IET Generation, Transmission & Distribution

Research Article

Modelling for switching impulse breakdown of ISSN 1751-8687

Received on 29th May 2019

live working gaps between equipotential

Revised 3rd September 2019
Accepted on 3rd January 2020
E-First on 28th February 2020
worker and transmission towers doi: 10.1049/iet-gtd.2019.0789
www.ietdl.org

Yaqi Fang1 , Linong Wang2, Rui Li3, Jiachen Gao2, Bin Song2, Kai Liu4
1School of Electrical and Electronic Engineering, Hubei University of Technology, Wuhan, People's Republic of China
2School of Electrical Engineering and Automation, Wuhan University, Wuhan, People's Republic of China
3Wuhan Power Supply Company, State Grid Hubei Electric Power Company Limited, Wuhan, People's Republic of China
4China Electric Power Research Institute, Wuhan, People's Republic of China

E-mail: fyq@hbut.edu.cn

Abstract: Air gap discharge voltage is the key parameter to determine the minimum approach distance for live working. In this
study, a mathematical–physical model for predicting 50% discharge voltage of the transmission line air gaps considering the
effect of live line worker is developed based on the continuous leader inception criterion. Also, the tower structure coefficient ST
is proposed to represent the influence of various tower structures and ST is calculated by finite element method. Finally, the
switching impulse discharge tests of the equipotential worker-tower structure gaps are carried out on 750 kV transmission lines
to verify the proposed model. The research results show that the errors between the calculation and experimental data are
within ±4.03% which demonstrates the feasibility of discharge voltage prediction for complicated live working gaps. This model
offers a possible way to determine the safety air gap distance for live working by analytical and numerical calculation rather than
costly and time-consuming experiments.

1 Introduction of discharge models has also been proposed. For engineering

application, the Critical Corona Radius (CCR) method proposed by
Live working is an important technical means to ensure the safety Carrara and Thione [12] and the continuous leader inception model
operation of the power grid. Due to the large tower size of extra- developed by Rizk [13, 14] are widely used. Carrara and Thione
high voltage (EHV) and ultra-high voltage (UHV) transmission [12] found that under the critical switching impulse, when the
lines, the bare-hand method is the most widely used method in live radius of the electrode is less than CCR, the breakdown voltage of
working on these transmission lines [1]. For the bare-hand method, the gap does not change significantly with the decrease of the
the worker is at the same potential with the live conductor. The electrode radius, and the minimum breakdown voltage of air gaps
existence of the live line worker can change the gap configuration, is determined by the continuous leader inception voltage and the
which significantly affects the insulation properties of the air gaps voltage drop of the leader channel. However, there is a limitation
[2]. In order to ensure the safety of the worker, the minimum of this model that the CCR must to be obtained via experiments.
approach distance (MAD) must be obtained before conducting live Later, for the electrode with a radius less than CCR, Rizk proposed
working. At present, the MAD is determined based on the the general model to calculate the air gap breakdown voltage with
acquisition of gap factor via full-scale discharge tests, and then the rod or conductor electrodes [14], the concept of ‘gap characteristic
MAD can be calculated using the method recommended by IEC function R’ is proposed to present the gap configuration on leader
standard 61472 [3]. inception voltage. However, it is hard to obtain function R with
Live working gap refers to the gaps between the equipotential irregular electrode. With the continuous understanding of discharge
worker and tower or the gaps between the bundle conductor and mechanism, based on the experimental observations and some
ground-potential worker. However, due to the existence of the reasonable simplifying assumptions, researchers attempt to
worker, the discharge characteristics of live working gap are describe the discharge processes occurring in each discharge phase
different from the conventional conductor-tower gaps. At present, (first corona, streamer propagation, leader formation and
the discharge voltage of live working gap is obtained by the full- propagation, final jump) in a mathematical way and some physical
scale test. With the construction of EHV and UHV transmission models are developed such as Bondiou–Gallimberti model [15],
projects, a series of experimental studies related to discharge Becerra–Cooray model [16] and Fofana–Beroual model [17].
characteristics for live working gaps have been conducted on a Physical models can reflect the physical characteristics of
single circuit, double-circuit and compact transmission lines [4– discharge development. However, due to the complexity of the
10]. For different live working conditions (the worker at different phenomena involved in the discharge, it is difficult to simulate the
locations) or new tower structures, discharge tests must be whole discharge processes accurately especially for the
performed repeatedly to obtain its 50% discharge voltage and the complicated electrode structures.
gap factor. However, these tests are extremely time-consuming and In this paper, considering the overall effect of live line worker,
costly because of the complicated gap configurations. Until now, bundle conductors and different tower structures, a mathematical–
there is no suitable prediction method that can be applied directly physical model for predicting the 50% discharge voltage of live
to analyse the discharge characteristics of live working gaps. working gap is developed based on the continuous leader inception
Therefore, it is necessary to establish a model to predict the criterion. Also, the switching impulse discharge tests are carried
discharge voltage of live working gaps so as to reduce the full- out on 750 kV transmission lines to verify the correctness of the
scale test task. model. The major contributions of this paper are given as follows:
Live working gaps of EHV and UHV transmission lines belong
to long air gaps. Since the 1970s, a large amount of studies about (i) Considering the influence of worker as well as the proximity
discharge characteristics of long air gaps including rod-plane, rod- effect of bundle conductor, gap characteristic function R of
rod and conductor-plane gaps has been conducted [11] and a series

IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278 1271
© The Institution of Engineering and Technology 2020
equipotential worker-tower leg gap is derived. Due to the difficulty coefficient related to high-voltage electrode type. For the rod
in obtaining function R of actual tower analytically, the tower electrode, Uc∞ = 1556 kV, A = 7.78 [13]. R is the gap characteristic
structure coefficient ST is introduced to present the effect of function related to the gap configuration, gap distance, and
complex towers and a numerical method for calculation ST is grounded electrode geometry. Q0 is the equivalent stream space
proposed based on finite element method (FEM) simulation. The charge located in the vicinity of the high-voltage electrode. Uin is
proposed model can be used to quantitatively analyse the influence the induced voltage generated by Q0 and its induced charge at the
of the worker and tower structure on discharge characteristic. streamer-stem tip; Uia is the induced potential at the streamer-stem
(ii) The proposed model can be applied to calculate the switching tip directly caused by Q0; Uib is the absolute induced potential at
impulse discharge voltages of the live working gap, under streamer-stem tip by the image charges induced on the opposite
simulated working conditions of a transmission line. It provides a electrode and the nearby objects due to Q0 and the charges Q0
possibility to obtain the MAD by numerical calculations instead of
costly and time-consuming tests, which can provide references for induces on the high-voltage electrode.
live working on HV transmission lines. Therefore, the acquisition of gap characteristic function R is the
main procedure to obtain the continuous leader inception voltage.
The rest of this paper is organised as follows. Section 2 introduces
the development of the model for discharge voltage prediction of 2.2 Equivalent of live line worker
live working gaps. Section 3 presents the live working test results For bare-hand method on HV transmission lines, the live line
and the comparisons between calculation of the test data. Section 4 worker is required to wear a full-set conductive clothing, remains
analyses the influence of worker's posture and tower structures on at the same potential with the energised conductors and works
the discharge voltage using the proposed model and gives some directly on the conductors. Our previous experimental results [4, 6–
safety recommendations. Section 5 presents conclusions and future 8, 10] indicate that when the worker is on the conductor or ready to
works of this study. get in or get out of equipotential, discharges usually occur between
the worker's head and tower, the worker's feet and tower, or the
2 Discharge model for equipotential worker – worker's back and tower. The discharge voltage of the gap between
tower gap worker's head (or feet) and tower are lower than that between
worker's back and tower due to the smaller radius of head and feet.
2.1 Discharge voltage calculation method Considering the most serious scenario, the minimum discharge
Continuous leader inception and propagation is the most important voltage of the live working gap should be obtained in order to
stage in long air gap discharge process. When the radius of maximise the safety margin of live working. Therefore, this paper
curvature for the high-voltage electrode is less than CCR, the focuses on the gap between worker's head or feet and tower as
continuous leader inception voltage almost remains the same and shown in Fig. 1.
the minimum breakdown voltage UB of the gap consists of the Previous research has found that when the gap length is within
continuous leader inception voltage Ulc and leader voltage drop 3 to 8 m, the CCR of the rod electrode is within 0.17 to 0.3 m [11].
The equivalent radius of the worker's head or feet is smaller than
ΔUl [12]:
0.15 m, which is less than the CCR of rod electrode. A three-
dimensional model corresponding to Fig. 1 is established and the
UB = Ulc + ΔUl (1) electric field distribution is calculated by the FEM. Fig. 2 displays
the simulated electric field strength along the different discharge
Since the leader can be regarded as an arc-like channel, the leader paths in Fig. 1 when the peak value of switching impulse voltage is
voltage drop can be expressed as [13]: applied. Since the electric field in the vicinity of the high stressed
electrode is the most intense and reduces rapidly as the distance
ΔUl = 50 d −
Ulc
+ 37.5ln 8 − 7e−1.33 d − Ulc / Es
(2) increasing, Fig. 2 only shows the electric field strength along the
Es discharge path within a range of 1 m. From Fig. 2, it can be
observed that there is slight difference of the electric field
where Es is the average electric field strength of the streamer, Es =  distribution between the worker's head and rod electrode in path 1
400 kV/m. because when the worker stands on the conductor the shape of the
Then the 50% breakdown voltage U50 of the air gaps under worker is more like a rod electrode. In path 2, the maximum
switching impulse can be calculated using the minimum electric field of rod is a little higher than worker's head or feet
breakdown voltage UB and standard deviation σ: when distance is <0.2 m due to the shielding effect of other body
parts of the worker. Therefore, the discharge parts such as head and
UB feet of the worker can be equivalent to a rod electrode, and the
U50 = (3) discharge voltage can be predicted based on the leader inception
1 − 3σ
mechanism.
Therefore, the key point is to calculate the continuous leader
inception voltage.
The general expression of the continuous leader inception
voltage of a long air gap is as follows [14]:

Uc∞
Ulc = (4)
1 + A/R

Q0
R= (5)
4πε0Uib

Uib = Uin − Uia (6)

Q0
Uia = (7)
4πε0s

where Ulc is the continuous leader inception voltage; Uc∞ is the Fig. 1 Diagram of discharge path between the equipotential worker and
tower
saturation value of continuous leader inception voltage; A is a

1272 IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278
© The Institution of Engineering and Technology 2020
 So the effect of the ground plane in Fig. 3a can be replaced by
an image charge −Q0 and the image of the conductor. In Fig. 3b, r0 
= s + h + a, r = 2d + a + h, thus the induced voltage Uib0 at point P
can be expressed as follows: (see (9)) . Since
s + h + a/2d + a + h <<1 and a/2d + a + h <<1, there is a
further approximation as follows:
∞
2 K0 s + h + a/2d + a + h n
π ∫ K0 a/2d + a + h n
K0(n)dn
0
(10)
K s + h + a/2d + a + h
≅ 0
K0 a/2d + a + h

For Bessel function K0(x), when x is small, it will be:

1
K0(x) = ln (11)
x

Thus the induced voltage Uib0 can be further approximated as: (see
(12)) . Now taking the effect of tower structure into consideration
as shown in Fig. 4. Experimental study shows that the gap factor of
conductor-tower leg gap is almost the same with the conductor-
cylinder gap [10, 18], therefore in this paper, the cylinder gap is
chosen as a basic structure to obtain its function R. The scheme for
determination R for a conductor + worker-cylinder gap is shown in
Fig. 4.
In Fig. 4, the resultant effect of Q0 and the charge it induces on
the conductor and rod can be expressed as follows: (see (13)) .
Therefore, allowing for the influence of the tower leg, Uib is the
absolute value of the induced voltage at the point P by the induced
charge on the cylinder due to the image charge −Qe: (see (14)) ,
where
∞
2 K02(n)
φ
b
b+d
=
π ∫ 0
K0 b/b + d n
dn (15)

For EHV and UHV transmission lines, the bundle conductor has
been adopted to reduce the corona loss. The bundle conductor can
be replaced by an equivalent conductor with an equivalent radius
ae, and ae is expressed as follows:

ae = (nalrbn − 1) 1/ n (16)

where n is the number of sub conductors; al is the radius of the

sub-conductor; rb is the bundle radius.
The numerical evaluation shows distance s has little influence
on the value of Uib, therefore the induced voltage of a bundle
conductor + rod-tower leg gap becomes as follows:
Fig. 2 Electric field of the worker-tower gap
(a) Worker's head-crossarm (path 1), (b) Worker's head-tower body (path 2), (c) Q0φ b/b + d
Worker's feet-tower body (path 2) Uib− leg =
4πε0(b + d)
2.3 Continuous leader inception voltage of live working gaps 2(d + b) ln ae + h/ae
× 1− 1−
2(d + b) + ae + h ln 2(d + b) + ae + h/ae
When the worker's head or feet become high-voltage electrode, (17)
these parts can be equivalent as a rod electrode. Considering a
conductor + rod-plane gap as shown in Fig. 3, the equivalent 1 φ b/b + d 2(d + b)
stream space charge Q0 is at a distance s from the surface of the rod = 1−
Rleg (b + d) 2(d + b) + ae + h
electrode. The point P at the streamer-stem tip is at a distance s0 ln ae + h/ae (18)
from the rod. Both s and s0 are much smaller than the gap distance 1−
ln 2(d + b) + ae + h/ae
d. The radius of the conductor is a, and the length of rod electrode
is h. The actual tower structures in transmission lines are more
For a long cylinder with diameter of 2a and a point charge Q complicated than a single tower leg, so it is difficult to calculate the
which is situated at a distance f from the axis of the cylinder. The gap characteristic function R with irregular tower structure
potential U(x) in points at any distance x from the axis of the mathematically. In this paper, we define the tower structure
cylinder is as follows [14]: coefficient ST to present the effect of different tower structures on
∞ the continuous leader inception voltage:
K0 f /x n
U(x) = −
Q
2π 2ε0 x ∫
0
K0 a/x n 0
K (n)dn (8)
Uib− tower
ST = (19)
Uib− leg
where K0(x) is the modified Bessel function in zero order.

IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278 1273
© The Institution of Engineering and Technology 2020
where Uib-tower is the induced voltage of any tower structure; Uib-
leg is the induced voltage of a single tower leg structure.
The value of ST can be calculated by the FEM [19, 20]. In this
paper, we assume that the point P in Fig. 3 is very close to the
surface of the high-voltage electrode. And we use a point charge Q
to simulate the effect of streamer space charge. By setting a point
charge Q along the discharge path at a certain distance s from the
high-voltage electrode, the overall induced voltage Uin can be
obtained by calculating the induced voltage on the surface of high-
voltage electrode. Then Uib can be obtained from (6) and (7). The
influence of the values of Q and s on ST has been studied by
establish a solid model of conductor-tower gap of 5 m spacing [20].
Firstly, s is kept as 0.5 m, and Q are set as 1 × 10−12, 1 × 10−6, 1 × 
10−3 and 1 C respectively to calculate Uib. Then Q is then kept as
1 C, and Uib is obtained when s ranges from 0.1 to 0.8 m. Potential
calculation results show that ST will not be affected by Q, and only
slightly affected by s, thus the value of Q is selected as 1 C and s is
determined as 0.5 m. The potentials of tower structures and outer
boundary are set to zero.
Therefore, considering the effect of the bundle-conductor and
different tower structures, when the worker's head or feet become
the high-voltage electrode, the general continuous leader inception
voltage of the live working gap can be expressed as follows: Fig. 3 Scheme for determining R of a conductor + rod-plane gap
(see (20))
(see (21))
Then the 50% breakdown voltage of live working gaps can be
calculated using (1)–(3).

3 Experimental validation
In this paper, the switching impulse discharge tests of the 750 kV
transmission line live working gaps are carried out in the UHV
outdoor test site of China Electric Power Research Institute. The
six-bundle conductors with 23 mm sub-conductor diameter and
400 mm bundle spacing are used in the test and the length of the
bundle conductor is 20 m. The simulated tower is made of high-
strength angle steel at a 1:1 ratio with the 750 kV side phase tower
structure, cup tower window and V type tower window. In live
working, the worker is required to wear full-set conductive
clothing. Therefore, in the test, the sim man made by aluminium-
alloy is used to simulate the worker and the sim man is produced
according to the real size and shape of the worker. During the test,
the air pressure is 100.0–101.1 kpa, the temperature is 18.5–
20.5°C, and the humidity is 9.7–11.3 g/m3. All the atmospheric
parameters are near to normal conditions. Fig. 4 Scheme for determining R of a conductor + rod-cylinder gap
During the discharge test, different gap configurations are
formed by adjusting the locations and the posture of the sim man. structure, worker's feet-cup tower window, worker's head-cup
There are five live working gaps including worker's feet-side phase tower crossarm, worker's feet-V tower window, worker's head-V

∞
2 K0 s + h + a/2d + a + h n
Uib0 =
Q0
−
Q0
4πε0 ⋅ 2d 4πε0(2d + a + h) π ∫ 0
K0 a/2d + a + h n
K0(n)dn (9)

Q0 Q0 ln 2d + a + h/s + h + a
Uib0 = − ⋅
4πε0 ⋅ 2d 4πε0(2d + a + h) ln 2d + a + h/a
(12)
Q0 1 1 ln s + h + a/a
= − 1−
4πε0 2d 2d + a + h ln 2d + a + h/a

Qe = Q0 × 2(d + b)
1 1 ln s + a + h/a (13)
× − 1−
2(d + b) 2(d + b) + a + h ln 2(d + b) + a + h/a

∞
K02(n)
Uib = 2
Qe
2π ε0(b + d) ∫ K0 b/b + d n
dn
(14)
0

Q 1 − 2(d + b)/2(d + b) + a + h 1 − ln s + a + h/a /ln 2(d + b) + a + h/a b

= 0 ⋅φ
4πε0(b + d) b+d
1274 IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278
© The Institution of Engineering and Technology 2020
 n
∑i = 1 Ui
U50 = (22)
n

n
∑i = 1 (Ui − U50)2 (23)
σ= × 100%
n−1

where Ui is the ith valid applied voltage, kV; n is the total number
of valid tests.
With the same high-voltage electrode, the induced voltage Uib
is dependent on the gap spacing and the grounded electrode
geometry. In order to obtain the value of ST in various scenarios,
the three-dimensional models of actual tower and the single tower
leg are established in SolidWorks according to the test arrangement
mentioned above. Fig. 6 shows the solid models of live working
gap between the worker and cup tower crossarm and the model of
gap between worker and single tower leg. Then the models are
imported to FEM-based software COMSOL. The relative
permittivity of bundle conductor, worker, tower structure and air
are set as 1. The equivalent streamer space charge Q of 1 C is set
along the discharge path at a distance s of 0.5 m from the
equipotential worker. The potentials of the tower structures and
outer boundary are set to zero. All the components are meshed with
tetrahedron. A workstation equipped with 16 × 3.07 GHz Intel
Xeon processors and 64 GB RAM is used to ensure smooth
simulation calculations. It takes about 20 min for the FEM
simulation of one model for different live working gaps (Fig. 7).
Consequently, the overall induced voltage Uin on conductor and
equipotential worker generated by Q and its induced charge can be
obtained, and Uib can be calculated via expressions (6) and (7).
Fig. 5 Switch impulse discharge test arrangement (worker's feet-tower) Table 1 presents the calculated value of ST and U50, and the
(a) Worker's feet-side phase, (b) Worker's feet-cup tower window, (c) Worker's feet-V tested results of U50 at different gap distance. The errors between
tower window calculation and test data are also displayed in Table 1.
It can be seen from Table 1 that the errors between the
calculated and tested discharge voltage are within − 2.81 and
0.37% for side phase tower structure, and within − 4.03 to 2.97%
for cup tower, and within − 3.43 to 4.02% for V tower under
different working conditions. The mean absolute percentage errors
between the calculated and test results for side phase tower
structure, cup tower and V tower are 1.25, 2.41 and 1.55%,
respectively. For all conditions, the maximum error is less than
±4.03%, which proves that the model developed in this paper has
good feasibility in the prediction of the 50% discharge voltage of
live working gaps.
The 50% discharge voltage of an air gap different from rod-
Fig. 6 Switch impulse discharge test arrangement (worker's head- plane gap can be predicted using the following expression:
crossarm)
(a) Worker's head-cup tower crossarm, (b) Worker's head-V tower crossarm U50 = K 500 d 0.6 (24)

tower crossarm. Fig. 5 presents the test arrangement of the worker's where K is the gap factor; d is gap distance, m.
feet-tower gaps and Fig. 6 shows the test arrangement of worker's Using the above expression to fit the experiment results from
head-crossarm gaps. Table 1, the gap factor of worker's feet- side phase gap, worker's
The standard switching impulse of 250/2500 μs is applied and feet- cup tower window gap, worker's feet- V tower window gap,
is generated by a 5400 kV impulse generator. The ‘up and down’ worker's head-cup tower crossarm gap and worker’ head-V tower
method, is adopted in the tests. During the test, the sim man is crossarm gap are 1.328, 1.265, 1.266, 1.134 and 1.142,
connected with the conductor to remain the same potential with the respectively. This means under the same gap distance, the
conductor. The tests are conducted with the gap distance ranging minimum 50% discharge voltage appears in the worker's head-cup
from 3 to 6 m, and 40 discharge tests are conducted at each tower crossarm gap as shown in Fig. 6a and the maximum 50%
distance. The 50% discharge voltage U50 and the standard discharge voltage occurs in the worker's feet-side phase gap as
deviation σ can be calculated by the following equation: shown in Fig. 5a. For example, when gap distance is 4 m, the 50%
discharge voltage of worker's feet-side phase gap is 17.1% higher

1556 1556
Ulc = =
1 + 7.78/Rtower 1 + Uib− tower /Uib− leg 7.78/Rleg
(20)
1556
=
1 + 7.78STφ b/b + d /b + d 1 − 2(d + b)/2(d + b) + ae + h 1 − ln ae + h/ae /ln 2(d + b) + ae + h/ae

b+d
Rtower =
STφ b/b + d 1 − 2(d + b)/2(d + b) + ae + h 1 − ln ae + h/ae /ln 2(d + b) + ae + h/ae (21)

IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278 1275
© The Institution of Engineering and Technology 2020
Fig. 7 Three-dimensional models of live working gap
(a) Worker's head-cup tower crossarm, (b) Worker's head-single tower leg

Table 1 Discharge voltage comparison of live working gaps

Gap structure Gap distance, m ST Calculated U50, kV Tested U50, kV Tested σ, % Error, %
Worker’ feet-side phase 3.13 1.06 1341 1336 3.8 0.37
3.5 1.09 1416 1415 4.0 0.07
4.0 1.1 1499 1527 5.7 −1.83
4.5 1.17 1557 1602 6.1 −2.81
Worker's feet-cup tower window 3.5 1.41 1352 1352 4.7 0.00
4.2 1.48 1459 1496 3.2 −2.47
4.5 1.38 1520 1575 5.0 −3.49
5 1.49 1573 1636 5.6 −3.85
Worker's feet-V tower window 3.5 1.31 1372 1360 5.1 0.88
4.2 1.26 1498 1497 4.2 0.07
4.5 1.21 1549 1562 5.4 −0.83
5 1.29 1599 1640 3.7 −2.50
Worker's head-cup tower crossarm 4.4 1.47 1388 1348 5.5 2.97
4.8 1.46 1449 1430 3.6 1.33
5.2 1.34 1526 1543 2.4 −1.10
5.6 1.36 1572 1638 4.7 −4.03
Worker’ head-V tower crossarm 4.4 1.30 1424 1369 4.3 4.02
4.8 1.40 1461 1458 3.6 0.21
5.2 1.32 1530 1537 5.2 −0.46
5.6 1.33 1577 1633 5.4 −3.43

than the 50% discharge voltage of worker's head-cup tower

crossarm gap. This is due to overall effect of the worker's posture
and the different tower structures. With the same tower structures,
the discharge voltage of the worker's feet-tower window gap is
11.6 and 10.9% higher than worker's head-tower crossarm gap in
cup tower and V tower, respectively. With the same worker's
posture, the discharge voltage of the worker's feet-side phase gap is
5.0% higher than worker's feet-cup tower window gap.

4 Discharge characteristics analysis

4.1 Influence of live line worker
The gap characteristic function R represents the effect of the gap
configuration on the continuous leader inception voltage. The
comparison of R and the continuous leader inception voltage Ulc
Fig. 8 Comparison of R under different gap configurations
under different gap configurations are shown in Figs. 8 and 9.
From Figs. 8 and 9, it can be seen that with the same tower
structure (like cup tower window), the difference in discharge discharge voltage. The shielding effect of the bundle conductor for
voltage is mainly affected by the posture of the worker. When the the worker will be reduced with the increasing value of h, which
discharge appears on the worker's feet, the length of the worker's will raise the possibility for air gap breakdown.
body part outside the conductor is 0.5 m (h = 0.5 m). When the Standing or riding on the conductor are the common posture of
discharge occurs on the worker's head, the length of the worker's the equipotential worker. Taking the cup tower as an example,
body part outside the conductor is 1 m (h = 1 m). Comparing the R when the length h various between 0.5 m (riding on conductor) to
values of cup tower and V tower, it can be seen that with the same 1.8 m (standing on conductor), the 50% discharge voltage of
gap length, the longer length the body exceeds the conductor worker's head- tower crossarm gap is calculated using the proposed
(which means larger h), the smaller the corresponding R value, and model when gap distance ranges from 4 to 7 m. The 50% discharge
the lower the corresponding leader inception voltage and the 50% voltage under different conditions is shown in Fig. 10. Also, the

1276 IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278
© The Institution of Engineering and Technology 2020
Fig. 9 Comparison of Ulc under different gap configurations

Fig. 10 Influence of worker's posture on 50% discharge voltage

test results of worker's feet-cup tower window from Table 1 is The difference is mainly caused by the tower areas and structures
plotted in Fig. 10 for comparison. It can be seen from Fig. 10 that surrounding the high-voltage electrode. The area of cup tower
the discharge voltage decreases with the increase of h. The window around the conductor and worker is larger than V tower
discharge voltage at h = 1.8 m is reduced by 8–12% compared with window and the side phase structure, resulting in the smallest R of
the discharge voltage at h = 0.5 m. With the increase in gap the cup tower window, therefore the magnitude of the induced
distance, the influence of the worker on the discharge voltage charge on cup tower window is the largest, which makes it easier
reduces. Fig. 10 also indicates that with the same length of h, the for the discharge development. So it can be concluded that the
discharge voltage between worker-tower crossarm gap is higher influence of the tower structure on the live working gap discharge
than worker-tower window gap. If the worker enters equipotential characteristics is mainly caused by the area of the grounded tower
from upper crossarm vertically (h = 1.8 m), the discharge voltage is structures around the high-voltage electrode. The larger area of the
reduced by about 14% compared with the worker entering tower surrounding the highly stressed electrode, the more induced
equipotential from the tower window. This is because the worker is charge is generated by the stream space charge, resulting in the
similar to a protruding rod electrode when standing or riding on lower discharge voltage of the gap.
conductor, which significantly distorts the electric field around the
worker, resulting in lower discharge voltage. If the worker enters 5 Conclusion
equipotential from the tower window (usually siting in a basket, h 
= 0.5 m), due to the shielding effect of the conductor, the discharge A discharge voltage calculation model of live working gaps is
voltage will be higher. presented in this paper, through the combination of analytical
The insulation capability of the gap will be significantly derivation and FEM calculation. In this model, both of the
reduced with the increasing length of the worker's body outside the influence of equipotential worker and various tower shapes are
bundle conductors. Therefore, in live working practice, the taken into consideration. In addition, the live working gap
equipotential worker should watch his posture and keep himself discharge test under switching impulse is carried out on 750 kV
close to the bundle conductor as much as possible. Based on the transmission lines. The correctness of the proposed model is
above analysis, it can be concluded that it is safer to enter the verified by the test data. The effects of the worker and tower
equipotential from the tower body side via basket than to enter structures on the discharges characteristics are also analysed and
from the crossarm above the conductor. safety recommendations are put forward. Based on the predicted
results and analysis, the conclusions of this study can be drawn as
4.2 Influence of tower structure follows:

Figs. 8 and 9 also indicate that the values of R and Ulc increase (i) The calculated 50% discharge voltage shows good agreement
with the gap distance increasing and the Ulc tends to be saturated with the test results. Under all circumstances, the errors between
with the rising gap spacing. With the same gap distance and the the calculated data and tested data are within 4.03%, which are
same worker's posture, for example, the worker's feet becomes the acceptable for engineering applications.
discharge electrode as shown in Figs. 5a–c, the discharge voltage (ii) The longer length the worker's body exceeds the conductor, the
of side phase tower structure is higher than cup tower and V tower. weaker the shielding effect of the conductor is, and the lower the

IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278 1277
© The Institution of Engineering and Technology 2020
gap discharge voltage is. Entering the equipotential from the tower [6] Xiao, B., Wu, T., Liu, K., et al.: ‘Experimental investigation on the minimum
approach distance on 1000 kV UHV compact transmission line’. 2016 IEEE
body side instead of from the crossarm is recommended. int. Conf. on High Voltage Engineering and Application, Chengdu, People's
(iii) The discharge voltage of worker-side phase gap is 5% higher Republic of China, September 2016, pp. 1–4
than worker-tower window gap, which means the larger area of the [7] Hu, Y., Liu, K., Wang, L.N., et al.: ‘Experimental research of live working on
tower surrounds the equipotential worker and conductor, the lower 1000 kV double circuit AC transmission line on the same tower’, High Volt.
Eng., 2010, 36, (11), pp. 2668–2673
the discharge voltage of the gap is. [8] Wang, L.N., Zhang, Q.S., Hu, J.X., et al.: ‘Research on complex gap
discharge model of live working on EHV and UHV high-voltage transmission
The proposed model preliminarily achieves discharge voltage lines’, Can. J. Electr. Comput. Eng., 2014, 37, (1), pp. 11–18
prediction of live working gaps, which contributes to reducing the [9] Liao, C.B., Ruan, J.J., Liu, C., et al.: ‘Helicopter live-line work on 1000-kV
UHV transmission lines’, IEEE Trans. Power Deliv., 2016, 31, pp. 982–989
full-scale test work and provides an effective means to determine [10] Zhang, Q.S.: ‘Research on complex gap discharge model of live working on
the MAD for live working on EHV and UHV transmission lines. It EHV and UHV high voltage transmission lines’, PhD Dissertation, Wuhan
should be pointed out that the model in this paper is applicable in University, Wuhan, People's Republic of China, 2018
normal atmospheric conditions. However, it should be pointed out [11] Les Renardieres Group: ‘Research on long air gap discharges at les
renardieres-1973 results’, Electra, 1974, (35), pp. 49–156
that the model in this paper is applicable in normal atmospheric [12] Carrara, G., Thione, L.: ‘Switching surge strength of large long air gaps: a
conditions, however during the implementation of live working, physical approach’, IEEE Trans. Power Appar. Syst., 1976, 95, (2), pp. 512–
variations in air density and humidity may lead to the diversity of 524
discharge voltage. Modelling accounting for the effects of [13] Rizk, F.A.M.: ‘A model for switching impulse leader inception and
breakdown of long air-gaps’, IEEE Trans. Power Deliv., 1989, 4, (1), pp.
atmospheric parameters still needs more work. 596–606
[14] Rizk, F.A.M.: ‘Switching impulse strength of air insulation: leader inception
criterion’, IEEE Trans. Power Deliv., 1989, 4, (4), pp. 2187–2195
6 References [15] Gallimberti, I., Bacchiega, G., Bondiou, C.A., et al.: ‘Fundamental processes
[1] Hu, Y., Liu, K., Wu, T., et al.: ‘Key technology research and application of in long air gap discharges’, C. R. Phys., 2002, 3, (10), pp. 1335–1359
live working technology on EHV/UHV transmission lines in China’. 2014 Int. [16] Diaz, O., Cooray, V., Arevalo, L.: ‘Numerical modeling of electrical
Conf. on Power System Technology, Chengdu, People's Republic of China, discharges in long air gaps tested with positive switching impulses’, IEEE
October 2014, pp. 2299–2309 Trans. Plasma Sci., 2018, 46, (3), pp. 611–621
[2] Ghassemi, M., Farzaneh, M.: ‘Effects of tower, phase conductors and shield [17] Fofana, I., Beroual, A., Rakotonandrasana, J.H.: ‘Application of dynamic
wires on the electrical field around a tower window during live-line work’, models to predict switching impulse withstand voltages of long air gaps’,
IEEE Trans. Dielectr. Electr. Insul., 2015, 22, (6), pp. 3413–3420 IEEE Trans. Dielectr. Electr. Insul., 2013, 20, (1), pp. 89–97
[3] IEC standard 61472:2013: ‘Live working – Minimum approach distances for [18] Paris, L.: ‘Influence of air gap characteristics on line-to-ground switching
a.c. systems in the voltage range 72.5 kV to 800 kV – a method of surge strength’, IEEE Trans. Power Appl. Syst., 1967, 86, (8), pp. 936–947
calculation’, 2013 [19] Gao, J.C., Wang, L.N., Zhang, Q.S., et al.: ‘Switching impulse discharge
[4] Hu, Y., Wang, L.N., Liu, K., et al.: ‘Experimental investigation on live characteristics of UHV transmission line air gaps’, IEEJ Trans. Electr.
working requirements for 1000 kV ultra high voltage AC transmission line’. Electron. Eng., 2019, 14, (5), pp. 735–741
2010 Asia-Pacific Power and Energy Engineering Conf., Chengdu, People's [20] Fang, Y.Q., Wang, L.N., Li, R., et al.: ‘A modified model for discharge
Republic of China, March 2010 voltage of AC transmission line-tower air gaps’, IEEE. Access., 2019, 7, pp.
[5] Lee, H.K., Lee, H.K., Koo, H.G.: ‘Application of the stepping pendulum 71472–71480
method to long tower arms in 765 kV transmission line for live working’.
11th Int. Conf. on Live Maintenance, Budapest, Hungary, May 2014, pp. 1–4

1278 IET Gener. Transm. Distrib., 2020, Vol. 14 Iss. 7, pp. 1271-1278
© The Institution of Engineering and Technology 2020