Servomotor cylinder body inner wall scratch robot self-adaptive additive and subtractive material repairing method based on three-dimensional measurement

Abstract

A servomotor cylinder inner wall scratch robot self-adaptive additive and subtractive material repairing method based on three-dimensional measurement is characterized in that a six-axis mechanical arm is integrated with a line laser sensor and milling, laser cladding and constant-force grinding and polishing tools, and automatic repairing is achieved. A regular groove body is milled through a slicing method and a rectangular-ambulatory-plane strategy; a workpiece coordinate system is built at the bottom of the groove, a bow-shaped layered alternate path is adopted for filling materials, and the laser power and the powder feeding rate are adjusted by measuring, matching and correcting the path according to the characteristics of the groove; and finally, a surplus height boundary is scanned and extracted, a rough grinding path is generated with the scratch trend as the benchmark, fine polishing is conducted through a reciprocating zigzag strategy, parameters are dynamically adjusted in combination with closed-loop feedback, and smooth transition between a cladding layer and the original surface is achieved. The method solves the problems of long construction period and low precision of traditional repair, integrates multiple processes and high-precision measurement, improves the repair efficiency and quality, and is suitable for on-site rapid maintenance of hydroelectric equipment.

Description

Technical Field

[0001] The present invention belongs to the technical field of hydroelectric power station equipment maintenance, and particularly relates to a method for adaptively adding and subtracting materials for repairing scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement by a robot. Background Technique

[0002] As a key control component of a hydro-generating unit, the servomotor is horizontally and symmetrically arranged under the movable guide vanes of the water turbine, responsible for receiving the hydraulic signal of the speed regulation system and converting it into piston displacement to control the operation of the guide vane mechanism. However, due to its horizontal layout characteristics, scratches are likely to occur on the piston and the bottom of the cylinder of the servomotor due to hard foreign objects, resulting in excessive oil leakage in the opening and closing chambers, making it impossible for the movable guide vanes of the unit to maintain a fixed opening degree, or even completely losing control, seriously affecting the safe operation of the equipment.

[0003] Traditional methods for repairing scratches on the servomotor mainly rely on returning to the factory for repair by welding and boring / turning, and there are many significant defects. First, the construction period is long, and usually, the repair of two sets of servomotors cannot be completed within a ten-day maintenance period. Second, the repair quality is difficult to guarantee. Traditional welding is prone to defects such as porosity, slag inclusion, and lack of fusion, and the surface roughness and geometric accuracy after repair often cannot meet the sealing requirements. Finally, there are no precise measurement and evaluation means in the repair process, relying on the experience of workers, resulting in poor consistency in the repair of different batches. The technical bottlenecks such as the inability of traditional repair methods to accurately restore the original surface topography often make the service life and reliability of the repaired servomotor difficult to achieve the expected effect.

[0004] To solve the above technical problems, the hydroelectric power industry urgently needs a servomotor scratch repair technology that can be implemented on-site, is efficient and accurate, and does not change the original dimensions. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for adaptively adding and subtracting materials for repairing scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement by a robot. By integrating technologies such as high-precision three-dimensional measurement, regularized milling, additive filling, and adaptive polishing, rapid in-situ repair of the scratched area is realized, which not only ensures the repair quality and dimensional accuracy but also significantly shortens the repair cycle, providing strong technical support for the safe and stable operation of hydroelectric equipment.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is as follows: A method for adaptively adding and subtracting materials for repairing scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement by a robot, the steps are as follows: S1. Integration of multi-process full-process repair equipment: Install a line laser sensor and a tool quick-change device at the end of a six-axis robotic arm, and integrate a milling tool, a laser cladding tool, and a constant-force polishing tool through a quick-change interface to realize the full-process repair processing of high-precision scanning measurement, milling regularization, material filling, and surface polishing of scratches on the inner wall of the servomotor cylinder. S2. Regularization treatment of the scratch surface: Obtain the point cloud data of the scratch area through a line laser sensor. Through feature point definition and geometric modeling, construct a grooving geometric model with the scratch direction as the benchmark, a trapezoidal cross-section, the original surface at the top, and a plane at the bottom. Use the slicing method and the zigzag pattern strategy to plan the milling path, and perform the milling operation to form a regularized groove.

[0007] S3. Material filling in the regularized area: Establish a workpiece coordinate system with the bottom of the groove as the origin. Use the bow-shaped layer-by-layer alternating method to plan the additive manufacturing path. After measuring the milled groove with a line laser and matching it with the theoretical model, correct the position of the additive manufacturing path. Adjust the laser power and powder feeding rate according to the groove characteristics, and implement precise material filling.

[0008] S4. Grinding and polishing treatment of the clad layer surplus height: Use a line laser sensor to scan the filled area, extract the surplus height contour boundary through cross-section analysis, generate a rough grinding path with the scratch direction as the benchmark, and generate a fine polishing path using the reciprocating zigzag strategy with boundary constraints. Dynamically match the grinding and polishing parameters according to the maximum surplus data, and achieve a smooth transition between the clad layer and the original surface through the cycle of measurement and fine polishing.

[0009] Preferably, the multi-process full-process repair equipment in step S1 specifically includes: S11. Install a line laser sensor and a tool quick-change device at the end flange of the six-axis robotic arm; milling tools, laser cladding tools, and constant-force grinding and polishing tools are all equipped with corresponding quick-change interfaces, enabling the robotic arm to achieve multi-process integrated processing through quick tool change.

[0010] S12. The line laser sensor is used to perform high-precision scanning and measurement on the scratch area on the inner wall of the relay cylinder, obtain the surface topography data, and provide basic data for the precise positioning of the repair area and the subsequent detection of the clad layer surplus height.

[0011] S13. The milling tool consists of a high-speed electric spindle and a milling cutter, and is used to mill the scratch area to form a regularized groove; the laser cladding tool consists of a laser and a cladding nozzle, and uses laser cladding technology to fill the material in the regularized groove area; the constant-force grinding and polishing tool consists of a force control unit, a motor, a grinding wheel, and a polishing wheel, and is used for grinding and polishing the clad layer surplus height to remove the excess material and restore the original surface topography of the workpiece.

[0012] Preferably, the scratch surface regularization treatment in step S2 includes: S21. Control the robotic arm to carry the line laser sensor to scan and measure the area around the scratch, obtain the surface point cloud data of the area, and use the point cloud processing algorithm to filter and reconstruct the data to generate an accurate surface mesh model.

[0013] S22. Manually select scratch feature points on the surface mesh model, and use the spline fitting method to calculate the feature curve representing the scratch trend.

[0014] S23. Construct a repair geometry based on the feature curve. This geometry takes the scratch trend spline curve as the benchmark, uses a trapezoidal cross-section, and adjusts the trapezoid height, upper base width, and sidewall angle according to the actual scratch situation; through the intersection operation of this geometry and the S21 surface mesh model, a regularized grooving geometry model with the original surface at the top, a plane at the bottom, and a trapezoidal cross-section is formed. S24. Plan the milling path with the grooving geometry model as the theoretical model. The path generation method adopts the slicing method and the zigzag pattern strategy.

[0015] First, discretize the three-dimensional grooving geometry model in the bottom plane direction to form a series of two-dimensional contours; then, adopt the zigzag pattern strategy, start from the contour boundary, and gradually calculate the offset inward along the contour shape to generate an equidistant contour sequence that shrinks inward. Finally, to ensure the machining quality, apply speed smoothing control at the corners to avoid sudden acceleration changes caused by sharp turns. At the same time, when connecting adjacent contours, optimize the tool lift and tool entry point selection to reduce the idle travel. Finally, control the robotic arm to perform the milling operation to achieve the regularization treatment of the scratch area.

[0016] Preferably, the material filling in the regularization area in step S3 includes: S31. Establish a theoretical workpiece coordinate system based on the edge contour and feature points of the bottom plane of the groove body according to the grooving geometry model.

[0017] S32. Plan the additive path of the robotic arm using the bow-shaped layer-by-layer alternating method.

[0018] This method first performs equal-height slicing of the three-dimensional model in the Z-axis direction to form a series of two-dimensional contours on parallel planes. The key of the bow-shaped trajectory lies in its unique "bow" - shaped motion mode: in the odd layers, it uses bow-shaped filling along the X-axis direction; while in the even layers, it is converted to vertical bow-shaped filling along the Y-axis direction to form a grid structure that intersects with the previous layer. This method can effectively improve the forming quality of additive manufacturing, reduce the deformation risk caused by the accumulation of thermal stress, and make the machining process more uniform and reliable.

[0019] S33. The robotic arm carries a line laser sensor to scan and measure the milled groove body, obtains the point cloud data of the actually machined groove body, and pairs the measured groove body model with the theoretical grooving model through point cloud template matching to obtain the actual workpiece coordinate system; S34. According to the matching result, correct the position of the additive path, and control the robotic arm to perform the laser cladding process. Adjust process parameters such as laser power and powder feeding rate according to the geometric characteristics of the groove body. For example, use a higher power and medium powder feeding rate for the bottom layer filling to ensure good bonding with the substrate; use medium power and slightly higher powder feeding rate for the middle layer to quickly fill the material; reduce the power and powder feeding rate for the surface layer to ensure surface flatness and achieve precise material filling of the regularized grooving area.

[0020] Preferably, the grinding treatment of the clad layer height in step S4 includes: S41. After the laser cladding filling is completed, the robotic arm carries a line laser sensor to scan and measure the surface of the filling area, uses the cross-section analysis method to identify the surface normal and curvature change characteristics, accurately extracts and fits the contour boundary of the clad layer height, and based on this, divides the original surface mesh model obtained in S21 to obtain the surface model of the clad layer to be ground and polished.

[0021] S42. The rough grinding path generation method based on the scratch direction projects the contour boundary of the clad layer height and the scratch direction curve onto the two-dimensional principal component plane. Taking the scratch direction curve as the reference, a series of grinding paths parallel to the scratch direction are generated under the constraint of the contour boundary by using the two-dimensional curve equidistant offset algorithm. Subsequently, through the orthogonal projection transformation, the two-dimensional path points on the principal component plane are accurately mapped back to the three-dimensional surface model, and the local surface normal vector at each path point is calculated. Based on this normal vector information, the spatial posture of the grinding tool is optimized and adjusted to keep the tool axis always consistent with the machining surface normal. S43. The full-coverage fine polishing path generation method adopts a reciprocating zigzag strategy with boundary constraints.

[0022] First, according to the closed contour boundary, set the main machining direction on the two-dimensional projection plane. Calculate the step distance between adjacent paths according to the effective width of the tool and the desired overlap rate, and generate a series of reciprocating straight line segments parallel to the main direction. Add connecting line segments between the endpoints of every two adjacent straight line segments to form a continuous "zigzag" path. The generated two-dimensional path is mapped onto the three-dimensional surface model through back-projection, and the tool posture is adjusted according to the surface normal vector of the mapped points.

[0023] S44. In the rough grinding stage, dynamically match the grinding process parameters according to the maximum allowance data of each area of the path to achieve effective material removal in a single process and significantly reduce the main allowance. Subsequently, enter the fine polishing stage. Use a line laser scanner to perform high-precision measurement on the machined area, generate a full-coverage fine polishing path based on the measurement results, and adaptively match the polishing process parameter combination, including polishing pressure, feed speed, and polishing wheel speed, to ensure the fine machining quality.

[0024] S45. Continuously execute the measurement and fine polishing steps in S44, and continuously optimize the processing process through a closed-loop feedback method until the residual height value within the area drops below the preset threshold. At this time, it indicates that the excess material of the cladding layer has been effectively removed, and the surface topography of the repaired area has achieved a smooth transition with the surface of the original workpiece, completing the overall repair of the scratch on the inner wall of the servomotor cylinder.

[0025] The present invention can achieve the following beneficial effects: 1. Realize the in-situ repair of the scratch on the inner wall of the servomotor cylinder. Complete the measurement, milling, cladding, and grinding processes through a multi-process integrated system, without the need for disassembly and return to the factory, significantly shortening the repair cycle, reducing safety risks and costs.

[0026] 2. In the additive filling stage, innovatively adopt the bow-shaped layered alternating additive path strategy to effectively disperse heat accumulation, reduce the deformation risk caused by thermal stress, and at the same time enhance the interlayer bonding strength to ensure sufficient filling.

[0027] 3. The vision guidance and point cloud template matching technology solve the path deviation problem during the tool switching process between milling and additive manufacturing, accurately compensate for the robot accuracy error and tool switching error, and ensure that the additive filling path coincides precisely with the actual milled groove body.

[0028] 4. The adaptive repair method based on closed-loop feedback realizes the intelligent adjustment of process parameters through the integrated processing of the entire process from measurement, milling, additive manufacturing to grinding, improves the repair efficiency and quality, and reduces the need for manual intervention. Description of the Drawings

[0029] The following further describes the present invention in conjunction with the drawings and embodiments: Figure 1 is the flowchart of the method of the present invention; Figure 2 is the effect diagram of point cloud data acquisition and processing of the present invention; Figure 3 is the effect diagram of scratch feature definition of the present invention; Figure 4 is the simulation diagram of the grooving geometric model of the present invention; Figure 5 is the simulation diagram of the formed geometric model of the present invention; Figure 6 is the simulation diagram of the milling grooving processing path of the present invention; Figure 7 is the structure diagram of the workpiece coordinate system of the present invention; Figure 8 is the effect diagram of the bow-shaped interlayer alternating filling path of the present invention; Figure 9 is the effect diagram of the residual height contour extraction of the cladding layer of the present invention Figure 10This is the effect diagram of the zoned grinding and process matching of the present invention; Figure 11 This is the structural diagram of the servomotor and repair system of the present invention; Figure 12 This is the structural diagram of the constant-force grinding and polishing tool of the present invention. Specific implementation manners

[0030] The preferred solution is as Figures 1 to 12 shown. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement, the steps are as follows: S1. Integrate the multi-process full-process repair equipment. Install a line laser sensor and a tool quick-change device at the end of the six-axis robotic arm. Integrate a milling tool, a laser cladding tool, and a constant-force grinding and polishing tool through a quick-change interface to achieve high-precision scanning measurement, milling regularization, material filling, and surface grinding and polishing of the scratches on the inner wall of the servomotor cylinder.

[0031] S11. Install a line laser sensor and a tool quick-change device on the flange at the end of the six-axis robotic arm; the milling tool, the laser cladding tool, and the constant-force grinding and polishing tool are all configured with corresponding quick-change interfaces, so that the robotic arm can achieve multi-process integrated processing by quickly changing tools.

[0032] S12. The line laser sensor is used to perform high-precision scanning and measurement on the scratch area on the inner wall of the servomotor cylinder to obtain surface topography data, providing basic data for the precise positioning of the repair area and the subsequent detection of the remaining height of the cladding layer. The X-axis resolution of the used sensor reaches 23.5 µm, the scanning frequency is 8 kHz, and the single-scan width can reach 100 mm.

[0033] S13. The milling tool consists of a high-speed electric spindle and a milling cutter, and is used to mill the scratch area to form a regular groove. The maximum rotational speed of the electric spindle can reach 12,000 rpm, and the power is 1.2 kW. The laser cladding tool consists of a laser and a cladding nozzle. The laser cladding technology is used to fill the material in the regular groove area. The cladding nozzle adopts a four-channel coaxial powder feeding design, and the cladding material is selected as a matching metal powder to ensure metallurgical bonding between the filling material and the substrate material. The constant-force grinding and polishing tool consists of a force control unit, a motor, a grinding wheel, and a polishing wheel, and is used to grind and polish the remaining height of the cladding layer, remove the excess material, and restore the original surface topography of the workpiece. The force control unit can achieve constant contact force control within the range of 3 - 30 N. The grinding wheel and the polishing wheel are configured with abrasives of different grits from 80 to 800 meshes according to the grinding and polishing stages, and the final surface roughness can reach Ra 0.4 µm, meeting the high-precision requirements of the sealing surface on the inner wall of the servomotor cylinder.

[0034] S2. Regularize the scratched surface. Obtain the point cloud data of the scratched area through a line laser sensor. Through feature point definition and geometric modeling, construct a grooving geometric model with the scratch direction as the reference, a trapezoidal cross-section, the original surface at the top, and a plane at the bottom. Use the slicing method and the zigzag pattern strategy to plan the milling path and perform the milling operation to form a regularized groove body.

[0035] S21. Control the robotic arm to carry a line laser sensor to scan and measure the area around the scratch, obtain the point cloud data of the area surface, and use the point cloud processing algorithm to filter and reconstruct the data to generate an accurate surface mesh model. In specific operations, the robotic arm moves along the set scanning path, and the scanning path spacing is set to 1 mm to ensure sufficient overlap of the point cloud data. The collected original point cloud data is denoised and filtered and downsampled through the processing algorithm, and finally the Poisson surface reconstruction algorithm is applied to generate a triangular mesh model with an accuracy of not less than 0.1 mm, retaining the fine morphological features of the inner wall surface of the servomotor cylinder.

[0036] S22. Manually select the scratch feature points on the surface mesh model and use the spline fitting method to calculate the feature curve representing the scratch direction. Specifically, the operator marks the starting point, ending point, and intermediate key turning points (usually 5 - 10 points) of the scratch in the software interface, and uses the cubic B-spline curve algorithm to fit these points to generate a smooth and continuous feature curve, which accurately represents the geometric direction of the scratch and provides a reference for the construction of the subsequent repaired geometry.

[0037] S23. Construct a repaired geometry based on the feature curve. This geometry uses the scratch direction spline curve as the reference, has a trapezoidal cross-section, and adjusts the trapezoid height, upper base width, and side wall angle according to the actual scratch situation; through the intersection operation of this geometry and the S21 surface mesh model, a regularized grooving geometric model with the original surface at the top, a plane at the bottom, and a trapezoidal cross-section is formed. Specifically, after setting the trapezoid parameters, subtract this repaired geometry from the original surface mesh model through Boolean operation to obtain a grooving geometric model with accurate boundaries and smooth transitions.

[0038] S24. Plan the milling processing path with the grooving geometric model as the theoretical model, control the robotic arm to perform the milling operation, and realize the regularization of the scratched area. The milling processing adopts a layer-by-layer cutting strategy, with the cutting depth per layer controlled within 1 mm, and the processing path adopts a zigzag pattern to ensure continuous tool paths and uniform cutting forces. The robotic arm calculates the joint angles through the inverse kinematics algorithm to ensure that the axis of the milling cutter is always consistent with the normal direction of the bottom plane of the groove body. By controlling the milling parameters and trajectories, a regularized groove body with sharp edges and a surface roughness Ra ≤ 1.6 μm is formed, creating ideal geometric conditions for subsequent laser cladding.

[0039] Milling slotting path generation method: First, perform layer slicing on the three-dimensional slotting model along the normal direction of the bottom surface, and its layer slicing mathematical expression is ; where h represents the slice thickness and k represents the current slice layer number.

[0040] Take A plane performs a slicing operation on the three-dimensional slotting geometric model M to obtain a set of two-dimensional contours ; where represents the i-th closed contour.

[0041] Adopt a zigzag path pattern. Based on the initial contour, generate a sequence of inner-shrinking equidistant contours by iteratively applying the offset function : ; where d represents the adjacent contour spacing and m represents the offset iteration times.

[0042] Meanwhile, by constructing the connection segments between the contours connect the adjacent offset contours and into a continuous curve. Considering the minimization of the idle travel, solve the connection point selection problem: ; where is the trajectory point of the j-th closed contour, is the trajectory point of the (j + 1)-th closed contour.

[0043] S3. Regularize the material filling in the area. Establish a workpiece coordinate system with the bottom of the slot. Adopt the bow-shaped layer-by-layer alternating method to plan the additive manufacturing path. Measure the slot after milling with a line laser and match it with the theoretical model. Correct the position of the additive manufacturing path. Adjust the laser power and powder feeding rate according to the slot characteristics, and implement precise material filling.

[0044] S31. According to the slotting geometric model, establish a theoretical workpiece coordinate system with the edge contour and characteristic points of the bottom plane of the slot.

[0045] S32. The additive manufacturing path of the robotic arm adopts the bow-shaped layer-by-layer alternating method. This method first performs contour slicing on the three-dimensional model along the Z-axis direction to obtain a series of two-dimensional contours on parallel planes. Then, for the filling trajectory of each layer, adopt the "bow"-shaped motion mode for planning: Given the two-dimensional closed contour C and the filling angle θ (the angle with the x-axis), the generation process of the bow-shaped trajectory follows the following steps: First, perform a coordinate transformation on the contour C according to the filling angle θ so that the filling direction is aligned with the coordinate axes: ; ; where represents the rotation transformation matrix with an angle of θ.

[0046] Then, determine the bounding box of the contour , and generate a series of scan lines parallel to the x-axis: ; ; where the parameter d is the spacing between adjacent scan lines.

[0047] Secondly, calculate the intersection set of each scan line and the contour : ; where is the number of intersections between the contour and the scan line (the number of intersection points of a closed contour must be even).

[0048] Connect the intersection pairs on adjacent scan lines into a bow shape, and transform the generated bow-shaped trajectory back to the original coordinate system. For the interlayer alternating bow-shaped trajectory, different filling angles are used for the odd and even layers. In the even layer, the trajectory direction is filled according to the given filling angle θ; in the odd layer, the trajectory is filled at the filling angle θ + 90°; to form a grid structure that intersects with the previous layer.

[0049] Set the layer thickness h = 0.8mm, the filling angle θ = 0°, and the row spacing d = 2mm, to obtain the interlayer alternating bow-shaped trajectory as Figure 8 shown

[0050] S33. The robotic arm carries a line laser sensor to scan and measure the milled groove body, obtains the point cloud data of the actually machined groove body, and pairs the measured groove body model with the theoretical grooving model through point cloud template matching to obtain the actual workpiece coordinate system. S34. According to the matching result, correct the position of the additive manufacturing path, control the robotic arm to perform the laser cladding process, and adjust process parameters such as laser power and powder feeding rate according to the geometric characteristics of the groove body to achieve precise material filling in the regularized grooving area. During the laser cladding process, a higher power and medium powder feeding rate are used for the bottom layer filling to ensure good bonding with the substrate; a medium power and slightly higher powder feeding rate are used for the middle layer to quickly fill the material; the power and powder feeding rate are reduced for the surface layer to ensure surface flatness.

[0051] S4. Grinding and finishing the height of the cladding layer. Use a line laser sensor to scan the filling area, extract the height contour boundary through cross-section analysis method, generate a rough grinding path based on the scratch direction, generate a fine polishing path using a reciprocating zigzag strategy with boundary constraints, dynamically match the grinding and polishing parameters according to the maximum allowance data, and achieve a smooth transition between the cladding layer and the original surface through the cyclic execution of measurement and fine polishing.

[0052] S41. After the laser cladding filling is completed, the robotic arm carries a line laser sensor to scan and measure the surface of the filling area, identify the surface normal change characteristics using cross-section analysis method, and accurately extract the height contour of the cladding layer. Specifically, the scanning area covers the cladding area and its surrounding 10 mm range. The collected point cloud data is first filtered to remove noise, and then a cross-section curve is extracted every 0.5 mm along the direction perpendicular to the cladding band. For each cross-section curve, calculate the curvature value of each point through the curvature analysis algorithm, calculate the curvature mutation point, and mark it as the boundary point between the cladding layer and the original surface. Connect all the boundary points on the cross-sections to form a complete height contour of the cladding layer.

[0053] S42. Method for generating a rough grinding path based on the scratch direction. Specifically, use the principal component analysis (PCA) algorithm to determine the main distribution direction of the contour boundary of the cladding layer height, and generate a two-dimensional plane composed of the first two principal components; then, orthogonally project the contour boundary and the scratch direction curve onto this two-dimensional plane; in the projection plane, with the scratch direction curve as a reference, use the equidistant offset algorithm to generate a series of grinding paths parallel to the scratch direction under the constraint of the contour boundary; finally, through the orthogonal projection transformation, accurately map the two-dimensional path points on the principal component plane back to the three-dimensional surface model, and calculate the local surface normal vector at each path point. Based on this normal vector information, optimize and adjust the spatial posture of the grinding tool to make the tool axis always consistent with the machining surface normal.

[0054] S43. Method for generating a full-coverage fine polishing path, using a reciprocating zigzag strategy with boundary constraints. First, set the main processing direction on the two-dimensional projection plane according to the closed contour boundary. Calculate the step distance between adjacent paths according to the effective width of the tool and the desired overlap rate, and generate a series of reciprocating straight line segments parallel to the main direction. Add connecting line segments between the endpoints of each adjacent two straight line segments to form a continuous "zigzag" path. The generated two-dimensional path is mapped to the three-dimensional surface model through back-projection, and the tool posture is adjusted according to the surface normal vector of the mapped point.

[0055] In the rough grinding stage, the grinding process parameters are dynamically matched according to the maximum allowance data of each area of the path to achieve effective material removal in a single process and significantly reduce the main allowance. Subsequently, it enters the fine polishing stage. A line laser scanner is used to perform high-precision measurement on the machined area. Based on the measurement results, a full-coverage fine polishing path is generated, and the polishing process parameter combination, including polishing pressure, feed speed, and polishing wheel speed, is adaptively matched to ensure the quality of fine machining.

[0056] S45. The measurement and fine polishing steps in S44 are executed cyclically, and the machining process is continuously optimized through a closed-loop feedback method until the residual height value in the area drops below the preset threshold. At this time, it indicates that the excess material of the cladding layer has been effectively removed, and the surface topography of the repaired area and the surface of the original workpiece achieve a smooth transition, completing the overall repair of the scratch on the inner wall of the servomotor cylinder. After each round of fine polishing, the surface state is re-scanned and evaluated, and the maximum value and root mean square value of the residual height are calculated. When the maximum height value is less than 0.05 mm and the root mean square value is less than 0.02 mm, it is determined that the repair meets the standard.

[0057] The method of the present invention has been actually applied to the repair of the scratch on the inner wall of the servomotor cylinder of a large hydropower station. Practice has proved that this method can shorten the repair work of the scratch on the inner wall of the servomotor cylinder from the traditional repair cycle of more than ten days to within 3 - 5 days. By establishing a machining strategy and a multi-process integration system based on three-dimensional measurement, high-quality in-situ repair can be achieved without disassembly and returning to the factory, significantly reducing safety risks and maintenance costs, effectively ensuring the safe and stable operation of the hydropower unit, and having important engineering application value and popularization prospects.

[0058] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations on the present invention. The protection scope of the present invention should be the technical solutions recorded in the claims, including the equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, the equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. An adaptive additive and subtractive repair method for the inner wall scratches of a servomotor cylinder based on three-dimensional measurement, characterized in that It includes the following steps: S1. Install a line laser sensor and a tool quick-change device at the end of the robotic arm, and integrate a milling tool, a laser cladding tool, and a constant-force grinding and polishing tool through a quick-change interface; S2. Obtain the point cloud data of the scratch area through the line laser sensor. Through feature point definition and geometric modeling, construct a grooving geometric model with the scratch direction as the reference, a trapezoidal cross-section, the original surface at the top, and a plane at the bottom. Adopt the slicing method and the zigzag pattern strategy to plan the milling path, and perform the milling operation to form a regularized groove; S3. Establish a workpiece coordinate system at the bottom of the groove. Adopt the bow-shaped layer-by-layer alternating method to plan the additive path. After measuring the milled groove with the line laser and matching it with the theoretical model, correct the position of the additive path, and adjust the laser power and powder feeding rate according to the groove characteristics; S4. Use the line laser sensor to scan the filling area, extract the remaining height contour boundary through cross-section analysis, generate a rough grinding path based on the scratch direction, generate a fine polishing path using the reciprocating zigzag strategy with boundary constraints, dynamically match the grinding and polishing parameters according to the maximum allowance data, and realize the smooth transition between the cladding layer and the original surface through the cycle execution of measurement and fine polishing.

2. The adaptive additive and subtractive repair method for the scratch on the inner wall of the servomotor cylinder of the robot based on three-dimensional measurement according to claim 1, wherein: The sub-steps of S1 are: S11. Install a line laser sensor and a tool quick-change device on the flange at the end of the robotic arm; the milling tool, the laser cladding tool, and the constant-force grinding and polishing tool are all equipped with corresponding quick-change interfaces, enabling the robotic arm to achieve multi-process integrated processing through quick tool change; S12. The line laser sensor is used to perform high-precision scanning and measurement on the scratch area on the inner wall of the actuator cylinder to obtain surface topography data; S13. Adopt the laser cladding technology to fill the material in the regularized groove area.

3. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 2, characterized in that: The described robotic arm adopts a six-axis robotic arm; The described constant-force grinding and polishing tool consists of a force control unit, a motor, a grinding wheel, and a polishing wheel, and is used for grinding and polishing the remaining height of the cladding layer to remove excess material and restore the original surface topography of the workpiece; The described milling tool consists of a high-speed electric spindle and a milling cutter, and is used for milling the scratch area to form a regularized groove; The described laser cladding tool consists of a laser and a cladding nozzle.

4. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 1, characterized in that: The sub-steps of S2 are: S21. Control the robotic arm to carry the line laser sensor to scan and measure the area around the scratch to obtain the surface point cloud data of the area. Adopt a point cloud processing algorithm to filter and reconstruct the data to generate an accurate surface mesh model; S22. Manually select the scratch feature points on the surface mesh model, and adopt the spline fitting method to calculate the feature curve representing the scratch direction; S23. Based on the feature curve, construct a repair geometry. This geometry takes the scratch direction spline curve as the reference, adopts a trapezoidal cross-section, and adjusts the trapezoidal height, upper base width, and sidewall angle according to the actual scratch situation; through the intersection operation of this geometry and the S21 surface mesh model, form a regularized grooving geometric model with the original surface at the top, a plane at the bottom, and a trapezoidal cross-section; S24. Plan the milling processing path with the grooving geometric model as the theoretical model. The path generation method adopts the slicing method and the zigzag pattern strategy.

5. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 4, characterized in that: The process of S24 based on the slicing method and the zigzag pattern strategy is as follows: Discretize and slice the three-dimensional grooving geometric model in the bottom plane direction to form a series of two-dimensional contours; Adopt a zigzag pattern strategy. Starting from the contour boundary, gradually calculate the offset inward along the contour shape to generate a sequence of equidistant contours that shrink inward; To ensure the machining quality, apply speed smoothing control at the corners to avoid sudden acceleration changes caused by sharp turns. At the same time, when connecting adjacent contours, optimize the tool lifting and cutting point selection to reduce the idle travel; Control the robotic arm to perform milling operations to achieve the regularization treatment of the scratched area.

6. The adaptive additive and subtractive repair method for the scratch on the inner wall of the servomotor cylinder by a robot based on three-dimensional measurement according to claim 1, wherein: The sub-steps of S3 are: S31. According to the grooving geometric model, establish a theoretical workpiece coordinate system based on the edge contour and feature points of the bottom plane of the groove; S32. Adopt a bow-shaped layer-by-layer alternating method to plan the additive path of the robotic arm; S33. The robotic arm carries a line laser sensor to scan and measure the milled groove body, obtain the point cloud data of the actually machined groove body, and pair the measured groove body model with the theoretical grooving model through point cloud template matching to obtain the actual workpiece coordinate system; S34. According to the matching result, correct the position of the additive path, control the robotic arm to perform the laser cladding process; and adjust process parameters including laser power and powder feeding rate according to the geometric features of the groove: (1) For the bottom layer filling, use a higher power and medium powder feeding rate to ensure good bonding with the substrate; (2) For the intermediate layer, use medium power and slightly higher powder feeding rate to quickly fill the material; (3) For the surface layer, reduce the power and powder feeding rate to ensure the surface flatness and achieve precise material filling of the regularized grooving area.

7. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 6, characterized in that: In S32, the bow-shaped layer-by-layer alternating method is adopted; Slice the three-dimensional model at equal heights in the Z-axis direction to form a series of two-dimensional contours on a series of parallel planes: Adopt bow-shaped filling along the X-axis direction for odd layers; For even layers, convert to vertical bow-shaped filling along the Y-axis direction to form a grid structure that intersects with the previous layer.

8. A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 4, characterized in that: The sub-steps of S4 are: S41. After the laser cladding filling is completed, the robotic arm carries a line laser sensor to scan and measure the surface of the filled area, use the sectional analysis method to identify the surface normal and curvature change characteristics, accurately extract and fit the contour boundary of the cladding layer residual height, and based on this, segment the original surface mesh model obtained in S21 to obtain the surface model of the cladding layer to be ground and polished; S42. Use a rough grinding path generation method based on the scratch direction. Project the cladding layer residual height contour boundary and the scratch direction curve onto the two-dimensional principal component plane, use the two-dimensional curve equidistant offset algorithm to generate a grinding path parallel to the scratch direction under the constraint of the contour boundary, then map it back to the three-dimensional surface model through orthogonal projection transformation, and optimize and adjust the spatial attitude of the grinding tool according to the local surface normal vector at the path points to make its axis consistent with the machining surface normal; S43. Set the main machining direction on the two-dimensional projection plane according to the closed contour boundary; Calculate the step distance between adjacent paths according to the effective width of the tool and the desired overlap rate, and generate a series of reciprocating straight line segments parallel to the main direction; Add connecting line segments between the endpoints of every two adjacent straight line segments to form a continuous "zigzag" path; the generated two-dimensional path is mapped onto the three-dimensional surface model through back-projection, and at the same time, the tool posture is adjusted according to the surface normal vector of the mapped points. S44. In the rough grinding stage, the grinding process parameters are dynamically matched according to the maximum allowance data of each area of the path to achieve effective material removal in a single process and significantly reduce the main allowance; then enter the fine polishing stage. A line laser scanner is used to perform high-precision measurement on the machined area, and a full-coverage fine polishing path is generated based on the measurement results, and the polishing process parameter combination, including polishing pressure, feed speed, and polishing wheel speed, is adaptively matched to ensure the quality of fine machining. S45. The measurement and fine polishing steps in S44 are executed in a loop, and the machining process is continuously optimized through a closed-loop feedback method until the residual height value in the area drops below the preset threshold; at this time, it indicates that the excess material of the cladding layer has been effectively removed, and the surface topography of the repaired area is smoothly transitioned to the surface of the original workpiece, completing the overall repair of the scratch on the inner wall of the servomotor cylinder.

9. An adaptive additive and subtractive repair method for scratches on the inner wall of a servomotor cylinder by a robot based on three-dimensional measurement according to claim 1, characterized in that: The steps of the rough grinding path generation method based on the scratch direction in S42 are as follows: Project the contour boundary of the cladding layer height and the scratch direction curve onto the two-dimensional principal component plane. Based on the scratch direction curve, a series of grinding paths parallel to the scratch direction are generated under the constraint of the contour boundary by using the two-dimensional curve equidistant offset algorithm. Through orthogonal projection transformation, the two-dimensional path points on the principal component plane are accurately mapped back to the three-dimensional surface model, and at the same time, the local surface normal vector at each path point is calculated. Based on this normal vector information, the spatial posture of the grinding tool is optimized and adjusted so that the tool axis is always consistent with the machining surface normal.

10. A robot adaptive additive and subtractive repair system for scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement, characterized in that Adopted: A method for adaptive additive and subtractive repair of scratches on the inner wall of a servomotor cylinder based on three-dimensional measurement according to any one of claims 1-9.

Images