Sei sulla pagina 1di 8

1

Automation of high voltage impulse tests

1HAEFELY

World Electrotechnical Congress 1977, Moskow

K. Feser, R. Niederhauser

El-37

Abstract

Rhum6

Nowadays two tendencies for test plants can be observed:


- on one side, highly specialised industrial plants for one specific
product
- on the other, universal laboratories for different research works
with a high flexibility.
These two extreme applications lead to different solutions in
automation.
Illustrated with examples of actual installations, a survey is given
on the various possibilities for automation of high voltage impulse
tests. Emphasis is placed not only on the endeavours to automate
the high voltage side (selection of resistors. choice of generator circuits, connection of test object), but also on the efforts to automate
the generator control systems, the measurements and the subsequent evaluation of the test results.
By using electronic components and computers in high voltage
laboratories, many grounding, screening and designing problems
on these instruments have to be solved. Discussions cover recently
developed control and measuring instruments, such as impulse
peak voltmeter, program controller, computer, trigatron-device and
special impulse oscilloscopes. The satisfactory functioning of all
these control and measuring instruments near or in connection with
high voltage test systems will be discussed as one of the most severe
problems for electronic devices used in high voltage impulse tests.

Deux tendances peuvent &tre observees actuellement:


- dune part, des installations industrielles hautement specialisees
pour un seul produit
- dautre part, des laboratoires ((umversels) a tres grande flexibilite
pour des travaux de recherche.
Ces deux applications conduisent a differentes solutions
dautomatisation.
Un apercu des diverses possibilites dautomatisation des essais
de choc a haute tension est discute, illustre par des exemples dinstallations deja realisees. Un effort particulier est port6 sur
lautomatisation non seulement du tote de la haute tension (choix
des resistances, choix du circuit de connexion et raccordement de
lobjet dessai), mais encore du cot& de la commande, de la mesure
et de Y&valuation des resultats.
Lutilisation de composants electroniques et le raccordement
dordinateurs au cot6 haute tension posent de nombreux problemes
de mise a la terre, de blindage et de construction des appareils Clectroniques. Les r&cents developpements dappareils de commande et
de mesure, comme le voltmetre de choc, le dispositif de commande
programmee, lordinateur, le trigatron et les oscilloscopes de choc
speciaux sont passes en revue. Le bon fonctionnement de tous ces
appareils a proximiti immediate et en liaison avec des sysdmes
dessais a haute tension constitue un des problemes les plus ardus
qui sera egalement examine.

Fig. 14 Automated impulse test system. [41

-.-

Automation of high voltage impulse tests


K. Feser. R. Niederhauser

I. Introduction

The progress on rationalised manufacturing is bound to the establishment of adequate test procedures for preliminary, intermediate and final production tests at favorable costs. In highly
developed industrial countries, the remarkably increasing personal
costs have to be taken into consideration. Beside the costs, quality
assurance influences the reflections on automation of testing
processes.
From todays view of testing practice, two different directions
become apparent with very different requirements to automation:
- the automation of a special testing procedure for a certain
product or
- a flexible automation, which can be adapted to changing
requirements.
Both possibilities have their specific fields of application.
As a rule, the first one can be realised economically with less
expenses for hardware and software for testing a product. The main
application therefore is testing mass products. On the contrary, the
second solution requires complicated hardware and software as
well as higher initial investments, but can be adapted much quicker
and generally at lower costs to new duties. Furthermore, service and
maintenance expenditure of the second solution are mostly more
important.
Nowadays industrial tests show a clear tendency towards
product specific solutions, whereas in the field of development and
research the universal solution is prefered.
For economical reasons, the manufacturer of test systems, e.g.
for impulse testing, will try to integrate both solutions into one
system, that can be enlarged at any later date. As an example, the
control of an impulse voltage system may be mentioned. Hereby the
step by step automation starts with manual control (first step), via
the automatic firing of the generator at a given voltage (second
step), to the automatically adjusted sphere gap distance and the
regulated charging of the generator (third step), to simple control
devices yielding specific predetermined processes (fourth step). The
fifth step, control by a computer, allows to solve any problem. As a
sixth step the impulse testing of e.g. a transformer (including
measurement such as resistance, loss factor, ac voltage test, partial
discharge, etc) can be integrated into a whole computer controlled
process.
From todays point of view automatic test equipment for impu!se
voltage tests is justified in following cases:
a. Tests of identical products, with same testing procedure and the
same or similar test values, e.g. testing of distribution
transformers for medium voltage system or testing of linear and
non-linear resistor discs for arresters. The main requirement in
these tests are automatically connecting and changing the test
object, as well as measuring, recording, evaluating and indicating
of faults.
b. Statistical measurements, e.g. determination of spark gap
characteristics (Ub, Tb) in gases or liquids, whereby the test object should not be destroyed. The principal requirements are
automatic adjustment and control of the test installation as well
as measuring, recording and evaluating of the test results.
By means of some realized examples, the possibilities of step by
step automation are described.

The procedure for impulse voltage and impulse current tests


respectively, thereby can be divided into the following four steps:
- connecting and changing of the test object
- adjustment of the test installation accordmg to the test
requirements
- control of the test installation according to the required test
procedure
- measuring, recording and evaluating the test results.
These four steps will be treated separately since they are independant from each other. Depending on the application, the central
point of automation may be situated differently.
2.Connection of the test object

In case of automatic connection of the test object. the lay-out is


determined by the test object.

Fig. I Arrangement of series- and parallel resistors in the stages of a 2.4 MV impulse voltage generator with internal service platforms.
3

3. Change-over in the test installation

For the automation of an impulse test on a determined product,


e.g. one type of distribution transformer, resp. a given lightning
arrester disc, a change-over of the voltage generator, resp. current
generator is not necessary. The test voltage, resp. the test current
shape remain always the same and the test data do not differ substantially.
An automation on the high voltage side of a generator is
necessary, if different impulse shapes are required on one test
object. This occurs mostly for research and development work. The
automation possibility depends upon the design of the impulse
generator. If the impulse generator is equiped with parallel gaps, the
time to half-value can be effected simply by removing the gaps.
The choice of the right front resistors can be effected with resistor
magazines whereby four positions, which can be controlled for instance pneumatically, seem to be sufficient. Further, the automatic
change of various front resistors will be particularly simple and
clear when the stages are stacked one upon another. Fig. 1 shows an
example of front and tail resistors arrangement in the stages.
As modern high voltage impulse generators are practically
equiped with internal resistors only, a security earthing of all the
capacitors and an easy accessibility to all stages are essential. According to the required comfort, an access to the stages from outside (through service platforms) or from the inside may be achieved.
4.Control of the test process

According to the application of a test plant, the control of the test


process may be realized semi- or fully automatically. Thereby certain conditions must be fulfilled on the high voltage side.
The most important condition on the high voltage side of an
automatic impulse test plant is an absolute linearity between the
selected charging voltage and the output voltage regarding the impulse shape and amplitude of the output voltage. This can be realized most effectively by using only passive and voltage independent elements in the high voltage circuit.
Further a negligible influence of the non-linear element, the
coupling sphere-gap, upon the impulse shape and amplitude of the
output voltage is of high importance. This condition for instance
can be attained by using a simple sphere-gap, tired by natural overvoltages [2].
A second important condition is a sufficiently wide trigger range
of the impulse voltage generator [21. Thereby a trigger range of
more than 20% should be attained for all important impulse shapes
and connections. In this working range of approx. 20 %, scattering
should be less than k50 ns [3].

With a semi-automatic control only certain test procedures can


be realized. The pre-adjusted values, e.g. charging voltage and
charging time, are set manually. For identical test objects a certain
test procedure can be executed automatically with a program controller. If, on the contrary, various test procedures are required, a
computer is recommended for fully automatic control.
The interface between a computer or a program controller and
an impulse test plant is effected by a control module, called the
Trigatron. This trigatron is designed so that the necessary values
(charging voltage, charging time, trigger times) can also be set
manually.
4.1 Trigatron

It fulfils the following operations:


- automatic adjustment of the sphere-gap distance in relation to
the charging voltage
- controlled charging of the impulse generator regarding voltage
and time
- release of several timely successive processes dependent on the
charging conditions of the generator
- synchronized release to an internal or external ac voltage.
In fig. 2 the basic diagram of a control module is given. The
charging voltage and the charging time are adjusted manually or by
a program controller.
The correct distance of the sphere-gaps is automatically derived
from the charging voltage. The related characteristic line lies approx. 10 % to 15 % below the static tiring voltage. The charging of
the generator is provided with a follow-up control in order to secure
an equalized load of the mains supply.
Fig. 3 shows the charging curve, determined by the charging time
and charging voltage, being linear up to approx. 95 % of the pre-set
charging voltage. The remaining 5 % of the charging voltage have a
five times decreased slope in order to secure a fully charged multistage generator. Once the selected charging voltage is attained, a
warning signal is released (e.g. a horn) and, unaffected by the
charging time, the generator will tire after two seconds.

Charging
rectifier

Fig. 2 Basic diagram Trigatmn.

triggering

Fig.4 shows the primary and secondary currents with nonregulated and regulated charging control. The decrease of the
charging current when 95 % of the charging voltage are reached, is
clearly noticeable. In order to secure a correct commutation of the
solid state controller with assymetrical capacitive load, an ohmic
load is introduced at the primary side. With regulated charging, the
max. primary current is reduced by approx. 30%.
The timely adjustable release impulses can be selected digitally
from 0 to 999,9 ps with a resolution of 100 ns. At time zero a trigger
reference signal and two independent signals, which can be used for
e.g. firing the generator and triggering the oscilloscope, are released. A third signal, adjustable from 0 to 999,9 ps, can be released by the firing of the generator and is used in the test technique for
firing a chopping device. Thereby the chopping device operates
even in case of an erratic firing of the generator so that the test
object is never subjected to a too high full impulse voltage.
The scattering of this trigger system at the output of the control
module (fig. 5 a and fig. 5 b)can be neglected. The time scattering of
the output voltage on the test object is also smaller than t50 ns
(fig. 5 a and fig. 5b). This behaviour guarantees the reproductibility
of impulse voltage tests, this being absolutely essential, e. g. for tests
on transformers, in order to evaluate the test results.
Fig. 6 shows a control rack of an impulse test plant, containing
several modules, including the trigatron.

Time-to-breakdown
meter

-t

-t
a

Fig. 4 Trigatron: Primary (I,) and secondary current (IJ


a - controlled charging
b - non controlled charging.

low voltage
output Trigatron

high voltage
output impulse generator

Impulse
peak voltmeter

Printer

Metering

module

Program

controller
high voltage
5 consecutive impulses chopped at
3 different delay times

Trigatron

Control

module

Fig. 5a Delay of generator output impulses.

Power supply

Fig.6 Control rack of an impulse voltage generator.

Fig. 5b Scattering of output impulses for 10 consecutive impulses.

4.2 Program Controller

5.1 Time-to-breakdown meter

To control an impulse test plant with standardized test


procedures, several program controllers have been developed,
which can be connected to the trigatron (fig. 7). Thus a program
controller for the measurement of the characteristic values of air
spark-gaps has been developed, with the following setting
possibilities (fig. 6):
- minimum and maximum charging voltage
- number of shots per determined charging voltage
- voltage rise per step in kV or in %
- selection of polarity
- total number of shots
- program until total number of shots is reached.
Four different programs are included which can all be used to
determine the characteristics of spark-gaps:
Program 1: begins at a preselected voltage level with a given
number of shots; then the voltage is raised step by step until the first
flashover occurs.
Program 2: as 1, but the program stops only if the number of
flashovers at a given step voltage corresponds to the preset number
of shots per step.
Program 3: as 2, but after reaching the 100% flashover level, the
voltage is decreased step-wise until reaching the nreselected
minimum voltage level.
Program 4: as 2, but after reaching the 100 % flashover level the
voltage is decreased step by step, until no flashover occurs
anymore. Then further raised step-wise, a. s. o. till the total preset
number of shots is reached.
The total number of shots and the number of shots for each step
are counted. For each step, the number of breakdowns, the charging voltage and the output voltage can also be printed.

Fig. 6 shows a time-to-breakdown meter displaying the value


digitally with 0.1 IJ.S accuracy. This device can be triggered directly
from the measuring signal. It has an input impedance of 2 Mn//
30 pF.
The 110 MHz impulses are added in a 4-decade counter till the
Stop condition stops the oscillator. The value can be recorded by a
printer (fig.6) or further processed by a data nrocessing system.

5. Measuring, recording and evaluating the test results

The extent of measuring, recording and evaluating is related to


the kind of test. For the statistical examination of the breakdown
behaviour of air spark-gaps it is often sufficient to know the peak
voltage value and the time to breakdown.
These values may be obtained with relatively simple analog or
digital instruments

5.2 Impulse Peak Voltmeter

Fig. 8 shows the block diagram of an impulse voltmeter for the


measurement of peak values of unipolar impulses. The measuring
principle is based upon the storage of the impulse voltage by means
of diode and storage capacitance (fig.9).
The unit is equiped with a multiplier, which multiplies the input
value by the HV ratio of the divider. The indicated measuring value
therefore equals directly the HV value in kVs or MVs (fig. 6).
The digitally displayed value is available for recording by a
printer or for further processing by a data system.
Very fast storage processes even in case of high input impedance
( 1 MO// 50 pF) are reached by todays operational amplifier. If
the op amps are driven at the slew rate stated by the manufacturer,
according to definition overshoot will occur. Fast voltage variation
leads to overshoot and therefor erroneous measurements by the
storage circuit.
Therefore the rise time of the signal to be measured must be
decreased until a practically aperiodical entering into the storage
circuit is assured.
Todays op amps with max. slew rates of approx. 1000 V/ps
make it possible to build storage circuits with minimum rise times of
30 ns.
The afore mentioned practically aperiodical entering into the
store can be obtained by a damping resistor RD in the input divider
and by a frequency compensation in the storage stage (fig. 9). The
op amps perform a decoupling of the different stages, whereby the
stages can be considered as delay circuits of first order. The travelling time delay within the stages does not introduce an additional
measuring error (fig. 10).
The time constant T of the storage circuit, which for the realized
device amounts to approx. 40 ns, can be evaluated by means of a
rectangular impulse of known amplitude (fig. 10). Fig. 11 shows the
measuring accuracy of the realized impulse voltmeter (fig. 8) for
different impulse shapes of the high voltage test technique. The
measuring accuracy stated was evaluated by low voltage
measurements. If this measuring accuracy is checked-up in the high

Digital
nalog
COnYme

L
Display

_jl

U C tkV)

i
Fig. 7 Basic diagram Program Controller.

Fig. 8 Basic diagram impulse peak voltmerer.

voltage test laboratory. for curves 1. 2 and 3 the same results as


stated in fig. 11 are obtained, whereas curve 4 depends upon the
shape of the impulse at the peak of the chopped impulse voltage.
The low voltage measurements (curve 4) given in fig. 11, were obtained by a rounding time of approx. 40 ns. The so called rounding
effect, which one counters at HV tests depends on the measuring
arrangements, on the transient behaviour of the voltage divider and
on the characteristics of the chopping gap, reduces the error. That
would reach its maximum in case of ideal triangular impulses.
In fig. 12 this rounding error is shown for a chopping time of 0.5
us and 1 us. Performing practical HV tests, rounding times of approx. 50 to 80 ns will occur. The rounding time increases with increasing impulse voltage.

Fig. 9 Circuit of voltage peak store.

5.3 Storage and comparison of wave shapes

Transformer testing requires storage and comparison of two


current shapes during the total time. Today, measuring and
evaluating are mostly performed by means of oscilloscopes. In the
past years, digital storage systems have been developed using fast
analog/digital converters. In the future they will replace the oscilloscope in many applications.

Fig. 10 Unit step responce of the peak store.

5.3.1 Oscilloscopes

5.3.2 Digital storage systems

Since ordinary standard oscilloscopes are too sensitive against


electromagnetic interference (mainly in case of chopped lightning
impulses) special oscilloscopes for high voltage impulse test
technique are built. Furthermore the brightness of the beam, during
the recording of non-repetitive impulses in the ps-range, is often not
sufficient with standard oscilloscopes.
The specially developed impulse oscilloscopes feature low sensitivity against interference. This characteristic is achieved on the
one hand by carefully screening and a low-ohmic grounding within
the device and on the other hand by using an unsensitive cathode
ray tube. The tendency of development in standard oscilloscopes
goes toward the opposite direction. Because ofthe band width more
and more sensitive tubes are manufactured.
Recording and evaluating of impulse curves is notably facilitated
by a built-in automatic device. The camera shutter is opened prior
to each impulse and sweep time, attenuation factor, number of exposure and, if necessary, the peak value of the impulse voltmeter
are recorded automatically.

Transient recorders scan the measuring signal at a defined


frequency, convert the values point by point into digital form and
store them into a shifting register. The scanning frequency is mainly
limited by the AD-converters. Today, devices with a resolution of
100 points per ps and storage capacities of 2000 8-bit words
(corresponding to a resolution of 256 parts) are available.
Also continuous shifting of the information, called pretriggering
mode, is possible. Thereby, e.g. chopping of an impulse voltage can
be stored at high time resolution.
Front time, time to half-value, peak-value and time-to-breakdown of an impulse voltage can easily be evaluated by means of
these devices in connection with a computer.
A complete evaluation of the total impulse shape by a computer
together with a fault analysis is still problematical. The problem
has been solved partially for distribution transformers. For every
transformer design new fault criteria have to be found.
The block diagram of a realized fully automatized installation for
testing distribution transformers is shown in fig. 13.

Lightning impulse I.2150


chapped after x 115

ps

I
0

0.2

0.5

10

-x PS

Fig. 11 Impulse peak voltmeter. Intluence of impulse shape on the measuring


accuracy. Measured with low voltage.

Fig. 12 Impulse peak voltmeter. Influence of rounding time taon the measuring accuracy at time to chop t, of 0.5 us and 1 ps.
Low voltage measurement ( T = 40 ns).
7

Due to the sensitivity of these electronic systems, in any case a


measuring cabin is necessary. Fig. 14 shows a computer controlled
installation [4]. Recording of the impulse voltage is performed by
transient recorders.

II

6. Sensitivity of electronic devices in HV test application

The main difficulty in automating HV impulse testing lies in the


sensitivity of electronic components against electromagnetic interference. Using integrated circuits, the signal levels are limited to
approx. 15 V and the limiting frequency amounts to some MHz.
Compared with the conventional control technique, the signal levels
drop an order of magnitude whereas the frequencies rise by a factor
104.
Interfering voltages in measuring and control systems are caused
by:
- direct electromagnetic interferences
- disturbances bound to leads
- currents in cable sheaths and grounding equalizing currents.
These induced currents go to ground across the measuring instrument.
Changing the interference behaviour requires a reduction of the
influence of these currents. In order to reduce the harmful effect of
interference signals, the following steps can be taken:
Direct electromagnetic interference can be reduced by completely encapsulated devices which, if necessary, are placed in measuring
cabins. The order of interference is also depending on the position of
the devices (magnetic coupling).
Disturbancies bound to leads are reduced by fdters, suitable
isolation transformers, relays or optocouplers. Interference within
the device can be reduced by local screening and correct wiring.
Currents in cable sheaths and grounding equalizing currents can
be decreased by double screened cables, increasing of cable sheath
impedance by placing ferrite cores on the cable, by winding the
measuring cable on a soft-magnetic iron core or by grounding the
cable sheath at several points.
The effect of direct electromagnetic interference is some orders
of magnitude smaller than interferences bound to leads. Hereby not
only power supply and measuring signal input are relevant. Every
lead (control or signal line), which is brought into the device, must
be filtered carefully. The grounding system must be of low impedance.
Nowadays, conditioned by frequency and linearity, optoelectronic couplers are not yet suitable for transmission of analog
measuring values. Mostly, only the computer with its pure digitally
signal lines is separated galvanically by optocouplers, whereby the
influence of coupling capacities must be considered.
Wide measurements in the high voltage test field, including
chopped impulse voltages, have proved that control and measuring
system using TTL and CMOS components are working satisfactorily, if these factors are observed.

Fig. 13 Basic diagram of computer&d

impulse test system.

7. Examination of devices or systems with respect to interference

Evaluating electronic devices with respect to their suitability for


high voltage test application, depends mainly upon the sensibility of
these devices against disturbance signals occuring in high voltage
laboratories. Since direct examination in the high voltage test field
represents a very expensive solution, test facilities for electronic
laboratories have been developed.
This solution is possible due to the fact, that the most disturbing
influences in high voltage test fields are interferences bound to
lines. These very impulse currents bound to lines may be simulated
by impulse current generators. The impulse currents are fed to the
device at different points. For the simulation, mainly disturbing
currents, occuring during fast current alterations (breakdown of
gaps, tiring of impulse generators), have to be observed.

Literature
[ 11 K. Few, R. Niederhauser: Trends towards Automation of Impulse Voltage Tests.
International High Voltage Symposium Zibich (1975), p. 301 to 306.
[Zl K. Few-t Gedanken zur Stonspannungserzeugung
und Priiftechnik firdie Entwickhmg van
UHV-Systemen.
Bull. SEV 64 (1973) 15, p. 911 to 917.
131 A. Rodewald, K. Feser: The generation of lightning and switching impulse voltages in the
UHV region with an improved Marx Circuit.
IEEE Trans. on Pow. App. and Syst. 93 (1974) 1, p. 414 to 420.
141 P.R. Wiesendanger: Automatic Data Acquisition in High Voltage Laboratories. Problems
of electromagnetic interference.
1st Symposium on Electromagnetic Compatibility, Montreux May 20-22 (19751, p, 150 to
155.

HIGH VOLTAGE TEST SYSTEMS


LEHENMATTSTRASSE 3 5 3 P 0 BOX
CH-4028 BASEL SWITZERLAND

TELEPHONE 061/41 18 17
TELEX 6 2 4 6 9 ebb ch
TELEGRAMS MICARTA BASE1

PrInted I Swtrerland
5 1977 2 0 0 0
E l - 3 7 1 Edition

Potrebbero piacerti anche