92 / 132
Technical article
March 2017
90
www.read-eurowire.comAn advantage of the online method is
the absence of reflections from the far
end. The breakdown causes a very low
impedance at its location and the signals
are reflected from here. A simplified circuit
for online measurements is shown in
Figure 1
.
The measurement on both cable ends with
two measuring devices improves the fault
location accuracy. Of course, this option
depends on the configuration of the
power cable system and the access to its
cable ends. This option is not considered in
the experimental tests yet.
Theoretical
Considerations
and Simulation
The physics of cables and their behaviour
is very complex and has been widely
discussed in literature. It shall not repeated
in this paper (example for reference see
[4]
) .
Only two basic equations are needed here:
When using this kind of TDR the
exact knowledge of the propagation
velocity
v
determines the accuracy of
the fault location. (It differs to the TDR
measurement
for
partial
discharge
(PD) fault location where only the time
relation of the reflections determines the
accuracy.) Therefore this propagation
velocity has to be known exactly to be
determined in advance. When the para-
meters
L’
and
C’
of the cable are effectually
known, the propagation velocity can be
calculated by
Equation 1
. However, if it is
possible, an initial measurement of the
propagation velocity should be done for
each commissioned cable.
The situation changes when the TDR
signals are measured on both cable ends.
Then the knowledge of the velocity is not
necessary (similar to the PD fault location)
and the fault location is calculated by:
with
T
x
and
T
y
as the signal propagation
measured from both cable ends. Of course,
the calculation by knowledge of the
propagation velocity it still valid and the
measurements can be verified when the
right cable length is also known.
The test circuit was simulated with
OrCAD PSpice and with realistic cable
parameters
[5]
. It allows the simulation of
the signal propagation in very long cables
and the signal distortion by the measuring
circuit on the cable end.
The simulation was made with a cable
length of 100km and a propagation
velocity of 171.25m/µs. The failure was
simulated at a distance of 83km from the
cable end where the measuring circuit was
connected.
The simulation results in
Figure 3
show a time
T
= 970 µs and with the
aforementioned velocity
v
the distance to
the failure is calculated to
l
x
= 83.06km. The
negligible deviation from the reference
value is the result of a slightly inaccurate
time measurement of the simulation
results.
Measuring Equipment
The measuring circuit consists of two
main components, the HV divider and
the transient recorder. While only one
type of transient recorder processes the
signals from measurements on AC and DC
cables, the HV dividers differ for AC and DC
applications.
A capacitive HV divider is preferably
used for measurements on AC cables.
For DC cables a broadband divider with
a resistive arm is necessary to achieve
the required response characteristic. This
response characteristic is also essential
when other voltage measuring devices are
taken for the online TDR measurements,
eg instrument transformers which are
installed in power nets. Their ability has
still to be approved.
The triggering of the signal processing is
also essential for the measuring quality
and accuracy.
The simulation and experiments confirmed
that a simple edge triggering is sufficient
on DC cable systems. On AC cable systems
the operational voltage itself prevents such
simple edge trigger. A very fast breakdown
detection device is therefore implemented
to release the signal processing.
The required components of the transient
recorder depend on its application for
cable testing or monitoring. The recorder
for cable testing is a small plug-in device
which operates by its own PC or by that
of the computerised HV test system.
It mainly contains the measurement
hardware (
Figure 4
).
The recorder for cable monitoring is a
robust and small stand-alone device.
Besides the measurement hardware it
contains a PC with especially adapted
software. This PC runs continuously over
years, can be restarted and operated
remotely, and must be supplied by an
uninterrupted power supply (UPS) in case
the cable fails (
Figure 4
).
Experimental Tests
Practical measurements on different cable
samples were carried out to prove the
measuring principle and the simulation
results. Thereby the AC or DC cable
samples were wound on cable drums or
turntables.
Capacity
to ground
HV
AC/DC
source
▲
▲
Figure 1
:
Principle circuit for online fault location
Equ. 1
Equ. 2
Equ. 3
▼
▼
Figure 2
:
Simulated circuit
▼
▼
Figure 3
:
Simulation results
▲
▲
Figure 4
:
3D models of transient recorders for cable
tests (left) and cable monitoring (right)
HVDivider