Method for automatic polishing of inner wall of runner chamber of water turbine based on three-dimensional measurement

Abstract

The present application belongs to the field of in-situ robot automatic polishing of hydraulic turbine, and specifically provides a method for in-situ robot automatic polishing of inner wall of runner chamber of hydraulic turbine based on three-dimensional measurement, comprising: setting the circumferential width of a single polishing area, dividing the inner wall surface into multiple polishing areas, arranging the identification bodies at equal intervals, and calculating the position and posture of the moving trolley and the mechanical arm at each area; controlling the moving trolley and the mechanical arm to scan and measure a single area, performing simplification processing on the measurement data, extracting the geometric features of the identification bodies, and establishing a coordinate system at the identification bodies; calculating and generating a polishing track according to the measurement point cloud data; repeating the above operation to obtain the polishing tracks of all polishing areas, then converting the coordinate systems of the identification bodies in the areas to each other, splicing the tracks of two adjacent polishing areas, and polishing according to the spliced polishing track by the robot. The method realizes polishing and repair work on a large area of the inner wall of the hydraulic turbine, and improves the maintenance efficiency of the hydraulic turbine.

Description

Technical Field

[0001] This invention belongs to the field of in-situ robotic automatic grinding and processing of water turbines, specifically, it relates to an in-situ robotic automatic grinding method for the inner wall of a water turbine runner based on three-dimensional measurement. Background Technology

[0002] As a core component of hydroelectric power generation equipment, the turbine is subjected to the impact, wear, and corrosion of water flow during long-term operation. This can lead to cracks or damage on the blades or the inner wall of the runner, thereby affecting the turbine's power generation efficiency and operational stability. Traditional maintenance methods usually require shutting down the turbine and then manually inspecting and repairing the problem. This method is not only time-consuming and labor-intensive, but may also affect the stable operation of the generator unit.

[0003] To improve the efficiency and quality of turbine maintenance and reduce maintenance costs, automated robotic systems have been introduced for the periodic inspection, maintenance, and repair of turbine runners. This automated grinding method utilizes a robotic system operating inside the turbine to precisely locate surface damage and cracks and perform efficient grinding and repair. By installing automated equipment within the turbine, the robot can move sequentially to each grinding area along a pre-set path. This automated grinding method requires the ability to generate grinding trajectories to perform automatic grinding; currently, no automated grinding method using an automated robotic system has been disclosed in the existing technology. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an in-situ robotic automatic grinding method for the inner wall of a water turbine runner based on three-dimensional measurement, so as to realize the grinding and repair operation of a large area of ​​the inner wall of the water turbine and improve the maintenance efficiency of the water turbine.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an in-situ robotic automatic grinding method for the inner wall of a water turbine runner based on three-dimensional measurement, comprising the following steps:

[0006] S1. Design multiple markers and distribute them evenly along the circumference of the inner wall of the turbine runner. Based on the working radius of the robotic arm, set the circumference width of a single grinding area, divide the inner wall surface into multiple grinding areas, arrange the markers at equal intervals, and calculate the position and posture of the moving trolley and robotic arm in each area.

[0007] S2. Control the mobile trolley and robotic arm to move to the designated position in sequence, perform single-area scanning measurement, simplify the measurement data, extract the geometric features of the markers, and establish a coordinate system at each marker.

[0008] S3. Calculate and generate the grinding trajectory based on the measurement point cloud data of a single area on the inner wall;

[0009] S4. Repeat steps S2 and S3 to obtain the grinding trajectory of all grinding areas. Then, according to the mutual transformation of the coordinate system of the marker body in the area, the trajectories of two adjacent grinding areas are spliced ​​together, and the robot grinds according to the spliced ​​grinding trajectory.

[0010] In a preferred embodiment, step S1 includes the following steps:

[0011] S11, Place three standard spheres of different diameters {S small S medi S large They are installed on the same base to form a single identifier I;

[0012] S12. Determine the circumferential area of ​​the turbine runner chamber that needs to be ground. Starting from a fixed starting point on the circumference, arrange markers at equal intervals to divide the grinding area {A1, A2, ..., A...}. n}, while ensuring that each polishing area has two markers, one on the left and one on the right {I}. left I right};

[0013] S12. Design the spacing between markers based on the working radius of the robotic arm. To ensure the position of the trolley on the circular guide rail is maintained during a single movement, the markers on both sides of the area must be covered during measurement. left I right};

[0014] S13. The three-dimensional measuring instrument is installed at the end of the robotic arm, which is fixed on a moving trolley platform that moves along the circumferential guide rail. For a single grinding area, the center position of the moving trolley platform is designed to be located on the vertical plane of the center of the arc-shaped area.

[0015] S14. Using robot virtual simulation software, teach the robotic arm's measurement posture to ensure that the robotic arm can complete the scanning and measurement of the markers on both sides while the moving trolley remains stationary, and obtain the measurement data for each polishing area {A}. i |i=1~n} and the measured position and orientation of the moving trolley and robotic arm.

[0016] In a preferred embodiment, step S2 involves simplifying the measurement data, extracting the geometric features of the markers, and establishing a coordinate system at each marker, comprising the following steps:

[0017] S21. Perform hand-eye calibration on the 3D measuring instrument installed at the end of the robotic arm. Using the hand-eye matrix and coordinate transformation, obtain point cloud data in the robot base coordinate system after each measurement. Remove noise interference points and filter the point cloud on the scanned measurement data to obtain simplified single-region point cloud data.

[0018] S22. In the CAD model of the sign, using a standard sphere S large The center of the ball Let be the origin of the coordinate system, with Let X be the direction of the coordinate system X-axis, with Construct the theoretical coordinate system of the marker body along the Y-axis of the coordinate system;

[0019] S23. Using the theoretical CAD model of the marker as a reference model, extract the feature point cloud of all markers in the measurement data through the template matching algorithm, and calculate the transformation matrix of the actual marker in the machine base coordinate system.

[0020] S24. For the initial grinding circumferential area A1, the measurement data only includes one right-side marker feature and its corresponding coordinate system. For other regions {A} i |i=2~n}, the measurement data includes the features of the left and right side markers and their corresponding coordinate systems.

[0021] In a preferred embodiment, step S3 includes the following steps:

[0022] S31. For each polishing area {A i |i=1~n}, all are based on the coordinate system of the right-hand marker within the region. Calculate the grinding trajectory within a single area;

[0023] S32. Based on the reference coordinate system of the right-side marker within a single area. And the length and width information of the area, extracting point cloud data within the area. And set the Y-axis of the reference coordinate system as the tool feed direction;

[0024] S33, using point cloud data Construct a kd-tree index for the surface point cloud, determine the range of points to be taken, traverse the point cloud to find the initial point, search for nearby points, take the centroids of multiple points to obtain control points, and re-solve the point cloud normal based on the control points;

[0025] S34. Determine the tool tilt angle and grinding disc radius based on the grinding process parameters. Solve the tool axis vector based on the tool tip position and normal. Using the tool axis vector as the Z-axis and the feed direction of the trajectory points as the X-axis, generate the grinding posture at the tool tip and calculate multiple grinding trajectories within a single area. The trajectory points are expressed in matrix form as follows:

[0026]

[0027] Where i represents the polishing trajectory of the i-th region, j represents the number of trajectory segments in the region, and k represents the number of trajectory points contained in a single polishing trajectory.

[0028] In the preferred embodiment, step S4 involves stitching together the trajectories of two adjacent grinding areas based on the mutual transformation of the coordinate systems of the marker bodies within the area. This includes the following steps:

[0029] S41. For the measurement data of each area, according to step S2, obtain the coordinate systems of the left and right markers. And calculate the transformation matrix.

[0030] S42, in the initial polishing area A1, with Using the coordinate system as a reference, following step S3, the grinding trajectory points within this area are generated and represented in matrix form as follows:

[0031]

[0032] Where i represents the polishing trajectory of the i-th region, j1 represents the number of trajectory segments within the polishing region A1, and k represents the number of trajectory points contained in a single polishing trajectory.

[0033] S43. For adjacent grinding areas {A i |i=2~n-1}, the measurement data within this region includes two marker coordinate systems, one on the left and one on the right. The sign on the left is the same sign as the one on the right in the previous area; therefore... Through coordinate transformation matrix Transform the endpoint of each trajectory segment in region i-1 to region i, that is:

[0034]

[0035] Using the endpoint of each trajectory segment in the (i-1)th region as the starting point of each trajectory segment in the i-th region, the grinding trajectory points in that region are generated according to step S3.

[0036] S44. For the final polishing area A n The left marker in this area is the same as the right marker in the (n-1)th area. Therefore, the final grinding trajectory is calculated by taking the end point of each trajectory segment in the (n-1)th area as the starting point of each trajectory segment in the nth area, and taking the starting point of each trajectory segment in the 1st area as the ending point of each trajectory segment in the nth area. The expression is:

[0037]

[0038] S45. Following steps S41 to S45, the robot automatically repeats the same processing procedure by splicing the upper and lower circumferential path points along the inner wall of the turbine runner chamber to complete the grinding work of multiple circumferential curved surfaces.

[0039] The present invention provides an in-situ robotic automatic grinding method for the inner wall of a water turbine runner based on three-dimensional measurement, which has the following beneficial effects:

[0040] 1. It has enabled automated grinding and repair of large areas of the inner wall of the turbine runner, significantly improving the maintenance efficiency of the turbine.

[0041] 2. Through the design of the marker body and coordinate system, the precise positioning and trajectory splicing of multiple grinding areas were achieved, ensuring the continuity and uniformity of grinding.

[0042] 3. High degree of automation reduces manual intervention, lowers labor intensity, and improves operational safety. It can adapt to the complex curved surface of the turbine runner chamber, achieving comprehensive grinding operations.

[0043] 4. By using digitalization and automation methods, the accuracy and consistency of turbine maintenance have been improved, maintenance downtime has been reduced, turbine utilization has been increased, and more stable equipment support has been provided for hydropower generation. Attached Figure Description

[0044] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0045] Figure 1 This is a flowchart of the present invention;

[0046] Figure 2 A schematic diagram of the logo;

[0047] Figure 3 Layout diagram of the signage;

[0048] Figure 4 This is a map showing the positions of the mobile vehicle and robot.

[0049] Figure 5 This serves as the theoretical identifier model and reference coordinate system.

[0050] Figure 6 Simulation diagram for identifier feature template matching and feature extraction;

[0051] Figure 7 A diagram showing the relationship between the regional coordinate system;

[0052] Figure 8 This is a schematic diagram of the grinding trajectory for a single area;

[0053] Figure 9 A path point distribution map;

[0054] Figure 10 A mosaic of the trajectory of adjacent regions;

[0055] Figure 11 A schematic diagram of the grinding trajectory of a single area for the robot;

[0056] Figure 12 A schematic diagram of the trajectory stitching of point clouds from adjacent processing areas. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0058] like Figures 1-12 As shown, an in-situ robotic automatic grinding method for the inner wall of a turbine runner based on three-dimensional measurement is applied to the repair and processing of the inner wall of a turbine runner, and includes the following steps:

[0059] S1. Design multiple markers, evenly distributed along the circumference of the turbine runner chamber wall. Based on the working radius of the robotic arm, set the circumference width of a single grinding area, divide the inner wall surface into multiple grinding areas, arrange the markers at equal intervals, and calculate the position and orientation of the moving trolley and robotic arm in each area. Specifically:

[0060] S11, such as Figure 2 As shown, three standard spheres {S} with different diameters small S me S large They are installed on the same base to form a sign body I.

[0061] S12, such as Figure 3 As shown, the circumferential area of ​​the turbine runner chamber that needs to be ground is determined. Starting from a fixed starting point on the circumference, markers are arranged at equal intervals to divide the grinding area {A1, A2, ..., A...}. n}, while ensuring that each area has two identifiers, one on the left and one on the right {I}. left I right}

[0062] S12. Design a reasonable spacing between markers based on the working radius of the robotic arm. To ensure the position of the trolley on the circular guide rail is maintained during a single movement, the markers on both sides of the area must be covered during measurement. left I right}

[0063] In this embodiment, for the Siasun GCR20-1100 robot, its working radius is 1100mm. Based on the robot's working height and the radius of the sphere space, the spacing between the markers is designed to be approximately L. I =800mm.

[0064] S13, such as Figure 4 As shown, a 3D measuring instrument is mounted at the end of a robotic arm, while the base of the robotic arm is fixed to a mobile trolley platform that can move along a circular guide rail. To ensure that the 3D measuring instrument can measure the markers on the left and right sides, for a single grinding area, the center of the mobile trolley platform is designed to be located on the vertical plane at the center of the arc-shaped area.

[0065] S14. Using robot virtual simulation software, teach the robotic arm's measurement posture and simulate its accessibility to ensure that the robotic arm can complete the scanning and measurement of the markers on both sides while the moving cart remains stationary. This allows for the acquisition of the measurement data for each polishing area {A}. i |i=1~n}, the measurement position and orientation of the moving trolley and the robotic arm.

[0066] S2. Control the mobile trolley and robotic arm to move sequentially to the designated positions and automatically perform single-area scanning measurements. Then, simplify the measurement data, extract the geometric features of the markers, and establish a coordinate system for each marker. Specifically:

[0067] S21. Perform hand-eye calibration on the 3D measuring instrument mounted at the end of the robotic arm. Using the hand-eye matrix and coordinate transformation, point cloud data in the robot's base coordinate system can be obtained after each measurement. Noise interference points are removed and point cloud filtering is performed on the scanned measurement data to obtain simplified single-region point cloud data.

[0068] In this embodiment, statistical outlier filtering is used to remove noise points, followed by voxel filtering for downsampling to reduce the amount of data. Next, bilateral filtering is used to smooth the point cloud while preserving edge features. Finally, PCA is used to estimate the point cloud normal vectors, and moving least squares is used for optimization to improve the accuracy of subsequent feature extraction and trajectory generation.

[0069] S22, such as Figure 5 As shown, in the CAD model of the marker, a standard sphere S... large The center of the ball Let be the origin of the coordinate system, with Let X be the direction of the coordinate system X-axis, with Construct the theoretical coordinate system of the marker body along the Y-axis.

[0070] S23, such as Figure 6As shown, the theoretical CAD model of the marker is used as a reference model. The feature point cloud of all markers in the measurement data is extracted by template matching algorithm, and the transformation matrix of the actual marker in the machine base coordinate system is calculated.

[0071] The template matching algorithm used in this embodiment can be the Iterative Closest Point (ICP) algorithm or an improved version based on the Iterative Closest Point algorithm, which is an existing technology widely used in point cloud registration.

[0072] S24, such as Figure 7 As shown, for each polishing circumferential region A i The measurement data includes the features of the left and right markers and their corresponding coordinate systems.

[0073] S3. Based on the measured point cloud data of a single area on the inner wall, calculate and generate the grinding trajectory. Specifically:

[0074] S31, such as Figure 8 As shown, for each polishing area {A} i |i=1~n}, all are based on the coordinate system of the right-hand marker within the region. To calculate the grinding trajectory within a single area.

[0075] S32. Based on the reference coordinate system of the right-side marker within a single area. And the length and width information of the area, extracting point cloud data within the area. And set the Y-axis of the reference coordinate system as the tool feed direction.

[0076] S33, such as Figure 9 As shown, point cloud data Construct a kd-tree index for the surface point cloud, determine the range of points to be selected, traverse the point cloud to find the initial point, search for nearby points, obtain control points by taking the centroids of multiple points, and re-solve the point cloud normal based on the control points.

[0077] S34. Determine the tool tilt angle and grinding disc radius based on the grinding process parameters. Solve the tool axis vector based on the tool tip position and normal. Using the tool axis vector as the Z-axis and the feed direction of the trajectory points as the X-axis, generate the grinding posture at the tool tip. From this, calculate multiple grinding trajectories within a single area. The trajectory points are expressed in matrix form.

[0078]

[0079] Where i represents the polishing trajectory of the i-th region, j represents the number of trajectory segments in the region, and k represents the number of trajectory points contained in a single polishing trajectory.

[0080] S4. Repeat steps S2 and S3, and according to the mutual transformation of the coordinate systems of the markers within the region, stitch together the trajectories of two adjacent grinding areas, such as... Figure 10 As shown. Specifically:

[0081] S41. For the measurement data of each area, first obtain the coordinate systems of the left and right markers according to the method in step S23. And calculate the transformation matrix.

[0082] S42. In the initial polishing circumferential area A1, with Using this coordinate system as a reference, and following steps S31 to S34, the grinding trajectory points within this area are generated and represented in matrix form as follows:

[0083]

[0084] S43. For adjacent grinding areas {A i |i=2~n-1}, the measurement data within this region includes two marker coordinate systems, one on the left and one on the right. The sign on the left is the same sign as the one on the right in the previous area; therefore... Through coordinate transformation matrix The endpoint of each trajectory segment in region i-1 can be transferred to region i, that is:

[0085]

[0086] Using the endpoint of each trajectory segment in region i-1 as the starting point of each trajectory segment in region i, and following steps S31-S34, the grinding trajectory points within that region are generated. The grinding trajectory for a single region is as follows: Figure 11 As shown.

[0087] S44. For the final polishing area A n The left marker in this area is the same as the right marker in the (n-1)th area. Therefore, referring to step S43, the final polishing trajectory is calculated by taking the end point of each trajectory segment in the (n-1)th area as the starting point of each trajectory segment in the nth area, and taking the starting point of each trajectory segment in the first area as the end point of each trajectory segment in the nth area. That is:

[0088]

[0089] S45. Following the steps above, the components are joined at the path points along the upper and lower circumference of the inner wall of the turbine runner chamber, as shown below. Figure 12 As shown, the robot can automatically repeat the same processing steps to complete the grinding work of multiple circumferential curved surfaces.

[0090] This invention provides an in-situ robotic automated grinding method for the inner wall of a turbine runner based on three-dimensional measurement. By utilizing three-dimensional measurement technology, it can efficiently and accurately acquire surface information of the inner wall, significantly reducing manual measurement costs. This method not only improves measurement efficiency but also provides a reliable data foundation for subsequent automated grinding operations. By precisely controlling the grinding trajectory and parameters, this method can effectively repair damage to the inner wall surface, improve surface quality, and simultaneously avoid unnecessary damage to the turbine structure. This automated grinding method significantly improves the efficiency and quality of turbine maintenance.

[0091] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatic polishing of the inner wall of a runner chamber of a hydraulic turbine based on in-situ robot measurement in three dimensions, characterized by: The method comprises the following steps: S1, a plurality of markers are designed to be evenly distributed along the circumferential direction of the inner wall of the runner chamber, the circumferential width of a single polishing area is set according to the working radius of the mechanical arm, the inner wall surface is divided into a plurality of polishing areas, the markers are arranged at equal intervals, and the positions and postures of the moving trolley and the mechanical arm at each area are calculated; S2, the moving trolley and the mechanical arm are controlled to move to the specified positions in turn, single-area scanning measurement is performed, the measurement data is simplified, the geometric features of the markers are extracted, and a coordinate system is established at each marker; S3, the polishing track is calculated and generated according to the single-area measurement point cloud data of the inner wall; S4, steps S2 and S3 are repeatedly performed to obtain the polishing tracks of all polishing areas, then the tracks of two adjacent polishing areas are spliced according to the mutual conversion of the marker coordinate systems in the areas, and the robot polishes according to the spliced polishing track; The mutual conversion of the marker coordinate systems in the areas is used to splice the tracks of two adjacent polishing areas, which comprises the following steps: S41. For the measurement data of each area, according to step S2, obtain the coordinate systems of the left and right markers. And calculate the transformation matrix. ; S42、in the starting polishing area , in With the coordinate system as the reference datum, the polishing track points in the area are generated according to step S3, which is expressed in the matrix form as follows: ； wherein, represents the polishing trajectory of a block region, 1 represents a polishing region number of inner trajectory segments, represents the number of trajectory points contained in a single polishing trajectory; S43, for the abutting polishing area , the measurement data in the area contains two identification body coordinate systems of left and right sides , wherein the left identification body is the same as the right identification body of the last area, therefore, ; through the coordinate conversion matrix , the i -1 area each track end is converted to area, namely: ； at the end of each trajectory in the region at the end of each trajectory in the region at the start of each trajectory in the region, the grinding trajectory points in the region are generated according to step S3; S44, for the final polishing area the left side marker in the area is the same marker as the right side marker in the n -1 area, so the polishing track of the final area is calculated with the end point of each track in the n -1 area as the start point of each track in the -1 area, and with the start point of each track in the -1 area as the end point of each track in the -1 area, and the expression is: ； S45, according to steps S41-S45, the robot automatically repeats the same processing procedure through the splicing of the upper and lower area path points along the circumference of the inner wall of the runner chamber, and the polishing work of the plurality of circumferential surfaces is completed.

2. The method according to claim 1, wherein, In the step S1, the following steps are included: S11. Three standard balls of different diameters are placed on a horizontal surface mounted on the same base, forming an identification body ; S12, determine the circumferential area of the inner wall of the runner chamber that needs to be polished, arrange the identification bodies one by one in an equidistant manner from a fixed starting point of the circumference to divide the polishing area , and ensure that each polishing area has left and right identification bodies ; S12, according to the working radius of the mechanical arm, design the distance between the identification bodies , to ensure that when the position of the trolley on the circular guide rail is moved once, the identification bodies on both sides of the area can be covered in the measurement ; S13, the three-dimensional measuring instrument is installed at the end of the mechanical arm, and the mechanical arm is fixed on the moving trolley platform moving along the circumferential guide rail, and the center position of the moving trolley platform is located on the vertical plane at the center of the arc-shaped area for a single polishing area; S14. Using robot virtual simulation software, teach the robotic arm's measurement posture to ensure that the robotic arm can complete the scanning and measurement of the markers on both sides while the moving trolley remains stationary, and obtain the measurement data for each polishing area. And the measurement position and orientation of the mobile cart and robotic arm.

3. The method according to claim 1, wherein, In the step S2, the measurement data is simplified, the geometric features of the markers are extracted, and a coordinate system is established at each marker, which comprises the following steps: S21, the three-dimensional measuring instrument installed at the end of the mechanical arm is hand-eye calibrated, the point cloud data in the robot base coordinate system is obtained after each measurement by using the hand-eye matrix and coordinate transformation, the noise interference points in the scanning measurement data are removed, and the point cloud is filtered to obtain the simplified single-area point cloud data; S22, in the CAD model of the identification body, a standard sphere with the sphere center as the coordinate system origin, with as the coordinate system X-axis direction, and with as the coordinate system Y-axis direction, a theoretical coordinate system of the identification body is constructed; S23, the theoretical CAD model of the marker is used as a reference model, the feature point cloud of all markers in the measurement data is extracted by using a template matching algorithm, and the transformation matrix of the actual marker in the robot base coordinate system is calculated; S24, for the starting polishing circumferential region , the measurement data only contains one right side identification body feature and the corresponding coordinate system ; for other regions , the measurement data contains left and right side identification body features and the corresponding coordinate system is .

4. The method according to claim 1, wherein, In the step S3, the following steps are included: S31、For each polishing area , taking the right side marker coordinate system in the area as the reference coordinate system , calculating the polishing track in the single area; S32, according to the single area right side identification body reference coordinate system And the area length-width information, intercept the point cloud data in the area And set the Y axis of the reference coordinate system as the tool feeding direction; S33, with point cloud data Constructing curved surface point cloud kd-tree index, determining point range, finding initial point by traversing point cloud, searching for nearby point, getting control point by taking multi-point centroid, and re-solving point cloud normal based on control point; S34, the tool inclination angle and the grinding disc radius are determined according to the polishing process parameters, the tool shaft vector is solved based on the tool tip point position and the normal line, the tool shaft vector is taken as the Z axis, the track point feeding direction is taken as the X axis, the polishing posture at the tool tip point is generated, and a plurality of polishing tracks in the single area are calculated, and the track points are expressed in the form of a matrix as follows: ； wherein, represents the number of polishing tracks of the block area, represents the number of track segments within the area, represents the number of track points contained in a single polishing track.

Images