Non-contact transformer temperature and vibration test fusion method and system

Abstract

The invention discloses a non-contact transformer temperature and vibration test fusion method and system, and the method comprises the steps: obtaining the steady-state temperature distribution data of a measurement array position of the surface of a transformer through a high-precision infrared imager, and generating a first-layer image result graph of the temperature data of the surface of the transformer; obtaining vibration data of a measurement array position on the surface of the transformer oil tank by adopting a Doppler vibration measurement technology, and generating a second-layer image result graph of the vibration data of the surface of the transformer; respectively carrying out Gaussian pyramid decomposition and Laplacian pyramid generation on the first-layer temperature image and the second-layer vibration image by adopting a Laplacian pyramid fusion method according to the position information of the measurement array, and carrying out weighted fusion and image reconstruction to obtain a fusion result image; according to the invention, a non-contact detection mode is adopted, the problems of installation complexity and environmental adaptability of a traditional sensor are avoided, and fusion processing of two types of images is realized by constructing an array of temperature images and vibration images and establishing a one-to-one correspondence relationship.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer monitoring, and more particularly to a non-contact transformer temperature and vibration test fusion method and system. Background Art

[0002] At present, the temperature and vibration measurement technology of power transformers is constantly developing towards high precision, non-contact and intelligent. By combining multiple sensing technologies with intelligent analysis methods, the comprehensive integration of temperature and vibration data can be achieved.

[0003] However, traditional sensors are complex to install and have poor environmental adaptability. In addition, multi-source information fusion requires processing a large amount of heterogeneous data, which may increase computational complexity and storage costs, while also requiring high synchronization of sensors.

[0004] In recent years, non-contact monitoring technology has gradually become a research hotspot in the field of power equipment status detection due to its advantages of high sensitivity, no installation interference and real-time monitoring.

[0005] Therefore, how to perform multimodal data fusion of temperature and vibration images based on non-contact monitoring technology in order to provide technical support for comprehensive monitoring of transformer operating status and early fault warning is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0006] In view of this, the present invention provides a non-contact transformer temperature and vibration test fusion method and system to solve some of the technical problems mentioned in the background technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] A non-contact transformer temperature and vibration test fusion method comprises the following steps:

[0009] S1. Obtain the steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager, and generate the first layer image result map of the transformer surface temperature data, where the temperature data is consistent with the measurement array position information;

[0010] S2. Use Doppler vibration measurement technology to obtain vibration data of the measurement array position on the surface of the transformer tank, generate a second layer image result map of the transformer surface vibration data, and synchronize the vibration data with the temperature data in space and time;

[0011] S3. According to the position information of the measurement array, the corresponding points in the first layer temperature image and the second layer vibration image are matched one by one, and the Gaussian pyramid decomposition and Laplacian pyramid generation are performed on the first layer temperature image and the second layer vibration image respectively by using the Laplacian pyramid fusion method;

[0012] S4. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed through inverse Laplace pyramid transformation.

[0013] Preferably, the specific contents of step S3 include:

[0014] S31. Preprocessing the first layer temperature image and the second layer vibration image, including standardizing image size, calibration and alignment, and processing noise;

[0015] S32. Mapping the pixel values ​​of the preprocessed temperature and vibration images to the same value range to eliminate data scale differences;

[0016] S33. Performing Gaussian pyramid decomposition on the normalized temperature image and vibration image, including Gaussian smoothing and downsampling, to generate a Gaussian pyramid;

[0017] S34. Perform upsampling and Gaussian convolution processing on each layer of the temperature and vibration Gaussian pyramid, calculate the difference with the previous layer, extract high-frequency details, and generate a Laplacian pyramid.

[0018] Preferably, the specific contents of generating the Gaussian pyramid in step S33 include:

[0019] S331. The temperature image G0 and the vibration image H0 are used as the source images at the bottom of the pyramid, corresponding to the 0th layer of the Gaussian pyramid respectively; l-1 , H l-1 Perform Gaussian smoothing convolution operation, and remove even rows and even columns from the result of the convolution operation to generate the l-th layer low-resolution image G l , H l ; Repeat the Gaussian smoothing and downsampling operations to build a Gaussian pyramid layer by layer until the top N layers are reached;

[0020] S332. During the construction process, image G l-1 , H l-1 is a high-level image G l , H l Magnified 4 times, and then G l+1 , H l+1 Gaussian smoothing and downsampling operations are performed, and the pyramid layer construction is repeated until the top layer image G is decomposed. N , H N , temperature images G0, G1, ..., G N and vibration images H0, H1, ..., H NThese are all the layers of the temperature image and vibration image, providing data for Laplacian pyramid construction.

[0021] Preferably, the generated Gaussian pyramid is represented as:

[0022]

[0023]

[0024] Among them, G l , H l are the temperature image and vibration image of the lth layer in the Gaussian pyramid, G l-1 and H l-1 are the temperature image and vibration image of the l-1th layer in the Gaussian pyramid, i.e., the previous layer image; C l and R l are the number of columns and rows of the l-th layer image respectively, ω(m,n) is the Gaussian kernel, m and n represent the offset of the Gaussian kernel relative to the image pixels when it is applied, and the Gaussian kernel is a two-dimensional matrix used to smooth the image.

[0025] Preferably, the specific contents of step S34 include:

[0026] S341. Temperature image G at the lth layer of the Gaussian pyramid l and vibration image H l Insert zero value 0 into the even rows and even columns to expand the image and restore it to the same level as the l-1th layer G l-1 , H l-1 The same resolution is then used to scale the expanded G using the same Gaussian kernel ω(m, n) as in the downsampling. l , H l Perform convolution operation to obtain a smoothed new gradient image

[0027] S342. Use each layer image G of the Gaussian pyramid l , H l With the previous layer image G l+1 , H l+1 Image after upsampling and Gaussian convolution A difference operation is performed to generate a Laplacian pyramid LPG corresponding to the temperature image and a Laplacian pyramid LPH corresponding to the vibration image.

[0028] Preferably, the new gradient image for:

[0029]

[0030]

[0031] Among them, G′ l , H′ l are the intermediate images expanded by interpolation based on the l-th layer image, ω(m, n) is the Gaussian kernel weight matrix, C l , R l The ratio is the number of columns and rows of the l-th layer image, are the smoothed new gradient images respectively;

[0032] The Laplace pyramid LPG corresponding to the temperature image is:

[0033]

[0034] The Laplacian pyramid LPH corresponding to the vibration image is:

[0035]

[0036] Among them, G l , H l are the lth layer images of the Gaussian pyramid, They are the previous layer images G l+1 , H l+1 After upsampling and smoothing, the image are the temperature and vibration images of the lth layer of the Laplace pyramid, and N is the top layer number of the Laplace pyramid.

[0037] Preferably, the specific content of step S4 is:

[0038] S41. In each layer of the temperature Laplace pyramid and the vibration Laplace pyramid, a dynamic weighted fusion rule is used to fuse the high-frequency detail information to generate a fused Laplace pyramid;

[0039] S42. Perform inverse pyramid transformation on the Laplacian pyramid obtained after fusion, start from the top layer of the pyramid, perform upsampling and addition operations layer by layer, gradually restore to the original resolution, superimpose and reconstruct the fused high-frequency information and low-frequency information, and generate the final fused image.

[0040] Preferably, the specific contents of step S1 include:

[0041] S11. Select a number of measurement points to form a temperature measurement array, that is, a temperature measurement point, and achieve uniform coverage of each measurement point in the transformer tank key area and the radiator area through an array distribution method;

[0042] S12. Adjust the angle and position of the infrared imager to accurately align it with the selected measurement array position, set relevant parameters, start the infrared imager when the transformer is in stable operation, and record the temperature image and data of the selected measurement point on the outer wall of the transformer tank in real time;

[0043] S13. The collected infrared images and temperature data are processed using temperature analysis software, the temperature values ​​of key measurement points are extracted, the temperature gradient and distribution are analyzed, and a temperature distribution map of each measurement point is generated;

[0044] S14. Import the collected array temperature image into the image processing program, analyze the temperature data, convert the temperature value of each measurement point into a visual image, show the temperature distribution on the transformer surface, and generate the first layer image result diagram of the transformer surface temperature data.

[0045] Preferably, the specific contents of step S2 include:

[0046] S21. Select several vibration measurement points to form a measurement array, the measurement points are arranged on the outer wall of the transformer tank, and the layout of the vibration measurement points is coordinated with the temperature measurement points, and the measurement array evenly covers the key areas and radiator areas on the surface of the transformer tank;

[0047] S22. Adjust the measuring head of the laser Doppler vibrometer to align with the selected vibration measurement point, set the acquisition parameters, and collect vibration data and temperature data simultaneously;

[0048] S23. Import the collected vibration measurement array data into the image data processing software. According to the distribution and intensity of the vibration data, use the image processing algorithm to convert the vibration signal into a visual image, display the spatial distribution and intensity variation of the transformer surface vibration, and generate the second layer image result diagram of the transformer surface vibration data.

[0049] A non-contact transformer temperature and vibration test fusion system, based on the non-contact transformer temperature and vibration test fusion method, includes a high-precision infrared imager, a data acquisition card, a laser Doppler vibrometer, a high-precision data acquisition system and a computer terminal;

[0050] The data acquisition card is used to obtain steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager;

[0051] A high-precision data acquisition system for obtaining vibration data of a measurement array position on the surface of a transformer tank using a laser Doppler vibrometer using Doppler vibration measurement technology;

[0052] The computer terminal stores a computer program for generating a first layer image result map of transformer surface temperature data and a second layer image result map of transformer surface vibration data. Based on the measurement array position information, the first layer temperature image is matched with the corresponding points in the second layer vibration image one by one. The temperature image and the vibration image are subjected to Gaussian pyramid decomposition and Laplace pyramid generation respectively by using the Laplace pyramid fusion method. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed and generated by inverse Laplace pyramid transformation.

[0053] It can be seen from the above technical solution that compared with the prior art, the present invention discloses a non-contact transformer temperature and vibration test fusion method and system, which can quickly collect the surface temperature distribution information of the transformer and the laser Doppler vibration sensing technology by adopting high-precision infrared imager thermal imaging technology and can capture the vibration characteristics of the transformer with high precision, and fuse the two temperature and vibration images through an array distribution method. Not only can the transformer surface temperature distribution and vibration characteristic data be obtained simultaneously through the fused image, but also the algorithm analysis can be combined to identify potential problems such as local overheating, insulation aging, winding deformation, etc., which can greatly improve the accuracy of fault diagnosis and early warning capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0055] Figure 1 A schematic diagram of a non-contact transformer temperature and vibration test fusion method provided by the present invention;

[0056] Figure 2 A schematic diagram of a non-contact transformer temperature and vibration test fusion system provided by the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] The embodiment of the present invention discloses a non-contact transformer temperature and vibration test fusion method, such as Figure 1 As shown, the following steps are included:

[0059] S1. Obtain the steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager, and generate the first layer image result map of the transformer surface temperature data, where the temperature data is consistent with the measurement array position information;

[0060] S2. Use Doppler vibration measurement technology to obtain vibration data of the measurement array position on the surface of the transformer tank, generate a second layer image result map of the transformer surface vibration data, and synchronize the vibration data with the temperature data in space and time;

[0061] S3. According to the position information of the measurement array, the corresponding points in the first layer temperature image and the second layer vibration image are matched one by one, and the Gaussian pyramid decomposition and Laplacian pyramid generation are performed on the first layer temperature image and the second layer vibration image respectively by using the Laplacian pyramid fusion method;

[0062] S4. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed through inverse Laplace pyramid transformation.

[0063] In order to further implement the above technical solution, the specific contents of step S3 include:

[0064] S31. Preprocessing the first layer temperature image and the second layer vibration image, including standardizing image size, calibration and alignment, and processing noise;

[0065] In this embodiment, the image size is standardized: the image information is read using an image processing tool to ensure that the pixel width and height of the temperature image and the vibration image as well as the acquisition physical area are consistent; the target resolution of the two images is set;

[0066] Calibration and alignment: Ensure that the pixels of the two images are in the same physical position. Use the calibration point alignment method to extract the corresponding feature points in the temperature image and the vibration image, and use these points to calculate the affine transformation or perspective transformation matrix to clarify the geometric relationship between the two images. Finally, apply the matrix to the image and rotate, translate or stretch it to ensure that the overlapping areas of the two images in space correspond to the same position.

[0067] Noise processing: Using Guided Filter (GF) to remove noise can effectively retain the detailed information of the edge part and ensure the accuracy of the data

[0068] S32. Mapping the pixel values ​​of the preprocessed temperature image and vibration image to the same numerical range to eliminate data scale differences and facilitate subsequent fusion calculations;

[0069] S33. Perform downsampling and convolution calculation on the temperature image and the vibration image to generate a Gaussian pyramid;

[0070] S34. Constructing a Laplacian pyramid describes generating the high-frequency information of the temperature image and the vibration image lost in the Gaussian pyramid, and generates a fused Laplacian pyramid.

[0071] In order to further implement the above technical solution, the specific contents of generating the Gaussian pyramid in step S33 include:

[0072] S331. The temperature image G0 and the vibration image H0 are used as the source images at the bottom of the pyramid, corresponding to the 0th layer of the Gaussian pyramid respectively; l-1 , H l-1 Perform Gaussian smoothing convolution operation, and remove even rows and even columns from the result of the convolution operation to generate the l-th layer low-resolution image G l , H l ; Repeat the Gaussian smoothing and downsampling operations to build a Gaussian pyramid layer by layer until the top N layers are reached;

[0073] S332. During the construction process, image G l-1 , H l-1 is a high-level image G l , H l Magnified 4 times, and then G l+1 , H l+1 Gaussian smoothing and downsampling operations are performed, and the pyramid layer construction is repeated until the top layer image G is decomposed. N , H N , temperature images G0, G1, ..., G N and vibration images H0, H1, ..., H N These are all the layers of the temperature image and vibration image, providing data for Laplacian pyramid construction.

[0074] In order to further implement the above technical solution, the generated Gaussian pyramid is represented as:

[0075]

[0076]

[0077] Among them, G l , H l are the temperature image and vibration image of the lth layer in the Gaussian pyramid, G l-1 and Hl-1 are the temperature image and vibration image of the l-1th layer in the Gaussian pyramid, i.e., the previous layer image; C l and R l are the number of columns and rows of the l-th layer image respectively, ω(m,n) is the Gaussian kernel, m and n represent the offset of the Gaussian kernel relative to the image pixels when applied. The Gaussian kernel is a two-dimensional matrix used to smooth the image. Each element ω(m,n) in the matrix represents the weight value of the kernel at the mth row and nth column.

[0078] In order to further implement the above technical solution, the specific content of step S34 includes:

[0079] S341. Temperature image G at the lth layer of the Gaussian pyramid l and vibration image H l Insert zero value 0 into the even rows and even columns to expand the image and restore it to the same level as the l-1th layer G l-1 , H l-1 The same resolution is then used to scale the expanded G using the same Gaussian kernel ω(m, n) as in the downsampling. l , H l Perform convolution operation to obtain a smoothed new gradient image

[0080] S342. Use each layer image G of the Gaussian pyramid l , H l With the previous layer image G l+1 , H l+1 Image after upsampling and Gaussian convolution Perform a difference operation to generate the Laplacian pyramid LPG corresponding to the temperature image and the Laplacian pyramid LPH corresponding to the vibration image. The calculation formula is:

[0081]

[0082] In order to further implement the above technical solution, the new gradient image for:

[0083]

[0084]

[0085] Among them, G′ l , H′ l are the intermediate images expanded by interpolation based on the l-th layer image, ω(m, n) is the Gaussian kernel weight matrix, C l , R l The ratio is the number of columns and rows of the l-th layer image, They are respectively obtained by l, H′ l Perform Gaussian convolution operation to recalculate all pixel values ​​and obtain a smooth upsampled image;

[0086] The Laplace pyramid LPG corresponding to the temperature image is:

[0087]

[0088] The Laplacian pyramid LPH corresponding to the vibration image is:

[0089]

[0090] Among them, G l , H l are the lth layer images of the Gaussian pyramid, They are the previous layer images G l+1 , H l+1 After upsampling and smoothing, the image are the temperature and vibration images of the lth layer of the Laplace pyramid, and N is the top layer number of the Laplace pyramid.

[0091] In order to further implement the above technical solution, the specific content of step S4 is:

[0092] S41. In each layer of the temperature Laplace pyramid and the vibration Laplace pyramid, a dynamic weighted fusion rule is used to fuse the high-frequency detail information to generate a fused Laplace pyramid;

[0093] S42. Perform inverse pyramid transformation on the Laplacian pyramid obtained after fusion, start from the top layer of the pyramid, perform upsampling and addition operations layer by layer, gradually restore to the original resolution, superimpose and reconstruct the fused high-frequency information and low-frequency information, and generate the final fused image.

[0094] Furthermore, the calculation method of inverse pyramid transformation, i.e. image reconstruction, is:

[0095]

[0096]

[0097] in, For the top layer G N and H N Directly take the corresponding Laplace pyramid value.

[0098] By performing an inverse pyramid transform (i.e., image reconstruction) on the Laplace pyramid obtained after fusing the temperature and vibration images, the advantageous local clear parts of each image can be taken into the fused image, ultimately achieving complementary advantages, reducing distortion, and reflecting more original details. The reconstructed image obtained is the fused image.

[0099] In order to further implement the above technical solution, the specific contents of step S1 include:

[0100] S11. Select a number of measurement points to form a temperature measurement array, that is, a temperature measurement point, and achieve uniform coverage of each measurement point in the transformer tank key area and the radiator area through an array distribution method;

[0101] In this embodiment, in order to fully capture the temperature characteristics and potential anomalies of the transformer, the measurement points need to be evenly distributed in key areas such as the top, side walls, bottom and radiator of the outer wall of the transformer oil tank, focusing on covering the winding, the outer wall of the core and the cooling oil path. 3-4 measurement points are arranged on the top to monitor the hot oil rising area, 3-4 measurement points are arranged on each side of the side wall to capture the temperature gradient, 3-4 measurement points are arranged on the bottom to reflect the cold oil sinking temperature, and at least 1 measurement point is arranged at the inlet and outlet of the radiator to evaluate the heat dissipation performance. 10-20 measurement points are selected to achieve uniform coverage through array distribution methods, while increasing the density of measurement points in key areas, combining infrared thermal imaging to supplement abnormal area monitoring, ensuring that the layout is easy to install and maintain, and providing comprehensive and reliable data support for status monitoring and fault diagnosis;

[0102] S12. Adjust the angle and position of the infrared imager to accurately align it with the selected measurement array position, set relevant parameters, start the infrared imager when the transformer is in stable operation, and record the temperature image and data of the selected measurement point on the outer wall of the transformer tank in real time;

[0103] To ensure the reliability and measurement accuracy of the high-precision infrared imager, it should be fixed on a tripod or bracket, and the angle should be adjusted to accurately align with the selected measurement array position; according to the operating conditions of the transformer, set the appropriate temperature range, resolution and measurement distance to ensure that the measurement area is fully covered and the temperature change details can be captured; during the installation process, it is necessary to ensure that the instrument is stable to avoid external vibrations that interfere with the measurement results. At the same time, regularly check the parameter settings of the fixed equipment and imager to ensure long-term stability and accuracy;

[0104] During the temperature data collection process, a certain collection time is maintained to ensure that the steady-state temperature distribution can be captured, the temperature changes are monitored in real time, and whether there are abnormally high temperature areas. The measurement parameters are adjusted in time according to the situation to ensure the accuracy and effectiveness of the data;

[0105] S13. The collected infrared images and temperature data are processed using temperature analysis software, the temperature values ​​of key measurement points are extracted, the temperature gradient and distribution are analyzed, and a temperature distribution map of each measurement point is generated;

[0106] In this embodiment, the focus is on temperature anomalies. By analyzing the distribution characteristics and change trends of abnormal areas, possible heat dissipation problems or potential failure risks are determined, providing data support and decision-making basis for transformer operation and maintenance.

[0107] Step S13 may also include comprehensively evaluating the thermal status of the transformer and generating a monitoring report; the temperature monitoring report includes the temperature value of each measurement point, the description of the abnormal location, and the temperature distribution map; the infrared image, analysis data, and report should be archived uniformly to ensure future reference and backtracking. The report not only helps to deeply understand the thermal characteristics of the transformer, but also provides key data support for subsequent equipment maintenance, monitoring, and fault diagnosis. These records will serve as an important basis for future operation evaluation and optimization strategies to ensure the long-term safety and reliability of transformer equipment;

[0108] S14. The collected array temperature data is imported into an image processing program, the temperature data is parsed by an algorithm, the temperature value of each measuring point is converted into a visual image, the temperature distribution of the transformer surface is displayed, and the first layer image result map of the transformer surface temperature data is generated;

[0109] Through different colors and temperature scales, the image can clearly reflect the temperature differences in various areas of the transformer surface, highlighting potential hot spots or temperature anomalies.

[0110] In this embodiment, before step S11, the equipment for temperature measurement is also prepared, including a high-precision infrared imager, a stable tripod or bracket, reflective materials or reflective markers, as well as data acquisition equipment and temperature analysis software; in actual applications, an infrared imager for power equipment is selected to ensure that it has high resolution and long-distance accurate measurement capabilities, so as to comprehensively monitor the temperature distribution on the transformer surface.

[0111] In order to further implement the above technical solution, the specific contents of step S2 include:

[0112] The specific contents of step S2 include:

[0113] S21. Select a number of vibration measurement points to form a measurement array. The measurement points should be arranged on the outer wall of the transformer tank, and ensure that the layout of the vibration measurement points is coordinated with the temperature measurement points. The measurement array evenly covers the key areas and radiator areas on the surface of the transformer tank;

[0114] S22. Adjust the measuring head of the laser Doppler vibrometer to align with the selected vibration measurement point, set the acquisition parameters, and ensure that the vibration data is collected synchronously with the temperature data;

[0115] These include: installing and calibrating the vibration meter, firmly mounting the vibration meter on the bracket to ensure that it can be aligned with all measurement points, calibrating according to the requirements of the vibration meter manual to ensure measurement accuracy, and ensuring the accuracy and reliability of subsequent data; setting measurement parameters, setting the sampling frequency (usually twice the target frequency) and sampling time according to the working characteristics of the transformer to ensure that the complete vibration signal of the equipment under different working conditions can be captured; ensuring the synchronization of multi-point data collection for subsequent analysis;

[0116] S23. After the collected array vibration data is imported into the image data processing software, the vibration signal is converted into a visual image using an image processing algorithm according to the distribution and intensity of the vibration data, showing the spatial distribution, intensity changes and potential abnormal areas of the transformer surface vibration, and generating a second layer of image result map of the transformer surface vibration data.

[0117] Before step S21, the method further includes preparing a laser Doppler vibrometer, a reflective film or a high-reflective patch, a high-precision data acquisition system, and signal processing software. The vibrometer is used to capture the surface vibration of the transformer, the reflective film improves the stability of the laser signal, and the data acquisition system and signal processing software are used to record and analyze data in real time.

[0118] A non-contact transformer temperature and vibration test fusion system, based on a non-contact transformer temperature and vibration test fusion method, such as Figure 2 As shown, it includes a high-precision infrared imager, a data acquisition card, a laser Doppler vibrometer, a high-precision data acquisition system and a computer terminal;

[0119] The data acquisition card is used to obtain steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager;

[0120] A high-precision data acquisition system for obtaining vibration data of a measurement array position on the surface of a transformer tank using a laser Doppler vibrometer using Doppler vibration measurement technology;

[0121] The computer terminal stores a computer program for generating a first layer image result map of transformer surface temperature data and a second layer image result map of transformer surface vibration data. Based on the measurement array position information, the first layer temperature image is matched with the corresponding points in the second layer vibration image one by one. The temperature image and the vibration image are subjected to Gaussian pyramid decomposition and Laplace pyramid generation respectively by using the Laplace pyramid fusion method. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed and generated by inverse Laplace pyramid transformation.

[0122] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0123] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A non-contact transformer temperature and vibration test fusion method, characterized in that: The following steps are involved: S1. Obtain the steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager, and generate the first layer image result map of the transformer surface temperature data, where the temperature data is consistent with the measurement array position information; S2. Use Doppler vibration measurement technology to obtain vibration data of the measurement array position on the surface of the transformer tank, generate a second layer image result map of the transformer surface vibration data, and synchronize the vibration data with the temperature data in space and time; S3. According to the position information of the measurement array, the corresponding points in the first layer temperature image and the second layer vibration image are matched one by one, and the Gaussian pyramid decomposition and Laplacian pyramid generation are performed on the first layer temperature image and the second layer vibration image respectively by using the Laplacian pyramid fusion method; S4. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed through inverse Laplace pyramid transformation.

2. A non-contact transformer temperature and vibration test fusion method according to claim 1, characterized in that: The specific contents of step S3 include: S31. Preprocessing the first layer temperature image and the second layer vibration image, including standardizing image size, calibration and alignment, and processing noise; S32. Mapping the pixel values ​​of the preprocessed temperature and vibration images to the same value range to eliminate data scale differences; S33. Performing Gaussian pyramid decomposition on the normalized temperature image and vibration image, including Gaussian smoothing and downsampling, to generate a Gaussian pyramid; S34. Perform upsampling and Gaussian convolution processing on each layer of the temperature and vibration Gaussian pyramid, calculate the difference with the previous layer, extract high-frequency details, and generate a Laplacian pyramid.

3. A non-contact transformer temperature and vibration test fusion method according to claim 2, characterized in that: The specific contents of step S33 generating the Gaussian pyramid include: S331. The temperature image G0 and the vibration image H0 are used as the source images at the bottom of the pyramid, corresponding to the 0th layer of the Gaussian pyramid respectively; l-1 , H l-1 Perform Gaussian smoothing convolution operation, and remove even rows and even columns from the result of the convolution operation to generate the l-th layer low-resolution image G l , H l ; Repeat the Gaussian smoothing and downsampling operations to build a Gaussian pyramid layer by layer until the top N layers are reached; S332. During the construction process, image G l-1 , H l-1 is a high-level image G l , H l Magnified 4 times, and then G l+1 , H l+1 Gaussian smoothing and downsampling operations are performed, and the pyramid layer construction is repeated until the top layer image G is decomposed. N , H N , temperature images G0,G1,…,G N and vibration images H0,H1,…,H N These are all the layers of the temperature image and vibration image, providing data for Laplacian pyramid construction.

4. A non-contact transformer temperature and vibration test fusion method according to claim 3, characterized in that: The generated Gaussian pyramid is represented as: Among them, G l , H l are the temperature image and vibration image of the lth layer in the Gaussian pyramid, G l-1 and H l-1 are the temperature image and vibration image of the l-1th layer in the Gaussian pyramid, i.e., the previous layer image; C l and R l are the number of columns and rows of the l-th layer image respectively, ω(m,n) is the Gaussian kernel, m and n represent the offset of the Gaussian kernel relative to the image pixels when applied, and the Gaussian kernel is a two-dimensional matrix used to smooth the image.

5. A non-contact transformer temperature and vibration test fusion method according to claim 2, characterized in that: The specific contents of step S34 include: S341. Temperature image G at the lth layer of the Gaussian pyramid l and vibration image H l Insert zero value 0 into the even rows and even columns to expand the image and restore it to the same level as the l-1th layer G l-1 , H l-1 The same resolution is then used to scale the expanded G using the same Gaussian kernel ω(m,n) as in the downsampling. l , H l Perform convolution operation to obtain a smoothed new gradient image S342. Use each layer image G of the Gaussian pyramid l , H l With the previous layer image G l+1 , H l+1 Image after upsampling and Gaussian convolution A difference operation is performed to generate a Laplacian pyramid LPG corresponding to the temperature image and a Laplacian pyramid LPH corresponding to the vibration image.

6. A non-contact transformer temperature and vibration test fusion method according to claim 5, characterized in that: New Gradient Image for: Among them, G' l , H' l are the intermediate images expanded by interpolation based on the l-th layer image, ω(m,n) is the Gaussian kernel weight matrix, C l , R l The ratio is the number of columns and rows of the l-th layer image, are the smoothed new gradient images respectively; The Laplace pyramid LPG corresponding to the temperature image is: The Laplacian pyramid LPH corresponding to the vibration image is: Among them, G l , H l are the lth layer images of the Gaussian pyramid, are the previous layer images G l+1 , H l+1 After upsampling and smoothing, the image are the temperature and vibration images of the lth layer of the Laplace pyramid, and N is the top layer number of the Laplace pyramid.

7. A non-contact transformer temperature and vibration test fusion method according to claim 1, characterized in that: The specific content of step S4 is: S41. In each layer of the temperature Laplace pyramid and the vibration Laplace pyramid, a dynamic weighted fusion rule is used to fuse the high-frequency detail information to generate a fused Laplace pyramid; S42. Perform inverse pyramid transformation on the Laplacian pyramid obtained after fusion, start from the top layer of the pyramid, perform upsampling and addition operations layer by layer, gradually restore to the original resolution, superimpose and reconstruct the fused high-frequency information and low-frequency information, and generate the final fused image.

8. The non-contact transformer temperature and vibration test fusion method according to claim 1 is characterized in that: The specific contents of step S1 include: S11. Select a number of measurement points to form a temperature measurement array, that is, a temperature measurement point, and achieve uniform coverage of each measurement point in the transformer tank key area and the radiator area through an array distribution method; S12. Adjust the angle and position of the infrared imager to accurately align it with the selected measurement array position, set relevant parameters, start the infrared imager when the transformer is in stable operation, and record the temperature image and data of the selected measurement point on the outer wall of the transformer tank in real time; S13. The collected infrared images and temperature data are processed using temperature analysis software, the temperature values ​​of key measurement points are extracted, the temperature gradient and distribution are analyzed, and a temperature distribution map of each measurement point is generated; S14. Import the collected array temperature image into the image processing program, analyze the temperature data, convert the temperature value of each measurement point into a visual image, show the temperature distribution on the transformer surface, and generate the first layer image result diagram of the transformer surface temperature data.

9. The non-contact transformer temperature and vibration test fusion method according to claim 1, characterized in that: The specific contents of step S2 include: S21. Select several vibration measurement points to form a measurement array, the measurement points are arranged on the outer wall of the transformer tank, and the layout of the vibration measurement points is coordinated with the temperature measurement points, and the measurement array evenly covers the key areas and radiator areas on the surface of the transformer tank; S22. Adjust the measuring head of the laser Doppler vibrometer to align with the selected vibration measurement point, set the acquisition parameters, and collect vibration data and temperature data simultaneously; S23. Import the collected vibration measurement array data into the image data processing software. According to the distribution and intensity of the vibration data, use the image processing algorithm to convert the vibration signal into a visual image, display the spatial distribution and intensity variation of the transformer surface vibration, and generate the second layer image result diagram of the transformer surface vibration data.

10. A non-contact transformer temperature and vibration test fusion system, characterized in that: A non-contact transformer temperature and vibration test fusion method based on any one of claims 1 to 9, comprising a high-precision infrared imager, a data acquisition card, a laser Doppler vibrometer, a high-precision data acquisition system and a computer terminal; The data acquisition card is used to obtain steady-state temperature distribution data of the measurement array position on the transformer surface through a high-precision infrared imager; A high-precision data acquisition system for obtaining vibration data of a measurement array position on the surface of a transformer tank using a laser Doppler vibrometer using Doppler vibration measurement technology; The computer terminal stores a computer program for generating a first layer image result map of transformer surface temperature data and a second layer image result map of transformer surface vibration data. Based on the measurement array position information, the first layer temperature image is matched with the corresponding points in the second layer vibration image one by one. The temperature image and the vibration image are subjected to Gaussian pyramid decomposition and Laplace pyramid generation respectively by using the Laplace pyramid fusion method. According to the fusion rule, the temperature and vibration image features of each level are weightedly fused to generate a fused Laplace pyramid. The fused temperature and vibration result map is reconstructed and generated by inverse Laplace pyramid transformation.

Images