A GIS fault diagnosis method and system based on acoustic monitoring

Abstract

The present invention relates to the technical field of electric power systems, and discloses a GIS fault diagnosis method and system based on acoustic monitoring, including: obtaining initial sound data of each monitoring point in the GIS; performing a first fault location according to the initial sound data to obtain the area where the fault source in the GIS is located; and extracting the positioning sound data; performing a second fault location according to the positioning sound data to obtain the location information of multiple fault sources in the GIS; and performing voiceprint feature analysis according to the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults; and performing GIS fault diagnosis according to the location information of the multiple fault sources and the fault types of the multiple fault sources. The GIS fault diagnosis method and system based on acoustic monitoring of the present invention significantly improves the accuracy of determining the location of the fault source, and simultaneously realizes the timeliness diagnosis of multiple fault sources.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power systems, and in particular to a GIS fault diagnosis method and system based on acoustic monitoring. Background Art

[0002] In modern power systems, gas insulated switchgear (GIS) is an efficient, reliable and space-saving power equipment widely used in substations and distribution systems. Due to its fully enclosed structure, GIS has excellent insulation performance and environmental adaptability. However, during long-term operation, its internal components may age, wear, loosen, discharge or other faults. These faults not only affect the normal operation of the equipment, but also may cause safety hazards to the power system. Therefore, real-time monitoring and fault diagnosis of GIS are particularly important.

[0003] Currently, the monitoring methods for GIS mainly rely on regular detection of equipment status and manual inspections. However, this method usually has problems such as delayed response, low efficiency, and inability to achieve real-time monitoring. In addition, with the development of smart grids and automation equipment, the monitoring methods for GIS need to be upgraded to meet new needs.

[0004] In addition, GIS status assessment is also carried out through electrical monitoring, temperature analysis, vibration testing and acoustic wave monitoring. Among them, acoustic wave monitoring is an effective fault diagnosis method, because the equipment will generate sound waves with a specific spectrum during operation. When the equipment fails, the sound wave characteristics will change. Therefore, by analyzing the acoustic wave signal of GIS, potential fault information can be effectively captured.

[0005] However, in existing research and applications, since the location of the fault source (i.e., the fault sound source) is uncertain, and the microphone array that collects sound data is usually evenly distributed, the accuracy of fault location based on the sound data collected by the microphone array is low, and usually only a rough diagnosis can be made, and the specific location of the fault cannot be accurately diagnosed. Summary of the invention

[0006] To this end, the technical problem to be solved by the present invention is to overcome the problem that the prior art cannot accurately diagnose the GIS fault location, and to provide a GIS fault diagnosis method and system based on acoustic monitoring, which significantly improves the accuracy of determining the fault source location and realizes the timely diagnosis of multiple fault sources.

[0007] In the first aspect, in order to solve the above technical problems, the present invention provides a GIS fault diagnosis method based on acoustic monitoring, comprising the following steps:

[0008] Acquiring initial sound data of each monitoring point in the GIS according to a sound detector array arranged on the GIS;

[0009] Performing a first fault location according to the initial sound data to obtain the area where the fault source in the GIS is located; and extracting the positioning sound data; wherein the positioning sound data is the initial sound data collected by the sound detector arranged in the area where the fault source is located;

[0010] Perform a second fault location according to the positioning sound data to obtain the location information of the fault source in the GIS; wherein the second fault location is based on a multi-input multi-output algorithm to build a model to locate the location information of multiple fault sources;

[0011] and performing voiceprint feature analysis based on the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults;

[0012] The fault diagnosis of the GIS is performed according to the location information of the multiple fault sources and the fault types of the multiple fault sources.

[0013] In one embodiment of the present invention, the first fault location is performed according to the initial sound data to obtain the area where the fault source of the fault in the GIS is located, including:

[0014] Constructing an initial sound matrix P according to the initial sound data;

[0015] Constructing a cross-spectrum matrix C according to the initial sound matrix; wherein, when constructing the cross-spectrum matrix C, the time delay of the sound signal of the initial sound matrix in the time domain is converted into a phase in the frequency domain;

[0016] Perform eigendecomposition on the cross-spectrum matrix C to obtain the non-zero eigenvalues ​​of the cross-spectrum matrix C, extract the minimum value of the non-zero eigenvalues, and calculate the total power threshold of the sound source corresponding to the minimum value of the non-zero eigenvalue ;

[0017] Construct a fault area location model based on the sparse deconvolution beamforming algorithm;

[0018] The total power threshold An optimal solution of the fault area location model is solved for the constraint conditions, and the area where the fault source is located is determined according to the optimal solution of the fault area location model.

[0019] In one embodiment of the present invention, constructing an initial sound matrix P according to the initial sound data includes:

[0020] Green's function is introduced to construct array vector and transfer matrix G is obtained;

[0021] Construct the initial sound matrix P according to the transfer matrix G: ; s is the sound source amplitude of the data vector at the grid point; n is the noise vector of the data.

[0022] In one embodiment of the present invention, the optimization problem and the constraint conditions for solving the fault area location model are respectively:

[0023] subject to ;

[0024] Among them, C is the cross-spectral matrix; G is the transfer matrix; is the covariance matrix of the unknown coherent sound source; is the conjugate matrix of the transfer matrix; is the unknown noise power; I is the unit matrix; Indicates the norm; is the power of the nth grid point of the sound source; N is the number of grid points.

[0025] In one embodiment of the present invention, performing a second fault location according to the positioning sound data to obtain location information of multiple fault sources in the GIS includes:

[0026] Constructing a positioning sound matrix according to the positioning sound data ;

[0027] Introduce Green's function to construct array vector and obtain transfer matrix ;

[0028] According to the positioning sound matrix and the transfer matrix Constructing the cross-spectral matrix ;

[0029] The cross-spectral matrix Perform eigenvalue decomposition to obtain the cross-spectral matrix The set of eigenvalues ​​of ;

[0030] For each eigenvalue in the eigenvalue set , both compare eigenvalues With the feature value threshold ,like , then the eigenvalue The corresponding sound source has a fault; otherwise, there is no fault;

[0031] According to the characteristic value Solve the total power threshold of the incoherent sound source corresponding to the fault sound source , is the total power threshold of the incoherent sound source of the i-th fault sound source; and solve the The total power threshold of incoherent sound sources corresponding to all eigenvalues ​​of is constructed to construct the total power threshold vector of incoherent sound sources of all eigenvalues ; K is the number of fault sound sources; is the total power threshold of the Kth fault sound source;

[0032] Construct a fault location model based on the sparse deconvolution beamforming algorithm;

[0033] According to the total power threshold vector of the incoherent sound source Setting constraint conditions to solve the optimal solution of the fault location positioning model, and determining the location of the fault sound source according to the optimal solution of the fault location positioning model ;in, is the location of the Kth fault sound source; Represents a three-dimensional matrix.

[0034] In one embodiment of the present invention, the total power threshold of the incoherent sound source is The calculation method is:

[0035] ;

[0036] Among them, i is the i-th fault sound source; is the eigenvalue corresponding to the i-th fault sound source; is the eigenvalue matrix remaining after removing the row and column where the fault sound source is located; M is the number of sound detectors; is the cross-spectral matrix The j-th eigenvalue of ; I is the identity matrix; represents trace operation; Represents a function for finding the position of an element in the corresponding matrix.

[0037] In one embodiment of the present invention, the optimization problem and the constraint conditions for solving the fault location model are respectively:

[0038] subject to ;

[0039] in, is the covariance matrix of the unknown coherent sound source of fault sound source i; is the unknown noise power; describe is a positive semidefinite matrix; represents trace operation; is the total power threshold of the incoherent sound source of the i-th fault sound source; is the unknown noise power.

[0040] In one embodiment of the present invention, the voiceprint feature analysis is performed based on the positioning sound data to obtain the fault types of the multiple fault sources in the GIS, including:

[0041] Performing empirical mode decomposition on the localization sound data and calculating the one-dimensional spectral entropy of each modal component;

[0042] The fault type is obtained according to the one-dimensional spectral entropy; wherein the one-dimensional spectral entropy is input into the SVM classifier as a feature element to obtain the fault type.

[0043] In one embodiment of the present invention, it also includes issuing an alarm tone when the GIS diagnoses a fault, and sending the diagnosed fault location and fault type to the display interface.

[0044] In a second aspect, in order to solve the above technical problems, the present invention also provides a GIS fault diagnosis system based on acoustic monitoring, comprising:

[0045] A data acquisition module, which is used to obtain initial sound data of each monitoring point in the GIS according to a sound detector array arranged on the GIS;

[0046] A fault area locating module, which is used to perform a first fault location according to the initial sound data, and obtain the area where the fault source of the fault in the GIS is located;

[0047] A positioning sound data extraction module, which is used to extract positioning sound data, wherein the positioning sound data is the initial sound data collected by the sound detector in the area where the fault source is located;

[0048] A fault location positioning module, which is used to perform a second fault location according to the positioning sound data to obtain location information of multiple fault sources in the GIS;

[0049] A fault type identification module, which is used to perform voiceprint feature analysis based on the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults;

[0050] A fault diagnosis module is used to perform GIS fault diagnosis based on the location information of the multiple fault sources and the fault types of the multiple fault sources.

[0051] The above technical solution of the present invention has the following beneficial effects compared with the prior art:

[0052] The GIS fault diagnosis method and system based on acoustic monitoring described in the present invention do not change the layout position and number of existing sound detectors (i.e., microphones), and the sound data of each monitoring point inside the GIS is monitored and detected by the arranged sound detector array in real time. The sound data is first located and analyzed to preliminarily locate the area where the fault source of the fault in the GIS is located; then a targeted second positioning analysis is performed on the sound data in the area where the preliminarily located fault source is located to accurately locate the specific location of the fault source; wherein the sound data of the first positioning analysis have high sparsity, while the sound data of the second positioning analysis have reduced sparsity, and a fault area positioning model and a fault location positioning model can be constructed with the help of a sparse deconvolution beamforming algorithm, and the sparse deconvolution beamforming algorithm can effectively separate and locate multi-channel audio signals, and the specific location of the fault source can be accurately located by changing the constraint conditions of the sparse deconvolution beamforming algorithm; the two fault positionings significantly improve the accuracy of determining the location of the fault source.

[0053] On the other hand, in the GIS fault diagnosis method and system based on acoustic monitoring described in the present invention, the sound data for the second positioning analysis only comes from the area where the fault source is located as determined by the preliminary positioning, the amount of sound data is reduced, and the fault type diagnosis is only performed on the sound data in the area where the fault source is located, thereby improving the data processing efficiency and accelerating the timeliness of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0055] Figure 1 A flowchart of a GIS fault diagnosis method based on acoustic monitoring provided in a preferred embodiment of the present invention;

[0056] Figure 2 A schematic diagram of mapping the array distribution of sound detectors in space to a grid plane;

[0057] Figure 3 This is a flow chart of the first fault location in a preferred embodiment of the present invention;

[0058] Figure 4 This is a flow chart of the second fault location in a preferred embodiment of the present invention;

[0059] Figure 5 The module block diagram of a GIS fault diagnosis system based on acoustic monitoring provided in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0061] Embodiment 1

[0062] Reference Figure 1 As shown, the embodiment of the present invention discloses a GIS fault diagnosis method based on acoustic monitoring, comprising the following steps:

[0063] The initial sound data of each monitoring point in the GIS is obtained according to the sound detector array arranged on the GIS; in a specific application scenario, a sound detector array is arranged on the shell of the GIS, for example, a microphone array is arranged on the shell of the GIS, and the arrangement position and number of the microphones are determined according to the actual situation, and can be evenly arranged in combination with the three-dimensional contour of the GIS, or the arrangement density and number of microphones can be increased in areas that are more prone to faults according to historical data; the sound signals of each arrangement point are monitored and detected in real time by the microphone as sound data for subsequent fault location analysis.

[0064] The first fault location is performed according to the initial sound data to obtain the area where the fault source in the GIS is located; and the positioning sound data is extracted; wherein the positioning sound data is the initial sound data collected by the sound detectors arranged in the area where the GIS fault source is located; in a specific application scenario, each sound detector detects one channel of sound data, and the sound detector array detects multiple channels of sound data at the same time; for different fault types, the detected sound data carries different information, for example, arc discharge is generated when a strong current passes through a gas, the gas is heated by electric energy and expands rapidly, generating sound waves, forming a "roaring" or "hissing" sound signal; lightning is a strong discharge phenomenon, and the accompanying loud noise is usually called "thunder"; electrostatic discharge (ESD) is generated by static electricity accumulation, when it is released, sound waves are generated, forming a short "pop" or "crackling" sound signal; and mechanical faults include poor contact of moving contacts, loose anchor bolts and internal residues; the initial sound data captured by the sound detector array carries sound information of various fault types, and the location and type of the fault can be diagnosed by analyzing the initial sound data.

[0065] Specifically, by analyzing multi-channel sound data, the area where the fault source of the fault in the GIS is located is located; a second fault location is performed based on the located sound data to obtain the location information of multiple fault sources in the GIS; and a voiceprint feature analysis is performed based on the located sound data to obtain the fault types of the multiple fault sources in the GIS; the fault types include local discharge faults and mechanical faults; in specific application scenarios, local discharge faults include arc discharges, lightning and electrostatic discharges; mechanical faults include poor contact of moving contacts, loose anchor bolts and internal residues, etc.; GIS fault diagnosis is performed based on the location information of the multiple fault sources and the fault types of the multiple fault sources.

[0066] The GIS fault diagnosis method based on acoustic monitoring described in the present invention does not change the layout position and number of existing sound detectors (i.e., microphones). The sound detectors arranged in an array monitor and detect the sound data of each monitoring point inside the GIS in real time. The sound data is first located and analyzed to preliminarily locate the area where the fault source of the fault in the GIS is located. Then, a targeted second location analysis is performed on the sound data in the area where the preliminarily located fault source is located to accurately locate the specific location of the fault source. Among them, the sound data of the first location analysis have high sparsity, while the sparsity of the sound data of the second location analysis is reduced. A fault area location model and a fault location location model can be constructed with the help of a sparse deconvolution beamforming algorithm. The sparse deconvolution beamforming algorithm can effectively separate and locate multi-channel audio signals. By changing the constraint conditions of the sparse deconvolution beamforming algorithm, the specific location of the fault source can be accurately located. The two fault locations significantly improve the accuracy of determining the location of the fault source.

[0067] On the other hand, in the GIS fault diagnosis method based on acoustic monitoring described in the present invention, the sound data for the second positioning analysis comes only from the area where the fault source is located as determined by the preliminary positioning, the amount of sound data is reduced, and the fault type diagnosis is only performed on the sound data in the area where the fault source is located, the data processing efficiency is improved, thereby accelerating the timeliness of fault diagnosis.

[0068] Specifically, refer to Figure 3 As shown, the first fault location is performed according to the initial sound data, and the area where the fault source of the fault in the GIS is located includes:

[0069] Constructing an initial sound matrix P according to the initial sound data;

[0070] Constructing a cross-spectrum matrix C according to the initial sound matrix; wherein, when constructing the cross-spectrum matrix C, the time delay of the sound signal of the initial sound matrix in the time domain is converted into a phase in the frequency domain;

[0071] In specific application scenarios, the initial sound matrix P is constructed by considering the detection frequency f of the sound detector, the spatial array distribution of the sound detector, and the initial sound data detected by the sound detector; the sound signal received by the sound detector needs to consider the phase difference between the signals, and also consider the attenuation of the sound signal during the propagation process. The Green function is introduced to construct the array vector and the transfer matrix G is obtained. The initial sound matrix P is represented by the transfer matrix G as follows: ; s is the sound source amplitude of the data vector at the grid point; n is the noise vector of the data.

[0072] Among them, constructing the array vector to obtain the transfer matrix G includes,

[0073] ;

[0074]

[0075] Reference Figure 2 As shown, the spatial array distribution of the sound detectors is mapped to the grid plane; wherein, is the distance from the nth grid point to the mth sound detector on the grid plane; is the distance from the nth grid point to the origin on the grid plane; is the angular frequency; Represents the element in the mth row and nth column of the transfer matrix G; Represents the speed of sound propagation.

[0076] In order to facilitate the solution and reduce the amount of calculation, the time delay of the sound signal in the time domain is converted to the phase in the frequency domain, and the cross-spectrum matrix C is constructed: ;in, represents the identity matrix; S represents the covariance matrix of the spatial sound source.

[0077] Perform eigendecomposition on the cross-spectrum matrix C to obtain the non-zero eigenvalues ​​of the cross-spectrum matrix C, extract the minimum value of the non-zero eigenvalues, and calculate the total power threshold of the sound source corresponding to the minimum value of the non-zero eigenvalue ;

[0078] Construct a fault area location model based on the sparse deconvolution beamforming algorithm;

[0079] The total power threshold An optimal solution of the fault area location model is solved for the constraint conditions, and the area where the fault source is located is determined according to the optimal solution of the fault area location model.

[0080] Furthermore, the optimization problem and the constraint conditions for solving the fault area location model are respectively:

[0081] subject to ;

[0082] Among them, C is the cross-spectral matrix; G is the transfer matrix; is the covariance matrix of the unknown coherent sound source; is the conjugate matrix of the transfer matrix; is the unknown noise power; I is the unit matrix; Indicates the norm; is the power of the nth grid point of the sound source; N is the number of grid points.

[0083] However, in the existing technology, fault diagnosis by acoustic wave monitoring is mostly limited to processing single-frame data, and cannot achieve timely identification of multiple fault sources. This limitation makes it difficult for the system to accurately diagnose the location of the fault source when facing complex faults, thereby affecting the efficiency and accuracy of troubleshooting.

[0084] In view of the fact that acoustic wave monitoring cannot realize the timely identification of multiple fault sources, the fault diagnosis method disclosed in the embodiment of the present invention performs a second fault location according to the positioning sound data to obtain the location information of multiple fault sources in the GIS; during the second fault location, a model is constructed based on a multiple-input multiple-output (MIMO) algorithm, and multiple groups of constraints for different fault sound sources are input at the same time to output the locations of multiple fault sources at the same time. The specific process is as follows:

[0085] Reference Figure 4 As shown, a positioning sound matrix is ​​constructed according to the positioning sound data In a specific application scenario, the detection frequency of the sound detector, the distance from the grid point to the sound detector, and the detected positioning sound data are considered to construct the positioning sound matrix .

[0086] According to the positioning sound matrix Constructing the cross-spectral matrix Specifically, construct the cross-spectral matrix The method is consistent with the method of constructing the cross-spectral matrix C;

[0087] The cross-spectral matrix Perform eigenvalue decomposition to obtain the cross-spectral matrix The set of eigenvalues ​​of ;

[0088] For each eigenvalue in the eigenvalue set , both compare eigenvalues With the feature value threshold ,like , then the eigenvalue There is a fault at the corresponding sound source; otherwise, there is no fault; traverse and compare each eigenvalue in the eigenvalue set With the feature value threshold , until the number of all fault sound sources is determined;

[0089] According to the characteristic value Solve the total power threshold of the incoherent sound source corresponding to the fault sound source , is the total power threshold of the incoherent sound source of the i-th fault sound source; and solve the The total power threshold of incoherent sound sources corresponding to all eigenvalues ​​of is constructed to construct the total power threshold vector of incoherent sound sources of all eigenvalues ; K is the number of fault sound sources; is the total power threshold of the Kth fault sound source.

[0090] Specifically, the total power value of the incoherent sound source of the fault sound source is The calculation method is:

[0091] ;

[0092] Among them, i is the i-th fault sound source; is the eigenvalue corresponding to the i-th fault sound source; is the eigenvalue matrix remaining after removing the row and column where the fault sound source is located; M is the number of sound detectors; is the cross-spectral matrix The j-th eigenvalue of ; I is the identity matrix; represents trace operation; Represents a function for finding the position of an element in the corresponding matrix.

[0093] Construct a fault location model based on the sparse deconvolution beamforming algorithm;

[0094] According to the total power threshold vector of the incoherent sound source Setting constraint conditions to solve the optimal solution of the fault location positioning model, and determining the location of the fault sound source according to the optimal solution of the fault location positioning model ;in, is the location of the Kth fault sound source; Represents a three-dimensional matrix.

[0095] Specifically, the optimization problem and the constraint conditions for solving the fault location model are: subject to ;

[0096] in, is the covariance matrix of the unknown coherent sound source of fault sound source i; is the unknown noise power; describe is a positive semidefinite matrix; represents trace operation; is the total power threshold of the incoherent sound source of the i-th fault sound source; is the unknown noise power.

[0097] In another embodiment of the present invention, the voiceprint feature analysis is performed based on the positioning sound data to obtain the fault types of the multiple fault sources in the GIS, including:

[0098] Performing empirical mode decomposition on the localization sound data and calculating the one-dimensional spectral entropy of each modal component;

[0099] The fault type is obtained according to the one-dimensional spectral entropy; wherein the one-dimensional spectral entropy is input into the SVM classifier as a feature element to obtain the fault type.

[0100] In another embodiment of the present invention, it also includes issuing an alarm tone when the GIS diagnoses a fault, and sending the diagnosed fault location and fault type to the display interface.

[0101] Embodiment 2

[0102] Reference Figure 5 As shown, the embodiment of the present invention also discloses a GIS fault diagnosis system based on acoustic monitoring, comprising:

[0103] A data acquisition module, which is used to obtain initial sound data of each monitoring point in the GIS according to a sound detector array arranged on the GIS;

[0104] A fault area locating module, which is used to perform a first fault location according to the initial sound data, and obtain the area where the fault source of the fault in the GIS is located;

[0105] A positioning sound data extraction module, which is used to extract positioning sound data, wherein the positioning sound data is the initial sound data collected by the sound detectors arranged in the area where the GIS fault source is located;

[0106] A fault location positioning module, which is used to perform a second fault location according to the positioning sound data to obtain location information of multiple fault sources in the GIS;

[0107] A fault type identification module, which is used to perform voiceprint feature analysis based on the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults;

[0108] The fault diagnosis module is used to perform fault diagnosis of the GIS according to the location information of the multiple fault sources and the fault types of the multiple fault sources.

[0109] The GIS fault diagnosis system based on acoustic monitoring described in the present invention does not change the layout position and number of existing sound detectors (i.e., microphones). The sound detectors arranged in an array monitor and detect the sound data of each monitoring point inside the GIS in real time. The sound data is first located and analyzed to preliminarily locate the area where the fault source of the fault in the GIS is located. Then, a targeted second location analysis is performed on the sound data in the area where the preliminarily located fault source is located to accurately locate the specific location of the fault source. Among them, the sound data of the first location analysis have high sparsity, while the sparsity of the sound data of the second location analysis is reduced. The fault area location model and the fault location location model can be constructed with the help of a sparse deconvolution beamforming algorithm. The sparse deconvolution beamforming algorithm can effectively separate and locate multi-channel audio signals. By changing the constraint conditions of the sparse deconvolution beamforming algorithm, the specific location of the fault source can be accurately located. The two fault locations significantly improve the accuracy of determining the location of the fault source.

[0110] On the other hand, in the GIS fault diagnosis system based on acoustic monitoring described in the present invention, the sound data for the second positioning analysis comes only from the area where the fault source is located as determined by the preliminary positioning, the amount of sound data is reduced, and the fault type diagnosis is only performed on the sound data in the area where the fault source is located, thereby improving the data processing efficiency and accelerating the timeliness of fault diagnosis.

[0111] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0112] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0113] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0115] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A GIS fault diagnosis method based on acoustic monitoring, characterized in that: The following steps are included: Acquiring initial sound data of each monitoring point in the GIS according to a sound detector array arranged on the GIS; Performing a first fault location according to the initial sound data to obtain the area where the fault source in the GIS is located; and extracting positioning sound data; wherein the positioning sound data is the initial sound data collected by the sound detector arranged in the area where the fault source is located; Perform a second fault location according to the positioning sound data to obtain the location information of the fault source in the GIS; wherein the second fault location is based on a multi-input multi-output algorithm to build a model to locate the location information of multiple fault sources; and performing voiceprint feature analysis based on the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults; The fault diagnosis of the GIS is performed according to the location information of the multiple fault sources and the fault types of the multiple fault sources.

2. The GIS fault diagnosis method based on acoustic monitoring according to claim 1 is characterized in that: Performing a first fault location according to the initial sound data to obtain the area where the fault source in the GIS is located includes: Constructing an initial sound matrix P according to the initial sound data; Constructing a cross-spectrum matrix C according to the initial sound matrix; wherein, when constructing the cross-spectrum matrix C, the time delay of the sound signal of the initial sound matrix in the time domain is converted into a phase in the frequency domain; Perform eigenvalue decomposition on the cross-spectrum matrix C to obtain the non-zero eigenvalues ​​of the cross-spectrum matrix C, extract the minimum value of the non-zero eigenvalues, and calculate the total power threshold of the sound source corresponding to the minimum value of the non-zero eigenvalue ; Construct a fault area location model based on the sparse deconvolution beamforming algorithm; The total power threshold An optimal solution of the fault area location model is solved for the constraint conditions, and the area where the fault source is located is determined according to the optimal solution of the fault area location model.

3. The GIS fault diagnosis method based on acoustic monitoring according to claim 2 is characterized in that: Constructing the initial sound matrix P according to the initial sound data comprises: Green's function is introduced to construct array vector and transfer matrix G is obtained; Construct the initial sound matrix P according to the transfer matrix G: ; s is the sound source amplitude of the data vector at the grid point; n is the noise vector of the data.

4. The GIS fault diagnosis method based on acoustic monitoring according to claim 3 is characterized in that: The optimization problem and the constraint conditions for solving the fault area location model are: subject to ; Among them, C is the cross-spectral matrix; G is the transfer matrix; is the covariance matrix of the unknown coherent sound source; is the conjugate matrix of the transfer matrix; is the unknown noise power; I is the unit matrix; Indicates the norm; is the power of the nth grid point of the sound source; N is the number of grid points.

5. The GIS fault diagnosis method based on acoustic monitoring according to claim 1 is characterized in that: Performing a second fault location based on the positioning sound data to obtain location information of multiple fault sources in the GIS includes: Constructing a positioning sound matrix according to the positioning sound data ; Introduce Green's function to construct array vector and obtain transfer matrix ; According to the positioning sound matrix and the transfer matrix Constructing the cross-spectral matrix ; The cross-spectral matrix Perform eigenvalue decomposition to obtain the cross-spectral matrix The set of eigenvalues ​​of ; For each eigenvalue in the eigenvalue set , both compare eigenvalues With the feature value threshold ,like , then the eigenvalue The corresponding sound source has a fault; otherwise, there is no fault; According to the characteristic value Solve the total power threshold of the incoherent sound source corresponding to the fault sound source , is the total power threshold of the incoherent sound source of the i-th fault sound source; and solve the The total power threshold of incoherent sound sources corresponding to all eigenvalues ​​of is constructed to construct the total power threshold vector of incoherent sound sources of all eigenvalues ; K is the number of fault sound sources; is the total power threshold of the Kth fault sound source; Construct a fault location model based on the sparse deconvolution beamforming algorithm; According to the total power threshold vector of the incoherent sound source Setting constraint conditions to solve the optimal solution of the fault location positioning model, and determining the location of the fault sound source according to the optimal solution of the fault location positioning model ;in, is the location of the Kth fault sound source; Represents a three-dimensional matrix.

6. The GIS fault diagnosis method based on acoustic monitoring according to claim 5 is characterized in that: The total power threshold of the incoherent sound source The calculation method is: ； Among them, i is the i-th fault sound source; is the eigenvalue corresponding to the i-th fault sound source; is the eigenvalue matrix remaining after removing the row and column where the fault sound source is located; M is the number of sound detectors; is the cross-spectral matrix The j-th eigenvalue of ; I is the identity matrix; represents trace operation; Represents a function for finding the position of an element in the corresponding matrix.

7. The GIS fault diagnosis method based on acoustic monitoring according to claim 5 is characterized in that: The optimization problem and the constraint conditions for solving the fault location model are: subject to ; in, is the covariance matrix of the unknown coherent sound source of fault sound source i; is the unknown noise power; describe is a positive semidefinite matrix; represents trace operation; is the total power threshold of the incoherent sound source of the i-th fault sound source; is the unknown noise power.

8. The GIS fault diagnosis method based on acoustic monitoring according to claim 1 is characterized in that: According to the positioning sound data, the voiceprint feature analysis is performed to obtain the fault types of multiple fault sources in the GIS, including: Performing empirical mode decomposition on the localization sound data and calculating the one-dimensional spectral entropy of each modal component; The fault type is obtained according to the one-dimensional spectral entropy; wherein the one-dimensional spectral entropy is input into the SVM classifier as a feature element to obtain the fault type.

9. The GIS fault diagnosis method based on acoustic monitoring according to claim 1 is characterized in that: It also includes sounding an alarm when the GIS diagnoses a fault, and sending the diagnosed fault location and fault type to the display interface.

10. A GIS fault diagnosis system based on acoustic monitoring, characterized in that: include, A data acquisition module, which is used to obtain initial sound data of each monitoring point in the GIS according to a sound detector array arranged on the GIS; A fault area locating module, which is used to perform a first fault location according to the initial sound data, and obtain the area where the fault source of the fault in the GIS is located; A positioning sound data extraction module, which is used to extract positioning sound data, wherein the positioning sound data is the initial sound data collected by the sound detector in the area where the fault source is located; A fault location positioning module, which is used to perform a second fault location according to the positioning sound data to obtain the location information of multiple fault sources in the GIS; wherein the second fault location is based on a multi-input multi-output algorithm to build a model to locate the location information of multiple fault sources; A fault type identification module, which is used to perform voiceprint feature analysis based on the positioning sound data to obtain the fault types of multiple fault sources in the GIS; the fault types include partial discharge faults and mechanical faults; A fault diagnosis module is used to perform GIS fault diagnosis based on the location information of the multiple fault sources and the fault types of the multiple fault sources.

Images