A Method and System for Localizing Partial Discharge in Substations Based on Directional Sensor Arrays

Abstract

This invention provides a method and system for locating partial discharge in substations based on a directional sensor array, comprising: Step S1: using an array of N directional antennas to receive and sample partial discharge signals from the partial discharge source to obtain a signal sequence x1, x2, ..., x N Based on the signal sequence x1,x2,…,x N Obtain the measurement matrix X; Step S2: Calculate the noise matrix E of the signal based on the measurement matrix X. n Step S3: Based on the antenna gain pattern function g(θ), according to the signal noise matrix E n The spatial spectral function P was calculated. mu (θ); Step S4: P is adjusted from 0° to 360°. mu (θ) performs a maximum search, where the angle corresponding to the maximum value is the directional angle of the local discharge source. This invention is based solely on signal strength information, has a fast calculation speed, does not have high sampling requirements, and has high directional accuracy. Furthermore, the directional antenna used is small in size, resulting in low equipment cost.

Description

Technical Field

[0001] This invention relates to a method for locating partial discharge in substations, specifically, to a method for locating partial discharge in substations based on a directional sensor array. Background Technology

[0002] Insulation defects in electrical equipment caused by manufacturing flaws, aging processes, and other factors are one of the main causes of equipment downtime or electrical accidents. Partial discharge (PD) is a discharge phenomenon caused by localized breakdown in the insulating medium. This discharge creates a localized short circuit between conductors without forming a conductive path. Each partial discharge affects the insulation strength, and over time, this eventually leads to insulation collapse, causing power outages or even system disconnection.

[0003] As a hub for power collection and distribution, substations require rigorous equipment condition inspection to ensure economic efficiency and reliability. Due to the large number of electrical devices in substations, traditional manual inspection methods suffer from low efficiency, limited testing methods, and high labor costs. The development of smart substations has placed higher demands on substation inspection. Intelligent inspection robots are an effective alternative to traditional methods and have broad application prospects. Existing inspection robots generally utilize infrared and high-definition cameras for temperature and condition detection. However, due to technological limitations, no inspection robot currently possesses the capability to effectively detect and locate partial discharges. Therefore, researching a miniaturized partial discharge location system that can be mounted on an inspection robot has both theoretical research value and significant engineering application value.

[0004] Currently, among partial discharge detection methods, ultra-high frequency (UHF) detection is widely used due to its strong anti-interference capability and high sensitivity. However, for substations, given the large number and variety of equipment and the low failure rate of individual devices, individual device detection and location suffer from low efficiency and high cost. To address this issue, researchers have proposed using omnidirectional UHF sensor arrays for station-area partial discharge detection and location, yielding significant results. Based on whether the omnidirectional antenna array is fixed, detection systems are categorized into fixed and vehicle-mounted systems. Fixed antenna array partial discharge detection systems are limited by the array's installation location, and their discharge detection sensitivity and location capabilities still have room for improvement. Vehicle-mounted partial discharge inspection systems require excessive personnel involvement, have long inspection cycles, and are difficult to implement for long-term online monitoring. Existing location algorithms rely on time difference information for positioning; constrained by the algorithm's principles, the system's positioning accuracy depends on the accuracy of the time difference calculation. Furthermore, the detection sensitivity of the equipment needs further improvement due to the limitations of omnidirectional antenna characteristics.

[0005] To address the problems of excessively large array size, difficulty in time delay calculation, high equipment cost, and low detection sensitivity in existing UHF positioning methods, and in light of the development trend of inspection robots, researching a lightweight, small-volume, high-gain antenna array that can be mounted on an inspection robot and a partial discharge positioning method based on signal strength information has potential application value and practical significance.

[0006] Patent document CN104198901A (application number: 201410398505.8) discloses a method and system for locating partial discharge signals in a substation. The method includes the following steps: acquiring partial discharge signals generated in the substation using at least four ultra-high frequency (UHF) sensors; denoising the partial discharge signals using a preset complex wavelet function; calculating the time delay of the denoised partial discharge signals at different UHF sensors using a preset higher-order cumulant function; and calculating the generation location of the partial discharge signals based on the time delays and the installation positions of each UHF sensor. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for locating partial discharge in substations based on a directional sensor array.

[0008] A substation partial discharge localization method based on a directional sensor array, according to the present invention, includes:

[0009] Step S1: Use an array of N directional antennas to receive and sample the partial discharge signals from the partial discharge source to obtain the signal sequence x1,x2,…,x N Based on the signal sequence x1,x2,…,x N Obtain the measurement matrix X;

[0010] Step S2: Calculate the noise matrix E of the signal based on the measurement matrix X. n ;

[0011] Step S3: Based on the antenna gain pattern function g(θ), according to the signal noise matrix E n The spatial spectral function P was calculated. mu (θ);

[0012] Step S4: Adjust P from 0° to 360° mu (θ) performs a maximum search, and the angle corresponding to the maximum value is the direction angle of the local discharge source.

[0013] Preferably, step S1 involves: sampling the signal sequence x1, x2, ..., x from the directional antenna array. N Stacked as a measurement matrix X;

[0014]

[0015] Wherein, signal sequence x i It is a 1×L vector, where L is the number of sampling points; T represents the matrix transpose.

[0016] Preferably, step S2 employs:

[0017] Step S2.1: Calculate the measurement covariance matrix R based on the measurement matrix X. X ;

[0018] R X =XX T (2)

[0019] Where T represents the matrix transpose;

[0020] Step S2.2: Based on the measurement covariance matrix R X Calculate R X eigenvectors λ1,λ2,…,λ N and the corresponding eigenvectors ν1,ν2,…,ν N ;

[0021] [EV,D]=eig(R X (3)

[0022] Where eig represents the function in Matlab for calculating eigenvalues ​​and eigenvectors, and D is an N×N diagonal matrix with diagonal elements λ1, λ2, ..., λ3. N EV=[ν1,ν2,…,ν N [A matrix consisting of N corresponding eigenvectors;]

[0023] Step S2.3: Set λ1, λ2, ..., λ N Sort by size from largest to smallest. At the same time, the column vectors of EV are reordered as follows:

[0024] Step S2.4: Delete the first column of matrix EV to obtain the noise matrix E n ;

[0025] E n =EV(:,2:N) (4).

[0026] Preferably, step S3 employs the following methods:

[0027] Step S3.1: Determine the antenna gain pattern function g(θ), which is based on the measured gain pattern and is represented by a third-order Gaussian function;

[0028]

[0029] Where a1, a2, a3, b1, b2, b3, c1, c2, and c3 all represent normalized fitting coefficients; θ represents the direction angle.

[0030] Step S3.2: Uniformly sample the orientation angles from 0° to 360° to construct an array manifold G(θ) for each θ.

[0031] G(θ) = [g1(θ), g2(θ), ..., g N (θ)] T (6)

[0032] Step S3.3: Based on the noise matrix E n And the guiding vector G(θ), and calculate its spatial spectral function P for each θ. mu (θ), its formula is (7):

[0033]

[0034] Where T represents the transpose of the matrix.

[0035] Preferably, step S4 employs the following methods:

[0036] P mu (θ m ) = maxP mu (θ) (8)

[0037] Where, θ m This represents an estimated value of the direction angle of the local discharge power source.

[0038] A substation partial discharge location system based on a directional sensor array, according to the present invention, includes:

[0039] Module M1: Utilizes an array of N directional antennas to receive and sample partial discharge signals from a partial discharge source to obtain a signal sequence x1, x2, ..., x N Based on the signal sequence x1,x2,…,x N Obtain the measurement matrix X;

[0040] Module M2: Calculates the noise matrix E of the signal based on the measurement matrix X. n ;

[0041] Module M3: Based on the antenna gain pattern function g(θ), and according to the signal noise matrix E n The spatial spectral function P was calculated. mu (θ);

[0042] Module M4: P can be positioned from 0° to 360°. mu (θ) performs a maximum search, and the angle corresponding to the maximum value is the direction angle of the local discharge source.

[0043] Preferably, module M1 employs: sampling the signal sequence x1, x2, ..., x obtained from the directional antenna array. N Stacked as a measurement matrix X;

[0044]

[0045] Wherein, signal sequence x i It is a 1×L vector, where L is the number of sampling points; T represents the matrix transpose.

[0046] Preferably, the module M2 adopts:

[0047] Module M2.1: Calculate the measurement covariance matrix R based on the measurement matrix X. X ;

[0048] R X =XX T (2)

[0049] Where T represents the matrix transpose;

[0050] Module M2.2: Based on the measurement covariance matrix R X Calculate R X eigenvectors λ1,λ2,…,λ N and the corresponding eigenvectors ν1,ν2,…,ν N ;

[0051] [EV,D]=eig(R X (3)

[0052] Where eig represents the function in Matlab for calculating eigenvalues ​​and eigenvectors, and D is an N×N diagonal matrix with diagonal elements λ1, λ2, ..., λ3. N EV=[ν1,ν2,…,ν N [A matrix consisting of N corresponding eigenvectors;]

[0053] Module M2.3: λ1, λ2, ..., λ N Sort by size from largest to smallest. At the same time, the column vectors of EV are reordered as follows:

[0054] Module M2.4: Delete the first column of matrix EV to obtain the noise matrix E n ;

[0055] E n =EV(:,2:N) (4).

[0056] Preferably, the module M3 adopts:

[0057] Module M3.1: Determines the antenna gain pattern function g(θ), which is based on the measured gain pattern and is represented by a third-order Gaussian function;

[0058]

[0059] Where a1, a2, a3, b1, b2, b3, c1, c2, and c3 all represent normalized fitting coefficients; θ represents the direction angle.

[0060] Module M3.2: Uniformly sample the orientation angle from 0° to 360° and construct an array manifold G(θ) for each θ.

[0061] G(θ)=[g1(θ), g2(θ),…,g N (θ)] T (6)

[0062] Module M3.3: Based on the noise matrix E n And the guiding vector G(θ), and calculate its spatial spectral function P for each θ. mu (θ), its formula is (7):

[0063]

[0064] Where T represents the transpose of the matrix.

[0065] Preferably, the module M4 adopts:

[0066] P mu (θ m ) = max P mu (θ) (8)

[0067] Where, θ m This represents an estimated value of the direction angle of the local discharge power source.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] 1. This invention uses the intensity information of partial discharge signals for positioning, which reduces the sampling rate requirement for positioning compared to positioning using time difference information;

[0070] 2. This invention uses a directional antenna array, which has the advantages of smaller size and lighter weight compared with an omnidirectional antenna array. Therefore, it can be used in handheld partial discharge detection equipment and substation inspection robots, and the equipment cost is greatly reduced accordingly.

[0071] 3. The positioning results of this invention have high accuracy, and still have high positioning accuracy in low signal-to-noise ratio environments. Attached Figure Description

[0072] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0073] Figure 1 This is a flowchart of a substation partial discharge localization method based on a directional sensor array.

[0074] Figure 2 This is a schematic diagram of an actual antenna array.

[0075] Figure 3 This is the measured gain pattern of the antenna.

[0076] Figure 4 This is a diagram showing the layout of the experimental site. Detailed Implementation

[0077] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0078] Example 1

[0079] According to the present invention, a method for locating partial discharge in a substation based on a directional sensor array is provided, such as... Figure 1 As shown, it includes:

[0080] Step S1: Use an array of N directional antennas to receive and sample the partial discharge signals from the partial discharge source to obtain the signal sequence x1,x2,…,x N Based on the signal sequence x1,x2,…,x N Obtain the measurement matrix X;

[0081] Specifically, step S1 involves: sampling the signal sequence x1, x2, ..., x from the directional antenna array. N Stacked as a measurement matrix X;

[0082]

[0083] Wherein, signal sequence x i It is a 1×L vector, where L is the number of sampling points; T represents the matrix transpose.

[0084] Step S2: Calculate the noise matrix E of the signal based on the measurement matrix X. n ;

[0085] Specifically, step S2 employs the following:

[0086] Step S2.1: Calculate the measurement covariance matrix R based on the measurement matrix X. X ;

[0087] R X =XX T (2)

[0088] Where T represents the matrix transpose;

[0089] Step S2.2: Based on the measurement covariance matrix R X Calculate R X eigenvectors λ1,λ2,…,λ N and the corresponding eigenvectors ν1,ν2,…,ν N ;

[0090] [EV,D]=eig(R X (3)

[0091] Where eig represents the function in Matlab for calculating eigenvalues ​​and eigenvectors, and D is an N×N diagonal matrix with diagonal elements λ1, λ2, ..., λ3. N EV = [ν1, ν2, ..., ν N [A matrix consisting of N corresponding eigenvectors;]

[0092] Step S2.3: Set λ1, λ2, ..., λ N Sort by size from largest to smallest. At the same time, the column vectors of EV are reordered as follows:

[0093] Step S2.4: Delete the first column of matrix EV to obtain the noise matrix E n ;

[0094] E n =EV(:,2:N) (4).

[0095] Where: : represents all rows; 2:N represents columns 2 to N;

[0096] Step S3: Based on the antenna gain pattern function g(θ), according to the signal noise matrix E n The spatial spectral function P was calculated. mu (θ);

[0097] Specifically, step S3 employs the following:

[0098] Step S3.1: Determine the antenna gain pattern function g(θ), which is based on the measured gain pattern and is represented by a third-order Gaussian function;

[0099]

[0100] Where a1=0.5255, b1=218.1, c1=51.73, a2=0.3405, b2=304.8, c2=41, a3=0.6251, b3=156.1, c3=109.1 represent normalized fitting coefficients; θ represents the direction angle;

[0101] Step S3.2: Uniformly sample the orientation angles from 0° to 360°. Taking 1° resolution as an example, let θ = 1:1:360, with the unit being degrees, and construct an array manifold G(θ) for each θ.

[0102] G(θ) = [g1(θ), g2(θ), ..., g N (θ)] T (6)

[0103] Among them, g i It is the gain pattern function of the i-th antenna; they differ only in orientation.

[0104] Step S3.3: Based on the noise matrix E n And the guiding vector G(θ), and calculate its spatial spectral function P for each θ. mu (θ), its formula is (7):

[0105]

[0106] Where T represents the transpose of the matrix.

[0107] Step S4: Adjust P from 0° to 360° mu (θ) performs a maximum search, and the angle corresponding to the maximum value is the direction angle of the local discharge source.

[0108] Specifically, step S4 employs the following:

[0109]

[0110] Where, θ m This represents an estimated value of the direction angle of the local discharge power source.

[0111] A substation partial discharge location system based on a directional sensor array, according to the present invention, includes:

[0112] Module M1: Utilizes an array of N directional antennas to receive and sample partial discharge signals from a partial discharge source to obtain a signal sequence x1, x2, ..., x N Based on the signal sequence x1,x2,…,x N Obtain the measurement matrix X;

[0113] Specifically, module M1 employs: sampling the signal sequence x1, x2, ..., x from the directional antenna array. N Stacked as a measurement matrix X;

[0114]

[0115] Wherein, signal sequence x i It is a 1×L vector, where L is the number of sampling points; T represents the matrix transpose.

[0116] Module M2: Calculates the noise matrix E of the signal based on the measurement matrix X. n ;

[0117] Specifically, module M2 adopts:

[0118] Module M2.1: Calculate the measurement covariance matrix R based on the measurement matrix X. X ;

[0119] R X =XX T (2)

[0120] Where T represents the matrix transpose;

[0121] Module M2.2: Based on the measurement covariance matrix R X Calculate R X eigenvectors λ1, λ2, ..., λ N and the corresponding eigenvectors ν1, ν2, ..., ν N ;

[0122] [EV, D] = eig(R) X (3)

[0123] Where eig represents the function in Matlab for calculating eigenvalues ​​and eigenvectors, and D is an N×N diagonal matrix with diagonal elements λ1, λ2, ..., λ N EV=[ν1,ν2,…,ν N [A matrix consisting of N corresponding eigenvectors;]

[0124] Module M2.3: λ1, λ2, ..., λ N Sort by size from largest to smallest. At the same time, the column vectors of EV are reordered as follows:

[0125] Module M2.4: Delete the first column of matrix EV to obtain the noise matrix E n ;

[0126] E n =EV(:,2:N) (4).

[0127] Where: : represents all rows; 2:N represents columns 2 to N;

[0128] Module M3: Based on the antenna gain pattern function g(θ), and according to the signal noise matrix E n The spatial spectral function P was calculated. mu (θ);

[0129] Specifically, module M3 adopts:

[0130] Module M3.1: Determines the antenna gain pattern function g(θ), which is based on the measured gain pattern and is represented by a third-order Gaussian function;

[0131]

[0132] Where a1=0.5255, b1=218.1, c1=51.73, a2=0.3405, b2=304.8, c2=41, a3=0.6251, b3=156.1, c3=109.1 represent normalized fitting coefficients; θ represents the direction angle;

[0133] Module M3.2: Uniform sampling is performed on the orientation angle from 0° to 360°. Taking 1° resolution as an example, let θ = 1:1:360, with the unit being degrees, and construct an array manifold G(θ) for each θ.

[0134] G(θ)=[g1(θ),g2(θ),…,g N (θ)] T (6)

[0135] Among them, g i It is the gain pattern function of the i-th antenna; they differ only in orientation.

[0136] Module M3.3: Based on the noise matrix E n And the guiding vector G(θ), and calculate its spatial spectral function P for each θ. mu (θ), its formula is (7):

[0137]

[0138] Where T represents the transpose of the matrix.

[0139] Module M4: P can be positioned from 0° to 360°. mu (θ) performs a maximum search, and the angle corresponding to the maximum value is the direction angle of the local discharge source.

[0140] Specifically, module M4 adopts:

[0141]

[0142] Where, θ m This represents an estimated value of the direction angle of the local discharge power source.

[0143] Example 2

[0144] Example 2 is a preferred example of Example 1.

[0145] The present invention provides a substation partial discharge localization method based on a directional sensor array. This method employs a circular uniform array of six directional antennas to monitor and locate partial discharge signals in the substation. Figure 2 As shown, the Dir-MUSIC algorithm based on intensity information is used to direct the local discharge source by taking advantage of the difference in received signal strength caused by different orientations of the directional antenna.

[0146] The gain pattern of the directional antenna was measured in a microwave anechoic chamber. The gain pattern of the antenna at 1.25G is shown below. Figure 3 As shown, the gain patterns of the antenna at other frequencies vary slightly, but their shapes are basically the same. The gain pattern exhibits good single-peak properties, which also demonstrates the feasibility of the positioning algorithm of this invention.

[0147] The signal sequences x1, x2, ..., x6 (each a 1×10000 vector, where 10000 represents the number of sampling points) obtained from the directional antenna array are stacked to form a measurement matrix X:

[0148]

[0149] Where T represents the matrix transpose.

[0150] Calculate the measurement covariance matrix R based on the measurement matrix X. X :

[0151] R X =XX T (2)

[0152] Based on the measurement covariance matrix R X Calculate R X The eigenvectors λ1,λ2,…,λ6 and their corresponding eigenvectors ν1,ν2,…,ν6 are:

[0153] [EV,D]=eig(R X (3)

[0154] Where eig is a function in Matlab to calculate eigenvalues ​​and eigenvectors, D is a 6×6 diagonal matrix with diagonal elements λ1, λ2, ..., λ6, and EV = [ν1, ν2, ..., ν6] is a matrix composed of N corresponding eigenvectors.

[0155] Sort λ1, λ2, ..., λ6 in descending order as follows: At the same time, the column vectors of EV are reordered as follows:

[0156] Deleting the first column of matrix EV yields the noise matrix E. n :

[0157] E n =EV(:,2:6)(4)

[0158] Determine the antenna gain pattern function g(θ), which is based on the measured gain pattern and is represented by a third-order Gaussian function:

[0159]

[0160] Where a1=0.5255, b1=218.1, c1=51.73, a2=0.3405, b2=304.8, c2=41, a3=0.6251, b3=156.1, c3=109.1 are normalized fitting coefficients.

[0161] Uniform sampling is performed on the orientation angle from 0° to 360°. Here, we take 1° resolution as an example for explanation. Let θ = 1:1:360, with the unit being degrees. Then, an array manifold G(θ) is constructed for each θ:

[0162] G(θ)=[g1(θ),g2(θ),…,g6(θ)] T (6)

[0163] Where g i It is the gain pattern function of the i-th antenna; they differ only in orientation.

[0164] According to the noise matrix E n And the guiding vector G(θ), and calculate its spatial spectral function P for each θ. mu (θ):

[0165]

[0166] Spatial spectral function P mu (θ) performs a maximum search:

[0167]

[0168] Where θ m This is the estimated value of the direction angle of the local discharge source.

[0169] This invention is first verified using simulation. 3600 incoming wave directions are randomly generated within a circular range [0°, 360°), and 3600 simulations are performed. The accuracy of the incoming wave direction estimation is defined as follows: if the error between the estimated angle and the actual angle is less than 2°, the positioning is considered successful; the accuracy is equal to the ratio of the number of successful positioning attempts to the total number of simulations. The accuracy rates under different signal-to-noise ratios are shown in the table below:

[0170] Table 1. Direction finding accuracy at different signal-to-noise ratios

[0171] Signal-to-noise ratio 10 5 0 -5 -10 accuracy 100％ 100％ 100％ 99.17％ 72.78％

[0172] The invention was then verified using field experiments, the field layout of which is shown in the diagram below. Figure 4 As shown, the partial discharge signal was simulated using an electrostatic gun. Table 2 summarizes the location results at 10 different locations:

[0173] Table 2. Field Test Positioning Results

[0174]

[0175] Simulation and experimental results fully demonstrate the effectiveness of the present invention.

[0176] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0177] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for locating partial discharge in a substation based on a directional sensor array, characterized in that, include: Step S1: Use an array of N directional antennas to receive and sample the partial discharge signal from the partial discharge source to obtain a signal sequence. Based on signal sequences Obtain the measurement matrix ; Step S2: Based on the measurement matrix Calculate the noise matrix of the signal ; Step S3: Based on the antenna gain pattern function According to the noise matrix of the signal The spatial spectral function was calculated. ; Step S4: In right Perform a maximum search; the angle corresponding to the maximum value is the direction angle of the local power source. Step S2 employs the following: Step S2.1: Based on the measurement matrix Calculate the measurement covariance matrix ; （2） Where T represents the matrix transpose; Step S2.2: Based on the measurement covariance matrix calculate eigenvectors and corresponding feature vectors ; （3） in, This represents a function in Matlab for calculating eigenvalues ​​and eigenvectors. for A diagonal matrix whose diagonal elements are , It is a matrix composed of N corresponding eigenvectors; Step S2.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require Sort by size from largest to smallest. At the same time The column vectors are reordered as follows ; Step S2.4: Delete the first column of matrix EV to obtain the noise matrix. ; （4）； Step S3 employs the following: Step S3.1: Determine the antenna gain pattern function , Based on the measured gain pattern, it is represented by a third-order Gaussian function; （5） in, , , , , , c1, c2, and c3 all represent normalized fitting coefficients; Indicates the direction angle; Step S3.2: Adjust the direction angle Perform uniform sampling to construct a model for each... Constructing an array manifold ; （6） Step S3.3: Based on the noise matrix and guide vector For each Calculate its spatial spectral function Its formula is (7): （7） Where T represents the transpose of the matrix; Step S4 employs the following: in, This represents an estimated value of the direction angle of the local discharge power source.

2. The substation partial discharge localization method based on a directional sensor array according to claim 1, characterized in that, Step S1 involves: sampling the signal sequence obtained from the directional antenna array. Stacked as a measurement matrix ; Among them, signal sequence for ;T represents matrix transpose.

3. A partial discharge location system for substations based on a directional sensor array, characterized in that, include: Module M1: Utilizes an array of N directional antennas to receive and sample partial discharge signals from the partial discharge source to obtain a signal sequence. Based on signal sequences Obtain the measurement matrix ; Module M2: Based on measurement matrix Calculate the noise matrix of the signal ; Module M3: Based on antenna gain pattern function According to the noise matrix of the signal The spatial spectral function was calculated. ; Module M4: In right Perform a maximum search; the angle corresponding to the maximum value is the direction angle of the local power source. The module M2 adopts: Module M2.1: Based on the measurement matrix Calculate the measurement covariance matrix ; （2） Where T represents the matrix transpose; Module M2.2: Based on the measurement covariance matrix calculate eigenvectors and corresponding feature vectors ; （3） in, This represents a function in Matlab for calculating eigenvalues ​​and eigenvectors. for A diagonal matrix whose diagonal elements are , It is a matrix composed of N corresponding eigenvectors; Module M2.3: will Sort by size from largest to smallest. At the same time The column vectors are reordered as follows ; Module M2.4: Delete the first column of matrix EV to obtain the noise matrix. ; （4）； The module M3 adopts: Module M3.1: Determine the antenna gain pattern function , Based on the measured gain pattern, it is represented by a third-order Gaussian function; （5） in, , , , , , c1, c2, and c3 all represent normalized fitting coefficients; Indicates the direction angle; Module M3.2: Direction Angle Perform uniform sampling to construct a model for each... Constructing an array manifold ; （6） Module M3.3: Based on the noise matrix and guide vector For each Calculate its spatial spectral function Its formula is (7): （7） Where T represents the transpose of the matrix; The module M4 adopts: in, This represents an estimated value of the direction angle of the local discharge power source.

4. The substation partial discharge location system based on a directional sensor array according to claim 3, characterized in that, The module M1 employs: the signal sequence sampled from the directional antenna array Stacked as a measurement matrix ; Among them, signal sequence for ;T represents matrix transpose.

Images