A Gear Surface Morphology Intelligent Detection and Analysis System

Abstract

This application discloses an intelligent detection and analysis system for gear surface morphology, which includes an image acquisition unit for acquiring images of the gear surface inside a gearbox. The system acquires these images and extracts surface morphology feature values ​​to generate gear surface morphology feature recording data. It also acquires cumulative unit operating time, cumulative unit power generation, gearbox usage time, and gearbox failure operating time to generate gear operating history data. Combining the gear surface morphology feature recording data and the gear operating history data, a gearbox life analysis model is established using a deep learning algorithm to analyze and calculate the gearbox life. This intelligent detection and analysis system for gear surface morphology has the advantages of convenient and simple detection methods and higher accuracy compared to purely manual inspection.

Description

Technical Field

[0001] This invention relates to the field of wind turbine maintenance technology, and in particular to an intelligent detection and analysis system for gear surface morphology. Background Technology

[0002] Wind turbine gearbox failures leading to downtime and out-of-line maintenance have caused significant economic losses. Conventional measures, such as periodically changing specified lubricating oil and replacing filters, cannot address the problem of continued wear. Furthermore, recent maintenance data shows an increasing number of gearbox failures requiring outage maintenance. Statistics indicate that on average, one gearbox is replaced annually in each wind farm's eight-year operational lifespan, with some farms replacing as many as four, representing a rate ranging from 3% to 10%. Traditional maintenance methods require disassembling and replacing the gearbox entirely or partially using large lifting equipment while the turbine is shut down. To reduce the economic losses associated with gearbox repair or replacement, various online gear repair and strengthening technologies are employed, such as "online metal surface strengthening and repair technology." This advanced surface treatment technology can repair wear and damage to mechanical parts without disassembling or shutting down the equipment. It has gained widespread attention from major wind power operators in China in recent years.

[0003] To ensure stable operation of the gearbox after online repair, the repair effect needs to be evaluated to calculate its lifespan. The evaluation of repair effectiveness still relies on traditional testing methods, such as vibration analysis, oil analysis, roughness measurement, oil temperature monitoring, endoscopic inspection, and mobile phone photography. Vibration analysis, as a primary assessment method, has low resolution and cannot reflect the effectiveness of "online repair technology," especially for micro-repairs of pitting corrosion. Endoscopic inspection can conveniently collect images of the gear surface condition and reflect the wear state, with low testing cost. However, traditional endoscopic inspection relies on expert opinion, lacks precise and quantitative calculations, and is subject to significant human factors. Therefore, the inspection results can only serve as a reference for other methods. Summary of the Invention

[0004] The purpose of this invention is to provide a system capable of detecting and analyzing damage to the surface of gears inside a gearbox, thereby enabling the assessment of the lifespan of gears inside the gearbox.

[0005] Therefore, this invention discloses an intelligent detection and analysis system for gear surface morphology, which includes an image acquisition unit for acquiring images of the gear surface inside the gearbox.

[0006] The image acquisition unit acquires the gear surface image and extracts the gear surface morphological feature values ​​to generate gear surface morphological feature recording data.

[0007] The system acquires the unit's cumulative operating time, cumulative power generation, gearbox usage time, and gearbox failure operating time to generate gear running history data.

[0008] By combining the recorded data on the surface morphology of the gears and the data on the gear running history, a gearbox life analysis model is established using a deep learning algorithm to analyze and calculate the life of the gearbox.

[0009] In some embodiments of this application, a method for extracting surface morphological feature values ​​of gears is disclosed, thereby realizing the extraction of surface morphological feature values ​​of gears. The method for extracting surface morphological feature values ​​of gears includes:

[0010] A pre-defined gear surface morphology feature recognition library is used to determine gear wear features in the gear surface image.

[0011] A gear surface image is acquired, and based on the gear surface morphology feature recognition library, the gear surface morphology feature values ​​that conform to gear wear characteristics in the gear surface image are scanned and determined.

[0012] In some embodiments of this application, a method for presetting the gear surface morphology feature recognition library is disclosed, thereby realizing the identification and determination of gear surface morphology feature values. The method for presetting the gear surface morphology feature recognition library includes:

[0013] Based on the surface morphological characteristics of gears, such as wear, fracture, pitting, spalling, plastic flow, scratches, scuffing, corrosion, and charring, the range of surface morphological characteristic values ​​of gears is divided.

[0014] This application discloses a more specific method for obtaining the surface morphological feature values ​​of a gear. The method for obtaining the surface morphological feature values ​​of a gear includes:

[0015] The gear surface image is scanned and analyzed using an image line detection algorithm to determine the gear body boundary;

[0016] Connect the defined boundaries of the gear body to determine the gear body detection area;

[0017] Based on the image edge detection algorithm, the detection area of ​​the gear body is scanned and analyzed to determine the outline of the damaged area and the area value of the damaged area outline.

[0018] Based on the image grayscale detection algorithm, the contour of the damaged area is scanned and analyzed to determine the grayscale value of the damaged area;

[0019] The damage type of the damaged area is determined based on its area and grayscale value.

[0020] The damage type, area value, and grayscale value of the damaged block within the gear body detection area are the surface morphological feature values ​​of the gear.

[0021] In some embodiments of this application, in order to determine the damage type and damage degree of the acquired gear surface morphology feature values, a method for dividing the range of gear surface morphology feature values ​​is disclosed. The method for dividing the range of gear surface morphology feature values ​​includes:

[0022] The range of characteristic values ​​for gear surface morphology includes the area range of the damaged area and the grayscale range of the damaged area.

[0023] For damage blocks of different damage types, obtain several sets of area values ​​and grayscale values ​​for the damage blocks;

[0024] The maximum and minimum area values ​​of several groups of the damaged block are determined as the endpoint values ​​of the area interval of the damaged block;

[0025] The maximum and minimum gray values ​​of several groups of the damaged block are determined as the endpoint values ​​of the gray range of the damaged block.

[0026] In some embodiments of this application, in order to determine the remaining life of the gearbox, a method for analyzing and calculating the remaining life of the gearbox is disclosed. The method for analyzing and calculating the remaining life of the gearbox may further include:

[0027] A gear running history analysis library is pre-set, which is used to determine the first gear life consumption value corresponding to different gear running history data;

[0028] A pre-set gear surface morphology feature analysis library is provided, which is used to determine the second gear life consumption value corresponding to different gear surface morphology feature record data.

[0029] The remaining lifespan of the gearbox is determined based on the lifespan consumption values ​​of the first gear and the second gear.

[0030] In some embodiments of this application, a method for presetting the gear running history analysis library is disclosed, thereby calculating the lifespan of the gearbox. The method for presetting the gear running history analysis library includes:

[0031] The cumulative operating time of the unit, the cumulative power generation of the unit, the gearbox usage time, and the gearbox failure operating time are all divided into segments, and a specific first gear life consumption value is set for each segment.

[0032] In some embodiments of this application, a method for presetting the gear surface morphology feature analysis library is disclosed, which enables the analysis and calculation of the lifespan of the gearbox. The method for presetting the gear surface morphology feature analysis library includes:

[0033] The data recorded for different gear surface morphology features are divided into segments, and a specific second gear life consumption value is set for each segment.

[0034] In some embodiments of this application, a unit capable of analyzing and processing gear surface images acquired by the image acquisition unit is disclosed. The system also employs an analysis and processing unit, which is used to analyze and process the gear surface morphology feature recording data and gear running history data to determine the lifespan of the gearbox.

[0035] This application discloses an intelligent detection and analysis system for gear surface morphology. It acquires images of the gear surface inside the gearbox through an image acquisition unit, generates gear surface morphology feature recording data, and combines the generated gear running history data with a deep learning algorithm to establish a gearbox life analysis model and then analyze and calculate the life of the gearbox. It has the advantages of convenient and simple detection methods, and is more accurate than general manual experience inspection.

[0036] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0037] Figure 1 This is a method and steps for analyzing the surface morphology of gears disclosed in the embodiments of this application.

[0038] Figure 2 This application discloses a method and steps for obtaining surface morphological feature values ​​of gears. Detailed Implementation

[0039] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0040] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0041] Example:

[0042] The purpose of this invention is to provide a system capable of detecting and analyzing damage to the surface of gears inside a gearbox, thereby enabling the assessment of the lifespan of gears inside the gearbox.

[0043] Therefore, this invention discloses an intelligent detection and analysis system for gear surface morphology, which includes an image acquisition unit for acquiring images of the gear surface inside the gearbox.

[0044] The method for detecting and analyzing the surface morphology of a gearbox using the gear surface image includes the following steps:

[0045] Step 1: Obtain the gear surface image acquired by the image acquisition unit, and extract the gear surface morphological feature values ​​to generate gear surface morphological feature record data.

[0046] Step 2: Obtain the unit's cumulative operating time, the unit's cumulative power generation, the gearbox's usage time, and the gearbox's fault operating time to generate gear running history data.

[0047] Step 3: Combining the recorded data of the gear surface morphology features and the gear running history data, a gearbox life analysis model is established using a deep learning algorithm to analyze and calculate the life of the gearbox.

[0048] Corresponding to step 1, in order to extract the surface morphological feature values ​​of the gear, some embodiments of this application disclose a method for extracting the surface morphological feature values ​​of the gear. The method for extracting the surface morphological feature values ​​of the gear includes:

[0049] a. A gear surface morphology feature recognition library is preset, which is used to determine the gear wear features in the gear surface image.

[0050] b. Acquire a gear surface image, and based on the gear surface morphology feature recognition library, scan and determine the gear surface morphology feature values ​​in the gear surface image that conform to gear wear characteristics.

[0051] Corresponding to step 1, in some embodiments of this application, a method for presetting the gear surface morphology feature recognition library is disclosed, thereby realizing the identification and determination of gear surface morphology feature values. The method for presetting the gear surface morphology feature recognition library includes: dividing the range of gear surface morphology feature values ​​according to the characteristics of gear surface morphology feature values ​​such as wear, fracture, pitting, peeling, plastic flow, scratch, scuffing, corrosion and charring.

[0052] Corresponding to step 1, this application discloses a more specific method for obtaining gear surface morphology feature values. The method for obtaining gear surface morphology feature values ​​includes:

[0053] Step 101: Scan and analyze the gear surface image according to the image line detection algorithm to determine the gear body boundary.

[0054] Step 102: Connect the determined boundaries of the gear body to determine the gear body detection area.

[0055] Step 103: Based on the image edge detection algorithm, the detection area of ​​the gear body is scanned and analyzed to determine the outline of the damaged area and the area value of the damaged area outline.

[0056] Step 104: Based on the image grayscale detection algorithm, the contour of the damaged area is scanned and analyzed to determine the grayscale value of the damaged area.

[0057] Step 105: Determine the damage type of the damaged area based on the area value and grayscale value of the damaged area.

[0058] The damage type, area value, and grayscale value of the damaged block within the gear body detection area are the surface morphological feature values ​​of the gear.

[0059] Corresponding to step 3, in some embodiments of this application, in order to determine the damage type and damage degree of the acquired gear surface morphology feature values, a method for dividing the range of gear surface morphology feature values ​​is disclosed. The method for dividing the range of gear surface morphology feature values ​​includes:

[0060] The range of characteristic values ​​for gear surface morphology includes the range of damaged area and the range of grayscale value of damaged area.

[0061] Step 301: For damage blocks of different damage types, obtain several sets of area values ​​and grayscale values ​​of the damage blocks.

[0062] Step 302: Determine the maximum and minimum area values ​​of several groups of the damaged area as the endpoint values ​​of the area range of the damaged area.

[0063] Step 303: Determine the maximum and minimum gray values ​​of several groups of the damaged block as the endpoint values ​​of the gray range of the damaged block.

[0064] Corresponding to step 3, in some embodiments of this application, in order to determine the remaining life of the gearbox, a method for analyzing and calculating the remaining life of the gearbox is disclosed. The method for analyzing and calculating the remaining life of the gearbox may further include:

[0065] A pre-set gear running history analysis library is provided, which is used to determine the first gear life consumption value corresponding to different gear running history data.

[0066] Step 3001: A gear surface morphology feature analysis library is preset. The gear surface morphology feature analysis library is used to determine the second gear life consumption value corresponding to different gear surface morphology feature record data.

[0067] Step 3002: Determine the remaining lifespan of the gearbox based on the lifespan consumption value of the first gear and the lifespan consumption value of the second gear.

[0068] Corresponding to step 3, in some embodiments of this application, a method for presetting the gear running history analysis library is disclosed, thereby calculating the lifespan of the gearbox. The method for presetting the gear running history analysis library includes: dividing the unit's cumulative running time value, the unit's cumulative power generation, the gearbox's usage time value, and the gearbox's fault running time value into segments, and setting a specific first gear lifespan consumption value for each segment.

[0069] Corresponding to step 3, in some embodiments of this application, a method for presetting the gear surface morphology feature analysis library is disclosed, which can then analyze and calculate the lifespan of the gearbox. The method for presetting the gear surface morphology feature analysis library includes: dividing different gear surface morphology feature recording data into segments, and setting a specific second gear lifespan consumption value for each segment.

[0070] In some embodiments of this application, a unit capable of analyzing and processing gear surface images acquired by the image acquisition unit is disclosed. The system also employs an analysis and processing unit, which is used to analyze and process the gear surface morphology feature recording data and gear running history data to determine the lifespan of the gearbox.

[0071] To further illustrate the technical solution of this application, a method and steps for analyzing the surface morphology of gears are now disclosed in conjunction with a specific application scenario, thereby enabling accurate and effective evaluation of the lifespan of gears in a gearbox.

[0072] Traditional inspection methods include vibration analysis, oil analysis, roughness measurement, oil temperature monitoring, endoscopic inspection, and mobile phone photography. Vibration analysis, as a primary assessment method, suffers from low resolution and cannot reflect the effectiveness of "online repair technology," especially for micro-repairs of pitting corrosion. Endoscopic inspection can conveniently collect images of gear surface conditions and reflect gear wear, with low inspection costs. However, traditional endoscopic inspection relies on expert opinion, lacks precise and quantitative calculations, and is heavily influenced by human factors. This means the inspection results can only serve as a reference for other methods.

[0073] To address the shortcomings of the aforementioned detection methods, such as Figure 1 The method steps include:

[0074] S1, acquire the gear surface image acquired by the image acquisition unit, and extract the gear surface morphological feature values ​​to generate gear surface morphological feature recording data.

[0075] S2, acquire the unit's cumulative operating time, the unit's cumulative power generation, the gearbox's usage time, and the gearbox's fault operating time to generate gear running history data.

[0076] S3. Combining the recorded data of the gear surface morphology features and the gear running history data, a gearbox life analysis model is established using a deep learning algorithm to analyze and calculate the life of the gearbox.

[0077] The gearbox life model collects, processes, analyzes, judges, and diagnoses the surface morphology of the gearboxes in the wind farm project, and evaluates the gearbox operating status and remaining service life.

[0078] In step S1, the method for extracting gear surface morphology feature values ​​includes: having a preset gear surface morphology feature recognition library, which is used to determine gear wear features in the gear surface image; acquiring a gear surface image, and scanning and determining gear surface morphology feature values ​​in the gear surface image that conform to gear wear features according to the gear surface morphology feature recognition library.

[0079] The method for pre-setting the gear surface morphology feature recognition library includes: dividing the range of gear surface morphology feature values ​​according to the characteristics of gear surface morphology feature values ​​such as wear, fracture, pitting, peeling, plastic flow, scratches, scuffing, corrosion and charring.

[0080] Corresponding to step S1, the method for obtaining the surface morphological feature values ​​of the gear includes the following steps:

[0081] S4. Based on the image line detection algorithm, the surface image of the gear is scanned and analyzed to determine the boundary of the gear body.

[0082] S5, connect the determined boundaries of the gear body to determine the gear body detection area.

[0083] S6. Based on the image edge detection algorithm, the detection area of ​​the gear body is scanned and analyzed to determine the outline of the damaged area and the area value of the damaged area outline.

[0084] S7. Based on the image grayscale detection algorithm, the contour of the damaged area is scanned and analyzed to determine the grayscale value of the damaged area.

[0085] S8. Determine the damage type of the damaged area based on the area value and gray value of the damaged area.

[0086] The damage type, area value, and grayscale value of the damaged block within the gear body detection area are the surface morphological feature values ​​of the gear.

[0087] The main principle of the image edge detection algorithm mentioned in step S6 is to identify pixels in a digital image that show significant changes in color or brightness. These significant changes in pixels often represent important changes in that part of the image's attributes, including discontinuities in depth, direction, and brightness.

[0088] Image edge detection algorithms first roughly detect some pixels of the image outline when detecting edges. Then, they connect these pixels using connection rules. Finally, they detect and connect previously unidentified boundary points, remove false pixels and boundary points, and form a unified edge. However, in real images, edges are often the edges of various types of objects or blurred landscapes. Furthermore, real images may contain noise, which, like edges, is high-frequency signal information. Therefore, traditional bandgap filtering methods are not very effective for edge detection.

[0089] Currently, there are many commonly used edge detection models: first-order models include the Roberts operator, Prewitt operator, Sobel operator, and Canny operator; second-order models include the Laplacian operator. Image edge detection is based on the image gradient, and obtaining the image gradient is transformed into performing convolution operations on the image using various operators.

[0090] This application discloses an intelligent detection and analysis system for gear surface morphology. It acquires images of the gear surface inside the gearbox through an image acquisition unit, generates gear surface morphology feature recording data, and combines the generated gear running history data with a deep learning algorithm to establish a gearbox life analysis model and then analyze and calculate the life of the gearbox. It has the advantages of convenient and simple detection methods, and is more accurate than general manual experience inspection.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A gear surface morphology intelligent detection and analysis system, characterized in that, The application includes an image acquisition unit, which is used to acquire images of the gear surface inside the gearbox; The system also employs an analysis and processing unit, which is used for: The image acquisition unit acquires the gear surface image and extracts the gear surface morphological feature values ​​to generate gear surface morphological feature recording data. The system acquires the unit's cumulative operating time, cumulative power generation, gearbox usage time, and gearbox failure operating time to generate gear running history data. By combining the recorded data of gear surface morphology features and gear running history data, a gearbox life analysis model is established using a deep learning algorithm to analyze and calculate the gearbox life. The method for analyzing and calculating the lifespan of the gearbox includes: A gear running history analysis library is pre-set, which is used to determine the first gear life consumption value corresponding to different gear running history data; A pre-set gear surface morphology feature analysis library is provided, which is used to determine the second gear life consumption value corresponding to different gear surface morphology feature record data. The remaining lifespan of the gearbox is determined based on the lifespan consumption values ​​of the first gear and the second gear. The method for presetting the gear running history analysis library includes: The cumulative operating time of the unit, the cumulative power generation of the unit, the gearbox usage time, and the gearbox failure operating time are all divided into segments, and a specific first gear life consumption value is set for each segment.

2. The intelligent detection and analysis system for gear surface morphology according to claim 1, characterized in that, Methods for extracting surface morphological feature values ​​of gears include: A pre-defined gear surface morphology feature recognition library is used to determine gear wear features in the gear surface image. A gear surface image is acquired, and based on the gear surface morphology feature recognition library, the gear surface morphology feature values ​​that conform to gear wear characteristics in the gear surface image are scanned and determined.

3. The intelligent detection and analysis system for gear surface morphology according to claim 2, characterized in that, The method for pre-setting the gear surface morphology feature recognition library includes: Based on the surface morphological characteristics of gears, such as wear, fracture, pitting, spalling, plastic flow, scratches, scuffing, corrosion, and charring, the range of surface morphological characteristic values ​​of gears is divided.

4. The intelligent detection and analysis system for gear surface morphology according to claim 3, characterized in that, Methods for obtaining the surface morphological feature values ​​of gears include: The gear surface image is scanned and analyzed using an image line detection algorithm to determine the gear body boundary; Connect the defined boundaries of the gear body to determine the gear body detection area; Based on the image edge detection algorithm, the detection area of ​​the gear body is scanned and analyzed to determine the outline of the damaged area and the area value of the damaged area outline. Based on the image grayscale detection algorithm, the contour of the damaged area is scanned and analyzed to determine the grayscale value of the damaged area; The damage type of the damaged area is determined based on its area and grayscale value. The damage type, area value, and grayscale value of the damaged block within the gear body detection area are the surface morphological feature values ​​of the gear.

5. The intelligent detection and analysis system for gear surface morphology according to claim 4, characterized in that, Methods for dividing the range of characteristic values ​​of gear surface morphology include: The range of characteristic values ​​for gear surface morphology includes the area range of the damaged area and the grayscale range of the damaged area. For damage blocks of different damage types, obtain several sets of area values ​​and grayscale values ​​for the damage blocks; The maximum and minimum area values ​​of several groups of the damaged block are determined as the endpoint values ​​of the area interval of the damaged block; The maximum and minimum gray values ​​of several groups of the damaged block are determined as the endpoint values ​​of the gray range of the damaged block.

6. The intelligent detection and analysis system for gear surface morphology according to claim 1, characterized in that, The method for presetting the gear surface morphology feature analysis library includes: The data recorded for different gear surface morphology features are divided into segments, and a specific second gear life consumption value is set for each segment.

Images