Vibration detection and analysis method of rotational flow self-priming pump based on machine vision

Abstract

The invention discloses a vibration detection and analysis method of a rotational flow self-priming pump based on machine vision. According to the method, the vibration signals in the time domain can be converted into the frequency domain to be analyzed by applying the fast Fourier transform to the processed data. FFT has the advantages that frequency components in vibration signals can be revealed through fast Fourier transform, fault features possibly occurring in equipment operation can be recognized, the faults are represented as specific frequency peak values or spectral lines in a frequency spectrum, and engineers can find potential fault sources in advance and take corresponding maintenance measures. By analyzing the vibration and strain response under different frequencies, the performance of the equipment under a specific working condition can be evaluated. The combination of the fast Fourier transform and the transverse vibration characteristic analysis provides comprehensive technical support for the operation state evaluation of the rotational flow self-priming pump, thereby being beneficial to realizing the optimal operation and maintenance of equipment, reducing the fault risk, prolonging the service life of the equipment and guaranteeing the production safety and efficiency.

Description

Technical Field

[0001] The invention belongs to the technical field of vibration detection and analysis of swirl self-priming pumps, and specifically relates to a vibration detection and analysis method of a swirl self-priming pump based on machine vision. Background Art

[0002] The swirl self-priming pump is a pump that uses the energy of rotating water flow to achieve self-priming. It has the characteristics of simple structure, easy use, good self-priming performance, etc. It is widely used in farmland irrigation, urban water supply and drainage, water conservancy projects, industrial production and other fields. This article will give you a detailed introduction to the working principle, structure, performance and characteristics of the swirl self-priming pump. The working principle of the swirl self-priming pump is based on the swirl principle in fluid mechanics. When the pump is started, the impeller rotates to drive the liquid to move, and the liquid will rotate to form a vortex. In the center of the vortex, due to the action of centrifugal force, the pressure decreases and a negative pressure area is formed. At this time, under the action of the external atmospheric pressure, the liquid enters the pump through the suction pipe to achieve self-priming. As the impeller continues to rotate, the liquid continues to accelerate in the pump cavity, the pressure gradually increases, and finally discharged from the pump through the exhaust pipe. The structure of the swirl self-priming pump is mainly composed of impeller, pump body, suction pipe, exhaust pipe and other parts. The impeller is the core component of the swirl self-priming pump, and its shape and size determine the performance of the pump. The pump body plays the role of containing the liquid and guiding the flow of the liquid. The suction pipe and exhaust pipe are responsible for the suction and discharge of the liquid respectively. Excessive vibration of the swirl self-priming pump during use may cause damage to the equipment, so it is necessary to conduct vibration detection and analysis on it.

[0003] However, common vibration detection and analysis methods are mostly carried out manually during use, which results in large errors in the detection process, which may cause problems to be unable to be discovered and handled in a timely manner, affecting the accuracy of the overall vibration detection of the swirl self-priming pump. Summary of the invention

[0004] The purpose of the present invention is to provide a vibration detection and analysis method of a swirl self-priming pump based on machine vision in order to solve the above-mentioned problems.

[0005] The technical solution adopted by the present invention is as follows: a vibration detection and analysis method of a cyclone self-priming pump based on machine vision, the method comprising the following steps:

[0006] S1: First, install the equipment for vibration and strain monitoring of the machine vision system and collect image information on site. In this step, it is necessary to select the appropriate lens and camera and adjust the distance between them to obtain high-definition equipment images. This step has a direct impact on the performance of the machine vision system, because the quality of the collected images will directly determine the accuracy and effectiveness of subsequent analysis;

[0007] S2: Next, the information transmission module is used to transmit the swirl self-priming pump vibration image data collected in step S1 to the vibration image preprocessing module. The wireless communication equipment in this module can effectively avoid the need for wiring and realize wireless data transmission between different control modules. This can not only simplify the equipment installation process, but also improve the efficiency and stability of data transmission;

[0008] S3: The host computer first stores the received data, and then uses image processing software to preprocess the image. This process includes standardizing the vibration image to eliminate the difference between images, and then performing image alignment preprocessing to eliminate the misalignment between images. Next, based on the rough matching of feature points, the RANSAC algorithm is used to obtain the homography matrix for pixel mapping. Finally, based on this matrix, the pixels of the polarized image are mapped to the corresponding single polarized image to complete the image alignment;

[0009] S4: After the preprocessing of step S3 is completed, the structural vibration and strain monitoring and analysis module is started to monitor and analyze the vibration and strain of the target equipment. The vibration and strain information of the equipment will be extracted based on the collected image data, so as to realize real-time monitoring and analysis of the equipment operation status;

[0010] S5: Apply fast Fourier transform (FFT) to the processed data to analyze the characteristics of each measurement point in the frequency domain. The specific operation is to extract the last cycle of the input and output time series at each frequency, perform FFT transformation, and obtain the principal components. The amplitude ratio and phase difference of these principal components are used to describe the frequency domain characteristics at that frequency. In this way, the vibration and strain response of the equipment at different frequencies can be obtained, thus providing a basis for the evaluation of the equipment's operating status;

[0011] S6: Calculate and analyze the vibration and strain monitoring data of the same batch of samples to obtain the corresponding strain amplitude. Combined with the sensor spectrum data with the DC component removed, statistically analyze the vibration of the equipment under stable working conditions. In addition, by calculating and analyzing the lateral vibration characteristics, the parameters of the equipment structure vibration and strain can be obtained.

[0012] In a preferred embodiment, in step S1, the machine vision-based structural vibration acquisition equipment uses the MER-503-40GC-P Gigabit Network CCD camera with high reliability and stable performance produced by Daheng Co., Ltd. In order to obtain higher-definition images and vibration signals, a corresponding lens is selected, wherein the function of the lens is to focus the reflected light of the actual object on the corresponding photosensitive element.

[0013] In a preferred embodiment, in step S1, the vibration image information acquisition device of the cyclone self-priming pump based on machine vision obtains a high-definition detection image by reasonably adjusting the lens distance; wherein, the resolution of the MER-503-40GC-P Gigabit CCD camera is 2448 (H) × 2048 (V), the maximum frame rate is 40fps, and the trigger mode is hardware trigger / software trigger; the M2518-MPW2 industrial lens selected by the image information acquisition device can automatically adjust the focal length, the focal length is 25mm, the target surface size is 16.9mm, the aperture range is F1.8, and the operating temperature is -20 to 50°C.

[0014] In a preferred embodiment, in step S2, the machine vision-based structural vibration communication module uses the lafalink brand LF-P691 model wireless bridge; the performance parameters of the wireless bridge are as follows: data transmission rate is 450M / s, wireless transmission distance is 3km, network standard is IEEE802.11an / a / ac, memory / flash memory is 64MB+8MB, and the number of cameras is 1 to 11.

[0015] In a preferred embodiment, in step S2, the image acquisition unit and the software processing unit are connected through a wireless bridge device, thereby realizing the medium and long-distance data transmission requirements. In actual operation, the wireless bridge shows strong adaptability to the geographical environment. It can span a long distance and maintain the high quality of the transmission signal. The wireless bridge device can operate in the 2.4G and 5.8G frequency bands, providing a wireless transmission capability of up to 3 kilometers.

[0016] In a preferred embodiment, in step S3, the vibration image normalization processing of the vortex self-priming pump based on machine vision first subtracts the mean of the grayscale of each structural vibration analysis sample, and then divides it by the standard deviation; and adds a small constant to the standard deviation to avoid the denominator being 0 and suppressing noise; for grayscale images in the range of [0,255], adding 10 to the variance is effective; sample-by-sample mean subtraction is mainly used in those data sets with stability, that is, the statistical properties of each dimension of those data are the same; the vibration and strain image to be processed is input into the vibration image alignment network of the vortex self-priming pump based on machine vision, and the aligned vibration image output by the vibration image alignment network of the vortex self-priming pump based on machine vision is obtained.

[0017] In a preferred embodiment, in step S3, during the training process, a batch of vibration and strain image samples are first input into the alignment network, and the network generates a series of predicted alignment images based on these inputs;

[0018] When performing a vibration image alignment operation of a cyclone self-priming pump based on machine vision, a training operation is repeatedly performed until the training of the vibration image alignment network of the cyclone self-priming pump based on machine vision is completed, and the training operation includes: inputting the training vibration and strain image into the vibration image alignment network of the cyclone self-priming pump based on machine vision, and obtaining a predicted aligned vibration and strain image related to the training vibration and strain image output by the vibration image alignment network of the cyclone self-priming pump based on machine vision; obtaining edge feature information of the predicted aligned vibration and strain image and edge feature information of a target vibration and strain image related to the training vibration and strain image; calculating the edge feature loss between the predicted aligned vibration and strain image and the target vibration and strain image based on the edge feature information of the predicted aligned vibration and strain image and the edge feature information of the target vibration and strain image, wherein the edge feature loss indicates the difference in edge features between the predicted aligned vibration and strain image and the target vibration and strain image; and updating the parameter value of the vibration image alignment network of the cyclone self-priming pump based on machine vision based on the edge feature loss.

[0019] In a preferred embodiment, in step S4, for the vibration image analysis and processing of the swirl self-priming pump, an IPC-610L host computer equipped with an IBM motherboard and an Intel I7 processor is selected. The host computer is equipped with 2G DDRII memory and a 16X DVD drive. These hardware resources ensure the smooth operation of the analysis software and meet the needs of data processing.

[0020] In a preferred embodiment, for the vibration signal x(n) of the swirl self-priming pump, the discrete Fourier transform X(k) calculation formula is:

[0021]

[0022] The vibration image strain data x(n) of the swirl self-priming pump based on machine vision is used as input, and the spectrum X(k) is calculated by the above method; the vibration image measurement data of the swirl self-priming pump of 8 sensors is used as the sample signal, and the frequency domain response of the relevant unit equipment under natural load is analyzed; the data collected under the stable working state of the relevant swirl self-priming pump unit equipment (a total of 10 hours) is taken, and one hour is cut as the sample length. After eliminating the interference such as the DC component in the sample and the periodic sinusoidal signal introduced by the rotation of the fan, the spectrum of 10 segments of sample data is obtained.

[0023] In a preferred embodiment, in step S6, the formula for calculating the lateral vibration characteristic analysis method of the structural vibration and strain-related equipment based on machine vision is:

[0024]

[0025] In the formula: fn is the natural frequency of the relevant unit equipment (Hz); T is the tension at both ends of the relevant unit equipment (N); L is the length of the wire rope (m); n is the order of the natural frequency, which is 1; E is the elastic modulus of the wire rope (Pa); I is the moment of inertia of the wire rope (m4); ρ is the linear density of the wire rope (kg / m); For the entire lifting system, the vibration of the relevant unit equipment of the hoist mainly depends on the natural vibration frequency of the relevant unit equipment and the operating speed of the lifting system.

[0026] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0027] 1. In the present invention, by applying fast Fourier transform to the processed data, the vibration signal in the time domain can be converted to the frequency domain for analysis. The beneficial effects of FFT include: the frequency components in the vibration signal can be revealed by fast Fourier transform, which helps to identify the fault characteristics that may occur during the operation of the equipment, such as imbalance, asymmetry, bearing failure, etc. These faults appear as specific frequency peaks or spectral lines in the spectrum, allowing engineers to discover potential sources of faults in advance and take corresponding maintenance measures. By analyzing the vibration and strain response at different frequencies, the performance of the equipment under specific working conditions can be evaluated. For example, the vibration level of the equipment at a specific speed can reflect the rationality of its design and its stability in actual operation. Understanding the vibration response of the equipment at different frequencies through fast Fourier transform helps to optimize the operating parameters of the equipment, such as adjusting the speed, load, etc., to reduce vibration and extend the life of the equipment. Frequency domain analysis provides rich information that can be used to establish data-driven maintenance strategies. By comparing real-time data with historical data, better decisions can be made on when to perform maintenance or replace parts. In summary, the combination of FFT and lateral vibration characteristic analysis provides comprehensive technical support for the operating status evaluation of swirl self-priming pumps, which helps to achieve optimal operation and maintenance of equipment, reduce the risk of failure, extend equipment life, and ensure production safety and efficiency.

[0028] 2. In the present invention, real-time monitoring and analysis of the vibration signal of the vortex self-priming pump is realized: by installing a machine vision system for vibration and strain monitoring, and using wireless communication equipment to realize real-time data transmission, the vibration and strain information of the vortex self-priming pump can be quickly and accurately obtained, thereby realizing real-time monitoring and analysis of the equipment operation status. By selecting highly reliable and stable performance cameras and lenses, as well as wireless bridge equipment, high-definition equipment images and high-quality vibration signals can be obtained, thereby improving the accuracy and stability of vibration detection. The use of wireless communication equipment avoids the need for wiring and realizes wireless data transmission between different control modules, thereby simplifying the installation and maintenance process of the equipment and improving the reliability and stability of the equipment. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0030] Example:

[0031] A vibration detection and analysis method for a cyclone self-priming pump based on machine vision, the method comprising the following steps:

[0032] S1: First, install the equipment for vibration and strain monitoring of the machine vision system and collect image information on site. In this step, it is necessary to select the appropriate lens and camera and adjust the distance between them to obtain high-definition equipment images. This step has a direct impact on the performance of the machine vision system, because the quality of the collected images will directly determine the accuracy and effectiveness of subsequent analysis;

[0033] S2: Next, the information transmission module is used to transmit the swirl self-priming pump vibration image data collected in step S1 to the vibration image preprocessing module. The wireless communication equipment in this module can effectively avoid the need for wiring and realize wireless data transmission between different control modules. This can not only simplify the equipment installation process, but also improve the efficiency and stability of data transmission;

[0034] S3: The host computer first stores the received data, and then uses image processing software to preprocess the image. This process includes standardizing the vibration image to eliminate the difference between images, and then performing image alignment preprocessing to eliminate the misalignment between images. Next, based on the rough matching of feature points, the RANSAC algorithm is used to obtain the homography matrix for pixel mapping. Finally, based on this matrix, the pixels of the polarized image are mapped to the corresponding single polarized image to complete the image alignment;

[0035] S4: After the preprocessing in step S3 is completed, the structural vibration and strain monitoring and analysis module is started to monitor and analyze the vibration and strain of the target equipment. This module will extract the vibration and strain information of the equipment based on the collected image data, thereby realizing real-time monitoring and analysis of the equipment's operating status;

[0036] S5: Apply fast Fourier transform (FFT) to the processed data to analyze the characteristics of each measurement point in the frequency domain. The specific operation is to extract the last cycle of the input and output time series at each frequency, perform FFT transformation, and obtain the principal components. The amplitude ratio and phase difference of these principal components are used to describe the frequency domain characteristics at that frequency. In this way, the vibration and strain response of the equipment at different frequencies can be obtained, thus providing a basis for the evaluation of the equipment's operating status;

[0037] S6: Calculate and analyze the vibration and strain monitoring data of the same batch of samples to obtain the corresponding strain amplitude. Combined with the sensor spectrum data with the DC component removed, statistically analyze the vibration of the equipment under stable working conditions. In addition, by calculating and analyzing the lateral vibration characteristics, the parameters of the equipment structure vibration and strain can be obtained.

[0038] In step S1, the machine vision-based structural vibration acquisition equipment uses the MER-503-40GC-P Gigabit CCD camera with high reliability and stable performance produced by Daheng Co., Ltd. In order to obtain higher-definition images and vibration signals, the corresponding lens is selected, wherein the function of the lens is to focus the reflected light of the actual object on the corresponding photosensitive element.

[0039] In step S1, the vibration image information acquisition device of the cyclone self-priming pump based on machine vision obtains a high-definition detection image by reasonably adjusting the lens distance; wherein, the resolution of the MER-503-40GC-P Gigabit CCD camera is 2448 (H) × 2048 (V), the maximum frame rate is 40fps, and the trigger mode is hardware trigger / software trigger; the M2518-MPW2 industrial lens selected by the image information acquisition device can automatically adjust the focal length, the focal length is 25mm, the target surface size is 16.9mm, the aperture range is F1.8, and the operating temperature is -20~50℃.

[0040] In step S2, the communication module of structural vibration based on machine vision uses the LF-P691 model wireless bridge of the lafalink brand; the performance parameters of the wireless bridge are as follows: the data transmission rate is 450M / s, the wireless transmission distance is 3km, the network standard is IEEE802.11an / a / ac, the memory / flash memory is 64MB+8MB, and the number of cameras is 1 to 11.

[0041] In step S2, the image acquisition unit and the software processing unit are connected through a wireless bridge device, realizing the medium and long-distance data transmission requirements. In actual operation, the wireless bridge shows strong adaptability to the geographical environment. It can span a long distance and maintain the high quality of the transmission signal. The wireless bridge device can operate in the 2.4G and 5.8G frequency bands, providing wireless transmission capabilities up to 3 kilometers.

[0042] In step S3, the vibration image normalization processing of the vortex self-priming pump based on machine vision first subtracts the mean grayscale of each structural vibration analysis sample, and then divides it by the standard deviation; and adds a small constant to the standard deviation to avoid the denominator being 0 and suppressing noise; for grayscale images in the range of [0,255], adding 10 to the variance is effective; sample-by-sample mean subtraction is mainly used in those data sets with stability, that is, the statistical properties of each dimension of those data are the same; the vibration and strain image to be processed is input into the vibration image alignment network of the vortex self-priming pump based on machine vision, and the aligned vibration image output by the vibration image alignment network of the vortex self-priming pump based on machine vision is obtained.

[0043] In step S3, during the training process, a batch of vibration and strain image samples are first input into the alignment network, and the network generates a series of predicted alignment images based on these inputs.

[0044] When performing a vibration image alignment operation of a swirl self-priming pump based on machine vision, a training operation is repeatedly performed until the training of a vibration image alignment network of a swirl self-priming pump based on machine vision is completed, and the training operation includes: inputting a training vibration and strain image into the vibration image alignment network of a swirl self-priming pump based on machine vision, obtaining a predicted aligned vibration and strain image related to the training vibration and strain image output by the vibration image alignment network of a swirl self-priming pump based on machine vision; obtaining edge feature information of the predicted aligned vibration and strain image and edge feature information of a target vibration and strain image related to the training vibration and strain image; calculating an edge feature loss between the predicted aligned vibration and strain image and the target vibration and strain image based on the edge feature information of the predicted aligned vibration and strain image and the edge feature information of the target vibration and strain image, wherein the edge feature loss indicates a difference in edge features between the predicted aligned vibration and strain image and the target vibration and strain image; and updating a parameter value of the vibration image alignment network of a swirl self-priming pump based on machine vision based on the edge feature loss.

[0045] In step S4, for the vibration image analysis and processing of the swirl self-priming pump, an IPC-610L host computer equipped with an IBM motherboard and an Intel I7 processor was selected. The host computer is equipped with 2GDDRII memory and a 16X DVD drive. These hardware resources ensure the smooth operation of the analysis software and meet the needs of data processing.

[0046] For the vibration signal x(n) of the swirl self-priming pump, the discrete Fourier transform X(k) calculation formula is:

[0047]

[0048] The vibration image strain data x(n) of the swirl self-priming pump based on machine vision is used as input, and the spectrum X(k) is calculated by the above method; the vibration image measurement data of the swirl self-priming pump of 8 sensors is used as the sample signal, and the frequency domain response of the relevant unit equipment under natural load is analyzed; the data collected under the stable working state of the relevant swirl self-priming pump unit equipment (a total of 10 hours) is taken, and one hour is cut as the sample length. After eliminating the interference such as the DC component in the sample and the periodic sinusoidal signal introduced by the rotation of the fan, the spectrum of 10 segments of sample data is obtained.

[0049] In step S6, the formula for calculating the lateral vibration characteristic analysis method of the structural vibration and strain-related equipment based on machine vision is:

[0050]

[0051] In the formula: fn is the natural frequency of the relevant unit equipment (Hz); T is the tension at both ends of the relevant unit equipment (N); L is the length of the wire rope (m); n is the order of the natural frequency, which is 1; E is the elastic modulus of the wire rope (Pa); I is the moment of inertia of the wire rope (m4); ρ is the linear density of the wire rope (kg / m); For the entire lifting system, the vibration of the relevant unit equipment of the hoist mainly depends on the natural vibration frequency of the relevant unit equipment and the operating speed of the lifting system.

[0052] In the present invention, by applying fast Fourier transform (FFT) to the processed data, the vibration signal in the time domain can be converted to the frequency domain for analysis. The beneficial effects of FFT include: the frequency components in the vibration signal can be revealed by fast Fourier transform, which helps to identify the fault characteristics that may occur during the operation of the equipment, such as imbalance, asymmetry, bearing failure, etc. These faults appear as specific frequency peaks or spectral lines in the spectrum, so that engineers can find potential sources of faults in advance and take corresponding maintenance measures. By analyzing the vibration and strain response at different frequencies, the performance of the equipment under specific working conditions can be evaluated. For example, the vibration level of the equipment at a specific speed can reflect the rationality of its design and its stability in actual operation. Understanding the vibration response of the equipment at different frequencies through fast Fourier transform helps to optimize the operating parameters of the equipment, such as adjusting the speed, load, etc., to reduce vibration and extend the life of the equipment. Frequency domain analysis provides rich information that can be used to establish data-driven maintenance strategies. By comparing real-time data with historical data, better decisions can be made on when to perform maintenance or replace parts. In summary, the combination of FFT and lateral vibration characteristic analysis provides comprehensive technical support for the operating status evaluation of the swirl self-priming pump, which helps to achieve the optimal operation and maintenance of the equipment, reduce the risk of failure, extend the life of the equipment, and ensure production safety and efficiency.

[0053] In the present invention, real-time monitoring and analysis of the vibration signal of the vortex self-priming pump is realized: by installing a machine vision system for vibration and strain monitoring, and using wireless communication equipment to realize real-time data transmission, the vibration and strain information of the vortex self-priming pump can be quickly and accurately obtained, thereby realizing real-time monitoring and analysis of the equipment operation status. By selecting highly reliable and stable performance cameras and lenses, as well as wireless bridge equipment, high-definition equipment images and high-quality vibration signals can be obtained, thereby improving the accuracy and stability of vibration detection. The use of wireless communication equipment avoids the need for wiring and realizes wireless data transmission between different control modules, thereby simplifying the installation and maintenance process of the equipment and improving the reliability and stability of the equipment.

[0054] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0055] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vibration detection and analysis method for a cyclone self-priming pump based on machine vision, characterized in that: The method comprises the following steps: S1: First, install the equipment used for vibration and strain monitoring of the machine vision system and collect image information on site. In this step, it is necessary to select appropriate lenses and cameras and adjust the distance between them to obtain high-definition equipment images. This step has a direct impact on the performance of the machine vision system, because the quality of the collected images will directly determine the accuracy and effectiveness of subsequent analysis. S2: Next, the vibration image data of the swirl self-priming pump collected in step S1 is transmitted to the vibration image preprocessing module by using the information transmission module; the wireless communication device in this module can effectively avoid the need for wiring and realize wireless data transmission between different control modules; this can not only simplify the installation process of the equipment, but also improve the efficiency and stability of data transmission; S3: The host computer first stores the received data, and then uses image processing software to perform image preprocessing; this process includes standardizing the vibration image to eliminate the difference between images, and then performing image alignment preprocessing to eliminate the misalignment between images; next, based on the rough matching of feature points, the RANSAC algorithm is used to obtain the homography matrix for pixel mapping; finally, based on the matrix, the pixels of the polarized image are mapped to the corresponding single polarized image to complete the image alignment; S4: After the preprocessing of step S3 is completed, the structural vibration and strain monitoring and analysis module is started to monitor and analyze the vibration and strain of the target equipment; the vibration and strain information of the equipment is extracted based on the collected image data, so as to realize real-time monitoring and analysis of the equipment operation status; S5: Apply fast Fourier transform (FFT) to the processed data to analyze the characteristics of each measurement point in the frequency domain; the specific operation is to extract the last cycle of the input and output time series at each frequency, perform FFT transformation, and obtain the principal components; use the amplitude ratio and phase difference of these principal components to describe the frequency domain characteristics at that frequency; in this way, the vibration and strain response of the equipment at different frequencies can be obtained, thereby providing a basis for evaluating the operating status of the equipment; S6: The vibration and strain monitoring data of the same batch of samples are calculated and analyzed to obtain the corresponding strain amplitude; combined with the sensor spectrum data with the DC component removed, the vibration of the equipment under stable working conditions is statistically analyzed; in addition, the parameters of the equipment structure vibration and strain can be obtained by calculating and analyzing the lateral vibration characteristics.

2. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In the step S1, the machine vision-based structural vibration acquisition device uses the MER-503-40GC-P Gigabit CCD camera with high reliability and stable performance produced by Daheng Co., Ltd. In order to obtain higher-definition images and vibration signals, a corresponding lens is selected, wherein the function of the lens is to focus the reflected light of the actual object on the corresponding photosensitive element.

3. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In the step S1, the vibration image information acquisition device of the cyclone self-priming pump based on machine vision obtains a high-definition detection image by reasonably adjusting the lens distance; wherein, the resolution of the MER-503-40GC-P Gigabit CCD camera is 2448 (H) × 2048 (V), the maximum frame rate is 40fps, and the trigger mode is hardware trigger / software trigger; the M2518-MPW2 industrial lens selected by the image information acquisition device can automatically adjust the focal length, the focal length is 25mm, the target surface size is 16.9mm, the aperture range is F1.8, and the operating temperature is -20 to 50°C.

4. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In the step S2, the communication module of the structural vibration based on machine vision uses the LF-P691 model wireless bridge of the lafalink brand; the performance parameters of the wireless bridge are as follows: the data transmission rate is 450M / s, the wireless transmission distance is 3km, the network standard is IEEE802.11an / a / ac, the memory / flash memory is 64MB+8MB, and the number of cameras is 1 to 11.

5. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In step S2, the image acquisition unit and the software processing unit are connected through a wireless bridge device, thereby realizing the medium and long-distance data transmission requirements; in actual operation, the wireless bridge shows strong adaptability to the geographical environment, it can span a long distance and maintain the high quality of the transmission signal; the wireless bridge device can operate in the 2.4G and 5.8G frequency bands, providing a wireless transmission capability of up to 3 kilometers.

6. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In the step S3, the vibration image normalization processing of the vortex self-priming pump based on machine vision first subtracts the mean of the grayscale of each structural vibration analysis sample, and then divides it by the standard deviation; and adds a small constant to the standard deviation to avoid the denominator being 0 and suppressing noise; for grayscale images in the range of [0,255], adding 10 to the variance is effective; sample-by-sample mean subtraction is mainly used in those data sets with stability, that is, the statistical properties of each dimension of those data are the same; the vibration and strain image to be processed is input into the vibration image alignment network of the vortex self-priming pump based on machine vision, and the vibration image after alignment processing output by the vibration image alignment network of the vortex self-priming pump based on machine vision is obtained.

7. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In step S3, during the training process, a batch of vibration and strain image samples are first input into the alignment network, and the network generates a series of predicted alignment images based on these inputs; When performing a vibration image alignment operation of a swirl self-priming pump based on machine vision, a training operation is repeatedly performed until the training of a vibration image alignment network of a swirl self-priming pump based on machine vision is completed, and the training operation includes: inputting a training vibration and strain image into the vibration image alignment network of a swirl self-priming pump based on machine vision, obtaining a predicted aligned vibration and strain image related to the training vibration and strain image output by the vibration image alignment network of a swirl self-priming pump based on machine vision; obtaining edge feature information of the predicted aligned vibration and strain image and edge feature information of a target vibration and strain image related to the training vibration and strain image; calculating an edge feature loss between the predicted aligned vibration and strain image and the target vibration and strain image based on the edge feature information of the predicted aligned vibration and strain image and the edge feature information of the target vibration and strain image, wherein the edge feature loss indicates a difference in edge features between the predicted aligned vibration and strain image and the target vibration and strain image; and updating a parameter value of the vibration image alignment network of a swirl self-priming pump based on machine vision based on the edge feature loss.

8. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In step S4, for the vibration image analysis and processing of the swirl self-priming pump, an IPC-610L host computer equipped with an IBM motherboard and an IntelI7 processor was selected; the host computer was configured with 2G DDRII memory and a 16X DVD optical drive. These hardware resources ensure the smooth operation of the analysis software and meet the needs of data processing.

9. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: For the vibration signal x(n) of the swirl self-priming pump, the discrete Fourier transform X(k) calculation formula is: The vibration image strain data x(n) of the swirl self-priming pump based on machine vision is used as input, and the spectrum X(k) is calculated by the above method; the vibration image measurement data of the swirl self-priming pump of 8 sensors is used as the sample signal, and the frequency domain response of the relevant unit equipment under natural load is analyzed; the data collected under the stable working state of the relevant swirl self-priming pump unit equipment (a total of 10 hours) is taken, and one hour is cut as the sample length. After eliminating the DC component in the sample and the periodic sinusoidal signal interference introduced by the rotation of the fan, the spectrum of 10 segments of sample data is obtained.

10. The vibration detection and analysis method of a cyclone self-priming pump based on machine vision as claimed in claim 1, characterized in that: In step S6, the formula for calculating the lateral vibration characteristic analysis method of the structural vibration and strain-related equipment based on machine vision is: In the formula: fn is the natural frequency of the relevant unit equipment (Hz); T is the tension at both ends of the relevant unit equipment (N); L is the length of the wire rope (m); n is the order of the natural frequency, which is 1; E is the elastic modulus of the wire rope (Pa); I is the moment of inertia of the wire rope (m4); ρ is the linear density of the wire rope (kg / m); For the entire lifting system, the vibration of the relevant unit equipment of the hoist mainly depends on the natural vibration frequency of the relevant unit equipment and the operating speed of the lifting system.

Images