UNSW - SCHOOL OF ELECTRICAL ENGINEERING AND TELECOMMUNICATIONS

ELEC4611 POWER SYSTEM EQUIPMENT LABORATORY

EXPERIMENT 2
SURGE PROPAGATION IN ELECTRICAL SYSTEMS

Overvoltage surges are generated frequently in electrical systems: they may be externally
generated as in the case of a lightning strike to a line or internally generated, usually by
some switching action. As the amplitude of these surges may be considerably greater
than the normal 50 Hz peak voltage of the line, knowledge of the behaviour of such
surges on transmission lines is important so that the system can be protected against
failure of insulation. It is necessary to know how fast the surge travels, whether there is
any damping and, particularly, the effect of terminations on the surge. This experiment is
primarily concerned with the latter behaviour.

The experiment is divided into two sections:

A. The propagation of short duration pulses in a length of coaxial cable (a distributed
parameter line).
B. The performance of a simulated line (using discrete L and C components) when
subjected to long duration pulses (a lumped parameter model).

The effects of various terminations are to be studied for both cases and related to the
simple theory of surge propagation. Over-voltage surges are generally of quite short
duration and fast rise time. Typically, rise times are 1-100 s and durations are 50-500 s
in practical situations.

1. THEORY

A brief review of the theory is given here (a more detailed description is contained in the
lecture notes issued). A lossless transmission line can be represented electrically by
distributed series inductance and shunt capacitance (a lossy line will have in addition
series resistance and shunt conductance). A surge incident on the line at any position will
generate surge voltages propagating in both directions along the line.

If l is the length of the transmission line, L is the series inductance per unit length, C is
the shunt capacitance per unit length and Lo and Co are total inductance and capacitance
of the line (Lo= l L, Co= l C), the velocity of propagation and transit time on a lossless
transmission line are:
1 l l
v= = and t o = = l. LC = Lo C o
LC Lo Co v

From the relevant expressions for L and C for the appropriate line geometry, it is easily
shown that the above reduce to:
1 c 1
v = = [c= ] and to = l . 
 rr oo

where c is the velocity of light in vacuum, r and r are the relative permittivity and
permeability of the insulating medium and l is the length of the line. Thus:
for overhead lines: r = r = 1 and v = c

ELEC4611 - Experiment 2: Surge propagation in electrical systems - Page 1 - 2020

 for coaxial cables: r > 1, r = 1 and v = c / r ( < c)

The voltage and current of the surge at any point on the line are related by the
characteristic, or surge, impedance of the line. For a lossless line, the surge impedance Zo
is a pure resistance and is given by:
L
Zo = C ohms

At the end of the line, the behaviour of the surge depends on the termination. If the
termination is equal to the surge impedance, the surge is completely absorbed: however,
for a general load, the voltage developed in a load ZT at the end of the transmission line,
due to an incident step voltage surge, V, is:
ZT
VT = 2 V Z + Z
o T

and the reflected portion of the surge is:

V (ZT - Zo)
VR = Z + Z
T o

For multiple reflections, the incident wave becomes that reflected from the opposite end.
A graphical representation may be developed by use of the Bewley Lattice diagram.

If the termination is not a pure resistance, there may be some other behaviour occurring:

(a) For another transmission line or cable of different surge impedance, an incident surge
will be transmitted and reflected according to the surge impedance of the other line.
The above formulae hold.

(b) If a transmission line or cable bifurcates or trifurcates, the reflection and transmission
coefficients will be determined by the effective surge impedance of the parallel
combination, e.g. for 50  cable bifurcating into two parallel 50  cables at the
termination, the effective surge impedance after the point is 25 . Note that the
transmitted voltage is the same in each parallel arm, even if they are of different Zo.

(c) If the line terminates in an energy storage element, C or L, the element voltage varies
with a time constant of ZoC or L/Zo.

2. PROCEDURE

Part A: Distributed Parameter Line - Coaxial Cable

In this part, we use 100 metres of RG58c/u 50 coaxial cable as the basic transmission
line. We are trying to observe the reflection in a relatively short length of line (100
metres) and the incident and reflection voltages are thus closely spaced in time. We
therefore use the oscilloscope at a very high sweep rate (about 1s/division) and repeat
the pulse frequently (about 100 kHz). The CRO pre-trigger facility enables it to be
triggered just prior to the creation of the pulse.

Connect the source, line and CRO as shown in Fig.1 (Ch.1 will be the sending end pulse
and Ch.2 the receiving end pulse). Observe the receiving and sending end pulses on the
CRO when the receiving end impedance on the 50  cable is:

ELEC4611 - Experiment 2: Surge propagation in electrical systems - Page 2 - 2020

 (i) ZT = 0 (short circuit) (ii) ZT = 18 
(iii) ZT = 50  (iv) ZT = 82 
(v) ZT = infinite (open circuit) (vi) ZT = 75 

[The loads are enclosed in BNC plugs with caps marked to indicate the values].

Fig.1: Circuit for Part A (note: set CRO and generator to suit short transmission times).

For each of the above load conditions, vary the pulse width for 3 different values (0.6 s,
1.5 s and 3 s) and observe the effect on the pulse shapes. Note that the transit time
along the 100 m of cable is 0.1 s < to < 3 s.

Calculate the reflection coefficients (VR/V) and transmission coefficients (VT/V) from
the Z values and compare with the measured values. Determine the dielectric constant r
of the line from the observed transit time and the length of the cable (100 metres). When
calculating the coefficients, ensure the pulse width used is long enough to give a constant
level of reflected voltage.

Also provided is a 70 m length of 75  cable. Connect this to the end of the 100m/50
cable and monitor the voltage at each end as before. Consider both open circuit and short
circuit terminations of the 75  cable. Note also the multiple reflections. Determine the
transmission and reflection coefficients at the joint and compare it to the theoretical
values.

Part B: Lumped Parameter (Artificial) Line

In this section we use an artificial line made up of 30 discrete sections of L  1.5

mH/section and C  9 nF/section. Thus, the characteristic impedance is about 400 ,
which is typical of an overhead power line.

As the line impedance is higher than that of the coaxial cable, the pulse unit is connected
through an internal resistor and when the input switch is on Zo, the input end of the line
is approximately matched.

ELEC4611 - Experiment 2: Surge propagation in electrical systems - Page 3 - 2020

 Fig.2: Circuit for Part B (note: set CRO and generator to suit long transmission times).

Connect the circuit as shown in Fig.2. Using pulses of 50 s and 3 ms duration, observe
the waveshapes at the end of the line and near the centre for the following conditions:
(i) Zs = Zo and ZT = Zo (ii) Zs = Zo and ZT = infinite
(iii) Zs = Zo and ZT = 0 (iv) Zs = Zo and ZT = 0.1 F

From the theory, explain the waveshapes and calculate and compare the velocity and the
transit time to the expected values, assuming this is used to model an air-insulated
transmission line. Compare the time constant of the voltage across the 0.1 F capacitor to
the expected value. (Note: It will help in understanding the waveforms if you vary the
pulse width between the two extremes and note the effect). Use clip probe to check the
progress of the pulse along the line (especially for pulse width of 50 s with ZT=Zo,
infinite and 0).

3. DISCUSSION AND CONCLUSIONS

(i) From the measured transit time determine, for the coaxial cable, the values of L and
C, assuming the usual expressions for L and C. You should not require the cable
dimensions.

Using the measured transit time of the artificial line and the given L and C values,
determine the equivalent length of two-wire, air insulated, overhead line
represented by the lumped parameter model.

(ii) Comment on any differences between the cable and the model line and any
limitations of the artificial model line (e.g. attenuation and phase properties, effect
of frequency, etc).

(iii) Discuss the differences between lightning and switching surges in electrical
systems.

ELEC4611 - Experiment 2: Surge propagation in electrical systems - Page 4 - 2020

REFERENCES

(i) E.W. Kimbark, Electrical Transmission of Power and Signals, Wiley, 1949.
(ii) A. Greenwood, Electrical Transients in Power Systems, 2nd ed., Wiley, 1991.
(iii) W.C. Johnson, Transmission Lines and Networks, McGraw-Hill, 1950.
(iv) J.D. Glover, T.J. Overbye and M.S. Sarma, Power System Analysis and Design, 6th
ed., Cengage Learning, 2016.

NOTES:

(i) For a coaxial cable:

 b
L= ln   H/m and C = 2 F/m
2  a  ln (b / a )

(ii) A pulse generator has been made up for this experiment and has a correctly
matched 50  impedance output. This circuit is housed in a small box.

==========

ELEC4611 - Experiment 2: Surge propagation in electrical systems - Page 5 - 2020