Automatic diagnostic methods, systems, devices, and storage media for GIS noise issues

Abstract

This invention relates to an automatic diagnostic method, system, device, and storage medium for abnormal noises in GIS equipment. The method includes the following steps: Step S1, collecting and analyzing acoustic parameters of various abnormal noises from GIS equipment to obtain multiple acoustic characteristic values ​​for each type of abnormal noise; Step S2, determining the range of each acoustic characteristic value for each type of abnormal noise based on the acoustic characteristic value data; Step S3, scoring and ranking the importance of each acoustic characteristic value based on the degree of difference between each acoustic characteristic value range and the abnormal noise of the GIS equipment; Step S4, establishing a fault diagnosis decision tree model based on the importance ranking of the acoustic characteristic values; Step S5, using the fault diagnosis decision tree model to diagnose the actual acoustic parameters of abnormal noises in the GIS equipment and outputting the diagnostic results of the abnormal noises in the GIS equipment. Compared with the prior art, this invention has the advantages of avoiding human hearing errors, avoiding the limitations of human experience, and automatically and accurately determining the type of abnormal noise in GIS equipment.

Description

Technical Field

[0001] This invention relates to the field of GIS equipment noise diagnosis technology, and in particular to an automatic diagnostic method, system, device and storage medium for GIS noise. Background Technology

[0002] Gas-insulated metal-enclosed switchgear (GIS) has been widely used in power systems both domestically and internationally due to its advantages such as high voltage, high current, compact structure, flexible layout, stable operation, long service life, superior technical specifications, and immunity to external influences. However, GIS malfunctions often produce abnormal noises. If these faults are not detected and resolved promptly, they can lead to serious safety accidents.

[0003] In existing technologies, the diagnosis of abnormal noises in GIS systems often relies on the hearing ability of experienced maintenance personnel. This method requires a high level of experience from the personnel, has significant limitations, and is not accurate enough, especially in the case of background noise or multiple abnormal noises, where it often cannot accurately determine the type of fault.

[0004] A search revealed Chinese invention patent publication number CN113919389A, which discloses a GIS fault diagnosis method and system based on acoustic signature imaging. The method includes: S1, acquiring acoustic signals radiated by the GIS; S2, performing time-frequency analysis and feature extraction on the acoustic signals acquired in step S1 to locate the abnormal sound source; S3, performing EMD decomposition on the acquired acoustic signals and calculating the one-dimensional spectral entropy of the decomposed IMF components, using these as training samples in the SVM model module to obtain the fault type; and S4, visualizing the fault diagnosis results obtained in steps S2 and S3, and simultaneously storing and providing alarm notifications. This existing patent has the problem of not disclosing specific diagnostic types.

[0005] How to accurately diagnose abnormal noises in GIS has become a technical problem that needs to be solved. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an automatic diagnostic method, system, device and storage medium for GIS abnormal noise.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] According to one aspect of the present invention, an automatic diagnostic method for abnormal noises in GIS is provided, the method comprising the following steps:

[0009] Step S1: Collect and analyze the acoustic parameters of various abnormal noises from GIS equipment to obtain multiple acoustic characteristic values ​​of various abnormal noises from GIS equipment.

[0010] Step S2: Based on the acoustic characteristic value data, determine the range of each acoustic characteristic value for various abnormal noises from the GIS equipment;

[0011] Step S3: Based on the degree of difference between each acoustic feature value range and the abnormal noise of the GIS equipment, score and rank the importance of the acoustic feature values;

[0012] Step S4: Establish a fault diagnosis decision tree model based on the importance ranking of acoustic feature values;

[0013] Step S5: Use the fault diagnosis decision tree model to diagnose the actual acoustic parameters of abnormal noises in the GIS equipment and output the diagnostic results of the abnormal noises in the GIS equipment.

[0014] Preferably, the various abnormal noises of the GIS equipment include noises from a loose shield and noises from loose anchor bolts.

[0015] Preferably, the acoustic characteristic values ​​include harmonic characteristic values, time-domain characteristic values, and frequency-domain characteristic values;

[0016] The harmonic characteristic values ​​mentioned include odd harmonic ratio, dominant frequency, and fundamental frequency ratio;

[0017] The time-domain feature values ​​include skewness;

[0018] The frequency domain eigenvalues ​​include spectral kurtosis, spectral entropy, and spectral flux.

[0019] Preferably, the larger the range of acoustic feature values, the greater the importance of the acoustic feature value, and the higher the corresponding importance score;

[0020] The scoring and ranking of the importance of acoustic feature values ​​is as follows: the acoustic feature values ​​are ranked from largest to smallest, and the top n feature values ​​are obtained in sequence, including spectral kurtosis, spectral entropy, fundamental frequency ratio, spectral flux, dominant frequency, odd harmonic ratio, and skewness.

[0021] Preferably, in step S4, the fault diagnosis decision tree model includes a diagnostic module for loose anchor bolts, a diagnostic module for loose shielding, and a diagnostic module for normal conditions.

[0022] The specific process of the diagnostic module for loose anchor bolts is as follows:

[0023] a1) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, the spectral entropy is less than or equal to the spectral entropy setting value, and the spectral flux is less than or equal to the spectral flux setting value, then the defect is diagnosed as loose anchor bolts.

[0024] Or a2) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, the spectral entropy is greater than the spectral entropy setting value, and the main frequency is greater than or equal to the main frequency setting value, then the defect is diagnosed as loose anchor bolts.

[0025] The specific process of the diagnostic module for the loose shield is as follows:

[0026] b1) If the following conditions are met simultaneously: spectral kurtosis is less than or equal to the spectral kurtosis setting value, spectral entropy is less than or equal to the spectral entropy setting value, spectral flux is less than or equal to the spectral flux setting value, and skewness is less than or equal to the skewness setting value, then the shielding cover is diagnosed as loose.

[0027] Or b2) If both the spectral kurtosis is less than or equal to the spectral kurtosis setting value and the fundamental frequency proportion is greater than the fundamental frequency proportion setting value, then the diagnosis is that the shielding cover is loose;

[0028] The specific process of the diagnostic module in the normal state is as follows:

[0029] If the following conditions are met simultaneously: spectral kurtosis is less than or equal to the spectral kurtosis setting value, fundamental frequency ratio is less than or equal to the fundamental frequency ratio setting value, and odd harmonic ratio is greater than the odd harmonic ratio setting value, then the diagnosis is normal.

[0030] Preferably, in step S5, the collected acoustic parameters of abnormal noise from the real GIS equipment are first filtered to remove background noise interference, and then analyzed to obtain multiple acoustic feature values.

[0031] Preferably, the diagnostic method further includes, when multiple GIS devices have abnormal noises, collecting real acoustic parameters of the abnormal noises of the GIS devices sequentially by frequency band, and then performing the processing in step S5.

[0032] According to another aspect of the present invention, an automatic diagnostic system for abnormal noises in GIS is provided. The system includes a host computer and a signal acquisition module, wherein the host computer includes a signal processing submodule and a storage submodule.

[0033] The signal acquisition module is used to collect the acoustic parameters of abnormal noises from GIS equipment and send them to the signal processing submodule for processing.

[0034] The storage submodule includes a database of acoustic characteristic values ​​for storing various abnormal noises from GIS equipment.

[0035] The signal processing submodule is used to analyze the acoustic parameters of abnormal noises of GIS equipment received from the signal acquisition module, obtain the acoustic characteristic values ​​of abnormal noises of GIS equipment, and store them in the storage word module 11.

[0036] The signal processing submodule is also used to analyze the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment, and to score and rank the importance of each acoustic feature value according to the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment;

[0037] The signal processing submodule is also used to establish a fault diagnosis decision tree model based on the sorting of each acoustic feature value;

[0038] The signal processing submodule is also used to output the results of the fault diagnosis decision tree model's analysis and diagnosis of the acoustic characteristic values ​​of the actual GIS equipment abnormal noise acoustic parameters.

[0039] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described thereon.

[0040] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] This invention is based on a large amount of acoustic feature data of GIS abnormal noises. By analyzing the range of each acoustic feature value of each type of abnormal noise in GIS equipment and the degree of difference between the abnormal noises, the importance of each acoustic feature value is scored and ranked to establish a fault diagnosis decision tree model. Then, the fault diagnosis decision tree model automatically analyzes the input real acoustic parameters of GIS abnormal noises and outputs the diagnosis results of GIS abnormal noises. This can avoid the listening error of human ears and the limitation of human experience, and can automatically and accurately determine the type of GIS abnormal noise. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the automatic diagnosis method for abnormal noises in GIS according to the present invention;

[0044] Figure 2 This is a schematic diagram of the automatic diagnostic system architecture for GIS abnormal noises in this invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0046] Example 1

[0047] This embodiment relates to an automatic diagnostic method for abnormal noises in GIS equipment. It collects a large amount of acoustic parameters of various types of abnormal noises from GIS equipment to establish a database of acoustic characteristic values ​​for GIS equipment abnormal noises. By analyzing the acoustic characteristic value data in the database, it derives the range of each acoustic characteristic value for each type of abnormal noise in the GIS equipment. Then, based on the correlation between the range of each acoustic characteristic value for each type of abnormal noise and the abnormal noise itself, it scores and ranks the importance of each acoustic characteristic value, thereby establishing a fault diagnosis decision tree model based on the ranking of each acoustic characteristic value. During GIS abnormal noise diagnosis, the collected acoustic parameters of the actual abnormal noises from the GIS equipment are analyzed to obtain multiple acoustic characteristic values, which are then input into the established fault diagnosis decision tree model for analysis, and the diagnostic results are output.

[0048] The diagnostic method in this embodiment is based on a large amount of acoustic feature data of GIS abnormal noises to diagnose the type of GIS abnormal noises. It can avoid the listening error of the human ear and the experience limitations of manual listening, and can accurately determine the type of GIS abnormal noises.

[0049] like Figure 1 The method includes the following steps:

[0050] Step S1: Collect and analyze the acoustic parameters of various abnormal noises of GIS equipment, obtain multiple acoustic characteristic values ​​of various abnormal noises of GIS equipment, and establish a database of acoustic characteristic values ​​of abnormal noises of GIS equipment based on the acoustic characteristic values ​​of abnormal noises of GIS equipment.

[0051] Step S2: Based on the acoustic characteristic value data of abnormal noises from GIS equipment, determine the range of each acoustic characteristic value for each type of abnormal noise from the GIS equipment.

[0052] Step S3: Based on the range of each acoustic feature value for each type of abnormal noise from the GIS equipment and the degree of difference between the abnormal noises from the GIS equipment, score and rank the importance of each acoustic feature value;

[0053] Step S4: Establish a fault diagnosis decision tree model based on the sorting of each acoustic feature value;

[0054] Step S5: Collect and analyze the actual acoustic parameters of abnormal noise from the GIS equipment to obtain multiple acoustic feature values. Input these values ​​into the fault diagnosis decision tree model established in step S4 for analysis and output the diagnostic results of the abnormal noise from the GIS equipment.

[0055] In step S1, the various abnormal noises of the GIS equipment refer to typical abnormal noise conditions of the GIS equipment, including abnormal noises from loose shielding covers and loose anchor bolts. The abnormal noises under different operating conditions have certain characteristics.

[0056] The loosening and abnormal noise of the shielding cover is caused by the electric field force between the shielding cover cylinder and the conductor inside. According to the principle of electric field force, the high-voltage conductor covered by the shielding cover can be regarded as the high potential plate of a capacitor, and the shielding cover cylinder can be regarded as the low potential plate of a capacitor. When the shielding cover becomes loose, a 300Hz harmonic will eventually appear due to the change in electric field force. The characteristics of the abnormal noise caused by the loose shielding cover are as follows: when the shielding cover is loose, the vibration amplitude of GIS is significantly increased compared to the normal state, with the maximum vibration amplitude increasing from 1.8mg to 10mg; under normal operating conditions, the maximum vibration frequency of GIS is 100Hz, while the maximum vibration frequency on the conductor is 300Hz when the shielding cover is loose, and there is also a significant difference in frequency distribution. When the loosening is severe, due to the nonlinear propagation of structural vibration, harmonics of 3n times the fundamental frequency (n is 1, 2, 3...), i.e., higher harmonics such as 300Hz, 600Hz, and 900Hz, will also appear; when the shielding cover becomes loose, the sound pressure level increases significantly, from 48.3dB(A) to 50.2dB(A).

[0057] The abnormal noise caused by loose anchor bolts is due to electromagnetic force. The characteristics of the abnormal noise are as follows: When the anchor bolts are loose, an impact signal appears in the surface vibration signal of the GIS. Because the stiffness of the connection decreases after the bolts are loose, the preload is unbalanced. Under the action of electromagnetic force, gaps will appear at the bolts, which will cause continuous impacts. When the loosening is large, the impact is obvious and the signal shows a broadband signal. When the loosening is not obvious, under the action of 100Hz electromagnetic force, continuous slight impacts will appear as 100Hz and its harmonic components on the signal. As can be seen from the sound field distribution, after the anchor bolts are loose, the sound pressure level increases significantly, from 48.3dB(A) to 52.3dB(A).

[0058] From the above analysis of different types of abnormal noises in GIS, it can be concluded that there are significant differences in frequency among various types of abnormal noises in GIS. The frequency of abnormal noises caused by loose shielding is generally the 3nth harmonic of the fundamental frequency (n is 1, 2, 3...), that is, higher harmonics such as 300Hz, 600Hz, and 900Hz. The abnormal noise caused by loose anchor bolts, under the action of 100Hz electrodynamic force, will be displayed as 100Hz and its harmonic components on the signal. Therefore, in order to improve the accuracy of GIS abnormal noise judgment, in step S5, The frequency characteristics in the acoustic parameters of GIS equipment can be used to help determine the type of abnormal noise: when the frequency of the abnormal noise is generally 3n times the fundamental frequency (n is 1, 2, 3..., and the fundamental frequency is 100Hz, i.e., higher harmonics such as 300Hz, 600Hz, 900Hz, etc.), the output diagnostic results will indicate the possibility of abnormal noise due to a loose shielding cover; when the abnormal noise is a higher harmonic of 100Hz, the output diagnostic results will indicate the possibility of abnormal noise due to a loose shielding cover or loose anchor bolts.

[0059] Sometimes, multiple types of GIS abnormal noises occur simultaneously. As can be seen from the above analysis of different types of GIS abnormal noises, there are certain differences in frequency among different types of GIS abnormal noises. In order to accurately identify different types of GIS abnormal noises, when multiple GIS equipment abnormal noises exist, the actual acoustic parameters of the GIS equipment abnormal noises are collected one by one according to the frequency band and processed in step S5, so as to diagnose the multiple different types of abnormal noises present on site.

[0060] By extracting features from the acoustic parameters of GIS abnormal noises as the basis for the analysis and judgment data of the diagnostic decision model, in order to facilitate data processing, step S1 also performs feature value normalization on multiple acoustic feature values.

[0061] The type of abnormal noise in GIS is diagnosed by analyzing the acoustic characteristic values ​​of the abnormal noise. There are many types of acoustic characteristic values. In step S1, the original vibration and acoustic signals of various operating states of GIS equipment are fused to obtain multiple acoustic characteristic values ​​of various abnormal noises of GIS equipment, including harmonic characteristic values, time domain characteristic values ​​and frequency domain characteristic values.

[0062] Among them, the harmonic characteristic values ​​include: odd harmonic ratio, dominant frequency, fundamental frequency ratio, and dominant frequency ratio;

[0063] Time-domain feature values ​​include: skewness, mean, root mean square value, peak value, and kurtosis value;

[0064] Frequency domain eigenvalues ​​include: spectral kurtosis, spectral entropy, spectral flux, spectral center of gravity, and spectral attenuation.

[0065] The definitions and calculation methods of these feature values ​​are existing techniques in speech signal processing and are common knowledge. Therefore, the calculation of each feature value will not be described in detail here. Among them:

[0066] The dominant frequency represents the frequency with the largest amplitude among the 50Hz harmonic frequencies of the sound signal. The calculation formula is:

[0067] fm A fm =max(A fi ))(1)

[0068] The dominant frequency ratio represents the proportion of the amplitude corresponding to the dominant frequency of the sound signal to the sum of the amplitudes of all harmonics. The calculation formula is:

[0069]

[0070] The fundamental frequency proportion represents the percentage of the amplitude corresponding to the fundamental frequency of the sound signal (50Hz) out of the sum of the amplitudes of all harmonics. The calculation formula is:

[0071]

[0072] The odd harmonic ratio represents the proportion of odd-numbered harmonic energy in a 50Hz sound signal, and is calculated using the following formula:

[0073]

[0074] In equations (1)-(4) above, fi represents the harmonic frequency; Afi represents the harmonic amplitude; p_fdt represents the harmonic proportion; A ff Afj represents the fundamental frequency amplitude, and Afj represents the odd harmonic amplitude.

[0075] In step S3, the seven feature values ​​are sorted from largest to smallest according to their importance: spectral kurtosis, spectral entropy, fundamental frequency ratio, spectral flux, main frequency, odd harmonic ratio, and skewness.

[0076] In step S4, a fault diagnosis decision tree model is established based on the ranking of each acoustic feature value, including the following diagnostic steps:

[0077] a) The specific diagnosis of loose anchor bolts is as follows: a1) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, and the spectral entropy is less than or equal to the spectral entropy setting value, and the spectral flux is less than or equal to the spectral flux setting value, then the defect is diagnosed as loose anchor bolts; a2) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, and the spectral entropy is greater than the spectral entropy setting value, and the main frequency is greater than or equal to the main frequency setting value, then the defect is diagnosed as loose anchor bolts.

[0078] b) The specific diagnosis of a loose shield is as follows: b1) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, and the spectral entropy is less than or equal to the spectral entropy setting value, and the spectral flux is less than or equal to the spectral flux setting value, and the skewness is less than or equal to the skewness setting value, then the shield is diagnosed as loose; b2) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, and the fundamental frequency proportion is greater than the fundamental frequency proportion setting value, then the shield is diagnosed as loose.

[0079] c) The normal state diagnosis is as follows: if the spectral kurtosis is less than or equal to the spectral kurtosis setting value, the fundamental frequency ratio is less than or equal to the fundamental frequency ratio setting value, and the odd harmonic ratio is greater than the odd harmonic ratio setting value, then it is diagnosed as a normal state.

[0080] In step S3, the range of each acoustic feature value for each type of abnormal noise from the GIS equipment is analyzed. The degree of difference between each feature value and different types of GIS abnormal noises is evaluated. The greater the difference in the distribution range of the feature value across different types of GIS abnormal noises, the greater the degree of difference between the feature value and different types of GIS abnormal noises, and thus the greater the importance of the feature value. The importance score of each feature value is evaluated based on the significance of the differences between each feature value. The importance of each feature value is ranked according to the scores, and then diagnosis is performed according to the ranking of feature value importance. For the feature value of a decision node in the decision tree model, the abnormal noise category is determined based on the difference in the range of each acoustic feature value for each type of abnormal noise from the GIS equipment, ultimately completing the determination of the abnormal noise category.

[0081] In step S4, a fault diagnosis decision tree model is established based on the ranking of each acoustic feature value. The feature value with higher importance has higher priority. The decision tree model is established according to the importance to diagnose abnormal noises in GIS.

[0082] In addition, there may be other noises in the background environment, which do not have obvious regularity. In order to reduce the impact of background noise on the diagnosis of abnormal noise in GIS, in step S5, the collected acoustic parameters of abnormal noise from the actual GIS equipment are first filtered to remove background noise interference before analysis to obtain multiple acoustic feature values. This can reduce the impact of environmental noise on the diagnosis of abnormal noise.

[0083] When multiple GIS devices emit abnormal noises, the actual acoustic parameters of the abnormal noises from the GIS devices are collected sequentially by frequency band and processed in step S5.

[0084] Example 2

[0085] This embodiment also relates to an automatic diagnostic system for abnormal noises in GIS systems, used to implement the method for diagnosing abnormal noises in GIS systems described in the above technical solution, such as... Figure 2 The system includes a host computer 1 and a signal acquisition module 2.

[0086] The host computer 1 includes a signal processing submodule 12 and a storage submodule 11. The storage submodule 11 is equipped with a database of acoustic characteristic values ​​of various abnormal noises of GIS equipment. The signal processing submodule 12 is used to analyze the acoustic parameters of abnormal noises of GIS equipment, obtain the acoustic characteristic values ​​of abnormal noises of GIS equipment, and store them in the storage submodule 11.

[0087] The signal processing submodule 12 is also used to derive the range of each acoustic characteristic value for each type of abnormal noise of the GIS equipment based on the GIS equipment abnormal noise acoustic characteristic value data in the established GIS equipment abnormal noise acoustic characteristic value database;

[0088] The signal processing submodule 12 is also used to analyze the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment, and to score and rank the importance of each acoustic feature value according to the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment;

[0089] Signal processing submodule 12 is also used to establish a fault diagnosis decision tree model based on the sorting of each acoustic feature value.

[0090] The signal acquisition module 2 is used to acquire real acoustic parameters of abnormal noise from GIS equipment and transmit them to the host computer 1; the signal processing submodule 12 of the host computer 1 is used to analyze the received real acoustic parameters of abnormal noise from GIS equipment to obtain multiple acoustic feature values, and input the obtained acoustic feature values ​​of real acoustic parameters of abnormal noise from GIS equipment into the fault diagnosis decision tree model.

[0091] The signal processing submodule 12 is also used to output the results of the fault diagnosis decision tree model's analysis and diagnosis of the acoustic characteristic values ​​of the actual GIS equipment abnormal noise acoustic parameters.

[0092] Signal acquisition module 2 can be integrated with a microphone array, which is used to acquire abnormal noise signals from GIS.

[0093] Example 3

[0094] This embodiment also relates to an electronic device, including: a processor and a memory; the memory is used to store a computer program; the processor is used to execute the computer program stored in the memory, so that the terminal executes the method for diagnosing abnormal noises in GIS as described in the above technical solution.

[0095] Example 4

[0096] This embodiment also relates to a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for diagnosing abnormal noises in GIS as described in the above technical solution.

[0097] The automatic diagnostic system, electronic device, and computer-readable storage medium for GIS noise of the present invention also have the beneficial effects of the automatic diagnostic method for GIS noise of the present invention described above, which will not be repeated here.

[0098] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An automatic diagnostic method for abnormal noises in GIS, characterized in that, The method includes the following steps: Step S1: Collect and analyze the acoustic parameters of various abnormal noises from GIS equipment to obtain multiple acoustic characteristic values ​​of various abnormal noises from GIS equipment. Step S2: Based on the acoustic characteristic value data, determine the range of each acoustic characteristic value for various abnormal noises from the GIS equipment; Step S3: Based on the degree of difference between each acoustic feature value range and the abnormal noise of the GIS equipment, score and rank the importance of the acoustic feature values; Step S4: Establish a fault diagnosis decision tree model based on the importance ranking of acoustic feature values; Step S5: Use the fault diagnosis decision tree model to diagnose the actual acoustic parameters of abnormal noises in the GIS equipment and output the diagnostic results of the abnormal noises in the GIS equipment.

2. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, The various abnormal noises from the GIS equipment include noises from loose shielding covers and noises from loose anchor bolts.

3. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, The acoustic characteristic values ​​include harmonic characteristic values, time-domain characteristic values, and frequency-domain characteristic values; The harmonic characteristic values ​​mentioned include odd harmonic ratio, dominant frequency, and fundamental frequency ratio; The time-domain feature values ​​include skewness; The frequency domain eigenvalues ​​include spectral kurtosis, spectral entropy, and spectral flux.

4. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, The larger the range of acoustic feature values, the greater the importance of the acoustic feature value, and the higher the corresponding importance score; The scoring and ranking of the importance of acoustic feature values ​​is as follows: the acoustic feature values ​​are ranked from largest to smallest, and the top n feature values ​​are obtained in sequence, including spectral kurtosis, spectral entropy, fundamental frequency ratio, spectral flux, dominant frequency, odd harmonic ratio, and skewness.

5. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, In step S4, the fault diagnosis decision tree model includes a diagnostic module for loose anchor bolts, a diagnostic module for loose shielding, and a diagnostic module for normal conditions. The specific process of the diagnostic module for loose anchor bolts is as follows: a1) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, the spectral entropy is less than or equal to the spectral entropy setting value, and the spectral flux is less than or equal to the spectral flux setting value, then the defect is diagnosed as loose anchor bolts. Or a2) If the spectral kurtosis is less than or equal to the spectral kurtosis setting value, the spectral entropy is greater than the spectral entropy setting value, and the main frequency is greater than or equal to the main frequency setting value, then the defect is diagnosed as loose anchor bolts. The specific process of the diagnostic module for the loose shield is as follows: b1) If the following conditions are met simultaneously: spectral kurtosis is less than or equal to the spectral kurtosis setting value, spectral entropy is less than or equal to the spectral entropy setting value, spectral flux is less than or equal to the spectral flux setting value, and skewness is less than or equal to the skewness setting value, then the shielding cover is diagnosed as loose. Or b2) If both the spectral kurtosis is less than or equal to the spectral kurtosis setting value and the fundamental frequency proportion is greater than the fundamental frequency proportion setting value, then the diagnosis is that the shielding cover is loose; The specific process of the diagnostic module in the normal state is as follows: If the following conditions are met simultaneously: spectral kurtosis is less than or equal to the spectral kurtosis setting value, fundamental frequency ratio is less than or equal to the fundamental frequency ratio setting value, and odd harmonic ratio is greater than the odd harmonic ratio setting value, then the diagnosis is normal.

6. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, In step S5, the collected acoustic parameters of abnormal noise from real GIS equipment are first filtered to remove background noise interference, and then analyzed to obtain multiple acoustic feature values.

7. The automatic diagnostic method for abnormal noises in GIS according to claim 1, characterized in that, The diagnostic method further includes, when multiple GIS devices are making abnormal noises, collecting real acoustic parameters of the abnormal noises of the GIS devices one by one according to frequency band, and then performing the processing described in step S5.

8. A system employing the automatic diagnostic method for GIS abnormal noise as described in claim 1, characterized in that, The system includes a host computer (1) and a signal acquisition module (2). The host computer (1) includes a signal processing submodule (12) and a storage submodule (11). The signal acquisition module (2) is used to acquire the acoustic parameters of abnormal noises from GIS equipment and send them to the signal processing submodule (12) for processing; The storage submodule (11) is equipped with a GIS equipment abnormal noise acoustic characteristic value database for storing the acoustic characteristic values ​​of various abnormal noises of GIS equipment; The signal processing submodule (12) is used to analyze the acoustic parameters of abnormal noise of GIS equipment received from the signal acquisition module (2), obtain the acoustic characteristic value of abnormal noise of GIS equipment and store it in the storage submodule (11). The signal processing submodule (12) is also used to analyze the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment, and to score and rank the importance of each acoustic feature value according to the correlation between the range of each acoustic feature value of each type of abnormal noise of the GIS equipment and the abnormal noise of the GIS equipment; The signal processing submodule (12) is also used to establish a fault diagnosis decision tree model based on the sorting of each acoustic feature value; The signal processing submodule (12) is also used to output the results of the fault diagnosis decision tree model's analysis and diagnosis of the acoustic characteristic values ​​of the real GIS equipment abnormal acoustic parameters.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images