A nondestructive testing device and method for insulation performance of locomotive converter

Abstract

The invention discloses a nondestructive testing device and method for insulation performance of a locomotive converter, relates to the technical field of electrical equipment performance testing, and aims to solve the technical problem that the signal obtained by the current nondestructive testing of insulation performance of a locomotive converter contains a large number of high-order harmonics, and the signal decays quickly, and the insulation state of the converter cannot be monitored in real time. The device and method include a testing device, an image extraction unit, an image conversion unit, an image processing unit, an image storage unit, a data analysis unit, a data analysis processing unit, and a testing result output unit. The invention adopts non-contact testing, and does not require power off, corona, or pressure during the testing process. The nondestructive testing of insulation performance can be performed during the normal operation of the locomotive converter, and does not affect the normal operation of the locomotive. It is convenient to timely discover possible problems of the insulation system, avoid secondary failures caused by insulation failures, and ensure the safety and reliability of locomotive operation.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical equipment performance detection, and more specifically, to a nondestructive detection device and method for insulation performance of a locomotive converter. Background Art

[0002] At present, for equipment such as locomotive converters, the insulation system is the most basic system in the complex system of locomotive converters, and it is also a system that is susceptible to various electrical, thermal, mechanical stresses and electromagnetic effects, resulting in failures and structural damage. As locomotive converters develop towards high power density, high integration and miniaturization, the structure and design of the insulation system have become more complex. During the normal operation of the locomotive converter, alternating voltage and current, mechanical vibration and thermal stress will cause partial discharge in the insulation system. Excessive partial discharge is not only a source of electromagnetic interference, but also causes deformation, cracks, breakdown and other phenomena in the converter insulation system, resulting in more serious failures, and then more complex secondary failures.

[0003] Therefore, in order to ensure the normal operation of the locomotive converter insulation system during long-term service, it is necessary to monitor the status of the locomotive converter insulation system in real time.

[0004] At present, the most commonly used insulation performance test method is to test the insulation performance such as partial discharge, dielectric loss tangent, absorption ratio, polarization index, etc. when the locomotive converter is working. However, these test methods are in a power-off state and require an external power supply. In addition to replacing the equipment to ensure the test accuracy, a lot of preparation work is required when measuring, and the timing of the test is required to be high.

[0005] Among them, partial discharge is the discharge process caused by the breakdown of local areas in the insulating medium, which can be regarded as a collection of many partial discharge sources. These partial discharges are usually pulsed and are usually received by ultra-high frequency sensors. The partial discharge detection method can be used to monitor both the frequency domain characteristics and the time domain characteristics of partial discharge, and the data obtained is comprehensive.

[0006] However, most of the local discharge detection methods in the existing technology use UHF sensors to detect the overall discharge signal and then analyze the data. However, from the measurement principle of the UHF sensor, it detects the envelope of the entire discharge signal. Therefore, the detected signal contains a large number of high-order harmonics, and in the initial stage of discharge, the signal monitored by the UHF sensor decays relatively quickly, and it is impossible to monitor the insulation status of the converter in real time. In view of this, we propose a non-destructive detection device and method for the insulation performance of a locomotive converter. Summary of the invention

[0007] The purpose of the present invention is to provide a locomotive converter insulation performance nondestructive testing device and method to solve the technical problem that the signal obtained by the current locomotive converter insulation performance nondestructive testing contains a large number of high-order harmonics and the signal decays quickly, making it impossible to monitor the converter insulation status in real time.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a non-destructive testing device for insulation performance of a locomotive converter, comprising:

[0009] A detection device, used for applying a partial discharge signal of a preset intensity to a locomotive converter and receiving a partial discharge signal fed back by the locomotive converter, the detection device comprising an electrical signal generating device and a partial discharge detection device, the electrical signal generating device is used for generating a partial discharge signal, and the partial discharge detection device is used for receiving a feedback signal;

[0010] An image extraction unit, connected to the detection device, for performing Fourier transform on the partial discharge signal to convert the partial discharge signal into a partial discharge image;

[0011] An image conversion unit, connected to the image extraction unit, for converting the partial discharge image into a partial discharge grayscale image, and further converting the partial discharge grayscale image into a partial discharge halftone image;

[0012] An image processing unit, the image processing unit is connected to the image conversion unit and is used to perform grayscale segmentation and image filtering on the partial discharge grayscale image;

[0013] An image storage unit, connected to the image processing unit, for storing the partial discharge grayscale image and the partial discharge halftone image respectively;

[0014] A data analysis unit, the data analysis unit is connected to the image processing unit and is used to pre-process the partial discharge halftone image to determine the insulation condition of the locomotive converter;

[0015] A data analysis and processing unit, the data analysis and processing unit is connected to the data analysis unit and is used to perform correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results;

[0016] The detection result output unit is connected to the data analysis and processing unit, and is used to display, store and analyze the analysis results, and output the final detection results of the non-destructive detection of the insulation performance of the locomotive converter in the form of images.

[0017] The present invention adopts a non-contact detection method through the arrangement of an electric signal generating device and a partial discharge detection device. During the detection process, there is no need to cut off power, no corona, and no pressure. During the normal operation of the locomotive inverter, the insulation performance can be efficiently and accurately non-destructively tested. Without affecting the normal operation of the locomotive, possible problems of the insulation system can be discovered in time, secondary faults caused by insulation faults can be avoided, and the safety and reliability of the locomotive operation can be guaranteed.

[0018] Preferably, the conversion of the partial discharge halftone image is performed in stages by the following formula:

[0019] ;

[0020] In the formula, is the pixel point of the partial discharge grayscale image, is the pixel of the partial discharge halftone image, are the number of columns and rows of pixels in the partial discharge grayscale image, respectively.

[0021] Preferably, the image processing unit performs grayscale segmentation of the partial discharge grayscale image including:

[0022] The first binary segmentation of the partial discharge grayscale image is performed by Locating the pixel points in the partial discharge grayscale image whose discharge times exceed a preset number;

[0023] In the formula, The grayscale image of partial discharge is The gray value at is the preset threshold, is the image after binarization;

[0024] The partial discharge grayscale image obtained by the first binarization segmentation is subjected to a second binarization segmentation to locate the rising edge and falling edge of each discharge pulse and obtain the rising edge coordinate data of each pulse. , And falling edge coordinate data , , and draw the rising edge or falling edge curve of each pulse. The rising edge and falling edge can be determined by the slope change;

[0025] The partial discharge grayscale image obtained by the second binary segmentation is segmented for the third time to determine the half-peak time of the partial discharge pulse The half-peak time data is obtained by calculating the time point when the grayscale value on both sides of the pulse peak drops to half.

[0026] Preferably, the coordinate data of the rising edge and falling edge of the partial discharge pulse detected by the second binary segmentation are recorded as: point A Point B , connect point A and point B to get a ray, and record the equation of the ray as ;

[0027] In the formula, , .

[0028] Preferably, the grayscale level of the partial discharge grayscale image after the second binarization segmentation is , the grayscale values ​​of all pixels between point A and point B for each discharge pulse are counted, and the statistical results are normalized to normalize the pixels to The value of the interval, the gray level corresponding to the pixel is recorded as , the formula is ;

[0029] In the formula, is the gray level corresponding to the pixel, is the value of any pixel of the partial discharge grayscale image The gray value of .

[0030] Preferably, the image processing unit performs filtering processing on the partial discharge grayscale image, including:

[0031] Performing median filtering and mean filtering on the partial discharge grayscale image obtained by the second binarization segmentation to obtain a filtered image;

[0032] The median filter formula is: , where is the original image pixel value, is the pixel value of the filtered image, is the filter window;

[0033] The mean filter formula is: , where is the filter window size, , ;

[0034] Calculate the local grayscale mean of the filtered image , where The filtered image is The gray value at is the image size;

[0035] According to the formula Calculate the local grayscale mean square value , enhance the image;

[0036] Calculate the difference between two adjacent discharge pulse half-peak time data , get the rising edge speed of the discharge pulse , where is the height difference between the rising edges of adjacent pulses.

[0037] Preferably, the data analysis and processing unit performs correlation analysis on the partial discharge grayscale image and the partial discharge halftone image, and obtaining the analysis result includes:

[0038] Set a pixel point of the partial discharge halftone image as point D, find the grayscale level corresponding to the horizontal coordinates of the two discharge pulses adjacent to point D, and record them as the first grayscale level and the second gray level ,like , then point D is located on the first discharge pulse. , then point D is located on the second discharge pulse. , then according to the rising edge speed of the discharge pulses on both sides of the first discharge pulse point D, the discharge pulse corresponding to the horizontal coordinate of point D is determined as the pulse to be analyzed;

[0039] If point D is located on the first discharge pulse, find the gray level corresponding to the horizontal coordinate of the discharge pulse closest to point D, and record it as the third gray level , if the vertical straight line extending from point D intersects with the third pulse curve, and the gray level corresponding to the intersection is , then point D is located on the first discharge pulse, otherwise, point D is located on the second discharge pulse.

[0040] Preferably, the data analysis unit preprocesses the partial discharge halftone image including:

[0041] The partial discharge halftone image is processed by the OTSU threshold segmentation method to obtain a binary image. The OTSU algorithm calculates the variance between classes. To determine the threshold , so that Maximum, where , is the proportion of two types of pixels, , is the mean of two types of pixels;

[0042] The obtained binary image is subjected to median filtering to retain the pixel information of the image, and the image after median filtering is subjected to morphological dilation operation to expand the image and obtain the area of ​​the insulation failure region. , the expansion operation formula is , where is the original image, is a structural element;

[0043] The partial discharge halftone image obtained by the first binarization segmentation is inverted to obtain an inverted image. The inverted color formula is: ;

[0044] In the formula, is the original image pixel value, is the pixel value of the image after inversion;

[0045] The pixel positions of the areas where the discharge pulse points D are located in the inverted image are located, and the located pixel positions are marked and deleted respectively to obtain the remaining inverted image, and the remaining inverted image is subjected to median filtering to obtain a filtered inverted image, and the filtered inverted image is subjected to morphological filling operation to obtain a filled inverted image. The filling operation can be achieved by scanning the image and filling the holes.

[0046] Preferably, the data analysis unit preprocesses the partial discharge halftone image further comprising:

[0047] Perform morphological corrosion operation on the filled inverted image to obtain the inverted image after corrosion. The corrosion formula is: , perform inversion processing on the inverted image after corrosion to obtain an inverted image;

[0048] Perform morphological closing operation on the image after inversion processing to obtain an insulation breakdown spot image. The closing operation is to dilate first and then erode. Perform median filtering on the insulation breakdown spot image to obtain a filtered insulation breakdown spot image.

[0049] Perform morphological filling operation on the filtered insulation breakdown spot image to obtain the largest connected area of ​​pixels, which is recorded as , The number of pixels is recorded as the number of pixels ,but The number of pixels in the area is the number of pixels in the breakdown spot area , then the number of pixels of the area that breaks through the spot area is , , where ,but ;

[0050] According to the formula Calculate the insulation breakdown rate ,when If it is greater than the preset breakdown threshold, the insulation failure of the locomotive converter is determined, where: is the insulation surface area of ​​the locomotive converter.

[0051] A nondestructive testing method for insulation performance of a locomotive converter comprises the following steps:

[0052] S1: Use the detection device to apply a local discharge signal of a preset intensity to the locomotive converter, generate a local discharge signal in the discharge excitation area by applying discharge voltage excitation on the high voltage side and the low voltage side, and monitor the local discharge signal;

[0053] S2: Perform Fourier transform on the partial discharge signal fed back by the locomotive converter to convert the partial discharge signal into a partial discharge image;

[0054] S3: converting the partial discharge image into a partial discharge grayscale image, which is further converted into a partial discharge halftone image;

[0055] S4: preprocessing the partial discharge grayscale image and the partial discharge halftone image;

[0056] S5: performing correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results. In the correlation analysis, the correlation coefficient between the partial discharge grayscale image and the partial discharge halftone image can be calculated:

[0057] ;

[0058] In the formula, are the values ​​of the corresponding pixels of the two images, is the mean value of image pixels;

[0059] S6: Display, store and analyze the analysis results, and output the final test results of the non-destructive test of the insulation performance of the locomotive converter in the form of images, so that users can intuitively understand the insulation performance status of the locomotive converter.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] 1. The present invention adopts a non-contact detection method through the arrangement of an electrical signal generating device and a partial discharge detection device. During the detection process, there is no need to cut off power, no corona, and no pressure. During the normal operation of the locomotive inverter, the insulation performance can be efficiently and accurately non-destructively tested. Without affecting the normal operation of the locomotive, possible problems in the insulation system can be discovered in time, secondary faults caused by insulation faults can be avoided, and the safety and reliability of the locomotive operation can be guaranteed.

[0062] 2. The present invention also obtains more accurate detection results by focusing on the key features of partial discharge pulses and performing correlation analysis on partial discharge grayscale images and halftone images. It can more sensitively capture subtle changes in the insulation system, help to discover potential insulation failure risks in advance, and provide strong support for maintenance decisions.

[0063] 3. The present invention can not only monitor the local discharge source in real time, accurately determine the specific location of insulation failure, and realize rapid fault location, but also can perform local discharge measurement on the locomotive inverter under normal insulation state, comprehensively covering many aspects of the insulation performance evaluation of the locomotive inverter, from daily routine detection to precise positioning when a fault occurs, providing a one-stop solution for the maintenance and management of the locomotive inverter insulation system, effectively improving the maintenance efficiency and reliability of the insulation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a diagram of the equipment of the present invention. DETAILED DESCRIPTION

[0065] Embodiment 1: Figure 1 As shown, the present invention relates to a nondestructive testing device for insulation performance of a locomotive converter, comprising:

[0066] A detection device, used for applying a partial discharge signal of a preset intensity to a locomotive converter and receiving a partial discharge signal fed back by the locomotive converter, the detection device comprising an electrical signal generating device and a partial discharge detection device, the electrical signal generating device is used for generating a partial discharge signal, and the partial discharge detection device is used for receiving a feedback signal;

[0067] An image extraction unit, connected to the detection device, for performing Fourier transform on the partial discharge signal to convert the partial discharge signal into a partial discharge image;

[0068] An image conversion unit, connected to the image extraction unit, for converting the partial discharge image into a partial discharge grayscale image, and further converting the partial discharge grayscale image into a partial discharge halftone image;

[0069] An image processing unit, the image processing unit is connected to the image conversion unit and is used to perform grayscale segmentation and image filtering on the partial discharge grayscale image;

[0070] An image storage unit, connected to the image processing unit, for storing the partial discharge grayscale image and the partial discharge halftone image respectively;

[0071] A data analysis unit, the data analysis unit is connected to the image processing unit and is used to pre-process the partial discharge halftone image to determine the insulation condition of the locomotive converter;

[0072] A data analysis and processing unit, the data analysis and processing unit is connected to the data analysis unit and is used to perform correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results;

[0073] The detection result output unit is connected to the data analysis and processing unit, and is used to display, store and analyze the analysis results, and output the final detection results of the non-destructive detection of the insulation performance of the locomotive converter in the form of images.

[0074] The present invention adopts a non-contact detection method through the arrangement of an electric signal generating device and a partial discharge detection device. During the detection process, there is no need to cut off power, no corona, and no pressure. During the normal operation of the locomotive inverter, the insulation performance can be efficiently and accurately non-destructively tested. Without affecting the normal operation of the locomotive, possible problems of the insulation system can be discovered in time, secondary faults caused by insulation faults can be avoided, and the safety and reliability of the locomotive operation can be guaranteed.

[0075] Embodiment 2: Figure 1 As shown, as another embodiment of the present invention, a method for nondestructive testing of insulation performance of a locomotive converter comprises the following steps:

[0076] S1: Use the detection device to apply a local discharge signal of a preset intensity to the locomotive converter, generate a local discharge signal in the discharge excitation area by applying discharge voltage excitation on the high voltage side and the low voltage side, and monitor the local discharge signal;

[0077] S2: Perform Fourier transform on the partial discharge signal fed back by the locomotive converter to convert the partial discharge signal into a partial discharge image;

[0078] S3: Convert the partial discharge image into a partial discharge grayscale image, which is further converted into a partial discharge halftone image. The conversion of the partial discharge halftone image is performed by Perform staged processing;

[0079] In the formula, is the pixel point of the partial discharge grayscale image, is the pixel of the partial discharge halftone image, are the number of columns and rows of pixels in the partial discharge grayscale image, respectively;

[0080] S4: preprocessing the partial discharge grayscale image and the partial discharge halftone image;

[0081] S401: partial discharge grayscale image preprocessing;

[0082] S401a: Perform the first binary segmentation on the partial discharge grayscale image. Locating the pixel points in the partial discharge grayscale image whose discharge times exceed a preset number;

[0083] In the formula, The grayscale image of partial discharge is The gray value at is the preset threshold, is the image after binarization;

[0084] S401b: Perform a second binarization segmentation on the partial discharge grayscale image obtained by the first binarization segmentation to locate the rising edge and falling edge of each discharge pulse and obtain the rising edge coordinate data of each pulse. , And falling edge coordinate data , , and draw the rising edge or falling edge curve of each pulse. The judgment of the rising edge and falling edge can be determined by the change of slope. The rising edge and falling edge coordinate data of the partial discharge pulse detected by the second binary segmentation are recorded as: point A Point B , connect point A and point B to get a ray, and record the equation of the ray as ;

[0085] In the formula, , ;

[0086] S401c: Assume that the grayscale level of the partial discharge grayscale image after the second binarization segmentation is , the grayscale values ​​of all pixels between point A and point B for each discharge pulse are counted, and the statistical results are normalized to normalize the pixels to The value of the interval, the gray level corresponding to the pixel is recorded as , the formula is ;

[0087] In the formula, is the gray level corresponding to the pixel, is the value of any pixel in the partial discharge grayscale image. Gray value of

[0088] S401d: Perform a third segmentation on the partial discharge grayscale image obtained by the second binarization segmentation to determine the half-peak time of the partial discharge pulse , the half-peak time data is obtained by calculating the time point when the grayscale value on both sides of the pulse peak drops to half;

[0089] S401e: performing median filtering and mean filtering on the partial discharge grayscale image obtained by the second binarization segmentation to obtain a filtered image;

[0090] The median filter formula is: , where is the original image pixel value, is the pixel value of the filtered image, is the filter window;

[0091] The mean filter formula is: , where is the filter window size, , ;

[0092] S401f: Calculate the local grayscale mean of the filtered image , where The filtered image is The gray value at is the image size;

[0093] S401g: According to the formula Calculate the local grayscale mean square value , enhance the image;

[0094] S401h: Calculate the difference between two adjacent discharge pulse half-peak time data , get the rising edge speed of the discharge pulse , where is the height difference of the rising edges of adjacent pulses;

[0095] S402: Partial discharge halftone image preprocessing includes:

[0096] S402a: The partial discharge halftone image is processed by the OTSU threshold segmentation method to obtain a binary image. The OTSU algorithm calculates the inter-class variance To determine the threshold , so that maximum;

[0097] In the formula, , is the proportion of two types of pixels, , is the mean of two types of pixels;

[0098] S402b: Perform median filtering on the obtained binary image to retain the pixel information of the image, perform morphological dilation operation on the median filtered image to expand the image and obtain the area of ​​the insulation failure region. , the expansion operation formula is , where is the original image, is a structural element;

[0099] S402c: Perform color inversion processing on the partial discharge halftone image obtained by the first binarization segmentation to obtain an inverted image. The inversion formula is: ;

[0100] In the formula, is the original image pixel value, is the pixel value of the image after inversion;

[0101] S402d: locating the pixel positions of the areas where the discharge pulse points D are located in the inverse color image, marking and deleting the located pixel positions respectively to obtain the remaining inverse color image, performing median filtering on the remaining inverse color image to obtain a filtered inverse color image, performing a morphological filling operation on the filtered inverse color image to obtain a filled inverse color image, and the filling operation can be achieved by scanning the image and filling the holes;

[0102] S402e: Perform a morphological corrosion operation on the filled inverted image to obtain an inverted image after corrosion. The corrosion formula is: , perform inversion processing on the inverted image after corrosion to obtain an inverted image;

[0103] S402f: performing a morphological closing operation on the image after the inversion processing to obtain an insulation breakdown spot image, wherein the closing operation is first dilation and then corrosion, and performing a median filter on the insulation breakdown spot image to obtain a filtered insulation breakdown spot image;

[0104] S402g: Perform morphological filling operation on the filtered insulation breakdown spot image to obtain the largest connected area of ​​pixels, denoted as , The number of pixels is recorded as the number of pixels ,but The number of pixels in the area is the number of pixels in the breakdown spot area , then the number of pixels of the area that breaks through the spot area is , , where ,but ;

[0105] S402h: According to the formula Calculate the insulation breakdown rate ,when If it is greater than the preset breakdown threshold, the insulation failure of the locomotive converter is determined, where: is the area of ​​the insulation surface of the locomotive converter;

[0106] S5: performing correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results. In the correlation analysis, the correlation coefficient between the partial discharge grayscale image and the partial discharge halftone image can be calculated:

[0107] ;

[0108] In the formula, are the values ​​of the corresponding pixels of the two images, is the mean value of image pixels;

[0109] S501: Set a pixel point of the partial discharge halftone image, record it as point D, find the grayscale levels corresponding to the horizontal coordinates of two discharge pulses adjacent to point D, and record them as the first grayscale levels respectively. and the second gray level ,like , then point D is located on the first discharge pulse. , then point D is located on the second discharge pulse. , then according to the rising edge speed of the discharge pulses on both sides of the first discharge pulse point D, the discharge pulse corresponding to the horizontal coordinate of point D is determined as the pulse to be analyzed;

[0110] S502: If point D is located on the first discharge pulse, find the gray level corresponding to the horizontal coordinate of the discharge pulse closest to point D, and record it as the third gray level , if the vertical straight line extending from point D intersects with the third pulse curve, and the gray level corresponding to the intersection is , then point D is located on the first discharge pulse, otherwise, point D is located on the second discharge pulse;

[0111] The present invention focuses on the key features of partial discharge pulses and performs correlation analysis on partial discharge grayscale images and halftone images to obtain more accurate detection results. It can more sensitively capture subtle changes in the insulation system, help to discover potential insulation failure risks in advance, and provide strong support for maintenance decisions.

[0112] S6: Display, store and analyze the analysis results, and output the final test results of the non-destructive test of the insulation performance of the locomotive converter in the form of images, so that users can intuitively understand the insulation performance status of the locomotive converter;

[0113] The present invention can not only monitor the local discharge source in real time, accurately determine the specific location of insulation failure, and realize rapid fault location, but also can perform local discharge measurement on the locomotive converter under normal insulation status, comprehensively covering many aspects of the insulation performance evaluation of the locomotive converter, from daily routine detection to precise positioning when a fault occurs, providing a one-stop solution for the maintenance and management of the locomotive converter insulation system, effectively improving the maintenance efficiency and reliability of the insulation system.

[0114] The embodiments of the present invention disclose preferred embodiments, but are not limited thereto. A person skilled in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not deviate from the spirit of the present invention, they are all within the protection scope of the present invention.

Claims

1. A nondestructive testing device for insulation performance of locomotive converter, characterized in that: include: A detection device, used for applying a partial discharge signal of a preset intensity to the locomotive converter and receiving the partial discharge signal fed back by the locomotive converter; An image extraction unit, connected to the detection device, for performing Fourier transform on the partial discharge signal to convert the partial discharge signal into a partial discharge image; An image conversion unit, connected to the image extraction unit, for converting the partial discharge image into a partial discharge grayscale image, and further converting the partial discharge grayscale image into a partial discharge halftone image; An image processing unit, connected to the image conversion unit, for performing grayscale segmentation and filtering processing on the partial discharge grayscale image; An image storage unit, connected to the image processing unit, for storing the partial discharge grayscale image and the partial discharge halftone image respectively; A data analysis unit, the data analysis unit is connected to the image processing unit and is used to pre-process the partial discharge halftone image to determine the insulation condition of the locomotive converter; A data analysis and processing unit, the data analysis and processing unit is connected to the data analysis unit and is used to perform correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results; A test result output unit, which is connected to the data analysis and processing unit and is used to display, store and analyze the analysis results, and output the final test results in the form of images; Among them, the grayscale segmentation of the partial discharge grayscale image includes: The first binary segmentation of the partial discharge grayscale image is performed by The pixel points whose discharge times exceed the preset times in the partial discharge grayscale image are located, where: The grayscale image of partial discharge is The gray value at is the preset threshold, is the image after binarization; The partial discharge grayscale image obtained by the first binarization segmentation is subjected to a second binarization segmentation to locate the rising edge and falling edge of each discharge pulse and obtain the rising edge coordinate data of each pulse. , And falling edge coordinate data , , and draw the rising edge or falling edge curve of each pulse; The partial discharge grayscale image obtained by the second binary segmentation is segmented for the third time to determine the half-peak time of the partial discharge pulse , the half-peak time data is obtained by calculating the time point when the grayscale value on both sides of the pulse peak drops to half; The PD halftone image is converted by Perform hierarchical processing, where is the pixel point of the partial discharge grayscale image, is the pixel of the partial discharge halftone image, are the number of columns and rows of pixels in the partial discharge grayscale image, respectively; Calculate the difference between two adjacent discharge pulse half-peak time data , get the rising edge speed of the discharge pulse , where is the height difference of the rising edges of adjacent pulses; The data analysis processing unit performs correlation analysis on the partial discharge grayscale image and the partial discharge halftone image, and obtains the analysis result, including: Set a pixel point of the partial discharge halftone image as point D, find the grayscale level corresponding to the horizontal coordinates of the two discharge pulses adjacent to point D, and record them as the first grayscale level and the second gray level ,like , then point D is located on the first discharge pulse. , then point D is located on the second discharge pulse. , then according to the rising edge speed of the discharge pulses on both sides of the first discharge pulse point D, the discharge pulse corresponding to the horizontal coordinate of point D is determined as the pulse to be analyzed; If point D is located on the first discharge pulse, find the gray level corresponding to the horizontal coordinate of the discharge pulse closest to point D, and record it as the third gray level , if the vertical straight line extending from point D intersects with the third pulse curve, and the gray level corresponding to the intersection is , then point D is located on the first discharge pulse, otherwise, point D is located on the second discharge pulse; The data analysis unit preprocesses the partial discharge halftone image, including: The partial discharge halftone image is processed by the OTSU threshold segmentation method to obtain a binary image. The OTSU algorithm calculates the variance between classes. To determine the threshold , so that Maximum, where , is the proportion of two types of pixels, , is the mean of two types of pixels; Perform median filtering on the binary image to retain the pixel information of the image, and perform morphological dilation operation on the median filtered image to obtain the area of ​​the insulation failure area. , the expansion operation formula is , where is the original image, is a structural element; The partial discharge halftone image obtained by the first binarization segmentation is inverted to obtain an inverted image. The inverted color formula is: , where is the original image pixel value, is the pixel value of the image after inversion; The pixel positions of the areas where the discharge pulse points D are located in the inverse color image are located, and they are marked and deleted respectively to obtain the remaining inverse color image, and the remaining inverse color image is subjected to median filtering to obtain a filtered inverse color image, and the filtered inverse color image is subjected to morphological filling operation to obtain a filled inverse color image.

2. The nondestructive testing equipment for insulation performance of a locomotive converter according to claim 1 is characterized in that: The coordinate data of the rising edge and falling edge of the partial discharge pulse detected by the second binary segmentation are recorded as: point A Point B , connect point A and point B to get a ray, and record the equation of the ray as , where , .

3. The nondestructive testing equipment for insulation performance of a locomotive converter according to claim 2 is characterized in that: Assume that the grayscale level of the partial discharge grayscale image after the second binarization segmentation is , the grayscale values ​​of all pixels between point A and point B for each discharge pulse are counted and normalized to The value of the interval, the gray level corresponding to the pixel is recorded as , the formula is , where is the gray level corresponding to the pixel, is the value of any pixel of the partial discharge grayscale image The gray value of .

4. The nondestructive testing equipment for insulation performance of a locomotive converter according to claim 3 is characterized in that: The image processing unit performs filtering processing on the partial discharge grayscale image, including: The partial discharge grayscale image obtained by the second binarization segmentation is subjected to median filtering and mean filtering to obtain the filtered image. The median filtering formula is: , where is the original image pixel value, is the pixel value of the filtered image, is the filtering window, and the mean filtering formula is , where is the filter window size, , ; Calculate the local grayscale mean of the filtered image , where The filtered image is The gray value at is the image size, according to the formula Calculate the local grayscale mean square value , enhance the image.

5. The nondestructive testing equipment for insulation performance of locomotive converter according to claim 4, characterized in that: The data analysis unit preprocesses the partial discharge halftone image further comprising: Perform morphological corrosion operation on the filled inverted image to obtain the inverted image after corrosion. The corrosion formula is: , performing inversion processing on the inverted image after corrosion to obtain an inverted image, performing morphological closing operation on the inverted image to obtain an insulation breakdown spot image, performing median filtering on the insulation breakdown spot image to obtain a filtered insulation breakdown spot image; Perform morphological filling operation on the filtered insulation breakdown spot image to obtain the largest connected area of ​​pixels, which is recorded as , The number of pixels is recorded as the number of pixels ,but The number of pixels in the area is the number of pixels in the breakdown spot area , then the number of pixels of the area that breaks through the spot area is , , where if ,but , according to the formula Calculate the insulation breakdown rate ,when If it is greater than the preset breakdown threshold, the insulation failure of the locomotive converter is determined, where: is the insulation surface area of ​​the locomotive converter.

6. A method for nondestructive testing of insulation performance of a locomotive converter, used for the nondestructive testing device for insulation performance of a locomotive converter as claimed in claim 5, characterized in that: The following steps are involved: S1: Use the detection device to apply a partial discharge signal of a preset intensity to the locomotive converter, and monitor the partial discharge signal; S2: Perform Fourier transform on the partial discharge signal fed back by the locomotive converter to convert the partial discharge signal into a partial discharge image; S3: converting the partial discharge image into a partial discharge grayscale image, which is further converted into a partial discharge halftone image; S4: preprocessing the partial discharge grayscale image and the partial discharge halftone image; S5: performing correlation analysis on the partial discharge grayscale image and the partial discharge halftone image to obtain analysis results. In the correlation analysis, the correlation coefficient between the partial discharge grayscale image and the partial discharge halftone image can be calculated: ； In the formula, are the values ​​of the corresponding pixels of the two images, is the mean value of image pixels; S6: Display, store and analyze the analysis results, and output the final test results of the non-destructive test of the insulation performance of the locomotive converter in the form of images, so that users can intuitively understand the insulation performance status of the locomotive converter.

Images