0 оценок0% нашли этот документ полезным (0 голосов)

1 просмотров9 страницAltitude Correction of Radio Interference of HVdc Transmission Lines Part I: Measured Data Analysis and Altitude Correction

Nov 25, 2016

Xie 2016

© © All Rights Reserved

PDF, TXT или читайте онлайн в Scribd

Altitude Correction of Radio Interference of HVdc Transmission Lines Part I: Measured Data Analysis and Altitude Correction

© All Rights Reserved

0 оценок0% нашли этот документ полезным (0 голосов)

1 просмотров9 страницXie 2016

Altitude Correction of Radio Interference of HVdc Transmission Lines Part I: Measured Data Analysis and Altitude Correction

© All Rights Reserved

Вы находитесь на странице: 1из 9

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY

Transmission Lines Part I: Converting Method of

Measured Data

Li Xie, Luxing Zhao, Jiayu Lu, Xiang Cui, Senior Member, IEEE, and Yong Ju

transmission technology in China that the high-voltage directcurrent (HVDC) transmission lines have to pass through the highaltitude areas. To satisfy the demands of environmental protection,

it is important to predict radio interference (RI) performance of the

HVDC transmission lines in the high-altitude areas. The existing

method to evaluate the altitude effect on RI of the HVDC transmission lines refers to the conclusion of the alternating-current (AC)

transmission lines, because of lack of supporting HVDC test data.

It was found by the measurement data on Qinghai-Tibet 500 kV

DC transmission line built in China that the ac altitude correction conclusion of RI is inappropriate for the HVDC transmission

lines in the high-altitude regions of China. In this paper, RI experiments were carried out on two full-scale HVDC test lines and four

reduced-scale dc test lines at different altitudes in China from 2010

to 2014. The method to convert the data on the test lines with different lengths and terminations into that of the infinite-long lines is

proposed and verified; then the measurement data of the full-scale

and reduced-scale test lines were converted by this method to derive the RI altitude correction of the HVDC transmission lines. It

is proved that the proposed method may provide a feasible way to

obtain RI altitude correction of the HVDC transmission lines.

Index TermsAltitude correction, HVDC transmission lines,

radio interference, test lines.

I. INTRODUCTION

N recent years, the HVDC transmission technology has

developed rapidly in China with the highest voltage promoted from 400, 500, 660, 800 kV to the forthcoming

1100 kV. The origins of these HVDC transmission lines in

China are mostly located in the western high-elevation region

with rich natural resource. For example, the highest altitude of

the 800-kV transmission line conveying the southwestern hydroelectric power to eastern China is close to 4000 m, and that

of the Qinghai-Tibet 500-kV transmission line is more than

5000 m above sea level.

Due to the thinner atmosphere at a higher altitude, the corona

discharges on the conductor surface occur more easily, and RI,

as one of the corona performances, is more obvious than that

Manuscript received May 07, 2016; revised July 01, 2016; accepted July 28,

2016. This work was supported in part by National Basic Research Program

of China (973 Program) under Grant 2011CB209402, and in part by SGCC

Science and Technology Project of China under Grant GY71-16-010.

L. Xie, J. Lu, and Y. Ju are with the China Electric Power Research

Institute, Beijing 100192, China (e-mail: xieli@epri.sgcc.com.cn; lujy@

epri.sgcc.com.cn; juyong@epri.sgcc.com.cn).

L. Zhao and X. Cui are with North China Electric Power University, Beijing

102206, China (e-mail: zhaolx@epri.sgcc.com.cn; x.cui@ncepu.edu.cn).

Color versions of one or more of the figures in this paper are available online

at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TEMC.2016.2597300

RI of the HVDC lines at high altitudes satisfies the requirement of environmental protection, the line parameters should be

determined in the design process, and it is usually effective to

increase the line height above ground, the subconductor number

per bundle or the cross section of the subconductors. However,

these measures will cause the construction cost of the project

rise significantly. To make a balance between expense saving

and environmental protection, it is of great importance to predict RI performance of the HVDC transmission lines at high

elevations with more accuracy.

Beginning in the 1950s, study of the altitude effect on corona

phenomena and performances has been carried out for the ac

transmission lines, including corona onset voltage, corona loss,

RI, audible noise, and television interference [2][6]. The most

popular altitude correction formula of RI for ac lines is proposed

by Robertson in 1961, that is, RI will increase by 1 dB with 300

m higher in altitude [4]. However, this ac conclusion is also

widely used in the elevation consideration of the HVDC lines

[7], because of lack of HVDC experimental validations. In recent

years, lots of survey results on the Qinghai-Tibet 500-kV DC

transmission line in China indicated that RI of the practical

HVDC lines at an altitude of around 4000 m is much smaller

than that calculated by the ac altitude correction method [8].

To investigate RI altitude characteristics of the HVDC transmission lines, the experiment data are collected on two full-scale

test lines and four reduced-scale test lines established at different altitudes in China, which are used, respectively, to obtain

the RI difference value between altitude 0 and 4300 m and to

describe how RI difference value varies with the altitude. Due

to the variety in lengths and terminal impedances of the test

lines, the RI measurement data on different test lines cannot be

compared directly. The propagation and reflection process of

corona currents along a transmission line with arbitrary length

and terminations are analyzed in the paper, and one method to

convert the RI data from diverse test lines to that of practical

infinite long lines is presented. Furthermore, by comparing the

converted RI data at different elevations, the RI altitude correction of the HVDC lines is derived, which will be reported in the

Part II, which is a companion paper written by the same authors.

II. TEST LINES AND EXPERIMENTAL DATA DESCRIPTION

A. Test Lines

The full-scale and reduced-scale test lines are two kinds of

fundamental utilities in this research. The full-scale ones, using

0018-9375 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.

See http://www.ieee.org/publications standards/publications/rights/index.html for more information.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2

Fig. 1.

Fig. 3.

Fig. 2.

the same bundles as the practical lines, may reach the identical

corona discharge intensity under the voltage of practical lines.

The reduced-scale ones with smaller cross-section subconductor and less splitting number, simulate the corona performance

of practical lines at the same average maximum surface voltage gradient by imposing a lower voltage. Generally speaking,

the full-scale test lines are the most realistic method except of

the consideration of the huge construction cost. To make a compromise, in this research, two full-scale test lines were built at

the altitudes of 50 and 4300 m, and four reduced-scale ones were

built at the altitudes of 50, 1700, 3400, and 4300 m, respectively.

The Yangbajing full-scale HVDC test line, located at an altitude of 4300 m in Tibet China, comprises two spans shown as

Fig. 1. The conductor bundles 8 JL/G1A-720/50 (diameter

3.623 cm) are strung between the first span of 200 m long as

the positive and negative polar power lead wires, where corona

discharges can hardly occur under the test voltage from 500

to 700 kV. The second section with a length of 300 m is the

main part for measurement purpose, which is strung with bipolar

4 JL/G1A-500/45 (diameter 3 cm) bundles with a pole spacing of 22 m and an average height of 20 m above ground.

The trap coils are installed between the power lead section and

the HVDC generators to prevent the high-frequency current of

the generator from flowing to the test line. The other terminal

of the measurement section is open, and the loop antennas to

measure RI data are arranged on the ground in a line in the

middle of measurement section.

The other full-scale HVDC test line is located in Beijing,

China, with the elevation of 50 m. Most dimensions and configurations of the test line are the same with the aforementioned

one, except of two parameters, shown in Fig. 2. One is that the

length of the measurement section is 605 m. The other one is the

situation of trap coils, which are fixed between the power lead

section and the measurement section, instead of the generators.

The test location section is 150 m away from the open end of

the test line.

The structure parameters of the four reduced-scale test lines

in Beijing, and Xiachayu, Bahe, Yangbajing in Tibet China, are

identical, and the altitudes of the four sites are 50, 1700, 3400,

and 4300 m, respectively. Bipolar 4 JL/G1A-95/15 (diameter

1.36cm) bundles with a pole spacing of 6 m, an average height

One terminal of all the test lines is open, and the other terminal

is connected with the trap coils. The scheme of the structure

configuration is shown in Fig. 3.

Although the RI altitude correction values of the conductors on the full-scale and reduced-scale test lines are somewhat

different under the same surface voltage gradient, the function

types to describe their nonlinear relations with the change of

elevation can be regarded as the same. Therefore, the purpose

of utilizing two kinds of test lines is to derive the deviation

of RI level between altitude 50 and 4300 m from the data of

the full-scale one, and to obtain the nonlinear function type

of the RI deviation and height above sea level by fitting the 50,

1700, 3400, and 4300 m RI measurement data from the reducedscale ones. Moreover, substituting the difference value between

50 and 4300 m into the derived nonlinear function to determine

the parameters yields the variation regularity of the RI altitude

correction values for the practical transmission lines.

In the following sections, the corona intensities on the conductor surfaces of these two kinds of test lines are supposed to

be identical under the same voltage gradient on the conductor

surface, and the corona performance difference of the two conductors is ignored, which is reasonable within the engineering

range of the conductor surface voltage gradient.

B. RI Measurement Results on Test Lines

The RI frequency spectral curves measured on the test lines

are different with that on the practical lines. The practical ones

may be regarded as infinitely long, so the ideal shape of their

RI frequency spectral curves is monotonic decreasing, and the

attenuation satisfies frequency modified formula [9]. However,

the electromagnetic waves propagating on the test lines with

finite lengths will reflect at both terminals and generate the

standing wave effect, and the RI frequency spectrum oscillates

with the frequency.

Figs. 4 and 5 show the RI frequency spectra under various different voltages measured separately in Beijing and Yangbajing

full-scale test lines, and Fig. 6 shows the measured RI frequency

spectra of the four reduced-scale test lines at different altitudes

under the same voltage.

Shown as these figures, it is obvious that the oscillation characteristics of RI spectra depend neither on the voltage nor on

the altitude, but on the lengths of the test lines and the terminal impedances. Consequently, it is necessary to convert the

measurement data on test lines with different lengths and terminations into RI level of infinitely long lines, instead of comparing them directly. The method to calculate the RI propagation

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: ALTITUDE CORRECTION OF RADIO INTERFERENCE OF HVDC TRANSMISSION LINES PART I

transmission line.

impedances is introduced in section III.

Fig. 4.

50 m).

TRANSMISSION LINES WITH ARBITRARY LENGTH AND

TERMINATIONS

Suppose that the HVDC transmission line is a n + 1conductor line with the length of L and the (n + 1)th conductor

is the ground. A single corona source I0i is injected at the position x1 of the ith conductor, shown as Fig. 7. Currents will also

be induced on the other conductors at the same location. The

propagation of currents and voltages on the line can be described

by the following the transmission line equations [10]:

dV (x)

+ (R + jL) I (x) = 0,

dx

dI (x)

+ (G + jC) V (x) = I 0 (x x1 ) .

dx

of 4300 m).

(1)

R,L,C, and G are the per-unit-length resistance, inductance, capacitance, and conductance matrices of the line, respectively; I 0 is the vector of all the induced current given by

I 0 = CI 0 /20 ; I 0 is the vector of the injected corona current,

where I0i is the ith element and the rest elements are zeroes; ()

is the Dirac function. The terms R + jL and G + jC are defined as the per-unit-length impedance matrix Z and admittance

matrix Y .

The line is connected with the impedances Z 1 and Z 2 on

both terminals, and the following relations exist:

V (L/ 2) = Z 1 I (L/ 2) ,

V (L/ 2) = Z 2 I (L/ 2)

Fig. 6. RI spectra measured on four reduced-scale test lines under the same

voltage.

(2)

Assuming that the matrix product ZY is diagonalizable and M is the transformation matrix, it can be written as H = M 1 ZY M . If H = D 2 ,D is also a diagonal matrix. Define the propagation matrix of the line as

= M DM 1 , the characteristic impedance matrix as Z c =

Y 1 = 1 Z, and the reflection coefficient matrices at

each terminal of the line as P 1 = (Z 1 Z c ) (Z 1 + Z c )1

and P 2 = (Z 2 Z c ) (Z 2 + Z c )1 . Then, the terminal voltage

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4

1

1 (U + P 1 )

P 1 e L

0

V (L/ 2)

=

0

(U + P 2 ) e L P 2

V (L/ 2)

2

(x +L / 2)

Z c I 0

e 1

(3)

e (L / 2x 1 ) Z c I 0

density of I0i . Due to the uniform distribution assumption of

injected currents, i is a constant along conductor i. Therefore,

the power spectral density of the RI electric field at (x, yj , z0 )

produced by all the corona sources along conductor i is

x

|qj i (x1 , x)|2 dx1

j (x) = i

L /2

at arbitrary point of the line can be written as

I (x) = Z 1

e (x+L / 2) P 1 e (x+L / 2)

c

(U +P 1 )1 V (L/ 2) , x < x1 ,

(L / 2x)

(L / 2x)

I (x) = Z 1

e

e

P

2

c

(U +P 2 )1 V (L/ 2) , x > x1 .

(4)

(5)

L /2

I (x) = G (x1 , x) I0 ,

x1 < x

(6)

I (x) = G (x1 , x) I0 ,

x1 > x.

(7)

Then, the vector consisted of the RI electric field x-axis components at different N points (x, yi , z0 ) (i = 1, 2, . . . , n) above

the ground can be written as [9]

[Ex (x, yi )]m = Z0 KG (x1 , x) I 0 ,

x1 < x,

x1 > x.

h z

h j +z 0 +2p

1

Kij = 2

(h z ) 2j+(y0 y ) 2 + (h +z +2p)

2

+(y

j

0j

2

0 j y i )

i, j = 1, 2, . . . , n

x1 < x

(8)

x1 > x.

(9)

random series pulse signals, which may be regarded as uniform

distribution on the surfaces, their mutual effect can be analyzed

in the term of power spectral density [10], [12]. Define 0i

and j (x) as the power spectral densities of the corona current

I0i injected into the ith conductor at the location of x1 and

produced RI electric field Ex (x, yj ) by I0i at (x, yj , z0 ), then

the relationship between j (x) and 0i are given by

j (x) = |qj i (x1 , x)|2 0i

2

j (x) = q j i (x1 , x) 0i

L /2

L /2

2

qj i (x1 , x) dx1 ,

j = 1, 2, . . . , n.

x1 < x,

x1 > x

Q(x1 , x) and Q (x1 , x).

(11)

Assume that the function i corresponds to the power spectrum density i , then i is the RI excitation function of conductor i, with the unit A/m1/2 . i is related to corona discharge

intensity on the conductor surface, so it is approximately equivalent on the whole conductor [13][15].

In this way, the RI electric field may be written as

x

2

|Ex (x, yi )|

=

|qij |2 dx1

L /2

L /2

+

x

where hj and y0j are the z and y coordinates of the jth conductor center, and p is the penetration depth of the field.

p= [0 /(0 f )]1/2 , where 0 is the resistivity of the earth.

Consequently, RI electric field in the above equations can be

denoted by

(10)

If corona also exists on the other conductors with the perunit-length power spectral densities of 1 , 2 , . . . ,n , the power

spectrum density of their RI electric fields at (x, yj , z0 ) is

x

n

i

|qj i (x1 , x)|2 dx1

j (x) =

i=1

Substituting the results of (3) into (4) and (5), the relation of

I(x) and I0 can be denoted by

qj i (x1 , x)2 dx1

2

qij dx1

2i n

m n

i, j = 1, 2, . . . , n.

(12)

terminal impedance has no effect on the RI electric field in measurement section. To simplify the analysis, it may be considered

that the line is connected with the characteristic impedance on

both terminals, where exists P 1 = P 2 = 0. According to the

aforesaid process, the RI electric field expression of infinitely

long transmission lines may be obtained.

IV. RI CONVERSION METHOD FROM SHORT TO LONG LINES

A. RI Conversion Method

It has been shown that the RI level mainly depends on two

properties of the lines: RI generation and propagation [16].

Among them, RI generation is the reflection of mean corona

intensity on per-unit-length line, expressed by the RI excitation

function in (12), and RI propagation includes the processes on

the line and in the space, shown by the m n matrix on the

right-hand side of (12). However, the influence of elevation to

RI is mainly embodied in the generation instead of propagation.

Therefore, the premise of RI conversion between long and short

lines is that they have the same RI excitation function.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: ALTITUDE CORRECTION OF RADIO INTERFERENCE OF HVDC TRANSMISSION LINES PART I

perpendicular to the line, the squares of the rms values of RI

electric field may be expressed by the linear combination of the

excitation function squares, that is

2 S

(13)

Ex,i n = F S 2i n ,

2 L

(14)

Ex,i n = F L i n

where F S and F L represent the transfer function matrices from

excitation functions to the RI electric field for short lines and

long lines, respectively. If the excitation functions in these two

equations are identical, the RI electric fields of long and short

lines have the following relationship:

2 S

2 L

Ex,i n .

(15)

Ex,i n = F L F 1

S

Due to the terminal reflection of the lines, the RI electric field

of a short test line vibrates with frequency. However, the frequency characteristics of the RI electric field of practical long

lines are smoothly declining, since the terminal reflection may

be ignored. The effect of these two aspects leads to the oscillation

1

of F L F 1

S with frequency. Theoretically, F L F S at a certain

frequency is the corresponding conversion coefficient matrix,

but the experiment results at a single frequency are subject to be

influenced by the external interference. Especially, at the reference frequency of 0.5 MHz, the unintentional interference from

human activities and natural environment are very common, so

it is a better choice to make the RI conversion in a certain band

of frequency. Moreover, during the process of RI conversion,

a smooth frequency spectral curve related to the oscillating RI

frequency spectrum of the short line is needed as the base line,

representing the overall level of the frequency spectrum. The

maximum and minimum envelope curves are easy to derive.

It was presented in [17] and [18] that RI frequency spectrum

of the long lines is equal to the geometric mean of the maximum

and minimum frequency spectral envelopes of the short openended lines. The result is deduced for the single-conductor lines,

which may cause errors in the case of multiconductor lines.

On the other hand, the method in [17] and [18] adopt modal

transformation for the multiconductor transmission lines and

make inverse transformation after analysis of power spectral

density. However, it is obvious that the latter one is not a linear

transformation, so the order of the two processes cannot be

exchanged anyway.

Additionally, the measurement data show that the maximum

values are more likely to be disturbed, so its envelope is not

suitable to be the base line. As a result, the minimum envelope

is adopted as follows.

The RI conversion method from short lines to long lines

includes the following five steps.

First, measure the RI electric field frequency spectra on a

certain frequency range at n ground points perpendicular to the

line, denoted by ES,1 (f ), ES,2 (f ), . . . , ES , n(f ), where n is the

bundle conductor number generating RI.

Second, calculate the frequency characteristics of every element in the transfer function matrices F S and F L , denoted by

FS,ij (f ) and FL,ij (f )(i, j = 1, 2, . . . , n).

wires at Yangbajing full-scale test line.

elements in F S and the RI electric field, search for the frequencies corresponding to their mutual minimum values, denoted by

f1 , f2 , . . . ,fk , on which the minimum values also have good

frequency attenuation characteristics.

Fourth, make curve fittings based on the values of every

element at these frequencies FS,ij (f1 ), FS, i j (f2 ), . . . , FS,ij (fk ),

then get the minimum frequency spectra envelope

FSM ,ij (f )(i, j = 1, 2, . . . , n); also, make curve fittings

based on the RI electric field values at these frequencies and

get the frequency spectra ESM ,i (f )(i = 1, 2, . . . , n).

Finally, arrange all the values of FSM ,ij (f )(i, j = 1, 2, . . . , n)

at the frequency f0 to form a matrix, denoted by F SM (f0 ), and

arrange all the values of ESM , i(f0 )(i = 1, 2, . . . , n) to form a

column vector, denoted by E SM (f0 ), then the RI electric fields

at all the locations at f0 of the long line may be expressed

by

E L (f0 ) = F L (f0 ) F 1

SM (f0 ) E SM (f0 ).

(16)

short lines to long lines is obtained by analysis with rms values,

while the RI measurement and evaluation always use the quasipeak values. According to [19] and [20], when the repetition

frequency of a pulse signal changes within a certain range, the

rate of quasi-peak value and rms value is almost a constant.

Therefore, altitude correction coefficient of RI obtained from

rms values may be regard as the same as that from the quasipeak values.

This section will introduce the RI conversion process based

on the measurement data of Yangbajing full-scale test line. The

background RI noise has been filtered out of the measurement

data.

It was found that the difference between RI levels under positive and negative conductors are always over 6 dB

whether in Beijing or Tibet, so the RI measurement should

be mainly carried out under the positive conductor, and the

RI produced by the negative conductor may be ignored in the

data analysis.

According to the structure parameters of Yangbajing fullscale test line, the reference directions of the physical quantities

of the power lead wires are shown in Fig. 8.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6

on both terminals of the lead wires is as follows [12]:

1

1

Zyc 0

U

y (Ly )

Iy1

=

Iy2

U

y (Ly )

0 Z 1

yc

Vy 1

U

y (Ly )

.

(17)

U

Vy 2

y (Ly )

I y 1 and I y 2 are current vectors on both terminals of the

wires; V y 1 and V y 2 are voltage vectors on both terminals of

the wires; Z y c is the characteristic impedance matrix of the lead

wires; y (Ly ) is the propagation matrix of the lead wires; and

Ly is the length of lead wires. Let X represents the following

matrix:

X 11 X 12

X=

X 21 X 22

1

Z 1

0

U

y (Ly )

yc

=

U

y (Ly )

0 Z 1

yc

U

y (Ly )

U

y (Ly )

1

.

(18)

wires there approximately exists I y 1 = 0. Substituting the relationship into (17), we may obtain V y 2 = X 22 I y2 . Moreover,

the equivalent impedance matrix of the power lead wires and

HVDC power supply is

Z 1 = X 22 .

(19)

For the measuring section of the test line, the lead wires

connected to the terminal may be regard as impedance Z1 , and

the another terminal is open.

Assuming that the excitation function of the positive polar

conductor under arbitrary frequency is equal to 0 dB(A/m1/2 ),

the comparison of calculation and experiment frequency spectral curves is shown in Fig. 9(a). It is found that the values of

the excitation functions used in the calculation do not change

with frequency, and part of the minimum values on the frequency

spectral curve remained numerically stable. When the actual excitation function declines with frequency, the minimum values

at the same frequencies on the RI measured curve also decrease.

Thus, it is inferred that the decreasing trend of the values on

these frequencies reflects the attenuation characteristics of the

excitation function, as well as the frequency spectral features

of the RI level of infinite long lines. Therefore, we may select

some minimum values from the calculated frequency spectrum

of transfer function F s s element, which are relatively stable,

shown in Fig. 9(b), and then correspond them to the RI measurement frequency spectral curve shown in Fig. 9(a) to get the

marked frequency points in the figure.

To make the curve fitting for the selected minimum values on

measurement RI frequency spectrum shown in Fig. 9(a), we may

adopt a similar form with the RI frequency modified formula

given by

Em in = a + b(lg 10f )2

the comparison of the calculated and measured data. (a) Comparison of the

calculated and measured RI frequency curves. (b) Selection of minimum values

on FS .

(20)

Fig. 10. RI electric field spectrum of long lines converted from the fitting

curve of minimum values on the measured RI spectrum of the test line.

where f is frequency with the unit MHz, and a, b are undetermined coefficients. The fitting curve for the selected values in

Fig. 9(a) is shown in Fig. 10. Then, using a constant to match

the values selected in Fig. 9(b) and according to the process presented in Section 4.1, the converted RI electric field frequency

spectrum of infinitely long lines is also shown in Fig. 10.

V. VERIFICATION

A. Verification of the RI Propagation Process

To verify the RI propagation process presented in this paper,

the calculated and measured results are compared under the

assumption that the excitation function of the positive conductor

is 0 dB(A/m1/2 ) for any frequency, and that of the negative one

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: ALTITUDE CORRECTION OF RADIO INTERFERENCE OF HVDC TRANSMISSION LINES PART I

Fig. 11. Comparison of the calculated and measured RI spectrums for Beijing

full-scale test line (altitude 50 m).

Fig. 13. Comparison of the calculated and measured RI spectrums for Bahe

reduced-scale test line (altitude 3400 m).

Fig. 14. Comparison of the RI data between the converted data from test line

and measurement data on Hami-Zhengzhou 800-kV transmission line.

Fig. 12. Comparison of the calculated and measured RI spectrums for

Yangbajing full-scale test line (altitude 4300 m).

data are all for the location just under the positive conductor.

The comparisons for Beijing full-scale test line (altitude

50 m), Yangbajing full-scale test line (altitude 4300 m) and

Bahe reduced-scale test line (altitude 3400 m) are shown in

Figs. 1113.

In these figures, the frequencies corresponding to the peak

and valley values of calculated and measured curves are very

close. Since the oscillation properties of RI frequency spectrum

is mainly determined by the propagation, the comparison results

can verify the effectiveness of the analysis method of the RI

propagation.

It should be emphasized that the differences between the amplitudes of calculated and measured curves in Figs. 1113 are

caused by the lack of knowledge of the practical RI excitation

function values in the model, and that it is only assumed as

0 dB(A/m1/2 ) in the calculating process. Therefore, we can

use the proposed minimum values envelope method to describe

the amplitude difference, that is, to determine the spectrum of

practical RI excitation function. If the derived spectrum is substituted into the calculating process again, the calculated and

measured curves will match with each other not only in the

curve trends but also in the amplitudes.

The amplitude difference between calculated and measured

data in Fig. 11 are less than the difference in Fig. 12, which

means the value of the RI excitation function is corresponding

with the elevation. The difference of the RI excitation function

values derived, respectively, from Figs. 11 and 12 reflects the

RI correction factor from altitude 50 to 4300 m. However, the

RI excitation function value derived from Fig. 13 cannot be

compared with Figs. 11 and 12, because it is the value for

reduced-scale test line with thinner conductor. Nevertheless, the

elevation correction factor of the reduced-scale test lines can be

compared with that of the full-scale test lines within a certain

range.

B. Verification of the RI Conversion Method From Short Lines

to Long Lines

The RI conversion method from short lines to long lines is

validated in this section. Shown in Fig. 14, the RI data measured

on a test line are converted to the RI level of a practical long

line, and it is compared with the RI data of Hami-Zhengzhou

800-kV HVDC transmission line in China, which has the same

configuration and parameters as the test line.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

8

spectrum of Hami-Zhengzhou 800-kV HVDC transmission

line due to the external interference. However, within the range

of 0.152 MHz, the RI frequency spectrum converted from the

experimental results of the short line is in a good consistency

with the undisturbed RI values of the practical long line. Therefore, the effectiveness of the presented RI conversion method is

verified.

In the Part II, which is a companion paper written by the same

authors, abundant RI measurement data on test lines at different

altitudes will be converted by the method proposed in this paper,

and the altitude correction will be discussed in detail.

VI. CONCLUSION

Two full-scale HVDC test lines at the altitudes of 50 and

4300 m, and four reduced-scale HVDC test lines at the altitudes

of 50, 1700, 3400, and 4300 m are utilized to get the RI experimental data in the paper. The experimental data of two kinds of

test lines are used, respectively, to obtain the RI deviation value

between altitude 0 and 4300 m and to describe the nonlinear

relation of the RI deviation to the altitude, then to conclude the

variation regularity of the RI altitude correction values for the

practical HVDC transmission lines.

The oscillation properties of the RI frequency spectral curve

are associated with lengths and terminal loads of the test lines,

and the amplitudes of the curve are related to the test voltages

and altitudes. The RI experimental data of the test lines with

different lengths and terminals should be converted to the RI

level of the infinite long lines before comparing to get the altitude

deviation.

The method to analyze the RI propagation on the transmission

lines with arbitrary lengths and loads is presented. Furthermore,

the RI conversion method from short lines to long lines is discussed. By comparison of the measured and calculated data, the

calculation method and the conversion method are verified. The

presented method to calculate the RI propagation and to convert

the RI data from short lines to long lines will be utilized to get

the altitude correction method in Part II, which is a companion

paper written by the same authors.

REFERENCES

[1] F. W. Peek, Dielectric Phenomena In High Voltage Engineering.

New York, NY, USA: McGraw-Hill, 1929.

[2] L. M. Robertson and J. K. Dillard, Leadville high-altitude extra-highvoltage test project, part Ireport on 4 years of testing, IEEE Trans.

Power Appl. Syst., vol. PAS-80, no. 3, pp. 715725, Dec. 1961.

[3] D. F. Shankle, J. C. Smith, and J. E. ONeil, Leadville high-altitude

extra-high-voltage test project, part IIcorona loss investigations, IEEE

Trans. Power Appl. Syst., vol. PAS-80, no. 3, pp. 725731, Dec. 1961.

[4] L. M. Robertson, W. E. Pakala, and E. R. Taylor, Leadville highaltitude extra-high-voltage test project, part IIIradio influence investigations," IEEE Trans. Power Appl. Syst., vol. PAS-80, no. 3, pp. 732743,

Dec. 1961.

[5] L. M. Robertson, C. F. Wagner, and T. J. Bliss, Colorado high-altitude

corona tests, part Iscope, tests and Instrumentation, AIEE Trans.,

vol. 76, no. 3, pp. 356365, Jun. 1957.

[6] V. L. Chartier, L. Y. Lee, L. D. Dickson, and K. E. Martin, Effect of high

altitude on high voltage AC transmission line corona phenomena, IEEE

Trans. Power Del., vol. PD-2, no. 1, pp. 225237, Jan.. 1987.

[7] EPRI, HVDC Handbook. Palo Alto, CA, USA: Electric Power Research

Institute, 1993.

Under 500kV DC Transmission Line From Qinghai to Tibet. Beijing,

China: China Electric Power Research Institute, Jan, 2010.

[9] Radio Interference Characteristics of Overhead Power Lines and High

Voltage Equipment, Part1: Description of Phenomena, IEC Std. CISPR

18-1, pp. 1314, 2010.

[10] P. S. Maruvada, Corona Performance of High-Voltage Transmission Lines.

Baldock, U.K.: Research Studies Press Ltd, 2000.

[11] F. M. Tesche, M. V. Ianoz, and T. Karlsson, EMC Analysis Methods and

Computational Models. New York, NY, USA: Wiley, 1997.

[12] A. Papoulis, Signal Analysis. New York, NY, USA: McGraw-Hill, 1977.

[13] C. H. Gary and M. R. Moreau, Predetermination of the radio noise level

under rain of an extra-high-voltage line, IEEE Trans. Power Appl. Syst.,

vol. PAS-88, no. 5, pp. 653660, May 1969.

[14] C. H. Gary, The theory of excitation function: A demonstration of its

physical meaning, IEEE Trans. Power Appl. Syst., vol. PAS-91, no. 1,

pp. 305310, Jan./Feb. 1972.

[15] M. R. Moreau and C. H. Gary, Predetermination of the radio interference level of high voltage transmission lines IPredetermination of

the excitation function, IEEE Trans. Power Appl. Syst., vol. 91, no. 1,

pp. 284291, 1972.

[16] Transmission Line Reference Book 345 kV and Above, Palo Alto, CA,

USA: Electric Power Research Institute, 1982.

[17] R. D. Dallaire and P. S. Maruvada, Analysis of radio interference from

short multiconductor lines Part 1. Theoretical analysis, IEEE Trans.

Power Appl. Syst., vol. PAS-100, no. 4, pp. 21002108, Apr. 1981.

[18] R. D. Dallaire and S. Maruvada, Analysis of radio interference from

short multiconductor lines Part 2. Analytical and test results, IEEE Trans.

Power Appl. Syst., vol. PAS-100, no. 4, pp. 21092119, Apr. 1981.

[19] F. Harber, Response of quasi-peak detector to periodic impulses with

random amplitudes, IEEE Trans. Electromagn. Compat., vol. EC-9,

no. 1, pp. 16, Mar. 1967.

[20] O. Nigol, Analysis of radio noise from high-voltage lines I meter response to corona pulses, IEEE Trans. Power Appl. Syst., vol. PAS-83,

no. 5, pp. 524533, Jan. 1964.

received the B.S., M.S., and Ph.D. degrees in electrical engineering from Beihang University, Beijing,

China, in 2002, 2005 and 2010, respectively.

She is currently at China Electric Power Research

Institute, Beijing. Her main research interest is electromagnetic environment in power systems.

received the B.S. and M.S. degrees in electrical engineering from Wuhan University, Wuhan, China, in

2006 and 2008, respectively. He is also currently

working toward the Ph.D. degree in North China

Electric Power University, Beijing, China.

His main research interest is electromagnetic environment in power systems.

received the B.S. degree in electrical engineering

from Shanghai Jiaotong University, Shanghai, China,

in 1982, the M.S. degree in electrical engineering

from Huazhong University of Science and Technology, Wuhan, China in 1986, and the Ph.D. degree in

electrical engineering from Xian Jiaotong University, Xian, China in 1996.

He is currently at China Electric Power Research

Institute, Beijing. His main research interests are the

theory and calculation of electromagnetic fields and

electromagnetic compatibility.

XIE et al.: ALTITUDE CORRECTION OF RADIO INTERFERENCE OF HVDC TRANSMISSION LINES PART I

Hebei Province, China, in 1960. He received the

B.Sc. and M.Sc. degrees in electrical engineering

from North China Electric Power University, Baoding, China, in 1982 and 1984, respectively, and

the Ph.D. degree in accelerator physics from the

China Institute of Atomic Energy, Beijing, China,

in 1988.

He is currently a Professor and the Vice Director of State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources,

North China Electric Power University, Beijing, China. His research interests

include computational electromagnetics, electromagnetic environment and electromagnetic compatibility in power systems, insulation, and magnetic problems

in high-voltage apparatus.

Dr. Cui is a Standing Council Member of the China Electrotechnical Society

and a Fellow of IET. He is also an Associate Editor of IEEE TRANSACTIONS ON

ELECTROMAGNETIC COMPATIBILITY.

Yong Ju was born in Jilin, China, in 1977. He received the B.S. and M.S. degrees in electrical engineering from North China Electric Power University,

Baoding, China, in 2000 and 2003, respectively.

He is currently at China Electric Power Research

Institute, Beijing, China. His main research interest

is electromagnetic environment in power systems.

## Гораздо больше, чем просто документы.

Откройте для себя все, что может предложить Scribd, включая книги и аудиокниги от крупных издательств.

Отменить можно в любой момент.