Method for predicting removal function of flexible bladder tool based on polishing pad wear

Abstract

A prediction method for judging the removal function of a flexible airbag tool based on polishing pad wear. The method obtains point cloud data of the polishing pad on the surface of the airbag tool through a laser line scanner, and realizes the detection of the polishing pad on the surface of the large-size airbag tool through point cloud data splicing. Finally, the wear amount is calculated by combining the airbag wear detection algorithm, and the removal function is predicted according to the wear amount of the polishing pad. This method solves the problem that the wear amount of the polishing pad on the surface of the airbag tool cannot be accurately identified. Compared with the traditional visual judgment of the polishing pad state, it can better quantify the tool state and predict the removal function in real time, thereby better shaping and polishing the surface of various elements.

Description

Technical Field

[0001] This invention relates to the field of ultra-precision airbag polishing technology, and in particular to a method for predicting the removal function of flexible airbag tools based on polishing pad wear. Background Technology

[0002] In the field of optical polishing, airbag polishing technology is used in the rapid polishing stage. Due to its excellent conformal polishing capabilities and high removal efficiency, airbag polishing is increasingly used in the rapid polishing processes of various components. However, the lifespan evaluation and removal stability of airbag tools have become bottlenecks limiting the development of airbag polishing technology. Currently, in production processes, relying solely on manual visual assessment of the wear state of the polishing pad on the airbag surface makes it impossible to quantify the wear condition and predict the different removal functions resulting from wear. Summary of the Invention

[0003] The purpose of this invention is to solve the problem in the prior art that the wear of the polishing pad in the airbag polishing process lacks a quantitative indicator, making it impossible to specifically assess the impact on polishing. This invention provides a prediction method for the removal function of flexible airbag tools based on the wear of the polishing pad.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A prediction method for the removal function of flexible airbag tools based on polishing pad wear includes the following steps:

[0006] 1) After the airbag finishes one polishing process, move the airbag to the laser line scanner area for scanning. When performing splicing scanning, it is necessary to ensure that there are overlapping areas at the seams.

[0007] 2) After obtaining point cloud data through scanning in step 1), the point cloud data is preprocessed to reduce noise, outliers are removed, and data is stitched and registered.

[0008] 3) After obtaining the surface data of the airbag in step 2), extract the cross-sectional profile of the airbag through the center of the sphere, calculate the center and radius of the fitted ideal circle according to the least squares method, obtain the ideal circular profile, and make the ideal circular profile coincide with the actual airbag cross-sectional profile. Take the difference between the ideal profile and the actual profile in the advance angle direction, which is the wear depth.

[0009] 4) Based on the detected wear depth of the polishing pad, compare it with the polishing pad's full lifespan state database to determine the current material removal capacity of the polishing pad, predict the removal function of the current state, and compensate for the residence time.

[0010] In step 1), the proportion of the repeated region is 20% to 40%.

[0011] In step 2), the point cloud data is preprocessed for noise reduction using voxel filtering and radius filtering methods.

[0012] In step 2), when scanning large workpieces, it is necessary to stitch together the scan data. The stitching and registration relationship between multiple surface data is calculated based on the correspondence of the data acquisition platform.

[0013] The calculation steps are as follows: obtain the translation amount of the point cloud coordinate axis according to the translation distance of the data acquisition platform; select three obvious feature points; measure the distance between the three obvious feature points in adjacent point cloud segments to determine the translation transformation matrix; and complete the data splicing and registration of the corresponding relationship.

[0014] After completing the data splicing and registration of the corresponding relationship, this invention also includes fine registration of point clouds: ICP algorithm registration based on KD nearest neighbor search, using point cloud fragments as target point clouds, constructing a bidirectional K-dimensional tree, traversing the source point cloud P and target point cloud Q after initial registration as the initial point set for fine registration of ICP algorithm, and obtaining the optimal rigid body transformation through the classic ICP algorithm to achieve fine registration of point clouds.

[0015] In step 4), the time term in the removal function of the Princeton equation is compensated, and the residence time of the residence point is modified to keep the removal capacity consistent.

[0016] Compared with the prior art, the beneficial effects achieved by the technical solution of this invention are:

[0017] This invention can quickly acquire the three-dimensional morphology of the airbag surface, monitor the state of the polishing pad, obtain the wear depth value, and then correct the removal function, which can improve the polishing accuracy and polishing efficiency of the component surface. Attached Figure Description

[0018] Figure 1 This is a flowchart of the prediction method of the present invention;

[0019] Figure 2 This is a schematic diagram of the mechanical structure of the present invention;

[0020] Figure 3 This is a flowchart of point cloud data processing.

[0021] Figure 4 A schematic diagram for point cloud data registration. Detailed Implementation

[0022] To make the technical problems, technical solutions and beneficial effects of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0023] This invention mounts a laser line scanner on a worktable. A robot moves an airbag mounted on its end effector above the scanner to scan, obtaining multiple sets of point cloud data of the surface morphology of a polishing pad attached to the airbag surface. After filtering the point cloud data, the multiple sets of data are stitched and registered to obtain accurate polishing pad surface data. Then, the central contour line passing through the center is extracted, and the ideal contour line with no wear depth is obtained using the least squares method. The actual contour line is subtracted from the ideal contour line along the precession angle; the difference is the wear depth of the polishing pad in the working area. Finally, by modifying the scaling constant of the removal function based on the Princeton equation, the removal function of the polishing pad under different wear states is predicted.

[0024] See Figures 1-4 The specific usage method and operation process of this invention are as follows:

[0025] 1. Tool Installation

[0026] Before polishing, a polyurethane polishing pad is adhered to the surface of airbag 2, then airbag 2 is installed onto servo motor 1, and finally the servo motor is fixed to the end flange of robot 5 using a clamp. After the polishing process, the polyurethane polishing pad is worn, which affects the polishing quality. Therefore, laser line scanner 3 is installed on worktable 4 for inspection.

[0027] 2. Start testing

[0028] The CNC program for the inspection process is input into the electrical control cabinet 7. The robot 5 moves the polishing pad to the identifiable area directly above the laser line scanner, and then begins a raster-like movement to inspect the entire surface of the polishing pad. The laser line scanner 3 transmits the surface morphology data of the polishing pad to the computer 6 for further data processing. During the splicing scanning process, it is necessary to ensure that there are overlapping areas at the seams, with the overlapping area accounting for 20% to 40%.

[0029] 3. Point cloud data filtering

[0030] When acquiring point cloud data, due to the influence of equipment accuracy, environmental factors, and other factors, some noise points and discrete points far from the main point cloud, i.e., outliers, will inevitably appear. These outliers and noise points will greatly affect the accuracy of point cloud registration algorithms. Therefore, in order to improve the accuracy and efficiency of the registration algorithm, filtering is performed in the preprocessing stage to remove outliers. Voxel filtering reduces the amount of point cloud data while maintaining the shape characteristics of the point cloud and does not destroy the geometric structure of the point cloud itself. Radius filtering has a good effect on removing some suspended isolated points or invalid points in the original laser point cloud data. Considering both the effect and efficiency of noise reduction, voxel filtering and radius filtering are used for preprocessing.

[0031] 4. Data splicing and registration

[0032] (1) Select three obvious feature points a, b, and c, measure the distances between the three points in adjacent point cloud segments, and thus determine the translation transformation matrix to complete the preliminary estimation of the corresponding relationship; (2) ICP algorithm registration based on K-D nearest neighbor search: Use the point cloud segment shown in Figure 4 (b) as the target point cloud, Figure 4 (a) and Figure 4 (c) to construct a two-way K-dimensional tree for the source point cloud. Traverse the source point cloud P and the target point cloud Q after initial registration as the initial point set for ICP fine registration, and find the optimal rigid body transformation through the classical ICP algorithm to achieve point cloud fine registration. Repeat the above process for three adjacent point clouds, and then splice all the source point clouds onto the target point cloud and perform merging and fusion to construct the complete three-dimensional surface contour of the airbag tool. If the registration error of the point cloud data is less than or equal to the positioning error of the data acquisition platform, it is considered qualified and the next step can be executed; if it is unqualified, re-detection is required.

[0033] 5. Extract the contour line

[0034] (1) Obtain the point cloud data through the data acquisition platform, select the X-axis coordinate and Y-axis coordinate in the three-dimensional point cloud data as the index of the Z-axis coordinate, construct a matrix, and represent the three-dimensional point cloud data in the form of a matrix;

[0035] (2) Calculate the number of rows and columns of the matrix, represent the number of rows by m and the number of columns by n, and find the median of m and n. If m and n are even, the median is m / 2, n / 2; if m and n are odd, the median is (m + 1) / 2, (n + 1) / 2;

[0036] (3) Extract the contour using the median respectively. The Z-axis coordinate value of the X-direction contour is represented by xprofile(i), where:

[0037] xprofile(i) = Z(m / 2, i) (i = 1, 2, 3....n) (1)

[0038] The Z-axis coordinate value of the Y-direction contour is represented by yprofile(i), where:

[0039] yprofile(i) = Z(i, n / 2) (i = 1, 2, 3....m) (2)

[0040] 6. Calculate the wear depth of the polishing pad

[0041] (1) Construct a theoretical contour through the general equation of a two-dimensional circle, the circle radius R, and the center coordinates (a, b). The radius of the circle determines the size of the circle, and the center determines the position of the circle. Therefore, it is necessary to ensure that R = 40 is consistent with the radius of the actual contour. The general equation of the circle is as follows:

[0042] (xa) 2 +(yb) 2 =R 2 (3)

[0043] Equation (3) can be transformed into:

[0044] y = (-x) 2 +2*a*xa 2 +R 2 ) 0.5 +b (4)

[0045] The theoretical profile in the X direction can be represented as:

[0046] x(i)=(-x(i) 2 +2*a1*x(i)-a1 2 +R 2 ) 0.5 +b1 (i=1、2、3....n) (5)

[0047] The theoretical profile in the Y direction can be represented as:

[0048] y(i)=(-x(i) 2 +2*a2*x(i)-a2 2 +R 2 ) 0.5 +b2 (i=1、2、3....m) (6)

[0049] (2) Calculate the difference between the actual and theoretical contours of the wear area on the Z-axis, and take the average of all differences within the wear area as the wear depth. The calculation process is as follows:

[0050] The wear depth in the X direction can be expressed as:

[0051] Z_x=x(i)-xprofile(i) (i is the x-axis coordinate of the wear area) (7)

[0052] The wear depth in the Y direction can be expressed as:

[0053] Z_y=y(i)-yprofile(i) (i is the x-axis coordinate of the wear area) (8)

[0054] 7. Predict the current removal function

[0055] The removal function for airbag polishing is dz = k·P·V·dt, where dz is the removal depth, k is a proportionality constant, P is the pressure, V is the velocity, and dt is the polishing time. Due to polishing pad wear, the same polishing process will produce different polishing effects. Therefore, the constant k in the removal function needs to be corrected for different polishing pad wear depths. After correcting the value of k, the removal function for the current polishing pad can be predicted.

[0056] The embodiments described above are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A prediction method for the removal function of a flexible airbag tool based on polishing pad wear, characterized in that, Includes the following steps: 1) After the airbag finishes one polishing process, move the airbag to the laser line scanner area for scanning. When performing splicing scanning, it is necessary to ensure that there are overlapping areas at the seams. 2) After obtaining point cloud data through scanning in step 1), the point cloud data is preprocessed to reduce noise, outliers are removed, and data is stitched and registered. 3) After obtaining the surface data of the airbag in step 2), extract the cross-sectional profile of the airbag through the center of the sphere, calculate the center and radius of the fitted ideal circle according to the least squares method, obtain the ideal circular profile, and make the ideal circular profile coincide with the actual airbag cross-sectional profile. Take the difference between the ideal profile and the actual profile in the advance angle direction, which is the wear depth. 4) Based on the detected wear depth of the polishing pad, compare it with the polishing pad's full lifespan state database to determine the current material removal capacity of the polishing pad, predict the removal function of the current state, and compensate for the residence time.

2. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 1, characterized in that: In step 1), the proportion of the repeated region is 20% to 40%.

3. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 1, characterized in that: In step 2), the point cloud data is preprocessed for noise reduction using voxel filtering and radius filtering methods.

4. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 1, characterized in that: In step 2), when scanning large workpieces, it is necessary to stitch together the scan data. The stitching and registration relationship between multiple surface data is calculated based on the correspondence of the data acquisition platform.

5. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 4, characterized in that, The calculation steps are as follows: obtain the translation amount of the point cloud coordinate axes based on the translation distance of the data acquisition platform; Select three prominent feature points; measure the distance between the three prominent feature points in adjacent point cloud segments to determine the translation transformation matrix; complete the data stitching and registration of the corresponding relationships.

6. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 5, characterized in that, After completing the data splicing and registration of the corresponding relationships, the process also includes fine registration of point clouds: ICP algorithm registration based on KD nearest neighbor search. Using point cloud fragments as target point clouds, a bidirectional K-dimensional tree is constructed. The source point cloud P and target point cloud Q, which have undergone initial registration, are traversed as the initial point set for fine registration of the ICP algorithm. The optimal rigid body transformation is obtained through the classic ICP algorithm to achieve fine registration of point clouds.

7. The prediction method for the removal function of flexible airbag tools based on polishing pad wear as described in claim 1, characterized in that: In step 4), the time term in the removal function of the Princeton equation is compensated, and the residence time of the residence point is modified to keep the removal capacity consistent.

Images