A multi-source partial discharge identification method based on PRPD images

Abstract

The present invention relates to the field of discharge fault detection technology, and specifically to a multi-source partial discharge identification method based on PRPD spectrum, comprising the following steps: reading sound signals; establishing signal representation and array directivity function; performing cross-correlation calculation to obtain horizontal azimuth and elevation angles; obtaining a composite signal representation; drawing a PRPD spectrum; calculating the phase difference between adjacent peaks, and determining the presence of discharge if it is within a preset range; performing phase synchronization to superimpose and merge every three cycles of the composite signal; calculating preset eigenvalues ​​respectively; comparing the preset eigenvalues ​​with representative eigenvalues ​​of known discharge types, and determining the presence of a discharge type if the similarity exceeds a preset threshold. The beneficial technical effects of the present invention include: being able to obtain the horizontal azimuth and elevation angles of each partial discharge sound source, thereby realizing the discovery and location of partial discharge situations; and being able to identify the discharge type of partial discharge, providing a reference for identifying and handling faults.

Description

Technical Field

[0001] The present invention relates to the technical field of discharge fault detection, and in particular to a multi-source partial discharge identification method based on PRPD spectra. Background Art

[0002] Most existing technologies identify a single type of partial discharge in ideal environments, while a small number identify partial discharge types in noisy environments. These recognition environments are limited to allowing only one object to generate a discharge, and the interference removed by these methods only includes some white noise and some low-frequency bandpass noise. However, if other objects are discharging in the environment, the new characteristics generated by the sounds emitted by the two different types of discharges will greatly reduce the recognition accuracy. Furthermore, the effective information generated by the discharge of another object also needs to be identified. If only one discharge signal is eliminated from the mixed signal, the corresponding discharge object cannot be matched. Therefore, it is necessary to develop technologies that can identify each discharge location when discharges occur in multiple locations.

[0003] The prior art discloses a device for identifying the type of partial discharge (PD). The device includes a detection circuit and a processing module. The detection circuit has an input terminal connected to the switchgear housing and an output terminal connected to the processing module. The detection circuit is configured to determine the equivalent impedance of the detection circuit based on electromagnetic waves generated by partial discharge in the switchgear. The processing module is configured to identify the output voltage waveform of the detection circuit and adjust the equivalent impedance of the detection circuit based on the identification result until the output voltage waveform of the detection circuit reaches a peak value. The processing module is further configured to determine the type of partial discharge in the switchgear based on the output voltage and frequency of the detection circuit when the output voltage waveform reaches a peak value. While this technical solution can produce a relatively stable and clear waveform when the output voltage waveform reaches a peak value, and can accurately obtain the output voltage value and frequency of the waveform based on this waveform, thereby improving the accuracy of partial discharge type identification, this technical solution does not address discharge identification in the case of multi-source partial discharge. Summary of the Invention

[0004] The technical problem to be solved by this invention is that there is currently a lack of a multi-source discharge identification solution that can identify discharges at multiple locations. A multi-source partial discharge identification method based on PRPD patterns is proposed, which can achieve multi-source partial discharge identification.

[0005] To solve the above technical problems, the present invention adopts the following technical solution: a multi-source partial discharge identification method based on PRPD spectrum, comprising the following steps:

[0006] Read the sound signal received by the array of N microphones, and the number of partial discharges is recorded as , the horizontal azimuth and elevation angles of partial discharge are respectively and , ;

[0007] Establish the signal representation received by N microphones at time t and array directivity function;

[0008] Perform cross-correlation calculation on the sound signals received by N microphones to obtain the horizontal azimuth angle of each partial discharge sound source and pitch angle ;

[0009] According to the horizontal azimuth angle of each partial discharge sound source and pitch angle Get the composite signal received by N microphones at time t ;

[0010] Based on the composite signal received by N microphones at time t Draw PRPD maps;

[0011] Calculate the phase difference between adjacent peaks of the composite signal. If the phase difference between adjacent peaks is within a preset range, it is determined that there is a discharge. Otherwise, if the phase difference between adjacent peaks is not within the preset range, it is determined that there is no discharge.

[0012] If there is discharge, the phase of the synthesized signal is synchronized with the grid signal, and every three cycles of the synchronized synthesized signal are superimposed and merged to form new sound signal data. ;

[0013] Calculate sound signal data separately The preset characteristic value of the sound signal received by each microphone;

[0014] Compare the preset characteristic values ​​with the known discharge type If the similarity exceeds the preset threshold, it is determined that there is a discharge type On the contrary, if the similarity does not exceed the preset threshold, it is determined that there is no discharge type. discharge.

[0015] As a preference, establish the signal representation of the signal received by N microphones at time t The methods include:

[0016] ;

[0017] in:

[0018] ;

[0019] , to Represent the coordinates of N microphones, , is the wavelength of the sound of partial discharge, represents white noise, to They represent the amplitudes of L partial discharge sound signals when they propagate to the microphone array.

[0020] Preferably, the method for establishing the array directivity function of N microphones includes:

[0021] ;

[0022] in Indicates the The distance between each microphone and the preset center microphone, Indicates the direction of focus The wave vector of the sound signal of a partial discharge, represents the wave vector in the incident direction, , , Set to 90°, 、 and represent the unit vectors of the x-axis, y-axis, and z-axis respectively, is the wavelength of the sound of partial discharge, Set to 0°.

[0023] As a preference, cross-correlation calculation is performed to obtain the horizontal azimuth angle of each partial discharge sound source. and pitch angle The methods include:

[0024] Select the preset center microphone as the reference microphone and compare the signal collected by the reference microphone with the The time delay is obtained by performing cross-correlation calculation on the signals collected by the microphones to be tested. , then the horizontal azimuth ,in represents the speed of sound, Indicates the horizontal distance and elevation angle of the microphone from the reference microphone. It can be expressed as ,in Indicates the vertical distance between the microphone and the reference microphone.

[0025] As a preference, based on the synthetic signal received by N microphones at time t Methods for mapping PRPD include:

[0026] calculate , with t as the horizontal axis Build a PRPD map for the vertical axis, where Indicates that Each row of elements is added together to form an N*1 matrix. Indicates that Add up all the elements of .

[0027] Preferably, the preset characteristic values ​​include discharge symmetry S, signal skewness , signal kurtosis R, signal average value and signal variance value, and compare the preset characteristic value with the known discharge type The representative eigenvalue comparison methods include:

[0028] Perform mRMR calculation on multiple preset characteristic values ​​to obtain preset The most representative eigenvalue is recorded as the representative eigenvalue;

[0029] Establish a discriminant, the discriminant is ,in ,in is the kkth microphone receiving the sound signal to be identified. represents the eigenvalue, For known discharge types No. Representative eigenvalues, For the represents the weight of the eigenvalue;

[0030] like >0, it is determined that a discharge type has occurred On the contrary, if ≤0, it is determined that no discharge type occurs discharge.

[0031] Preferably, the method for calculating the discharge symmetry S is: ;

[0032] Calculating signal skewness The method is: ,in for The mean of for The variance of the function Indicates the target expected value;

[0033] Calculate signal kurtosis The method is: .

[0034] Preferably, the method for calculating the phase difference between adjacent peaks of the composite signal includes:

[0035] Intercept the first four cycles of the synthesized signal after phase synchronization and find the peak value of each cycle signal;

[0036] The phase differences between the four peaks are calculated. If the phase differences between the peaks are all outside the preset range, it is determined that no discharge occurs. Conversely, if the phase differences between the four peaks are within the preset range, it is determined that there is a discharge.

[0037] Preferably, the preset range corresponding to the phase difference is [330°, 390°].

[0038] Preferably, the method for phase synchronization of the composite signal and the grid signal includes: taking the phase of the first peak as a reference, multiplying the subsequent phase by a scaling factor so that the second peak moves to a phase difference of 85° relative to the first peak, thereby completing the phase synchronization of the composite signal and the grid signal.

[0039] The beneficial technical effects of the present invention include: by establishing signal representation and array directivity functions, the functional representation of the microphone array signal is realized, and then cross-correlation calculation is performed to obtain the horizontal azimuth and pitch angle of each local discharge sound source, thereby realizing the discovery and positioning of local discharge conditions; by establishing representative characteristic values ​​and comparing them with known discharge types, the discharge type of local discharge can be identified, providing a reference for identifying and handling faults.

[0040] Other features and advantages of the present invention will be disclosed in detail in the following specific embodiments and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The present invention will be further described below with reference to the accompanying drawings:

[0042] Figure 1 2 is a flow chart of a method for identifying multi-source partial discharge according to an embodiment of the present invention.

[0043] Figure 2 Schematic diagram of a flow chart of a method for calculating phase differences between adjacent peak values ​​of a composite signal according to an embodiment of the present invention.

[0044] Figure 3 This is a schematic diagram of the PRPD spectrum of suspended discharge according to an embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram of the PRPD spectrum of corona discharge in an embodiment of the present invention.

[0046] Figure 5 This is a schematic diagram of the PRPD spectrum of surface discharge in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following is an explanation and description of the technical solutions of the embodiments of the present invention in conjunction with the drawings of the embodiments of the present invention. However, the following embodiments are only preferred embodiments of the present invention and are not exhaustive. Based on the embodiments in the implementation manner, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention.

[0048] In the following description, terms such as "inside", "outside", "up", "down", "left", "right", etc. that indicate directions or positional relationships are only used to facilitate the description of the embodiments and simplify the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0049] A multi-source partial discharge identification method based on PRPD spectrum, please refer to the attached Figure 1 , including the following steps:

[0050] Step A01) Read the sound signal received by the array of N microphones, and record the number of partial discharges as , the horizontal azimuth and elevation angles of partial discharge are respectively and , ;

[0051] Step A02) Create a signal representation of the signals received by N microphones at time t and array directivity function;

[0052] Step A03) Perform cross-correlation calculation on the sound signals received by N microphones to obtain the horizontal azimuth angle of each partial discharge sound source and pitch angle ;

[0053] Step A04) Based on the horizontal azimuth angle of each partial discharge sound source and pitch angle Get the composite signal received by N microphones at time t ;

[0054] Step A05) Based on the composite signal received by N microphones at time t Draw PRPD maps;

[0055] Step A06) Calculating the phase difference between adjacent peaks of the composite signal. If the phase difference between adjacent peaks is within a preset range, it is determined that a discharge has occurred. Conversely, if the phase difference between adjacent peaks is not within the preset range, it is determined that no discharge has occurred.

[0056] Step A07) If there is discharge, the synthesized signal is synchronized with the grid signal, and every three cycles of the synchronized synthesized signal are superimposed and merged to form new sound signal data. ;

[0057] Step A08) Calculate the sound signal data separately The preset characteristic value of the sound signal received by each microphone;

[0058] Step A09) Compare the preset characteristic value with the known discharge type If the similarity exceeds the preset threshold, it is determined that there is a discharge type On the contrary, if the similarity does not exceed the preset threshold, it is determined that there is no discharge type. discharge.

[0059] Among them, the signal received by N microphones at time t is represented as The methods include:

[0060] ;

[0061] in:

[0062] ;

[0063] , to Represent the coordinates of N microphones, , is the wavelength of the sound of partial discharge, represents white noise, to They represent the amplitudes of L partial discharge sound signals when they propagate to the microphone array.

[0064] On the other hand, this embodiment provides a method for establishing an array directivity function of N microphones, including:

[0065] ;

[0066] in Indicates the The distance between each microphone and the preset center microphone, Indicates the direction of focus The wave vector of the sound signal of a partial discharge, represents the wave vector in the incident direction, , , Set to 90°, 、 and represent the unit vectors of the x-axis, y-axis, and z-axis respectively, is the wavelength of the sound of partial discharge, Set to 0°.

[0067] Angle of incident direction Setting it to 90° means that the microphone array only receives sound from the front.

[0068] Perform cross-correlation calculation to obtain the horizontal azimuth angle of each partial discharge sound source and pitch angle The methods include:

[0069] Select the preset center microphone as the reference microphone and compare the signal collected by the reference microphone with the The time delay is obtained by performing cross-correlation calculation on the signals collected by the microphones to be tested. , then the horizontal azimuth ,in represents the speed of sound, Indicates the horizontal distance and pitch angle of the microphone from the reference microphone It can be expressed as ,in Indicates the vertical distance between the microphone and the reference microphone.

[0070] Based on the composite signal received by N microphones at time t Methods for mapping PRPD include:

[0071] calculate , with t as the horizontal axis Build a PRPD map for the vertical axis, where Indicates that Each row of elements is added together to form an N*1 matrix. Indicates that Add up all the elements of .

[0072] The preset characteristic values ​​include discharge symmetry S, signal skewness , signal kurtosis R, signal average value and signal variance value, and compare the preset characteristic value with the known discharge type The representative eigenvalue comparison methods include:

[0073] Perform mRMR calculation on multiple preset characteristic values ​​to obtain preset The most representative eigenvalue is recorded as the representative eigenvalue;

[0074] Establish a discriminant, the discriminant is ,in ,in is the kkth microphone receiving the sound signal to be identified. represents the eigenvalue, For known discharge types No. Representative eigenvalues, For the represents the weight of the eigenvalue;

[0075] like >0, it is determined that a discharge type has occurred On the contrary, if ≤0, it is determined that no discharge type occurs discharge.

[0076] The mRMR operation (Max-Relevance and Min-Redundancy) refers to a maximum relevance and minimum redundancy algorithm. It finds a set of features within the original feature set that has the highest relevance (Max-Relevance) for the final output, but the lowest relevance (Min-Redundancy) for each other. The mRMR calculation process is conventionally used. Simply providing features with associated results, the publicly available calculation steps can be followed to obtain a feature ranking based on maximum relevance and minimum redundancy. In this example, the first four features are selected as representative feature values.

[0077] The method for calculating the discharge symmetry S is: ;

[0078] Calculating signal skewness The method is: ,in for The mean of for The variance of the function Indicates the target expected value;

[0079] Calculate signal kurtosis The method is: .

[0080] Please see the attached Figure 2 , the method for calculating the phase difference between adjacent peaks of the synthetic signal includes:

[0081] Step B01) intercepting the first four cycles of the phase-synchronized composite signal and finding the peak value of each cycle signal;

[0082] Step B02) Calculate the phase differences between the four peaks. If the phase differences between the peaks are not within the preset range, it is determined that there is no discharge. Conversely, if the phase differences between the four peaks are within the preset range, it is determined that there is discharge.

[0083] The preset range for the phase difference is [330°, 390°]. Calculate the phase difference between peaks. If the phase difference between the maximum peaks of adjacent cycles is not between 330 and 390 degrees, it is considered that there is no discharge in these four cycles, and the next four cycle signals are intercepted.

[0084] The method for synchronizing the phase of the composite signal with the grid signal includes: using the phase of the first peak as a reference, multiplying the subsequent phase by a scaling factor so that the second peak is shifted to a phase difference of 85 degrees relative to the first peak, thereby completing the phase synchronization of the composite signal with the grid signal. Given the 50 Hz operating frequency of the national grid, the received signal can be operated in cycles of 20 milliseconds. The received acoustic signal and the grid signal are phase-synchronized according to the operating cycle. If the phase difference between the maximum peaks of adjacent cycles is within 330-390 degrees, it is assumed that a discharge signal exists within four cycles. The phase of the peak of the second cycle is then calculated, and the data of the first cycle is used as a phase buffer area. The phase of the second peak is shifted to a phase of 85 degrees relative to its cycle. After completing the above steps, the grid signal and the acoustic signal are considered to be phase-synchronized or differ by an integer number of cycles.

[0085] This embodiment provides several examples of discharge types, including suspension discharge, corona discharge, and surface discharge. Suspension discharge refers to the circuit not being effectively grounded. Suspension discharge can lead to voltage instability due to the lack of a voltage zero point, and may cause discharge phenomena in weak insulation. The PRPD spectrum of suspension discharge is shown in the figure below. Figure 3 As shown. Corona discharge refers to the local self-sustaining discharge of a gas medium in an inhomogeneous electric field, and is the most common form of gas discharge. Near a sharp electrode with a very small radius of curvature, the local electric field strength exceeds the ionization field strength of the gas, causing the gas to ionize and excite, resulting in corona discharge. When corona occurs, light can be seen around the electrode, accompanied by a hissing sound. Corona discharge can be a relatively stable discharge form, or it can be an early development stage in the breakdown process of an inhomogeneous electric field gap. The PRPD spectrum of corona discharge is shown as follows: Figure 4 shown.

[0086] Surface discharge refers to the discharge phenomenon along the interface of dielectrics with different aggregate states. What usually occurs more frequently is the discharge along the surface of solid dielectrics in gas or liquid dielectrics. All charged conductors cannot be suspended in the atmosphere, and they must be suspended or supported by solid insulating devices. When a charged conductor needs to pass through a wall or an oil tank of an electrical equipment, it must also be fixed and insulated with a wall bushing or equipment bushing. These solid insulating devices have both mechanical fixing and electrical insulating functions. They are all surrounded by a gas medium, often with one electrode connected to a high voltage and the other electrode grounded. There are two possibilities for the loss of insulation function between the two poles: one is the breakdown of the same dielectric itself, and the other is the flashover along the surface of the solid dielectric. The PRPD spectrum of surface discharge is as follows: Figure 5 shown.

[0087] Under laboratory conditions, a partial discharge simulator is used to simulate the corresponding discharge type. An array of N microphones is used to collect sound signals and calculate the representative eigenvalues ​​of the sound signals. The actual sound signals are compared with the representative eigenvalues ​​of known discharge types to identify the corresponding discharge type.

[0088] The beneficial technical effects of this embodiment include: by establishing a signal representation and array directivity function, a functional representation of the microphone array signal is realized, and then cross-correlation calculation is performed to obtain the horizontal azimuth and pitch angle of each local discharge sound source, thereby realizing the discovery and positioning of the local discharge situation; by establishing a representative characteristic value and comparing it with a known discharge type, the discharge type of the local discharge can be identified, providing a reference for identifying and handling faults.

[0089] 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. Those skilled in the art will understand that the present invention includes, but is not limited to, the contents described in the drawings and the above specific embodiments. Any modifications that do not deviate from the functional and structural principles of the present invention are intended to be included within the scope of the claims.

Claims

1. A multi-source partial discharge identification method based on PRPD spectrum, characterized in that: The following steps are involved: Read the sound signal received by the array of N microphones, and the number of partial discharges is recorded as , the horizontal azimuth and elevation angles of partial discharge are respectively and , ; Establish the signal representation received by N microphones at time t and array directivity function; Perform cross-correlation calculation on the sound signals received by N microphones to obtain the horizontal azimuth angle of each partial discharge sound source and pitch angle ; According to the horizontal azimuth angle of each partial discharge sound source and pitch angle Get the composite signal received by N microphones at time t ; Based on the composite signal received by N microphones at time t Draw PRPD maps; Calculate the phase difference between adjacent peaks of the composite signal. If the phase difference between adjacent peaks is within a preset range, it is determined that there is a discharge. Otherwise, if the phase difference between adjacent peaks is not within the preset range, it is determined that there is no discharge. If there is discharge, the phase of the synthesized signal is synchronized with the grid signal, and every three cycles of the synchronized synthesized signal are superimposed and merged to form new sound signal data. ; Calculate sound signal data separately The preset characteristic value of the sound signal received by each microphone; Compare the preset characteristic values ​​with the known discharge type If the similarity exceeds the preset threshold, it is determined that there is a discharge type On the contrary, if the similarity does not exceed the preset threshold, it is determined that there is no discharge type. discharge.

2. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1 is characterized in that: Establish the signal representation received by N microphones at time t The methods include: ； in: ； , to Represent the coordinates of N microphones, , is the wavelength of the sound of partial discharge, represents white noise, to They represent the amplitudes of L partial discharge sound signals when they propagate to the microphone array.

3. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1 is characterized in that: The method of establishing the array directivity function of N microphones includes: ； in Indicates the The distance between each microphone and the preset center microphone, Indicates the direction of focus The wave vector of the sound signal of a partial discharge, represents the wave vector in the incident direction, , , Set to 90°, 、 and represent the unit vectors of the x-axis, y-axis, and z-axis respectively, is the wavelength of the sound of partial discharge, Set to 0°.

4. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1, characterized in that: Perform cross-correlation calculation to obtain the horizontal azimuth angle of each partial discharge sound source and pitch angle The methods include: Select the preset center microphone as the reference microphone and compare the signal collected by the reference microphone with the The time delay is obtained by performing cross-correlation calculation on the signals collected by the microphones to be tested. , then the horizontal azimuth ,in represents the speed of sound, Indicates the horizontal distance and pitch angle of the microphone from the reference microphone It can be expressed as ,in Indicates the vertical distance between the microphone and the reference microphone.

5. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1, characterized in that: Based on the composite signal received by N microphones at time t Methods for mapping PRPD include: calculate , with t as the horizontal axis Build a PRPD map for the vertical axis, where Indicates that Each row of elements is added together to form an N*1 matrix. Indicates that Add up all the elements of .

6. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1, characterized in that: The preset characteristic values ​​include discharge symmetry S, signal skewness , signal kurtosis R, signal mean and signal variance values, and compare the preset characteristic values ​​with the known discharge type The representative eigenvalue comparison methods include: Perform mRMR calculation on multiple preset characteristic values ​​to obtain preset The most representative eigenvalue is recorded as the representative eigenvalue; Establish a discriminant, the discriminant is ,in ,in is the kkth microphone receiving the sound signal to be identified. represents the eigenvalue, For known discharge types No. Representative eigenvalues, For the represents the weight of the eigenvalue; like >0, it is determined that a discharge type has occurred On the contrary, if ≤0, it is determined that no discharge type occurs discharge.

7. The multi-source partial discharge identification method based on PRPD spectrum according to claim 6, characterized in that: The method for calculating the discharge symmetry S is: ; Calculating signal skewness The method is: ,in for The mean of for The variance of the function Indicates the target expected value; Calculate signal kurtosis The method is: .

8. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1, characterized in that: Methods for calculating the phase difference between adjacent peaks of the composite signal include: Intercept the first four cycles of the synthesized signal after phase synchronization and find the peak value of each cycle signal; The phase differences between the four peaks are calculated. If the phase differences between the peaks are all outside the preset range, it is determined that no discharge occurs. Conversely, if the phase differences between the four peaks are within the preset range, it is determined that there is a discharge.

9. The multi-source partial discharge identification method based on PRPD spectrum according to claim 1, characterized in that: The preset range corresponding to the phase difference is [330°, 390°].

10. The multi-source partial discharge identification method based on PRPD spectrum according to claim 8, characterized in that: The method for phase-synchronizing the composite signal with the grid signal includes: taking the phase of the first peak as a reference, multiplying the subsequent phase by a scaling factor so that the second peak moves to a phase difference of 85° relative to the first peak, thereby completing the phase synchronization of the composite signal with the grid signal.

Images