering in recent years.).J. It has been used to solve Maxell’s equations that govern the distribution of electric and magnetic fields. Liu refers to the frequency response of the distributed series inductance and resistance of a transmission line. It must be emphasized that Carson’s method does not provide a closed-form solution. voltage drop calculations etc. On the other hand. and proved that Gary’s formulae were indeed a closed-form approximation of Carson’s improper integral and were valid over the whole frequency range (Deri et al. In 1926. there was no rigorous theoretical proof. many outstanding electrical engineers have studied this problem over the past 100 years and obtained accurate solutions using various methods. An error as large as 15% is possible. the complex depth of earth return method and the finite element method.g. a professor at Budapest University in Hungary. a research engineer with the French Electricity Board (Electricité de France. the line impedance varies considerably at high frequencies. and to show the distribution of the magnetic field around the conductors. The resulting error is less than 0. 1999. or when they cross a river. Since then. the well known solution that the internal inductance of a straight solid cylindrical conductor is equal to µ0/(8π) H/m. Fortunately. C. This assumption. Deri. It is able to display the distribution of earth return current in the ground.J. no longer holds in analyses involving harmonics and transients in which the voltage and current in the transmission line are rich in high frequency components. Gary’s method does not encounter this problem and is a powerful tool that can replace Carson’s method.Y. Car-son’s method is still the standard method for calculation of the frequency-dependent impedance of overhead transmission lines. Furthermore. That is.. Hence. Both the self and mutual impedances are influenced by the earth return current. we must increase the term number of the infinite series to avoid truncation errors. proposed an approach in which the earth could be replaced by a set of earth return conductors located directly under the overhead lines at a complex depth (Gary. load flow studies. the d/h ratio of typical power transmission lines is very small. EDF). On the other hand. derived a mathematical relation between Carson’s method and the complex ground return. Wang and S. Gary’s proposition was found to be equally accurate over the whole frequency range than Carson’s method. the d/h ratio can be very large. necessitating calculation of the transmission line impedance over the whole frequency range. Electrical engineers at that time still doubted the validity of Gary’s proposition. which may be acceptable in analyses of only the fundamental frequency (e. J.5%. The advantage of the FEM lies in its ability to handle irregular terrain geometry in the vicinity of transmission lines. There might be two reasons. Alvarado and Betancourt (1983) proposed a modified method that reduced the error of the complex depth of earth return method to less than 2. This was a milestone in the development of power engineering. The inconveniences of the FEM include the high cost of the software. Five years later. the FEM is capable of considering the effect of the terrain variation on the line impedance. Unfortunately. Pollaczek in Germany (Pollaczek. When calculating high frequency impedance using Carson’s method. Surprisingly. That is. Triantafyllidis et al. when the transmission lines are located besides a lake or cliff. is derived under DC conditions. Transmission line impedance can be separated into two components: self impedance and mutual impedance. in Gary’s paper. it refers to the distributed resistance and inductance parameters as functions of frequency. Carson’s method and the complex depth of earth return method are not applicable to problems involving terrain irregularities. In 1976. including Carson’s method. 1981)..R. the transmission line impedance can be written in a simple algebraic form.5%. Influenced by the skin and proximity effects. For example. The error of the complex depth of earth return method increases with the ratio of the horizontal distance between conductors to their heights (d/h). Carson’s method may cause considerable truncation errors at high frequencies. However. for cases involving crosstalk between power lines and communication lines. For example. Gary. The mutual impedance Zij between conductors i and j is the ratio of the induced voltage per unit length in conductor i to the current in conductor j. 1926) and F. The FEM can also be used to calculate the frequencydependent impedance of transmission lines. The finite element method (FEM) (Cristina and Feliziani. the distances between the imagined earth return conductors and the overhead lines are “complex numbers”! As a result. In most textbooks on power system engineering. the transmission line parameters are calculated based on the assumption that DC or very low frequency conditions prevail. A. 50 years after the publication of Carson’s method. The earth can be considered as a semi-infinitely extended non-ideal conductor. This method is now called the Carson/Pollaczek model. Gary’s proposition did not receive much attention from the power engineering community. 1926. Because of the symmetry of the circuit. The former is the ratio of the voltage drop per unit length to the current flowing in the conductor and returning through the earth. Carson in the US (Carson. if not properly applied. It expresses the impedance by means of an improper integral that has to be expanded into a infinite series for computation. The currents flowing in the transmission line cause a magnetically induced earth return current that spreads out under the lines. seeking the lowest impedance return path. the distribution of the induced earth return current in the ground is hard to determine. 1931) almost simultaneously published a method for determining the AC transmission line frequency-dependent impedance considering earth return. much longer − 330 − . However. or simply Carson’s method. the complex depth of earth return method has found widespread application in electric power engineering. Papagiannis et al. 1976).

ber x + j bei x = I 0(x j ) . (ber′ mr + j bei ′ mr) + φ(ker ′ mr + j kei′ mr) (3) where Rd = 1/[πσ(r2 − q2)] is the DC resistance of the conductor. (7) B. III. ω is the radian frequency. q and r are the inner and outer radii of the conductor. Calculation of Zg will be discussed in the next section.f. ker′ and kei′ are their derivatives. ker x + j kei x = K 0(x j ) . It is noteworthy that the inner radius q equals zero for a solid line. Kelvin’s functions are often defined as ′ where Dij is the distance between conductors i and j. m = ωµσ is a variable related to the frequency. The variable φ in Eq. bei. Mutual Impedance The mutual impedance Zij of two conductors i and j. The mutual inductance Lij is given by (8) j Z c = R d mr(1 – S 2) 2 × (ber mr + j bei mr) + φ(ker mr + j kei mr) . The internal impedance Zc of the conductor is calculated using a tubular conductor model as shown in Fig. ker ′ mq + j kei′ mq ′ µ 0 D ij ln . equal to 4π × 10−7 H/m by definition. Carson’s Method A. The internal impedance of conductor Zc is given by where I0 and K0 are the modified Bessel functions of the 1st and the 2nd kinds of order zero. 2π D ij (9) (4) where ber. (10) − 331 − . Mathematical Formulation 1. (3) is given by L ij = φ =– ber′ mq + j bei ′ mq . 1992). ri is the radius of the conductor and hi is the average height of the conductor above the ground. bei′. which is zero for a solid line. 2). ker and kei are Kelvin’s functions which belong to the Bessel function family. respectively. The solid line indicates a special case of the model for q equal to zero. The loop inductance is given by (1) Fig. 2π r i (2) in which µ0 is the permeability constant. and the impedance of the ground Zg. 1. respectively (Andrews. leading to the result φ = 0.Impedance of Overhead Transmission Lines computational time and the necessity of training individuals to use the FEM software. The second component is the impedance of the earth return path Zgm that is common to the currents in conductors i and j. and D ij is the distance between conductor i and the image of conductor j (c. which makes application of the tubular conductor model easier. The ground impedance Zg is expressed by Zg = Rg + jXg. The first component is the mutual inductance Lij between the two conductors when the conductors and the ground are perfectly conductive. and ber′. Many scientific Fortran libraries provide subroutines for calculation of Kelvin’s functions with satisfactory accuracy. The impedance of the earth return path Zgm is Zgm = Rgm + jXgm. The tubular conductor model. the internal impedance of line Zc. (5) (6) L ii = µ 0 2h i ln . has two components. both parallel to the ground with their respective heights above the ground being hi and hj. σ is the conductivity and µ is the permeability of the conductor.. Fig. The impedance of conductor i can then be written as Zii = jωLii + Zc + Zg. 1. in which the inner and outer radii q and r of a hollow cylindrical conductor are indicated. S = q/r is the ratio of the inner to the outer radius. The mutual impedance Zij can be written as Zij = jωLij + Zgm. Self Impedance Self impedance includes three components: the reactance of the loop inductance Lii assuming that the line and the earth are perfect conductors.

. 2π D ij where f refers to the frequency and ρ to the resistivity of the ground. Eqs.Y. 6 ′). the earth can be replaced by a set of earth return conductors.}. 3.}. 4. 6. The angle θ is indicated in Fig. sign = +1 when i = 1. Wang and S.. Variables k and km in Eqs.6159315 − ln k) + b1k − d2k2 2 + b3k3 − b4[(C4 − ln k)k4] + . the self and the mutual impedances can be written as b2 = 1 . (17). 7.. the sign of coefficient bi changes every four terms. Allocation of conductors i and j and their images i′ and j′. (12) (13) Xgm = 4ω × 10−7{ 1 (0... 2(hi + p)). sign = −1 when i = 5. Thus. p is the skin depth of the soil..}. (11) − (14)) converge very quickly at low frequencies.J. i (i + 2) (17) (18) (19) (20) Ci = Ci –2 + 1 + 1 . they converge slowly at high frequencies. However. 4 In Eq. (14) Fig. 2. The distance between a conductor and its imagined earth return conductor equals twice its height above the ground plus the skin depth of the ground (i.. The number of terms required to obtain accurate results increases rapidly with the frequency. (11) Xg = 4ω × 10−7{ 1 (0. in which i′′ refers to the imagined earth return conductor of conductor i and p to the skin depth of the ground. 1988) Rg = 4ω × 10−7{ π − b1k + b2[(C2 − ln k)k2] 8 + b3k3 − d4k4 − . Meliopoulos. ri 2π (23) ′′ µ 0 D ij ln . i i +2 C2 = 1.f. Liu Carson’s correction terms for the self and mutual impedances due to the earth path impedances Zg and Zgm are given by (Carson.J. 1926. 2. where b1 = 2 .e. θ = sin– 1(x ij /D ij (22) (15) (16) 2. ′ f /ρ . That is. 2. di = πbi . It is the angle between i − i′ and i − j′..}. The Complex Depth of Earth Return Carson’s infinite series for calculating Zg and Zgm (c.. It must be emphasized that this distance is a complex number since the skin depth p is a complex number.3659315. 8 and so on. The complex depth of earth return method assumes that the current in conductor i returns through an imagined earth path located directly under the original conductor at a depth of (hi + 2p) as shown in Fig. Rgm = 4ω × 10−7{ π − b1kmcosθ 8 2 2 + b2[(C2 − ln km)k m cos2θ + θk m sin2θ] 3 4 + b3k m cos3θ − d4k m cos4θ − . (11) − (14) are frequency-related and are given by k = 4π 5 × 10– 4(2h i ) f /ρ ..6159315 − ln km) + b1kmcosθ 2 − d2k m cos2θ + b3k m cos3θ 2 3 − b4[(C4 − ln 4 km)k m cos4θ + θk 4 m sin4θ] + . 16 bi = bi –2 sign . and is expressed by − 332 − Z ij = j ω µ0 ln 2π (h i + h j + 2p )2 + x 2 ij (h i – h j ) + x 2 ij 2 = jω (24) . 2. k m = 4π 5 × 10 D ij –4 (21) (22) Z ii = j ω µ 0 2(h i + p ) ln + Zc . In other words.

A. The boundary condition of the magnetic field is set to be the Dirichlet condition. The current density of the ground induced current attenuates exponentially as depth increases. Calculation of the transmission line impedance using the FEM requires that an appropriate boundary be defined first. eddy current. Ij (28) The complex power can be written as the product of the voltage − 333 − . In this paper. The geometric mode. For convenience of formulation. When the depth reaches 6|p|. hj. thermal conduction etc. energy storage in electromagnetic fields. a current Ii is applied to conductor i and returns through the ground. electromagnetic forces. power dissipation. we can simultaneously apply currents Ii and Ij to conductors i and j. Fig. Then. Let the voltage drops per unit length of conductors i and j be denoted by Vi and Vj. (27) B. in which the voltage drop per unit length on each conductor is obtained using the FEM and is then divided by the driving current to find the self or mutual impedance. Dij and D ij in Fig. and the current in conductor j is set to be zero. respectively. A similar method can be used to find the self impedance of conductor j: Z jj = R jj + jX jj = P j + jQ j I2 j Ii =0 . assuming that there is no normal component of the magnetic field at the boundary. the current density is reduced to e−6 (approximately 0. permeability. boundary radius and boundary conditions of the FEM. we take Ii = Ij = I and use the FEM to calculate the complex power Pij + jQij dissipated within the boundary. Under this condition. Although Fig. j ωµ 0 (25) ′′ are clearly indicated and the distances hi. When calculating the self impedance of conductor i. the air and the ground. Vi and Vj can be written as Z ii Vi = Vj Z ji Z ij Z jj Ii . We use Fig. the boundary radius has been chosen to be six times (the absolute value of) the skin depth of the ground p. The Finite Element Method The finite element method (FEM) has developed rapidly as a topic of research in electrical engineering.25%) of the density at the surface and can be neglected without causing perceptible errors. All areas are subdivided into first order triangular finite elements.Impedance of Overhead Transmission Lines of two conductors i and j. All bodies (including the air and the ground) included in the boundary are modeled as areas. p. (1999) and Papagiannis et al. the active power Pi and reactive power Qi dissipated within the whole boundary can be used to find the self impedance of conductor i by where p= ρ . 3 to show how the self and the mutual impedances of conductors i and j are formulated using the FEM. the method of calculation is general one and can be extended to calculating the impedances of any number of conductors. Mutual Impedance Having obtained the self impedances Zii and Zjj. The ground can be considered as a semi-infinite conductor. xij. (26) includes the impedances of the conductor and the earth return path. Ij =0 (26) The impedance given by Eq. Self Impedance The FEM is able to calculate the active and reactive power that is dissipated in each area included in the defined boundary at sinusoidal steady-state at a specified frequency. This paper proposes a new approach that combines the energy method and the two-port network theory to obtain the impedance in an indirect way. 3. nonlinearity of magnetization etc. The boundary conditions must then be given to allow Maxwell’s equations to be solved correctly according to the nature of the problem. such as conductivity. 2. Figure 3 shows the boundary and the geometric models Z ii = R ii + jX ii = P i + jQ i I2 i . The FEM formulation proposed in this paper is different from those proposed by Triantafyllidis et al. 3. 3 shows only two conductors. (2000). It is a numerical method that can be used to calculate the electromagnetic fields. Each area has its own geometric shape and physical properties.

Using the relations Ii = Ij = I and Zij = Zji. The shield conductor’s diameter is 14.108 Ω/km. In our illustrative example. Numerical Examples Figure 4 shows the allocation of the conductors of a 500 kV.84 mm. the positive. The bundle spacing is 0. 5. Any of the previously discussed three methods can be used to find the self and mutual impedance of the 14 conductors and to then obtain a 14 × 14 primitive impedance matrix Zp. therefore.374 Ω/km. (32). and that those far away from the conductors are larger. and the DC resistance of each bundle conductor is 0. 4. Primitive Impedance Matrix and Combination of Bundle Conductors The transmission line shown in Fig. all the conductors of the kth group can be treated as a single conductor whose self admittance equals the sum of all the elements of [Ypkm]. given by (30) [Y p 11]4 × 4 [ Y p 21]4 × 4 1 Yp = Z– p = [Y p 31]4 × 4 [Y p 41]2 × 4 [Y p 12]4 × 4 [Y p 22]4 × 4 [Y p 32]4 × 4 [Y p 42]2 × 4 [Y p 13]4 × 4 [Y p 23]4 × 4 [Y p 33]4 × 4 [Y p 43]2 × 4 [Y p 14]4 × 2 [Y p 24]4 × 2 . and the DC resistance of each conductor is 0. Each phase is composed of a four-conductor bundle. the conductors are numbered sequentially from phases A. For the case of k = m. for the case of k ≠ m. It can be seen that the elements near the conductors are smaller.Y. Wang and S. (31) IV. negative. The inverse of Zp yields the primitive admittance matrix Yp. 1. Liu mission line. and current vectors: Pij + jQ ij = [V i V j ] I* i I* j (29) 2 * * = Z ii I 2 i + Z jj I jj + Z ij I i I j + Z ji I j I i . In general. 4 is composed of 14 conductors. so that its impedance matrix was symmetric and could be reduced to the positive-. Two shield wires are located at the top.7 mm. The mutual impedance is.and zero-sequence impedances. Yp has been partitioned according to the conductor allocation into several minor matrices [Ypkm]. The shield wires are numbered 13 and 14. because of their equal voltage. (29) can be reduced to Pij + jQij = (Zii + Zjj + 2Zij)I2. Allocation of conductors for a 500 kV transmission line (unit: m).J. B and C. Meshing of the conductors and their surroundings for finite element analysis. The soil resistivity ρ is assumed to be 100 Ω-m.and the negative-sequence impedances are equal. For convenience of comparison. where the asterisk refers to the conjugate operation. which can be written as Fig. Eq. It is noted here that for a trans- In Eq. Similarly. The bundle conductor’s diameter is 23. which refer to the self admittance matrix of the kth group of conductors of the same voltage if k = m or to the mutual admittance matrix between the kth and the mth groups of conductors if k ≠ m. we assumed that the transmission line fully transposed. [Y p 34]4 × 2 [Y p 44]2 × 2 (32) Z ij = 1 [ 2 P ij + jQ ij I2 – Z ii – Z jj ] .J.45 m. the elements must be made smaller at places where the field gradient or the current density gradient is high to maintain a high level of computational accuracy. Figure 5 shows the meshing of the conductors and their surroundings. the mutual admittance between the kth and the mth groups is the sum of all the elements Fig. three-phase transmission line. The frequency-dependent impedance of the 500 kV line was calculated using the aforementioned three methods. − 334 − .

the positive-sequence resistance and inductance are related to the positive-sequence currents that sum to zero and do not flow through the ground. The FEM. zero-sequence resistance R0 and positive-sequence resistance R1 based on frequencies varying from DC to 2. Sequence Impedances It is assumed in our example that the line is fully transposed. zm zm zs (36) (37) (38) The sequence impedance matrix of the line becomes z0 0 0 Z sys = A – 1Z tp A = 0 z 1 0 . takes into account the proximity effect and may give more accurate results. 6 reveal that the three methods are equally accurate in calculating the ground impedance. on the other hand. Both the FEM and the complex depth of earth Since the electric potential of the shield wires is zero. the conductors of the same voltage can be combined. The differences between the positive-sequence resistances obtained using the three methods increase with frequency. which is given by where A= y r 11 y r 21 Yr = y r 31 y r 41 y r 12 y r 22 y r 32 y r 42 y r 13 y r 23 y r 33 y r 43 y r 14 y r 24 y r 34 y r 44 1 1 1 1 a2 a 1 a a2 (40) . 0 0 z2 (39) − 335 − . (41) (42) where yrkm is equal the sum of all the elements of [Ypkm].000 Hz. z1 = z2 = zs − zm. This is sometimes called the ground mode. The zero-sequence resistance and inductance are related to the zero-sequence currents that flow in the phase conductors and return through the ground. z ca z cb z cc (35) 2. zs zm zm = zm zs zm .000 Hz were calculated using the three methods. Figure 7 is similar to Fig. the complex depth of earth return method and the FEM produced results that are in good agreement with those obtained using Carson’s method. positive-sequence and negativesequence impedances. The resulting impedance matrix is Z tp = 1 3 z aa z ab z ac z bb z bc z ba z ba z bb z bc + z cb z cc z ca z ca z cb z cc z ab z ac z aa z cc z ca z cb + z ac z aa z ab z bc z ba z bb where zs = (zaa + zbb + zcc)/3 zm = (zab + zbc + zca)/3. which is the major part of the zero-sequence impedance. we have z0 = zs + 2zm. Over the frequency range from DC to 2. and z0. 6 except for a wider frequency range from 1 Hz to 105 Hz. 4×4 (33) and a = exp(j2π/3) is the Fortescue operator. On the other hand. and the 14 × 14 admittance matrix Yp can be reduced to a 4 × 4 matrix Yr. respectively. in which both the conductors and the ground are excited. Zr can be reduced to a 3 × 3 matrix Z by eliminating the shield wires: z aa z ab z ac 1 Z = Z LL – Z LT Z – Z = z TT TL ba z bb z bc . Figure 7(b) shows the positive-sequence inductance. The zero-sequence resistance and inductance results obtained using the three methods are in very good agreement. positive-sequence inductance L1. Comparison of the Calculation Results The zero-sequence inductance L0. (39).Impedance of Overhead Transmission Lines of [Ypkm]. The zero-sequence resistance and inductance shown in Fig. The discrepancies in the zero-sequence inductance and resistance results produced by the three methods are very small. The greater discrepancies in the positive-sequence resistance and inductance results may be explained by the proximity effect between conductors that is not considered in the combination of bundle conductors when using Carson’s method or the complex earth return method. Z TL Z TT (34) 3. z1 and z2 are the zero-sequence. 6. The inverse of Yr yields z r 11 z r 21 1 Zr = Y– r = z r 31 z r 41 z r 12 z r 22 z r 32 z r 42 z r 13 z r 23 z r 33 z r 43 z r 14 z r 24 z r 34 z r 44 = 4×4 Z LL Z LT . From Eq. Consequently. The discrepancy in the positive-sequence inductance results is evident at frequencies near 500 Hz. and the results are shown in Fig. This is called the line mode since the ground is not excited.

Y. the complex depth of earth return method (dashed lines) and the FEM (scattered dots) over the frequency range from 0 to 2. 6. Liu Fig. 7. (b) positive-sequence inductance. the complex depth of earth return method (dashed lines) and the FEM (scattered dots) over the frequency range from 1 to 105 Hz. − 336 − . (c) zero-sequence resistance and (d) positive-sequence resistance results obtained using Carson’s method (solid lines).J. Wang and S. Comparison of (a) zero-sequence inductance. Fig.J. (b) positive-sequence inductance. (c) zero-sequence resistance and (d) positive-sequence resistance results obtained Carson’s method (solid lines).000 Hz. Comparison of (a) zero-sequence inductance.

The inconveniences of the FEM include the high cost of software and hardware. This result stems from the use of Carson’s correction terms in Eqs. in the air and in the ground. 1999). We deliberately used only the first eight terms from the infinite series in the calculations. We expect that the current generation of electrical engineers will find a new method for dealing with this old problem. The increased speed and memory capacity of modern computers has greatly reduced the difficulty of calculating infinite series. It applies a numerical method to calculate the distribution of the magnetic fields in the conductors. The complex depth of earth return method is indeed a closed-form approximation of Carson’s method. Conclusions Three methods for calculating the frequency-dependent impedance of overhead power transmission lines have been reviewed in this paper. long computation time and the need Nomenclature Zii Zij µ0 ri hi hj Zc m ω µ Rd r q Zgm Zg dij ′ D ij self impedance of conductor i (Ω/m) mutual impedance of conductors i and j (Ω/m) permeability of the free space (H/m) radius of conductor i (m) height of conductor i (m) height of conductor j (m) internal impedance of conductor (Ω/m) variable used in the Bessel function (m−1) radian frequency (Rad/s) permeability of conductor (H/m) DC resistance (Ω/m) outer radius (m) inner radius (m) Carson’s correction terms of mutual impedance (Ω/m) Carson’s correction terms of self impedance (Ω/m) distance between conductors i and j (m) distance between conductors i and the image of conductor j (m) distance between conductor i and the complex image of conductor j (m) angle included between i − i′ and i − j′ (deg) coefficients of Carson’s infinite series soil resistivity (Ω-m) horizontal distance between conductors i and j (m) self resistance of conductor i (Ω/m) mutual resistance of conductors i and j (Ω/m) self inductance of conductor i (H/m) mutual inductance of conductors i and j (H/m) skin depth of the ground (m) conductor conductivity (Ω-m)−1 active power (W/m) reactive power (Var/m) primitive impedance matrix (Ω/m) primitive admittance matrix (S/m) reduced primitive impedance matrix (Ω/m) reduced primitive admittance matrix (S/m) impedance matrix after eliminating shield wires (Ω/m) impedance matrix of transposed line (Ω/m) sequence impedance matrix (Ω/m) zero-sequence impedance (Ω/m) positive-sequence impedance (Ω/m) ′′ D ij θ bi. it provides an explanation for the imagined earth return conductors. (11) − (14). and the latest method. including the early Carson’s method (Carson.Impedance of Overhead Transmission Lines return method gave very similar results. It can replace Carson’s method in almost all cases in which power line impedance is calculated. The problem of determining the frequency-dependent impedance of an overhead transmission line as discussed in this paper is an example. Besides its contribution to the approximation theory. It is derived neither from Carson’s method. the authors also wish to remind the reader to note the historical path along which the three methods developed. L. Figure 7(d) also shows that the positive-sequence resistance obtained using Carson’s method diverges from those obtained using the FEM and the complex depth of earth return method. but the inductance obtained using Carson’s method failed to converge to a correct value at frequencies beyond 40 kHz. The FEM attacks the problem from a completely different angle.. Numerous efforts have been made by electrical engineers over the past 75 years to find solutions. The FEM is powerful. the FEM (Triantafyllidis et al. di ρ xij Rii Rij Lii Lij p σc P Q Zo Yp Zp Yr Z Ztp Zsys Z0 Z1 References Alvarado. The FEM is new and promising. The complex depth of earth return method is simple. nor from the complex depth of earth return. Carson’s method is important because it gives insight into how and to what degree the earth impedance is influenced by the frequency and the electrical properties of soil. Carson’s method is the classic method for dealing with this problem. 1926). The series converge slowly at high frequencies. Having given a detailed introduction to the three methods and compared them. To obtain acceptable accuracy when using Carson’s method at high frequencies. However. The electric power dissipation and magnetic energy storage in all the areas in the boundary are then obtained and used to indirectly find the impedance. care still needs to be taken when using Carson’s method to calculate high frequency impedance to avoid possible truncation errors. Betancourt (1983) An accurate closed-form approxi- − 337 − . the more recent complex depth of earth return method (Gary. V. An interesting and important point to be made is that it is always possible to find a newer and better solution to an old problem. but the solution is expressed in terms of improper integrals that have to be expanded into infinite series to allow calculation. 1976). It is able to consider the skin effect and proximity effect simultaneously and gives accurate results. ci. F. easy to use and accurate. and R. more terms from the infinite series must be calculated. A 500 kV three-phase transmission line has also been analyzed using the three methods for the purpose of comparison. for experience in using the software. It is still the only method that provides a complete analytical solution to the problem.

!"#$%&'()*+. Cristina. Triantafyllidis.A. IEEE Transactions on Power Systems. 339-359. J.<=>?500 !"#$%&'()* !"#$%&'()"%&*+. Triantafylids. Power Engineering Journal. Meliopoulos. Andrews. 25(4). 2nd Ed. Boston. R./2 !"#$%&'()*+.-.-. G. G. Pollaczek.!"#$ !"#$%&'()* !"#$%&'()*+. P. IEEE Transactions on Magnetics. (1926) Uber das feld einer unendlich langen wechselstromdurchflossenen Einfachleitung. Glover.Y. Gary. MA. and D.. K. Deri.<=>?@ABCDEFG. McGraw-Hill. A. Sarma (1989) Power System Analysis and Design. EDF Bulletin de la Direction des Études et Recherches. 47-56.J. Papagiannis.S.-. and A.<=>?!@ABCDEF"GHI !"#$%&'()*+. 287-293. J. L. F. (1931) Sur le champ produit par un conducteur simple infiniment long parcouru par un courant alternatif. Marcel Dekker. J. Liu mation for ground return impedance calculations. and M. NY. Wang and S. Semlyen. F. C. S. and M. 14(1).S. IEEE Transactions on Power Delivery. P. 279-280. Série B-Réseaux Électriques Matériels Électriques. G. Castanheira (1981) The complex ground return plane: a simplified model for homogeneous and multilayer earth return. A. R. Tevan. Feliziani (1989) A finite element technique for multiconductor cable parameter calculation.S. A. G. Labridis (1999) Calculation of overhead transmission line impedances: a finite element approach. (1988) Power System Grounding and Transients./0!12345"6789:. 100(8). 5-20. New York. U.J.. 15 (1). /0 123456789:. (1976) Approche complète de la propagation multifilaire en haute fréquence par utilisation des matrices complexes.-. C. K. PWS-KENT. NY.A. U. (1992) Special Functions of Mathematics for Engineers. Proceedings of the IEEE. Hill./01234563789:.89:.4567-. Revue Générale d’Électricité. 2986-2988.. (1926) Wave propagation in overhead wires with ground return. 8(1). 29(22). New York. 3686-3693. 33-38. 851-867. 5. G. D. P.3<=>?@3A BCD!E. (1994) Electric railway traction − Part 1 Electric traction and DC traction motor drives. Pollaczek.6H52I !"#$% − 338 − . Papagiannis. U. 3/4. S. 71(2). IEEE Transactions on Power Apparatus and Systems. Labrids (2000) A onestep finite element formulation for the modeling of single and doublecircuit transmission lines. D. 539-554. 3(9). D. Bell System Technical Journal. Carson./0123. Elektrische Nachrichlen Technik. and D.A.