Partial Discharge Detection Method and Device Based on Tunable Optical Fiber

Abstract

This application relates to the field of fiber optic discharge detection technology, specifically to a method and apparatus for partial discharge detection based on tunable optical fiber. The method includes: determining the fluctuation difference value by calculating the difference between the diameter of each test location and adjacent locations, combined with the diameter fitting error within the neighborhood; dividing the area into marked intervals based on the dispersion of the fluctuation difference value within the neighborhood and the differences between adjacent locations; comparing the diameter distribution characteristics and dispersion within adjacent intervals, merging intervals to form marked segments; and fitting the diameter of the markers within each marked segment to establish an accurate relationship model between the diameter of the fiber optic interference loop and the location of the partial discharge. This application solves the problem of overfitting that traditional methods easily produce during the overall fitting process, improving the sensitivity and reliability of partial discharge detection.

Description

Technical Field

[0001] This application relates to the field of optical fiber discharge detection technology, specifically to a method and apparatus for partial discharge detection based on tunable optical fiber. Background Technology

[0002] Partial discharge detection in optical fibers works by utilizing the ultrasonic signals generated by partial discharges. These ultrasonic signals induce mechanical vibrations in the fiber, which in turn alter the Rayleigh scattering signal, thus changing the data obtained from fiber interference loop detection. This allows for the determination of the location and intensity of the partial discharge. Due to the high sensitivity and electromagnetic interference resistance of optical fiber technology, fiber interference loop detection is rapidly being applied to the detection of partial discharges in power transmission lines such as cables and cables.

[0003] Traditional fiber optic interferometer loop partial discharge detection methods typically employ a fixed diameter design and establish a relationship model between the partial discharge location and the interferometer loop diameter through nonlinear fitting. However, due to the significant nonlinear characteristics of the attenuation effect of ultrasonic waves propagating in air, traditional methods are prone to overfitting during the overall fitting process. This results in poor model generalization ability and an inability to accurately reflect the diameter variation at different distances. This deficiency directly leads to inaccurate adjustment of the interferometer loop diameter, thereby reducing the sensitivity and reliability of partial discharge detection. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method and apparatus for detecting partial discharge based on tunable optical fiber. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a partial discharge detection method based on tunable optical fiber, the method comprising the following steps:

[0006] Ultrasonic signals were acquired after simulated discharge at various test locations using fiber optic interference rings of different diameters.

[0007] At each test location, the center frequency of the ultrasonic signal under all diameters is compared with the preset frequency, and the marker diameter for each test location is selected from all diameters. Based on the difference between the marker diameter of each test location and the previous test location, the diameter characteristic value at each test location is determined. The diameter characteristic value of each test location is compared with the diameter characteristic value of the next test location, and combined with the fitting error when fitting the diameter characteristic value of all test locations in the neighborhood of each test location, the fluctuation difference value of each test location is determined.

[0008] Based on the dispersion of the fluctuation difference values ​​of all test locations within the neighborhood, and by comparing the fluctuation difference values ​​of each test location with its adjacent test locations, all test locations are divided into identification intervals. The distribution of the diameters of all test location markers between each identification interval and its next identification interval is compared, and the identification intervals are merged to obtain identification segments by combining the dispersion of the diameters of all test location markers within each identification interval.

[0009] By fitting the diameters of all markers within each marker segment, a model relating the diameter of the fiber interference loop to the location of partial discharge is established.

[0010] Preferably, the step of selecting the marker diameter for each test location from all diameters includes:

[0011] For each test location, the diameter with the center frequency of the ultrasonic signal that is greater than and closest to the preset frequency among all diameters of the light interference ring is taken as the measured diameter for each test location.

[0012] Based on the measured diameter, the values ​​are increased sequentially, and the measured diameter and the results of the increases are fitted with the corresponding center frequency. The fitted value at the position corresponding to the preset center frequency in the fitted curve is used as the mark diameter for each test position.

[0013] Preferably, the diameter characteristic value at each test position is the result of the difference between the marker diameter at each test position and the previous test position, multiplied by the corresponding marker diameter.

[0014] Preferably, the method for determining the fluctuation difference value at each test location is as follows:

[0015] Compare the diameter feature values ​​of each test location with the next test location to filter out homomorphic locations from all test locations in the neighborhood of each test location;

[0016] Calculate the difference in quantity between all test locations and all homomorphic locations in the neighborhood of each test location, and denote it as the quantity difference. The product of the fitting error when fitting the diameter feature value of all test locations in the neighborhood of each test location and the quantity difference is used as the fluctuation difference value of each test location.

[0017] Preferably, the homomorphic position is the test position whose diameter feature value is less than a preset threshold among all test positions in the neighborhood of each test position.

[0018] Preferably, the step of dividing all test locations into identified intervals includes:

[0019] Based on the dispersion of the fluctuation difference values ​​of all test locations in the neighborhood of each test location, and the difference in fluctuation difference values ​​between each test location and its neighboring test locations, the discrimination at each test location is determined.

[0020] The discrimination of all test locations is sorted in ascending order according to the distance from the test location to the fiber optic test ring. The peak and trough in the sorting result are extracted using the peak detection algorithm. All test locations between each peak and its adjacent previous trough are combined to form an identification interval.

[0021] Preferably, the expression for the discrimination at each test location is: In the formula, Indicates the discrimination at test position i; This indicates the degree of dispersion of the fluctuation difference values ​​among all test locations within the neighborhood of test location i; , These represent the difference in fluctuation value between test position i and its next test position, and the difference in fluctuation value between i and its previous test position, respectively. This indicates a constant that is preset to be greater than 0.

[0022] Preferably, the step of merging the identifier intervals to obtain the identifier segment includes:

[0023] Based on the difference in the average distribution of the marker diameters at all test locations between each marker interval and its successor, and the dispersion of the marker diameters at all test locations within each marker interval, the non-merging coefficient between each marker interval and its successor is determined.

[0024] If the non-merging coefficient between the m-th identifier interval and its next identifier interval is less than 0, then the m-th identifier interval and its next identifier interval are merged into one identifier segment; otherwise, no merging is performed, and all identifier intervals are traversed to obtain all identifier segments.

[0025] Preferably, the expression for the non-merging coefficient between each identifier interval and its subsequent identifier interval is: In the formula, This represents the uncombined coefficient between interval j and its subsequent identified interval; This represents the difference in the mean diameter of all test location markers between marker interval j and its subsequent marker interval; This indicates the dispersion of the marker diameter at all test locations within the identification interval j. Secondly, embodiments of this application also provide a partial discharge detection device based on tunable optical fiber, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the aforementioned partial discharge detection methods based on tunable optical fiber.

[0026] This application has at least the following beneficial effects:

[0027] This application, by combining marker diameter screening, diameter characteristic value calculation, and fluctuation difference value analysis, achieves a quantitative assessment of the nonlinearity of the trend of fiber optic interferometer ring diameter changing with partial discharge distance. This effectively reflects the complexity and propagation stability of ultrasonic wave attenuation, providing a reliable basis for optimizing the selection of fiber optic ring diameter in partial discharge detection, and improving the accuracy and adaptability of partial discharge detection. Furthermore, this application achieves precise segmentation of the test position by combining the dispersion of fluctuation difference values ​​with the discrimination calculation of adjacent differences, and intelligently merges the marked intervals using a non-merging coefficient, ultimately obtaining a marked segment that accurately reflects the nonlinear gradual change characteristics of ultrasonic wave attenuation. This effectively improves the adaptability and accuracy of fiber optic interferometer ring diameter selection and partial discharge detection. Furthermore, this application effectively overcomes the overall overfitting problem caused by the nonlinearity of ultrasonic wave attenuation by piecewise fitting the marker diameter within the marked segment, improving the accuracy and generalization ability of the fiber optic interferometer ring diameter-partial discharge position relationship model, significantly improving the accuracy of partial discharge detection, thus providing a more precise basis for diameter adjustment in partial discharge detection, and enhancing the sensitivity and reliability of partial discharge detection. Attached Figure Description

[0028] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A flowchart illustrating the steps of a partial discharge detection method based on tunable optical fiber provided in one embodiment of this application;

[0030] Figure 2 This is a flowchart of a homomorphic location filtering process provided in one embodiment of this application. Detailed Implementation

[0031] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the partial discharge detection method and apparatus based on tunable optical fiber proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0033] The following description, in conjunction with the accompanying drawings, details the specific scheme of the partial discharge detection method and device based on tunable optical fiber provided in this application.

[0034] Please see Figure 1 The diagram illustrates a flowchart of a partial discharge detection method based on tunable optical fiber according to an embodiment of this application. The method includes the following steps:

[0035] Step S1: Obtain ultrasonic signals from simulated discharges at various test locations using fiber optic interference rings of different diameters.

[0036] With the position of the fixed fiber optic interference ring unchanged, partial discharge is simulated at preset test locations at different distances. The diameter of the fiber optic interference ring is changed multiple times to measure the ultrasonic signal at each test location. The detailed process is as follows:

[0037] In this embodiment, the position of the fixed fiber optic interferometer ring remains unchanged. Partial discharge is simulated at a distance L from the fiber optic interferometer ring. The diameter of the fiber optic interferometer ring is changed multiple times to obtain ultrasonic signals at the test positions with different diameters. Furthermore, partial discharge is simulated at a distance 2L from the fiber optic interferometer ring, and ultrasonic signals at the test positions with different diameters of the fiber optic interferometer ring are obtained at this test position. The acquisition duration is set to t and the acquisition frequency is f. This process is repeated N times to obtain ultrasonic signals at N different test positions at distances from the fiber optic interferometer ring. The diameter of the fiber optic interferometer ring is adjusted within the range of [5, 30], with the diameter unit being cm. The ultrasonic signals at each test position are tested starting from an initial value of 5 cm, and the diameter is increased by 1 cm from the previous value. In this embodiment, N is 100. In actual applications, as other implementation methods, the implementer can set the values ​​according to specific circumstances, and the implementer can also set the test positions themselves. This embodiment does not impose any special restrictions.

[0038] It should be noted that the values ​​of the acquisition duration and acquisition frequency are set manually. In this embodiment, the acquisition duration t is 0.01s and the acquisition frequency f is 100kHz. In actual applications, as other implementation methods, implementers can also set them according to specific circumstances. This embodiment does not impose any special restrictions.

[0039] Furthermore, in order to remove noise interference to the ultrasonic signal, this embodiment uses a wavelet threshold denoising algorithm to denoise the ultrasonic signal. In practical applications, implementers may also use other denoising algorithms depending on the specific circumstances. This embodiment does not impose any special restrictions on the selection of denoising algorithms.

[0040] Among them, the wavelet threshold denoising algorithm is a well-known technology, and the specific process of using it to denoise ultrasonic signals will not be elaborated here.

[0041] Step S2: At each test location, compare the center frequency of the ultrasonic signal with the preset frequency for all diameters, and select the marker diameter for each test location from all diameters; based on the difference between the marker diameter of each test location and the previous test location, determine the diameter characteristic value at each test location; compare the diameter characteristic value of each test location with the diameter characteristic value of the next test location, and combine the fitting error when fitting the diameter characteristic value of all test locations in the neighborhood of each test location to determine the fluctuation difference value of each test location.

[0042] Since the ultrasonic signal at the partial discharge point is a high-frequency signal, with a frequency distribution between 6 and 40 kHz and a center frequency around 7 kHz, a Fourier transform algorithm is used to obtain the center frequency of the ultrasonic signal. Furthermore, at each test location, the difference between the center frequency of the ultrasonic signal for all diameters of the fiber optic interference loop and the preset frequency is compared to select the marker diameter for each test location from all diameters. Specifically:

[0043] In this embodiment, for each test location, the diameter with the ultrasonic signal center frequency that is greater than and closest to the preset frequency among all diameters of the light interference ring is taken as the measured diameter for each test location.

[0044] It should be noted that the preset frequency is set manually, and in this embodiment, the preset frequency is 6kHz.

[0045] Fiber optic interference loops are devices manufactured using the Sagnac effect. The Sagnac effect is directly proportional to the loop's area and rotation angle, and inversely proportional to the wavelength. The center frequency is also inversely proportional to the wavelength. Therefore, as the diameter of the fiber optic interference loop increases, higher frequencies can be detected. Consequently, the detected center frequency also increases with the increase in the fiber optic interference loop diameter. However, due to the varying degrees of attenuation caused by interference from other materials during the propagation of ultrasound in air or other substances, the relationship between the fiber optic interference loop diameter and the detected center frequency is not linear, but rather nonlinear.

[0046] Therefore, in this embodiment, the measured diameter at each test location is increased sequentially, and the measured diameter and the increase thereon are fitted with the corresponding center frequency. The center frequency is used as the independent variable in the fitting process, and the measured diameter and the increase thereon are used as the dependent variable in the fitting process. The fitted value at the position corresponding to the preset center frequency in the fitting curve is used as the mark diameter at each test location.

[0047] It should be noted that the preset center frequency is set manually, and in this embodiment, the preset center frequency is 7kHz.

[0048] It should be noted that there are many commonly used fitting algorithms. This embodiment uses the nonlinear least squares method to fit the diameter and center frequency. In practical applications, as other implementation methods, implementers may also use other fitting algorithms such as the polynomial function fitting method according to the specific situation. This embodiment does not impose any special restrictions on the selection of fitting algorithms.

[0049] The process of obtaining the signal center frequency using the Fourier transform algorithm and the process of fitting using the nonlinear least squares method are well-known techniques and will not be elaborated further.

[0050] It should be noted that, unless otherwise specified, all fitting operations in this embodiment employ the nonlinear least squares method.

[0051] Furthermore, ultrasonic waves attenuate with distance during propagation. While the attenuation exhibits a linear relationship at shorter distances, it intensifies with increasing distance, resulting in a non-linear attenuation pattern. Therefore, the degree of ultrasonic wave attenuation varies at different distances. This embodiment determines the diameter characteristic value at each test location based on the difference in the marker diameter between each test location and the previous test location, indirectly reflecting the changing trend of the ultrasonic signal data at the same test location. Specifically:

[0052] In this embodiment, the difference between the marker diameters of each test position and the previous test position is compared with the marker diameters of each test position to obtain the diameter characteristic value at each test position. This value is used to characterize the rate at which the diameter of the fiber interference loop changes with the partial discharge distance. If the diameter characteristic value at the current test position is larger, it indicates that the diameter increases sharply with the distance, indicating that the ultrasonic wave attenuation is very severe. This may be due to the propagation medium, obstacles, or distance response causing the signal to weaken rapidly. Conversely, if the diameter characteristic value at the current test position is smaller, it indicates that the diameter increases slowly with the distance, indicating that the ultrasonic wave attenuation is relatively gentle and the signal propagation is relatively stable. This may be due to the uniform propagation medium, fewer obstacles, or a better linear distance response, resulting in slower signal attenuation.

[0053] Furthermore, this embodiment uses the difference in diameter feature values ​​between each test location and its subsequent test location to filter out homomorphic locations from all test locations within the neighborhood of each test location. The flowchart of the homomorphic location filtering process provided in this embodiment is as follows: Figure 2 As shown, the specific process of homomorphic location filtering is as follows:

[0054] A neighborhood W is divided with each test position as the center. The radius of the neighborhood W is 2. That is, the test positions corresponding to the two tests before each test position and the test positions corresponding to the two tests after each test position, together with each test position, constitute the neighborhood of each test position. The radius of the neighborhood W is set by human intervention. In actual application, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0055] Furthermore, in this embodiment, the test positions whose diameter feature value is less than a preset threshold among all test positions in the neighborhood of each test position are taken as the homomorphic positions of each test position. This is used to characterize the consistency of the diameter change trend of the fiber interference ring at adjacent test positions. The number of homomorphic positions indicates that the more consistent the diameter change trend of the fiber interference ring at adjacent test positions, the better.

[0056] It should be noted that the preset threshold value is set manually. In this embodiment, the preset threshold value is 0.05. In actual applications, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0057] Furthermore, this embodiment determines the fluctuation difference value of each test position based on the fitting error when fitting the diameter feature values ​​of all test positions in the neighborhood of each test position, and the difference in quantity between all test positions in the neighborhood and all homomorphic positions, in order to measure the nonlinearity of the trend of the fiber interference loop diameter change. Specifically:

[0058] As a specific implementation method, in this embodiment, the difference in quantity between all test locations and all homomorphic locations in the neighborhood of each test location is calculated and denoted as the quantity difference. The product of the fitting error when fitting the diameter feature value of all test locations in the neighborhood of each test location and the quantity difference is used as the fluctuation difference value of each test location.

[0059] It should be noted that there are many methods to measure the differences between data. In this embodiment, the absolute value of the difference in quantity between all test locations and all homomorphic locations in the neighborhood of test location i is taken as the difference in quantity between all test locations and all homomorphic locations in the neighborhood of test location i. In practical applications, as other implementation methods, implementers may also use other methods to measure the differences between data, such as the square or ratio of the difference, depending on the specific circumstances. This embodiment does not impose any special restrictions on the selection of methods to measure the differences between data.

[0060] It should be noted that, unless otherwise specified, in this embodiment, all content involving the measurement of differences between data uses the method of taking the absolute value of the difference.

[0061] The method for calculating the fitting error is a well-known technique, and the specific calculation process will not be elaborated here.

[0062] Based on the fluctuation difference values ​​at each test location, it can be understood that the fluctuation difference value combines the local fitting error and the number of homomorphic positions, and is used to measure the nonlinearity of the diameter change trend of the fiber interference loop. If the fitting error when fitting the diameter characteristic value of all test locations in the neighborhood of the current test location is larger, it indicates that the difference between the diameter characteristic value and the fitted value is larger, indicating that the nonlinearity of the diameter change rate of the fiber interference loop is stronger and the diameter change is irregular. This means that the ultrasonic attenuation is more complex, and therefore, the corresponding fluctuation difference value is larger. At the same time, if the difference in the number of all test locations and all homomorphic positions in the neighborhood of the current test location is larger, that is, the greater the difference in the number of homomorphic positions, the greater the difference in the diameter trend change of the fiber interference loop, indicating that the ultrasonic attenuation nonlinearity is more significant, and the corresponding fluctuation difference value is larger.

[0063] Conversely, if the fitting error in the neighborhood of the current test position is smaller, it indicates that the difference between the diameter characteristic value and the fitted value is smaller, suggesting that the nonlinearity of the diameter change rate of the fiber interference ring is weaker and the diameter change is more regular. This means that the ultrasonic attenuation is relatively stable, and therefore, the corresponding fluctuation difference value is smaller. At the same time, if the difference in the number of all test positions and all homomorphic positions in the neighborhood of the current test position is smaller, that is, the smaller the difference in the number, it indicates that there are more homomorphic positions and the more consistent the diameter trend change of the fiber interference ring is. This indicates that the nonlinearity of ultrasonic attenuation is less significant, and the fluctuation difference value is correspondingly smaller.

[0064] Thus, this embodiment, by combining marker diameter screening, diameter characteristic value calculation, and fluctuation difference value analysis, achieves a quantitative assessment of the nonlinearity of the trend of fiber interference ring diameter changing with partial discharge distance. It effectively reflects the complexity and propagation stability of ultrasonic wave attenuation, provides a reliable basis for optimizing the selection of fiber ring diameter in partial discharge detection, and improves the accuracy and adaptability of partial discharge detection.

[0065] Step S3: Based on the dispersion of the fluctuation difference values ​​of all test locations within the neighborhood, and comparing the fluctuation difference values ​​of each test location with its adjacent test locations, all test locations are divided into identification intervals; the distribution of the diameters of all test location markers between each identification interval and its next identification interval is compared, and the identification intervals are merged to obtain identification segments based on the dispersion of the diameters of all test location markers within each identification interval.

[0066] Because subtle differences still exist in the data changes during the same phase of change, but the attenuation trend of ultrasound remains constant with increasing distance, the nonlinear fluctuations of the data in the same phase exhibit a certain gradual characteristic. That is, the data changes with the same trend, but the amount of change differs in different phases, i.e., the gradient of the gradual change is different. As the distance increases, the diameter of the fiber optic interference loop becomes larger. This is because a larger diameter is needed to detect the center frequency of partial discharge faults and confirm the occurrence of partial discharge faults in the line.

[0067] Based on the above analysis, this embodiment divides all test locations into labeled intervals by comparing the dispersion of fluctuation difference values ​​of all test locations within the neighborhood and comparing the fluctuation difference values ​​of each test location with its adjacent test locations; it then compares the distribution of the diameters of all test location markers between each labeled interval and its next labeled interval, and combines this with the dispersion of the diameters of all test location markers within each labeled interval to merge the labeled intervals into labeled segments, specifically:

[0068] In this embodiment, firstly, based on the dispersion of the fluctuation difference values ​​of all test locations within the neighborhood of each test location, and the difference in fluctuation difference values ​​between each test location and its adjacent test locations, the discrimination at each test location is determined. This is used to characterize whether the test location can serve as a dividing point for the ultrasonic attenuation trend, i.e., whether the diameter change rate undergoes a significant inflection point. Specifically:

[0069] As one implementation method, in this embodiment, the discrimination at test location i is... The expression is: In the formula, This indicates the degree of dispersion of the fluctuation difference values ​​among all test locations within the neighborhood of test location i; , These represent the difference in fluctuation value between test position i and its next test position, and the difference in fluctuation value between i and its previous test position, respectively. This indicates a preset constant greater than 0, used to prevent the denominator from being 0. The value is set manually, in this embodiment. The value of is 0.01. Provided that the denominator is not zero and does not excessively affect the calculation result, the implementer may also set it according to the specific situation. This embodiment does not impose any special restrictions.

[0070] It should be noted that there are many methods to measure the dispersion of a set of data. In this embodiment, the standard deviation of the fluctuation difference values ​​of all test locations in the neighborhood of test location i is used as the dispersion of the fluctuation difference values ​​of all test locations in the neighborhood of test location i. In practical applications, as other implementation methods, implementers may also use other methods such as variance or coefficient of variation to measure the dispersion of data, depending on the specific circumstances. This embodiment does not impose any special restrictions on the selection of methods for measuring the dispersion of data.

[0071] It should be noted that, unless otherwise specified, in this embodiment, all content involving the measurement of data dispersion uses the standard deviation calculation method.

[0072] Based on the discriminability at each test location, it can be understood that the discriminability characterizes whether a test location can serve as a dividing point for the ultrasonic attenuation trend, i.e., whether the diameter change rate undergoes a significant inflection point. If the dispersion of the fluctuation difference values ​​among all test locations in the neighborhood of test location i is greater, it indicates a more significant difference in the diameter change trend between test location i and its preceding and following test locations, with inconsistent fluctuations, making it suitable as a segmentation node, and correspondingly, a greater discriminability. Simultaneously, if the difference in fluctuation difference value between test location i and its subsequent test location is greater than the difference in fluctuation difference value between test location i and its preceding test location, i.e. The larger the value, the greater the difference between the diameter feature value of the current test position and the next test position is than the difference with the previous test position. This indicates that the diameter change trend before and after the current test position has changed significantly, making it suitable as a segmentation node with a correspondingly high discrimination.

[0073] Conversely, if the dispersion of the fluctuation difference values ​​among all test positions in the neighborhood of test position i is smaller, it indicates that the diameter change trend between test position i and its preceding and following test positions is relatively consistent, the fluctuation is stable, and it is not suitable as a segmentation node, resulting in lower discrimination. At the same time, if the ratio of the fluctuation difference value between test position i and its following test position to the fluctuation difference value between test position i and its preceding test position is smaller, that is... The smaller the value, the closer the difference between the diameter feature value of the current test position and the next test position is to or less than the difference between the current test position and the previous test position. This indicates that there is no obvious change in the diameter change trend before and after the current test position, and it is not suitable as a segmentation node. The corresponding discrimination is relatively small.

[0074] Furthermore, based on the aforementioned discrimination index, this embodiment divides all test locations into identified intervals to characterize several continuous intervals obtained after dividing all test locations according to the node discrimination index. Each interval represents a relatively consistent diameter variation trend, while the trends between different identified intervals show significant differences. The specific process for determining the identified intervals is as follows:

[0075] In this embodiment, the discrimination of all test locations is arranged in ascending order according to the distance from the test location to the fiber optic test ring. The peaks and valleys in the arrangement result are extracted by an automatic multi-scale peak detection algorithm, and all test locations between each peak and its adjacent previous valley form an identification interval.

[0076] There are many commonly used peak detection algorithms. In this embodiment, the automatic multi-scale peak detection algorithm is used. In practical applications, as other implementation methods, implementers may also use other peak detection algorithms according to specific circumstances. This embodiment does not impose any special restrictions.

[0077] Among them, the automatic multi-scale peak detection algorithm is a well-known technology, and the specific process of using it to detect peaks and valleys will not be described in detail.

[0078] Furthermore, this embodiment determines the non-merging coefficient between each identification interval and its successor based on the difference in the average distribution of the marker diameters at all test locations between each identification interval and its successor, as well as the dispersion of the marker diameters at all test locations within each identification interval. This coefficient is then used to merge the identification intervals into identification segments. Specifically:

[0079] In this embodiment, the non-merging coefficient between the identifier interval j and its subsequent identifier interval The expression is: In the formula, This represents the difference in the mean diameter of all test location markers between marker interval j and its subsequent marker interval; This indicates the dispersion of the marker diameter at all test locations within the identification interval j.

[0080] Based on the non-merging coefficient between each marker interval and its successor, we can understand that the non-merging coefficient reflects the similarity of the diameter trends between two adjacent marker intervals. If the difference between the mean diameter of all test positions of marker interval j and its successor is greater, it indicates that the diameter trends of marker interval j and its successor are significantly different and not suitable for merging. Therefore, the corresponding non-merging coefficient is relatively large. At the same time, if the dispersion of the diameter of all test positions within marker interval j is smaller, it indicates that the data fluctuation within the marker interval is small and the trend is stable. It may be affected by noise and tends to remain independent and not merge with adjacent intervals. The corresponding non-merging coefficient is relatively large.

[0081] Conversely, if the difference between the mean diameter of all test positions in the identification interval j and the next identification interval is smaller, it indicates that the diameter trends of the identification interval j and the next identification interval are similar, making them suitable for merging. Therefore, the corresponding non-merging coefficient is relatively small. At the same time, if the dispersion of the diameter of all test positions in the identification interval j is larger, it indicates that the data within the identification interval fluctuates greatly and the trend is unstable. It may be significantly affected by noise and tends to merge with adjacent intervals to reduce noise interference. The corresponding non-merging coefficient is relatively small.

[0082] Furthermore, in this embodiment, based on the non-merging coefficient, the identified intervals are merged to obtain identified segments, which are used to determine the relationship between the diameter and the location of partial discharge. Specifically:

[0083] In this embodiment, if the non-merging coefficient between the m-th identifier interval and the next identifier interval is less than 0, then the m-th identifier interval and the next identifier interval are merged into one identifier segment; otherwise, no merging is performed, and all identifier intervals are traversed to obtain all identifier segments.

[0084] The final diameter segmentation is obtained after evaluation and merging of non-merging coefficients in the identification segment. The diameter change trend within each segment is highly consistent. By segmented fitting, the overfitting problem caused by overall fitting is avoided, the generalization ability of the model is improved, and the diameter of the fiber ring can be adjusted more accurately in partial discharge detection.

[0085] Thus, this embodiment achieves precise segmentation of the test location by combining the dispersion of fluctuation difference values ​​with the discrimination of adjacent differences, and intelligently merges the marked intervals using a non-merging coefficient, ultimately obtaining marked segments that can accurately reflect the nonlinear gradual change characteristics of ultrasonic attenuation, effectively improving the adaptability and accuracy of fiber optic interferometer ring diameter selection and partial discharge detection.

[0086] Step S4: Fit the diameter of all markers in each marker segment to establish a model relating the diameter of the fiber interference loop to the location of partial discharge.

[0087] Based on step S3, the identification segment is divided. By using piecewise fitting, overfitting caused by overall fitting is avoided, improving the model's generalization ability. This allows for more precise adjustment of the fiber optic ring diameter in partial discharge detection. The specific fitting process is as follows:

[0088] The diameter of all test locations within each marker segment is used as the input to the polynomial function fitting method. Since the attenuation of ultrasound follows an exponential attenuation model, a larger interference ring diameter is required during detection. Therefore, in this embodiment, the independent variable of the polynomial fitting is an exponential function value with the natural constant as the base and the interference ring diameter as the exponent, and the distance between the fiber interference ring and the test location is the dependent variable. The diameter-partial discharge location fitting curve of each marker segment is output to determine the relationship between the diameter of the fiber interference ring and the partial discharge location, thereby determining the location of the partial discharge based on the fitting curve model.

[0089] Among them, the polynomial function fitting algorithm is a well-known technique, and the specific process of using it to fit data will not be described in detail.

[0090] Thus, this embodiment effectively overcomes the overall overfitting problem caused by the nonlinearity of ultrasonic attenuation by segmenting the diameter of the marker within the marked segment. This improves the accuracy and generalization ability of the model relating the diameter of the fiber interference loop to the location of the partial discharge, significantly enhancing the accuracy of partial discharge detection. Consequently, it provides a more precise basis for diameter adjustment in partial discharge detection, thereby improving the sensitivity and reliability of partial discharge detection.

[0091] Based on the same inventive concept as the above methods, this application also provides a partial discharge detection device based on tunable optical fiber, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described partial discharge detection methods based on tunable optical fiber.

[0092] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments of this specification have been described above. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0093] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0094] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A partial discharge detection method based on tunable optical fiber, characterized in that, The method includes the following steps: Ultrasonic signals were acquired after simulated discharge at various test locations using fiber optic interference rings of different diameters. At each test location, the center frequency of the ultrasonic signal under all diameters is compared with the preset frequency, and the marker diameter for each test location is selected from all diameters. Based on the difference between the marker diameter of each test location and the previous test location, the diameter characteristic value at each test location is determined. The diameter characteristic value of each test location is compared with the diameter characteristic value of the next test location, and combined with the fitting error when fitting the diameter characteristic value of all test locations in the neighborhood of each test location, the fluctuation difference value of each test location is determined. Based on the dispersion of the fluctuation difference values ​​of all test locations within the neighborhood, and by comparing the fluctuation difference values ​​of each test location with its adjacent test locations, all test locations are divided into identification intervals. The distribution of the diameters of all test location markers between each identification interval and its next identification interval is compared, and the identification intervals are merged to obtain identification segments by combining the dispersion of the diameters of all test location markers within each identification interval. By fitting the diameters of all markers within each marker segment, a model relating the diameter of the fiber interference loop to the location of partial discharge is established.

2. The partial discharge detection method based on tunable optical fiber as described in claim 1, characterized in that, The process of selecting the marker diameter for each test location from all diameters includes: For each test location, the diameter with the center frequency of the ultrasonic signal that is greater than and closest to the preset frequency among all diameters of the light interference ring is taken as the measured diameter for each test location. Based on the measured diameter, the values ​​are increased sequentially, and the measured diameter and the results of the increases are fitted with the corresponding center frequency. The fitted value at the position corresponding to the preset center frequency in the fitted curve is used as the mark diameter for each test position.

3. The partial discharge detection method based on tunable optical fiber as described in claim 1, characterized in that, The diameter characteristic value at each test position is the ratio of the difference between the marker diameter at each test position and the previous test position to the corresponding marker diameter.

4. The partial discharge detection method based on tunable optical fiber as described in claim 1, characterized in that, The method for determining the fluctuation difference value at each test location is as follows: Compare the diameter feature values ​​of each test location with the next test location to filter out homomorphic locations from all test locations in the neighborhood of each test location; Calculate the difference in quantity between all test locations and all homomorphic locations in the neighborhood of each test location, and denote it as the quantity difference. The product of the fitting error when fitting the diameter feature value of all test locations in the neighborhood of each test location and the quantity difference is used as the fluctuation difference value of each test location.

5. The partial discharge detection method based on tunable optical fiber as described in claim 4, characterized in that, The homomorphic position is the test position whose diameter feature value is less than a preset threshold among all test positions in the neighborhood of each test position.

6. The partial discharge detection method based on tunable optical fiber as described in claim 1, characterized in that, The process of dividing all test locations into identified intervals includes: Based on the dispersion of the fluctuation difference values ​​of all test locations in the neighborhood of each test location, and the difference in fluctuation difference values ​​between each test location and its neighboring test locations, the discrimination at each test location is determined. The discrimination of all test locations is sorted in ascending order according to the distance from the test location to the fiber optic test ring. The peak and trough in the sorting result are extracted using the peak detection algorithm. All test locations between each peak and its adjacent previous trough are combined to form an identification interval.

7. The partial discharge detection method based on tunable optical fiber as described in claim 6, characterized in that, The expression for the discrimination at each test location is: In the formula, Indicates the discrimination at test position i; This indicates the degree of dispersion of the fluctuation difference values ​​among all test locations within the neighborhood of test location i; , These represent the difference in fluctuation value between test position i and its next test position, and the difference in fluctuation value between i and its previous test position, respectively. This indicates a constant that is pre-defined as being greater than 0.

8. The partial discharge detection method based on tunable optical fiber as described in claim 1, characterized in that, The process of merging the identifier intervals to obtain the identifier segment includes: Based on the difference in the average distribution of the marker diameters at all test locations between each marker interval and its successor, and the dispersion of the marker diameters at all test locations within each marker interval, the non-merging coefficient between each marker interval and its successor is determined. If the non-merging coefficient between the m-th identifier interval and its next identifier interval is less than 0, then the m-th identifier interval and its next identifier interval are merged into one identifier segment; otherwise, no merging is performed, and all identifier intervals are traversed to obtain all identifier segments.

9. The partial discharge detection method based on tunable optical fiber as described in claim 8, characterized in that, The expression for the non-merging coefficient between each identifier interval and its subsequent identifier interval is as follows: In the formula, This represents the uncombined coefficient between interval j and its subsequent identified interval; This represents the difference in the mean diameter of all test location markers between marker interval j and its subsequent marker interval; This indicates the dispersion of the marker diameter at all test locations within the identification interval j.

10. A partial discharge detection device based on tunable optical fiber, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the partial discharge detection method based on tunable optical fiber as described in any one of claims 1-9.

Images