Journal of Mathematics and Statistics 6 (3): 221-225, 2010

ISSN 1549-3644
© 2010 Science Publications

The Magnetic Field Induced by a Lightning Strike’s Indirect

Effect Double Exponential Current Waveform
1
Scott L. Meredith, 1Susan K. Earles, 1Ivica N. Kostanic and 2Niescja E. Turner
1
Department of Electrical and Computer Engineering,
2
Department of Physics and Space Sciences,
Florida Institute of Technology, Melbourne, Florida, USA

Abstract: Problem statement: Develop a new formula which describes the magnetic field induced by
a lightning strike’s indirect effect double exponential current waveform. Approach: A novel approach
for developing a closed-form solution for the magnetic field from the indirect effect double
exponential current waveform will be presented. In the literature, models typically employ the pulse
waveform to derive the corresponding electromagnetic fields. However, given the Department of
Defense (DoD) has incorporated the double exponential current waveform as part of their
“Electromagnetic Environmental Effects Requirements For Systems”, we felt it important to develop a
solution for the magnetic field which utilized this waveform. In order to facilitate the integration
required for deriving the field, Taylor series expansion was used for all variable dependent exponential
terms. In many publications, the dipole and monopole techniques have been used when solving for the
magnetic field. However, for this study the dipole technique was deemed the preferred method for
evaluating the field. A derivation of the magnetic field will be presented along with a graphical
illustration of the field’s distribution over time. Results: The equation presented utilized Taylor series
to augment the integration required to solve for the magnetic field. Conclusion: A new method for
deriving the magnetic field induced by a lightning strike’s indirect effect double exponential has been
presented. By approximating the variable dependent exponential terms, we were able to minimize the
complexity of the mathematics required to solve for the magnetic field in closed-form.

Key words: Magnetic field, double exponential waveform, lightning and dipole technique

INTRODUCTION For the purposes of this study we will be

utilizing the severe stroke current waveform
Bewley (1929) proposed using the double
parameters given in (MIL-STD-464, 1997). The
exponential to describe the wave shape a lightning
current waveform may have. What he found were values for I0, α and β are 218,810 A, 11,354 s-1 and
measurements obtained in both the field and lab closely 647,265 s-1 respectively.
matched wave shapes obtained by analysis. This Given this waveform’s usage in literature, a
discovery gave rise to an empirical equation which validation of the parameters used was warranted. Jia
closely resembled lightning’s characteristic shape. and Xiaoqing (2006) used numerical trial and error to
Currently, the Department of Defense (DoD) solve the double exponential parameters α and β. Upon
utilizes the double exponential current waveform as substituting these values into the equation they found
part of their Electromagnetic Environmental Effects that the resulting waveform closely models the one
document MIL-STD-464 to describe the lightning’s which utilized the fixed parameters. Karwowski and
indirect effect waveform. The parameters (I0, α and β) Zeddam (1995) developed a method that utilized the
which make up the waveform are contingent upon the mixed-potential Electric Field Integral Equation (EFIE)
type and severity of the lightning strike. The double in the frequency domain to solve for the electric field
exponential current waveform i(t) described in (MIL- numerically using method of moments. However, this
STD-464, 1997) is shown by: study will introduce a closed-form solution for the
magnetic field which is induced by the double
i ( t ) = I0 ( e −αt − e −βt ) (1) exponential current waveform.
Corresponding Author: Scott L. Meredith, Department of Electrical and Computer Engineering, Florida Institute of Technology,
Melbourne, Florida, USA
221
 J. Math. & Stat., 6 (3): 221-225, 2010

MATERIALS AND METHODS This Retarded time, t’ is used as a means of describing

the time delay that exists between a photon being
Retarded time: In differential form Maxwell’s emitted and when it is perceived by an observer at a
equations for a homogeneous, time variant and linear distance R some R/c time later. The Retarded time or
medium can be written (Jackson, 1999), where D is the time delay is defined as:
electric displacement, E is the electric field, B is the
magnetic field, H is the magnetic field strength, J is the t' = t − R / c (10)
current density, ρ is the charge distribution per unit
volume, µ 0 the magnetic permeability and ε0 is the Where:
electric permittivity: t’ = The actually time of photon emission
t = The time it’s perceived by an observer
∇ ⋅ E = ρ / ε0 (2) R/c = The time it took to travel

∂H The inclusion of the retarded time is used to

∇ × E = −µ0 (3)
∂t account for the delay that exists between the lightning
channel and observer. By differentiating both sides of
∂E (7) and inserting (10) to account for the Retarded time,
∇ × H = J + ε0 (4) one could now describe the potential A with:
∂t

∇ ⋅ µ0H = 0 µ I ( rs , t − R / c ) dz '
(5) dA z ( rs , t ) = (11)
4π R
Given Eq. 5, one can solve for the magnetic field in
Note: By canceling out the areas we can write the
terms of the vector potential alone. Upon utilizing
current density per unit area J in terms of I and dV’ in
some substitutions and vector identities one would
terms of dz’.
obtain:
Upon substituting (11) into (9), one can develop a
general expression for the magnetic field at any point in
µ0H = ∇ × A (6)
(r, z) space such that:

where the potential A in terms of z’ equals: 1  r ∂i ( t − R / c ) r 

dH Φ =  + 3 i ( t − R / c )  dz ' (12)
4π  cR 2 ∂t R 
µ I ( r , t ) dz '
A z ( rs , t ) = 0 ∫ s (7)
4π z ' R Image theory: Since the conductivity of the earth’s
surface increases with frequency, it is logical to assume
Upon applying the Curl operator in cylindrical that above a certain frequency the conductivity σ,
coordinates for (6), we can now re-write the magnetic approaches infinity. Therefore, if we treat the earth as a
field such that: perfect conductor, we are able to use image theory in
our analysis. These images, or virtual sources, account
1  1 ∂A z ∂A z  for reflections which can then be added to the real
H= ir − iϕ (8)
µ 0  r ∂ϕ ∂r  source constituents to form a general equation. An
illustration of this phenomenon is depicted in Fig. 1
which captures the lightning channel and its
Given the radial symmetry of the problem, it is
corresponding image. In order to account for the
apparent that Az cannot depend on the coordinate φ thus
current waveform that moves upward at a velocity ν, we
(8) simplifies to:
now must include the term |z’|/v (Rubinstein and Uman,
1989). With this additional term we can now describe
1 ∂A z
H=− iϕ (9) the current waveform (1) with the following:
µ 0 ∂r
( )
i(t − z ' / v ) = I 0 e − α(t − z ' / v ) − e −β(t − z ' / v ) ⋅ u (t − z ' / v ) (13)
As pointed out by Nevels and Shin (2001), Lorenz
proposed that the standard Neumann potentials in terms of where, |z’| is written to include all points along the z-
instantaneous charge, current density and position be axis both real and reflected while the Heaviside
modified to include the propagation of time from the source. function, u(t) is required to bound the current waveform.
222
 J. Math. & Stat., 6 (3): 221-225, 2010

After some diligent algebraic manipulations, the

height can now be expressed by:

( ct − β z ) ± ( z − βct ) + (1 − β2 ) r 2
2

h =β (18)
1 − β2

where the quantity β = v/c is the ratio of the current

propagation speed along the lightning channel and the
speed of light. Given the formula is in quadratic form,
two possible solutions exist for h. One solution is
Causal while the other is non-Causal. Causal signals
are zero for all negative values of t while non-Causal
signals are zero for all positive values of t. In this case,
only the Causal form of h will be considered as one of
Fig. 1: Application of image theory used for lightning the participating parameters for evaluating the magnetic
channel field.
Approximating the double exponential: Given the
As Rubinstein and Uman (1989) pointed out, by complexity of the integration required to solve for the
assuming no current flows above height h in the magnetic field, we must separate the exponential terms
lightning channel, allows one to set the Heaviside into their constant and variable constituents. Once the
function equal to zero. The known absence of current two have been separated, Taylor series expansion can
permits one to conclude that time at this point must be used to approximate the exponential terms which
equal zero as well. Given the propagating current contain the variable z’. The expansion for the
waveform shifts in time along both the vertical and exponential terms can be approximated by:
horizontal axis, allows one to include both of these
conditions when defining the time at the top of the x2 x3
ex ≈ 1 + x + + + ... (19)
channel. With this in mind, one should now encompass 2! 3!
the horizontal or “Retarded time” shift, along with the
waveform’s vertical movement |z’|/v, to develop a time- In principle, there are two double exponential
based expression at the top of the channel. In doing so, waveforms that must be approximated. These include:
we can expand Eq. 13 to include the Retarded time such the original waveform and differentiated waveform.
that: Each of which will have been shifted in time by R/c +
z’/v. From (14), one can separate the exponentials in
(
i(z ' , t − R / c ) = I 0 e − α(t − R / c − z ' / v ) − e −β(t − R / c − z ' / v ) ) (14)
terms of their constants and variables constituents.
⋅ u (t − R / c − z ' / v )
Doing so allows one to utilize Taylor series to account
for the variable terms while constants can be set aside
and incorporated at a later time. Therefore from (14),
Given: we can distribute α and β then regroup the exponential
terms shown by (20-21):
t − R / c − | z ' | /v = 0 (15)
i ( z ', t − R / c ) = I0  e −αt ⋅ eαR / c ⋅ eαz '/ v − e −βt ⋅ eβR / c ⋅ eβz '/ v 
and (20)
⋅ u ( t − R / c − z '/ v )

R= ( z − z ') + r2
2
i ( z ', t − R / c ) = I0 e −αt ⋅ e ( )
(16) α R / c + z '/ v β ( R /c + z '/ v )
−e −βt ⋅ e 
  (21)
⋅ u ( t − R / c − z '/ v )
one can derive an equation in terms of some height h.
Plugging (16) into (15), setting z’ equal to an arbitrary
Given (21), we can make the following
height h and solving for t one would obtain:
assignments:

(z − h) + r2
2
h  R z'
t= + (17) xα = α  +  (22)
c v c v
223
 J. Math. & Stat., 6 (3): 221-225, 2010

larger magnitude and takes longer to decay. However,

when this approximated waveform is modified with the
seminal constituent 1.06*exp(-1.25αt), the resulting
solution reveals a waveform that closely models the
original double exponential shifted by R/c + z’/v. By
realizing the acquiescent similarities between the
original and approximated waveforms allows one to
leverage the latter to derive the magnetic field in
closed-form with a high level of confidence.

RESULTS

Calculating the fields: The methodology used in

deriving the magnetic field will forgo using the
standard square pulse presented in literature (Rubinstein
and Uman, 1989) and replace it with the double
Fig. 2: Modified and unmodified double exponential exponential waveform. Let’s first start off by re-
current waveforms shifted by t = R/c + z’/v writing (12) such that:

1  r ∂i ( t − R / c ) r 
and: dH Φ =  + 3 i ( t − R / c )  dz ' (27)
4π  cR 2 ∂t R 
 R z' 
xβ = β  +  (23)
c v Next, substitute (25, 26) into the second and first
terms of (27) respectively. Upon doing so, one will
Therefore, one can leverage the first two terms arrive at:
from (19) where (22, 23) equal the variable x as given

( ( )
by: I 0dz '  r
dH Φ = βe −βt − αe −αt + β 2 e −βt − α 2 e −αt
4 π  cR 2
  R z' 
i ( z ', t − R / c ) = I0 e −αt ⋅  1 + α  +    R z'  
  c v  ⋅  +   ⋅ u (t − R / c − z ' / v )
 c v 
  R z '  
−e −βt ⋅ 1 + β  +   
  c v  
(24)
+
r
cR 2
( (
e −αt − e −βt + αe −αt − βe −βt )
⋅ u ( t − R / c − z '/ v ) (28)
 R z'  
⋅  +   ⋅ δ(t − R / c − z ' / v )
 c v 
Upon regrouping the “like” terms, (24) can now be
written as: +
r −αt
R3
e ( (
− e −βt + αe −αt − βe −βt )
i ( z ', t − R / c ) = I0  e −αt − e −βt + ( αe −αt − βe −βt )  R z'  
⋅  +   ⋅ u (t − R / c − z ' / v )]
(25)  c v 
 R z' 
⋅  +  ⋅u ( t − R / c − z '/ v ) 
c v
Note the additional term added to (28) is due in
with differentiated form of (25) being: part to the following product rule:

i ( z ', t − R / c ) = I0 β e −βt − αe −αt + ( β2 e−βt − α 2 e −αt ) f ( t ) ⋅ u ( t )  = f ( t ) ⋅ u ( t ) + f ( t ) ⋅ δ ( t )

' '
(29)
 R z'  (26)
⋅  +  ⋅u ( t − R / c − z '/ v ) 
c v Although the terms bound by the delta function are
shown in (28), they vanish once the current waveforms
As the Fig. 2 shows, the modified waveform which are evaluated at t = R/c + z’/v. Upon integrating Eq. 28
utilized Taylor series expansion tends to yield a slightly from -h to h along the z = 0 plane, we can obtain a
224
 J. Math. & Stat., 6 (3): 221-225, 2010

closed-form solution for the magnetic field to include CONCLUSION

the correction factor as shown by the following:
A closed-form solution for the magnetic field
I0 (1.06e −1.25 αt
)  r (β e
2 −βt
−α e 2 −αt
) induced by the double exponential current waveform
H Φ ( r,0, t ) = has been presented. The approach proposed here is
2π  c 2
unique in that it utilized Taylor series to approximate
h2 + r2 h the variable dependent exponential terms within the
⋅ ln + current waveform. This greatly simplified the
r r
mathematics which helped facilitate the integration
r ( β2 e −βt − α 2 e −αt ) r required to solve for the solution in a closed-form.
− ⋅ ln (30) However, this approximation caused the modified
cv h + r2
2
current waveform to grow in both magnitude and decay
+
(e −αt
− e −βt )
⋅
h time. With this in mind, it is logical to presume that the
r h2 + r2 resulting fields would be subject to this same anomaly.

+
( αe −αt
− βe −βt ) 
⋅ 1 −
r 
Realizing a departure existed between the modified and
unmodified waveforms, warranted the use of a

v  h + r  
2 2
correction factor to recapture its original form. Doing
so ensured the magnetic fields which followed, would
Figure 3 illustrates how the magnetic field’s tend to more closely model those from the original
magnitude is distributed as time varies from 0-1×10−5 unmodified current waveform.
sec.
REFERENCES
DISCUSSION
Bewley L.V., 1929. Traveling waves due to lightning.
Equation 30 used a methodology which greatly AIEE Trans., 49: 1050-1064.
simplified the integration required to solve for the Department of Defense Interface Standards, 1997.
magnetic field in a closed-form. By utilizing Taylor Electromagnetic Environmental Effects
series to approximate the exponential terms, the Requirements for Systems. MIL-STD-464, pp: 1-10.
solution’s accuracy became a function of the number of Jackson, J.D., 1999. Classical Electrodynamics. 3rd Edn.,
terms used from (19). Consequently, a correction factor John Wiley and Sons Inc, New Jersey, USA,
needed to be added in order to help reshape the ISBN: 0-471-30932-X, pp: 24-63.
magnetic field which resulted from the modified current Jia, W. and Z. Xiaoqing, 2006. Double-Exponential
waveform. Expression of Lightning Current Waveforms.
CEEM Dalian 3A1-09, pp: 320-323.
Karwowski A. and A. Zeddam, 1995. Transient
currents on lightning protection systems due to the
indirect effect. IEE Proc. Sci. Meas. Technol.,
142: 231-222.
Nevels, R. and C. Shin, 2001. Lorenz, Lorentz and the
gauge. IEEE Antenn. Propagat. Mag., 43: 70-72.
Rubinstein, M. and M.A. Uman, 1989. Methods for
calculating the electromagnetic fields from a
known source distribution: Application to
lightning. IEEE Trans. Elect. Compat., 31: 183-189.

Fig. 3: Magnetic fields from the approximated double

exponential current waveforms at various
heights z
225