A method and system for separating partial discharge signals of a transformer

Abstract

The present invention belongs to the technical field of electrical variable measurement and processing, and particularly relates to a method and system for separating partial discharge signals of a transformer. The method includes: synchronously sampling multiple measurement points of a transformer to be measured at different times to obtain multiple aliased test signals; decomposing each aliased test signal to obtain multiple independent source signals; calculating the rising fluctuation parameters of each independent source signal at different times to determine the target aliased test signal with partial discharge signals; determining the feature vectors of each independent source signal in the target aliased test signal, and performing clustering processing on all independent source signals in the target aliased test signal to obtain multiple clustering clusters; determining the density eigenvalue of each clustering cluster to determine the target clustering cluster corresponding to the internal partial discharge signal; and separating the internal partial discharge signal in the target aliased test signal according to the target clustering cluster. The solution provided by the present invention realizes the accurate separation of the internal partial discharge signal.

Description

Technical Field

[0001] The present invention belongs to the technical field of electrical variable measurement and processing, and particularly relates to a method and system for separating partial discharge signals of a transformer. Background Art

[0002] In the construction of a substation, a transformer is an indispensable and important hub device, which plays the role of voltage conversion and energy transmission. Its insulation performance directly affects the safe operation of the power system. During long-term operation, the transformer will be affected by various factors, resulting in equipment failures or even damages, and causing partial discharge phenomena inside or on the surface of the insulating material, which is called partial discharge. Usually, it is manifested as weak electromagnetic waves, sound waves or optical signals. By conducting partial discharge tests on the transformer, the insulating defects inside the transformer can be effectively detected, the fault problems of the transformer can be discovered in time, and effective maintenance measures can be implemented, so as to extend the service life of the equipment and ensure the safe and stable operation of the power system.

[0003] When conducting partial discharge tests on the transformer, due to the influence of on-site external pulse interferences such as the test power supply, leads, and discharges of other equipment around the laboratory, the measurement results will show the phenomenon of aliasing of interference pulses, discharge pulses of external transformer equipment, and internal discharge pulses of the transformer. In the traditional separation scheme, only the distribution state of outliers after clustering is used to judge the abnormal discharge state, so as to separate the partial discharge signals. However, in the traditional separation scheme, the outlier power supply state formed by external partial discharges cannot be distinguished from internal partial discharges, resulting in inaccurate separation results of partial discharge signals and making it difficult to effectively separate internal partial discharge signals.

[0004] Therefore, the traditional separation scheme has the technical problem that it cannot distinguish external partial discharge signals from internal partial discharge signals, resulting in low accuracy and reliability of the separation results. Summary of the Invention

[0005] In order to solve the technical problem that the traditional separation scheme cannot distinguish external partial discharge signals from internal partial discharge signals, resulting in low accuracy and reliability of the separation results, the purpose of the present invention is to provide a method and system for separating partial discharge signals of a transformer. The specific technical solutions adopted are as follows:

[0006] On the one hand, the present invention provides a method for separating partial discharge signals of a transformer, and the method includes:

[0007] Synchronously sample multiple measurement points of the transformer to be measured at different times to obtain multiple aliased test signals;

[0008] Decompose each aliased test signal to obtain multiple independent source signals in each aliased test signal;

[0009] Calculate the rising fluctuation parameters of each independent source signal at different times, and determine the target aliased test signal with partial discharge signals based on the rising fluctuation parameters of each independent source signal in each aliased test signal at different times;

[0010] Determine the feature vectors of each independent source signal in the target aliased test signal, and perform clustering processing on all the independent source signals in the target aliased test signal based on the feature vectors to obtain multiple clustering clusters;

[0011] Determine the density eigenvalue of each clustering cluster, and determine the target clustering cluster corresponding to the internal partial discharge signal based on the density eigenvalue;

[0012] Separate the internal partial discharge signals in the target aliased test signal based on the target clustering cluster.

[0013] According to a method for separating partial discharge signals of a transformer provided by the present invention, calculating the rising fluctuation parameters of each independent source signal at different times includes:

[0014] Obtain the first amplitude rising duration of each independent source signal at different times and the second amplitude rising duration of each independent source signal in the aliased test signal where each independent source signal is located at different times;

[0015] Subtract the first amplitude rising duration from each of the second amplitude rising durations respectively and take the absolute value to calculate multiple absolute differences in rising durations;

[0016] Take the mean of the multiple absolute differences in rising durations to calculate the rising fluctuation parameters of each independent source signal at different times.

[0017] According to a method for separating partial discharge signals of a transformer provided by the present invention, determining the target aliased test signal with partial discharge signals based on the rising fluctuation parameters of each independent source signal in each aliased test signal at different times includes:

[0018] Compare the rising fluctuation parameters of each independent source signal in each aliased test signal at different times with a preset fluctuation threshold to obtain a comparison result;

[0019] Determine the aliased test signal whose comparison result meets the preset determination condition as the target aliased test signal with partial discharge signals; wherein, the preset determination condition is that there is at least one comparison result that meets the condition that the rising fluctuation parameter of the independent source signal is above the preset fluctuation threshold.

[0020] According to a method for separating partial discharge signals of a transformer provided by the present invention, determining the feature vectors of each independent source signal in the target aliased test signal includes:

[0021] Determine the energy concentration parameters of each independent source signal at different times;

[0022] Calculate the attenuation exponent of each independent source signal at different times based on the rising fluctuation parameters of each independent source signal at different times and the energy concentration parameters of each independent source signal at different times;

[0023] Average the rising fluctuation parameters of each independent source signal at different times to obtain the average rising fluctuation parameter of each independent source signal;

[0024] Average the energy concentration parameters of each independent source signal at different times to obtain the average energy concentration parameter of each independent source signal;

[0025] Average the attenuation exponents of each independent source signal at different times to obtain the average attenuation exponent of each independent source signal;

[0026] Based on the average rising fluctuation parameter, average energy concentration parameter, and average attenuation exponent of each independent source signal, establish the feature vector of each independent source signal in the target aliased test signal.

[0027] According to a method for separating partial discharge signals of a transformer provided by the present invention, based on the density eigenvalue, determine the target clustering cluster corresponding to the internal partial discharge signal, including:

[0028] Compare the density eigenvalue of each clustering cluster with a preset density eigenvalue threshold respectively to obtain a comparison result;

[0029] Determine the clustering cluster with a density eigenvalue higher than the preset density eigenvalue threshold in the comparison result as the target clustering cluster corresponding to the internal partial discharge signal.

[0030] According to a method for separating partial discharge signals of a transformer provided by the present invention, after separating the internal partial discharge signal in the target aliased test signal based on the target clustering cluster, the method further includes:

[0031] Determine the signal source position of the internal partial discharge signal;

[0032] Perform regional annotation on the separated internal partial discharge signal based on the signal source position.

[0033] According to a method for separating partial discharge signals of a transformer provided by the present invention, determining the signal source position of the internal partial discharge signal includes:

[0034] Determine the position coordinates of each measurement point, the signal detection value at the current moment, the initial value of the partial discharge signal, and the signal attenuation exponent at the current moment;

[0035] Calculate the distance weight of each measurement point relative to the signal source position based on the detected value of the current moment signal, the initial value of the partial discharge signal, and the signal attenuation exponent at the current moment;

[0036] Establish a plurality of distance weight ratio equations based on the distance weight and the position coordinates of each measurement point;

[0037] Jointly solve through the plurality of distance weight ratio equations to calculate the signal source position of the internal partial discharge signal.

[0038] According to a method for separating partial discharge signals of a transformer provided by the present invention, after regionally annotating the separated internal partial discharge signals according to the signal source position, the method further includes:

[0039] Obtain the signal parameters of the separated internal partial discharge signals and the operating parameters of the transformer to be measured;

[0040] Perform discharge diagnosis on the signal source position based on the signal parameters and the operating parameters to obtain a discharge diagnosis result.

[0041] According to a method for separating partial discharge signals of a transformer provided by the present invention, after obtaining the discharge diagnosis result, the method further includes:

[0042] Generate a discharge analysis report and maintenance suggestion information based on the discharge diagnosis result and the signal source position.

[0043] On the other hand, the present invention also provides a system for separating partial discharge signals of a transformer, and the system includes:

[0044] A sampling module, configured to synchronously sample a plurality of measurement points of the transformer to be measured at different times to obtain a plurality of aliased test signals;

[0045] A decomposition module, configured to decompose each aliased test signal to obtain a plurality of independent source signals in each aliased test signal;

[0046] A calculation module, configured to calculate the rising fluctuation parameters of each independent source signal at different times, and determine the target aliased test signal with a partial discharge signal based on the rising fluctuation parameters of each independent source signal in each aliased test signal at different times;

[0047] A clustering module, configured to determine the feature vectors of each independent source signal in the target aliased test signal, and perform clustering processing on all the independent source signals in the target aliased test signal based on the feature vectors to obtain a plurality of clustering clusters;

[0048] A processing module, configured to determine the density eigenvalue of each clustering cluster, and determine the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue;

[0049] A separation module, configured to separate the internal partial discharge signal in the target aliased test signal according to the target clustering cluster.

[0050] The present invention has the following beneficial effects:

[0051] By decomposing each aliased test signal to obtain multiple independent source signals in each aliased test signal, then calculating the rising fluctuation parameters of each independent source signal at different times, and determining the target aliased test signal with local discharge signals according to the rising fluctuation parameters of each independent source signal in each aliased test signal at different times. Then, determine the feature vector of each independent source signal in the target aliased test signal, and perform clustering processing on all independent source signals in the target aliased test signal according to the feature vector to obtain multiple clustering clusters. Subsequently, determine the density eigenvalue of each clustering cluster, and determine the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue. Finally, separate the internal partial discharge signal in the target aliased test signal according to the target clustering cluster. Since the target aliased test signal with local discharge signals is located according to the rising fluctuation parameters before the clustering operation, then clustering is performed according to the feature vector of each independent source signal, and finally the target clustering cluster corresponding to the internal partial discharge signal is determined according to the density eigenvalue, the external partial discharge signal and the internal partial discharge signal can be effectively distinguished in the clustering process, the target clustering cluster where the internal partial discharge signal is located can be accurately located, and thus the accurate separation of the internal partial discharge signal in the target aliased test signal can be realized, effectively improving the accuracy and reliability of the separation result. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0053] Figure 1 It is a method flow chart of a method for separating local discharge signals of a transformer provided by an embodiment of the present invention;

[0054] Figure 2 It is a schematic diagram of the position distribution of each measurement point;

[0055] Figure 3 It is a spectral image of the theoretically local discharge signal;

[0056] Figure 4It is the spectral image of the aliasing test signal of the multi-source pulse aliasing actually detected;

[0057] Figure 5 It is the system structure diagram of a transformer partial discharge signal separation system provided by an embodiment of the present invention;

[0058] Reference numerals:

[0059] 201, ground wire of capacitive voltage divider; 202, ground wire of low-voltage winding; 203, ground wire of high-voltage winding neutral point; 204, ground wire of high-voltage bushing end screen; 310, sampling module; 320, decomposition module; 330, calculation module; 340, clustering module; 350, processing module; 360, separation module. Detailed implementation manners

[0060] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the drawings and preferred embodiments to detail the specific implementation manners, structures, features and effects of a transformer partial discharge signal separation method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0062] During the transformer partial discharge test, the pulse sources are mainly of three types: externally invasive pulses, external partial discharge pulses, and internal partial discharge pulses. That is to say, during the transformer partial discharge test, there will be a situation where externally invasive noise signals, external partial discharge signals, and internal partial discharge signals are aliased. Among them, the externally invasive noise signal usually refers to the noise signal caused by the test power supply, lead wires, and other equipment interference around the laboratory. The external partial discharge signal usually refers to the partial discharge signal caused by the discharge of external equipment of the transformer. The internal partial discharge signal usually refers to the partial discharge signal caused by the internal discharge of the transformer. Since the main purpose of this embodiment is to separate the partial discharge signal from the aliasing test signal for subsequent fault diagnosis, the partial discharge signal mentioned in this embodiment mainly refers to the internal partial discharge signal.

[0063] Due to multi-source pulse aliasing, in traditional separation schemes, the partial discharge signals are separated by directly using the distribution to judge the discharge state. However, during the separation process, since there are similar information in the time domain and frequency domain between some internal partial discharge signals and external partial discharge signals, they will be marked as outlier noise points during the clustering process and are difficult to effectively distinguish. Therefore, the traditional separation scheme has problems of low accuracy and reliability of the separation results.

[0064] Accordingly, the present invention provides a solution to the above problems. Considering the differences in the transmission process of internal and external partial discharge signals, the present invention analyzes the cluster outlier signals by continuously monitoring the propagation characteristics from the positions where different pulse signals are generated to the measurement points, and further determines the area where partial discharge occurs by analyzing the attenuation change of the influence degree of the internal partial discharge signals on different measurement points. The following specifically describes the detailed solutions of a method and system for separating partial discharge signals of a transformer provided by the present invention with reference to the accompanying drawings.

[0065] Please refer to Figure 1 , which shows a flowchart of a method for separating partial discharge signals of a transformer provided by an embodiment of the present invention. As shown in the accompanying Figure 1 drawings, the above method for separating partial discharge signals of a transformer specifically includes:

[0066] Step 110: Synchronously sample multiple measurement points of the transformer to be measured at different times to obtain multiple aliased test signals.

[0067] As Figure 2 shown, the transformer to be measured is arranged and wired. Among them, A, B, and C represent the line terminals of the high-voltage winding, and a, b, and c represent the line terminals of the low-voltage winding. In this embodiment, the ground wires 201 of the capacitive voltage divider, the ground wire 202 of the low-voltage winding, the ground wire 203 of the neutral point of the high-voltage winding, and the ground wire 204 of the end screen of the high-voltage bushing are respectively selected as the measurement points to ensure that multiple measurement points can cover the entire transformer area. High-frequency current sensors are arranged at each measurement point, and multi-segment synchronous measurement technology is used to synchronously collect the high-frequency pulse signals of the four measurement points, and the acquisition time of each high-frequency pulse signal is recorded.

[0068] In practical applications, it is necessary to filter the high-frequency pulse signals recorded at the four measurement points at the same time respectively to remove the background noise interference in the high-frequency pulse signals, and perform timestamp alignment and normalization processing on each high-frequency pulse signal to ensure the comparability of the high-frequency pulse signals at the same time between different measurement points. After filtering, timestamp alignment, and normalization processing respectively, the high-frequency pulse signals recorded at the four measurement points at the same time can be combined into an aliased test signal. The aliased test signal can be specifically expressed as , where t represents the acquisition time, m represents the number of measurement points, denotes the aliased test signal obtained at time t, denotes the high-frequency pulse signal corresponding to the first measurement point at time t, denotes the high-frequency pulse signal corresponding to the second measurement point at time t, denotes the high-frequency pulse signal corresponding to the m-th measurement point at time t. In practical applications, 1000 aliased test signals measured under different test voltages can be selected for partial discharge signal separation analysis.

[0069] Step 120: Decompose each aliased test signal to obtain multiple independent source signals in each aliased test signal.

[0070] Figure 3 and Figure 4 exemplarily show the spectral images of the partial discharge signal in theory and the spectral image of the aliased test signal with multi-source pulse aliasing actually detected, respectively. Among them, Figure 3 and Figure 4 in the shown spectral images, the abscissa is the sampling time in μs; the ordinate is the signal voltage in mV. In practical applications, the collected aliased test signals need to be separated to obtain the time-frequency performance characteristics of each separated signal. Since the internal partial discharge signal is usually directly emitted from a specific point or defect point inside the transformer, the transmission distance of the signal is short, and the time for the sensor to receive the signal is short, resulting in a fast rising speed of the signal amplitude and a short rise time; while the external partial discharge signal will experience attenuation and interference from several components on the transformer during the propagation process, and the time for the signal to reach the sensor will be extended. Therefore, in this embodiment, by combining the time-domain characteristics of the signals at different moments of a measurement point, the time-domain distribution characteristics of the signals are analyzed.

[0071] Since the collected aliased test signals contain data of various source components, and the data of each source component are independent of each other, the independent component analysis method can be applied to decompose the collected aliased test signals. Specifically, the independent component analysis method can be used to decompose a group of aliased test signals to obtain several separated signals, and perform normalization processing on each separated signal to obtain the independent source signals after normalization processing.

[0072] Step 130: Calculate the rising fluctuation parameters of each independent source signal at different moments, and determine the target aliased test signal with a partial discharge signal according to the rising fluctuation parameters of each independent source signal in each aliased test signal at different moments.

[0073] It can be understood that since the main purpose of signal separation is to separate the internal partial discharge signal, and further provide data support for the partial discharge state diagnosis of the transformer to be tested, therefore, the partial discharge signal mentioned in this embodiment mainly refers to the internal partial discharge signal generated inside the transformer.

[0074] Step 140: Determine the eigenvectors of each independent source signal in the target aliased test signal, and cluster all the independent source signals in the target aliased test signal according to the eigenvectors to obtain multiple clusters.

[0075] It can be understood that in this embodiment, the eigenvector of each independent source signal includes the rising fluctuation parameter, the mean value of the energy concentration parameter, and the mean value of the attenuation index of each independent source signal, which can comprehensively characterize the propagation characteristics of each independent source signal from the generation position to the measurement point.

[0076] Step 150: Determine the density eigenvalue of each cluster, and determine the target cluster corresponding to the internal partial discharge signal according to the density eigenvalue.

[0077] It can be understood that since the internal partial discharge signals are concentrated in a relatively small internal space of the transformer and the frequency and amplitude characteristics are relatively consistent, the clustering result shows a high-density characteristic. Therefore, in this embodiment, the target cluster corresponding to the internal partial discharge signal can be determined through the density eigenvalue.

[0078] Step 160: Separate the internal partial discharge signals in the target aliased test signal according to the target cluster.

[0079] The solution provided in this embodiment improves the clustering process. By continuously monitoring the propagation characteristics of different pulse signals from the generation position to the measurement point and analyzing the outlier signals in the clusters, the target cluster corresponding to the internal partial discharge signal can be accurately located. Furthermore, the internal partial discharge signals in the target aliased test signal can be accurately separated according to the target cluster, improving the accuracy and reliability of the separation result.

[0080] In one embodiment, calculating the rising fluctuation parameter of each independent source signal at different times specifically includes:

[0081] First step, obtain the first amplitude rising duration of each independent source signal at different times and the second amplitude rising duration of each independent source signal in the aliased test signal where each independent source signal is located at different times.

[0082] Second step, subtract the first amplitude rising duration from each of the second amplitude rising durations and take the absolute value to calculate a plurality of absolute differences in rising durations.

[0083] Third step, calculate the mean value of the plurality of absolute differences in rising durations to obtain the rising fluctuation parameter of each independent source signal at different times.

[0084] It can be understood that the amplitude of the independent source signal can characterize the signal intensity of the signal source at the current moment. Since partial discharge is a process of releasing energy, the amplitude of the internal partial discharge signal will increase rapidly within a short period of time. Calculate the rising duration of the amplitude of the current independent source signal. For example , represents the duration of the process from the amplitude reaching 10% to the amplitude reaching 90% in the j-th independent source signal, reflecting the change speed of the signal represents the time taken for the amplitude of the j-th independent source signal to reach 90% represents the time taken for the amplitude of the j-th independent source signal to reach 10%

[0085] Internal partial discharge signals usually exhibit large transient fluctuations and instability, and will increase rapidly within a short period of time, while external partial discharge signals have a long duration and a slow signal change period. Therefore, there are often large fluctuations between the rising duration of the internal partial discharge signal and the external partial discharge signal. For the aliased test signal at the same moment, by calculating the volatility of the rising duration of the amplitudes of different independent source signals, the change speed of the independent source signal at different measurement points can be identified, and the rising fluctuation parameters of the current independent source signal can be analyzed

[0086] In this embodiment, the rising fluctuation parameter of the j-th independent source signal at time t can be expressed as follows

[0087] (1)

[0088] Where represents the rising fluctuation parameter of the j-th independent source signal at time t represents the first rising duration of the j-th independent source signal at time t represents the second rising duration of the r-th independent source signal at time t in the aliased test signal where the j-th independent source signal is located n represents the number of independent source signals in the aliased test signal where the j-th independent source signal is located

[0089] It can be understood that both the first rising duration and the second rising duration refer to the duration of the signal amplitude changing from 10% to 90%. The larger the absolute difference in the rising duration, the faster the change speed of the current independent source signal. Since the relative change frequency of the internal partial discharge signal is relatively fast, when the value of is larger, the greater the possibility that the aliased test signal contains an internal partial discharge signal

[0090] In one embodiment, based on the rising fluctuation parameters of each independent source signal in each aliased test signal at different times, the target aliased test signal with a partial discharge signal is determined, which specifically includes

[0091] First, compare the rising fluctuation parameters of each independent source signal in each aliasing test signal at different times with a preset fluctuation threshold to obtain a comparison result.

[0092] Then, determine the target aliasing test signal with a partial discharge signal among the aliasing test signals whose comparison results meet the preset determination conditions; wherein, the preset determination condition is that there is at least one comparison result satisfying that the rising fluctuation parameter of the independent source signal is above the preset fluctuation threshold.

[0093] In this embodiment, the preset fluctuation threshold can take a value of 0.5. When there is a situation where the rising fluctuation parameter of several independent source signals obtained after decomposing a certain aliasing test signal is greater than or equal to the preset fluctuation threshold, it indicates that there is an internal partial discharge signal in the aliasing signal.

[0094] In one embodiment, determining the feature vector of each independent source signal in the target aliasing test signal specifically includes:

[0095] The first step is to determine the energy concentration parameter of each independent source signal at different times.

[0096] Since different independent source signals exhibit different signal characteristics in the frequency domain, spectral feature analysis can be performed based on the frequency domain information of several independent source signals.

[0097] Specifically, perform a fast Fourier transform on the independent source signal to obtain the spectral data of the independent source signal. The specific formula is as follows:

[0098] (2)

[0099] Wherein, represents the vibration energy of the i-th vibration frequency at time t in the j-th independent source signal; represents the spectrum of the i-th vibration frequency at time t in the j-th independent source signal.

[0100] Since the source of the external partial discharge signal is complex and the frequency components of the signal are relatively wide, it may contain both high-frequency and low-frequency information at the same time; while the internal partial discharge signal comes from the partial discharge caused by the discharge defect of the equipment, and the frequency components are related to the discharge type and medium, usually concentrated in a specific frequency band. Therefore, spectral feature analysis based on the frequency domain information of the independent source signal can calculate the energy concentration parameters of each independent source signal at different times.

[0101] In this embodiment, the energy concentration parameter of the j-th independent source signal at time t can be expressed as follows:

[0102] (3)

[0103] Wherein, represents the energy concentration parameter of the j-th independent source signal at time t; nj represents the number of different vibration frequencies in the j-th independent source signal; represents the vibration energy of the i-th vibration frequency in the j-th independent source signal at time t; represents the vibration energy of the i-th vibration frequency in the s-th independent source signal other than the j-th independent source signal at the current measurement point at time t; represents the contribution degree of each vibration frequency in the j-th independent source signal to the current independent source signal at time t; represents the sum of the energies of all vibration frequencies in the j-th independent source signal at time t, represents the sum of the energies of all vibration frequencies of all independent source signals at the current measurement point at time t, [ ] represents the normalization function, exp ( ) represents the exponential function with the natural constant e as the base.

[0104] In the second step, according to the rising fluctuation parameter of each independent source signal at different times and the energy concentration parameter of each independent source signal at different times, the attenuation exponent of each independent source signal at different times is calculated.

[0105] It can be understood that there are significant changes in the signal degrees of the attenuation characteristics of independent source signals from different sources during the transmission process. When the discharge source is located inside the transformer (such as inside the winding, near the iron core, or in the oil-paper insulation part, etc.), the signal energy is significantly weakened, and the internal partial discharge signal will be attenuated and distorted to a large extent during the transmission process; while the external partial discharge signal, due to transmission through external connections or air, contains less attenuation of high-frequency components, and the signal waveform remains relatively complete. Therefore, the attenuation characteristics of the signal can be obtained by analyzing the variation law of the signal. Specifically, the attenuation exponent of the j-th independent source signal at time t can be expressed as follows:

[0106] (4)

[0107] Among them, represents the attenuation exponent of the j-th independent source signal at time t; represents the rising fluctuation parameter of the j-th independent source signal at time t; represents at the energy concentration parameter of the j-th independent source signal detected at time; represents at the energy concentration parameter of the j-th independent source signal detected at time; represents the time interval (in seconds) between the reference time and the corresponding time , and the preset reference time is within five seconds after the target time; It represents the difference in the frequency-domain characteristic factors of the j-th independent source signal before and after the reference moment, reflecting the degree of change in the signal's frequency-domain characteristics over time; It represents the attenuation degree of the signal over time; as the time interval continues to increase, the attenuation of the signal becomes more and more obvious, indicating that the current independent source signal belongs to the internal partial discharge signal to a greater extent.

[0108] In the third step, the rising fluctuation parameters of each independent source signal at different moments are averaged to obtain the average rising fluctuation parameter of each independent source signal.

[0109] In the fourth step, the energy concentration parameters of each independent source signal at different moments are averaged to obtain the average energy concentration parameter of each independent source signal.

[0110] In the fifth step, the attenuation exponents of each independent source signal at different moments are averaged to obtain the average attenuation exponent of each independent source signal.

[0111] In the sixth step, based on the average rising fluctuation parameter, the average energy concentration parameter, and the average attenuation exponent of each independent source signal, a feature vector of each independent source signal in the target aliased test signal is established.

[0112] Taking the j-th independent source signal as an example, a feature vector can be constructed , where represents the average energy concentration parameter of the j-th independent source signal, represents the average rising fluctuation parameter of the j-th independent source signal, represents the average attenuation exponent of the j-th independent source signal.

[0113] It can be understood that the internal partial discharge signal has a high pulse frequency and significant amplitude variations, and the change period is relatively stable; while the external pulse noise has an irregular shape, random change period and frequency spectrum distribution, and uniform energy distribution. When directly clustering and analyzing several independent source signals using time-domain and frequency-domain information, due to the similar information of some internal partial discharge signals and external partial discharge signals in the time domain and frequency domain, they will be marked as outlier noise points during the clustering process and are difficult to effectively distinguish. However, the information change stability of the internal partial discharge signal and the external partial discharge signal is different. Therefore, in this embodiment, a feature vector is constructed in combination with the signal change law of the independent source signal, and then the outlier noise points are weighted.

[0114] In practical applications, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm can be used to classify according to the feature vectors of independent source signals, obtaining several clustering clusters, and each clustering cluster represents a type of pulse signal with similar characteristics.

[0115] In one embodiment, according to the density eigenvalue, determining the target clustering cluster corresponding to the internal partial discharge signal specifically includes:

[0116] First, compare the density eigenvalue of each clustering cluster with a preset density eigenvalue threshold respectively to obtain a comparison result.

[0117] Then, determine the clustering cluster with a comparison result that the density eigenvalue is higher than the preset density eigenvalue threshold as the target clustering cluster corresponding to the internal partial discharge signal.

[0118] It can be understood that since the internal partial discharge signals are concentrated in a relatively small internal space of the transformer and the frequency and amplitude characteristics are relatively consistent, the clustering result of the DBSCAN algorithm shows a high-density area. Therefore, the density eigenvalue of the clustering cluster can be calculated , and a preset density eigenvalue threshold is set for screening. In practical applications, the preset density eigenvalue threshold can take a value of T = 0.75. When the density eigenvalue of the clustering cluster is such, it can be determined that this clustering cluster is the target clustering cluster with internal partial discharge signals.

[0119] In one embodiment, after separating the internal partial discharge signals in the target aliasing test signal according to the target clustering cluster, the above transformer partial discharge signal separation method may further include:

[0120] First, determine the signal source position of the internal partial discharge signal.

[0121] In a specific implementation, determining the signal source position of the internal partial discharge signal specifically includes:

[0122] The first step is to determine the position coordinates of each measurement point, the signal detection value at the current moment, the initial value of the partial discharge signal, and the signal attenuation index at the current moment.

[0123] It is understandable that the intensity and characteristics of internal partial discharge signals are closely related to the signal source location. When internal partial discharge signals propagate in the transformer under test, they will be affected by the medium, and there is a large attenuation characteristic during the transmission process. External partial discharge signals are pulse signals caused by external lines or air. In the partial discharge test, due to different collection times of the measuring point sensors and different signal propagation paths, the receiving effects of the sensors will also vary. Therefore, the signal amplitude ratio information in different clustering clusters can be used to narrow down the positioning range of the signal source and achieve precise positioning of the signal source of partial discharge in the transformer under test.

[0124] In this embodiment, in the two-dimensional plane image of the transformer under test, the position coordinates of m measuring points can be determined, that is , ,… 。

[0125] In the second step, according to the signal detection value at the current moment, the initial value of the partial discharge signal, and the signal attenuation exponent at the current moment, the distance weight of each measuring point relative to the signal source position is calculated.

[0126] In this embodiment, using the signal amplitude ratio information of the signal attenuation characteristic, the distance weight from the signal source to each measuring point can be deduced, and the signal source position can be further determined by using the triangulation method. Specifically, the distance weight of the mth measuring point relative to the signal source position can be expressed as follows:

[0127] (5)

[0128] Where represents the distance weight of the mth measuring point relative to the signal source position, represents the initial value of the partial discharge signal, represents the signal detection value of the mth measuring point at the current moment, represents the signal attenuation exponent of the mth measuring point at the current moment.

[0129] In the third step, based on the distance weight and the position coordinates of each measuring point, multiple distance weight ratio equations are established.

[0130] In the fourth step, through the joint solution of multiple distance weight ratio equations, the signal source position of the internal partial discharge signal is calculated.

[0131] In this embodiment, multiple distance weight ratio equations can be expressed as follows:

[0132] (6)

[0133] (7)

[0134] (8)

[0135] Among them, , , , … represent the position coordinates of each measurement point, represents the position of the signal source, , , , represent the distance weights of each measurement point relative to the position of the signal source.

[0136] By jointly solving the above equations, the position of the signal source of the internal partial discharge signal can be obtained .

[0137] Then, based on the position of the signal source, regional annotation is performed on the separated internal partial discharge signal.

[0138] In the regional annotation step, the position of the signal source of the internal partial discharge signal can be marked on the two-dimensional plane image of the transformer under test.

[0139] In one embodiment, after performing regional annotation on the separated internal partial discharge signal according to the position of the signal source, the above method for separating transformer partial discharge signals may further include:

[0140] First, obtain the signal parameters of the separated internal partial discharge signal and the operating parameters of the transformer under test.

[0141] In this embodiment, the signal parameters specifically include parameters such as signal amplitude and signal frequency that can characterize the signal attributes, and the operating parameters of the transformer under test specifically include parameters such as the operating voltage and load size of the transformer under test that affect partial discharge.

[0142] Then, based on the signal parameters and operating parameters, discharge diagnosis is performed on the position of the signal source to obtain a discharge diagnosis result.

[0143] In practical applications, the discharge degree can be determined based on the signal parameters of the internal partial discharge signal, and the operating state of the transformer under test can be determined based on the operating parameters of the transformer under test. The severity of the discharge is comprehensively evaluated through the discharge degree and the operating state to obtain a discharge diagnosis result.

[0144] For example, the proportion of high-frequency components can be determined according to the signal frequency in the signal parameters of the internal partial discharge signal. After normalizing the proportion of high-frequency components, it is used as the discharge degree.

[0145] In some embodiments, multiple discharge severity levels can be pre-divided according to the discharge degree and operating status. For example, if the discharge degree is in the first interval and the operating status is normal, the discharge severity level is low, and the potential safety hazard is relatively small at this time. If the discharge degree is in the second interval and the operating status is abnormal, the discharge severity level is medium, and the potential safety hazard is average at this time. If the discharge degree is in the third interval and the operating status is abnormal, the discharge severity level is high, and the potential safety hazard is relatively large at this time.

[0146] Subsequently, the discharge degree and operating status are respectively compared with the reference intervals at each level, and then the discharge severity level of the transformer to be tested is determined, and the determined discharge severity level is used as the discharge diagnosis result.

[0147] In one embodiment, after obtaining the discharge diagnosis result, the above-mentioned transformer partial discharge signal separation method may further include:

[0148] Generating a discharge analysis report and maintenance suggestion information based on the discharge diagnosis result and the signal source location.

[0149] In this embodiment, the discharge diagnosis result and the signal source location can be summarized, and a discharge analysis report can be generated in the form of a table. At the same time, corresponding maintenance suggestion information can be generated according to the discharge severity level in the discharge diagnosis result. For example, when the discharge severity level is high, the corresponding maintenance suggestion information is shutdown diagnosis; when the discharge severity level is low or medium, the corresponding maintenance suggestion information is continuous monitoring. Subsequently, the discharge analysis report and maintenance suggestion information can be sent to the terminal of the supervisor, so that the supervisor can timely understand the discharge situation of the transformer to be tested and can perform maintenance or monitoring in a timely manner in combination with the maintenance suggestion information.

[0150] Based on the same inventive concept, the present invention also protects a transformer partial discharge signal separation system. The transformer partial discharge signal separation system provided by the present invention is described below, and the transformer partial discharge signal separation system described below can be correspondingly referred to the transformer partial discharge signal separation method described above.

[0151] As Figure 5 shown, the embodiment of the present invention also provides a transformer partial discharge signal separation system, which specifically includes:

[0152] A sampling module 310, configured to synchronously sample multiple measurement points of the transformer to be tested at different times to obtain multiple aliased test signals.

[0153] A decomposition module 320, configured to decompose each aliased test signal to obtain multiple independent source signals in each aliased test signal.

[0154] The calculation module 330 is configured to calculate the rising fluctuation parameters of each independent source signal, and determine the target aliased test signal with partial discharge signals according to the rising fluctuation parameters of each independent source signal in each aliased test signal.

[0155] The clustering module 340 is configured to determine the feature vectors of each independent source signal in the target aliased test signal, and perform clustering processing on all the independent source signals in the target aliased test signal according to the feature vectors to obtain multiple clustering clusters.

[0156] The processing module 350 is configured to determine the density eigenvalue of each clustering cluster, and determine the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue.

[0157] The separation module 360 is configured to separate the internal partial discharge signals in the target aliased test signal according to the target clustering cluster.

[0158] It can be seen that the transformer partial discharge signal separation system provided by the embodiment of the present invention can effectively distinguish internal partial discharge signals and external partial discharge signals by combining the signal attenuation characteristics of multiple measurement points, and accurately locate the signal source position of the internal partial discharge signal through the analysis of the propagation path, which can provide more accurate information for subsequent fault diagnosis.

[0159] Regarding the system in the above embodiment, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0160] It should be noted that the above sequence of the embodiments of the present invention is only for description, and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0161] Each embodiment in this specification is described in a progressive manner, and the same or similar parts among the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for separating partial discharge signals of a transformer, characterized in that, The method includes: Synchronously sampling multiple measurement points of the transformer under test at different times to obtain multiple aliased test signals; Decomposing each aliased test signal to obtain multiple independent source signals in each aliased test signal; Calculating the rising fluctuation parameters of each independent source signal at different times, and determining the target aliased test signal with partial discharge signals according to the rising fluctuation parameters of each independent source signal in each aliased test signal at different times; Determining the feature vectors of each independent source signal in the target aliased test signal, and clustering all the independent source signals in the target aliased test signal according to the feature vectors to obtain multiple clustering clusters; Determining the density eigenvalue of each clustering cluster, and determining the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue; Separating the internal partial discharge signals in the target aliased test signal according to the target clustering cluster; Calculating the rising fluctuation parameters of each independent source signal at different times, including: Obtaining the first amplitude rising duration of each independent source signal at different times and the second amplitude rising duration of each independent source signal in each aliased test signal at different times; Taking the difference between the first amplitude rising duration and each of the second amplitude rising durations and taking the absolute value to calculate multiple absolute differences in rising durations; Calculating the mean of the multiple absolute differences in rising durations to obtain the rising fluctuation parameter of each independent source signal at different times; Determining the feature vectors of each independent source signal in the target aliased test signal, including: Determining the energy concentration parameter of each independent source signal at different times; Calculating the attenuation index of each independent source signal at different times according to the rising fluctuation parameter of each independent source signal at different times and the energy concentration parameter of each independent source signal at different times; Taking the mean of the rising fluctuation parameters of each independent source signal at different times to obtain the mean rising fluctuation parameter of each independent source signal; Taking the mean of the energy concentration parameters of each independent source signal at different times to obtain the mean energy concentration parameter of each independent source signal; Taking the mean of the attenuation indices of each independent source signal at different times to obtain the mean attenuation index of each independent source signal; Establishing the feature vector of each independent source signal in the target aliased test signal according to the mean rising fluctuation parameter, mean energy concentration parameter and mean attenuation index of each independent source signal; Performing a fast Fourier transform on the independent source signal to obtain the spectrum data of the independent source signal. The specific formula is as follows: ； Among them, represents the vibration energy of the i-th vibration frequency in the j-th independent source signal at time t; represents the frequency spectrum of the i-th vibration frequency in the j-th independent source signal at time t; The energy concentration parameter of the jth independent source signal at time t is expressed as follows: ； Among them, represents the energy concentration parameter of the j-th independent source signal at time t; nj represents the number of different vibration frequencies in the j-th independent source signal; represents the vibration energy of the i-th vibration frequency in the j-th independent source signal at time t; represents the vibration energy of the i-th vibration frequency in the s-th independent source signal other than the j-th independent source signal at the current measurement point at time t; represents the contribution degree of each vibration frequency in the j-th independent source signal to the current independent source signal at time t; represents the sum of the energies of all vibration frequencies in the j-th independent source signal at time t, represents the sum of the energies of all vibration frequencies of all independent source signals at the current measurement point at time t, [ ] represents the normalization function, exp ( ) represents the exponential function with the natural constant e as the base; The attenuation index of the jth independent source signal at time t is expressed as follows: ； Among them, represents the attenuation exponent of the j-th independent source signal at time t; represents the rising fluctuation parameter of the j-th independent source signal at time t; represents at the energy concentration parameter of the j-th independent source signal detected at time; represents at the energy concentration parameter of the j-th independent source signal detected at time; represents the time interval between the reference time and the corresponding time , and the preset reference time is within five seconds after the target time; represents the difference in the frequency domain characteristic factors of the j-th independent source signal before and after the reference time, reflecting the degree of change in the signal frequency domain characteristics over time; represents the attenuation degree of the signal over time.

2. The method for separating partial discharge signals of a transformer according to claim 1, wherein Determining the target aliased test signal with partial discharge signals according to the rising fluctuation parameters of each independent source signal in each aliased test signal at different times, including: Comparing the rising fluctuation parameters of each independent source signal in each aliased test signal at different times with a preset fluctuation threshold to obtain a comparison result; Determine the aliasing test signal whose comparison result meets the preset determination condition as the target aliasing test signal with a partial discharge signal; wherein, the preset determination condition is that there is at least one comparison result such that the rising fluctuation parameter of the independent source signal is above the preset fluctuation threshold.

3. A method for separating partial discharge signals of a transformer according to claim 1, characterized in that, Determine the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue, including: Compare the density eigenvalue of each clustering cluster with the preset density eigenvalue threshold respectively to obtain a comparison result; Determine the clustering cluster with a comparison result that the density eigenvalue is higher than the preset density eigenvalue threshold as the target clustering cluster corresponding to the internal partial discharge signal.

4. A method for separating partial discharge signals of a transformer according to claim 1, characterized in that After separating the internal partial discharge signal in the target aliasing test signal according to the target clustering cluster, the method further includes: Determine the signal source position of the internal partial discharge signal; Perform area annotation on the separated internal partial discharge signal according to the signal source position.

5. A method for separating partial discharge signals of a transformer according to claim 4, characterized in that, Determine the signal source position of the internal partial discharge signal, including: Determine the position coordinates of each measurement point, the signal detection value at the current moment, the initial value of the partial discharge signal, and the signal attenuation index at the current moment; Calculate the distance weight of each measurement point relative to the signal source position according to the signal detection value at the current moment, the initial value of the partial discharge signal, and the signal attenuation index at the current moment; Establish multiple distance weight ratio equations according to the distance weight and the position coordinates of each measurement point; Solve the equations jointly through the multiple distance weight ratio equations to calculate the signal source position of the internal partial discharge signal.

6. A method for separating partial discharge signals of a transformer according to claim 4, characterized in that After performing area annotation on the separated internal partial discharge signal according to the signal source position, the method further includes: Obtain the signal parameters of the separated internal partial discharge signal and the operating parameters of the transformer under test; Perform discharge diagnosis on the signal source position according to the signal parameters and the operating parameters to obtain a discharge diagnosis result.

7. A method for separating partial discharge signals of a transformer according to claim 6, characterized in that After obtaining the discharge diagnosis result, the method further includes: Generate a discharge analysis report and maintenance suggestion information according to the discharge diagnosis result and the signal source position.

8. A partial discharge signal separation system for a transformer, characterized in that, The system is used to execute the steps of a method for separating partial discharge signals of a transformer as described in any one of claims 1-7, including: A sampling module for synchronously sampling multiple measurement points of the transformer under test at different times to obtain multiple aliasing test signals; A decomposition module for decomposing each aliasing test signal to obtain multiple independent source signals in each aliasing test signal; A calculation module for calculating the rising fluctuation parameter of each independent source signal at different times, and determining the target aliasing test signal with a partial discharge signal according to the rising fluctuation parameters of each independent source signal in each aliasing test signal at different times; A clustering module for determining the feature vector of each independent source signal in the target aliasing test signal, and performing clustering processing on all the independent source signals in the target aliasing test signal according to the feature vector to obtain multiple clustering clusters; A processing module for determining the density eigenvalue of each clustering cluster, and determining the target clustering cluster corresponding to the internal partial discharge signal according to the density eigenvalue; A separation module, configured to separate the internal partial discharge signals in the target aliased test signal according to the target clustering cluster.

Images