Engine vibration perception and fault diagnosis method, system, equipment and medium

Abstract

The present invention provides an engine vibration perception and fault diagnosis method, system, equipment and medium, which relates to the field of vibration perception technology, including: step S1: collecting video of the engine running, processing the collected video, and obtaining a processing result; step S2: according to the processing result, amplifying the vibration amplitude without changing the vibration frequency, and then extracting the engine vibration information; step S3: according to the extracted engine vibration information, diagnosing whether a fault occurs. The present invention can avoid the problem that the traditional sensor obtains vibration information when measuring engine vibration, which affects the normal operation of the equipment.

Description

Technical Field

[0001] The present invention relates to the field of vibration sensing technology, and in particular to an engine vibration sensing and fault diagnosis method, system, equipment and medium. Background Art

[0002] By analyzing the engine vibration information, we can understand the service status of the engine and intervene in time for the engine with abnormal operation status. Since the engine vibration is relatively weak, it is difficult to accurately and quickly perceive the engine vibration information and identify whether the vibration information corresponds to the fault state. It is necessary to select a suitable detection principle and design a corresponding feasible detection scheme to realize online monitoring of the engine operation status.

[0003] Currently, laser displacement sensors, accelerometers and other sensors are commonly used to measure vibration information. Laser displacement sensors use the principle of triangulation to shoot visible red lasers at the surface of an object through a lens. The laser scattered by the surface is received by the internal CCD linear camera, and the distance between the sensor and the object is calculated based on the angle and known distance information. Accelerometers first measure the vibration acceleration, and then obtain the vibration displacement through secondary integration. Traditional measurement methods have good measurement accuracy, but because the sensor needs to be deployed on or near the surface of the device being measured, the measured vibration information has deviations due to the introduction of additional mass, and the sensor deployment affects the normal operation of the equipment.

[0004] The wear particles in the oil are detected by ferrography, and the faulty components are located by analyzing the wear particle elements. On this basis, the correlation between the vibration signal morphology and the fault type is established. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides an engine vibration perception and fault diagnosis method, system, equipment and medium.

[0006] According to an engine vibration sensing and fault diagnosis method, system, device and medium provided by the present invention, the scheme is as follows:

[0007] In a first aspect, a method for engine vibration perception and fault diagnosis is provided, the method comprising:

[0008] Step S1: collecting video when the engine is running, and processing the collected video to obtain a processing result;

[0009] Step S2: according to the processing result, the vibration amplitude is amplified without changing the vibration frequency, thereby extracting the engine vibration information;

[0010] Step S3: Diagnose whether a fault occurs based on the extracted engine vibration information.

[0011] Preferably, the step S1 comprises:

[0012] Step S1.1: Calibrate the camera and use the calibrated camera to collect the running video of the engine area to be analyzed;

[0013] Step S1.2: Decompose the collected video into time-sequential single-frame images according to the frame rate, and use a spatial decomposition filter to separate the foreground and background of the single-frame image to obtain foreground features and background features.

[0014] Preferably, the step S2 comprises:

[0015] Step S2.1: After the foreground features and background features are obtained by processing the video, a linear amplification method is used to amplify the separated foreground features;

[0016] Step S2.2: Fusing the amplified foreground features with the background features and reconstructing the image output;

[0017] Step S2.3: Extract the region of interest from each reconstructed image to obtain the strip data features. The region of interest should be selected according to the actual situation, where the color difference is obvious and the edge features are present, so as to eliminate the calculation error.

[0018] Step S2.4: The acquired strip data of the region of interest are merged in time sequence to form a new data composite image; the reconstructed image is binarized using threshold segmentation to obtain a binary image, and the interference of occasional factors is avoided by setting a dynamic threshold, and the boundary attributes are enhanced to facilitate feature recognition;

[0019] Step S2.5: extract boundary features based on the binary image, and determine the upper and lower boundary positions in pixels by combining the reconstructed image when the color value changes;

[0020] Step S2.6: Export the extracted upper and lower boundary pixel position information, transform it into the real distance in the world coordinate system based on the camera calibration parameters, and calculate the time interval between two adjacent image frames based on the video frame rate.

[0021] Preferably, step S3 comprises: recording the real position coordinates of the world coordinate system corresponding to the upper and lower boundary pixels in each reconstructed image, and arranging them into engine vibration signals in chronological order, and calculating the vibration frequency, amplitude, skewness, and kurtosis characteristic parameters based on the acquired engine vibration signals to perform engine fault diagnosis;

[0022] Among them, the vibration frequency is the reciprocal of the time interval between two adjacent maximum or minimum values ​​of the signal, the amplitude is half of the difference between the maximum value and the minimum value of the adjacent position of the signal, and the skewness measures the direction and degree of skewness of the data distribution, reflecting the degree of asymmetry of the data distribution. The calculation formula is;

[0023]

[0024] Where, E represents the mathematical expectation; X represents the engine vibration signal; μ represents the mean value of the engine vibration signal; σ represents the standard deviation of the engine vibration signal;

[0025] Kurtosis measures the statistical characteristics of the sharpness of data distribution, reflecting the steepness of the data distribution shape. The calculation formula is:

[0026]

[0027] Under normal operating conditions, the skewness and kurtosis of the vibration signal should be close to 0. An increase in skewness indicates friction or collision in the engine, and a shift in kurtosis indicates impact vibration in the engine. A ferroscope is used to detect wear particles in the oil, and the faulty components are located by analyzing the wear particle elements, and a correlation between the vibration signal morphology and the fault type is established.

[0028] In a second aspect, an engine vibration sensing and fault diagnosis system is provided, the system comprising:

[0029] Module M1: collects the video when the engine is running, processes the collected video, and obtains the processing result;

[0030] Module M2: Based on the processing results, the vibration amplitude is amplified without changing the vibration frequency, thereby extracting the engine vibration information;

[0031] Module M3: Diagnose whether a fault occurs based on the extracted engine vibration information.

[0032] Preferably, the module M1 comprises:

[0033] Module M1.1: Calibrate the camera and use the calibrated camera to collect the running video of the engine area to be analyzed;

[0034] Module M1.2: Decompose the collected video into time-sequential single-frame images according to the frame rate, use the spatial decomposition filter to separate the foreground and background of the single-frame image, and obtain the foreground features and background features.

[0035] Preferably, the module M2 comprises:

[0036] Module M2.1: After the foreground features and background features are obtained by processing the video, a linear amplification method is used to amplify the separated foreground features;

[0037] Module M2.2: Fuse the enlarged foreground features with the background features and reconstruct the image output;

[0038] Module M2.3: Extract the region of interest for each reconstructed image and obtain the strip data features. The region of interest should be selected according to the actual situation, with obvious color differences and edge features, so as to eliminate calculation errors;

[0039] Module M2.4: Merge the acquired band data of the region of interest in time sequence to form a new data composite image; use threshold segmentation to binarize the reconstructed image to obtain a binary image, set a dynamic threshold to avoid interference from occasional factors, enhance boundary attributes, and facilitate feature recognition;

[0040] Module M2.5: Extract boundary features based on the binary image. When the color value changes, the upper and lower boundary positions are determined in pixels by combining the reconstructed image.

[0041] Module M2.6: Export the extracted upper and lower boundary pixel position information, transform it to the real distance in the world coordinate system based on the camera calibration parameters, and calculate the time interval between two adjacent image frames based on the video frame rate.

[0042] Preferably, the module M3 includes: recording the real position coordinates of the world coordinate system corresponding to the upper and lower boundary pixels in each reconstructed image, and arranging them into engine vibration signals in chronological order, and calculating the vibration frequency, amplitude, skewness, and kurtosis characteristic parameters based on the acquired engine vibration signals to perform engine fault diagnosis;

[0043] Among them, the vibration frequency is the reciprocal of the time interval between two adjacent maximum or minimum values ​​of the signal, the amplitude is half of the difference between the maximum value and the minimum value of the adjacent position of the signal, and the skewness measures the direction and degree of skewness of the data distribution, reflecting the degree of asymmetry of the data distribution. The calculation formula is;

[0044]

[0045] Where, E represents the mathematical expectation; X represents the engine vibration signal; μ represents the mean value of the engine vibration signal; σ represents the standard deviation of the engine vibration signal;

[0046] Kurtosis measures the statistical characteristics of the sharpness of data distribution, reflecting the steepness of the data distribution shape. The calculation formula is:

[0047]

[0048] Under normal operating conditions, the skewness and kurtosis of the vibration signal should be close to 0. An increase in skewness indicates friction or collision in the engine, and a deviation in kurtosis indicates impact vibration in the engine.

[0049] In a third aspect, a computer-readable storage medium storing a computer program is provided, and when the computer program is executed by a processor, the steps of the engine vibration perception and fault diagnosis method are implemented.

[0050] In a fourth aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of the engine vibration perception and fault diagnosis method when executed by the processor.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The present invention only needs to collect engine operation video, without the need for complex sensor installation and deployment;

[0053] 2. The in-situ non-contact measurement method of the present invention can accurately and quickly obtain engine vibration information;

[0054] 3. The present invention can be applied to different engine models to achieve rapid deployment and application.

[0055] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0057] Figure 1 Amplify the algorithm process for the weak vibration information of the engine;

[0058] Figure 2 The motion amplification algorithm is constructed using a neural network. DETAILED DESCRIPTION

[0059] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0060] The embodiment of the present invention provides an engine vibration sensing and fault diagnosis method, which is mainly used to obtain engine vibration information and monitor the service status of the engine. The purpose of the present invention is to realize in-situ non-contact vibration sensing and fault diagnosis, and avoid the problem that when traditional sensors measure engine vibration, the vibration information obtained is biased and affects the normal operation of the equipment.

[0061] Reference Figure 1 and Figure 2 As shown, the method of the present invention includes: integrating image processing, neural network and other methods, collecting video when the engine is running, processing the collected video, amplifying the vibration amplitude without changing the vibration frequency, and then extracting the engine vibration information to diagnose whether a fault occurs.

[0062] Specifically, the steps are as follows:

[0063] S1, calibrate the camera, use the calibrated camera to collect the running video of the engine area to be analyzed;

[0064] S2, decompose the collected video according to the frame rate, that is, obtain video screenshots of the frame rate at equal time intervals per second, including but not limited to using ffmpeg command line, OpenCV video frame reading, Adobe Premiere. For the time series single frame image, use the spatial decomposition filter to separate the foreground and background of the image. The spatial decomposition filter includes a convolution module and a residual module. After the input of two adjacent frames of images, the shape features and texture features of the image are obtained after being processed by the convolution module and the residual module. The shape features are the required foreground features, and the texture features are the background features;

[0065] S3, using a linear magnification method, that is, obtaining a differential image of two adjacent input frames, directly multiplying the differential image by the magnification factor and adding it to the original image. Amplify the separated foreground motion features;

[0066] S4, fuses the amplified foreground features with the background features and reconstructs the image output. The reconstruction module includes an upsampling module, a merging module, a residual module, and a convolution module. The upsampling module is used to reduce memory usage and receptive field size. The merging module combines the foreground features with the background features. The residual module is used to improve the output quality. The convolution module enables the method to run at any resolution.

[0067] S5, extracting the region of interest (ROI) from each reconstructed image to obtain the strip data features. The region of interest should be selected according to the actual situation, with obvious color differences and edge features, so as to eliminate calculation errors;

[0068] S6, merge the acquired band data of the region of interest in time sequence to form a new data composite image. Use threshold segmentation to binarize the reconstructed image, set a dynamic threshold to avoid interference from occasional factors, set the threshold to 30% of the overall brightness of the reconstructed image, enhance the boundary attributes, and facilitate feature recognition;

[0069] S7, extracting boundary features based on the binary image, and determining the upper and lower boundary positions in pixels in combination with the reconstructed image when the color value changes;

[0070] S8, derive the extracted upper and lower boundary pixel position information, and transform it to the real distance in the world coordinate system based on the camera calibration parameters. Calculate the inverse of the video frame rate, and the calculation result is the time interval between two adjacent frames of images;

[0071] S9, calculating characteristic parameters such as vibration frequency, amplitude, skewness, kurtosis, etc. based on the acquired engine vibration information, and performing engine fault diagnosis.

[0072] Among them, the vibration frequency is the reciprocal of the time interval between two adjacent maximum or minimum values ​​of the signal, the amplitude is half of the difference between the maximum value and the minimum value of the adjacent position of the signal, and the skewness measures the direction and degree of skewness of the data distribution, reflecting the degree of asymmetry of the data distribution. The calculation formula is;

[0073]

[0074] Among them, E represents the mathematical expectation; X represents the engine vibration signal; μ represents the mean value of the engine vibration signal; σ represents the standard deviation of the engine vibration signal;

[0075] Kurtosis measures the statistical characteristics of the sharpness of data distribution, reflecting the steepness of the data distribution shape. The calculation formula is:

[0076]

[0077] Under normal operating conditions, the skewness and kurtosis of the vibration signal should be close to 0. Increased skewness indicates friction or collision in the engine, and kurtosis deviation indicates impact vibration in the engine. Ferrograph is used to detect wear particles in the oil, locate faulty components by analyzing wear particle elements, and establish a correlation between the vibration signal morphology and the fault type.

[0078] The present invention also provides an engine vibration sensing and fault diagnosis system, which can be implemented by executing the process steps of the engine vibration sensing and fault diagnosis method, that is, those skilled in the art can understand the engine vibration sensing and fault diagnosis method as a preferred implementation of the engine vibration sensing and fault diagnosis system.

[0079] The system is mainly used to obtain engine vibration information and monitor the service status of the engine. The purpose of the present invention is to achieve on-site non-contact vibration sensing and fault diagnosis, and avoid the problem that when traditional sensors measure engine vibration, there is deviation in the vibration information obtained, which affects the normal operation of the equipment. It includes integrated image processing, neural network and other methods to collect video of the engine when it is running, process the collected video, amplify the vibration amplitude without changing the vibration frequency, and then extract the engine vibration information to diagnose whether a fault has occurred. There is no need for complex sensor installation and deployment, and it will not affect the operation of the engine. The use of this method can achieve accurate and rapid perception of weak engine vibration information, and then perform fault diagnosis. The details are as follows:

[0080] M1, calibrate the camera, use the calibrated camera to collect the running video of the engine area that needs to be analyzed;

[0081] M2, decomposes the collected video into time-sequential single-frame images according to the frame rate, and uses a spatial decomposition filter to separate the foreground and background of the image. The spatial decomposition filter includes a convolution module and a residual module;

[0082] M3, uses a linear amplification method to amplify the separated foreground motion features;

[0083] M4, which fuses the amplified foreground features with the background features and reconstructs the image output. The reconstruction module includes an upsampling module, a merging module, a residual module, and a convolution module;

[0084] M5, extract the region of interest (ROI) for each reconstructed image to obtain the strip data features. The region of interest should be selected according to the actual situation, with obvious color differences and edge features, so as to eliminate calculation errors;

[0085] M6, merge the acquired band data of the region of interest in time sequence to form a new data composite picture. Use threshold segmentation to binarize the image, set a dynamic threshold to avoid interference from occasional factors, enhance boundary attributes, and facilitate feature recognition;

[0086] M7, extracts boundary features based on binary images, and determines the upper and lower boundary positions based on the image when the color value changes, in pixels;

[0087] M8, derives the extracted upper and lower boundary pixel position information, transforms it to the real distance in the world coordinate system based on the camera calibration parameters, calculates the inverse of the video frame rate, and the calculation result is the time interval between two adjacent frames of images;

[0088] M9 calculates characteristic parameters such as vibration frequency, amplitude, kurtosis, skewness, and kurtosis based on the acquired engine vibration information to perform engine fault diagnosis.

[0089] The embodiments of the present invention provide an engine vibration sensing and fault diagnosis method, system, device and medium, which solves the problem that the measurement results of traditional sensors such as laser displacement sensors and acceleration sensors are biased and affect the operation of the equipment under test when sensing engine vibration information and diagnosing faults. The present invention integrates image processing, neural network and other methods, and only needs to shoot a video of the engine running to obtain engine vibration information and diagnose faults in time, realizing in-situ non-contact measurement.

[0090] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0091] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for engine vibration perception and fault diagnosis, characterized in that: include: Step S1: collecting video when the engine is running, and processing the collected video to obtain a processing result; Step S2: according to the processing result, the vibration amplitude is amplified without changing the vibration frequency, thereby extracting the engine vibration information; Step S3: diagnose whether a fault occurs according to the extracted engine vibration information; The step S2 comprises: Step S2.1: After the foreground features and background features are obtained by processing the video, a linear amplification method is used to amplify the separated foreground features; Step S2.2: Fusing the amplified foreground features with the background features and reconstructing the image output; Step S2.3: Extract the region of interest from each reconstructed image to obtain the strip data features. The region of interest should be selected according to the actual situation, where the color difference is obvious and the edge features are present, so as to eliminate the calculation error. Step S2.4: The acquired strip data of the region of interest are merged in time sequence to form a new data composite image; the reconstructed image is binarized using threshold segmentation to obtain a binary image, and the interference of occasional factors is avoided by setting a dynamic threshold, and the boundary attributes are enhanced to facilitate feature recognition; Step S2.5: extract boundary features based on the binary image, and determine the upper and lower boundary positions in pixels in combination with the reconstructed image when the color value changes; Step S2.6: Export the extracted upper and lower boundary pixel position information, transform it to the real distance in the world coordinate system based on the camera calibration parameters, and calculate the time interval between two adjacent image frames based on the video frame rate.

2. The engine vibration sensing and fault diagnosis method according to claim 1, characterized in that: The step S1 comprises: Step S1.1: Calibrate the camera and use the calibrated camera to collect the running video of the engine area to be analyzed; Step S1.2: Decompose the collected video into time-sequential single-frame images according to the frame rate, and use a spatial decomposition filter to separate the foreground and background of the single-frame image to obtain foreground features and background features.

3. The engine vibration sensing and fault diagnosis method according to claim 2, characterized in that: The step S3 comprises: recording the real position coordinates of the world coordinate system corresponding to the upper and lower boundary pixels in each reconstructed image, and arranging them into engine vibration signals in chronological order, and calculating the vibration frequency, amplitude, skewness, and kurtosis characteristic parameters based on the acquired engine vibration signals to perform engine fault diagnosis; Among them, the vibration frequency is the reciprocal of the time interval between two adjacent maximum or minimum values ​​of the signal, the amplitude is half of the difference between the maximum value and the minimum value of the adjacent position of the signal, and the skewness measures the direction and degree of skewness of the data distribution, reflecting the degree of asymmetry of the data distribution. The calculation formula is; Where, E represents the mathematical expectation; X represents the engine vibration signal; μ represents the mean value of the engine vibration signal; σ represents the standard deviation of the engine vibration signal; Kurtosis measures the statistical characteristics of the sharpness of data distribution, reflecting the steepness of the data distribution shape. The calculation formula is: Under normal operating conditions, the skewness and kurtosis of the vibration signal should be close to 0. An increase in skewness indicates friction or collision in the engine, and a deviation in kurtosis indicates impact vibration in the engine.

4. An engine vibration sensing and fault diagnosis system, characterized in that: include: Module M1: collects the video when the engine is running, processes the collected video, and obtains the processing result; Module M2: Based on the processing results, the vibration amplitude is amplified without changing the vibration frequency, thereby extracting the engine vibration information; Module M3: diagnose whether a fault occurs based on the extracted engine vibration information; The module M2 comprises: Module M2.1: After the foreground features and background features are obtained by processing the video, a linear amplification method is used to amplify the separated foreground features; Module M2.2: Fuse the enlarged foreground features with the background features and reconstruct the image output; Module M2.3: Extract the region of interest for each reconstructed image and obtain the strip data features. The region of interest should be selected according to the actual situation, with obvious color differences and edge features, so as to eliminate calculation errors; Module M2.4: Merge the acquired band data of the region of interest in time sequence to form a new data composite image; use threshold segmentation to binarize the reconstructed image to obtain a binary image, set a dynamic threshold to avoid interference from occasional factors, enhance boundary attributes, and facilitate feature recognition; Module M2.5: Extract boundary features based on binary images. When the color value changes, combine the reconstructed image to determine the upper and lower boundary positions in pixels. Module M2.6: Export the extracted upper and lower boundary pixel position information, transform it to the real distance in the world coordinate system based on the camera calibration parameters, and calculate the time interval between two adjacent image frames based on the video frame rate.

5. The engine vibration sensing and fault diagnosis system according to claim 4, characterized in that: The module M1 comprises: Module M1.1: Calibrate the camera and use the calibrated camera to collect the running video of the engine area to be analyzed; Module M1.2: Decompose the collected video into time-sequential single-frame images according to the frame rate, use the spatial decomposition filter to separate the foreground and background of the single-frame image, and obtain the foreground features and background features.

6. The engine vibration sensing and fault diagnosis system according to claim 4, characterized in that: The module M3 includes: recording the real position coordinates of the world coordinate system corresponding to the upper and lower boundary pixels in each reconstructed image, and arranging them into engine vibration signals in chronological order, and calculating the vibration frequency, amplitude, skewness, and kurtosis characteristic parameters based on the acquired engine vibration signals to perform engine fault diagnosis; Among them, the vibration frequency is the reciprocal of the time interval between two adjacent maximum or minimum values ​​of the signal, the amplitude is half of the difference between the maximum value and the minimum value of the adjacent position of the signal, and the skewness measures the direction and degree of skewness of the data distribution, reflecting the degree of asymmetry of the data distribution. The calculation formula is; Where, E represents the mathematical expectation; X represents the engine vibration signal; μ represents the mean value of the engine vibration signal; σ represents the standard deviation of the engine vibration signal; Kurtosis measures the statistical characteristics of the sharpness of data distribution, reflecting the steepness of the data distribution shape. The calculation formula is: Under normal operating conditions, the skewness and kurtosis of the vibration signal should be close to 0. An increase in skewness indicates friction or collision in the engine, and a shift in kurtosis indicates impact vibration in the engine. A ferroscope is used to detect wear particles in the oil, and the faulty components are located by analyzing the wear particle elements, and a correlation between the vibration signal morphology and the fault type is established.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the engine vibration perception and fault diagnosis method according to any one of claims 1 to 3 are implemented.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the engine vibration perception and fault diagnosis method according to any one of claims 1 to 3 are implemented.

Images