Water cooling wall contour wear detection system

Abstract

The invention discloses a water-cooled wall contour wear detection system, which is used for realizing automatic detection of water-cooled wall outer contour wear, and belongs to the field of automatic detection. The system comprises a platform support, a servo driver, an upper computer PC, an industrial personal computer, a lead screw guide rail mechanism, a sliding table base, a sliding rail, a guide rail sliding block connecting base, a hand-cranking lead screw guide rail mechanism, a 405 nm wavelength linear laser, a multi-position quick release plate, an industrial camera, a movable bearing platform, a motor connecting base and a direct-current servo motor. The system is operated at an upper computer PC end, an industrial personal computer serves as a core, a platform is controlled to move, the surface of the water cooling wall is scanned through linear structured light, an industrial camera synchronously collects deformation stripes, and after image preprocessing, a center line extraction algorithm is combined with the position of a motor to generate a three-dimensional point cloud. The system adopts a standard template to match and calculate the wear depth of each position, realizes non-contact rapid detection of water-cooled wall contour wear, and effectively solves the problems of low efficiency, high cost and the like of traditional manual detection.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic detection, and particularly relates to a water-cooled wall contour wear detection system. Background Art

[0002] With the development of thermal power plants towards large capacity and high parameters, the water-cooled walls used to cool boilers are long-term exposed to high-temperature flames and flue gases. Affected by various corrosion mechanisms, poor air distribution, and pulverized coal erosion in high-speed flow fields, local wear is likely to occur, seriously affecting the operation safety and service life of boilers. Maintenance personnel will choose to repair or replace the water-cooled walls according to the wear situation to prevent large local wear from causing pipe explosion and resulting in abnormal shutdown of the unit, causing serious safety problems and economic losses. Therefore, the wear detection of water-cooled walls is of great value.

[0003] In order to achieve the optimal maintenance cost under the premise of ensuring the safe operation of the boiler, it is necessary to regularly measure the change in the wall thickness. When the local wear amount approaches the critical threshold of the material strength, maintenance personnel will replace the water-cooled wall pipes; for the wear areas that have not reached the dangerous value, periodic detection and protective coating repair are adopted to avoid waste of resources caused by excessive maintenance.

[0004] Traditional water-cooled wall maintenance requires scaffolding or aerial platforms for visual inspection sampling or single-point ultrasonic thickness measurement, which has problems such as low efficiency, difficulty in full inspection, and high cost. Therefore, an automated detection method is needed to achieve real-time tracking of the change trend of the wall thickness and realize the automated wear detection of water-cooled wall pipes.

[0005] With the rapid development of optical three-dimensional reconstruction technology, the contour detection method based on line structured light active sensing has shown significant advantages in the field of industrial detection. By projecting high-precision laser stripes onto the surface of the object to be measured and using the mapping relationship between stripe deformation and three-dimensional curvature, the contour change of the surface depth can be calculated in real time to achieve dynamic scanning and full-region point cloud reconstruction. Compared with traditional single-point measurement, the line structured light technology has the core advantages of non-contact and anti-environmental interference. Summary of the Invention

[0006] Aiming at the problems existing in the water-cooled wall detection and the advantages of three-dimensional reconstruction technology, the present invention discloses a water-cooled wall contour wear detection method, mainly by a center line extraction method for the surface characteristics of the water-cooled wall, three-dimensionally reconstructing the water-cooled wall model by line structured light and calculating the outer contour wear value to determine whether the water-cooled wall is worn. Based on the above method, a water-cooled wall contour wear detection system is constructed, which can realize the automatic detection of the wear of water-cooled wall pipes, improve the detection efficiency, reduce the detection cost, and ensure the safe operation of the unit.

[0007] To achieve the above object, the present invention discloses a water-cooled wall profile wear detection system, which comprises a platform support (1), a servo driver (2) installed on the platform support, a host computer PC (3), an industrial control computer (4), a lead screw guide mechanism (5), a slide base (6), a slide rail (7), a guide rail slider connection seat (8), a hand-operated lead screw guide mechanism (9), a 405-nm wavelength linear laser (10), a multi-position quick-release plate (11), an industrial camera (12), a moving carrier platform (13), a motor connection seat (14), and a DC servo motor (15); wherein the host computer PC (3) is communicatively connected to the industrial control computer (4) through the Ethernet protocol; the industrial control computer (4) communicates bidirectionally with the servo driver (2) through the CAN bus protocol and is connected to the industrial camera (12) through a USB interface; the servo driver (2) receives the rotation speed instruction from the industrial control computer (4) and controls the DC servo motor (15) to drive the linear displacement of the slide base (6) of the lead screw guide mechanism (5); the hand-operated lead screw guide mechanism (9) is slidably engaged with the slide rail (7) through the guide rail slider connection seat (8); the industrial camera (12) is fixed to the multi-position quick-release plate (11) and is equipped with a 405-nm narrow-band filter; the moving carrier platform (13) is used to support the frame and prevent the slide rail (7) from being deformed under force.

[0008] Based on the above method and system composition, the present invention is implemented through the following technical solutions: the industrial control computer (4) runs a real-time operating system and deploys a real-time control program; the host computer PC (3) and the industrial control computer (4) achieve bidirectional communication through the Ethernet protocol; wherein, the host computer PC (4) is used to control the system operation, display the system initialization result, display the states and information during the operation of the industrial camera (12) and the DC servo motor (15), and the result of the extraction of the center line of the water-cooled wall line structure light; after issuing a stop instruction, it receives and analyzes the point cloud data uploaded by the industrial control computer (4) through the network port and generates a three-dimensional model of the water-cooled wall and a detection result; wherein, the industrial control computer (4) runs a real-time operating system and deploys a real-time program, which is used to receive the instruction from the host computer PC (4), initialize the system, control the motor operation, collect images through the industrial camera (12), and extract the center line of the water-cooled wall line structure light, including: image preprocessing, laser center line region segmentation, and sub-pixel level center line coordinate extraction; calculate and record the three-dimensional coordinates of the water-cooled wall profile, and perform wear detection; when receiving the stop instruction sent by the host computer PC (3), the industrial control computer (4) uploads all the recorded three-dimensional point cloud data packets to the host computer PC (3) through the network port, and the system stops working and returns to the initial position;

[0009] The method for extracting the center line of the line structure light for the water wall includes the following steps: image preprocessing, laser center line area segmentation, and sub-pixel level center line coordinate extraction. Among them, the image preprocessing step is: gamma correction (set γ = 1.2) to enhance the contrast of the image contour; use a 3×3 Gaussian kernel to perform convolution operation with each point in the acquired image for Gaussian filtering to effectively suppress the abnormal fluctuation of the gray value at the contour edge; perform a single erosion operation using a 3×3 pixel cross-shaped structure kernel, and perform a morphological operation optimization with a 3×3 cross-shaped kernel for two iterative dilation operations. Among them, the laser center line area segmentation step is: use the Sobel operator for edge detection; perform conditional dilation operation based on a 3×3 square structuring element, and perform a logical AND operation with the original image after iterative seed point dilation (terminate when the current iteration result is the same as the previous iteration result) to achieve the extraction of the connected region; screen the connected region, and combine the characteristics of the water wall contour image, and take the region with a horizontal length greater than 80 pixels and a length-width ratio greater than 4 in the connected region as the effective region; finally, aiming at the problem of center line breakage caused by the low gray value in the water wall gap region, perform a dilation operation using a 5×20 rectangular kernel to expand the effective region by 5 pixels in the horizontal direction on both the left and right and 10 pixels in the vertical direction up and down, and segment the laser center line area. Among them, the sub-pixel level center line coordinate extraction step is: use the Otsu algorithm to determine the threshold for calculating the laser center line area by traversing the gray levels from 0 to 255 according to the criterion of maximizing the between-class variance; divide the water wall contour into a high-brightness region and a low-brightness region through the threshold; select the Gaussian model for sub-pixel center fitting for the high-brightness region, and select the Lorentz model for sub-pixel center fitting for the low-brightness region to obtain the center line pixel coordinates.

[0010] The steps of the wear detection algorithm are as follows: read the three-dimensional coordinate data of the current frame; use the extreme value positioning within a certain step length to locate the positions of the wave peaks (water wall wave peaks) and wave valleys (pipe gaps), and divide the three-dimensional coordinates into independent pipe segments; uniformly select a predetermined number of feature points from the wave peak region, pipe gap region, and complete contour respectively, and perform three independent circle fitting calculations using the least squares method to obtain the corresponding center coordinates and radii; select a suitable standard contour. If the center deviation and radius difference of the three fitting results are small, then use the result with the smallest root mean square error in the three fittings as the standard contour template. If there is a situation where the center height decreases significantly and the radius increases (wave peak wear) or the center shifts horizontally and the radius decreases (unilateral wear) in the three fitting results, then select the fitting result of the pipe gap region (using the unworn data at the bottom) or the unworn side (using the unworn data on the other side) as the standard template respectively; perform a radial comparison between the real-time acquired contour point cloud and the standard template, calculate the wear depth at each location, and mark the over-limit region.

[0011] The platform bracket (1) includes: an aluminum base (0101) serving as an overall load-bearing matrix; two identical 40 aluminum profiles (0102, 0103) fixedly arranged in parallel on both sides of the upper end of the aluminum base (0101); two identical 40 aluminum profiles (0104, 0105) respectively fixedly arranged vertically on the inner sides of the 40 aluminum profiles (0102, 0103) for structural reinforcement; a DC servo motor (15) fixedly arranged on the 40 aluminum profile (0105); two 30 aluminum profiles (0106, 0107), and the 30 aluminum profile (0106) is horizontally installed on the slide base (6).

[0012] The industrial control computer (4) is equipped with an Intel Celeron processor J1900, has 4GB of memory and 256GB of disk storage space, and is equipped with a Linux real-time operating system.

[0013] The industrial camera (12) is a high-speed CMOS black-and-white camera with a resolution of 1280*960; it is equipped with a low-distortion lens with a focal length of 6 - 12mm, the distortion rate at the wide-angle end is ≤0.1%, and at the telephoto end is ≤0.05%; and it is equipped with a 405nm narrow-band filter, with a half-peak width of ±10nm and a center-wavelength transmittance of ≥85%.

[0014] The mobile carrying platform (13) includes: a platform base (1301) provided with a connecting structure extending in two directions; two symmetrically arranged supports (1305) forming a rigid frame with the platform base through the connecting structure; four groups of ball casters (1303) distributed in a 2×2 symmetric array on both sides of the supports; a magnetic attraction assembly including an armature (1304) fixedly connected to the supports and a pair of permanent magnets (1302), wherein the axis of the ball caster forms an orthogonal spatial layout with the center line of the permanent magnet.

[0015] The hand-operated lead screw guide mechanism (9) is installed along the vertical direction (Z-axis) with a stroke of 300mm to adjust the heights of the 405nm wavelength one-dimensional laser (10) and the industrial camera (12), and the height range is 480 - 780mm, corresponding to a field-of-view width range of 350 - 700mm. Generally, a height of about 700mm is adopted, and at this time the field-of-view width range is 350mm; the slide rail (7) is arranged along the horizontal longitudinal direction (X-axis) with a stroke of 570mm, and its direction coincides with the axis of the field-of-view width of the industrial camera (12), so that the total field-of-view width of the system extends to 870 - 1280mm, and the scanning width extends to 920mm at a height of 700mm; the lead screw guide mechanism (5) is installed along the horizontal transverse direction (Y-axis) with a stroke of 510mm and its movement trajectory is parallel to the extending direction of the water-cooled wall tube; the three-axis motion system is connected through the slide base (6) and the guide rail slider connecting seat (8) to form a spatial orthogonal layout; the X-axis slide rail carries the overall lateral translation of the detection system, and the Y-axis guide rail realizes the pose matching between the detection reference plane and the measured pipe system.

[0016] The DC servo motor (15) introduces feedback control and adopts a PI control algorithm to keep the system running stably. Description of the Drawings

[0017] Figure 1 It is a schematic diagram of a water wall profile wear detection system of the present invention.

[0018] Figure 2 It is a schematic diagram of the structure of the mobile carrying platform of the present invention.

[0019] Figure 3 It is a schematic diagram of the structure of the platform bracket of the present invention Detailed Embodiments

[0020] As Figure 1 shown, fix the system on the water wall, and the direction of the water wall tube is the same as that of the lead screw guide mechanism (5); connect the upper computer PC (3) and the industrial control computer (4) through a network cable; connect the industrial control computer (4) and the servo driver (2) through a CAN bus; connect the industrial control computer (4) and the industrial camera (12) through USB; power on the 405nm wavelength one-dimensional laser (10) and the servo driver (2); slide the hand-cranked lead screw guide mechanism (9) to adjust the working height (Z-axis) and longitudinal (X-axis) position; open the front panel program of the upper computer PC (3) and start the system; after starting the program, the overall operation process of the system is as follows: the industrial control computer (4) automatically initializes each module, and after the initialization is completed, a prompt is given in the front panel; control the DC servo motor (15) to drive the lead screw guide mechanism (5) to linearly displace at a preset speed; start the industrial camera (12) to collect the original image of the deformed laser center line; extract the coordinates of the water wall line structured light center line from the image; generate the three-dimensional coordinates of the water wall profile in combination with the internal parameters of the industrial camera (12), the pose of the laser (10) and the motor position and record them; calculate the wear depth through the wear detection algorithm; after the detection is completed: click stop in the front panel program, and the upper computer PC (3) sends a stop command to the industrial control computer (4); the industrial control computer (4) uploads the three-dimensional point cloud data packet through the network port, the system stops and returns to the initial position; the upper computer PC (3) analyzes the data to generate a three-dimensional model of the water wall and the detection result, automatically saves the scanning time and the result, and supports viewing of historical records; the detection is ended; the water wall profile wear detection system described in the present invention can replace the currently relatively backward manual detection method, realize the automatic detection of the wear of the water wall pipeline, not only can significantly improve the detection efficiency and reduce the missed detection rate, but also can save resources and protect the environment.

[0021] The above are only the preferred embodiments of the present invention, which demonstrate the main functional characteristics and innovative advantages of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A water-cooled wall profile wear detection system, which consists of a platform bracket (1), a servo driver (2) installed on the platform bracket, a host computer PC (3), an industrial control computer (4), a lead screw guide mechanism (5), a slide base (6), a slide rail (7), a guide rail slider connection seat (8), a hand-operated lead screw guide mechanism (9), a 405nm wavelength linear laser (10), a multi-position quick-release board (11), an industrial camera (12), a moving carrier platform (13), a motor connection seat (14), and a DC servo motor (15); it is characterized in that: The industrial control computer (4) communicates bidirectionally with the servo driver (2) through the CAN bus protocol and is connected to the industrial camera (12) through a USB interface; it runs a real-time operating system and deploys a real-time program, which is used to receive instructions from the host computer PC (4), initialize the system, control the operation of the motor, collect images through the industrial camera (12), and extract the center line of the structured light of the water-cooled wall, including: image preprocessing, laser center line area segmentation, and sub-pixel level center line coordinate extraction; calculate and record the three-dimensional coordinates of the water-cooled wall profile for wear detection; when receiving the stop instruction sent by the host computer PC (3), the industrial control computer (4) uploads all the recorded three-dimensional point cloud data packets to the host computer PC (3) through the network port, and the system stops working and returns to the initial position; The host computer PC (3) is communicatively connected to the industrial control computer (4) through the Ethernet protocol; it is used to control the operation of the system, display the results of system initialization, display the status and information during the operation of the industrial camera (12) and the DC servo motor (15), and the results of the extraction of the center line of the structured light of the water-cooled wall; after sending the stop instruction, it receives and analyzes the point cloud data uploaded by the industrial control computer (4) through the network port, and generates a three-dimensional model of the water-cooled wall and the detection results.

2. The water wall profile wear detection system according to claim 1, characterized in that The servo driver (2) receives the speed instruction from the industrial control computer (4) and controls the DC servo motor (15) to drive the linear displacement of the slide base (6) of the lead screw guide mechanism (5).

3. The water-cooled wall profile wear detection system according to claim 1, characterized in that The industrial camera (12) is equipped with a 405nm narrow-band filter.

4. The water wall profile wear detection system according to claim 1, characterized in that The platform bracket (1) includes: an aluminum profile base (0101), which serves as the overall load-bearing matrix; two identical 40 aluminum profiles (0102, 0103), which are fixedly arranged in parallel on both sides of the upper end of the aluminum profile base (0101); two identical 40 aluminum profiles (0104, 0105), which are respectively vertically fixed to the inner sides of the 40 aluminum profiles (0102, 0103) for structural reinforcement; the DC servo motor (15), which is fixed to the 40 aluminum profile (0105); two 30 aluminum profiles (0106, 0107), and the 30 aluminum profile (0106) is horizontally installed on the slide base (6).

5. The water-cooled wall profile wear detection system according to claim 1, characterized in that The mobile bearing platform (13) is used to support the frame and prevent the slide rail (7) from deforming under force, and includes: a platform base (1301) provided with a connecting structure extending in two directions; two symmetrically arranged supports (1305) forming a rigid frame with the platform base through the connecting structure; four groups of bull's-eye wheels (1303) distributed symmetrically on both sides of the support in a 2×2 array; a magnetic attraction assembly including an armature (1304) fixed on the support and a pair of permanent magnets (1302), wherein the axis of the bull's-eye wheel forms an orthogonal spatial layout with the center line of the permanent magnet.

6. A method for extracting the center line of line structured light for a water-cooled wall, characterized in that: A water-cooled wall contour wear detection system according to claim 1, comprising the following steps: image preprocessing, laser center line region segmentation, and sub-pixel level center line coordinate extraction; Among them, the image preprocessing step is: gamma correction (set γ = 1.2) to enhance the contrast of the image contour; perform Gaussian filtering by convolving each point in the acquired image with a 3×3 Gaussian kernel to effectively suppress abnormal fluctuations in the gray value at the contour edge; perform a single erosion operation using a 3×3 pixel cross-shaped structural kernel, and perform a morphological operation optimization of two iterative dilation operations using a 3×3 cross-shaped kernel; Among them, the laser center line region segmentation step is: perform edge detection using the Sobel operator; perform conditional dilation operation based on a 3×3 square structuring element, and perform a logical AND operation with the original image after iterative seed point dilation (terminate when the current iteration result is the same as the previous iteration result) to achieve connected region extraction; screen the connected regions, and combine the characteristics of the water-cooled wall contour image to select the region with a horizontal length greater than 80 pixels and an aspect ratio greater than 4 in the connected regions as the effective region; finally, to address the problem of the center line breakage caused by the low gray value in the water-cooled wall gap region, perform a dilation operation using a 5×20 rectangular kernel to expand the effective region by 5 pixels on the left and right in the horizontal direction and 10 pixels up and down in the vertical direction to segment the laser center line region; Among them, the sub-pixel level center line coordinate extraction step is: use the Otsu algorithm to determine the threshold for calculating the laser center line region by traversing the gray levels from 0 to 255 according to the criterion of maximizing the between-class variance; divide the water-cooled wall contour into a high-brightness region and a low-brightness region through the threshold; select a Gaussian model for sub-pixel center fitting for the high-brightness region and a Lorentz model for sub-pixel center fitting for the low-brightness region to obtain the center line pixel coordinates.

7. A method for detecting the profile wear of a water wall, characterized in that: A water-cooled wall profile wear detection system according to claim 1, comprising the following steps: reading the three-dimensional coordinate data of the current frame; using extreme value positioning within a certain step size to locate the positions of the wave peaks (water-cooled wall wave peaks) and wave valleys (pipe gaps), and dividing the three-dimensional coordinates into independent pipe segments; uniformly selecting a predetermined number of feature points from the wave peak region, the pipe gap region, and the complete profile respectively, and performing three independent circle fitting calculations using the least squares method to obtain the corresponding center coordinates and radii; selecting a suitable standard profile. If the center deviation and radius difference of the three fitting results are small, then use the result with the smallest root mean square error in the three fittings as the standard profile template. If there is a situation where the center height decreases significantly and the radius increases (wave peak wear) or the center shifts laterally and the radius decreases (unilateral wear) in the three fitting results, then respectively select the fitting result of the pipe gap region (using the unworn data at the bottom) or the unworn side (using the unworn data on the other side) as the standard template; radially compare the real-time collected profile point cloud with the standard template, calculate the wear depth at each location, and mark the over-limit area.

Images