A partial discharge positioning method for insulators

Abstract

The application relates to the technical field of partial discharge source positioning, in particular to a partial discharge positioning method for insulators, which comprises the following steps: collecting positioning signals generated by partial discharge of an insulator by using a polygonal three-dimensional sensor array; decomposing and denoising the positioning signals to obtain time-domain positioning signals after denoising; obtaining frequency-domain positioning signals based on the time-domain positioning signals, carrying out weighted cross-correlation processing, and obtaining signal time delay; combining four sensors at random to establish a target function for solving, with each solution corresponding to a positioning initial value; clustering all the positioning initial values by using a clustering algorithm, and the clustering center of the cluster with the most elements in the clustered cluster being the position of the partial discharge of the insulator. The method can reduce errors that may occur in the positioning process, improve the reliability of overall positioning, and find a balance between global search and local search by dynamically adjusting the search speed of particles.

Description

Technical Field

[0001] This invention relates to the field of partial discharge source location technology, specifically a method for locating partial discharge in an insulator. Background Technology

[0002] During long-term operation, insulators may develop defects such as aging, contamination, or damage. These defects can lead to a decline in the insulation performance of the insulator and the generation of partial discharge. Partial discharge serves as an early warning signal of insulator faults and plays a crucial role in early warning. If not addressed or repaired promptly, partial discharge may progress, eventually causing flashover or even breakdown of the insulator. Therefore, accurately locating the site of partial discharge in an insulator is of great significance for preventing insulator faults.

[0003] Currently, commonly used methods for locating partial discharge insulators include infrared thermal imaging and ultrasonic positioning. Infrared thermal imaging uses an infrared thermal imager to scan the surface of the insulator and determines the location of the fault point by analyzing the generated thermal image. This method is simple in principle and easy to implement, but in practical applications it is easily affected by factors that can cause changes in the surface temperature of the insulator, such as sunlight, rain, snow, and strong winds.

[0004] Ultrasonic location methods determine the location of partial discharge sources based on the time difference or beam propagation direction of the ultrasonic signals generated by partial discharge reaching different sensors. However, because ultrasonic signals attenuate rapidly in air and their wave velocity is unstable, the sensitivity of ultrasonic detection is greatly reduced, often requiring the use of other techniques. Therefore, an effective method is urgently needed to determine the location of partial discharge in insulators. Summary of the Invention

[0005] To address the technical problems existing in the background art, the present invention provides a method for locating partial discharge in insulators.

[0006] The technical solution of this invention is as follows: A method for locating partial discharge in an insulator, the specific method includes the following steps: S1. A polygonal three-dimensional sensor array is used to collect the positioning signals generated by partial discharge of the insulator.

[0007] S2. The acquired positioning signal is adaptively decomposed from high to low frequency to obtain intrinsic mode function components at different frequencies. The correlation coefficient of each intrinsic mode function component is calculated, and correlation coefficients greater than the correlation coefficient threshold are selected as high-quality coefficients. Wavelet packet decomposition is performed on the intrinsic mode function components corresponding to the high-quality coefficients. The wavelet packet coefficients obtained by wavelet packet decomposition are processed using an adjustable threshold function to reconstruct the denoised high-quality intrinsic mode function components. All denoised high-quality intrinsic mode function components are synthesized to obtain the denoised time-domain positioning signal.

[0008] S3. Based on the denoised time-domain positioning signal, the frequency-domain positioning signal is obtained by Fourier transform. Based on any two frequency-domain positioning signals, weighted cross-correlation processing is performed to obtain the signal delay between the two sensors corresponding to any two frequency-domain positioning signals in the polygonal stereo sensor array.

[0009] S4. Randomly select four sensors from the polygonal 3D sensor array and combine them. Based on the signal delay and the distance from the positioning signal to the sensor, establish the objective function corresponding to each combination. Randomly initialize a group of particles, with each particle representing a potential solution to the objective function. Use a combination of particle swarm optimization and gravity search algorithm to update the velocity and position of the particles. Repeat the update steps and calculate the objective function value after each update. After a preset number of iterations, output the global optimal solution, which is the solution to the objective function. Each solution to the objective function corresponds to an initial positioning value.

[0010] S5. Use a clustering algorithm to cluster all initial positioning values. The cluster center of the cluster with the most elements in the cluster is the partial discharge location of the insulator.

[0011] When updating particle velocities using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: , in, and These represent the velocities of the i-th particle in the (u+1)-th and u-th iterations, respectively, where u is the current iteration number. Indicates inertia weight, and As a learning factor, and Let be a random variable between [0, 1]. This represents the best historical position of the i-th particle. It is the globally optimal position among all particles. As the gravitational factor, This represents the position of the i-th particle in the u-th iteration. Let be the acceleration of the i-th particle during the u-th iteration.

[0012] Gravity factor The calculation formula is as follows; , in, The minimum gravitational factor, The maximum gravitational factor, This is the preset number of iterations.

[0013] Inertia weight The specific calculation method is as follows; , in, Indicates the maximum inertia weight. This represents the minimum inertia weight. The preset number of iterations is given, and u is the current number of iterations.

[0014] When updating the position of a particle using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: , in, and These represent the positions of the i-th particle in the (u+1)-th and u-th iterations, respectively. Let represent the velocity of the i-th particle in the (u+1)-th iteration, where u is the current iteration number.

[0015] Choose any four sensors , , , Form combinations, and with The objective function established using a reference sensor for: , In the formula, The speed at which the positioning signal travels through the air is given by a value of [value]. , For the first The time difference between when each sensor receives the positioning signal and when the reference sensor receives the positioning signal, i.e., the delay. Indicates that the power supply is discharged to the first The distance between the sensors, when h takes the values ​​1, 2, 3, and 4, yields the result. , , and , , , , Among them, the power supply is to the first The formula for calculating the distance between the sensors is: , In the formula, x, y, and z are the coordinate values ​​of the discharge source on the x, y, and z axes, respectively. h y h z h These are the coordinates of the h-th sensor on the x, y, and z axes, respectively.

[0016] In S3, weighted cross-correlation processing is performed based on any two frequency domain positioning signals, and the specific weighting function is as follows: , in, For weighted functions, Angular frequency, and Let m and n represent the self-power spectral density of the two frequency domain positioning signals, respectively, and m and n represent the sensor serial numbers, respectively. These are weighting coefficients, and .

[0017] The polygonal stereo sensor array can be a cone array, sphere array, or cylinder array, and the number of sensors N in the polygonal stereo sensor array is N ≥ 5.

[0018] The method for calculating the correlation coefficient in S2 is as follows; , in, This represents the correlation coefficient of the i-th intrinsic mode function component. L It is the length of the positioning signal. It is the average value of the positioning signal. It is the mean of the i-th intrinsic mode function component. Indicates the location signal is in The value at time, Indicates that the i-th intrinsic mode function component is in The value at time, Indicates the time.

[0019] The number of clusters in S5 is selected according to the elbow rule.

[0020] Compared with the prior art, the present invention has the following beneficial effects: By designing an N-element polygonal 3D sensor array, signal focusing can be achieved in the spatial domain, thereby increasing the received signal strength and reducing the impact of background noise. The collaborative operation of multiple sensors effectively improves positioning accuracy. When one sensor malfunctions or is interfered with, other sensors can still provide valid information, reducing potential errors during positioning and enhancing overall positioning reliability.

[0021] Adaptive decomposition of the positioning signal allows for independent analysis of intrinsic mode function (IMF) components at different frequencies. Wavelet packet decomposition, performed by selecting high-quality IMF components with high correlation coefficients, enables more accurate extraction of key signal features and reduces redundant information. An adjustable threshold function allows for flexible adjustment based on actual signal characteristics during the denoising process, significantly improving denoising performance.

[0022] The weighting function considers signals from multiple channels, thus effectively utilizing the correlation between these channels and enhancing signal processing performance. By adjusting the weighting coefficients, the main components of the signal can be highlighted, particularly in cases of low signal-to-noise ratio, thereby improving the ability to estimate signal delay.

[0023] By dynamically adjusting the search speed of particles, the algorithm's flexibility in global and local searches can be improved. This allows the algorithm to explore the solution space extensively in the early stages of the search, while focusing on a fine search for the optimal solution in the later stages, thus improving the convergence speed and global optimization capability of the search algorithm.

[0024] By clustering the multiple initial location values ​​obtained through clustering algorithms, outliers can be effectively identified and excluded from the location results. By selecting the cluster center of the cluster with the most elements as the final location result, the accuracy of the location can be improved. Attached Figure Description

[0025] In the attached diagram: Figure 1 The arrangement is a five-element cone sensor array; Figure 2 The positioning signal collected by the sensor; Figure 3 This is the result image after noise reduction; Figure 4 This is a schematic diagram of signal delay principle; Figure 5 This is a diagram showing the effect of signal delay estimation. Detailed Implementation

[0026] Example 1 The technical solution of this invention is as follows: A method for locating partial discharge in an insulator, the specific method includes the following steps: S1. A polygonal three-dimensional sensor array is used to collect the positioning signals generated by partial discharge of the insulator.

[0027] The polygonal stereo sensor array can be a cone array, sphere array, or cylinder array, and the number of sensors N in the polygonal stereo sensor array is N ≥ 5.

[0028] In this embodiment, a five-element cone sensor array is used to collect the positioning signal generated by the partial discharge of the insulator. The arrangement of the five-element cone sensor array is as follows: Figure 1 As shown.

[0029] The sensor antenna is a spherical antenna with a diameter of 8cm, which can effectively detect partial discharge positioning signals in the frequency band between 100MHz and 1000MHz. Compared with ultrasonic signals, the signals in this detection frequency band have good frequency response characteristics and propagation characteristics.

[0030] Each sensor is electrically connected to an oscilloscope via a feeder. The oscilloscope has a sampling rate of 10 GSa / s. Figure 2 Understand the positioning signals collected by the sensors.

[0031] By working collaboratively with multiple sensors, positioning accuracy can be effectively improved. When one sensor malfunctions or is interfered with, other sensors can still provide valid information, thereby reducing potential errors during the positioning process and enhancing overall positioning reliability.

[0032] S2. The acquired positioning signal is adaptively decomposed from high to low frequency to obtain intrinsic mode function components at different frequencies. The correlation coefficient of each intrinsic mode function component is calculated, and correlation coefficients greater than the correlation coefficient threshold are selected as high-quality coefficients. Wavelet packet decomposition is performed on the intrinsic mode function components corresponding to the high-quality coefficients. The wavelet packet coefficients obtained by wavelet packet decomposition are processed using an adjustable threshold function to reconstruct the denoised high-quality intrinsic mode function components. All denoised high-quality intrinsic mode function components are synthesized to obtain the denoised time-domain positioning signal.

[0033] The acquired positioning signals are adaptively decomposed from high to low frequency to obtain intrinsic mode function components at different frequencies. The correlation coefficient of each intrinsic mode function component is calculated. The correlation coefficient is calculated as follows: , in, This represents the correlation coefficient of the i-th intrinsic mode function component. L It is the length of the positioning signal. It is the average value of the positioning signal. It is the mean of the i-th intrinsic mode function component. Indicates the location signal is in The value at time, Indicates that the i-th intrinsic mode function component is in The value at time, Indicates the time.

[0034] The closer the correlation coefficient of an intrinsic mode function component is to 1, the stronger the correlation between that intrinsic mode function component and the original data.

[0035] In this embodiment, the correlation coefficient threshold is set to 0.5, that is, intrinsic mode function components with a correlation coefficient greater than 0.5 are considered to be high-quality intrinsic mode function components.

[0036] The selected high-quality intrinsic mode function components are subjected to wavelet packet decomposition. The wavelet basis function used for decomposition is the "db45" wavelet basis, and the number of decomposition levels is selected as 3. Apply an adjustable threshold function to the wavelet packet coefficients obtained from the decomposition. Quantization processing is performed, and the threshold function is adjustable. The expression is as follows: , in, As a regulating factor, Let B be the coefficient of the B-th wavelet packet at scale A obtained from wavelet packet decomposition. These are the new wavelet packet coefficients obtained after processing with a threshold function. For a fixed threshold, the mathematical expression is: , It is the variance of the noise. The length of the positioning signal.

[0037] In this embodiment The value is 5.

[0038] The new wavelet packet coefficients are subjected to inverse wavelet packet transform to reconstruct the denoised high-quality intrinsic mode function components. By reconstructing all the denoised high-quality intrinsic mode function components, the denoised time-domain positioning signal is obtained. Figure 3 This is a diagram showing the effect of noise reduction using the method of the present invention.

[0039] S3. Based on the denoised time-domain positioning signals, Fourier transforms are performed to obtain frequency-domain positioning signals. Weighted cross-correlation processing is then performed on any two frequency-domain positioning signals to obtain the signal delay between the two sensors corresponding to any two frequency-domain positioning signals in the polygonal stereo sensor array. Figure 4 Indication.

[0040] Let the two time-domain positioning signals after denoising be respectively , t represents the position of the positioning signal in the time domain, and m and n represent the sensor serial numbers, respectively.

[0041] The time-domain positioning signal is transformed into a frequency-domain positioning signal using Fourier transform. The specific steps are as follows: , , in, express The frequency domain positioning signal obtained after Fourier transform, express The frequency domain positioning signal obtained after Fourier transform, It is the imaginary unit, satisfying , ω is the angular frequency.

[0042] Therefore, two frequency domain positioning signals and The cross-power spectral density between them can be expressed as: , Indicates taking The conjugate of complex numbers, This represents the cross-power spectral density.

[0043] The obtained cross-power spectral density is then weighted, and the specific weighting function is as follows: , in, For weighted functions, Angular frequency, and Let m and n represent the self-power spectral density of the two frequency domain positioning signals, respectively, and m and n represent the sensor serial numbers, respectively. These are weighting coefficients, and .

[0044] In this embodiment Taking a value of 0.85, the obtained generalized cross-power spectral density is: .

[0045] Performing an inverse Fourier transform on the obtained generalized cross-power spectral density yields the generalized cross-correlation function in the time domain. for: , The time difference between the signals received by the two sensors, i.e., the signal delay, is determined by... Peak detection can be performed to obtain the time delay between the positioning signal and the two sensors.

[0046] In digital communication systems, cross-correlation-based time delay estimation methods can estimate the time delay between the reference signal and the received signal, and are divided into integer-level and fractional-level delay estimations. To improve the accuracy of the time delay estimation in this invention, a frequency-domain weighted approach is used to process the cross-power spectral density of the signal, combined with... Figure 5Understand the signal delay estimation results.

[0047] Similarly, the time delay between any two other sensors in receiving the positioning signal can be obtained.

[0048] S4. Randomly select four sensors from the polygonal 3D sensor array and combine them. Based on the signal delay and the distance from the positioning signal to the sensor, establish the objective function corresponding to each combination. Randomly initialize a group of particles, with each particle representing a potential solution to the objective function. Use a combination of particle swarm optimization and gravity search algorithm to update the velocity and position of the particles. Repeat the update steps and calculate the objective function value after each update. After a preset number of iterations, output the global optimal solution, which is the solution to the objective function. Each solution to the objective function corresponds to an initial positioning value.

[0049] For an N-element polygonal stereo sensor array, combining any quaternions (one of which is used as a reference) yields a total of There are several combinations, each corresponding to a target function. By searching and solving the obtained target functions, we can obtain... Initial values ​​for each location.

[0050] In this embodiment, for the quintuple cone array, it is assumed that the coordinates of the insulator partial discharge source are... The coordinates of the five positioning sensors are as follows: , , , , .

[0051] Choose any four sensors , , , Form combinations, and with The objective function established using a reference sensor for: , In the formula, The speed at which the positioning signal travels through the air is given by a value of [value]. , For the first The time difference between when each sensor receives the positioning signal and when the reference sensor receives the positioning signal, i.e., the delay. Indicates that the power supply is discharged to the first The distance between the sensors, when h takes the values ​​1, 2, 3, and 4, yields the result. , , and , , , .

[0052] Among them, the power supply is to the first The formula for calculating the distance between the sensors is: , In the formula, x, y, and z are the coordinate values ​​of the discharge source on the x, y, and z axes, respectively. h y h z h These are the coordinates of the h-th sensor on the x, y, and z axes, respectively.

[0053] Each particle represents a potential solution to the objective function. The objective function value of each particle is calculated and called its fitness value.

[0054] Update the individual optimal solution and the global optimal solution for each particle based on the objective function value.

[0055] When updating particle velocities using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: , in, and These represent the velocities of the i-th particle in the (u+1)-th and u-th iterations, respectively, where u is the current iteration number. Indicates inertia weight, and As a learning factor, and Let be a random variable between [0, 1]. This represents the best historical position of the i-th particle. It is the globally optimal position among all particles. As the gravitational factor, This represents the position of the i-th particle in the u-th iteration. Let be the acceleration of the i-th particle during the u-th iteration.

[0056] Gravity factor The calculation formula is as follows; , in, The minimum gravitational factor, The maximum gravitational factor, This is the preset number of iterations.

[0057] Compared to traditional numerical solutions, dynamically adjusting the particle search speed by using a gravity factor can find a balance between global and local searches. This allows the algorithm to explore the solution space extensively in the early stages and focus on a fine search for the optimal solution in the later stages, thus improving convergence speed and global optimization capabilities.

[0058] Inertia weight The specific calculation method is as follows; , in, Indicates the maximum inertia weight. This represents the minimum inertia weight. The preset number of iterations is given, and u is the current number of iterations.

[0059] The inertia weights can be dynamically adjusted according to the needs of the algorithm. In the early stages of iteration, larger inertia weights are beneficial for global search; in the later stages of iteration, smaller inertia weights are beneficial for local search.

[0060] This dynamic adjustment strategy allows the algorithm to flexibly adjust its search strategy according to the needs of different stages, thereby improving the overall performance of the algorithm.

[0061] When updating the position of a particle using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: , in, and These represent the positions of the i-th particle in the (u+1)-th and u-th iterations, respectively. Let represent the velocity of the i-th particle in the (u+1)-th iteration, where u is the current iteration number.

[0062] Repeat the particle velocity and position update steps described above, while updating the fitness value of each particle, until the preset number of iterations is reached. Then, output the global optimal solution, which yields a set of initial positioning values. Similarly, by solving the objective functions obtained under the other four combinations, the initial values ​​for positioning are obtained respectively. , , , .

[0063] The solution method achieves alternating iterations by using a search strategy that combines particle swarm optimization and gravity search algorithms.

[0064] This alternating iterative approach can fully utilize the advantages of both algorithms, avoid getting trapped in local optima, and improve the convergence of the algorithm.

[0065] S5. Use a clustering algorithm to cluster all initial positioning values. The cluster center of the cluster with the most elements is the partial discharge location of the insulator.

[0066] In this embodiment, the five sets of initial positioning values ​​obtained when N=5 are used. , , , , Clustering is performed using clustering algorithms.

[0067] The number of clusters in S5 is selected according to the elbow rule.

[0068] The specific operation of S5 is as follows: By calculating the sum of squared intra-cluster errors under different cluster numbers and plotting the relationship between the number of clusters and the sum of squared intra-cluster errors, the "elbow" position in the plot (i.e. the point where the rate of decrease of the sum of squared intra-cluster errors slows down) is found as the cluster number K.

[0069] In this embodiment, the number of clusters K is 2, that is, two data points are randomly selected as the initial cluster centers.

[0070] For each data point, calculate its distance to all cluster centers and assign it to the cluster containing the nearest cluster center.

[0071] For each cluster, a new cluster center is selected such that the total distance from the new cluster center to all other data points within the cluster is minimized. This step is called distance iteration.

[0072] Repeat the assignment and distance iteration steps until the cluster centers no longer change or the preset number of iterations is reached.

[0073] The cluster center of the cluster with the most elements is selected as the final location result.

[0074] By clustering the multiple initial location values ​​obtained through clustering algorithms, outliers can be effectively identified and excluded from the location results. By selecting the cluster center of the cluster with the most elements as the final location result, the accuracy of the location can be improved.

[0075] By designing an N-element polygonal 3D sensor array, signal focusing can be achieved in the spatial domain, thereby increasing the received signal strength and reducing the impact of background noise. The collaborative operation of multiple sensors effectively improves positioning accuracy. When one sensor malfunctions or is interfered with, other sensors can still provide valid information, reducing potential errors during positioning and enhancing overall positioning reliability.

[0076] Adaptive decomposition of the positioning signal allows for independent analysis of intrinsic mode function (IMF) components at different frequencies. Wavelet packet decomposition, performed by selecting high-quality IMF components with high correlation coefficients, enables more accurate extraction of key signal features and reduces redundant information. An adjustable threshold function allows for flexible adjustment based on actual signal characteristics during the denoising process, significantly improving denoising performance.

[0077] The weighting function considers signals from multiple channels, thus effectively utilizing the correlation between these channels and enhancing signal processing performance. By adjusting the weighting coefficients, the main components of the signal can be highlighted, particularly in cases of low signal-to-noise ratio, thereby improving the ability to estimate signal delay.

[0078] By dynamically adjusting the search speed of particles, the algorithm's flexibility in global and local searches can be improved. This allows the algorithm to explore the solution space extensively in the early stages of the search, while focusing on a fine search for the optimal solution in the later stages, thus improving the convergence speed and global optimization capability of the search algorithm.

[0079] By clustering the multiple initial location values ​​obtained through clustering algorithms, outliers can be effectively identified and excluded from the location results. By selecting the cluster center of the cluster with the most elements as the final location result, the accuracy of the location can be improved.

Claims

1. A method for locating partial discharge in an insulator, characterized in that, The specific method includes the following steps: S1. A polygonal three-dimensional sensor array is used to collect the positioning signal generated by the partial discharge of the insulator; S2. The acquired positioning signal is adaptively decomposed from high to low frequency to obtain intrinsic mode function components at different frequencies. The correlation coefficient of each intrinsic mode function component is calculated, and correlation coefficients greater than the correlation coefficient threshold are selected as high-quality coefficients. Wavelet packet decomposition is performed on the intrinsic mode function components corresponding to the high-quality coefficients. The wavelet packet coefficients obtained by wavelet packet decomposition are processed using an adjustable threshold function to reconstruct the denoised high-quality intrinsic mode function components. All denoised high-quality intrinsic mode function components are synthesized to obtain the denoised time-domain positioning signal. S3. Based on the denoised time-domain positioning signal, the frequency-domain positioning signal is obtained by Fourier transform. Based on any two frequency-domain positioning signals, weighted cross-correlation processing is performed to obtain the signal delay between the two sensors corresponding to any two frequency-domain positioning signals in the polygonal stereo sensor array. S4. Randomly select four sensors from the polygonal stereo sensor array and combine them. Based on the signal delay and the distance from the positioning signal to the sensor, establish the objective function corresponding to each combination. Randomly initialize a group of particles, each particle representing a potential solution of the objective function. Use a combination of particle swarm optimization and gravity search algorithm to update the velocity and position of the particles. Repeat the update steps and calculate the objective function value after each update. After a preset number of iterations, output the global optimal solution, which is the solution of the objective function. Each solution of the objective function corresponds to an initial positioning value. S5. Use a clustering algorithm to cluster all initial positioning values. The cluster center of the cluster with the most elements in the cluster is the partial discharge location of the insulator.

2. The method for locating partial discharge in an insulator according to claim 1, characterized in that, When updating particle velocities using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: ， in, and These represent the velocities of the i-th particle in the (u+1)-th and u-th iterations, respectively, where u is the current iteration number. Indicates inertia weight, and As a learning factor, and Let be a random variable between [0, 1]. This represents the best historical position of the i-th particle. It is the globally optimal position among all particles. As the gravitational factor, This represents the position of the i-th particle in the u-th iteration. Let be the acceleration of the i-th particle during the u-th iteration.

3. The method for locating partial discharge in an insulator according to claim 2, characterized in that, The gravity factor The calculation formula is as follows; ， in, The minimum gravitational factor, The maximum gravitational factor, This is the preset number of iterations.

4. The method for locating partial discharge in an insulator according to claim 2, characterized in that, The inertial weight The specific calculation method is as follows; ， in, Indicates the maximum inertia weight. This represents the minimum inertia weight. The preset number of iterations is given, and u is the current number of iterations.

5. The method for locating partial discharge in an insulator according to claim 1, characterized in that, When updating the position of a particle using a combination of particle swarm optimization and gravity search algorithms, the specific update formula is as follows: ， in, and These represent the positions of the i-th particle in the (u+1)-th and u-th iterations, respectively. Let represent the velocity of the i-th particle in the (u+1)-th iteration, where u is the current iteration number.

6. The method for locating partial discharge in an insulator according to claim 1, characterized in that, Choose any four sensors , , , Form combinations, and with The objective function established using a reference sensor for: ， In the formula, The speed at which the positioning signal travels through the air is given by a value of [value]. , For the first The time difference between when each sensor receives the positioning signal and when the reference sensor receives the positioning signal, i.e., the delay. Indicates that the power supply is discharged to the first The distance between the sensors, when h takes the values ​​1, 2, 3, and 4, yields the result. , , and , , , ; Among them, the power supply is to the first The formula for calculating the distance between the sensors is: ， In the formula, x, y, and z are the coordinate values ​​of the discharge source on the x, y, and z axes, respectively. h y h z h These are the coordinates of the h-th sensor on the x, y, and z axes, respectively.

7. The method for locating partial discharge in an insulator according to claim 1, characterized in that, The weighted cross-correlation processing based on any two frequency domain positioning signals described in S3 is specifically weighted as follows: , in, For weighted functions, Angular frequency, and Let m and n represent the self-power spectral density of the two frequency domain positioning signals, respectively, and m and n represent the sensor serial numbers, respectively. These are weighting coefficients, and .

8. The method for locating partial discharge in an insulator according to claim 1, characterized in that, The polygonal stereo sensor array is a cone array, sphere array, or cylinder array, and the number of sensors N in the polygonal stereo sensor array is N ≥ 5.

9. A method for locating partial discharge in an insulator according to claim 1, characterized in that, The method for calculating the correlation coefficient described in S2 is as follows: ， in, This represents the correlation coefficient of the i-th intrinsic mode function component. L It is the length of the positioning signal. It is the average value of the positioning signal. It is the mean of the i-th intrinsic mode function component. Indicates the location signal is in The value at time, Indicates that the i-th intrinsic mode function component is in The value at time, Indicates the time.

10. A method for locating partial discharge in an insulator according to claim 1, characterized in that, The number of clusters mentioned in S5 is selected according to the elbow rule.

Images