Method for partial discharge identification in high voltage applications and unit using the same

Abstract

The invention relates to a method for partial discharge identification in high voltage applications and a high voltage unit using the method, the method comprising the steps of: detecting a signal; transforming the signal from time domain to frequency domain; cutting off frequencies above a defined threshold; and retransforming the cut-off signal from frequency domain to time domain. The information content of the detected signal and the information content of the cut-off signal are determined and compared.

Description

[0001] The present invention relates to a method for partial discharge identification in high-voltage applications and a high-voltage unit using the method, the method comprising the steps of: detecting a signal; transforming the signal from the time domain to the frequency domain; truncating frequencies above a defined threshold; and transforming the truncated signal back from the frequency domain to the time domain.

[0002] Partial discharge is the localized dielectric breakdown of electrical insulation under high voltage. In high-voltage applications, such as in instrumentation transformers, for voltages ranging from several kilovolts to several kilovolts, partial discharge can lead to faults, measurement errors, damage, and even complete unit failure and grid failure. Dielectric breakdown occurs between conductors that cross an insulator, particularly a solid insulator such as ceramics, silicon, and / or non-conductive composite materials, or a fluid such as oil and / or gas (e.g., SF6 or clean air), which does not span the entire space between the conductors. In fluids such as oil, partial discharge may cause bubbles, but these are generally invisible. In or across solid insulators, partial discharge is usually invisible until irreversible damage to the equipment.

[0003] Partial discharges in or within solid insulators begin in voids, cracks, contaminants, or inclusions, particularly at the conductor-dielectric interface. In liquids, partial discharges begin, especially in bubbles, contaminants, or inclusions. Because partial discharges are confined to a portion of the insulation, the discharge only partially spans the distance between conductors. When a partial discharge begins, a high-frequency transient current pulse will appear and last from several nanoseconds to several microseconds.

[0004] The current pulses then repeatedly disappear and reappear as the sine wave crosses zero. Due to the high voltage and high levels of electrical background noise, the detection and measurement of partial discharges are difficult, complex, and costly. Partial discharges are particularly challenging to identify and determine in high-voltage applications during operation, and the damage and / or electrical losses can be substantial.

[0005] In particular, distinguishing partial discharge from noise introduced by sources such as corona signals, crosstalk signals, electromagnetic signals, and others in substations is difficult, and accurate partial discharge measurements in the field and online are extremely complex due to the high noise levels that obscure the true signal. Because of the short duration and the rise time of the partial discharge current, which is measured in nanoseconds, the typical way to quantify partial discharge magnitude is in picocoulombs. To study partial discharge, the intensity is displayed relative to time. For example, on an oscilloscope, partial discharge appears as uniformly spaced bursts at the peaks of a sine wave, while arcs and sparks occur randomly.

[0006] Partial discharge measurements are performed for both quality assessment and problem diagnosis. Partial discharge serves as an indicator of the condition of the equipment under test. Partial discharge measurements are performed, for example, in factory quality testing, particularly in Faraday cages and under limited power supply conditions, or in operation within a power grid under conditions of significant background noise. Different methods have been developed, especially for the testing and / or online monitoring of high-voltage units (e.g., high-voltage switches, bushings, and / or transformers). One method is based on transient grounding voltage, i.e., induced voltage spikes on the surface of surrounding grounded metal components.

[0007] Other methods for detecting and measuring partial discharges utilize ultrasonic acoustic sensors, ultra-high frequency sensors / antennas, high-frequency current transformer sensors, and / or directional couplers. Temporal reflectance techniques, which observe reflected waveforms (especially those displayed in partial discharge mapping formats), enable the spatial localization of insulation irregularities. Phase-resolved partial discharge patterns correlated with recorded signals and applied voltages, time-frequency plots that differentiate different sources using characteristics of the recorded signals, and cross-correlation analysis of the recorded signals facilitate the study of partial discharges.

[0008] Especially during field measurements, the methods described above offer only low sensitivity and require costly equipment with high complexity.

[0009] The purpose of this invention is to overcome the aforementioned problems. In particular, the purpose of this invention is to reduce costs, improve the sensitivity of partial discharge identification and determination, thereby reducing noise and improving measurement results by using less complex equipment to perform partial discharge measurements.

[0010] The above objective is achieved by a method for partial discharge identification in high-voltage applications, the method comprising the following steps:

[0011] -Detection signal

[0012] - Transform the signal from the time domain to the frequency domain

[0013] -Remove frequencies above a defined threshold

[0014] - Transform the truncated signal from the frequency domain back to the time domain. Determine and compare the information content of the detected signal with that of the truncated signal.

[0015] In high-voltage applications, distinguishing between noise signals and partial discharge signals is extremely difficult. Partial discharge signals are small but contain information compared to pure noise signals. Direct detection, identification, and measurement, especially monitoring, of partial discharge signals requires expensive and complex equipment and is not always reliably achievable. By measuring, separately, detecting the signal, transforming the signal to the frequency domain, truncating frequencies above a defined threshold, and transforming the truncated signal back to the time domain, and determining and comparing the information content of the detected signal and the truncated signal, noise and partial discharge can be distinguished with less measurement work, higher reliability, and lower cost. Partial discharge can be identified more reliably in appearance, and measures such as shutting down units and / or equipment and / or disconnecting electrical connections to the power grid can be initiated to avoid damage and / or failure. Partial discharge can be identified and determined with reduced cost and improved sensitivity. The method according to the invention identifies and distinguishes partial discharge and noise by analyzing the information content of the signal after truncating higher frequencies above a specific threshold, achieving partial discharge measurement using less complex equipment, thereby reducing noise and improving measurement results.

[0016] The method may include: identifying the signal as noise when the detected signal and the truncated signal have comparable, particularly identical, information content, and / or identifying the signal as a signal including partial discharge information when the truncated signal has lower information content compared to the detected signal. This enables reliable differentiation between noise signals and partial discharge signals while having the advantages described above.

[0017] The information content of the detected signal and the information content of the truncated signal can be calculated, and / or the information content can be compared by subtraction and / or division. Subtraction and division are easy to perform, especially automatically and within a short time (separately, simultaneously with the measurement). The determination and / or comparison of the information content of the detected signal and / or the information content of the truncated signal can be performed using a computer and / or online, especially in real time and / or automatically. The rapid and particularly automatic determination and / or comparison of the information content enables, for example, the monitoring of the timing of partial discharge events, and enables a rapid response to protect the equipment, for example, by cutting off the wire. Maintenance can be initiated for equipment (separately, units) that displays one or more partial discharge events. Warning signals can be generated and / or sent online and / or at one side, for example, via the Internet or a mobile network, or partial discharge events can be monitored online for a period of time, for example, via the Internet.

[0018] The steps, particularly all steps, of the method according to the invention can be repeated, especially periodically, within a fixed time period to increase the reliability of the results.

[0019] The amount of information can be determined using Shannon entropy. Entropy is a measure of the average information content of a signal. The amount of information in a signal can be calculated. Claude Elwood Shannon defined the entropy of a discrete, memoryless source of information (specifically, a discrete random variable H) for a given number of characters z. The information content is assigned to each probability p of the event z (specifically, the signal value) as I(z) = I(z) = -log₂p. z The entropy of character z is defined as the expected value of its information content, H1 = -∑p z log2p z The probability of the character z appearing is p. z Addends with different probabilities will not be added to the sum as defined. The entropy of a symbol w of length n is H. n =-∑p w log2p w The probability of the symbol w appearing is p. w The probability H is derived as the limit n→∞, where H=lim(n→∞)Hn / n.

[0020] Since the values ​​of noise signals are random, they contain no information. Truncating high frequencies does not change the Shannon entropy because no information is lost. For partial discharge signals, truncating the frequency reduces the Shannon entropy because information is lost. For pure noise, comparing the Shannon entropy of the measured signal with the Shannon entropy of the truncated signal yields the same value, and for signals including partial discharge, it yields a lower Shannon entropy value for the truncated signal compared to the measured signal. The occurrence of partial discharge can be identified with the advantages described above, and the value of the partial discharge can be calculated.

[0021] The detected signal can be a voltage signal and / or a current signal, and / or may include voltage signals and / or current signals. For example, high-frequency transient current and / or voltage pulses, lasting from several nanoseconds to several microseconds, can be measured, repeatedly disappearing and reappearing as the sine wave crosses zero. It is not necessary to identify partial discharge pulses within noisy signals using complex and expensive equipment with high resolution. Using the method according to the invention, partial discharges, even those with picocoulomb values, can be detected and measured using readily available and cost-effective equipment, even under high voltage and high levels of electrical background noise. High voltages can range from more than several kilovolts to 1200 kV, particularly from 15 kV to 1200 kV.

[0022] In the event of partial discharge being identified and / or detected, an alarm signal can be generated and / or sent, particularly online via the Internet and / or mobile networks. To prevent damage and / or destruction of the unit, further steps can be taken, such as shutting off the high voltage and / or disconnecting the unit from the power grid, or personnel can be dispatched to inspect and / or maintain the equipment (specifically, the electrical unit) where the partial discharge occurred.

[0023] The value of partial discharge can be evaluated by the following: an iterative method, removing signals identified as noise and / or determining a cutoff frequency above which signals are identified as noise.

[0024] The high-voltage unit according to the present invention, particularly the high-voltage unit using the method described above, is an instrument transformer, a high-voltage switch, a surge arrester, a bushing, and / or a medium-voltage or high-voltage transformer, or includes an instrument transformer, a high-voltage switch, a surge arrester, a bushing, and / or a medium-voltage or high-voltage transformer.

[0025] The unit may include an insulator, particularly an insulator housing with a rip on its outer surface.

[0026] The unit can be designed for voltages ranging from several kilovolts to 1200kV.

[0027] The advantages of the high-voltage unit described in accordance with the present invention, particularly using the method described above, are similar to those described in the previous method for partial discharge identification in high-voltage applications.

[0028] The invention is further described below with reference to the illustrated embodiments shown in the accompanying drawings, in which:

[0029] Figure 1 The diagram schematically illustrates a measured signal with a partial discharge signal transformed to the frequency domain, and...

[0030] Figure 2 The diagram schematically illustrates a measurement signal that is entirely composed of noise and has been transformed into the frequency domain.

[0031] exist Figure 1The diagram illustrates a signal measured by partial discharge, transformed to the frequency domain. The signal is measured, for example, in the form of a voltage and / or current varying over time. The alternating voltage and / or current are introduced through a partial discharge event between two electrodes, for example, an insulator, particularly the insulator of a high-voltage device (specifically, an application / unit). As previously mentioned, the partial discharge signal is relatively small compared to the high voltage between the electrodes, where the high voltage, for example, ranges from several kilovolts to 1200 kV, compared to the small voltage and current signals of the partial discharge, which are typically microcoulomb signals. Partial discharge occurs across, for example, the outer surface of a solid insulator. The solid insulator is, for example, composed of or includes ceramics, silicone resins, and / or composite materials. In particular, the insulator is part of a high-voltage unit (e.g., an instrument transformer, switch, bushing, transformer, and / or search and discharge device). The insulator is designed, for example, in the form of a housing and / or mounting structure of the high-voltage unit, such as a hollow cylinder with slits on its outer surface. Partial discharge can also occur, for example, in fluid insulators, particularly oil in high-voltage units.

[0032] Partial discharge is the breakdown of a partially insulating dielectric material under high voltage between electrodes. Partial discharges at or within an insulator can begin, for example, in voids, cracks, contaminants, or inclusions, particularly at the conductor-dielectric interface, and in liquids, particularly in bubbles, contaminants, and / or inclusions. Partial discharges are confined to a portion of the insulation. The discharge only partially spans the distance between the electrodes (specifically, between electrical conductors). When a partial discharge begins, a high-frequency transient current pulse appears and lasts for several nanoseconds to several microseconds. The current pulse then repeatedly disappears and reappears, for example, as a sine wave crosses zero. Partial discharge signals are short in duration and exhibit current rise times in the nanosecond range. High levels of electrical background noise, such as due to corona, crosstalk, and other effects, make accurate measurement, as well as simple identification and determination, of partial discharges within the signal of high voltage / current measurements difficult.

[0033] The measurement of a signal S (e.g., current and / or voltage signal) over time is performed. Due to the short duration and the nanosecond-range rise time of partial discharge current in high-voltage applications, visualization and differentiation between noise and partial discharge in voltage / current versus time graphs are difficult. To identify partial discharge and / or to determine its value, according to the invention, the information content I of the signal S is studied. For example, the measured signal S(t) over time is transformed from the time domain to the frequency domain by Fourier transform and / or discrete cosine transform. This is achieved by removing frequencies above a predefined frequency limit f. lim The signal components, f lim The frequency f above is removed to obtain the truncated signal. Frequency limit f limThe value is defined, for example, at the start of each monitoring activity (separately, measurement). This is achieved by removing the predefined frequency limit f. lim In the steps following the frequency f above, the signal is transformed from the frequency domain back to (and respectively, transformed back to) the time domain.

[0034] exist Figure 1 The image shows an example of a partial discharge signal S(f) in the frequency domain (i.e., after transformation from the time domain), where the signal S(f) i Each component of ) has information I i For comparison, Figure 2 The signal S(f) consisting entirely of noise is shown. i Here, each component S(f) i It contains no information because the signal is completely random. Signal information I, for example, is obtained as...

[0035]

[0036] That is, by adding the information associated with each component. For example... Figure 1 and Figure 2 As shown, the partial discharge signal contains information, while the noise signal does not. The noise signals are completely uncorrelated with each other. By switching from the time domain to the frequency domain, each component has a portion of the global information I. i If the measured signal includes partial discharge and not just noise, then by limiting the signal at frequency f... lim The above portion is truncated, resulting in information loss. According to the present invention, the method (specifically, the algorithm) extracts information from the sampled signal, for example, by calculating Shannon entropy, and combines the information content of the sampled signal with information obtained by limiting the signal at a frequency f. lim The information content of the truncated signal obtained by removing the above frequencies is compared.

[0037] The information content of the measured (specifically, sampled) signal is specifically calculated by the computer locally and / or in the cloud, as described above, for example, by calculating Shannon entropy, and the information content of the truncated signal (specifically, the signal after frequency truncation) is specifically calculated by the computer locally and / or in the cloud, also as described above, for example, by calculating Shannon entropy. Both calculations are performed, for example, concurrently with the measurement in time, or separately one after the other, or in a predefined order. After removing the frequency components, the time-domain signal of the truncated signal is obtained, for example, by inverse integral transform, particularly by Fourier transform and / or discrete cosine transform. The determination of the information content of the measured signal and the truncated signal is similar, for example, by calculating Shannon entropy in both cases.

[0038] According to the present invention, a comparison of the information content of the measured signal and the information content of the truncated signal leads to the identification and / or determination of partial discharge. The comparison is performed, for example, automatically and / or manually via a computer and / or in the cloud. Methods for comparing the signals include, for example, division and / or subtraction of the information content of the measured signal and the information content of the truncated signal. If the measured signal includes partial discharge, then for the division, the result is...

[0039]

[0040] Less than 1. In the absence of partial discharge, it is...

[0041]

[0042] The result of the division is exactly or approximately 1 because the signal contains only noise and no information is lost by truncating the frequency.

[0043] The identification and / or determination of partial discharges has the advantages described above, such as being able to generate and / or send warnings and trigger further actions, such as shutting off the voltage and / or disconnecting the high-voltage unit from the power grid to avoid damage and / or failure / malfunction of the unit. Partial discharge measurements also provide quality assessment and diagnostic results. Partial discharges are, for example, indicators of the condition of the equipment under test. Partial discharge measurements are performed during factory quality testing to demonstrate the proper functioning of the manufactured units. In operation on the power grid, high-voltage units (e.g., high-voltage switches, bushings, and / or transformers) can be monitored online in environments with significant background noise.

[0044] For example, signal detection can be performed based on transient ground voltage, i.e., induced voltage spikes on the surface of a surrounding grounded metal object. Other methods for detecting and measuring partial discharge use ultrasonic acoustic sensors, ultra-high frequency sensors / antennas, high-frequency current transformer sensors, and / or directional couplers.

[0045] The embodiments of the present invention described above can also be used in combination and combined with embodiments known from the prior art. For example, partial discharge can be detected by performing the steps, particularly all the steps as described above, once. Higher reliability can be achieved by repeatedly performing the steps, particularly all the steps as described above. An iterative method for determining partial discharge may include determining a frequency limit f. lim The value of f. The iterative method may include repeating the steps according to the invention and changing the cutoff frequency limit f. lim Until the Shannon entropy changes and / or with the frequency constraint f lim Despite the change, Shannon entropy remained unchanged.

Claims

1. A method for partial discharge identification in high-voltage applications, comprising the following steps: - Detect signal, - Transform the signal from the time domain to the frequency domain. - The truncated signal is obtained by removing frequencies above a defined threshold. - Transform the truncated signal from the frequency domain back to the time domain. Its characteristic is that it determines and compares the information content of the detected signal with the information content of the truncated signal. When the detected signal and the truncated signal have comparable information content, the signal is identified as noise; when the truncated signal has lower information content compared to the detected signal, the signal is identified as a signal including partial discharge information. The amount of information is determined by calculating Shannon entropy.

2. The method according to claim 1, characterized in that, Calculate the information content of the detected signal and the information content of the truncated signal, and perform a comparison of the information content by subtraction and / or division.

3. The method according to claim 1 or 2, characterized in that, Perform real-time determination and / or comparison of the information content of detected signals and / or truncated signals.

4. The method according to claim 1 or 2, characterized in that, Repeat the steps within a fixed time period to increase the reliability of the results.

5. The method according to claim 1 or 2, characterized in that, The detected signal is a voltage signal and / or a current signal, or includes both voltage and / or current signals.

6. The method according to claim 1 or 2, characterized in that, When partial discharge is identified and / or detected, an alarm signal is generated and / or sent.

7. The method according to claim 1 or 2, characterized in that, The value of partial discharge is evaluated by using an iterative method to remove signals identified as noise or to identify a cutoff frequency above which signals are identified as noise.

8. A high-voltage unit using the method according to any one of claims 1 to 7, characterized in that, The unit is an instrument transformer, a high-voltage switch, a surge arrester, a bushing, and / or a medium-voltage or high-voltage transformer, or includes an instrument transformer, a high-voltage switch, a surge arrester, a bushing, and / or a medium-voltage or high-voltage transformer.

9. The high-voltage unit according to claim 8, characterized in that, The unit includes an insulator.

10. The high-voltage unit according to claim 8 or 9, wherein, The unit is designed for voltages ranging from over 1 kV to 1200 kV.

Images