Based on Deutsch assumption, a calculation method on the electric field over the ground surface under HVDC transmission lines in the wind is proposed. Analyzing the wind effects on the electric field and the space charge density the existing method based on Deutsch assumption is improved through adding the wind speed to the ion flow field equations. The programming details are illustrated. The calculation results at zero wind speed are compared with available data to validate the code program. Then the ionized fields which resulted from corona of ±800kV HVDC lines are analyzed. Both the electric field and the current density on the ground level are computed under different wind direction and speed. The computation results are in good agreement with measurements. The presented method and code program can be used to rapidly predict and evaluate the wind effects in HVDC transmission engineering.
1. Introduction
Corona discharges usually occur around the high voltage overhead transmission lines. The space charges created by AC corona are constrained to the vicinity of conductors because of the periodical reversal of electric field. However, the charges created by DC corona fill in the entire space and drift toward either the conductor of opposite polarity or the ground plane. If the space charges move to the ground plane, the groundlevel electric field will be enhanced and can be several times larger than the spacechargefree electric field. Such fields are called ion flow fields.
In the design and construction of high voltage direct current (HVDC) transmission lines, it is necessary to control the electric field over the ground surface and make it satisfy the environment protection requirements
[1]
. The calculation on the electric field has received wide attentions
[2

15]
involving the methods based on Deutsch assumption
[2

7]
and the meshbased methods
[8

14]
. Deutsch assumption supposes that space charges affect only the magnitude of the electric field without changing its direction. In the calculation, the Deutsch assumption based methods avoid mesh division and complex computations and they have fast calculation speed. They have been applied to doublecircuit HVDC transmission lines on the same tower in china
[4]
and achieved engineering benefits. Because the geographical and climatic conditions of the regions covered by the overhead transmission lines are various, they should not be neglected in analyzing accurately the electric fields around the lines. As for the existing meshbased methods, the air factors such as the wind
[10
,
11]
, the rainfall
[12]
and the fog
[13]
have been taken into account. However, they were ignored in the Deutsch assumption based methods.
In this paper, we propose a Deutsch assumption based method to solve the bipolar DC ion flow field within the wind. The equations and boundary conditions describing the ion flow field are presented. A program is coded to solve the boundary value problem. The effectiveness of the program is validated by some measurement results. Finally some characteristics of ion flow field considering the wind speed and direction are given.
2. Equations Considering Wind Effects
The ion flow field of HVDC transmission lines is assumed as follows
[2

7]

(1) The field is twodimensional.

(2) Drifting particles are positive around the positive conductor and negative around the negative one. The particles over the ground surface are either positive or negative and the ion flow field is unipolar.

(3) The thickness of the ion layer is negligible, and the electric field at the conductor surface remains constant which is the corona onset value.

(4) The diffusion of the ions is neglected, and the ion mobility keeps constant.
The equations on the electric field
E_{s}
and the ion current density
J
when the corona appears are described as
where
ρ
is the space charge density,
w
is the wind speed,
ε
_{0}
is the permittivity of air, and
K
is the ion mobility.
The above equations describe a positive ionized field. In the region filled with negative charges, Eq. (2) can be rewritten as
J
=
ρ
(−
K
E
_{s}
+
w
)
The space charges in the ion flow field move along the electric flux lines. The moving speed is
v
=
K
E
_{s}
. Generally,
K
is in the range of (1.2~2.0)×10
^{−4}
m
^{2}
/(V·s), and
E_{s}
is in (0~50)kV/m, so
v
is in (0~10)m/s.
Moreover, according to (2) the wind may well change the ion current density through changing the moving speed of space charges as similar as
E_{s}
does. Historically, Beaufort wind force scale provides an empirical description of wind speed based on observed sea conditions.
Table 1
illustrates the wind speed and the ground state from level 0 to 7. Apparently, the wind speed is comparable with the charge moving speed in the ion flow field. Thus its effects should not be ignored.
Wind classification
3. Solving Method on the Field
 3.1. Boundary value problem
According to the Deutsch assumption, the space charges affect only the magnitude of the electric field rather than its direction. Therefore,
where
E
is the spacechargefree electric field,
A
is a scalar function depending on the distribution of space charges.
E
and
E_{s}
can be expressed by the spacechargefree potential
φ
and the potential
ϕ
in the presence of space charges
The field flux lines are not changed by the presence of space charges in Deutsch assumption, so
where
s
is the distance along the flux line of the spacechargefree field.
From (4), (7) and (8),
Substitute (4) into (1), and then (noting that ∇⋅
E
= 0 )
Replacing
E
⋅∇
A
by
E
d
A
/ d
s
=
ρ
/
ε
_{0}
and according to the relation:
Eq. (10) is transformed into
The space charge density along the flux line may also be derived in a similar way. From (2) and (3),
where
and
Therefore,
Eqs. (9), (11) and (12) describe the distribution of the electric field and the charges within the wind. When the wind speed
w
is set to 0, these equations are in agreement with those without the wind
[14]
. The boundary conditions required to solve these equations are given as follows
where
V
_{0}
is the onset voltage of corona,
A_{e}
is the value of the scalar function
A
at the conductor surface.
The boundary value problem defined by (9), (11), (12) and (13) has no analytical solutions and it should be solved numerically.
 3.2. Programming method
Referring to the program described in
[4]
which has been validated and now widely used in electric power engineering survey and design in China, a program solving the boundary value problem considering wind effects is coded to calculate the electric field over the ground surface under HVDC transmission lines shown in
Fig. 1
.
U
is the voltage applied to the conductors,
N
is the number of subconductors of each bundle conductor,
R
is the radius of each subconductor,
D
is the bundle span,
L
is the polar span and
H
is the height of the bundle conductor axial. The programming details are described as follows.
Configuration of the transmission lines
1) Calculation on the SpaceChargeFree Electric Field: The bundled conductors are usually used in HVDC transmission lines. We use the successive image method to calculate their spacechargefree electric fields as in
[16]
.
2) Calculation on the Electric Field Lines: In the Deutsch assumption based methods, the calculation on the electric field strength and the charge density is along the electric field lines
[2
,
3]
. The electric field at any point outside the pole wires satisfies
where
E_{x}
and
E_{y}
are the field strengths in x and y direction.
Suppose a point
P
(
x
,
y
) is on the electric field line and its close point is
P
′(
x
′,
y
′) . Taking the length
along the field line as a step the coordinates of
P
′ are expressed as
where
λ
= 1 when
P
′ is along the field direction and
λ
= −1 when it is along the opposite field direction.
3) Calculation on the Corona Onset Voltage: The corona onset field strength
E
_{on}
of each subconductor in the transmission line is calculated by the formula in
[17]
where
E
_{on}
is measured by kV/cm,
m
( 0 <
m
≤ 1 ) is the roughness coefficient of the conductor surface which reflects the environment influence on corona discharge, and
R
(cm) is the subconductor radius. Eq. (16) is obtained by modifying the empirical formula proposed in
[18]
according to the practical transmission lines in China and
m
is taken as 0.5 by many tests.
When the maximal field strength
E
_{max}
of each subconductor surface reaches to
E
_{on}
, the corresponding voltage is the corona onset voltage. From (4), the scalar function at each pole surface is
where
is the average value of
E
_{max}
. If the transmission line uses two bundle conductors, and the maximal field strengths on their surfaces are
E
_{max1}
and
E
_{max2}
, then
The corona onset voltage is described as
V
_{0}
=
A
_{e}
V
, where
V
is the voltage applied to the transmission line.
4) Calculation on the Electric Field over the Ground Surface: To obtain the electric field at
P
(
x
,
y
) over the ground surface, (9), (11) and (12) are solved by the secant method.
a) Calculate the potential at
P
(
x
,
y
) and determine the electric field line through this point and the intersection point
O
(
x
,
y
) of this line and the pole conductor surface.
A_{e}
at
O
(
x
,
y
) is calculated by (17).
b) Two initial values of the charge density at
O
(
x
,
y
) are chosen as
ρ
_{e1}
and
ρ
_{e2}
where
f
_{1}
and
f
_{2}
are positive constants and their values can be chosen freely. Here we adopt
f
_{1}
=1.5 and
f
_{2}
=3 which can make the computation have a faster convergence.
ρ
_{m}
is the average charge density
[2]
.
c) For the initial value
ρ
_{e1}
, taking
O
(
x
,
y
) as a start point and considering the boundary condition (13), the two order of RungeKutta method is used to solve (9), (11) and (12).
Finally, the corresponding spacechargefree potential
at
P
(
x
,
y
) is obtained. Similarly,
corresponding to
ρ
_{e2}
is also obtained.
d) The new estimated value
ρ
_{e3}
is calculated by the secant iteration method and its general iteration form is
where let
,
t_{k}
=
ρ
_{ek}
(
k
= 1, 2,3...), thus
e) According to c),
and
A
_{p3}
corresponding to
ρ
_{e3}
are obtained. If the following equations are satisfied, then
ρ
_{e3}
is the estimated true value of charge density
Otherwise, let
ρ
_{e1}
=
ρ
_{e2}
and
ρ
_{e2}
=
ρ
_{e3}
and repeat 3)~5) until both (22) and (23) are satisfied at the same time.
f) Using the true value of the charge density, the electric field and the ion flow density at
P
(
x
,
y
) are calculated out by (1)~(3).
4. Results and Discussion
Using the above programming method, a program has been coded by MATLAB. By setting
w
=
0
, the program is used to calculate the model in
Fig. 1
where
U
= 500 kV,
R
= 2.37 cm,
N
= 4,
D
= 45 cm,
L
= 14 m and
H
= 12.7 m. The calculation results on the electric field over the ground surface are shown in
Fig. 2
, which are in agreement with the results in literature
[14]
, but are a bit different from experimental data. This is mainly attributed to neglecting the different corona characteristics of positive and negative conductors. In addition, the accuracy of experimental data is limited by the test equipment and the environment conditions.
Comparison on the calculation results and experimental data without wind
Inclusive of the wind, the calculation results are compared with the experimental data of ±800 kV HVDC transmission lines at the National Engineering Laboratory for Ultra High Voltage Technology (Kunming) in China Southern Power Grid.
R
=1.68 cm,
N
=6,
D
=45 cm,
L
=22 m and
H
=18 m.
Fig. 3
and
4
show the calculation results on the electric field and the ion current density on the ground level.
Calculation results and experimental data within wind
Calculation results on the ion current density
In the experiments of measuring the fields, the wind speed and direction were not very stable. The wind speed was kept in the range of 13 m/s. The wind remained nearly the same direction blowing from the negative pole to the positive one in two hours. The measured values shown in
Fig.3
at the same point varied randomly.
From
Fig. 3
, when the wind direction changes, the electric field at a distance of more than 20 meters from the center varies greatly. When the wind blows from ‘−’ pole to ‘+’ pole,
w
is assumed to be positive, and we find that the absolute values of the calculated fields under ‘+’ and ‘−’ poles have different variation tendency. They become larger under ‘+’ pole and smaller under ‘−’ pole with the increase of the wind speed. When the wind direction reverses, the variation tendency is opposite to the situations above.
From
Fig. 4
, the variation of ion current density is similar to the electric field. However, when the value of wind speed changes, the maximum value of ion current density varies more significantly, which goes upward due to the increase of wind speed value.
5. Conclusions
We have presented the boundary value problems based on the Deutsch assumption and coded the calculation program on the electric field and the ion flow density caused by HVDC transmission lines in the wind. The conclusions are as follows.

(1) The wind has little effect on the electric field within a certain distance from the center of the transmission lines, while the ion current density on the ground level is easier to be affected by wind. And ion current density tends to grow with the increase of wind speed.

(2) On the lateral sides from the center, the effects of wind speed and direction on the electric field are different under positive and negative conductors. As an external force, the wind tends to move the space charges. The ion current density decreases under the negative pole and increases under the positive pole.

(3) The research work in this paper can also be expanded to evaluate other environment effects concerning dust and atmospheric pressure on HVDC transmission line by the method based on Deutsch assumption.
BIO
Wu Jing She received her Ph.D in Electrical Engineering from Tsinghua University, Beijing, China in 2002. She is working as Associate Professor in School of Automation Science and Electrical Engineering at Beihang University, Beijing, China. Her research interests include the electromagnetic environment of high voltage transmission engineering and the design on the electric machines and apparatus.
Sheng Gao He received his B. S degree in Electrical Engineering from Beihang University, Beijing, China in 2013. He is currently a Master candidate in Electrical Engineering at Beihang University, China. His research interests include highvoltage corona discharge characteristic and electric field calculation.
Yuxiao Liu She received her M. S. degree in Electrical Engineering from Beihang University, Beijing, China in 2014. Her research interests include highvoltage corona discharge characteristic, electric field calculation, ion mechanism in special environment, and nonlinear control.
Sun Z. Y.
,
Zeng L. S.
,
Lu J. Y.
2005
“Guide for HVDC overhead transmission lines, DL/T4362005,” Power industry standard of the People’s Republic of China Tech
Maruvada P. S.
,
Janischewskyj W.
1969
“Analysis of corona losses on DC transmission lines: I  unipolar lines,”
IEEE Trans. Power Apparatus and Systems
88
718 
731
Maruvada P. S.
,
Janischewskyj W.
1969
“Analysis of corona losses on DC transmission lines: II  bipolar lines,”
IEEE Trans. Power Apparatus and Systems
88
1476 
1491
Yang Y.
,
Lu J. Y.
,
Lei Y. Z.
2008
“A calculation method for the electric field under doublecircuit HVDC transmission lines,”
IEEE Trans. Power Delivery
23
1736 
1742
DOI : 10.1109/TPWRD.2008.923051
Yang Y.
,
Lu J. Y.
,
Lei Y. Z.
2010
“A calculation method for the hybrid electric field under UHVAC and UHVDC transmission lines in the same corridor,”
IEEE Trans. Power Delivery
25
1146 
1153
DOI : 10.1109/TPWRD.2009.2036359
1982
“EL2419 HVDC transmission line research,”
EPRI
Zhao T.
,
Sebo S. A.
,
Kasten D. G.
1996
“Calculation of single phase AC and monopolar DC hybrid corona effects,”
IEEE Trans. Power Delivery
11
1454 
1463
DOI : 10.1109/61.517504
Janischewskyj W.
,
Gela G.
1979
“Finite Element Solution for Electric Fields of Coronating DC Transmission Lines,”
IEEE Trans. Power Apparatus and Systems
98
1000 
1012
Takuma T.
,
Tsutomu I.
,
Tadashi K.
1981
“Calculation of ion flow fields of HVDC transmission lines by the finite element method,”
IEEE Trans. Power Apparatus and Systems
100
4802 
4810
Shemshadi A.
,
Niayesh K.
,
Akbari A.
2012
“Influence of the airflow speed along transmission lines on the DC corona discharge loss, using finite element approach,”
Physics of Plasmas
19
073506 
073515
DOI : 10.1063/1.3664310
Yin H.
,
Zhang B.
,
He J.
,
Zeng R.
2012
“Timedomain finite volume method for ionflow field analysis of bipolar highvoltage direct current transmission lines,”
IET Generation, Transmission & Distribution
6
785 
791
DOI : 10.1049/ietgtd.2011.0814
Brahami M.
,
Gourbi A.
,
Tilmatine A.
2011
“Numerical analysis of the induced corona vibrations on highvoltage transmission lines affected by rainfall,”
IET Generation, IEEE Trans. Power Delivery
26
617 
624
DOI : 10.1109/TPWRD.2010.2093606
Zhao Y. S.
,
Zhang W. L.
2013
“Effects of fog on ion flow field under HVDC transmission lines,”
Proceedings of the CSEE
33
194 
199
Yang Y.
2008
“Analysis and calculation of the electric field under Ac and DC transmission lines in the same transmission corridor,”
Beihang University
Beijing, China
Maruvada P. S.
2012
“Electric field and ion current environment of HVDC Transmission lines: comparison of calculations and measurements,”
IEEE Trans. Power Delivery
27
401 
410
DOI : 10.1109/TPWRD.2011.2172003
Maruvada P. S.
1969
“Electrostatic field of a system of parallel cylindrical conductors,”
IEEE Trans. Power Apparatus and Systems
88
1069 
1079
Salam M. A.
,
Shamloul D.
1992
“Computation of ionflow fields of accoronating wires by charge simulation techniques,”
IEEE Trans. Electrical Insulation
27
352 
361
DOI : 10.1109/14.135606
Sarma M. P.
2000
“Corona performance of highvoltage transmission lines.”
Research Studies Press
New York
82 
83