You are on page 1of 10

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, D13101, doi:10.

1029/2006JD007958, 2007

On return stroke currents and remote electromagnetic fields


associated with lightning strikes to tall structures:
1. Computational models
D. Pavanello,1 F. Rachidi,1 M. Rubinstein,2 J. L. Bermudez,3 W. Janischewskyj,4
V. Shostak,5 C. A. Nucci,6 A. M. Hussein,4,7 and J. S. Chang8
Received 24 August 2006; revised 8 February 2007; accepted 6 April 2007; published 3 July 2007.

[1] In this paper, analytical expressions relating far fields and currents associated with
lightning strikes to tall towers are derived. The derived equations are general and can be
used with any engineering model. It is shown that the far field can be decomposed
into three terms, namely (1) contribution of the main return stroke pulse along the
lightning channel, (2) contribution of the multiple-reflection process along the elevated
strike object, including the contribution of upward propagating pulses transmitted into the
channel, and (3) contribution of the so-called ‘‘turn-on’’ terms, associated with the current
discontinuity at the return stroke wavefront. This discontinuity is associated with
neglecting reflections at the upward moving front. It is also shown that only the first term
is model-dependent and that the far field–current relationship does not significantly
depend on the adopted engineering model. It is therefore possible to use the analytical
equation relating the current peak and the associated distant electric or magnetic fields,
derived by Bermudez et al. (2005) for the transmission line (TL) model, for any
engineering model extended to include a tall strike object. It is also shown that the peak
amplitude of the electromagnetic field radiated by a lightning strike to a tall structure is
relatively insensitive both to the value of the return stroke speed and the top reflection
coefficient. These results and findings of this paper emphasize the key differences between
return strokes initiated at ground level (or from short strike objects) and those striking tall
towers: (1) The electromagnetic field from lightning strikes to tall towers is largely
determined by the tower and only to a lesser extent by the channel, (2) electromagnetic
fields associated with tall strike objects are less model-sensitive than those corresponding
to a strike to ground; in particular, the early time response of the field is nearly
model-independent, and (3) unlike ground-initiated strikes, for which the far field peak is
strongly dependent on the return stroke speed (proportional according to the TL model),
far field peaks associated with strikes to tall strike objects are little sensitive to the
return stroke speed. This is a particularly interesting result when lightning currents are
measured directly on instrumented towers to calibrate the performance of lightning
location systems, since in most practical cases the value of the return stroke speed is
unknown. Note that points (2) and (3) can be considered as corollaries of point (1).
Citation: Pavanello, D., F. Rachidi, M. Rubinstein, J. L. Bermudez, W. Janischewskyj, V. Shostak, C. A. Nucci, A. M. Hussein, and
J. S. Chang (2007), On return stroke currents and remote electromagnetic fields associated with lightning strikes to tall structures:
1. Computational models, J. Geophys. Res., 112, D13101, doi:10.1029/2006JD007958.

1. Introduction erably facilitates the collection of data on the lightning


return stroke current without having to instrument towers or
[2] The determination of the peak return stroke current
from remotely measured electric or magnetic fields consid-
4
Department of Electrical and Computer Engineering, University of
1
EMC Group, Swiss Federal Institute of Technology (EPFL), Lausanne, Toronto, Toronto, Ontario, Canada.
5
Switzerland. Department of Electrical Engineering, Kyiv Polytechnic Institute,
2
IICT Department, University of Applied Sciences of Western Kyiv, Ukraine.
6
Switzerland (HEIG-VD), Yverdon-les-Bains, Switzerland. Department of Electrical Engineering, University of Bologna,
3
Diagnostic and Monitoring Group Service, ABB Sécheron SA, Bologna, Italy.
7
Genève, Switzerland. Electrical and Computer Engineering Department, Ryerson University,
Toronto, Ontario, Canada.
8
Copyright 2007 by the American Geophysical Union. Department of Engineering Physics, McMaster University, Hamilton,
0148-0227/07/2006JD007958 Ontario, Canada.

D13101 1 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

between lightning return stroke currents and far electromag-


netic fields, taking into account the presence of an elevated
strike object.
[5] The paper is organized as follows: In section 2, the
generalized return stroke models taking into account the
presence of an elevated strike object are summarized. In
section 3, equations relating lightning return stroke currents
and far electromagnetic fields are derived. Simulation
results obtained using engineering models are presented
and compared in section 4. Section 5 is devoted to the
relation between far field peak and current peak. In partic-
ular, the sensitivity of the inferred current peak (from far
field peak) to the parameters involved (e.g., the return stroke
speed) will be discussed. General conclusions are presented
in section 6.

2. Generalized Return Stroke Models Taking


Into Account an Elevated Strike Object
[6] Considering the elevated strike object as a vertically
extended transmission line characterized by constant reflec-
tion coefficients at its top rt and at its bottom rg (see
Figure 1), the general expressions for the spatial-temporal
current distribution along the lightning channel and along
the tower, respectively, are given by Rachidi et al. [2002]:
    
z0  h z0  h
iðz0 ; t Þ ¼ Pðz0  hÞio h; t   rt io h; t 
v* c
X1  #
nþ1 n h þ z0 2nh
þ ð1  rt Þð1 þ rt Þ rg rt io h; t  
n¼0
c c
 0

z h
u t for h < z0 < H0 ð1Þ
v

X 1   
Figure 1. Strike object and channel geometry (H0 is the h  z0 2nh
iðz0 ; t Þ ¼ ð1  rt Þ rnt rng io h; t  
actual height of the return stroke wavefront). n¼0
c c
 0

h þ z 2nh
þ rnt rnþ1
g io h; t  
c c
trigger the lightning artificially, and without the inherent  
h þ z0 2nh
relative inefficiency associated with those methods. This is u t  for 0 z0 h ð2Þ
c c
especially true nowadays because of the widespread use of
lightning location systems. Indeed, such systems are already
where h is the height of the tower, H0 is the height of the
used to provide also estimates of lightning current param-
extending return stroke channel, io(h, t) is the so-called
eters [e.g., Cummins et al., 1998; Herodotou et al., 1993].
‘‘undisturbed current’’, defined as the idealized current that
[3] The theoretical estimation of return stroke currents
would be measured at the tower top if the current reflection
from remote electromagnetic fields depends on the adopted
coefficients at its both extremities were equal to zero, u(t) is
return stroke model. Expressions relating radiated fields and
the Heaviside function, P(z0) is the height-dependent current
return stroke channel-base currents have been derived for
attenuation factor within the lightning channel, v is the
various ‘‘engineering’’ return stroke models for a ground
return stroke speed, v* is the current-wave propagation
initiated lightning [e.g., Rachidi and Thottappillil, 1993].
speed and c is the speed of light. Table 1 summarizes P(z0)
[4] Recently, Bermudez et al. [2005] derived expressions
and v* for the considered five engineering models (BG,
relating lightning return stroke currents and far radiated
TCS, TL, MTLL, and MTLE) [Rakov and Uman, 1998], in
electromagnetic fields, taking into account the effect of an
which, Htot is the total channel height and l is the current
elevated strike object, whose presence is included as an
decay constant.
extension to the transmission line (TL) model. The aim of
[7] Note that equations (1) and (2), which are obtained by
this paper is to generalize the work of Bermudez et al. and
using a distributed source representation of the lightning
provide, for various engineering models, namely Bruce-
return stroke channel [Rachidi et al., 2002], can be equiv-
Golde (BG), transmission line (TL), traveling current source
alently expressed in terms of the ‘‘short-circuit current’’
(TCS), and modified transmission line models MTLL and
isc(t) = 2io(h, t) [Baba and Rakov, 2005b; Rakov, 2001], or
MTLE [Rakov and Uman, 1998], the analytical relationship
the ‘‘reference current’’ [Shigihara and Piantini, 2006].

2 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

Table 1. P(z0) and v* For Different Return Stroke Modelsa 2 H


Z
1 4 r 0
Modelb P(z0) v* H8 ðr; t Þ ¼ iðz ; t  R=cÞdz0
2p R3
BG 1 1 0
TCS 1 c 3
ZH
TL 1 v r @iðz0 ; t  R=cÞ 0 5
MTLL 1  z0/Htot v þ dz ð4Þ
cR2 @t
MTLE exp(z0/l) v 0
a
Adapted from Rakov and Uman [1998].
b
BG is Bruce-Golde; TCS is traveling current source; TL is transmission where r is the horizontal distance between the channel and
line; MTLL and MTLE are modified transmission line models. the observation point, R is the distance between a single
dipole located at a height z0 ffi above ground and the
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
observation point (R = r þ z02 ). H is the height of the
2
return stroke wavefront as seen by the observer (at a given
Note that the ‘‘reference current’’ and the ‘‘short-circuit’’ time t, H is the solution of the equation H/v + R(H)/c = t, where
current coincide when the reflection coefficient at ground is R is the distance between the return stroke wavefront and
equal to 1. Note also that equation (2) does not include any the observation point). Contributions to remote electric and
dependence on P(z0) and v* because the current distribution magnetic fields from currents in the lightning channel and in
along the tower does not depend upon the adopted engi- the CN tower have been previously studied by Kordi et al.
neering model [Pavanello et al., 2004c]. [2003] who used an Antenna Theory (AT) model.
[9] Let us consider here only the radiated electromagnetic
3. Far Field–Current Relationship field. For distant observation points, neglecting the static
and induction components of the electric field, and consid-
[8] The general expressions for the vertical electric field ering R ffi r and r H, the general expression for the
and the azimuthal magnetic field from a vertical antenna electric and magnetic field for an observation point located
above a perfectly conducting ground, for an observation at ground level (see Figure 2) reduces to
point at ground level (see Figure 2), are given by Uman
[1985] ZH
1 @iðz0 ; t  r=cÞ 0
2 Ez ðr; t Þ ffi  dz ð5Þ
ZH Zt 2peo c2 r @t
1 6 2z02  r2 0
Ez ðr; tÞ ¼ 4 iðz0 ; t  R=cÞdtdz0
2peo R5
0 R=c

ZH ZH
2 z02  r2 0 1 @iðz0 ; t  r=cÞ 0
þ iðz ; t  R=cÞdz0 H8 ðr; t Þ ffi dz ð6Þ
cR4 2pcr @t
0 0
3
ZH 2
r @iðz0 ; t  R=cÞ 0 5 [10] Introducing the general expressions for the spatial-
 dz ð3Þ
c2 R3 @t temporal distribution of the current (1) and (2) into (5)
0

Figure 2. Geometry for the far field calculation (H is the height of the return stroke wavefront, as seen
by the observer).

3 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

and (6), and after appropriate mathematical manipulations, it into the tower from its top is reflected back and forth at its
can be obtained [Pavanello et al., 2004a] ends, and portions of it are transmitted into the channel;
these transmitted pulses, which are assumed to travel at the
speed of light, catch up with the return stroke wavefront
Ez ðr; t Þ ffi Ezrs ðr; t Þ þ Ezeso ðr; tÞ þ Ezto ðr; t Þ ð7Þ
traveling at a lower speed and no current is allowed to flow
in the leader region above the front [Pavanello et al.,
2004b]. The return stroke wavefront is assumed in this
H8 ðr; t Þ ffi H8rs ðr; tÞ þ H8eso ðr; tÞ þ H8to ðr; t Þ ð8Þ study not to produce any reflection when reached by faster-
moving current waves produced by transient processes
along the tower and injected into the lightning channel
in which traveling upward at the speed of light. The effect of
[11] 1. Ers rs
z (r, t) and H8 (r, t) are electric and magnetic including reflections at the return stroke wavefront has been
fields due to the main return stroke pulse within the described in previous works [e.g., Heidler and Hopf, 1994;
lightning channel, that is the first term on the right hand side Shostak et al., 1999].
of equation (1), given by [14] The discontinuity which affects the current distribu-
tion at the wavefront, although not physically conceivable,
ZH is predicted by the extended models which take into account
1
Ezrs ðr; t Þ ¼  Pðz0  hÞ the presence of an elevated strike object and needs to be
2peo c2 r
h carefully treated when calculating the radiated electromag-
     netic field through an additional ‘‘turn-on’’ term in the
@ z0  h z0  h
 io h; t  u t dz0 ð9Þ electromagnetic field equations, given by
@t v* v
Ifront ð H Þ 1
Ezto ¼    ð13Þ
ZH 2pe0 c2 r 1 H
þ
1 v cR
H8rs ðr; t Þ ¼ Pðz0  hÞ
2pcr
h
    
@ z0  h z0  h
 io h; t  u t dz0 ð10Þ Ifront ð H Þ 1
@t v* v H8to ¼   ð14Þ
2pcr 1 H
þ
v cR
[12] 2. Eeso eso
z (r, t) and H8 (r, t) are electric and magnetic
fields resulting from the contribution of multiple-reflection where Ifront is the amplitude of the current discontinuity at
process along the elevated strike object, including the the wavefront [Pavanello et al., 2004b].
contribution of pulses transmitted into the channel. They are [15] It is interesting to note that the second and the third
given by Bermudez et al. [2005] terms of the electromagnetic fields, namely, Eeso z (r, t),
Heso to to
8 (r, t) and Ez (r, t), H8 (r, t) are independent of the
1
Ezeso ðr; t þ r=cÞ ¼  ð1  2rt Þio ðh; t Þ adopted model. The only model-dependent terms are Ers z (r, t)
2peo c2 r for the electric field, and Hrs 8 (r, t) for the magnetic field.
8   9 Tables 2 and 3 summarize the specific expressions for
> n n ð2n þ 1Þh >
> r  1 r r i h; t  > ð11Þ Ers rs
z (r, t) and H8 (r, t) developed for various models [e.g.,
1 >< g g t o >
=
ð1  rt Þ X c
Rachidi and Thottappillil, 1993].
  
>
2peo cr n¼0 > 2ðn þ 1Þh > > [16] Additionally, it is important to note that the BG and
>
: þ2rnþ1 nþ1
io h; t  >
;
g rt the TCS models exhibit an inherent discontinuity at the
c
return stroke wavefront. This discontinuity gives rise to a
turn-on term which is already included in the main pulse
1
H8eso ðr; t þ r=cÞ ¼ ð1  2rt Þio ðh; t Þ (Ers rs
z (r, t) and H8 (r, t)) contributions.
2pr
8   9
> n n ð2n þ 1Þh > 4. Simulation Results
>
> g r  1 r r
g t o i h; t  >
> ð12Þ
1 <
X c =
ð1  rt Þ [17] Figures 3 and 4 present simulation results for the
þ  
2peo cr n¼0 >
> 2ðn þ 1Þh >> electric field at a distance of 100 km, for the five considered
>
: þ2rnþ1 nþ1
io h; t  >
;
g rt models and considering two different elevated strike
c
objects. At this distance, the fields are essentially radiation
[13] 3. Finally, Eto to
z (r, t) and H8 (r, t) are the so-called fields, and electric and magnetic fields have the same
‘‘turn-on’’ electric and magnetic field terms [Pavanello et waveshape. The computations are performed using the same
al., 2004b], associated with the current discontinuity at the undisturbed current io(t), given by Nucci et al. [1990]:
return stroke wavefront. Indeed, the current distribution
associated with the extended models (models extended to Io1 ðt=t 1 Þ2  
ðt=t 2 Þ t=t 3 t=t 4
io ðh; t Þ ¼ e þ I o2 e  e ð15Þ
take into account the presence of the elevated strike object) h 1 þ ðt=t 1 Þ2
exhibits a discontinuity at the return stroke wavefront. This
discontinuity arises from the fact that the current injected

4 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

Table 2. Expressions to Calculate Ezrs(r, t) for Different Return [22] 3. The main contribution (about 70%) to the field
Stroke Models, Far Field Conditionsa peak is given by the elevated strike object Eeso z (r, t). Then
Model Ezrs(r, t) comes the contribution of the main return stroke pulse
  Ers
z (r, t) (about 20– 25%), and finally, the contribution due
BG v dio ðh; t Þ
 2
io ðh; t Þ þ t to the turn-on term is only about 15% or less.
2peo c r dt [23] 4. The fast decay of the field right after the first peak
v occurs, respectively, at 1.1 ms for the 168-m tall tower and
TL  io ðh; t Þ 3.7 ms for the 553-m tall one (namely, twice the traveltime
2peo c2 r
along the respective towers) and such a decay is produced
TCS
v by the reflection of the current wave at ground level.
 ½kio ðh; ktÞ  io ðh; t Þ
2peo c2 r
v
5. Far Field Peak–Current Peak Relationship
MTLL  ½io ðh; t  r=cÞ [24] As can be seen from the expressions in Tables 2 and
2peo c2 r 3
Zt 3, among the considered models, the TL is the only model
v for which it is possible to derive a simple, closed form
 io ðh; t  r=cÞdt 5
Htot expression relating the distant electric and magnetic field
0 peaks and the associated return stroke current peaks. This
 rs  has recently been done by Bermudez et al. [2005] who
MTLE
Ez ðr; t þ r=cÞ 1 dEzrs ðr; t þ r=cÞ derived specific expressions for tall and short strike objects.
þ ¼
l v dt [25] For the case when the round-trip propagation time of
Solution of
1 dio ðh; t Þ the tall structure is greater than the zero-to-peak risetime tf,
 2
2peo c r dt the relation between far electric and magnetic field peaks
a and the associated undisturbed current peak at the top of the
Note: k = 1 + v/c.
elevated object is given by Bermudez et al. [2005]. (Note
that Bermudez et al. [2005] disregarded in their derivation
the effect of the discontinuity at the return stroke
where the values of the parameters chosen are Io1 = 9.9 kA, wavefront.)
h = 0.845, t 1 = 0.072 ms, t 2 = 5.0 ms, Io2 = 7.5 kA, t 3 = v h c i
100.0 ms, t 4 = 6.0 ms. These values correspond to the Ezfarpeak ¼  2
1 þ ð1  2rt Þ Io peak ð16Þ
2peo c r v
channel-base current adopted by Nucci et al. [1990] (and in
several other papers) to compare ground-initiated lightning
return stroke models. Other considered parameters are Htot = v h c i
8 km (for the MTLL model), l = 2 km (for the MTLE H8farpeak ¼ 1 þ ð1  2rt Þ Io peak ð17Þ
2pcr v
model) and the return stroke front speed v = 150 m/ms.
[18] The two considered elevated strike objects corre-
spond respectively to simplified models of the Peissenberg
Tower in Germany and the CN Tower in Toronto. The
Peissenberg Tower is characterized by a height h = 168 m Table 3. Expressions to Calculate H8rs(r, t) for Different Return
and reflection coefficients of rt = 0.53 and rg = 0.7 Stroke Models, Far Field Conditionsa
[Heidler et al., 2001]. The adopted parameters of the CN Model rs
H8 (r, t)
Tower are given by h = 553 m, rt = 0.366 and rg = 0.8  
BG v dio ðh; t Þ
[Janischewskyj et al., 1996]. io ðh; t Þ þ t
[19] In the same figures (Figures 3 and 4), the contribu- 2pcr dt
tions of the three terms in equation (7) are also reproduced. TL
v
io ðh; t Þ
The following conclusions can be drawn from the results 2pcr
presented in Figures 3 and 4:
v
[20] 1. No significant differences are found among the TCS ½kio ðh; kt Þ  io ðh; t Þ
various models, especially in the early time region where 2pcr
the peak field occurs (see also Table 4). This conclusion v
MTLL
 ½io ðh; t  r=cÞ
confirms the findings of Pavanello et al. [2004c]. Note that 2pcr
the field peaks associated with the 168-m tall tower are 3
Zt
slightly larger than those associated with the 553-m tall v
 io ðh; t  r=cÞdt 5
tower. Since the risetime of the adopted current (about Htot
0.5 ms) is comparable to the traveltime along the 168-m tall 0
 rs 
tower, the contributions of the current and its reflection at MTLE
H8 ðr; t þ r=cÞ 1 dH8rs ðr; t þ r=cÞ
ground add constructively to the field peak. Whereas for the þ ¼
Solution of
l v dt
553-m tall tower, the current peak occurs before the current 1 dio ðh; t Þ
has reached the ground. 
[21] 2. As predicted by the theory, the contributions 2pcr dt
Eeso to
z (r, t) and Ez (r, t) are independent of the considered
a
Note: k = 1 + v/c.
models.

5 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

Figure 3. Electric field calculated at a distance of 100 km from a lightning return stroke to a 168-m
tower for five different engineering models.

6 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

Figure 4. Electric field calculated at a distance of 100 km from a lightning return stroke to a 553-m
tower for five different engineering models.

7 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

0
Table 4. Electric Field Peaks at 100 km as Predicted by where ktall is defined as
Engineering Models
168-m Tall Tower 553-m Tall Tower 0 1 þ ð1  2rt Þc=v
ktall ¼ ð23Þ
Model Ez peak (V/m) Ez peak(V/m) 1  rt
TL 15.18 13.56
MTLL 15.16 13.54
MTLE 15.11 13.48
and Ipeak is the peak amplitude of the current that would be
BG 15.32 13.69 measured at the top of the elevated strike object.
TCS 15.23 13.59 [30] A short discussion on the terminology is in order.
Bermudez
0
et al. [2005] used the notation ktall instead of
ktall and called this factor the ‘‘tower enhancement factor’’.
In the present paper, we followed the notation and definition
where Io peak is the first peak of the undisturbed current io(h, t). of Baba and Rakov [2005a] in which the term ‘‘tower
Note that, because of the condition tf < 2h/c imposed on the enhancement factor’’ is attributed to the ratio of the far
current, the above are independent of the structure’s height h and field associated with a strike to a tall tower to the far field
of the ground reflection coefficient rg [Bermudez et al., 2005]. associated with a similar strike to ground, expressed by ktall,
[26] For the case of a strike to ground when the reflec- and given by (20) within the model adopted in this study.
tions at ground level are taken into account, the relation- Results obtained using (22) will be compared in part 2
ships between far field peaks and the associated undisturbed [Pavanello et al., 2007] with experimental data consisting
current peak become [Bermudez et al., 2005] of simultaneous measurements of magnetic fields and of the
v h c i return stroke current associated with lightning strikes to the
Ezfarpeak ¼  2
1 þ rchg Io peak ð18Þ Toronto CN Tower (553 m) obtained during the summer of
2peo c r v
2005. 0
[31] Figure 5 presents the variation of the factor ktall as a
v h c i function of the reflection coefficient at the tower top. The
H8farpeak ¼ 1 þ rchg Io peak ð19Þ adopted value for the return stroke speed is v = 150 m/ms. It
2pcr v
can be seen that0
a variation of rt from 0 to 1 results in a
in which rchg = (Zch  Zg)/(Zch + Zg) is the current variation of ktall in the range 3 to 3.5. 0
reflection coefficient at ground level. [32] Figure 6 presents the variation of the factor ktall as a
[27] Comparing (16) – (17) with (18) – (19), we can define function of the return stroke speed. This time, the value of rt
the far field enhancement factor due to the presence of a tall has been chosen
0
to be equal to 0.4. Note that, in Figure 6,
strike object as follows the factor ktall is somewhat more sensitive to the return
stroke speed. However, it is important to realize that, in the
c expression relating the far electric field peak and the current
1 þ ð1  2rt Þ
ktall ¼ v ð20Þ
peak (see0
(22)), the return stroke speed v appears not only
c within ktall but also as a separate proportionality factor. The
1 þ rchg
v overall effect of the return stroke speed on the far field–
current relation is presented in Figure 7. In this figure, two
[28] Baba and Rakov [2005a] have derived a similar typical values for the return stroke speed have been con-
expression for the far field enhancement factor using their sidered, namely, 100 m/ms and 200 m/ms. It can be seen that
model, and they have presented a thorough analysis of the the peak E field and H field radiated by a lightning return
distance dependences of electric and magnetic fields due to
a lightning strike to a tall object and due to the same
lightning strike to flat ground. (The expression for the far
field enhancement factor in the model of Baba and Rakov is
slightly different from (20). As shown by Baba and Rakov
[2005a], the difference is due to the fact that the speed of
current waves propagating along the lightning channel is
assumed to be the speed of light in the present study, which
is based on the work of Rachidi et al. [2002], whereas in the
model by Baba and Rakov, the current waves along the
channel are assumed to travel at the return stroke speed v.)
[29] In what follows, let us express the relationship
between the far electromagnetic field peak and the peak
current that would be measured at the top of a tall strike
object [Bermudez et al., 2005]
v
Ezfarpeak ¼  k 0 Ipeak ð21Þ
2peo c2 r tall
0
Figure 5. Variation of ktall for a tall tower as a function of
v 0
H8farpeak ¼ k Ipeak ð22Þ the top reflection coefficient rt. The return stroke speed is
2pcr tall assumed to be v = 1.5  108 m/s.

8 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

the top of an elevated strike object is relatively insensitive


both to the value of the return stroke
0
speed and the top
reflection coefficient. The factor ktall could find a useful
application in correcting current peak estimates obtained
from remote field measurements for strikes to tall structures,
particularly when these are used to calibrate the perfor-
mance of lightning location systems.
[38] Results and findings of this paper emphasize the
differences between return strokes initiated at ground level
(or from short strike objects) and those striking tall towers:
(1) Radiated electromagnetic field peaks can be significantly
enhanced by the presence of a tall strike object. (2) Electro-
magnetic fields from lightning strikes to tall towers are
largely determined by the tower and only to a lesser extent
by the channel. Consequently, electromagnetic fields asso-

0
Figure 6. Variation of ktall for a tall tower as a function of
the return stroke speed. The top reflection coefficient rt is
assumed to be equal to 0.4.

stroke to a tall structure is little sensitive to the return stroke


speed.

6. Conclusions
[33] In this paper, analytical expressions relating far fields
and currents associated with lightning strikes to tall towers
have been derived. The derived equations are general and
can be used with any engineering model expressed in terms
of P(z0) (height-dependent current attenuation factor), and
v* (current-wave propagation speed).
[34] It is shown that the far field can be decomposed into
three terms, namely (1) contribution of the main return
stroke pulse along the lightning channel, (2) contribution of
the multiple-reflection process along the elevated strike
object, including the contribution of upward propagating
pulses transmitted into the channel, and (3) contribution of
the so-called ‘‘turn-on’’ terms, associated with the current
discontinuity at the return stroke wavefront. This disconti-
nuity is associated with neglecting reflections at the upward
moving front. Only the first term is model-dependent.
[35] It is found that the computed electromagnetic fields
associated with a strike to a tall tower are not very sensitive
to the adopted model for the return stroke. This is especially
true for the early time response of the field and, in
particular, the peak field value. It is also found that the
electromagnetic field from lightning strikes to tall towers is
largely determined by the tower and only to a lesser extent
by the channel.
[36] It is demonstrated, in addition, that the TL is the only
model, among the considered ones, for which it is possible
to derive simple analytical formulas relating current peak
and far field peak values. Such derivation has been recently
presented by Bermudez et al. [2005]. The electromagnetic Figure 7. (a) Electric field peak as a function of return
field peak value being nearly independent of the adopted stroke current peak and (b) magnetic field peak as a function
model, the TL expression becomes a general expression that of return stroke current peak. The observation point is
can be applied for any of the considered engineering return located at a distance of 100 km from the stricken tall
stroke models extended to include a tall strike object. structure. The top reflection coefficient rt is assumed to be
[37] It is shown that the relationship between far electro- equal to 0.4. The adopted return stroke speeds are v = 2 
magnetic fields and the current that would be measured at 108 m/s (solid line) and v = 1  108 m/s (dashed line).

9 of 10
D13101 PAVANELLO ET AL.: RETURN STROKE CURRENTS AND EM FIELDS D13101

ciated with strikes to tall objects are less model-sensitive channel-base current: A review and comparison, J. Geophys. Res., 95,
20,395 – 20,408.
than those corresponding to a strike to ground. In particular, Pavanello, D., F. Rachidi, J. L. Bermudez, and M. Rubinstein (2004a),
the early time response of the field is nearly model-inde- Engineering lightning return stroke models including an elevated strike
pendent. In addition, and unlike ground-initiated strikes, for object: Far field-current relationship, paper presented at the International
Symposium on Electromagnetic Compatibility, EMC EUROPE 2004,
which the far field peak is strongly dependent on the return Eindhoven, The Netherlands, September.
stroke speed (proportional, according to the TL model), far Pavanello, D., F. Rachidi, M. Rubinstein, J. L. Bermudez, and C. A. Nucci
field peaks associated with strikes to tall objects, where a (2004b), Electromagnetic field radiated by lightning to tall towers: Treat-
major contribution comes from the current propagating ment of the discontinuity at the return stroke wave front, J. Geophys.
Res., 109, D06114, doi:10.1029/2003JD004185.
practically at the speed of light within the tall object, are Pavanello, D., F. Rachidi, V. A. Rakov, C. A. Nucci, and J. L. Bermudez
not very sensitive to the return stroke speed within the (2004c), Return stroke current profiles and electromagnetic fields asso-
channel. This is an interesting result when lightning currents ciated with lightning strikes to tall towers: Comparison of engineering
models, paper presented at International Conference on Lightning Protec-
are measured directly on instrumented towers to calibrate tion (ICLP 2004), Avignon, France, September.
the performance of lightning location systems, since in most Pavanello, D., et al. (2007), On return-stroke currents and remote electro-
practical cases the value of the return stroke speed is magnetic fields associated with lightning strikes to tall structures: 2. Ex-
periment and model validation, J. Geophys. Res., doi:10.1029/
unknown. 2006JD007959, in press.
Rachidi, F., and R. Thottappillil (1993), Determination of lightning currents
[39] Acknowledgments. This work has been financially supported by from far electromagnetic fields, J. Geophys. Res., 98, 18,315 – 18,320.
the Swiss National Science Foundation (grant 2000-0681479) and the Rachidi, F., V. A. Rakov, C. A. Nucci, and J. L. Bermudez (2002), Effect of
Natural Sciences and Engineering Research Council of Canada (NSERC). vertically extended strike object on the distribution of current along the
The authors would like to thank V. Rakov and two anonymous reviewers lightning channel, J. Geophys. Res., 107(D23), 4699, doi:10.1029/
for their valuable comments and suggestions. 2002JD002119.
Rakov, V. A. (2001), Transient response of a tall object to lightning, IEEE
References Trans. Electromagn. Compat., 43, 654 – 661.
Rakov, V. A., and M. A. Uman (1998), Review and evaluation of lightning
Baba, Y., and V. A. Rakov (2005a), Lightning electromagnetic environment return stroke models including some aspects of their application, IEEE
in the presence of a tall grounded strike object, J. Geophys. Res., 110, Trans. Electromagn. Compat., 40, 403 – 426.
D09108, doi:10.1029/2004JD005505. Shigihara, M., and A. Piantini (2006), Estimation of lightning currents from
Baba, Y., and V. A. Rakov (2005b), On the use of lumped sources in measurements performed on elevated strike objects, paper presented at
lightning return stroke models, J. Geophys. Res., 110, D03101, International Conference on Lightning Protection (ICLP 2006),
doi:10.1029/2004JD005202. Kanazawa, Japan.
Bermudez, J. L., F. Rachidi, W. Janischewskyj, V. Shostak, M. Rubinstein, Shostak, V., W. Janischewskyj, A. M. Hussein, J. S. Chang, and B. Kordi
D. Pavanello, A. M. Hussein, J. S. Chang, C. A. Nucci, and M. Paolone (1999), Return-stroke current modeling of lightning striking a tall tower
(2005), Far-field – Current relationship based on the TL model for light- accounting for reflections within the growing channel and for upward-
ning return strokes to elevated strike objects, IEEE Trans. Electromagn. connecting discharges, paper presented at 11th International Conference
Compat., 47, 146 – 159. on Atmospheric Electricity, Global Hydrol. and Clim. Cent., Guntersville,
Cummins, K. L., E. P. Krider, and M. D. Malone (1998), The US National Ala.
Lightning Detection Network (TM) and applications of cloud-to-ground Uman, M. A. (1985), Lightning return stroke electric and magnetic fields,
lightning data by electric power utilities, IEEE Trans. Electromagn. Com- J. Geophys. Res., 90, 6121 – 6130.
pat., 40, 465 – 480.
Heidler, F., and C. Hopf (1994), Influence of the lightning channel termina- 
tion on the lightning current and lightning electromagnetic impulse, paper J. L. Bermudez, Diagnostic and Monitoring Group Service, ABB
presented at International Aerospace and Ground Conference on Light- Sécheron SA, Rue de Sablières 4-6, CH-1211 Genève 2, Switzerland.
ning and Static Electricity, Mannheim, Germany. J. S. Chang, Department of Engineering Physics, McMaster University,
Heidler, F., J. Wiesinger, and W. Zischank (2001), Lightning currents mea- 1280 Main Street West, Hamilton, Ontario, Canada L8S 4L7.
sured at a telecommunication tower from 1992 to 1998, paper presented A. M. Hussein, Electrical and Computer Engineering Department,
at 14th International Zurich Symposium on Electromagnetic Compatibil- Ryerson University, 350 Victoria Street, Toronto, Ontario, Canada M5B
ity, Zurich, Switzerland, 20 – 22 February. 2K3.
Herodotou, N., W. A. Chisholm, and W. Janischewskyj (1993), Distribution W. Janischewskyj, Edward S. Rogers Sr Department of Electrical and
of lightning peak stroke currents in Ontario using an LLP system, IEEE Computer Engineering, University of Toronto, 10 King’s College Road,
Trans. Power Delivery, 8, 1331 – 1339. Toronto, Ontario, Canada, M5S 3G4.
Janischewskyj, W., V. Shostak, J. Barratt, A. M. Hussein, I. Rusan, and J. S. C. A. Nucci, Department of Electrical Engineering, University of
Chang (1996), Collection and use of lightning return stroke parameters Bologna, Viale Risorgimento, 2 40136 Bologna, Italy.
taking into account characteristics of the struck object, paper presented at D. Pavanello and F. Rachidi, EMC Group, Swiss Federal Institute of
23rd International Conference on Lightning Protection (ICLP), Florence, Technology (EPFL), EPFL-STI-ISE-LRE, Station 11, CH-1015 Lausanne,
Italy. Switzerland. (farhad.rachidi@epfl.ch)
Kordi, B., R. Moini, W. Janischewskyj, A. M. Hussein, V. O. Shostak, and
M. Rubinstein, IICT Department, University of Applied Sciences of
V. A. Rakov (2003), Application of the antenna theory model to a tall
Western Switzerland (HEIG-VD), Rte de Cheseaux 1, CH-1401, Yverdon-
tower struck by lightning, J. Geophys. Res., 108(D17), 4542,
les-Bains, Switzerland.
doi:10.1029/2003JD003398.
V. Shostak, Department of Electrical Engineering, Kyiv Polytechnic
Nucci, C. A., G. Diendorfer, M. Uman, F. Rachidi, M. Ianoz, and
Institute, KPI-1670, 37 Prospect Peremohy, Kyiv, 03056, Ukraine.
C. Mazzetti (1990), Lightning return stroke current models with specified

10 of 10

You might also like