High-voltage equipment partial discharge on-line detection method and system

Abstract

The invention relates to the technical field of partial discharge detection, and discloses a high-voltage equipment partial discharge online detection method and system. The method comprises the steps of obtaining a first voltage signal and extracting a voltage peak value, and calculating environment reference voltage; monitoring a second voltage signal in real time, and intercepting a signal segment when the second voltage signal exceeds an environment reference voltage; judging whether the signal accords with a partial discharge characteristic or not, and marking the signal as an effective signal; extracting phase angles and amplitudes of the effective signals as features, and carrying out gridding statistics; calculating the occurrence rate of each partial discharge based on the statistical result, and judging whether the partial discharge exists or not; and if yes, extracting a time sequence feature vector in a periodic monitoring mode, and generating trend diagnosis and early warning information. The system comprises a corresponding threshold setting unit, a voltage acquisition unit, an effective identification unit, a feature extraction unit, a discharge identification unit and a trend early warning unit. According to the invention, the accuracy and timeliness of detection can be improved.

Description

Technical Field

[0001] This application relates to the field of partial discharge detection technology, and in particular to an online detection method and system for partial discharge in high-voltage equipment. Background Technology

[0002] During long-term operation, high-voltage equipment is prone to partial discharge due to insulation aging, manufacturing defects, or external environmental influences. Partial discharge is a significant precursor to insulation degradation; if not detected and addressed promptly, it can lead to equipment failure or even major accidents. Therefore, online monitoring of partial discharge in high-voltage equipment is of paramount importance.

[0003] Current common partial discharge detection methods mostly rely on threshold triggering and manual analysis, which have the following shortcomings: First, environmental noise interference is significant, easily leading to false triggering and misjudgment, affecting detection reliability; second, they lack intelligent recognition capabilities for discharge signal waveforms, making it difficult to effectively distinguish between real discharges and interference pulses; third, the extraction and analysis of discharge characteristics are relatively simple, mostly limited to single dimensions such as amplitude and frequency, making it difficult to support accurate identification of discharge types and risk assessment; fourth, they lack long-term trend tracking and early warning mechanisms, making it impossible to quantitatively assess and provide early warnings of the evolution of insulation status.

[0004] Therefore, there is a need for an online detection method and system that can adapt to complex field environments, has high anti-interference capabilities, can achieve in-depth extraction of discharge characteristics and type identification, and supports trend analysis and early warning, so as to improve the intelligence level and operational safety of high-voltage equipment condition monitoring.

[0005] Therefore, the present invention provides an online detection method and system for partial discharge in high-voltage equipment. Summary of the Invention

[0006] The embodiments in this specification provide the following technical solutions: Step S1: Obtain the first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. Step S2: Monitor the second voltage signal in real time. When the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within each first time period before and after the starting point. Step S3: Determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If yes, mark the corresponding second voltage signal as a valid partial discharge signal. Step S4: Extract the signal features of the effective partial discharge signals. The signal features include the first feature and the second feature. For all effective partial discharge signals within the preset second time period, perform signal feature statistics based on the signal features to obtain the signal feature statistics results. Step S5: Calculate the first ratio for each type of partial discharge based on the statistical results of signal characteristics, and determine whether partial discharge exists based on the first ratio; Step S6: When partial discharge is determined to exist, a preset third time period is used as the monitoring cycle. The effective partial discharge signal in each monitoring cycle is monitored, the time-series feature vector is extracted and calculated, and trend diagnosis and early warning information is generated based on the time-series feature vector.

[0007] The embodiments in this specification also provide that the signal features for extracting effective partial discharge signals include: From the power supply circuit of the high-voltage equipment, the power frequency voltage signal synchronized with the working voltage of the high-voltage equipment is obtained as a reference phase signal. The effective partial discharge signal is time-aligned with the reference phase signal to obtain the reference point in the reference phase signal where the start time of the effective partial discharge signal is collected. The voltage zero point closest to the reference point is obtained. The time difference between the voltage zero point and the reference point is calculated. The time difference is divided by the power supply voltage period and then multiplied by 360 degrees to obtain the phase angle of the effective partial discharge signal as the first feature. The maximum voltage value in the effective partial discharge signal is used as the second feature.

[0008] The embodiments in this specification also provide that determining whether the second voltage signal conforms to preset partial discharge characteristics includes: Find the first moment when the voltage signal first crosses the zero voltage value from the second voltage signal. Starting from the first moment, count the total number of times the second voltage signal crosses the zero voltage value and obtain all the peak points of the second voltage signal. Perform exponential fitting on all the peak points to obtain the first coefficient. If the total number of times is greater than the preset second threshold and the first coefficient is less than the preset third threshold, it is determined that the second voltage signal meets the preset partial discharge characteristics; otherwise, it is determined that it does not meet the characteristics.

[0009] The embodiments in this specification also provide that signal feature statistics are performed based on signal features, including: The two-dimensional space composed of continuous first and second features is discretized into a cell grid. The signal features corresponding to all effective partial discharge signals are traversed, and each signal feature is classified into its own cell grid. The first number of signal features classified into each cell grid is counted. Each cell grid is represented by a first feature interval, a second feature interval, and a corresponding first number.

[0010] The embodiments in this specification also provide that, based on the statistical results of signal characteristics, a first ratio for each type of partial discharge is calculated, including: For each type of partial discharge, multiple corresponding interest analysis intervals are preset. For each interest analysis interval, based on the statistical results of signal characteristics, the sum of the first number of cell grids appearing in each interest analysis interval is obtained as the second number. The second number of all interest analysis intervals is added together to obtain the third number. The third number is used as the first number of occurrences of the corresponding partial discharge type. The first number is divided by the number of all effective discharge signals in the second time period to obtain the first ratio.

[0011] The embodiments in this specification also provide that, based on the first count, determining whether partial discharge exists includes: Obtain the first ratio of all partial discharge types. If any first ratio is greater than a preset fourth threshold, it is determined that a partial discharge exists.

[0012] The embodiments in this specification also provide that extracting and calculating time-series feature vectors includes: Signal features are extracted from the effective partial discharge signals in each monitoring cycle to obtain the corresponding signal features. The signal features generated in multiple monitoring cycles form a time-series feature vector, which includes a first time-series feature vector, a second time-series feature vector, and a third time-series feature vector.

[0013] The embodiments in this specification also provide that trend diagnosis and early warning information is generated based on time-series feature vectors, including: Multiple trend indicators are calculated based on time series feature vectors. The trend indicators include the distribution stability of the first time series feature vector, the growth trend of the second time series feature vector, and the growth trend of the third time series feature vector. A corresponding trend threshold is set for each trend indicator. The degradation status is classified based on trend indicators, and corresponding early warning information is generated.

[0014] This specification also provides embodiments that the system includes: The threshold setting unit is used to acquire a first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. The voltage acquisition unit is used to monitor the second voltage signal in real time, and when the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within a preset first time period after the starting point of the voltage signal. The effective identification unit is used to determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If it does, the corresponding second voltage signal is marked as an effective partial discharge signal. The feature extraction unit is used to extract the signal features of the effective partial discharge signals. The signal features include a first feature and a second feature. For all effective partial discharge signals within a preset second time period, the unit performs signal feature statistics based on the signal features to obtain the signal feature statistics results. The discharge identification unit is used to calculate the first ratio of each partial discharge type based on the statistical results of signal characteristics, and to determine whether a partial discharge exists based on the first ratio. The trend warning unit is used to determine the presence of partial discharge. It uses a preset third time period as the monitoring cycle, monitors the effective partial discharge signal in each monitoring cycle, extracts and calculates the time-series feature vector, and generates trend diagnosis and warning information based on the time-series feature vector.

[0015] Compared with the prior art, the beneficial effects of the present invention are at least as follows: The technical solution provided in this application effectively distinguishes partial discharge signals from environmental noise by dynamically setting the environmental reference voltage and identifying waveform features, thereby improving the accuracy of triggering and identification and reducing the false alarm rate. By introducing phase-amplitude two-dimensional feature statistics and gridded analysis methods, the discharge distribution pattern can be intuitively reflected, supporting quantitative identification and activity assessment of different discharge types. Through the extraction of time-series feature vectors and the calculation of trend indicators, continuous tracking and state classification of the partial discharge development process can be achieved, providing a scientific basis for early warning and preventive maintenance. By integrating functions such as threshold setting, feature extraction, type identification, and trend warning, the entire process of automated online monitoring from signal acquisition to intelligent diagnosis can be realized, improving monitoring efficiency and reliability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of one embodiment of an online partial discharge detection method for high-voltage equipment in this application. Figure 2 This is a schematic diagram of one embodiment of an online partial discharge detection system for high-voltage equipment in this application. Detailed Implementation

[0018] This application provides a method and system for online detection of partial discharge in high-voltage equipment. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0019] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the online detection method for partial discharge in high-voltage equipment in this application includes: Step S1: Obtain the first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. Specifically, to avoid misinterpreting fixed environmental noise as a discharge event, stray current flowing through the metal surface near the high-voltage equipment installation location is measured and converted into a first voltage signal. This quantifies the inherent background noise voltage level, which is unrelated to the operation of the high-voltage equipment. The first voltage signal represents the environmental noise of the corresponding monitoring environment. To obtain the original voltage signal containing partial discharge pulses and provide a data source for subsequent analysis, a current sensor is installed at the sensitive part of the high-voltage equipment to sense the high-frequency current generated by the partial discharge pulses. This current is explicitly converted into a second voltage signal through a signal conditioning circuit. The second voltage signal contains the partial discharge pulses and environmental noise. To set a reasonable voltage threshold based on real-time environmental noise and avoid storing a large amount of invalid data, an environmental reference voltage is set as the trigger threshold. The environmental reference voltage needs to be higher than the environmental noise to ensure that a significant voltage signal is triggered. Therefore, multiple voltage peaks are extracted from the first voltage signal, and the largest voltage peak is multiplied by a preset coefficient (e.g., 1.2) to obtain the value as the environmental reference voltage. Only when the second voltage signal is greater than the environmental reference voltage is it determined that a partial discharge may have occurred; otherwise, it may all be environmental noise.

[0020] Step S2: Monitor the second voltage signal in real time. When the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within each first time period before and after the starting point. Specifically, in order to obtain a complete signal of a single suspected partial discharge time, provide an accurate analysis object for subsequent partial discharge identification, and avoid recording lengthy and invalid data, the second voltage signal is monitored in real time. When the second voltage signal is greater than the ambient reference voltage, the second voltage signal is acquired within a preset first time period before and after the starting point of the ambient reference point. For example, when a certain instantaneous voltage is detected to be greater than the ambient reference voltage, the second voltage signal is immediately acquired within 2 microseconds before and after the starting point.

[0021] Step S3: Determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If yes, mark the corresponding second voltage signal as a valid partial discharge signal. Specifically, in order to filter out signals that are truly generated by partial discharge from all acquired second voltage signals, eliminate noise caused by external environmental interference, and ensure the accuracy of subsequent analysis data, identification is performed based on the physical characteristics of the partial discharge signal (it exhibits decaying oscillation in a homogeneous medium). It is then determined whether the second voltage signal conforms to the preset partial discharge characteristics. Only when the second voltage signal conforms to the preset partial discharge characteristics is the second voltage signal marked as a valid partial discharge signal.

[0022] Step S4: Extract the signal features of the effective partial discharge signals. The signal features include the first feature and the second feature. For all effective partial discharge signals within the preset second time period, perform signal feature statistics based on the signal features to obtain the signal feature statistics results. Specifically, in order to achieve online intelligent monitoring more efficiently, complex time-domain voltage signals are transformed into feature data with clear physical meaning, and signal features of effective partial discharge signals are extracted. The signal features include a first feature and a second feature. The first feature refers to the phase of the waveform corresponding to the effective partial discharge signal, and the second feature refers to the maximum peak value of the effective partial discharge signal. In order to perform in-depth analysis of discrete partial discharge time, for all effective partial discharge signals within a preset second time period (which is longer than the first time period), signal feature statistics are obtained based on the signal features.

[0023] Step S5: Calculate the first ratio for each type of partial discharge based on the statistical results of signal characteristics, and determine whether partial discharge exists based on the first ratio; Specifically, in order to accurately determine whether partial discharge exists, different types of partial discharge caused by different reasons exhibit different characteristics. Environmental noise or interference does not have this characteristic. Therefore, based on the statistical results of signal characteristics, the first ratio of each type of partial discharge is calculated. Only when the first ratio of any type of partial discharge is greater than a certain preset threshold is it determined that this type of partial discharge actually exists. Only then can the existence of partial discharge be confirmed. Otherwise, it may be due to environmental noise or interference.

[0024] Step S6: When partial discharge is determined to exist, a preset third time period is used as the monitoring cycle. The effective partial discharge signal in each monitoring cycle is monitored, the time-series feature vector is extracted and calculated, and trend diagnosis and early warning information is generated based on the time-series feature vector.

[0025] Specifically, after determining that partial discharge does exist, it is known that partial discharge may be caused by various reasons such as insulation defects. Partial discharge will develop slowly and gradually deteriorate. In order to track the deterioration process and quantify the trend, a preset third time period is used as the monitoring cycle. Effective partial discharge signals within each monitoring cycle are monitored, and the corresponding time-series feature vectors are extracted. Based on the time-series feature vectors, trend diagnosis and early warning information are generated to achieve early warning. For example, problems can be detected in advance before insulation completely fails and causes damage, providing sufficient space for planned maintenance and avoiding larger failures caused by partial discharge.

[0026] In one specific embodiment, extracting the signal features of the effective partial discharge signal specifically includes the following steps: From the power supply circuit of the high-voltage equipment, the power frequency voltage signal synchronized with the working voltage of the high-voltage equipment is obtained as a reference phase signal. The effective partial discharge signal is time-aligned with the reference phase signal to obtain the reference point in the reference phase signal where the start time of the effective partial discharge signal is collected. The voltage zero point closest to the reference point is obtained. The time difference between the voltage zero point and the reference point is calculated. The time difference is divided by the power supply voltage period and then multiplied by 360 degrees to obtain the phase angle of the effective partial discharge signal as the first feature. The maximum voltage value in the effective partial discharge signal is used as the second feature.

[0027] Specifically, to extract the signal characteristics of the effective partial discharge signal, and to establish a correlation between the randomly occurring discharge pulse signal and the periodic changes in the power supply voltage of the high-voltage equipment, thereby obtaining the first characteristic of the effective partial discharge signal, namely the phase information, a power frequency signal with the same frequency and phase as the operating voltage of the high-voltage equipment is obtained from the power supply circuit of the high-voltage equipment (usually through a voltage transformer). This power frequency voltage signal is used as the reference phase signal. Since the effective discharge signal and the reference phase signal are strictly aligned on the time axis, the effective partial discharge signal is time-aligned with the reference phase signal. By finding the starting time point that triggers the acquisition of the discharge pulse segment (i.e., the "effective partial discharge signal") and locating the same moment on the time axis of the reference phase signal, assuming that at time t1, the second voltage signal exceeds the loop... The ambient reference voltage triggered data acquisition. The system recorded the discharge signal 2 microseconds before and after time t1. Simultaneously, on the time axis of the reference phase, the point corresponding to time t1 was found as the reference point. Using the reference point as a reference, the system searched forward or backward in the reference phase signal for the closest point where the voltage changed from negative to positive as the voltage zero point. Then, the time difference between the voltage zero point and the reference point was calculated. The time difference was divided by the power supply voltage period and then multiplied by 360°. Assuming the time difference is 510 microseconds and the power supply voltage period is 0.02s (corresponding to a 50Hz power grid), the phase angle of the effective partial discharge signal = (510 microseconds / 0.02s)*360°≈9.18°, indicating that the partial discharge occurred about 9 degrees after the voltage zero point. The phase angle was used as the first feature of the effective partial discharge signal.

[0028] To quantify the intensity of partial discharge pulses, the discharge amplitude is used as another key feature. The entire range of the acquired effective partial discharge signals is traversed to find the maximum positive peak value or the minimum negative peak value of the voltage (the one with the largest absolute value) as the second feature.

[0029] In one specific embodiment, determining whether the second voltage signal conforms to a preset partial discharge characteristic includes the following steps: Find the first moment when the voltage signal first crosses the zero voltage value from the second voltage signal. Starting from the first moment, count the total number of times the second voltage signal crosses the zero voltage value and obtain all the peak points of the second voltage signal. Perform exponential fitting on all the peak points to obtain the first coefficient. If the total number of times is greater than the preset second threshold and the first coefficient is less than the preset third threshold, it is determined that the second voltage signal meets the preset partial discharge characteristics; otherwise, it is determined that it does not meet the characteristics.

[0030] Specifically, to distinguish between real partial discharge oscillation pulses and non-oscillating electromagnetic interference pulses, the oscillation behavior and attenuation mode of the pulse waveform after the trigger point are analyzed to quantify its consistency with the real partial discharge physical process. To locate the actual starting point of the oscillation waveform, it is used as a unified time reference for subsequent zero-crossing techniques and waveform analysis. The trigger point may be located at any position on the oscillation waveform; therefore, the first moment when the voltage signal first crosses the zero voltage value is found from the second voltage signal. To quantify the oscillation activity of the waveform, real partial discharge pulses, due to oscillation in the RLC equivalent circuit, will produce multiple positive and negative alternations. Many narrow pulse interferences often only cross zero once or not at all. Starting from the first moment, the total number of times the second voltage signal crosses the zero voltage value is counted. Due to the presence of dielectric loss and circuit resistance, the peak value of real partial discharge oscillations decays exponentially with time. To analyze the attenuation characteristics of the oscillation waveform, all peak points of the second voltage signal are obtained, and an exponential fit is performed on all peak points to obtain the first coefficient. The first coefficient is the attenuation coefficient, which indicates the rate of attenuation; the larger the attenuation coefficient, the faster the attenuation. To comprehensively consider both oscillation persistence and attenuation regularity, ensuring that the selected signal possesses both time-domain and frequency-domain characteristics of partial discharge, the following conditions are met: the total number of oscillations exceeds a preset second threshold (representing the minimum number of zero-crossings required), and the first coefficient is less than a preset third threshold (representing the upper limit of the attenuation coefficient, used to ensure that attenuation is not too rapid; excessively rapid attenuation might correspond to extremely high-frequency interference that is quickly damped. Since real partial discharge oscillations have a certain time constant, the attenuation coefficient will be within a reasonable range. This determines that the second voltage signal conforms to the preset partial discharge characteristics. This step effectively filters out a large number of high-amplitude but monotonous instantaneous interferences, excluding falsely triggered signals from subsequent processing and improving the purity of the effective partial discharge signal.

[0031] In one specific embodiment, signal feature statistics are performed based on signal features, which specifically includes the following steps: The two-dimensional space composed of continuous first and second features is discretized into a cell grid. The signal features corresponding to all effective partial discharge signals are traversed, and each signal feature is classified into its own cell grid. The first number of signal features classified into each cell grid is counted. Each cell grid is represented by a first feature interval, a second feature interval, and a corresponding first number.

[0032] Specifically, after feature extraction of the effective partial discharge signals within the second time period, a large number of signal features are collected. These signal features are discrete. To utilize these features later, signal feature statistics are performed on the signal features of all effective partial discharge signals. The original, scattered signal features that cannot be directly applied are transformed into a gridded statistical representation that reflects the distribution density and pattern of partial discharge in phase and amplitude. First, based on the data ranges of the first and second features, the value ranges of the first and second features are determined. The continuous value ranges are uniformly divided into small intervals. The small intervals of the two-dimensional features can be discretized into multiple... The cell grid iterates through all valid partial discharge signals and classifies each signal feature into its respective cell grid. To quantify the frequency of discharge activity within each cell grid, the first number of signal features classified into each cell grid is counted. The first number reflects the statistical probability of partial discharge under specific phase and voltage conditions. The signal feature statistics include the corresponding records of all cell grids. The corresponding record of each cell grid includes the first feature interval (phase range), the second feature interval (voltage amplitude range), and the first number of the cell grid, which fully describes the distribution of partial discharge in the two-dimensional space of phase and voltage in each monitoring cycle.

[0033] In one specific embodiment, the first ratio for each partial discharge type is calculated based on the statistical results of signal characteristics, specifically including the following steps: For each type of partial discharge, multiple corresponding interest analysis intervals are preset. For each interest analysis interval, based on the statistical results of signal characteristics, the sum of the first number of cell grids appearing in each interest analysis interval is obtained as the second number. The second number of all interest analysis intervals is added together to obtain the third number. The third number is used as the first number of occurrences of the corresponding partial discharge type. The first number is divided by the number of all effective discharge signals in the second time period to obtain the first ratio.

[0034] Specifically, in order to solve the problem of how to automatically and quantitatively identify possible partial discharge types from the statistical results of signal features, a typical phase amplitude distribution range, also known as the analysis range of interest, is preset for each partial discharge type. The occurrence ratio of each partial discharge type falling within these analysis ranges of interest is statistically analyzed to quantify the activity of each partial discharge type.

[0035] Suppose there are two types of partial discharges, A and B. The analysis intervals for type A are intervals A1 and A2, and the analysis interval for type B is interval B1. For each analysis interval, calculate the first number of cell grids falling within that interval. Sum the first numbers of all cell grids within the analysis interval to obtain the second number. Sum the second numbers of all corresponding analysis intervals for each partial discharge type to obtain the third number. Use the third number as the first occurrence count of the corresponding partial discharge type. The first count represents the total number of occurrences of each partial discharge type. Divide the first count by the number of all valid discharge signals in the second time period to obtain the first ratio. The first ratio is used to represent the activity index of each partial discharge type, which can serve as a direct scientific basis for subsequent judgments on the existence of partial discharges.

[0036] In one specific embodiment, determining whether partial discharge exists based on the first count includes the following steps: Obtain the first ratio of all partial discharge types. If any first ratio is greater than a preset fourth threshold, it is determined that a partial discharge exists.

[0037] Specifically, partial discharge usually occurs continuously and significantly. In order to accurately determine whether there is partial discharge activity, the first ratio of all partial discharge types is obtained. If any first ratio is greater than a preset fourth threshold, it indicates that the corresponding partial discharge type has a high activity level and the possibility of partial discharge is very high. Therefore, it is judged that partial discharge exists. If all first ratios are less than or equal to the preset fourth threshold, it indicates that the partial discharge may be caused by some kind of environmental interference and occurs less frequently. Therefore, it is judged that there is no partial discharge.

[0038] In one specific embodiment, extracting and calculating the time-series feature vector further includes: Signal features are extracted from the effective partial discharge signals in each monitoring cycle to obtain the corresponding signal features. The signal features generated in multiple monitoring cycles form a time-series feature vector, which includes a first time-series feature vector, a second time-series feature vector, and a third time-series feature vector.

[0039] Specifically, after identifying the presence of partial discharge, the confirmed partial discharge activity is continuously tracked and quantitatively assessed over a long period. Based on the quantitative assessment results, the rate and severity of degradation are graded and early warnings are issued. First, a preset third time period is used as the monitoring cycle. Valid partial discharge signals within each monitoring cycle are acquired. Signal features are extracted from the valid partial discharge signals within each monitoring cycle to obtain the corresponding signal features. The signal features generated from multiple monitoring cycles form a time-series feature vector, which includes a first time-series feature vector, a second time-series feature vector, and a third time-series feature vector. The first time-series feature vector is a vector composed of multiple first features within the monitoring cycle; the second time-series feature vector is a vector composed of multiple second features within the monitoring cycle; and the third time-series feature vector is a vector composed of the total number of discharges within the monitoring cycle.

[0040] In one specific embodiment, generating trend diagnosis and early warning information based on time-series feature vectors includes the following steps: Multiple trend indicators are calculated based on time series feature vectors. The trend indicators include the distribution stability of the first time series feature vector, the growth trend of the second time series feature vector, and the growth trend of the third time series feature vector. A corresponding trend threshold is set for each trend indicator. The degradation status is classified based on trend indicators, and corresponding early warning information is generated.

[0041] Specifically, the distribution stability of the first time-series feature vector over multiple monitoring periods can be analyzed. Specifically, the mean square error of the first feature over multiple monitoring periods can be calculated as the corresponding distribution stability. For each second feature in the second time-series feature vector, its growth trend with the monitoring period can be analyzed. For example, its growth rate with the monitoring period can be calculated and the growth rate can be used as the corresponding trend indicator. For the third time-series feature vector, its growth rate with the monitoring period can also be calculated and the growth rate can be used as the corresponding trend indicator.

[0042] Finally, based on the significance and persistence of changes in trend indicators, the deterioration state is classified, and a multi-level early warning rule is established. For example, a Level 1 early warning is established when only one trend indicator exceeds the threshold; a Level 2 early warning is established when two trend indicators exceed the preset threshold; and a Level 3 early warning is established when all three trend indicators exceed their corresponding preset thresholds. Corresponding warning information is generated for each level. For example: a Level 1 warning indicates a slight increasing trend in partial discharge activity, suggesting increased monitoring; a Level 2 warning indicates a significant increase in partial discharge activity, suggesting an upcoming inspection; and a Level 3 warning indicates a drastic deterioration in discharge activity, posing a high risk of failure, suggesting immediate inspection. These warnings alert relevant personnel to take appropriate preventative measures.

[0043] The above describes an online partial discharge detection method for high-voltage equipment according to an embodiment of this application. The following describes an online partial discharge detection system for high-voltage equipment according to an embodiment of this application. Please refer to [link / reference]. Figure 2 One embodiment of the online partial discharge detection system for high-voltage equipment in this application includes: The threshold setting unit is used to acquire a first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. The voltage acquisition unit is used to monitor the second voltage signal in real time, and when the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within a preset first time period after the starting point of the voltage signal. The effective identification unit is used to determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If it does, the corresponding second voltage signal is marked as an effective partial discharge signal. The feature extraction unit is used to extract the signal features of the effective partial discharge signals. The signal features include a first feature and a second feature. For all effective partial discharge signals within a preset second time period, the unit performs signal feature statistics based on the signal features to obtain the signal feature statistics results. The discharge identification unit is used to calculate the first ratio of each partial discharge type based on the statistical results of signal characteristics, and to determine whether a partial discharge exists based on the first ratio. The trend warning unit is used to determine the presence of partial discharge. It uses a preset third time period as the monitoring cycle, monitors the effective partial discharge signal in each monitoring cycle, extracts and calculates the time-series feature vector, and generates trend diagnosis and warning information based on the time-series feature vector.

[0044] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0045] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0046] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for online detection of partial discharge in high-voltage equipment, characterized in that, The method includes: Step S1: Obtain the first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. Step S2: Monitor the second voltage signal in real time. When the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within each first time period before and after the starting point. Step S3: Determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If yes, mark the corresponding second voltage signal as a valid partial discharge signal. Step S4: Extract the signal features of the effective partial discharge signals. The signal features include the first feature and the second feature. For all effective partial discharge signals within the preset second time period, perform signal feature statistics based on the signal features to obtain the signal feature statistics results. Step S5: Calculate the first ratio for each type of partial discharge based on the statistical results of signal characteristics, and determine whether partial discharge exists based on the first ratio; Step S6: When partial discharge is determined to exist, a preset third time period is used as the monitoring cycle. The effective partial discharge signal in each monitoring cycle is monitored, the time-series feature vector is extracted and calculated, and trend diagnosis and early warning information is generated based on the time-series feature vector.

2. The method according to claim 1, characterized in that, The signal features of the effective partial discharge signal are extracted, including: From the power supply circuit of the high-voltage equipment, the power frequency voltage signal synchronized with the working voltage of the high-voltage equipment is obtained as a reference phase signal. The effective partial discharge signal is time-aligned with the reference phase signal to obtain the reference point in the reference phase signal where the start time of the effective partial discharge signal is collected. The voltage zero point closest to the reference point is obtained. The time difference between the voltage zero point and the reference point is calculated. The time difference is divided by the power supply voltage period and then multiplied by 360 degrees to obtain the phase angle of the effective partial discharge signal as the first feature. The maximum voltage value in the effective partial discharge signal is used as the second feature.

3. The method according to claim 1, characterized in that, Determine whether the second voltage signal conforms to the preset partial discharge characteristics, including: Find the first moment when the voltage signal first crosses the zero voltage value from the second voltage signal. Starting from the first moment, count the total number of times the second voltage signal crosses the zero voltage value and obtain all the peak points of the second voltage signal. Perform exponential fitting on all the peak points to obtain the first coefficient. If the total number of times is greater than the preset second threshold and the first coefficient is less than the preset third threshold, it is determined that the second voltage signal meets the preset partial discharge characteristics; otherwise, it is determined that it does not meet the characteristics.

4. The method according to claim 1, characterized in that, Signal feature statistics based on signal characteristics include: The two-dimensional space composed of continuous first and second features is discretized into a cell grid. The signal features corresponding to all effective partial discharge signals are traversed, and each signal feature is classified into its own cell grid. The first number of signal features classified into each cell grid is counted. Each cell grid is represented by a first feature interval, a second feature interval, and a corresponding first number.

5. The method according to claim 1, characterized in that, The first ratio for each partial discharge type is calculated based on the statistical results of signal characteristics, including: For each type of partial discharge, multiple corresponding interest analysis intervals are preset. For each interest analysis interval, based on the statistical results of signal characteristics, the sum of the first number of cell grids appearing in each interest analysis interval is obtained as the second number. The second number of all interest analysis intervals is added together to obtain the third number. The third number is used as the first number of occurrences of the corresponding partial discharge type. The first number is divided by the number of all effective discharge signals in the second time period to obtain the first ratio.

6. The method according to claim 1, characterized in that, Determining the presence of partial discharge based on the first count includes: Obtain the first ratio of all partial discharge types. If any first ratio is greater than a preset fourth threshold, it is determined that a partial discharge exists.

7. The method according to claim 1, characterized in that, Extracting and calculating time-series feature vectors also includes: Signal features are extracted from the effective partial discharge signals in each monitoring cycle to obtain the corresponding signal features. The signal features generated in multiple monitoring cycles form a time-series feature vector, which includes a first time-series feature vector, a second time-series feature vector, and a third time-series feature vector.

8. The method according to claim 1, characterized in that, Trend diagnosis and early warning information is generated based on time-series feature vectors, including: Multiple trend indicators are calculated based on time series feature vectors. The trend indicators include the distribution stability of the first time series feature vector, the growth trend of the second time series feature vector, and the growth trend of the third time series feature vector. A corresponding trend threshold is set for each trend indicator. The degradation status is classified based on trend indicators, and corresponding early warning information is generated.

9. A high-voltage equipment partial discharge online detection system, used to implement the high-voltage equipment partial discharge online detection method as described in any one of claims 1-8, characterized in that, The system includes: The threshold setting unit is used to acquire a first voltage signal, extract multiple voltage peaks from the first voltage signal, and multiply the largest voltage peak by a preset coefficient to obtain the value as the environmental reference voltage. The voltage acquisition unit is used to monitor the second voltage signal in real time, and when the second voltage signal is greater than the ambient reference voltage, acquire the second voltage signal within a preset first time period after the starting point of the voltage signal. The effective identification unit is used to determine whether the second voltage signal sequence conforms to the preset partial discharge characteristics. If it does, the corresponding second voltage signal is marked as an effective partial discharge signal. The feature extraction unit is used to extract the signal features of the effective partial discharge signals. The signal features include a first feature and a second feature. For all effective partial discharge signals within a preset second time period, the unit performs signal feature statistics based on the signal features to obtain the signal feature statistics results. The discharge identification unit is used to calculate the first ratio of each partial discharge type based on the statistical results of signal characteristics, and to determine whether a partial discharge exists based on the first ratio. The trend warning unit is used to determine the presence of partial discharge. It uses a preset third time period as the monitoring cycle, monitors the effective partial discharge signal in each monitoring cycle, extracts and calculates the time-series feature vector, and generates trend diagnosis and warning information based on the time-series feature vector.

Images