CA2072183A1 - Process for measuring partial discharges - Google Patents
Process for measuring partial dischargesInfo
- Publication number
- CA2072183A1 CA2072183A1 CA002072183A CA2072183A CA2072183A1 CA 2072183 A1 CA2072183 A1 CA 2072183A1 CA 002072183 A CA002072183 A CA 002072183A CA 2072183 A CA2072183 A CA 2072183A CA 2072183 A1 CA2072183 A1 CA 2072183A1
- Authority
- CA
- Canada
- Prior art keywords
- spectrum
- sensed
- process according
- pulses
- digital
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
- G01R23/16—Spectrum analysis; Fourier analysis
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Measurement Of Resistance Or Impedance (AREA)
- Testing Relating To Insulation (AREA)
- Electrical Discharge Machining, Electrochemical Machining, And Combined Machining (AREA)
Abstract
ABSTRACT:
A process for the measuring of partial discharges on a plurality of homogeneous device-under-tests by processing sensed digital TE-pulses as follows: storing a reference frequency spectrum of TE-pulses in a digital computer;
transforming in said computer by a Fourier Transform at least one of said digital sensed TE-pulses from the time domain into the frequency range as a sensed frequency spectrum; scaling said sensed spectrum to the amplitude values of the reference spectrum; comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
Interference signals can be sensed and compensated mathematically or suppressed so that also a device-under-test TE-pulse with impressed interference signals can be used for the evaluation.
A process for the measuring of partial discharges on a plurality of homogeneous device-under-tests by processing sensed digital TE-pulses as follows: storing a reference frequency spectrum of TE-pulses in a digital computer;
transforming in said computer by a Fourier Transform at least one of said digital sensed TE-pulses from the time domain into the frequency range as a sensed frequency spectrum; scaling said sensed spectrum to the amplitude values of the reference spectrum; comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
Interference signals can be sensed and compensated mathematically or suppressed so that also a device-under-test TE-pulse with impressed interference signals can be used for the evaluation.
Description
A METHOD FOR MEASURING PARTIAL DISCHARGES OF A DEVICE-UNDER-TEST
The present invention relates to a process for measuring par1:ial discharges on a plurality of homogeneous device-under-tests .
The measuring of partial discharges of a device-under-tes1: includes applying a variable testing voltage and mea~;uring the partial discharge on a series measuring impedance. The measured discharge pulse, hereinafter called TE-pulse, is displayed, for example on an oscilloscope.
Measuring the TE-pulses is difficult because the TE-pulses are relatively small, the fed testing voltage is very high and interfering signals are superimposed by the network, the measuring cir~uit, the environment, or the like. The evaluation of a measured TE-pulse displayed, for example, by a storage oscilloscope, particularly by a digital memory oscilloscope, therefore requires a lot of experience in recognizing interference signals and is therefore relatively work-intensive and strenuous. A fast evaluation of a series production of components of a certain type or even an automating of the measuring and evaluating operations of partial discharges is therefore not possible or very imprecise. In order to increase the suppression of interferences, a measuring system may be used, but information concerning the course of the TE-pulse will be lost. Thus a differentiation between TE-pulses and interference pulses is no longer possible.
It is an object of the present invention to provide a proce~ for measuring and evaluating of TE-pulses fast and particularly automatically, and independent of occurring interference voltage with high precision.
This object is achieved by processing the sensed digit;al TE-pulses as follows: storing a reference frequency spectrum of TE-pulses in a digital computer; transforming in said computer by a Fourier Transform at least one of said digital sensed TE-pulses from the time domain into the frequency range as a sensed frequency spectrum; scaling said sensed spectrum to the amplitude values of the reference spectrum; comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum. This determines in a simple manner whether the TE-pulses are present with or without interference pulses. The present and determined interference signals can be compensated mathematically or suppressed so that also a device-under-test TE-pulse with impressed interference signals can be used for the evaluation.
As a result, series products of a uniform construction type can be tested rapidly and automatically by the TE-pulse measuring according to the invention.
For automatic comparative measuring, the TE-pulse of a device-under-test is coupled out in a wide-band manner and is subsequently amplified by a wide-band amplifier with an input voltage protection. By an anti-aliasing filter, the amplified signal is fed to an analog-to-digital converter with a digital memory. The stored data corresponding to the T~-pulse are fed to a computer and are transformed by this computer by the Fowrier Transform from the time domain into the frequency range, and the calculated digital values are stored. Before the automatic measuring operation, the mentioned process ~teps are carried out on several sample components of the same construction as the device-under-tests to be measured and, Prom the stored sample transformed data, a TE-pulse spectrum i8 aYeraged and ~tored that i5 typical of the component series, and is made available as the reference TE-pulse spectrum. Subsequently, a device-under-test is then applied to the testing voltage, and one or several TE~pulses are coupled out individually corresponding to the above-mentioned process steps and are finally transformed. Each measured TE-pulse is scaled to the amplitude values of the reference TE-pulse and is compared with it.
Other objects, advantages and novel features of the present invention will become apparent from the following detailed description of the invention when considered in conjunction with the accompanying drawings.
BRIEF DESCRIPTION OF T~E DRAWINGS
Figure 1 is a schematic of a measuring apparatus for carrying out the process according to the invention;
Figures 2 to 8 are diagrams of different TE-pulses.
BEST MODES FOR CARRYING OUT THE INVENTION
As illu~trated in Figure 1, a device-under-test 1 is connected to ground 3 by a measuring impedance 2 of, for example, 50 to 200 ohms, which, in the case of the embodiment is a measuring resistor.
The measuring impedance may also include an inductance or the parallel or series connection of an ohmic resistance with an inductance. The input 4 terminal of the device-under-test 1 is connected to the output 5 of a high-voltage generator 6, which can emit a testing pulse, for example, a high testing voltage, of a defined wave shape. The output 5 is ~lso connected to a voltage divider 7 including resistors 8 and 9 connected to ground 3. The voltage divider 7 divides the high testing voltage, for example, at a ratio of 1:1,000. The divider point 10 of the voltage divider 7 is connected by a lead through a lead-through 12 provided in a shield housing 11, by a buffer amplifier 13 to a voltmeter 14, particularly a digital voltmeter.
The voltage of the high-voltage generator 6 can be adjusted in steps or continuously by a control unit 15, for example, between 0 and 12 kV. The control unit 15 can be digitally controlled from exterior the housing 11 by a manual control device 16 and/or electronically by an interface control 17.
The output 5 of the high-voltage generator 6 is preferably shunted to ground 3 by a coupling capacitor 18.
The terminal 19 of the measuring resistor 2 is connected to a wide-band measuring amplifier 22 by line 20, for example, a coaxial cable having an impedance of approximately 50 ohms and, if necessary, by a high pass device 21, such as a high pass of the 6th order. The high pass device 21 is preferably designed to suppress a 50 hertz network frequency in the coupled-out TE-pulse which originates from the high-voltage generator 6. The cut-off frequency of the high pass device 21 amounts to approximately 1.0 to 1.5 KHz. If necessary, at least one additional input voltage divider 23 may be provided prior to the wide-band measuring amplifier 22. In the case of several input voltage dividers 23, the adjustment of different amplification factors is possible. Furthermore, the wide-band measuring amplifier 22 is protected against excess voltage, particularly against pulse-shaped excess voltage of short risP
times by an input protection device 23a shown as a protective diode.
The input stage for the wide-band measuring amplifier 22 is preferably a differential amplifier having a FET-input or a corresponding operations amplifier with a high common mode rejection .
A low-pass filter in the form of an anti-aliasing Pilter 24 is connected after the wide-band measuring amplifier 22.
For example, its upper cut-off frequency may be between 3 MHz and 10 MHz and for example at 5 MHz. The low-pass filter 24 is connected by a coaxial lead-though 25 to an input of an analog-to-digital converter with a digital memory, such as the digital memory oscilloscope 26. An interface bus 27 connects the oscilloscope with a computer 28, for example, a microcomputer with a timing frequency that is as high as possiblQ, for example, of 20 MHz to 33 MHz.
The interface bus 27 al~o connects the digital voltmeter 14 and the control unit 15, by the interface control 17, with the computer 28. The computer 28 contains, among others, a program for the transformation of a time domain data of the TE-pulse into a frequency range, for example, for the Fourier Transform and particularly for the Fast Fourier Transform. The data of the TE-pulse i8 from the analog-to-digital converter with the digital memory or from the digital memory oscilloscope 26. As a result, the obtained digitized pulse-time diagram can ~e computed into a frequency spectrum.
The generally known Fast Fourier Transform is adapted to the measuring-site-specific data formats in order to reduce the computing time. In addition, the computer 28 calculates the charge of a pulse, and the obtained values are stored for further processing. Furthermore, the computer 28 or the program contain routines for the recognition and elimination of interference signals, preferably by autocorrelation computing and crosscorrelation computing.
Using the measuring circuit of Figure 1, the sequence of the measuring process takes place essentially as follows:
The high-voltage generator 6 is controlled by the manual control device 16 or the computer 28. Then the high-voltage generator 6, on its output 5, emits to the device-under-test 1 a testing voltage of a defined wave form as a testing pulse with a voltage that increases from pulse to pulse. When partial discharges occur, a corresponding current flows to ~ ~ O ~ ~ g ~
ground 3 by the measuring resistor 2. At terminal 19 of the measuring resistor 2, a voltage drop occurs, which is the measured voltage UM, and is fed to the wide-band measuring amplifier 22 by the coaxial cable 20, the high pass device 21, the input voltage divider 23 and the input protection device 23. The amplified measured voltage UM'from the wide-band measuring amplifier is fed, by the anti-aliasing filter 24 and the lead-through 25, to the input of an analog-to-digital converter with a memory or to the input of the digital memory oscilloscope 26. Therein, the curve line or series of measured voltages UM' is digitized and stored. By the interface bus 27, the digital values are fed to the computer 2~ and are transformed into the frequency range by it by a discrete Fourier Transform or a Fast Fourier Transform. It is known that, for this purpose, the amounts of the discrete frequency are calculated. In the case of a scanning frequency of, for example, 20 MHz and a transmission of 2,048 measuring points, while ta~ing into account the direct-current parts and the symmetry of the Fast Fourier Transform FFT, the frequency spectrum of the pulse is represented as frequency amplitudes of 10 KHz to 10.23 MHz at a distance of 10 KHz. However, because the low~pass filter 24 is constructed as an anti-aliasing filter, only frequencies below its cut-off frequency, thus, for example, below 5 MHz, may be used for the signal evaluation. The calculated frequency spectrum can be sent to a video screen by the computer 28 for display.
For automated measuring operation, first several sample TE~pulses are recorded manually or automatically by the measuring apparatu~ of Figure 1 and the mean i8 taken for each amplitude of the frequency spectrum by a family. The obtained data are stored and are used as a reference TE-pulse spectrum for a device-under-test group, for example, for optocouplers, transformers or printed circuit boards, in each case, of a certain construction type. The reference TE-pulse spectrum is compared with the device-under-test TE-pulses occurring at a device-under-test.
As an alternative, synthetically generated pulses may also be used for the generating of the reference TE-pulse spectrum. For example, the synthetically generated pulses are fed as a mathematical function or in the form of a value table directly into the computer 28 or are applied to the testing structure, instead of the TE-pulses created in the device-under-test, by a pattern pulse generator 29, preferably when the testing voltage is switched off, switched in parallel to the device-under-test, as indicated in Figure 1 by interrupted lines.
Since the amplitudes of the individual frequencies of the frequency spectrum change with the amplitude of a TE-pulse, a scaling of the amplitude values of the frequency spectrum is required. For this purpose, a 100% amplitude value is calculated for each measured TE-pulse, and the amplitudes of all frequencies are applied to it.
For the formation of the 100% amplitude value, preferably only those frequencies are used which are characteristic of the reference TE-pulse spectrum. Thus, for example, starting from a rectangular pulse as an input signal and a (sin *x)/x-function as the frequency spectrum, the amplitudes offrequencies from 10 KHz to 100 KHz and from 3.01 to 3.1 MHz can be selected as a reference quantity.
The first zero point of the function (sin*x)/x for a measuring structure of the above-mentioned type is, for example, 2 MHz. A typical frequency spectrum of tested low-volt:age transformers is shown by the thick solid Line 29 ~ trated in Figure 2. Lines 30 and 31 above and below Line 29 indicate the selected tolerance range in which the frequency spectrum of a device-under-test TE-pulse may be situated. In the illustrated example, the tolerance is ~/-8%, relative to Line 29. For forming the mean value of a reference TE-pulse spectrum, five TE-pulses may, for example, be required. As a result of the selected narrow tolerance, the separation of TE-pulses and interferences pulses becomes possible. An automatic series testing of component~ which are of the same type as the device-under-test which was used for the reference TE-pulse spectrum formation can therefore be achieved in a ~imple and reliable manner.
Basically, the following interference signals may occur as interference pulses:
al Digital Ouantitazation Noise, triggered by the analog-to-digital conversion in the digital memory oscilloscope 26.
This has a particularly intensive effect in the case of constant values in the time domain, for example, as noise values about the zero line. A TE-pulse 32 of this type is illustrated in Figure 3. The influence of the quantitazation noise can be reduced by evaluating pulses by a window function. The Fast Fourier Transform of such a pulse 33 is indicated in Figure 4.
b~ Periodic Intex~erence Sianals, as illustrated, for example, in Figure 5. It shows an interfer~nce signal 34 of a ~requency of 1 MHz which is superimposed on the TE-pulse in the time representation. The interference signal 34.1 is clearly visible in the frequency spectrum of the TE-pulse 35 after the Fourier Transform in Figure 6. This interference can easily be eliminated by a correcting routine of the computer progra~. For this purpose, the pulses are examined, for example, with respect to pronounced peaks of the frequency spectrum which are larger than the tolerance range, and these are then adapted to the amounts of the adjacent spectral parts. Further~ore, if necessary, periodic interference signals may be recognized by a signal processing, preferably an autocorrelation calculation, and can be eliminated by a correcting routine, preferably of a digital filter.
c~ ~ide-band Interference Siqnals, a~ a rule, can hardly be avoided. However, by the comparison of the device-under-test TE-pulses with the reference TE-pulse spectrum, their recc~nition is possible in that, for example, by a program routine. A device-under-test TE-pulse will be evaluated as such only after possible interference signals according to a) and b) are eliminated and more than a certain percentage, for example, 75% to 95%, of the amounts of the spectral parts are within the tolerance range.
d) Noise: In the case of very low TE-charge intensities, a pronounced noise is superimposed on the useful signal. By a ~2~ ~3 computation of the pulses, preferably a crossrelation computation, with a low-noise, possibly synthetically generated sample pulse, the signal/noise ratio of the TE-pulses can be improved, if required.
e) Switchina Interference Sianals 36, for example, procluced by the switching-on of the motor for the adjustment of the kestlng voltage of the high-voltage generator 6, relay or contact controls, fluorescent lamps, thermal time delay switches, or the like, also show a time sequence, as indicated in Figure 7, which deviates significantly from the reference TE-pulse spectrum. Correspondingly, the frequency spectrum illustrated in Figure 8, thus after the Fourier Transform, can be recognized to clearly differ from the reference spectrum.
An interference pulse 36.1 of this type is therefore also reliably recognized and thus the measured pulse 37 of Figure 8 is not evaluated as a device-under-test TE-pulse.
The automatic measuring of device-under-tests starts with a low testing voltage which is increased per second, for example by 100 V. This leads, for example, to a measuring resolution of 50 V, and, in the case of a maximal testing voltage of 12 kV, the testing time is two minutes. When after the interference suppression, adaptation and comparison with the reference TE-pulse spectrum, a device-under-test TE-pulse is recognized as such, its charge can still be determined, for example, by the integration of scanned peak values, and these peak values as well as the pertaining operating voltage can be stored. Since, in the case of routine check tests, the testing voltage is often preset, the automatic TE-measurement may be limited to the recognition and ~toring of the operating voltage of the evaluated device-under-test TE-pulses.
A6 indicated above, the automatic measuring may be carried out very rapidly by the Fast Fourier Transform, which i5 known per se. A further shortening of the measuring time is possible by the corresponding selection of digital units with a high timing frequency or by specialized structural elements, such as digital signal processors.
It may be expedient to use a coaxial tube resistance or preferably a pyramid resistance as a measuring impedance 2.
It may also be advantageous to construct the measuring impedance as a metal film resistor.
Although the variable testing voltage is preferably in the lower high-voltage range of up to approximately 12 kV, the present invention is not limited to this testing voltage range. For achieving the object according to the invention, testing voltages of significantly higher voltage levels may be used with the same or comparable advantages.
Although the present invention has been described and illustrated in detail, it is to be clearly understood that the same is by way of illustration and example only, and is not to be taken by way of limitation. The spirit and scope of the present invention are to be limited only by the terms of the appended claims.
The present invention relates to a process for measuring par1:ial discharges on a plurality of homogeneous device-under-tests .
The measuring of partial discharges of a device-under-tes1: includes applying a variable testing voltage and mea~;uring the partial discharge on a series measuring impedance. The measured discharge pulse, hereinafter called TE-pulse, is displayed, for example on an oscilloscope.
Measuring the TE-pulses is difficult because the TE-pulses are relatively small, the fed testing voltage is very high and interfering signals are superimposed by the network, the measuring cir~uit, the environment, or the like. The evaluation of a measured TE-pulse displayed, for example, by a storage oscilloscope, particularly by a digital memory oscilloscope, therefore requires a lot of experience in recognizing interference signals and is therefore relatively work-intensive and strenuous. A fast evaluation of a series production of components of a certain type or even an automating of the measuring and evaluating operations of partial discharges is therefore not possible or very imprecise. In order to increase the suppression of interferences, a measuring system may be used, but information concerning the course of the TE-pulse will be lost. Thus a differentiation between TE-pulses and interference pulses is no longer possible.
It is an object of the present invention to provide a proce~ for measuring and evaluating of TE-pulses fast and particularly automatically, and independent of occurring interference voltage with high precision.
This object is achieved by processing the sensed digit;al TE-pulses as follows: storing a reference frequency spectrum of TE-pulses in a digital computer; transforming in said computer by a Fourier Transform at least one of said digital sensed TE-pulses from the time domain into the frequency range as a sensed frequency spectrum; scaling said sensed spectrum to the amplitude values of the reference spectrum; comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum. This determines in a simple manner whether the TE-pulses are present with or without interference pulses. The present and determined interference signals can be compensated mathematically or suppressed so that also a device-under-test TE-pulse with impressed interference signals can be used for the evaluation.
As a result, series products of a uniform construction type can be tested rapidly and automatically by the TE-pulse measuring according to the invention.
For automatic comparative measuring, the TE-pulse of a device-under-test is coupled out in a wide-band manner and is subsequently amplified by a wide-band amplifier with an input voltage protection. By an anti-aliasing filter, the amplified signal is fed to an analog-to-digital converter with a digital memory. The stored data corresponding to the T~-pulse are fed to a computer and are transformed by this computer by the Fowrier Transform from the time domain into the frequency range, and the calculated digital values are stored. Before the automatic measuring operation, the mentioned process ~teps are carried out on several sample components of the same construction as the device-under-tests to be measured and, Prom the stored sample transformed data, a TE-pulse spectrum i8 aYeraged and ~tored that i5 typical of the component series, and is made available as the reference TE-pulse spectrum. Subsequently, a device-under-test is then applied to the testing voltage, and one or several TE~pulses are coupled out individually corresponding to the above-mentioned process steps and are finally transformed. Each measured TE-pulse is scaled to the amplitude values of the reference TE-pulse and is compared with it.
Other objects, advantages and novel features of the present invention will become apparent from the following detailed description of the invention when considered in conjunction with the accompanying drawings.
BRIEF DESCRIPTION OF T~E DRAWINGS
Figure 1 is a schematic of a measuring apparatus for carrying out the process according to the invention;
Figures 2 to 8 are diagrams of different TE-pulses.
BEST MODES FOR CARRYING OUT THE INVENTION
As illu~trated in Figure 1, a device-under-test 1 is connected to ground 3 by a measuring impedance 2 of, for example, 50 to 200 ohms, which, in the case of the embodiment is a measuring resistor.
The measuring impedance may also include an inductance or the parallel or series connection of an ohmic resistance with an inductance. The input 4 terminal of the device-under-test 1 is connected to the output 5 of a high-voltage generator 6, which can emit a testing pulse, for example, a high testing voltage, of a defined wave shape. The output 5 is ~lso connected to a voltage divider 7 including resistors 8 and 9 connected to ground 3. The voltage divider 7 divides the high testing voltage, for example, at a ratio of 1:1,000. The divider point 10 of the voltage divider 7 is connected by a lead through a lead-through 12 provided in a shield housing 11, by a buffer amplifier 13 to a voltmeter 14, particularly a digital voltmeter.
The voltage of the high-voltage generator 6 can be adjusted in steps or continuously by a control unit 15, for example, between 0 and 12 kV. The control unit 15 can be digitally controlled from exterior the housing 11 by a manual control device 16 and/or electronically by an interface control 17.
The output 5 of the high-voltage generator 6 is preferably shunted to ground 3 by a coupling capacitor 18.
The terminal 19 of the measuring resistor 2 is connected to a wide-band measuring amplifier 22 by line 20, for example, a coaxial cable having an impedance of approximately 50 ohms and, if necessary, by a high pass device 21, such as a high pass of the 6th order. The high pass device 21 is preferably designed to suppress a 50 hertz network frequency in the coupled-out TE-pulse which originates from the high-voltage generator 6. The cut-off frequency of the high pass device 21 amounts to approximately 1.0 to 1.5 KHz. If necessary, at least one additional input voltage divider 23 may be provided prior to the wide-band measuring amplifier 22. In the case of several input voltage dividers 23, the adjustment of different amplification factors is possible. Furthermore, the wide-band measuring amplifier 22 is protected against excess voltage, particularly against pulse-shaped excess voltage of short risP
times by an input protection device 23a shown as a protective diode.
The input stage for the wide-band measuring amplifier 22 is preferably a differential amplifier having a FET-input or a corresponding operations amplifier with a high common mode rejection .
A low-pass filter in the form of an anti-aliasing Pilter 24 is connected after the wide-band measuring amplifier 22.
For example, its upper cut-off frequency may be between 3 MHz and 10 MHz and for example at 5 MHz. The low-pass filter 24 is connected by a coaxial lead-though 25 to an input of an analog-to-digital converter with a digital memory, such as the digital memory oscilloscope 26. An interface bus 27 connects the oscilloscope with a computer 28, for example, a microcomputer with a timing frequency that is as high as possiblQ, for example, of 20 MHz to 33 MHz.
The interface bus 27 al~o connects the digital voltmeter 14 and the control unit 15, by the interface control 17, with the computer 28. The computer 28 contains, among others, a program for the transformation of a time domain data of the TE-pulse into a frequency range, for example, for the Fourier Transform and particularly for the Fast Fourier Transform. The data of the TE-pulse i8 from the analog-to-digital converter with the digital memory or from the digital memory oscilloscope 26. As a result, the obtained digitized pulse-time diagram can ~e computed into a frequency spectrum.
The generally known Fast Fourier Transform is adapted to the measuring-site-specific data formats in order to reduce the computing time. In addition, the computer 28 calculates the charge of a pulse, and the obtained values are stored for further processing. Furthermore, the computer 28 or the program contain routines for the recognition and elimination of interference signals, preferably by autocorrelation computing and crosscorrelation computing.
Using the measuring circuit of Figure 1, the sequence of the measuring process takes place essentially as follows:
The high-voltage generator 6 is controlled by the manual control device 16 or the computer 28. Then the high-voltage generator 6, on its output 5, emits to the device-under-test 1 a testing voltage of a defined wave form as a testing pulse with a voltage that increases from pulse to pulse. When partial discharges occur, a corresponding current flows to ~ ~ O ~ ~ g ~
ground 3 by the measuring resistor 2. At terminal 19 of the measuring resistor 2, a voltage drop occurs, which is the measured voltage UM, and is fed to the wide-band measuring amplifier 22 by the coaxial cable 20, the high pass device 21, the input voltage divider 23 and the input protection device 23. The amplified measured voltage UM'from the wide-band measuring amplifier is fed, by the anti-aliasing filter 24 and the lead-through 25, to the input of an analog-to-digital converter with a memory or to the input of the digital memory oscilloscope 26. Therein, the curve line or series of measured voltages UM' is digitized and stored. By the interface bus 27, the digital values are fed to the computer 2~ and are transformed into the frequency range by it by a discrete Fourier Transform or a Fast Fourier Transform. It is known that, for this purpose, the amounts of the discrete frequency are calculated. In the case of a scanning frequency of, for example, 20 MHz and a transmission of 2,048 measuring points, while ta~ing into account the direct-current parts and the symmetry of the Fast Fourier Transform FFT, the frequency spectrum of the pulse is represented as frequency amplitudes of 10 KHz to 10.23 MHz at a distance of 10 KHz. However, because the low~pass filter 24 is constructed as an anti-aliasing filter, only frequencies below its cut-off frequency, thus, for example, below 5 MHz, may be used for the signal evaluation. The calculated frequency spectrum can be sent to a video screen by the computer 28 for display.
For automated measuring operation, first several sample TE~pulses are recorded manually or automatically by the measuring apparatu~ of Figure 1 and the mean i8 taken for each amplitude of the frequency spectrum by a family. The obtained data are stored and are used as a reference TE-pulse spectrum for a device-under-test group, for example, for optocouplers, transformers or printed circuit boards, in each case, of a certain construction type. The reference TE-pulse spectrum is compared with the device-under-test TE-pulses occurring at a device-under-test.
As an alternative, synthetically generated pulses may also be used for the generating of the reference TE-pulse spectrum. For example, the synthetically generated pulses are fed as a mathematical function or in the form of a value table directly into the computer 28 or are applied to the testing structure, instead of the TE-pulses created in the device-under-test, by a pattern pulse generator 29, preferably when the testing voltage is switched off, switched in parallel to the device-under-test, as indicated in Figure 1 by interrupted lines.
Since the amplitudes of the individual frequencies of the frequency spectrum change with the amplitude of a TE-pulse, a scaling of the amplitude values of the frequency spectrum is required. For this purpose, a 100% amplitude value is calculated for each measured TE-pulse, and the amplitudes of all frequencies are applied to it.
For the formation of the 100% amplitude value, preferably only those frequencies are used which are characteristic of the reference TE-pulse spectrum. Thus, for example, starting from a rectangular pulse as an input signal and a (sin *x)/x-function as the frequency spectrum, the amplitudes offrequencies from 10 KHz to 100 KHz and from 3.01 to 3.1 MHz can be selected as a reference quantity.
The first zero point of the function (sin*x)/x for a measuring structure of the above-mentioned type is, for example, 2 MHz. A typical frequency spectrum of tested low-volt:age transformers is shown by the thick solid Line 29 ~ trated in Figure 2. Lines 30 and 31 above and below Line 29 indicate the selected tolerance range in which the frequency spectrum of a device-under-test TE-pulse may be situated. In the illustrated example, the tolerance is ~/-8%, relative to Line 29. For forming the mean value of a reference TE-pulse spectrum, five TE-pulses may, for example, be required. As a result of the selected narrow tolerance, the separation of TE-pulses and interferences pulses becomes possible. An automatic series testing of component~ which are of the same type as the device-under-test which was used for the reference TE-pulse spectrum formation can therefore be achieved in a ~imple and reliable manner.
Basically, the following interference signals may occur as interference pulses:
al Digital Ouantitazation Noise, triggered by the analog-to-digital conversion in the digital memory oscilloscope 26.
This has a particularly intensive effect in the case of constant values in the time domain, for example, as noise values about the zero line. A TE-pulse 32 of this type is illustrated in Figure 3. The influence of the quantitazation noise can be reduced by evaluating pulses by a window function. The Fast Fourier Transform of such a pulse 33 is indicated in Figure 4.
b~ Periodic Intex~erence Sianals, as illustrated, for example, in Figure 5. It shows an interfer~nce signal 34 of a ~requency of 1 MHz which is superimposed on the TE-pulse in the time representation. The interference signal 34.1 is clearly visible in the frequency spectrum of the TE-pulse 35 after the Fourier Transform in Figure 6. This interference can easily be eliminated by a correcting routine of the computer progra~. For this purpose, the pulses are examined, for example, with respect to pronounced peaks of the frequency spectrum which are larger than the tolerance range, and these are then adapted to the amounts of the adjacent spectral parts. Further~ore, if necessary, periodic interference signals may be recognized by a signal processing, preferably an autocorrelation calculation, and can be eliminated by a correcting routine, preferably of a digital filter.
c~ ~ide-band Interference Siqnals, a~ a rule, can hardly be avoided. However, by the comparison of the device-under-test TE-pulses with the reference TE-pulse spectrum, their recc~nition is possible in that, for example, by a program routine. A device-under-test TE-pulse will be evaluated as such only after possible interference signals according to a) and b) are eliminated and more than a certain percentage, for example, 75% to 95%, of the amounts of the spectral parts are within the tolerance range.
d) Noise: In the case of very low TE-charge intensities, a pronounced noise is superimposed on the useful signal. By a ~2~ ~3 computation of the pulses, preferably a crossrelation computation, with a low-noise, possibly synthetically generated sample pulse, the signal/noise ratio of the TE-pulses can be improved, if required.
e) Switchina Interference Sianals 36, for example, procluced by the switching-on of the motor for the adjustment of the kestlng voltage of the high-voltage generator 6, relay or contact controls, fluorescent lamps, thermal time delay switches, or the like, also show a time sequence, as indicated in Figure 7, which deviates significantly from the reference TE-pulse spectrum. Correspondingly, the frequency spectrum illustrated in Figure 8, thus after the Fourier Transform, can be recognized to clearly differ from the reference spectrum.
An interference pulse 36.1 of this type is therefore also reliably recognized and thus the measured pulse 37 of Figure 8 is not evaluated as a device-under-test TE-pulse.
The automatic measuring of device-under-tests starts with a low testing voltage which is increased per second, for example by 100 V. This leads, for example, to a measuring resolution of 50 V, and, in the case of a maximal testing voltage of 12 kV, the testing time is two minutes. When after the interference suppression, adaptation and comparison with the reference TE-pulse spectrum, a device-under-test TE-pulse is recognized as such, its charge can still be determined, for example, by the integration of scanned peak values, and these peak values as well as the pertaining operating voltage can be stored. Since, in the case of routine check tests, the testing voltage is often preset, the automatic TE-measurement may be limited to the recognition and ~toring of the operating voltage of the evaluated device-under-test TE-pulses.
A6 indicated above, the automatic measuring may be carried out very rapidly by the Fast Fourier Transform, which i5 known per se. A further shortening of the measuring time is possible by the corresponding selection of digital units with a high timing frequency or by specialized structural elements, such as digital signal processors.
It may be expedient to use a coaxial tube resistance or preferably a pyramid resistance as a measuring impedance 2.
It may also be advantageous to construct the measuring impedance as a metal film resistor.
Although the variable testing voltage is preferably in the lower high-voltage range of up to approximately 12 kV, the present invention is not limited to this testing voltage range. For achieving the object according to the invention, testing voltages of significantly higher voltage levels may be used with the same or comparable advantages.
Although the present invention has been described and illustrated in detail, it is to be clearly understood that the same is by way of illustration and example only, and is not to be taken by way of limitation. The spirit and scope of the present invention are to be limited only by the terms of the appended claims.
Claims (23)
1. In a process for the measuring of partial discharges of a plurality of homogeneous device-under-tests by feeding a variable testing voltage of a defined wave form to a respective device-under-test connected in series with a measuring impedance of a testing apparatus, feeding a partial discharge pulse sensed at the measuring impedance by a wide-band amplifier to an analog-to-digital converter which converts and stores digital sensed partial discharge pulses, the improvement comprising:
storing a reference frequency spectrum of partial discharge pulses in a digital computer;
transforming in said computer by a Fourier Transform at least one of said digital sensed partial discharge pulses from the time domain into the frequency range as a sensed frequency spectrum;
scaling said sensed spectrum to the amplitude values of the reference spectrum;
subsequently comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
storing a reference frequency spectrum of partial discharge pulses in a digital computer;
transforming in said computer by a Fourier Transform at least one of said digital sensed partial discharge pulses from the time domain into the frequency range as a sensed frequency spectrum;
scaling said sensed spectrum to the amplitude values of the reference spectrum;
subsequently comparing the scaled spectrum with the reference spectrum; and determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
2. A process according to Claim 1, wherein the scaling includes:
determining a 100% amplitude value for said reference spectrum;
determining a 100% amplitude value for said sensed spectrum;
calculating a scaling factor; and scaling the amplitude values for all frequencies of the sensed spectrum.
determining a 100% amplitude value for said reference spectrum;
determining a 100% amplitude value for said sensed spectrum;
calculating a scaling factor; and scaling the amplitude values for all frequencies of the sensed spectrum.
3. A process according to Claim 2, wherein for determining said the 100% amplitude value, a limited number of characteristic frequency ranges of the spectrum is used.
4. A process according to Claim 3, wherein the frequency range of approximately 10 KHz to 100 KHz and the frequency range of approximately 3 MHz to 3.1 MHz are used.
5. A process according to Claim 1, including synthetically generating, in said computer using a formula, said reference spectrum to be stored.
6. A process according to Claim 1, including synthetically generating said reference spectrum to be stored as a table of values.
7. A process according to Claim 1, including feeding pulses of a defined curve shape generated by a sample pulse generator as partial discharge pulses into said testing apparatus.
8. A process according to Claim 7, including connecting said sample pulse generator in parallel to the device-under-test.
9. A process according to Claim 1, including synthetically generating, in said computer, a simulated reference spectrum to be stored; and wherein said scaling includes using at least one characteristic frequency or at least one characteristic frequency range of the simulated reference spectrum for the determination of 100% reference amplitude value for scaling.
10. A process according to Claim 1, including eliminating pronounced peak values outside tolerance limits from the sensed spectrum mathematically by a correction routine, before the comparing and before the scaling.
11. A process according to Claim 10, wherein the scaling includes:
determining a 100% amplitude value for said reference spectrum;
determining a 100% amplitude value for said sensed spectrum after elimination of said pronounced peaks;
calculating a scaling factor; and scaling the amplitude values for all frequencies of the sensed spectrum.
determining a 100% amplitude value for said reference spectrum;
determining a 100% amplitude value for said sensed spectrum after elimination of said pronounced peaks;
calculating a scaling factor; and scaling the amplitude values for all frequencies of the sensed spectrum.
12. A process according to Claim 1, including eliminating periodic interference signals from said digital sensed pulses mathematically by an autocorrelation routine, before transforming.
13. A process according to Claim 1, including improving the signal-to-noise ratio of the digital sensed pulses mathematically by a crosscorrelation routine, before transforming.
14. A process according to Claim 1, including reducing the influence of digital quantitazation noise on the sensed spectrum by evaluating the digital sensed pulses by a window function.
15. A process according to Claim 1 including smoothing pronounced peaks of the sensed spectrum which exceed a tolerance range to adjacent spectral portions.
16. A process according to Claim 1 , including determining which scaled spectrum have a predetermined percentage of the amounts of the spectral portions within a tolerance range.
17. A process according to Claim 1, wherein said transforming uses a Fast Fourier Transform.
18. A process according to Claim 1, wherein storing a reference frequency spectrum includes:
performing said feeding and converting on several sample components to store sample digital sensed pulses for each sample;
transforming each of said sample digital sensed pulses into a sample frequency spectrum; and averaging said sample spectrum to produce said reference frequency spectrum.
performing said feeding and converting on several sample components to store sample digital sensed pulses for each sample;
transforming each of said sample digital sensed pulses into a sample frequency spectrum; and averaging said sample spectrum to produce said reference frequency spectrum.
19. An apparatus for the measuring of partial discharges of 21 plurality of homogeneous device-under-tests comprising:
means for applying a variable testing voltage of a defined wave form to a respective device-under-test connected in series with a measuring impedance;
means for feeding a partial discharge pulse sensed at the measuring impedance by a wide-band amplifier to an analog-to-digital converter which converts and stores digital sensed partial discharge pulses;
means for storing a reference frequency spectrum of partial discharge pulses in a digital computer;
means for transforming in said computer by a Fourier Transform at least one of said digital sensed partial discharge pulses from the time domain into the frequency range as a sensed frequency spectrum;
means for scaling said sensed spectrum to the amplitude values of the reference spectrum;
means for comparing the scaled spectrum with the reference spectrum; and means for determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
means for applying a variable testing voltage of a defined wave form to a respective device-under-test connected in series with a measuring impedance;
means for feeding a partial discharge pulse sensed at the measuring impedance by a wide-band amplifier to an analog-to-digital converter which converts and stores digital sensed partial discharge pulses;
means for storing a reference frequency spectrum of partial discharge pulses in a digital computer;
means for transforming in said computer by a Fourier Transform at least one of said digital sensed partial discharge pulses from the time domain into the frequency range as a sensed frequency spectrum;
means for scaling said sensed spectrum to the amplitude values of the reference spectrum;
means for comparing the scaled spectrum with the reference spectrum; and means for determining scaled spectrum which are within a defined tolerance range of the reference spectrum.
20. An apparatus according to Claim 19, wherein said measuring impedance includes a coaxial tube resistor.
21. An apparatus according to Claim 19, wherein said measuring impedance includes a coaxial pyramid resistor.
22. An apparatus according to Claim 19, wherein said measuring impedance includes a metal film resistor.
23. An apparatus according to Claim 19, wherein said wide-band amplifier includes a differential amplifier input stage and an input protection device.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE4120821A DE4120821A1 (en) | 1991-06-24 | 1991-06-24 | METHOD FOR MEASURING PARTIAL DISCHARGES |
| DEP4120821.8 | 1991-06-24 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2072183A1 true CA2072183A1 (en) | 1992-12-25 |
Family
ID=6434626
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002072183A Abandoned CA2072183A1 (en) | 1991-06-24 | 1992-06-23 | Process for measuring partial discharges |
Country Status (5)
| Country | Link |
|---|---|
| EP (1) | EP0520193B1 (en) |
| JP (1) | JPH05240902A (en) |
| AT (1) | ATE151537T1 (en) |
| CA (1) | CA2072183A1 (en) |
| DE (2) | DE4120821A1 (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE29518286U1 (en) * | 1995-11-17 | 1996-01-18 | Siemens Ag | Device for locating partial discharges in dynamoelectric high-voltage machines and / or high-voltage systems |
| DE19548466A1 (en) * | 1995-12-22 | 1997-06-26 | Abb Research Ltd | Sensor for decoupling partial discharge pulses from a high-voltage electrical system |
| US6032524A (en) * | 1997-12-03 | 2000-03-07 | Raytheon Aircraft Company | Method and apparatus for detecting p-static interference source |
| JP2009047831A (en) * | 2007-08-17 | 2009-03-05 | Toshiba Corp | Feature quantity extracting device, program and feature quantity extraction method |
| ES2370613T3 (en) * | 2008-04-14 | 2011-12-20 | Abb Research Ltd. | DETERMINATION OF THE INSULATED CAPACITY DEGRADED IN AN EXPECTED ISOLATION BETWEEN TWO OBJECTS OF AN INDUCTIVE OPERATING ELEMENT. |
| US9482710B2 (en) * | 2013-09-11 | 2016-11-01 | GM Global Technology Operations LLC | Inspection system for evaluating electrical parts for unwanted partial discharge |
| KR101787901B1 (en) * | 2016-06-14 | 2017-11-15 | 엘에스산전 주식회사 | Power equipment diagnostic apparatus |
| CN111142001B (en) * | 2020-01-10 | 2022-04-22 | 三峡大学 | Transformer multi-source partial discharge mode identification method based on parallel characteristic domain |
| EP3961227A1 (en) | 2020-08-31 | 2022-03-02 | General Electric Company | Online and offline partial discharge detection for electrical drive systems |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE2856354C2 (en) * | 1978-12-27 | 1986-06-26 | Calor-Emag Elektrizitäts-Aktiengesellschaft, 4030 Ratingen | Device for testing metal-enclosed high-voltage systems for partial discharges |
| DE3148735A1 (en) * | 1981-12-09 | 1986-10-09 | Fried. Krupp Gmbh, 4300 Essen | Method and device for frequency analysis |
| DE3441938A1 (en) * | 1984-11-16 | 1986-05-22 | Preh, Industrieausrüstungen GmbH, 8740 Bad Neustadt | DEVICE FOR COMPARING VALUE PAIRS TO BE MEASURED WITH A TARGET CHARACTERISTIC |
| US4744041A (en) * | 1985-03-04 | 1988-05-10 | International Business Machines Corporation | Method for testing DC motors |
| SU1337837A1 (en) * | 1985-12-27 | 1987-09-15 | Ленинградский Политехнический Институт Им.М.И.Калинина | Method of nondestructive check of production quality of h.v.insulation |
| DD253333A1 (en) * | 1986-10-13 | 1988-01-13 | Robotron Messelekt | PART DISCHARGE MEASURING ASSEMBLY WITH SUPPRESSION OF SUPPRESSED RF INTERFERENCE VOLTAGES |
| DD276217A3 (en) * | 1987-06-09 | 1990-02-21 | Robotron Elektronik | Method and device for non-destructive dielectric strength testing |
| DE3733040A1 (en) * | 1987-09-30 | 1989-04-13 | Siemens Ag | Device and method for testing electrical conduction networks, especially printed circuit boards |
| US4918382A (en) * | 1989-03-20 | 1990-04-17 | Tektronix, Inc. | Method for distinguishing between real and spurious responses in a spectrum analyzer |
-
1991
- 1991-06-24 DE DE4120821A patent/DE4120821A1/en not_active Withdrawn
-
1992
- 1992-05-24 EP EP92108749A patent/EP0520193B1/en not_active Expired - Lifetime
- 1992-05-24 DE DE59208310T patent/DE59208310D1/en not_active Expired - Fee Related
- 1992-05-24 AT AT92108749T patent/ATE151537T1/en not_active IP Right Cessation
- 1992-06-23 CA CA002072183A patent/CA2072183A1/en not_active Abandoned
- 1992-06-24 JP JP4165950A patent/JPH05240902A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| DE59208310D1 (en) | 1997-05-15 |
| DE4120821A1 (en) | 1993-01-07 |
| ATE151537T1 (en) | 1997-04-15 |
| EP0520193A1 (en) | 1992-12-30 |
| JPH05240902A (en) | 1993-09-21 |
| EP0520193B1 (en) | 1997-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6195614B1 (en) | Method of characterizing events in acquired waveform data from a metallic transmission cable | |
| US5633801A (en) | Pulse-based impedance measurement instrument | |
| US5698984A (en) | Adaptive digital filter for improved measurement accuracy in an electronic instrument | |
| US8760171B2 (en) | Device and method for determining partial discharges at an electrical component | |
| JPH10319056A (en) | Measuring front-end for measuring device, and method for providing plural measuring parameters from signal voltage | |
| US4820973A (en) | Impedance measuring system | |
| CA2072183A1 (en) | Process for measuring partial discharges | |
| US4568878A (en) | Spectrum analyzers | |
| EP3658925A1 (en) | Method and testing device for measuring partial discharge pulses of a shielded cable | |
| WO2015154587A1 (en) | Measurement apparatus and method for ripple current | |
| WO2000073747A1 (en) | Method for the noise-free evaluation of radar signals | |
| US5528134A (en) | AC power analyzer | |
| US20190025344A1 (en) | Monitoring device under test waveform on signal generator | |
| DE102006005595B4 (en) | Apparatus and method for measuring spurious emissions in real time | |
| US4126037A (en) | Compression test using battery voltage waveform during cranking | |
| DE10308279B4 (en) | Interference cancellation in an antenna test | |
| US5036477A (en) | Method for the interference suppression in ultrasonic distance measurements | |
| KR101820040B1 (en) | Apparatus and method for measuring a partial discharging pulse | |
| JP2000097982A (en) | Coil testing and evaluating device | |
| EP0430256A2 (en) | Method and equipment for cablibrating output levels of waveform analyzing apparatus | |
| WO1992021984A1 (en) | Apparatus for measuring average value of impulsive signal | |
| US3222598A (en) | Electroencephalograph wave analyzer | |
| Slomovitz et al. | Review of digital techniques applied to high-voltage transformer impulse-test. | |
| KR101832190B1 (en) | Apparatus and method for diagnosing a partial discharging having a calibration funtiality | |
| Schaefer | Signal detection with EMI receivers |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FZDE | Discontinued |