Water turbine top cover cavitation area robot polishing control method and system

Abstract

The invention relates to the technical field of electric digital processing, and discloses a water turbine top cover cavitation area robot polishing control method and system, and the method comprises the steps: constructing a hydropower station maintenance environment sensing module with multi-modal information fusion; a composite grinding robot is constructed; reconstructing a three-dimensional scene facing the cavitation part of the top cover of the water turbine and identifying a to-be-polished area; planning a polishing path of the composite robot facing the cavitation characteristics of the top cover; constructing a contact force model between the robot and the to-be-polished cavitation region of the top cover; combining the contact force model with a robot control law, and calculating a force tracking steady-state error; constructing a self-adaptive force control algorithm combined with a six-dimensional force sensor; robot self-adaptive force control grinding pre-experiment and parameter simulation setting are carried out; the invention discloses robot self-adaptive constant-force closed-loop feedback real-time control for a cavitation area. Constant force self-adaptive adjustment between the grinding head at the tail end of the robot and the curved surface of the top cover is achieved, and the machining quality of the robot during top cover cavitation area grinding is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of electronic digital processing technology, and in particular to a robotic grinding control method and system for cavitation areas of a turbine top cover. Background Technology

[0002] With the rapid development of robotics technology and the popularization of intelligent manufacturing, robotic polishing, due to its advantages such as consistent processing quality, high operating efficiency, and flexible deployment, has gradually replaced manual polishing and semi-automatic polishing machines, and is beginning to be applied in industries such as automotive, shipbuilding, and hydropower. Robotic polishing can significantly reduce the labor intensity of humans, minimize the impact of dust and noise generated during polishing on human health, and improve the working environment. Therefore, automated robotic polishing is the main trend in the future development of polishing processes.

[0003] Six-degree-of-freedom (6DOF) robotic arms have been widely used in automated robotic grinding due to their simple structure, convenient deployment, and high flexibility. Particularly in the field of intelligent maintenance of hydropower equipment, 6DOF robotic arms have been applied to machining scenarios such as milling cracks in turbine runner blades, grinding cracks in turbine runner blades, laser enhancement of turbine runner blades, grinding of the inner wall of the runner, and in-situ addition and subtraction of materials for the top cover. However, a standalone 6DOF robotic arm lacks mobility and cannot achieve the mobile processing of large-sized workpieces in the hydropower field. Researchers have increased the working space and reach of the robotic arm by installing a moving mechanism below it or deploying it on a sliding guide rail. In the currently published invention patent CN119217153A, "Automatic In-situ Robotic Grinding Method for the Inner Wall of a Turbine Runner Based on Three-Dimensional Measurement," Wu Tao et al. mounted a robotic arm on a ring guide rail to grind the inner wall of the turbine runner. While this method effectively increases the robot's range of motion, it also limits the working space of the robotic arm to the area around the ring guide rail. In the currently published invention patent CN116550990A, "A Mobile Laser Additive Processing Method and Device for a Large Turbine Top Cover," Liu Hui et al. mounted a robotic arm on a mobile trolley, enabling the arm to move. However, this device lacks an environmental perception module, preventing the mobile grinding robot system from actively sensing the complex maintenance environment beneath the turbine top cover. Consequently, it cannot perform obstacle avoidance planning or autonomous path planning during grinding operations. Therefore, the robot system lacks omnidirectional mobile processing capabilities and currently requires manual remote control for station changes before subsequent processing can commence. Furthermore, existing research on cavitation area detection for turbine top cover components typically relies on manual visual inspection or single-sensor detection. In contrast, the currently published invention patent CN119580069A, "A Method for Identifying Cracks and Cavitation Defects in Mixed-Flow Turbine Runners," utilizes visual image information combined with target detection algorithms to detect cavitation areas in turbine top cover components. However, this method only collects two-dimensional image information and cannot obtain the three-dimensional morphological features of the cavitation area surface. It can only preliminarily locate the cavitation area of ​​the turbine roof and cannot provide a reference for determining the amount of material to be removed during subsequent grinding. Manual inspection is still required to determine the amount of material to be removed. In existing roof cavitation detection tasks, it is urgent to construct the three-dimensional morphological features of the cavitation area surface to guide the design and selection of subsequent processing parameters. During the grinding process of the cavitation area of ​​the turbine roof, the force control accuracy between the robot's end mill and the environment has a significant impact on the grinding quality and effect. To improve the force control accuracy, existing methods usually install a six-dimensional force sensor at the robot's end mill to measure the force on the robot's end mill in real time, thereby providing closed-loop feedback for the control algorithm.

[0004] Currently, some scholars, such as Shen Yichao et al., have proposed robot force control methods based on fuzzy impedance control. This method can achieve contact force tracking of the desired force between the robot and the environment even when the environmental stiffness parameters and environmental position are unknown. However, to eliminate force tracking errors, this method requires real-time estimation of the environmental stiffness, and online estimation is prone to convergence problems. Furthermore, fuzzy logic and real-time estimation algorithms place high demands on the computational power of the robot control system. In the paper "Adaptive Variable Impedance Force Tracking Control Method for Robots in Unstructured Environments," Gan Yahui et al. designed an adaptive variable impedance control method. By adaptively adjusting the damping control parameters, they adjusted the dynamic characteristics of the robot's end effector when in contact with the environment, thereby improving the force control accuracy of the end effector. Duan Jinjun et al. designed a force control algorithm for online adjustment of impedance control parameters. However, the adjustment capability of the single damping parameter in this algorithm is limited, and improper adjustment of the damping parameter can easily lead to fluctuations in the contact force during robot grinding, resulting in over-grinding or under-grinding. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a robotic grinding control method and system for the cavitation area of ​​a turbine top cover, which addresses the shortcomings of the above-mentioned methods and devices in terms of omnidirectional mobile processing capability, environmental feature perception and three-dimensional reconstruction capability, and adaptive constant force control.

[0006] To achieve the above objectives, in a first aspect, this application provides a robotic grinding control method for cavitation areas of a turbine top cover, comprising the following steps: Step 1: Construct a multimodal information fusion module for the hydropower station maintenance environment perception module; Step 2: Construct a composite grinding robot; Step 3: 3D scene reconstruction and identification of the area to be polished for the cavitation area of ​​the turbine top cover; Step 4: Composite robot grinding path planning for cavitation characteristics of the top cover; Step 5: Construct a contact force model between the robot and the cavitation area to be polished on the top cover; Step Six: Combine the contact force model and the robot control law to calculate the steady-state error of force tracking; Step 7: Construct an adaptive force control algorithm that combines a six-dimensional force sensor to adjust the position of the robot's end effector in order to control the normal contact force between the robot and the cavitation area to be polished. Step 8: Pre-experiment and parameter simulation tuning of robot adaptive force control grinding; Step 9: Real-time adaptive constant force closed-loop feedback control of the robot for cavitation zones.

[0007] In step one, the hydropower station maintenance environment perception module includes a lidar, a binocular depth camera, and an IMU sensor. By fusing point cloud information collected by the lidar, visual and depth map information collected by the binocular depth camera, and multi-axis acceleration / angular velocity information collected by the IMU sensor, it perceives the surrounding environment of the composite grinding robot. The integrated 3D point cloud and visual texture information is used for scene reconstruction of the hydropower station maintenance area and identification of the cavitation area to be ground. Among them, the visual image information collected by the binocular depth camera is used to compensate for the texture features of the 3D point cloud information, and the IMU sensor compensates for the motion distortion of the point cloud and image by providing high-frequency motion state data.

[0008] In step two, the constructed hydropower station maintenance environment perception module is installed on an omnidirectional mobile platform equipped with a robotic arm, thereby constructing a composite grinding robot. The end of the robotic arm is equipped with a grinding device, which consists of an electric grinding head and a six-dimensional force sensor connected by an adapter.

[0009] The algorithm module of the hydropower station maintenance environment perception module includes a laser odometer module, a visual odometer module, an image frame association module, and an inter-frame map association module. In step three, after receiving the point cloud data input from the lidar, the laser odometer module performs point cloud motion distortion correction and then updates the point-to-plane laser odometer module, thereby adding map points to the global map. After receiving the visual image information input, the visual odometer module updates the visual odometer module through the image frame association module. During the global map update process, the inter-frame map association module achieves visual texture rendering of the global map by minimizing photometric errors. The input from the IMU sensor is used for state propagation, providing position information for the laser odometer module and the visual odometer module, and constructing constraints for matching point clouds and visual feature points, thereby realizing the 3D scene reconstruction of the global map.

[0010] In step three, the identification of the area to be polished includes the following steps: 1) Collect point cloud information and visual image information of the scene below the turbine top cover, construct a three-dimensional scene below the top cover, and fuse visual texture and point cloud information to construct the three-dimensional morphological features of the lower surface of the top cover. 2) Combine visual image information to identify and segment the cavitation area to be polished, perform fine modeling of the area to be polished, determine the amount of material to be removed from the cavitation part of the top cover, and provide a reference for the planning of subsequent polishing process parameters.

[0011] In step four, the composite robot grinding path planning for the cavitation characteristics of the top cover includes the following steps: 1) Based on the point cloud of the cavitation area collected by the robot's environmental perception module, fit the approximate curved surface and plane of the cavitation area; 2) After completing the surface reconstruction, select robot polishing points on the reconstructed surface and connect the polishing points to obtain the preset polishing path; wherein, the preset polishing path is a reciprocating traversal path. 3) Smooth the preset polishing path; 4) During the actual grinding process, the grinding path is adjusted differently according to the severity of the cavitation. For shallow cavitation areas, i.e., areas with cavitation depth ≤ 2mm, a parallel reciprocating path is used, and the spacing is selected based on the geometric characteristics of the cavitation area and the grinding process requirements. For deep cavitation pit areas, i.e. areas with cavitation depth > 2mm, a spiral inward path is used, and the angle of the grinding head needs to be dynamically adjusted during the grinding process.

[0012] In step five, an impedance control model is used to model the dynamic relationship of the contact force between the robot's end effector position and the cavitation area to be polished on the top cover. The model expression is as follows: ; In the formula, , , These represent the robot's inertial parameters, damping parameters, and stiffness parameters, respectively. For the desired contact force, For environmental contact force, For the robot's reference position, This refers to the robot's actual position. Environmental contact force The expression is: ; In the formula, , These represent the stiffness and damping parameters of the environment, respectively. This represents the position of the robot relative to the surface of the contact environment; it is the equilibrium position where the robot's end effector is exactly free from force. The first derivative represents the robot's actual position. The first derivative represents the robot's reference position. The second derivative representing the robot's actual position. The second derivative represents the robot's reference position. The first derivative represents the contact position.

[0013] In step six, force tracking steady-state error The expression is: ; In the formula, Indicates the desired contact force; Indicates environmental contact force; Indicates force tracking error; The reference position trajectory for the robot represents the pre-generated offline grinding position trajectory used to execute the plan; This represents the stiffness parameter in the impedance control model; Represents the stiffness parameters of the external environment; This indicates the position where the robot just touches the surface of the external environment, which is the equilibrium position where the robot's end effector is just free from force. Represents the Laplace operator; t Indicates time.

[0014] In step seven, the adaptive force control algorithm expression based on the six-dimensional force sensor is as follows: ; In the formula, express t The acceleration of the robot's end effector at any given moment; express t The second derivative of the external environment position at time; Represents the inertial parameters in the impedance control model; This represents the proportional parameter in the PD control law; express t The expected contact force at any moment; express t The contact force between the robot's end effector and the environment at any given moment; This represents the stiffness parameter in the impedance control model; express t The first derivative of the expected contact force at all times; express t The first derivative of the contact force between the robot's end effector and the environment at any given moment; This represents the damping parameter in the impedance control model; This is the first damping parameter in the designed control law that is adjusted based on the first-order difference of the force deviation; This is the second damping parameter in the designed control law, adjusted based on the first-order difference of the force deviation. express t The speed of the robot's end effector at time -1; express t The first derivative of the external environment position at a given time; express t The speed of commands given at the robot's end effector at any given moment; This represents the system communication cycle between the robot controller and the servo driver; express t The position of the robot's end effector at any given moment; express t The command position of the robot's end effector at time -1.

[0015] In step eight, the robot adaptive force control grinding pre-experiment and parameter simulation tuning includes the following steps: 1) In the pre-grinding experiment, the robot's grinding head was not started initially. After setting the initial impedance control parameters, the robotic arm was controlled to conduct a fixed-point force control experiment in the cavitation area of ​​the top cover using a trial-and-error method. Force and position information were collected during the dynamic contact process between the robot's end effector and the top cover. Multiple contact points were selected, the above experiment was repeated, and data were recorded. 2) Control the end effector of the robotic arm to move at a constant speed over the cavitation area of ​​the top cover, and collect relevant force and position information; 3) After completing the above force contact pre-experiment, the collected information is imported into data analysis and processing software to build a simulation model, thereby completing system identification and estimating the stiffness and damping parameters of the environment. 4) Based on the known information of environmental stiffness and damping, the robot controller parameters are initially adjusted. The initial robot grinding process parameters are first set, and then the robot grinding head is started to conduct a constant force control experiment. Multiple areas are selected for trial grinding, and then the robot grinding process parameters are adjusted to select the process parameters. 5) After determining the process parameters, repeat the previous robot force contact experiment, record the force and displacement information of the robot end, export it, and perform system analysis and simulation again to determine the optimal robot controller and impedance control parameters.

[0016] In step nine, the robot's adaptive constant force closed-loop feedback real-time control for cavitation zones includes the following steps: 1) Based on the constructed robot environment perception module, the scene features of the area to be polished on the top cover of the water turbine are collected to complete the path planning and intelligent obstacle avoidance of the composite polishing robot; 2) Compile the adaptive force control algorithm built in step seven into a file that can be directly executed by the robot hardware, and burn it into the robot controller; 3) In the algorithm execution layer, the position, speed, current and robot end force sensing information signals of the servo motor are read based on the EtherCAT bus, and the information signals are input into the module compiled by the adaptive control algorithm for calling, calculation and execution. The obtained position command output is sent to the servo motor for execution through the real-time EtherCAT bus.

[0017] Secondly, this application provides a robotic grinding system for cavitation areas of a turbine roof, used to implement the aforementioned robotic grinding control method for cavitation areas of a turbine roof, the robotic grinding system comprising: A robotic arm, the end of which is used to mount grinding equipment; A lifting platform, wherein a robotic arm is mounted on the top of the lifting platform for adjusting the height of the robotic arm; An omnidirectional mobile platform, wherein a lifting platform is mounted on the top of the omnidirectional mobile platform, and the omnidirectional mobile platform is used for moving the load lifting platform and the robotic arm; The lifting device has its lower end mounted on an omnidirectional moving platform, and its top is equipped with a driver. The power output end of the driver is equipped with a hydropower station maintenance environment perception module, which is used to drive the hydropower station maintenance environment perception module to rotate. The hydropower station maintenance environment perception module is used to perceive the surrounding environment of the composite grinding robot, and integrates three-dimensional point cloud and visual texture information to complete the scene reconstruction of the hydropower station maintenance area and the identification of the cavitation area to be ground.

[0018] Compared with the prior art, the above-conceptual technical solution conceived in this application has the following beneficial effects: 1. At the system level, this invention constructs a composite robot system with an expandable workspace through the high integration of a six-degree-of-freedom robotic arm and an omnidirectional mobile platform. It can support precise motion control of the composite grinding robot system at a control frequency of 1kHz, ensuring that the robot can complete the end-effector pose adjustment with high dynamics during force-controlled grinding, thereby improving the quality of force-controlled grinding.

[0019] 2. At the perception layer, this invention designs a robot environment perception module that integrates multi-source sensor data from LiDAR, binocular cameras, and IMU. This effectively compensates for the lack of point cloud texture in single LiDAR data and the inability of single visual sensing information to reflect the three-dimensional characteristics of the environment, thus effectively enabling robot environment reconstruction and autonomous localization. Furthermore, the environment perception module of this invention adopts a modular design, is compatible with multiple communication protocols, and can be used plug-and-play on different robot platforms, effectively avoiding the redundant development of robot environment perception algorithms.

[0020] 3. In the robot force control algorithm, the present invention adopts dual damping parameters to prevent overshoot or oscillation caused by inaccurate setting of single damping coefficient. It can also avoid problems such as inability to track dynamic environmental changes due to excessive single damping coefficient. It can effectively overcome the problem of limited adjustment capability of single damping parameter, realize constant force adaptive adjustment between robot end mill grinding head and top cover curved surface, and avoid over-grinding or under-grinding caused by contact force fluctuation during robot grinding.

[0021] 4. Regarding omnidirectional mobile processing capability, this invention deploys an environmental perception module above the mobile platform, enabling the composite grinding robot to actively perceive complex maintenance environments. This allows the robot system to possess obstacle avoidance planning and autonomous path planning functions, thus achieving omnidirectional mobile processing capability. In terms of environmental feature perception and 3D reconstruction, this method fuses 2D image information and 3D point cloud information, thereby constructing the 3D morphological features of the cavitation area surface, providing a reference for determining subsequent grinding process parameters. Regarding adaptive constant force control, this method employs variable damping control, leveraging the force control adjustment capability of a dual-damping parameter enhancement algorithm to achieve constant force adaptive adjustment between the robot's end-effector grinding head and the top cover curved surface, ensuring the processing quality of the robot when grinding the top cover cavitation area. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0023] Figure 1 This is a three-dimensional structural diagram of the overall device according to a preferred embodiment of the present invention.

[0024] Figure 2 This is a structural diagram of the hardware control system of a preferred embodiment of the present invention.

[0025] Figure 3 This is a general flowchart of a preferred embodiment of the present invention.

[0026] Figure 4 This is a three-dimensional structural diagram of a robot environment perception module according to a preferred embodiment of the present invention.

[0027] Figure 5 This is a flowchart of an environmental perception algorithm according to a preferred embodiment of the present invention.

[0028] Figure 6 This is a flowchart of a preferred embodiment of the robot adaptive constant force control algorithm of the present invention.

[0029] Figure 7 This is a simulated contact force variation diagram between the robot end mill and the workpiece being ground, according to a preferred embodiment of the present invention.

[0030] Figure 8 This is a diagram showing the actual contact force variation between the robot end mill and the workpiece being ground, according to a preferred embodiment of the present invention.

[0031] Figure 9 This is a scene reconstruction and point cloud topography map collected by the environment perception module in a preferred embodiment of the present invention.

[0032] Figure label: 1. Robotic arm; 2. Grinding equipment; 3. Hydropower station maintenance environment perception module; 4. Lifting device; 5. Lifting platform; 6. Omnidirectional mobile platform; 7. LiDAR; 8. Binocular depth camera; 9. IMU sensor. Detailed Implementation

[0033] To more clearly illustrate the purpose, technical solution, and beneficial effects of this application, a further detailed description of this application is provided below in conjunction with illustrations and specific embodiments. It should be specifically noted that the specific embodiments described below are only for illustrating the technical content of this application and do not constitute a limitation on the scope of protection of this application.

[0034] Regarding the explanation of terminology: In this application, "and / or" is used to describe the relationship between related objects, covering three possible situations: taking "A and / or B" as an example, it can indicate the situation where only A exists, A and B exist simultaneously, or only B exists; the symbol " / " indicates the "or" relationship between related objects, such as "A / B" which refers to A or B.

[0035] Regarding the description of the embodiments: The terms "exemplary" and "for example" appearing in this application are only used to illustrate the technical solutions through specific examples. It should be particularly emphasized that any implementation method or design scheme marked as "exemplary" or "for example" should not be construed as having an advantage over other solutions. Such expressions are only used to present the technical concepts more intuitively.

[0036] Example 1: See Figure 1 This invention provides a robotic grinding control method for cavitation areas of a turbine top cover, which can be used for constant force grinding of cavitation areas near the leak-stopping ring below the turbine top cover.

[0037] Step 1: Construct a multimodal information fusion hydropower station maintenance environment perception module 3.

[0038] Specifically, see Figure 4 The hydropower station maintenance environment perception module 3 includes a lidar 7, a binocular depth camera 8, and an IMU sensor 9. By fusing the point cloud information collected by the lidar 7, the visual and depth map information collected by the binocular depth camera 8, and the multi-axis acceleration / angular velocity information collected by the IMU sensor 9, it perceives the surrounding environment of the composite grinding robot. It integrates the three-dimensional point cloud and visual texture information for scene reconstruction of the hydropower station maintenance area and identification of the cavitation area to be ground. Among them, the visual image information collected by the binocular depth camera is used to compensate for the texture features of the three-dimensional point cloud information, and the IMU sensor (inertial measurement sensor) compensates for the motion distortion of the point cloud and image by providing high-frequency motion state data.

[0039] In this embodiment, the output frequency of the point cloud of the lidar 7 in the hydropower station maintenance environment perception module 3 is 20Hz, the output frequency of the image of the binocular depth camera 8 is 30Hz, and the output frequency of the IMU sensor is 20Hz. Multi-source sensor data fusion is performed through the embedded main control board, and an external independent power supply ensures stable operation of the system.

[0040] In the hydropower station maintenance environment perception module 3 designed in this embodiment, the point cloud information collected by the lidar 7 can be used for three-dimensional reconstruction of the scene around the composite grinding robot, the visual image information collected by the binocular depth camera 8 can make up for the weakness of the lack of texture features in the three-dimensional point cloud information, and the IMU sensor 9 can make up for the motion distortion of the point cloud and the image by providing high-frequency motion state data, enhance the robustness of the system, and provide relevant pose constraints for the association of point cloud / visual data.

[0041] The hydropower station maintenance environment perception module 3 designed in this embodiment can independently run the robot environment perception algorithm on the embedded main control board, complete the real-time positioning and map construction for the hydropower maintenance environment, and communicate and interact with the main control system of the composite grinding robot to realize the autonomous positioning and navigation of the composite grinding robot. The collected point cloud information can be used to detect obstacles around the composite grinding robot, enabling the system to have obstacle avoidance function.

[0042] Step 2: Construct a composite grinding robot.

[0043] Specifically, see Figure 1 The constructed hydropower station maintenance environment perception module 3 is installed on an omnidirectional mobile platform 6 equipped with a robotic arm 1, thereby constructing a composite grinding robot. A grinding device 2 is installed at the end of the robotic arm, consisting of an electric grinding head and a six-dimensional force sensor connected by an adapter. In this embodiment, the robotic arm is a six-degree-of-freedom robotic arm.

[0044] Step 3: 3D scene reconstruction and identification of the area to be polished for the cavitation area of ​​the turbine top cover.

[0045] In the hydropower station maintenance environment perception module 3 designed in this embodiment, the robot perception algorithm structure is as follows: Figure 5As shown, the algorithm modules of the hydropower station maintenance environment perception module include a laser odometry module, a visual odometry module, an image frame association module, and an inter-frame map association module. After receiving point cloud data from the lidar, the laser odometry module performs point cloud motion distortion correction and then updates the point-to-plane laser odometry module, thereby adding map points to the global map. After receiving visual image information, the visual odometry module updates its odometry module through the image frame association module. During the global map update process, the inter-frame map association module achieves visual texture rendering of the global map by minimizing photometric errors. The input from the IMU sensor is used for state propagation, providing position information to the laser odometry and visual odometry modules, and constructing constraints for matching point clouds and visual feature points, thereby achieving accurate reconstruction of the global map.

[0046] The environmental perception algorithm proposed in this embodiment projects image pixel information onto a radar point cloud map, achieving a close correlation between camera and radar data. When constructing the image photometric error, the accuracy of visual positioning is improved by minimizing the photometric error between the image frame and the point cloud map. The process of environmental perception and cavitation area identification for the composite grinding robot is as follows: First, point cloud information and visual image information of the scene below the turbine top cover are collected. Based on the adopted robot perception algorithm, such as... Figure 5 As shown, a three-dimensional scene under the top cover is constructed, and visual texture and point cloud information are fused to construct the three-dimensional morphological features of the underside of the top cover. Then, the cavitation area to be polished is identified and segmented by combining visual image information. The area is then finely modeled to determine the amount of material to be removed during polishing of the cavitation area of ​​the top cover. This provides a reference for the planning of subsequent polishing process parameters, avoids manual intervention, and improves polishing efficiency.

[0047] The robot perception algorithm used in this embodiment can run independently within the environment perception module. The environment perception module, via a wired connection, can interact with the robot's main controller for advanced applications of the environmental perception results. Furthermore, the environment perception module includes a memory card that can independently store or download the collected scene data. This data can be used for global map construction of the hydropower maintenance environment, and also for subsequent visual recognition of cavitation areas.

[0048] Step 4: Planning the composite robot grinding path for the cavitation characteristics of the top cover.

[0049] Based on the 3D reconstruction of the cavitation area of ​​the top cover and the identification results of the area to be polished in step three, this embodiment plans and generates the robot polishing path. Considering that the cavitation area mostly occurs near the sealing ring of the top cover, the radial shape of this area can be approximated as a cylindrical surface, and the lower end face of the top cover can be approximated as a plane. Therefore, the approximate curved surface and plane of this area can be fitted based on the point cloud of the cavitation area collected by the robot's environmental perception module. After the surface reconstruction is completed, suitable robot polishing points can be selected on the reconstructed surface, and the polishing points can be connected to obtain the preset polishing path. The preset polishing path is a ZigZag traversal path to ensure that the end polishing head can traverse the surface of the cavitation area of ​​the top cover component.

[0050] Because this mode involves many large curvature angles, it easily causes frequent and abrupt acceleration and deceleration of the robot joints, which can lead to robot vibration and reduce grinding efficiency and quality. Therefore, it is necessary to use a high-order continuity parameter curve (G2 or higher) for path smoothing. In actual grinding, the grinding path needs to be adjusted differently depending on the severity of the cavitation. For shallow cavitation areas (depth ≤ 2mm), a parallel reciprocating path can be used with appropriate spacing. For deep cavitation pits (depth > 2mm), a spiral inward path is used, and the angle of the grinding head needs to be dynamically adjusted during the grinding process.

[0051] Step 5: Construct a contact force model between the robot and the cavitation area to be polished on the top cover.

[0052] In this embodiment, an impedance control model is used to model the dynamic relationship of the contact force between the robot's end effector position and the cavitation area to be polished on the top cover. When the robot comes into contact with the external environment, the impedance control model of the robot's end effector position and the contact force with the environment can be regarded as a physical model composed of mass-spring-damping. In this embodiment, only the robot's end effector contacts the environment in the normal direction. The above model can be expressed as: ; in, , , These represent the robot's inertial parameters, damping parameters, and stiffness parameters, respectively. For the desired contact force, For environmental contact force, For the robot's reference position, This represents the robot's actual position.

[0053] In the robotic polishing process, two environmental contact models exist: a stiffness model and a stiffness-damped model. The stiffness-damped model is more complex than the stiffness model but more closely approximates reality. In this embodiment, the environmental model adopts the stiffness-damped model, which more closely reflects the actual polishing contact situation, and considers the environmental damping term. The environmental model consists of springs and damping elements, and the environmental contact force... It can be represented as: ; in, , These represent the stiffness and damping parameters of the environment, respectively. This represents the position of the robot relative to the surface of the contact environment; it is the equilibrium position where the robot's end effector is exactly free from force. The first derivative represents the robot's actual position. The first derivative represents the robot's reference position. The second derivative representing the robot's actual position. The second derivative represents the robot's reference position. The first derivative represents the contact position.

[0054] Step 6: Combine the contact force model with the robot control law to calculate the steady-state error of force tracking.

[0055] In this embodiment, the desired contact force is set as Environmental contact force is The force tracking error can be calculated. ,set up The reference position trajectory for the robot represents the pre-generated offline position trajectory for executing the planned grinding operation. The position in the positive pressure direction of the workpiece is composed of the offline trajectory position and the output position of the impedance control model, while the positions in other directions are controlled by the offline trajectory. The output position along the normal direction of the workpiece surface is adjusted by the impedance control model, which receives the actual contact force read from the force sensor. and the given desired contact force The output trajectory correction amount is calculated in real time based on the impedance parameters. Robot reference position After the impedance control model calculation, a trajectory correction factor is added. Later became ,available , This represents the robot's commanded position trajectory. In position control mode, the robot's servo motors have high trajectory tracking accuracy; therefore, the robot's commanded position can be considered equal to its actual position. .

[0056] Force tracking error With position correction amount The dynamic relationship between them can be expressed as: ; After applying the Laplace transform, the force tracking error With position correction amount The relationship between them is as follows: ; Combining the impedance control model and the environmental stiffness-damping model, and substituting the controller parameters and environmental parameters into the expression for the force tracking error, we can obtain: ; According to the final value theorem of the Laplace transform, the steady-state error of the system in force tracking is calculated. : ; In the formula, Indicates the desired contact force; Indicates environmental contact force; Indicates force tracking error; The reference position trajectory for the robot represents the pre-generated offline grinding position trajectory used to execute the plan; This represents the stiffness parameter in the impedance control model; Represents the stiffness parameters of the external environment; This indicates the position where the robot just touches the surface of the external environment, which is the equilibrium position where the robot's end effector is just free from force. Represents the Laplace operator; t Indicates time.

[0057] Step 7: Construct an adaptive force control algorithm that incorporates a six-dimensional force sensor to adjust the robot's end effector position in order to control the normal contact force between the robot and the cavitation area to be polished.

[0058] This embodiment uses a six-dimensional force sensor at the robot's end effector to measure the contact force information between the robot and the environment. Combined with the adaptive force control algorithm proposed in this embodiment, the position of the robot's end effector is adjusted to control the normal contact force between the robot and the cavitation area to be polished. The control law of the robot adaptive force control algorithm designed in this embodiment will be introduced next.

[0059] Based on the system's force tracking steady-state error The expression shows that the force tracking steady-state error is satisfied. There are two methods to make it zero: the first method is to set the stiffness parameter of the impedance controller to zero. The second method is to adjust the robot's reference position trajectory. To satisfy ,Right now However, this method requires accurate knowledge of the environmental location and environmental stiffness to calculate the precise reference position trajectory. This allows the force tracking error to converge to zero in steady state. In practical robotic constant force grinding applications, it is difficult to obtain an accurate workpiece surface model, and the changes in the workpiece's position and stiffness are unknown. Therefore, there is a deviation between the calculated reference position trajectory value and the theoretical value, which makes it impossible to eliminate the steady-state error of force tracking.

[0060] To avoid the problems associated with the second method, the adaptive impedance control algorithm designed in this embodiment is based on the first method, setting the controller stiffness parameter... At this point, the system impedance control equation becomes: ; Because the environmental position and stiffness are unknown during constant force grinding, it is difficult to obtain an accurate reference position trajectory for the robot's end effector. Therefore, a reference position trajectory is set. relative to the initial environmental location Similarly, the system impedance control equation then becomes: ; Considering the reference position trajectory relative to the initial environmental location The same, will Substituting into the above formula, we get: ; At this point, the system impedance control equation becomes: ; Considering that the robot's commanded position is equal to its actual position, Substituting the environmental stiffness-damping model selected in this embodiment, the expression for the environmental contact force can be obtained as follows: ; Substituting the above expression for environmental contact force into the system's impedance control model yields: ; As can be seen from the above formula, even with environmental stiffness and environmental damping The unknown can also be solved by selecting appropriate inertial parameters. and damping parameters To satisfy the above equation, thereby reducing the system's force tracking steady-state error. The value is zero. Assuming the external environment is an ideal plane and there are no deformations, then... , At this point, we can obtain: ; When the system is in a stable state, there exists ,but This holds true consistently, and the system's force tracking error remains zero. However, in actual grinding processes, the surface position of the workpiece being ground is dynamically changing, and the environmental position... It is a time-varying function, and at this time it exists ,or and Let the estimated value of the environmental location deformation be... The following equation holds true: ; Estimated value of position correction Satisfy the following formula: ; Estimated value of environmental location deformation Substituting into the system impedance control equation, we get: ; It can be observed that, due to and Both are time-varying functions, and force tracking error always exists. In order to incorporate the force tracking error term... To eliminate and ensure system stability, this embodiment introduces a damping parameter adjustment term. The damping parameter is adjusted based on the force deviation information and the first-order difference information of the force deviation. The adaptive impedance control equation designed in this embodiment can be written as: ; It can be adjusted and To compensate for errors caused by changes in the external environment, its form is as follows: ; in, The damping parameters are adjusted according to the force deviation. The damping parameters are adjusted based on the first-order difference of the force deviation. and Corresponding to and The amount of compensation, Representation function The update coefficient, Representation function The update coefficient. This represents the time of each sampling period. Let be a constant close to zero. To avoid the impedance control parameter tending to infinity when the denominator is zero, the adaptive impedance control method proposed in this embodiment introduces an improved impedance control parameter update coefficient. By incorporating the differential information of force deviation into the adjustment of damping parameters, the force overshoot during system response is reduced, thereby improving the dynamic performance of the constant force tracking algorithm. The control algorithm block diagram of the adaptive impedance control method proposed in this embodiment is shown below. Figure 6 As shown.

[0061] To facilitate the deployment of the adaptive impedance control method proposed in this embodiment in the robot controller, the adaptive impedance control law can be converted into a discrete form, the general form of which can be expressed as: ; In the formula, express t The acceleration of the robot's end effector at any given moment; express t The second derivative of the external environment position at time; Represents the inertial parameters in the impedance control model; This represents the proportional parameter in the PD control law; express t The expected contact force at any moment; express t The contact force between the robot's end effector and the environment at any given moment; This represents the stiffness parameter in the impedance control model; express t The first derivative of the expected contact force at all times; express t The first derivative of the contact force between the robot's end effector and the environment at any given moment; This represents the damping parameter in the impedance control model; This is the first damping parameter in the designed control law that is adjusted based on the first-order difference of the force deviation; This is the second damping parameter in the designed control law, adjusted based on the first-order difference of the force deviation. express t The speed of the robot's end effector at time -1; express t The first derivative of the external environment position at a given time; express t The speed of commands given at the robot's end effector at any given moment; This represents the system communication cycle between the robot controller and the servo driver; express t The position of the robot's end effector at any given moment; express t The command position of the robot's end effector at time -1.

[0062] Step 8: Pre-experiment and parameter simulation tuning of robot adaptive force control grinding.

[0063] Before conducting formal robotic grinding of cavitation areas, it is necessary to first tune the relevant control parameters of the adaptive force control algorithm proposed in this embodiment, and select appropriate key process parameters such as grinding head speed, normal contact force, and grinding head material and type. Therefore, this embodiment designs a robotic grinding pre-experiment. In the pre-experiment, the robotic grinding head is not started initially. After setting the initial impedance control parameters, the robotic arm is controlled to conduct a fixed-point force control experiment in the cavitation area of ​​the top cover using a trial-and-error method. Force and position information during the dynamic contact process between the robot end effector and the workpiece are collected. Multiple contact points are selected, the above experiment is repeated, and data is recorded. Then, the robotic arm end effector is controlled to move at a constant speed on the cavitation area of ​​the top cover, and relevant force and position information is collected. After completing the above force contact pre-experiment, the collected information is imported into data analysis and processing software to construct a simulation model, thereby completing system identification and estimating the stiffness and damping parameters of the environment. Based on the known information of environmental stiffness and damping, the robot controller parameters can be initially adjusted. First, initial robot grinding process parameters are set. Then, the robot grinding head is started to conduct a constant force control experiment, selecting multiple areas for trial grinding. The robot grinding process parameters are then adjusted to select suitable parameters. After determining the process parameters, the previous robot force contact experiment is repeated, recording the force and displacement information of the robot end effector. This information is then exported and used for system analysis and simulation again to determine the optimal robot controller and impedance control parameters. In this embodiment, a set of optimal control parameters is used to illustrate the simulated contact force variation between the robot end effector grinding head and the workpiece being ground. Figure 7 As shown.

[0064] Step 9: Real-time adaptive constant force closed-loop feedback control of the robot for cavitation zones.

[0065] First, based on the constructed robot environment perception module, scene features of the area to be polished on the turbine top cover are collected to complete the path planning and intelligent obstacle avoidance of the composite polishing robot. Then, using TwinCAT software, the adaptive force control algorithm designed in step five is compiled into a file that can be directly executed by the robot hardware and burned into the robot controller. In this embodiment, Beckhoff industrial PC is selected as the robot controller. Based on the Beckhoff real-time kernel, the PLC cycle time is set to 1ms, thereby executing the robot end force control algorithm at a frequency of 1kHz. The communication frequency of the robot end six-dimensional force sensor is also set to 1kHz. In the algorithm execution layer designed in this embodiment, signals such as the position, speed, current of the execution layer servo motor and the robot end force sensing information are read based on the EtherCAT bus and input into the module compiled by the adaptive control algorithm in step five for calling, calculation and execution. The obtained position command output can be sent to the servo motor for execution through the real-time EtherCAT bus. In this embodiment, the actual contact force variation between the robot end polishing head and the workpiece under a set of optimal control parameters is shown in the figure. Figure 8 As shown.

[0066] Example 2: See Figure 1 A robotic grinding system for cavitation areas of a turbine roof is provided to implement the aforementioned robotic grinding control method for cavitation areas of a turbine roof. The robotic grinding system includes: A robotic arm, the end of which is used to mount a grinding device 2; A lifting platform 5, wherein a robotic arm 1 is mounted on the top of the lifting platform 5 for adjusting the height of the robotic arm 1; An omnidirectional mobile platform 6 is provided, with a lifting platform 5 mounted on its top. The omnidirectional mobile platform 6 is used for moving the lifting platform 5 and the robotic arm 1. The lifting device 4 is mounted on the omnidirectional moving platform 6 at its lower end. A driver is mounted on the top of the lifting device 4. A hydropower station maintenance environment perception module 3 is mounted on the power output end of the driver. The driver is used to drive the hydropower station maintenance environment perception module 3 to rotate. The hydropower station maintenance environment perception module 3 is used to perceive the surrounding environment of the composite grinding robot. By integrating three-dimensional point cloud and visual texture information, it completes the scene reconstruction of the hydropower station maintenance area and the identification of the cavitation area to be ground.

[0067] Specifically, in this embodiment, a self-built six-degree-of-freedom robotic arm is used as the processing execution mechanism of the composite grinding robot, such as... Figure 1As shown, all joints employ a rotary structure design. The drive system uses Panasonic MINAS A6 series servo motors and matching drivers, and a Beckhoff industrial controller is used to achieve precise motion control. Based on the controller's real-time PLC core, the control system can acquire status information such as the position, speed, and current of the servo motors at a frequency of 1kHz and issue corresponding control commands. The hardware control system structure of this six-degree-of-freedom robotic arm is as follows. Figure 2 As shown. The robotic arm's end effector uses a standardized interface, allowing for the quick installation or replacement of modular tools such as grinding heads and laser cladding nozzles.

[0068] In this embodiment, an omnidirectional moving platform 6 is integrated below the base of the six-degree-of-freedom robotic arm, such as... Figure 2 As shown, the mobile platform communicates with the main control system using the EtherCAT communication protocol, and can feed back the position / velocity / attitude data of the mobile platform at a frequency of 1kHz, while synchronously receiving control commands. This design enables the dynamic expansion of the workspace of the grinding robot.

[0069] The grinding tool used in this embodiment is a Bosch electric grinding head. The maximum no-load speed can reach 15000 r / min, and the selected grinding head is an 80-grit flap wheel grinding head.

[0070] The lifting device 4 can be a servo electric cylinder with the telescopic end of the servo electric cylinder facing upwards. The driver is installed on the top of the telescopic end of the servo electric cylinder and can be a servo motor.

[0071] The multimodal sensors selected in this embodiment are the Livox MID-360 lidar, the Realsense D435i binocular depth camera, and the Wheeltec N300 inertial measurement unit (IMU), and the corresponding control board is the RK3588 embedded main control board.

[0072] The hydropower station maintenance environment perception module 3 can independently run the robot's environment perception algorithm on the embedded board, completing real-time localization and map construction for the hydropower maintenance environment. It also communicates and interacts with the main control system of the composite grinding robot, enabling the robot's autonomous localization and navigation. The collected point cloud information can be used to detect obstacles around the composite grinding robot, giving the system obstacle avoidance capabilities. After scene reconstruction, the environment perception module can autonomously plan its moving processing path based on obstacle information below the turbine top cover, achieving full coverage processing of the cavitation area to be ground.

[0073] The hydropower station maintenance environment perception module 3 designed in this invention can be used not only in the grinding and processing of the cavitation area below the turbine top cover, but also in other hydropower station unit maintenance scenarios, such as the reconstruction of scenarios like the runner chamber, wind tunnel, lower frame, and stator base. Figure 9 As shown.

[0074] At the system level, this invention aims to address the lack of omnidirectional mobile processing capability in existing turbine roof cavitation area grinding robots. It designs a composite grinding robot integrating an omnidirectional mobile platform and a six-degree-of-freedom robotic arm, and performs collaborative motion planning for the robotic arm and omnidirectional mobile platform to achieve autonomous navigation and obstacle avoidance planning functions, thereby completing uninterrupted coverage processing on the large-size turbine roof surface to be ground.

[0075] At the perception level, this invention aims to solve the challenges of accurate perception and detailed reconstruction in complex hydropower maintenance environments. It overcomes the technical difficulties of visual sensing information lacking three-dimensional features, point cloud feedback information lacking semantic texture, and single sensors failing to reflect the full picture of the site environment. The invention achieves data fusion from multiple sources and modal sensors (LiDAR, vision, and IMU), thereby enabling high-precision perception of turbine roof maintenance scenarios and identification of cavitation feature areas. Furthermore, this invention addresses the shortcomings of existing robot environmental perception devices, such as weak versatility, low integration, and lack of plug-and-play functionality. This invention designs a high-precision environmental perception device with high multimodal sensor integration, compatibility with multiple communication protocols, and good modularity. It can be used not only in turbine roof cavitation area repair scenarios but also transferred to other mobile robot hydropower maintenance application scenarios without requiring repeated algorithm development, thus achieving accurate and reliable maintenance scenario perception and robot autonomous positioning and navigation.

[0076] At the force control algorithm level, this invention aims to address the problem of non-convergence of force tracking error during force-controlled grinding by the robot. To avoid environmental stiffness and position estimation, this invention proposes to employ variable damping control. Considering the limited adjustment capability of a single damping parameter, which can easily lead to force overshoot or oscillation during control, this invention uses dual damping parameters to enhance the force control adjustment capability of the algorithm, achieving constant force adaptive adjustment between the robot's end-effector grinding head and the top cover curved surface, thus ensuring the processing quality when the robot grinds the cavitation area of ​​the top cover.

[0077] It is understood that the various numerical designations used in the embodiments of this application are merely for the convenience of description and are not intended to limit the scope of the embodiments of this application.

[0078] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

[0079] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the invention. Modifications and variations made by those skilled in the art in accordance with the spirit of the invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A robotic grinding control method for cavitation areas of a water turbine roof, characterized in that, Includes the following steps, Step 1: Construct a multimodal information fusion module for the hydropower station maintenance environment perception module; Step 2: Construct a composite grinding robot; Step 3: 3D scene reconstruction and identification of the area to be polished for the cavitation area of ​​the turbine top cover; Step 4: Composite robot grinding path planning for cavitation characteristics of the top cover; Step 5: Construct a contact force model between the robot and the cavitation area to be polished on the top cover; Step Six: Combine the contact force model and the robot control law to calculate the steady-state error of force tracking; Step 7: Construct an adaptive force control algorithm that combines a six-dimensional force sensor to adjust the position of the robot's end effector in order to control the normal contact force between the robot and the cavitation area to be polished. Step 8: Pre-experiment and parameter simulation tuning of robot adaptive force control grinding; Step 9: Real-time adaptive constant force closed-loop feedback control of the robot for cavitation zones.

2. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step one, the hydropower station maintenance environment perception module includes a lidar, a binocular depth camera, and an IMU sensor. By fusing point cloud information collected by the lidar, visual and depth map information collected by the binocular depth camera, and multi-axis acceleration / angular velocity information collected by the IMU sensor, it perceives the surrounding environment of the composite grinding robot. The integrated 3D point cloud and visual texture information is used for scene reconstruction of the hydropower station maintenance area and identification of the cavitation area to be ground. Among them, the visual image information collected by the binocular depth camera is used to compensate for the texture features of the 3D point cloud information, and the IMU sensor compensates for the motion distortion of the point cloud and image by providing high-frequency motion state data.

3. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step two, the constructed hydropower station maintenance environment perception module is installed on an omnidirectional mobile platform equipped with a robotic arm, thereby constructing a composite grinding robot. The end of the robotic arm is equipped with a grinding device, which consists of an electric grinding head and a six-dimensional force sensor connected by an adapter.

4. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 2, characterized in that, The algorithm module of the hydropower station maintenance environment perception module includes a laser odometer module, a visual odometer module, an image frame association module, and an inter-frame map association module. In step three, after receiving the point cloud data input from the lidar, the laser odometer module performs point cloud motion distortion correction and then updates the point-to-plane laser odometer module, thereby adding map points to the global map. After receiving the visual image information input, the visual odometer module updates the visual odometer module through the image frame association module. During the global map update process, the inter-frame map association module achieves visual texture rendering of the global map by minimizing photometric errors. The input from the IMU sensor is used for state propagation, providing position information for the laser odometer module and the visual odometer module, and constructing constraints for matching point clouds and visual feature points, thereby realizing the 3D scene reconstruction of the global map.

5. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 4, characterized in that, In step three, the identification of the area to be polished includes the following steps: 1) Collect point cloud information and visual image information of the scene below the turbine top cover, construct a three-dimensional scene below the top cover, and fuse visual texture and point cloud information to construct the three-dimensional morphological features of the lower surface of the top cover. 2) Combine visual image information to identify and segment the cavitation area to be polished, perform fine modeling of the area to be polished, determine the amount of material to be removed from the cavitation part of the top cover, and provide a reference for the planning of subsequent polishing process parameters.

6. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step four, the composite robot grinding path planning for the cavitation characteristics of the top cover includes the following steps: 1) Based on the point cloud of the cavitation area collected by the robot's environmental perception module, fit the approximate curved surface and plane of the cavitation area; 2) After completing the surface reconstruction, select robot polishing points on the reconstructed surface and connect the polishing points to obtain the preset polishing path; wherein, the preset polishing path is a reciprocating traversal path. 3) Smooth the preset polishing path; 4) During the actual grinding process, the grinding path is adjusted differently according to the severity of the cavitation. For shallow cavitation areas, i.e., areas with cavitation depth ≤ 2mm, a parallel reciprocating path is used, and the spacing is selected based on the geometric characteristics of the cavitation area and the grinding process requirements. For deep cavitation pit areas, i.e. areas with cavitation depth > 2mm, a spiral inward path is used, and the angle of the grinding head needs to be dynamically adjusted during the grinding process.

7. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step five, an impedance control model is used to model the dynamic relationship of the contact force between the robot's end effector position and the cavitation area to be polished on the top cover. The model expression is as follows: ； In the formula, , , These represent the inertial parameters, damping parameters, and stiffness parameters in the impedance control model exhibited when the robot's end effector comes into contact with the external environment, respectively. For the desired contact force, For environmental contact force, For the robot's reference position, This refers to the robot's actual position. Environmental contact force The expression is: ； In the formula, , These represent the stiffness and damping parameters of the external environment, respectively. This indicates the position where the robot contacts the surface of the external environment; it is the equilibrium position where the robot's end effector is just free from force. The first derivative represents the robot's actual position. The first derivative represents the robot's reference position. The second derivative representing the robot's actual position. The second derivative represents the robot's reference position. The first derivative represents the contact position.

8. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step six, force tracking steady-state error The expression is: ； In the formula, Indicates the desired contact force; Indicates environmental contact force; Indicates force tracking error; The reference position trajectory for the robot represents the pre-generated offline grinding position trajectory used to execute the plan; This represents the stiffness parameter in the impedance control model; Represents the stiffness parameters of the external environment; This indicates the position where the robot just touches the surface of the external environment, which is the equilibrium position where the robot's end effector is just free from force. Represents the Laplace operator; t Indicates time.

9. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step seven, the adaptive force control algorithm expression based on the six-dimensional force sensor is as follows: ； In the formula, express t The acceleration of the robot's end effector at any given moment; express t The second derivative of the external environment position at time; Represents the inertial parameters in the impedance control model; This represents the proportional parameter in the PD control law; express t The expected contact force at any moment; express t The contact force between the robot's end effector and the environment at any given moment; This represents the stiffness parameter in the impedance control model; express t The first derivative of the expected contact force at all times; express t The first derivative of the contact force between the robot's end effector and the environment at any given moment; This represents the damping parameter in the impedance control model; This is the first damping parameter in the designed control law that is adjusted based on the first-order difference of the force deviation; This is the second damping parameter in the designed control law, adjusted based on the first-order difference of the force deviation. express t The speed of the robot's end effector at time -1; express t The first derivative of the external environment position at a given time; express t The speed of commands given at the robot's end effector at any given moment; This represents the system communication cycle between the robot controller and the servo driver; express t The position of the robot's end effector at any given moment; express t The command position of the robot's end effector at time -1.

10. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step eight, the robot adaptive force control grinding pre-experiment and parameter simulation tuning includes the following steps: 1) In the pre-grinding experiment, the robot grinding head was not started first. After setting the initial impedance control parameters, the robotic arm was controlled to conduct a fixed-point force control experiment in the cavitation area of ​​the top cover by means of trial contact. Force and position information during the dynamic contact process between the robot end and the top cover were collected. Multiple contact points were selected, the above experiment was repeated and the data were recorded. 2) Control the end effector of the robotic arm to move at a constant speed over the cavitation area of ​​the top cover, and collect relevant force and position information; 3) After completing the above force contact pre-experiment, the collected information is imported into data analysis and processing software to build a simulation model, thereby completing system identification and estimating the stiffness and damping parameters of the environment. 4) Based on the known information of environmental stiffness and damping, the robot controller parameters are initially adjusted. First, the initial robot grinding process parameters are set, then the robot grinding head is started to conduct a constant force control experiment. Multiple areas are selected for trial grinding, and then the robot grinding process parameters are adjusted to select the corresponding process parameters. 5) After determining the process parameters, repeat the previous robot force contact experiment, record the force and displacement information of the robot end, export it, and perform system analysis and simulation again to determine the optimal robot controller and impedance control parameters.

11. The robotic grinding control method for cavitation zone of a turbine top cover according to claim 1, characterized in that, In step nine, the robot's adaptive constant force closed-loop feedback real-time control for cavitation zones includes the following steps: 1) Based on the constructed robot environment perception module, the scene features of the area to be polished on the top cover of the water turbine are collected to complete the path planning and intelligent obstacle avoidance of the composite polishing robot; 2) Compile the adaptive force control algorithm built in step seven into a file that can be directly executed by the robot hardware, and burn it into the robot controller; 3) In the algorithm execution layer, the position, speed, current and robot end force sensing information signals of the servo motor are read based on the EtherCAT bus, and the information signals are input into the module compiled by the adaptive control algorithm for calling, calculation and execution. The obtained position command output is sent to the servo motor for execution through the real-time EtherCAT bus.

12. A robotic grinding system for cavitation areas of a turbine roof, used to implement the robotic grinding control method for cavitation areas of a turbine roof as described in any one of claims 1 to 11, wherein the robotic grinding system comprises: A robotic arm (1), the end of which is used to mount a grinding device (2); A lifting platform (5) is provided with a mechanical arm (1) installed on its top for adjusting the height of the mechanical arm (1); An omnidirectional mobile platform (6) is provided, with a lifting platform (5) mounted on its top. The omnidirectional mobile platform (6) is used to move the load lifting platform (5) and the robotic arm (1). The lower end of the lifting device (4) is installed on the omnidirectional moving platform (6), and the top of the lifting device (4) is equipped with a driver. The power output end of the driver is equipped with a hydropower station maintenance environment perception module (3). The driver is used to drive the hydropower station maintenance environment perception module (3) to rotate. The hydropower station maintenance environment perception module (3) is used to perceive the surrounding environment of the composite grinding robot, integrate three-dimensional point cloud and visual texture information, and complete the scene reconstruction of the hydropower station maintenance area and the identification of the cavitation area to be ground.

Images