Volume 56, Number 1-2, 2015 9

Investigation into Tower Model Effect

on Fast-Front Overvoltages in
Transmission Lines
Levente Czumbil, Dan D. Micu, Denisa Şteţ, Andrei Ceclan
Dept. of Electrotechnics and Measurements, Technical University of Cluj-Napoca, Romania

Abstract: The proper modelling of transmission line towers and conductors plays an essential part in the travelling-wave analysis
of fast front overvoltages due to lightning surges in overhead power lines. This paper investigates the effect of different simulation
component models like power line towers, lightning current or tower footing resistance on the transient overvoltages produced by
back flashover phenomena. A new combined tower model that takes into consideration both the bracings and the damping effect of
each tower section is proposed.

keywords: fast front overvoltages, tower models, lightning current models, transmission lines, back flashover.

1. INTRODUCTION 2. TRANSMISSION LINE TOWER

MODELLING
One of the principal causes of Fast Front
Overvoltages (FFO) in power systems is lightning One of the most important steps in the evaluation
stroke to transmission lines. Lightning overvoltages of fast front overvoltages produced by lightning strikes
have a wave head of several microseconds and are an to overhead power lines, is the accurate modelling of
important factor in determining the insulation design of transmission line towers. Although the lightning
substation equipment [1, 2]. response of a transmission line tower is an
Lightning overvoltages could be produced by electromagnetic phenomenon, the representation of a
direct lightning strokes to phase wires or by back tower is usually made in circuit terms.
flashover, when the lightning stroke hits the tower or
the ground wire and the voltage across the phase wire
insulator string exceeds its withstanding capability.
A direct strike to a power line is very rare, and
most of the lightning strikes the top of the transmission
tower. The lightning back flashover effects are
recognized as one of the major causes of transmission
line (TLine) outages and may also result in substation
outages, caused by incoming surges with amplitude
exceeding the insulation level of the substation
equipment [2].
It is very difficult to observe the lightning
overvoltage experimentally, thus a numerical
Fig. 1. a) Single lossless transmission line model; b) Multi-Story
simulation is adopted in order to investigate it [3]. This transmission line model;
paper studies the effect of different simulation
component implementation (e.g. tower model, footing 2.1. Single lossless transmission line models
resistance, lightning current model and waveform, etc.)
on the transient overvoltages produced by back The simplest representation of power line tower is
flashover phenomena in case of overhead power lines. a lossless distributed-parameter transmission line,
characterized by its surge impedance and travel time.
Table I shows the different tower surge impedance
formulas presented in the literature, based on an
electromagnetic theory approach or on experimental
measurements. Due the multiple paths through

© 2015 – Mediamira Science Publisher. All rights reserved

10 ACTA ELECTROTEHNICA

crossarms and the lattice structure of the tower, the The evaluation error produced by each of the
propagation velocity of the lightning current and above presented tower surge impedance formula are
voltage waves across the tower is reduced to 290 ∙ 106 presented in Table III. It can be observed that the most
m/s. accurate results are provided by Ametani’s formula.
TABLE I
However, due the fact that the tower response to a
DIFFERENT TOWER SURGE IMPEDANCE FORMULAS lightning strike is directly dependent on its surge
Author Proposed Formula
impedance value one could also use the corrected
CIGRE formula in order to obtain a higher margin on
Wagner  H 
Z S = 60 ⋅ ln 2 2 ⋅ 

the evaluated overvoltage values at GIS entrance.
[4]
 RB 
TABLE III.
  OBTAINED EVALUATION ERRORS
 RB 2 + H 2 
Sargent [4] Z S = 60 ⋅ ln 2 ⋅ 
 RB 
 
 H 
Jordan [4] Z S = 60 ⋅ ln  − 60

 RB 
 1  R Avg   
CIGRE [2]
Z S = 60 ⋅ ln cot ⋅ arctan 
IEEE [5]  2  H 
   
Corrected    
π   Ravg   
CIGRE [4]
ZT = ⋅ 60 ⋅  ln cot ⋅ arctan
1   − ln 2 
Baba & 4    2   2.2. Ishii Multi-Story Tower Model
   H  
Ishii
For a more realistic representation of transmission
line towers, Ishii proposed a multi-story tower model
TABLE I (CONTINUES) where, in order to take into account the attenuation of
DIFFERENT TOWER SURGE IMPEDANCE FORMULAS the traveling wave long tower height, each tower
Author Proposed Formula section is represented by a lossless TLine in series with
  
2  a parallel R–L circuit (see Fig. 2).
  R Eq + H + H 
2 2

   
ln + 
 
Z S = 60 ⋅  R Eq ⋅  R Eq + 4 H + 2 H  
2 2
Ametani [6]    
 
 3R Eq + R Eq 2 + 4 H 2 − 4 R Eq 2 + H 2 
+ 
 2H 
 2 2 ⋅H 
Hara [7] Z S = 60 ⋅  ln − 2
 REq 
 

where: H is tower height; RB, RAvg and REq are

tower base, average and equivalent radius by
considering the tower as system of multiple vertical Fig. 2. Ishii Multi-Story Tower Model [4].
parallel conductors, respectively.
Due the fact that most of the tower surge The values of the parameters, and the model itself,
impedance equations are empirical ones or have been have been revised in more recent years by Ametani [9].
developed by imposing a set of simplifying conditions, Thus, the damping resistances and inductances could be
in order to determine the most accurate evaluation evaluated based on tower section surge impedance,
formula the authors have developed MathCad algorithm according to the following equations:
which implement and test the above presented
equations for different tower geometries. Obtained
1 hi
results have been compared to measured surge Ri = 2 ⋅ Z T 1 ⋅ ln  ⋅ (i = 1, 2, 3) (1)
impedance values presented in literature [6, 8]. Table II  γ  l1 + l 2 + l 3
contains details regarding the investigate tower
1 H
geometries: R 4 = 2 ⋅ Z T 2 ⋅ ln  , Li = 2 ⋅ Ri ⋅ τ , τ = (2)
TABLE II. γ  υT
INVESTIGATED TOWER GEOMETRIES [6, 8]
where: ZT1 and ZT2 are the surge impedance of the
three upper tower sections, and of the tower foot
section respectively, hi is the height of each tower
section, γ is an attenuation coefficient (γ = 0.8944) and
τ is the wave travel time across the tower.
 Volume 56, Number 1-2, 2015 11

2.3. Hara Multi-Story Tower Model In order to eliminate unwanted wave reflections
the authors propose to use the same surge impedance
Measurements on scaled and real size
value considering one single section for the upper part
transmission line towers showed that the surge
of the tower. To obtain accurately the surge impedance
impedance of conductors is reduced about 10% by
of each tower section the Ametani formula, should to
adding the bracings to the main legs. To take into
be used.
account this reduction in the real value of tower surge
impedance, Hara proposed [7] a multi-story tower 2.5. Tower Footing Impedance
model (see Fig. 3) where introduced for each tower
section an additional parallel surge impedance: In order to take into account soil ionization
phenomena, tower footing could be represented by a
current dependent variable resistance driven by
Z Lk = 9 ⋅ Z Tk , k = (1,2,3,4) (3) equation (5) proposed by CIGRE [2] and IEEE [5]:

where: ZT is the surge impedance of each tower R0 1 E0 ⋅ ρ

RT = ; IC = ⋅
section. I 2π R0 2 (5)
1+
IC

where: RT is the current dependent tower footing

resistance, R0 is the low current and low frequency
footing resistance, IC is the limiting current to initiate
soil ionization, I is the strike current, ρ is the soil
resistivity and E0 is the soil ionization gradient.

3. LIGHTNING CURRENT MODELLING

Fig. 3. Hara Multi-Story Tower Model. The lightning current produced by a lightning
stroke to an overhead transmission line could be model
Also, for each tower crossarm, Hara introduced an using a variable current source in parallel with a 400 Ω
equivalent transmission line component based on the resistance representing the air path between the clouds
surge impedance expression for a conventional and the transmission line. To describe the lightning
horizontal conductor: impulse current, usually manufactures use a Ramp, eq.
(6), or a Double Exponential, eq. (7), function which
2 ⋅ hk l are easier to implement in test equipment.
Z Ak = 60 ⋅ ln , rAk = Ak (4)
rAk 4
 I LC
 ⋅t t ≤ tf
 tf
where: hAk and rAk are respectively the height and I Ramp (t ) =  , (6)
)( )
I LC
the equivalent radius of the kth crossarm (¼ of the  ⋅ 2th − t f − t t > tf
crossarm length) 
 2 ⋅ t(h −tf

2.4. Proposed Multi-Story Tower Model (

I DExp (t ) = I LC ⋅ e − A⋅t − e − B ⋅t ) (7)

In order to take into account both the presence of

tower crossarms, the damping effect of each tower where: ILC is the amplitude of the lightning
section and the surge impedance values reduction current, tf is the time to front, th is the time to half,
produced by bracings, a new combined tower model while A and B are exponential coefficients for the
has been implemented by the authors based on the double exponential model.
previously presented two multi-story tower models, see In 1991, based on statistical analysis of recorder
Fig. 4. lightning strikes around the world, CIGRE [2] proposed
a more complex formulation of lightning current which
to take into account the concave form of the rise up
part:

 A1 ⋅ t + A2 ⋅ t n t ≤ tc

I CIGRE (t ) =   t − tc   t − tc  (8)
   
 B1 ⋅ exp − t  − B2 ⋅ exp − t  t > tc
  1   2 

Fig. 4. Proposed Multi-Story Tower Model.

12 ACTA ELECTROTEHNICA

where: tc is the time to crest, t1 and t2 are time TABLE V

CONDUCTORS POSITION ON TOWER
parameters and A1, A2, B1, B2 are coefficient
parameters which describe the front and the tail part of Conductor Position [m] Height [m]
Phase Wire A1 -3.2 34.4
lightning current waveform [2].
Phase Wire B1 -5.3 28.2
Recently, in literature, the Heilder lightning Phase Wire C1 -3.2 23.2
current function [10] is used for the study of fast front Phase Wire A2 3.2 23.2
overvoltages produced by lightning strikes to overhead Phase Wire B2 5.3 28.2
power lines: Phase Wire C2 3.2 34.4
Sky Wire 0 39.5
n The 110 kV power line has a total length of 50 km
 t  and an average 35 MVA power load (approximately
 
I LC  t f   t  160 A on each phase wire). The produced fast-front
I Heidler (t ) = ⋅ ⋅ exp −  (9) overvoltages due to back flashover are analysed
h  t 
n
 th 
1+   considering different tower models, different lightning
 tf  currents and different tower footing resistance values.
 

where: n is the order of the implemented Heidler 5. SIMULATION RESULTS

function and η is a scaling coefficient.
Due the fact that the insulator string may
withstand a high transient voltage for a short duration,
but it could fail to withstand a lower transient voltage
with a longer duration, the volt-time characteristic
proposed by CIGRE [2] has been implemented for the
back flashover simulations.

K2
VWIns = K 1 + (10)
t 0.75
Fig. 5. 10 kA, 8/20 µs lightning current implementation with different
mathematical functions. where: VWIns is the insulator string withstand
voltage in (kV), K 1 = 400 ⋅ L , K 2 = 710 ⋅ L , L is the
4. CASE STUDY
length of the insulator string in (1.5 m for the
In order to determine the effect of the different investigated power line), and t is the elapsed time in
simulation component models presented above on the (µs) from the lighting strike occurring.
produced fast-front overvoltages a 110 kV double
circuit power line connecting two distribution
substations is considered (see Fig. 5). The overhead
power line consists in one sky wire and six phase wires
placed in an anti-symmetrical phase arrangement on
ITSn110.256 type transmission line towers. Table V
presents the geometry of the power line tower, while in
Table VI the position of each transmission line
conductor on the tower is presented.
Fig. 7. Lightning current flowing through the tower.

Considering the worst case scenario, when a 200

kA lightning hits the sky wire next to a tower, the
produced FFO waveforms, due to back flashover on left
side upper phase wire (Phase A1), are evaluated along
the power line length are evaluated by modeling the
first five tower in both direction from the lightning
Fig. 6. Investigate 380 kV overhead power line.
discharge location.
From Fig. 6 it can be observed that most of the
TABLE IV lightning stroke current flows directly to the ground
POWER LINE TOWER STRUCTURE GEOMETRY
through footing resistance of the tower that has been
Section Width Conductor Radius Section Height
hit. However, a part of the lightning current flows
DT DM DB rT rM rB H1 H2
[m] [m] [m] [mm] [mm] [mm] [m] [m]
through the insulator string air gap to left side upper
0.4 1.28 4.78 30 50 85 23.2 16.3
phase as a result of the back flashover phenomena, and
another part flows through the sky wire to the nearest
towers and that way to the ground. (see Fig. 7).
 Volume 56, Number 1-2, 2015 13

considering a
200 kA, 8/20 µs lightning current and the back
flashover appears on the left side upper conductor (A1)
when the phase voltage riches its negative peak value.
It can be observed that, by neglecting the different
overvoltage levels produced at tower crossarms, the
single TLine model provides an overestimation of the
Fig. 8. Current flowing through insulator string air gap and nearby produced FFO. On the other hand, by taking into
towers. consideration the presence of tower crossarms and
bracings, the Hara multi-story model introduces a small
Due to electromagnetic coupling, the amount of delay when the FFO waveform reaches its peak values.
current flowing into the faulted phase wire and through The proposed tower model makes o good correlation
the sky wire, will not only produce overvoltages on the between the evaluated overvoltage levels and the
upper left side conductor but in all of them. The introduced oscillations by crossarm reflections.
produced FFO waveforms propagate over the
transmission line towards the distribution substation. 5.2. Effect of different lightning current
Fig. 8 presents the evaluated FFO waveforms on Phase implementations
A1 (faulted phase) and Phase A2 (symmetrically
To determine the effect of using different
opposite on power line towers) at different locations
mathematical function implementation of the lightning
along transmission line length considering no surge
stroke current, the produced FFO waveform at
arresters connected at substation entrance. In this case
flashover location has been analyzed for a 200 kA, 8/20
even if the lightning hits at the middle of the power line
µs lightning current waveform.
length the overvoltage value that reaches the
distribution substations is far over their basic insulation
level.

Fig. 11. Produced black flash FFO for different lightning current
functions.
Fig. 9. Produced FFO waveform along power line length.
From Fig. 10 it can be observed that for the
investigated power line geometry the Ramp and GIGRE
5.1. Influence of Different Tower Models lightning current function implementations provide
higher overvoltage values. However, due their
To identify the effect of using different tower
exponential rise up curve the Double Exponential and
models in the simulations of a back flashover on the
the Heidler function implementations provide an earlier
110 kV power line, the transmission line tower are
overvoltage peak value.
model using the equivalent circuit configuration
Fig. 12 presents the obtained FFO waveforms, at
proposed by the authors. Obtained FFO values and
fault location, in case of a 200 kA lightning current,
waveforms are compared with the situation when the
with different waveforms using a Heidler function
towers were modeled using the single and multi-story
implementation.
transmission line models presented in section II: the
simple multi-story model, the Ishii and respectively
Hara multi-story model.

Fig. 12 Produced black flash FFO at fault location for different

lightning current waveforms.

Fig. 10. Produced back flashover FFO at GIS entrance considering

5.3. Effect of different tower footing resistance
different tower models and no surge arresters connected.
values
Fig. 9 presents the evaluated FFO waveform at To show the effect of different tower footing
fault location for each implemented tower model impedance values, the produced FFO waveforms have
14 ACTA ELECTROTEHNICA

been evaluated considering for each transmission line oscillations by crossarm reflections, has been proposed
tower footing resistance a fixed 20 Ω, 10 Ω, and 5 Ω by the authors. Obtained FFO waveforms have been
value, respectively the CIGRE soil ionization model compared to other existing tower models.
with a 20 Ω low frequency footing resistance and a 300 Based on obtained results, the authors conclude
kV/m soil ionization gradient. Obtained simulation that besides tower modeling, the proper tower footing
results are presented in Fig. 11. It can be observed that resistance and possible soil ionization processes
the value of tower footing resistance plays an important representation consists in one of the most important
role in the produced maximum FFO values during steps regarding the evaluation of fast front overvoltages
black flashover phenomena. produced by lightning strikes to overhead power lines.
Also, it is recommended that in frequent thunder storm
areas to provide a low grounding resistance to
transmission line tower in order to protect the line
against back flashover phenomena.

ACKNOWLEDGEMENTS

This paper was supported by the Post‐Doctoral

Fig. 13. Produced back flashover FFO at GIS entrance for tower Programme POSDRU/159/1.5/S/137516, project co-
footing impedance values.
funded from European Social Fund, through the Human
Resources Sectorial Operational Program 2007‐2013.
Taking into consideration the proper soil
ionization process could prevent overestimations of the
REFERENCES
produced overvoltage values and an overdesign of
power line and substation protection equipment. 1. IEEE Fast Front Transients Task Force, “Modeling Guidelines
for Fast Front Transients”, IEEE Trans. on Power Delivery, vol.
5.4. Effect of surge arrester on FFO 11, no. 1,
pp. 493-506, January, 1996.
Finally the effect of surge arresters on the 2. CIGRE, Guide to Procedures for Estimating the Lightning
produced back flashover FFO is analyzed. Usually Performance of Transmission Lines, CIGRE Technical Report
063, October, 1991.
surge arresters are connected at substation entrance in 3. IEC 60071-4, “Insulation Coordination - Part 4: Computational
order to protect substation equipment. However, in Guide to Insulation Coordination and Modelling of Electrical
frequent thunder storms area could be placed also along Networks”, International Electrotechnical Commission, (IEC),
the transmission line in order to protect it Fig. 12 p. 122, June, 2004.
4. J.A. Martinez-Velasco, Power Systems Transients. Parameter
presents the FFO waveform at substation entrance with Deter-mination, Boca Raton: CRC Pres, 2010.
and without surge arresters connected and respectively 5. IEEE Working Group, “Estimating Lightning Performance of
at the third tower from fault location if additional surge Transmission Lines. II. Updates to Analytical Models”, IEEE
arresters would be placed there: Trans. on Power Delivery, vol. 8, no. 3, pp. 1254-1267, 1993.
6. A. Ametani, Y. Kasai, J. Sawada, A. Mochizuki, and T.
Yamada, “Frequency-Dependent Impedance of Vertical
Conductors and A Multiconductor Tower Model”, IEE
Proceedings - Generation, Transmission and Distribution, vol.
141, no. 4, pp. 339 -345, July, 1994.
7. T. Hara, and O. Yamamoto, “Modelling of a Transmission
Tower for Lightning Surge Analysis”, IEE Proceedings -
Generation, Transmission and Distribution, vol. 143, no. 3, pp.
283 - 289, May, 1996.
8. Y. Baba, and M. Ishii, “Numerical Electromagnetic Field
Analysis on Measuring Methods of Tower Surge Impedance”,
IEEE Trans. on Power Delivery, vol. 14, no. 2, pp. 630 - 635,
April, 1999.
Fig. 14. Produced black flash FFO at substation entrance without and 9. A. Ametani, and T. Kawamura, “A Method of a Lightning
with surge arresters connected to power line. Surge Analysis Recommended in Japan Using EMTP,” IEEE
Trans. on Power Delivery, vol. 20, no. 2, pp. 867 - 895, April,
2005.
10. A. Ceclan, V. Topa, D. D. Micu, and A. Andreotti, “Lightning-
6. CONCLUSIONS Inverse Reconstruction by Remote Sensing and Numerical-Field
Synthesis,” IEEE Trans. on Magnetics, vol. 49, no. 5, pp. 1657 -
The influence of different simulation component 1660, May, 2013.
models (like transmission line towers, footing
Levente Czumbil
impedance or lightning current) on fast front transient Department of Electrotechnics and Measurements
overvoltage evaluation, in case of lighting strokes to Faculty of Electrical Engineering, Technical University of
overhead power lines has been investigated. Cluj-Napoca
A new more realistic transmission line tower Cluj-Napoca, Romania
levente.czumbil@ethm.utcluj.ro
model that makes o good correlation between the
evaluated overvoltage levels and the produced