GB2066967A

GB2066967A - Partial discharge detection in the presence of interference

Info

Publication number: GB2066967A
Application number: GB7943148A
Authority: GB
Original assignee: COUNTING INSTRUMENTS Ltd
Current assignee: COUNTING INSTRUMENTS Ltd
Priority date: 1979-12-14
Filing date: 1979-12-14
Publication date: 1981-07-15
Also published as: GB2066967B

Abstract

Partial discharge in a test component (CX) is detected by arranging the test component (CX) in series with a reference impedance (ZSB) across a source of high a.c. voltage (S). A second reference impedance (ZSA) and reference capacitance (CB) can also be connected in series across the source (S). A detector monitors transient pulses across the reference impedance or impedances to identify partial discharges. To enable partial discharges in test component (CX) to be detected in the presence of interference signal (RF1) a second signal (RF2) is injected into the detection circuit whose frequency is of the same order as the interference signal and which is not phase-locked to the interference signal. The two signals combine algebraically to produce a net resulting voltage whose magnitude is, at certain times, sufficiently small to allow a discharge pulse to be detected. The second signal can be injected either in the high voltage or preferably the low voltage side of the detection circuit. <IMAGE>

Description

SPECIFICATION Partial discharge detection in the presence of interference The present invention is concerned with the detection of partial discharges in the presence of continuous interference.

The term "partial discharge" refers to the phenomenon met in connection with high voltage equipment, such as capacitors, cables and transformers, where an electric discharge occurs which does not bridge the electrodes of the component in question, for example in the case of internal discharges in cavities in a dielectric, surface discharges along an insulator, and corona discharges around a sharp edge.

Methods are known for detecting such partial discharges. In the simplest method, the component under test is connected in series with a monitoring unit across the terminals of a high voltage a.c. source, the monitoring unit comprising a series impedance formed by the parallel combination of an inductor or a capacitor and a damping resistor. The monitoring unit provides an output potential which is dependent at any instant upon the supply potential and the characteristics of the test component in that, in the event of any partial discharge in the dielectric of the test component, e.g. in the case of a test capacitor, a transient pulse appears at the output of the monitoring unit which can be fed directly, or via an impedance matching transformer, to a display and/or recording device.

This basic techique is taken further in our earlier U.K. Patent No. 1425825 where a method is described of eliminating detection of any transient pulses which might be superimposed upon the supply voltage.

A problem in the known methods of detecting partial discharges is, however, that detection of such discharges is not possible, or is rendered very unreliable, in the presence of large magnitude, common mode interference signals, particularly when such interference signals are of a continuous nature, such as radio interference.

It is an objective of the present invention to provide a method and apparatus whereby partial discharges can be detected even in the presence of such large magnitude interference.

In accordance with the present invention, partial discharges are detected in the presence of an interference signal by injecting a second signal whose frequency is of the same order as the interference signal and which is not phase-locked to the interference signal, whereby the two signals combine algebraically to produce a net resulting voltage whose magnitude is, at certain times, sufficiently small to allow a discharge pulse to be detected.

Experiments have shown that it is best to inject a signal whose frequency is of the same order as the interference signal, but the exact value is not critical. The magnitude of the injected signal should ideally be the same as that of the interference signal but again in practice this is not critical and the technique works as long at the injected signal is between approximately 50 and 150% of the interference level.

If the magnitude of the interference signal is not known accurately, then the injected signal can be varied or modulated to ensure that at some time the two signals cancel.

Similarly, the frequency of the injected signal could be modulated although it is believed that in most circumstances this is unlikely to be necessary.

It has been found possible using this technique to detect partial discharge pulses in the presence of interference signals which were about ten times the height of the discharge pulses, i.e. an improvement in signal to noise ratio of about 20dB.

Obviously, when using this technique may discharge pulses are lost since they occur during times when the net resulting voltage of the two signals is high. However, the use of this technique enables one to detect discharges when it would otherwise be impossible to do so and information regarding the overall discharge characteristic can be gained by statistical methods from those discharges which are detected.

A measure of the frequency of the interfering signal can be gained using, for example, a spectrum analyser or a radio interference voltage meter which needs to be tuned manually through the appropriate frequency band, the amplitude of signals within that band being indicated on an analogue meter. Generally speaking, however, the interference does not have to be continuously monitored.

The invention is described further hereinafter, by way of example, with reference to the accompanying drawings, in which: Figures la to id diagrammatically illustrate the operating principle used in the present technique; and Figures 2a and 2b show two techniques for injecting the second signal.

Figs. 1 a to 1 c demonstrate how the present technique works. Fig. la shows an interference signal in the form of a first sinusoidal signal A,sin W,t. Fig. 1 b shows the injected signal in the form of a second sinusoidal signal A,sin(W2t + 8) of substantially the same frequency and magnitude as the first signal but which is not phase-locked in relation thereto. As a result of the algebraic summation or cross-modulation of these two signals there results a signal shown in Fig. 1c which exhibits periods Xl, X2 etc. where the resultant magnitude is sufficiently small for partial discharges occuring during these periods to be detected. In Fig. 1 c, the value of

is large. Fig. 2c shows the resultant signal envelope in the case when the value of

is small.

A number of methods of injecting the second signal are possible two of which are illustrated in Figs. 2a and 2b. Fig. 2a illustrates the case where the injection is effected from the high voltage end. RF1 is the common mode interference picked up by the circuit. A high voltage a.c. source is connected across the two input terminals 10, 1 2 between which is also connected a high grade reference component CB in series with an impedance ZSA and the component under test Cx in series with a second impedance ZSB. The terminal 1 2 is earthed.The high voltage injection signal RF2 is fed to the circuit via a mains frequency decoupling circuit CD LD. The impedance of the latter branch should be much higher than that of the other two test circuit arms, otherwise the partial (internal) discharge magnitude in the two measuring unit inputs may differ. This requirement can be satisfied by means of the series inductor LD which should be large enough to stop discharge frequencies getting through but small enough to shunt the mains voltage across the current source RF2. RD is included to overdamp the LD CD circuit.

The aforegoing arrangement has, however, the disadvantage that a small value highvoltage capacitor CD is required which increases the overall cost of the test circuit.

Fig. 2b shows a low voltage injection technique where Cx' and CB' are two low voltage capacitors. In this case the interference is modulated at the low voltage side of the test circuit before entering the measuring units, the effect being the same as before.

Claims

1. A method of detecting partial discharges in the presence of an interference signal comprising injecting a second signal whose frequency is of the same order as the interference signal and which is not phaselocked to the interference signal, whereby the two signals combine algebraically to produce a net resulting voltage whose magnitude is, at certain times, sufficiently small to allow a discharge pulse to be detected.

A Amethod as claimed in claim 1, in which the magnitude of the injection signal is substantially the same as that of the interference signal.

3. A method as claimed in claim 1 or 2, in which the magnitude of the injected signal is varied or modulated to ensure that at some time the two signals cancel.

4. A method as claimed in claim 1, 2 or 3 in which the frequency of the injected signal is varied or modulated.

5. Apparatus for enabling partial discharges to be detected in the presence of an interference signal comprising means forinjecting a second signal whose frequency is of the same order as the interference signal and which is not phase-locked to the interference signal, whereby the two signals combine algebraically to produce a net resulting voltage whose magnitude is, at certain times, sufficiently small to allow a discharge pulse to be detected.

6. Apparatus as claimed in claim 4 including a monitoring unit for connection in series with a test component across the terminals at a high voltage a.c. source, the monitoring unit comprising a series impedance formed by the parallel combination of an inductor or a capacitor and a damping resister and a display and/or recording device for displaying and/or recording any transient pulses appearing at the output of the monitoring unit.

7. An apparatus for detecting partial discharges in a test component wherein the test component is connected in series with a first reference impedance across a source of high frequency voltage and wherein a second reference impedance is connected in series with a reference capacitance, across the voltage source, the voltages appearing across the two impedances being monitored to detect partial discharges in the test component and wherein, in use, an unwanted interference signal is present on the applied voltage, the apparatus including a means for injecting a further signal onto the voltage applied to the two impedances whose frequency is of the same order as the interference signal whereby the two signals combine algebraically to produce a net resulting voltage whose magnitude is, at certain times, sufficiently small to allow a discharge pulse to be detected. -

8. Apparatus for enabling partial discharges to be detected, substantially as herein before described with reference to arld as illustrated in Fig. 2a of the accompanying drawings.

9. Apparatus for enabling partial discharge to be detected, substantially as hereinbefore described with reference to and as illustrated in Fig. 2b of the accompanying drawings.