Intelligent wind power tower cylinder outer wall polishing system and method

Abstract

The application discloses an intelligent wind power tower drum outer wall polishing system and method, and relates to the technical field of outer wall polishing.The system comprises the following steps: determining a surface topography topology map; generating a preset outer wall polishing path, combining the hardness of the tower drum material with the grinding tool wear coefficient to determine the grinding head pressure-rotation speed matching parameters; introducing mechanical arm pose data, combining the tower drum outer wall corrosion depth data to determine the first error compensation data; through the wind power tower drum outer wall polishing task, combining the tower drum outer wall weld to determine the second error compensation data; connecting multiple mechanical arms for collaborative work, combining the first error compensation data and the second error compensation data to configure a master-slave control architecture, and performing outer wall polishing multi-end collaborative control on the target wind power tower drum by the master-slave control architecture.The application solves the technical problem that the prior art cannot realize high-precision collaborative polishing of the wind power tower drum outer wall under complex topography, and achieves the technical effect of improving the wind power tower drum outer wall polishing precision and operation consistency.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of outer wall polishing, and in particular to an intelligent wind power tower drum outer wall polishing system and method. BACKGROUND

[0002] The wind power tower drum outer wall is prone to rust, uneven welds and other complex surface defects in the long-term operation process. Traditional polishing operations are usually completed by manual work or single machine mechanical equipment, and it is difficult to accurately adapt to the changes in the appearance of different regions, especially in the face of irregular rust depth and weld protrusions. The polishing track is unstable, and the grinding pressure control is not accurate, resulting in uneven surface treatment quality, which is difficult to meet the maintenance requirements of high-quality wind power structural parts. SUMMARY

[0003] The application provides an intelligent wind power tower drum outer wall polishing system and method, which is used to solve the technical problem that the prior art cannot realize high-precision collaborative polishing of the wind power tower drum outer wall under complex topography.

[0004] In view of the above problems, the application provides an intelligent wind power tower drum outer wall polishing system and method.

[0005] The first aspect of the application provides an intelligent wind power tower drum outer wall polishing system, which comprises:

[0006] The mapping module is used to map the surface topography topology of the tower drum outer wall weld by the surface feature of the target wind power tower drum, and the parameter determination module is used to generate a preset outer wall polishing path associated with the wind power tower drum outer wall polishing task according to the surface topography topology, and to determine the matching parameters of the grinding head pressure and speed in combination with the tower drum material hardness and the grinding tool wear coefficient. The first error compensation data determination module is used to introduce the mechanical arm pose data dynamically adjusted by the fuzzy PID controller, and to determine the first error compensation data by combining the preset outer wall polishing path to compensate the tower drum outer wall rust depth data in the surface topography topology. The second error compensation data determination module is used to compensate the tower drum outer wall weld in the surface topography topology by combining the preset outer wall polishing path through the wind power tower drum outer wall polishing task, and to determine the second error compensation data. The cooperative control module is used to connect multiple mechanical arms for collaborative work, to configure a master-slave control architecture in combination with the first error compensation data and the second error compensation data, and to perform multi-end collaborative control on the outer wall polishing of the target wind power tower drum by using the master-slave control architecture.

[0007] The second aspect of the application provides an intelligent wind power tower drum outer wall polishing method, which comprises:

[0008] By the surface morphology characteristics of the target wind power tower, a surface topography topological map mapping the corrosion depth data of the tower outer wall and the tower outer wall weld is drafted; according to the surface topography topological map, a preset outer wall polishing path associated with the wind power tower outer wall polishing task is generated, and the matching parameters of the grinding head pressure and speed are determined simultaneously in combination with the hardness of the tower material and the grinding tool wear coefficient; the mechanical arm pose data dynamically adjusted by the fuzzy PID controller is introduced, the corrosion depth data of the tower outer wall in the surface topography topological map is compensated in combination with the preset outer wall polishing path, and the first error compensation data is determined; the tower outer wall weld in the surface topography topological map is compensated in combination with the preset outer wall polishing path through the wind power tower outer wall polishing task, and the second error compensation data is determined; the multiple mechanical arms in collaborative work are connected, the master-slave control architecture is configured in combination with the first error compensation data and the second error compensation data, and the target wind power tower is subjected to outer wall polishing multi-end collaborative control by the master-slave control architecture.

[0009] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0010] By the surface morphology characteristics of the target wind power tower, a surface topography topological map mapping the corrosion depth data of the tower outer wall and the tower outer wall weld is drafted; according to the surface topography topological map, a preset outer wall polishing path associated with the wind power tower outer wall polishing task is generated, and the matching parameters of the grinding head pressure and speed are determined simultaneously in combination with the hardness of the tower material and the grinding tool wear coefficient; the mechanical arm pose data dynamically adjusted by the fuzzy PID controller is introduced, the corrosion depth data of the tower outer wall in the surface topography topological map is compensated in combination with the preset outer wall polishing path, and the first error compensation data is determined; the tower outer wall weld in the surface topography topological map is compensated in combination with the preset outer wall polishing path through the wind power tower outer wall polishing task, and the second error compensation data is determined; the multiple mechanical arms in collaborative work are connected, the master-slave control architecture is configured in combination with the first error compensation data and the second error compensation data, and the target wind power tower is subjected to outer wall polishing multi-end collaborative control by the master-slave control architecture. The present application solves the technical problem that the existing technology cannot realize high-precision collaborative polishing of the wind power tower outer wall under complex topography, and achieves the technical effect of improving the polishing precision and work consistency of the wind power tower outer wall by introducing a topological map fitting and multi-arm error compensation collaborative control mechanism. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0012] Figure 1 A schematic diagram of an intelligent wind turbine tower outer wall grinding system provided in this application embodiment;

[0013] Figure 2 This is a schematic diagram of an intelligent wind turbine tower outer wall grinding method provided in an embodiment of this application.

[0014] Explanation of reference numerals in the attached figures: 11 for proposed mapping module, 12 for parameter determination module, 13 for first error compensation data determination module, 14 for second error compensation data determination module, and 15 for cooperative control module. Detailed Implementation

[0015] This application provides an intelligent wind turbine tower outer wall grinding system and method, which addresses the technical problem that existing technologies cannot achieve high-precision collaborative grinding of wind turbine tower outer walls under complex morphologies. By introducing a topology graph fitting and multi-arm error compensation collaborative control mechanism, the technical effect of improving the grinding accuracy and operational consistency of wind turbine tower outer walls is achieved.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0018] Example 1, as Figure 1 As shown in the figure, this application provides an intelligent wind turbine tower outer wall grinding system, which includes:

[0019] The mapping module 11 is used to generate a surface topology map of the tower outer wall corrosion depth data and the tower outer wall weld seam based on the surface morphology features of the target wind turbine tower.

[0020] In the embodiment of the present application, the mapping module 11 constructs a surface topography topology map containing corrosion depth data and weld geometry based on the actual surface morphology characteristics of the outer wall of the wind turbine tower. Specifically, first, a three-dimensional laser radar (LiDAR) system is used to scan the outer wall of the tower circumferentially to obtain high-density point cloud data, capturing its macroscopic contour and microscopic undulation information; then a structured light vision sensor is introduced to collect high-resolution grayscale images, and image feature extraction algorithms (such as SIFT or ORB) are used to extract features of surface cracks, corrosion boundaries and weld textures. Through the registration method of point cloud and image (such as multi-modal alignment based on ICP), the fusion of geometric data and texture information is realized. For the corrosion area, the local surface reconstruction algorithm combined with the depth profile difference technique is used to quantitatively analyze the Z-axis depth variation of the recessed area, and the corrosion depth data is extracted; for the weld area, the curvature extraction and high convex area identification method is used to reconstruct the weld continuous surface and its boundary features. Finally, the corrosion depth distribution and weld topography parameters are mapped to a global coordinate system, and a surface topography topology map is constructed.

[0021] The parameter determination module 12 is configured to generate a preset outer wall polishing path associated with the wind turbine tower outer wall polishing task according to the surface topography topology map, and simultaneously determine the matching parameters of the polishing head pressure and speed in combination with the tower material hardness and the abrasive tool wear coefficient.

[0022] In the embodiment of the present application, the parameter determination module 12 first identifies the wind turbine tower outer wall polishing task, extracts the key process requirements defined in the task, including the target polishing area (such as the corrosion concentration area, the weld repair area), the required surface treatment level (such as the roughness level Ra), the processing priority order and the overall path coverage. The polishing task not only limits the spatial boundary of path planning, but also determines the path density, trajectory direction and control target for path adaptability, thereby constituting the constraint conditions and optimization targets of subsequent path planning.

[0023] Then based on the three-dimensional point cloud data in the surface topography topology map, the cylindrical outer wall of the wind turbine tower is converted into a two-dimensional expanded plane by using the grid expansion method, and by analyzing the corrosion depth distribution, weld height fluctuation and local curvature change in the topology map, a partition algorithm is used to construct multiple polishing areas with different processing requirements. The division of polishing areas is strictly guided by the processing requirements in the polishing task, realizing the fusion of task constraints and geometric characteristics.

[0024] Based on the grinding area, a variable pitch spiral path planning method is used to generate a preset outer wall grinding path associated with the grinding task. During the specific path generation process, the path pitch is automatically adjusted according to the process requirements of different areas. For example, the path density is increased in severely corroded areas to enhance the cleaning effect, the path guide in the weld transition area follows the weld direction to avoid edge weakening, and the path spacing is increased in ordinary substrate areas to improve grinding efficiency. To ensure trajectory continuity and fit, a curvature-guided interpolation method is introduced to smooth the path, making the path adaptable to complex curved surfaces and improving operational stability.

[0025] After completing the path planning, based on the tower structure material type, the corresponding tower material hardness (such as Brinell hardness HB) is extracted from the material performance database, and the grinding wheel wear coefficient corresponding to the currently used abrasive (such as alumina, ceramic grinding belt) is obtained from the grinding wheel database.

[0026] Finally, the preset outer wall grinding path, tower material hardness, and grinding wheel wear coefficient are input together, and the grinding process parameter matching database is called to query the grinding head contact pressure and spindle speed corresponding to various grinding areas. If no completely matching combination is found in the search results, linear interpolation is used to interpolate between existing samples to fill in the parameter gaps, thereby simultaneously determining the grinding head pressure-speed matching parameters.

[0027] The first error compensation data determination module 13 is used to introduce the robot arm pose data dynamically adjusted by the fuzzy PID controller, and combine it with the tower outer wall corrosion depth data in the surface topology diagram to determine the first error compensation data.

[0028] In this embodiment, the first error compensation data determination module 13 introduces the robotic arm pose data dynamically adjusted by a fuzzy PID controller, and combines it with the tower outer wall rust depth data in the preset outer wall grinding path compensation surface topology map to determine the first error compensation data. Specifically, firstly, using rotary encoders installed at each joint of the robotic arm and the end effector inertial measurement unit (IMU), the current attitude, position, and angle information of each axis of the robotic arm are obtained through forward kinematics calculation, forming continuous robotic arm pose data that reflects the actual movement trajectory of the grinding head in real time. To dynamically adapt to the height changes of the rusted area surface, the operating state of the robotic arm is adjusted by a fuzzy PID controller. Based on the input trajectory error and error change rate, the controller constructs a fuzzy inference process using a fuzzy rule base, dynamically adjusting the proportional, integral, and derivative parameters to make the control commands smoother and more flexible to adapt to the surface undulations of the rusted area, enhancing tracking stability.

[0029] Then the adjusted mechanical arm end trajectory is compared with the preset outer wall polishing path point by point, the Euclidean distance measurement method is used to calculate the error value of each corresponding point in space, and an error set containing all position offset information is constructed. In order to eliminate the actual polishing deviation caused by the change of corrosion depth, the topological mapping method is introduced, and each space error point is corresponded with the corrosion depth data in the surface topological map. For the error corresponding point, the cubic spline interpolation method is used to compensate the corrosion depth value, and the trajectory offset is mapped to the local corrosion depth correction, so as to realize the high-precision compensation repair of the corrosion area in the topological map.

[0030] Through the above steps, the first error compensation data is finally formed according to the adjusted mechanical arm pose data, the fuzzy PID control amount and the corrected corrosion depth compensation data.

[0031] The second error compensation data determination module 14 is used to compensate the weld of the outer wall of the tower drum in the surface topography topological map through the wind power tower drum outer wall polishing task, and determine the second error compensation data in combination with the preset outer wall polishing path.

[0032] In the embodiments of the present application, the second error compensation data determination module 14 is used to compensate the weld of the outer wall of the tower drum in the surface topography topological map through the wind power tower drum outer wall polishing task, and determine the second error compensation data in combination with the preset outer wall polishing path. In the implementation process, the actual contour change around the weld surface and the path execution deviation are combined with the process target of the polishing task, and steps such as weld feature extraction, path deviation calculation, position and attitude adjustment and compensation data generation are adopted.

[0033] Specifically, according to the weld processing requirements defined in the wind power tower drum outer wall polishing task, including weld excess height control (such as keeping 11.5mm), weld polishing width range (such as 10~20mm) and weld direction keeping consistent, the specified area to be polished is extracted. Then the three-dimensional point cloud data in the surface topography topological map is called, and the weld center line is identified by using the method based on curvature calculation. Specifically, the main curvature value of each point is calculated to extract continuous high curvature line segments from the tower drum surface point cloud; then the two side boundaries of the weld are determined by using the expansion method based on neighborhood density, and the contour range and height information of the weld are formed.

[0034] Then the preset outer wall polishing path corresponding to the weld area is loaded, and the trajectory points in the path are projected to the weld area, and the nearest neighbor matching method is used to obtain the corresponding position of each path point in the weld surface. By calculating the Euclidean distance between each trajectory point and its projection point in the weld area, the transverse offset error of the path is obtained; at the same time, the angle between the path line direction and the weld direction is analyzed to obtain the attitude deviation information.

[0035] To correct the error, a normal-guided path adjustment method is introduced, the normal direction of each point on the weld surface is calculated, and the normal direction is taken as the basis for path adjustment direction, the position coordinates of each path point are repositioned to fit the weld surface; At the same time, according to the weld contour height data, the feed depth of the grinding head is dynamically adjusted, for example, the feed depth is reduced in the area with large weld excess height, and the feed depth is appropriately increased in the weld thinning area, to maintain the uniformity of grinding.

[0036] Considering that the polishing path is prone to trajectory mutation at the start and end of the weld, a curve continuity processing method is used for these edge path segments, and a spline interpolation method is used to smooth the path point sequence, so that the transition between the weld path and the overall path is natural, and the problem of trajectory jump or path discontinuity is avoided.

[0037] Finally, the spatial correction value (such as coordinate adjustment along the normal direction) of each path point in the weld area, the path direction angle fine adjustment value (attitude adjustment), and the feed depth adjustment value of the corresponding position are integrated to form the second error compensation data.

[0038] The cooperation control module 15 is used for connecting multiple mechanical arms for collaborative work, combining the first error compensation data and the second error compensation data, configuring a master-slave control architecture, and performing multi-end collaborative control on the outer wall polishing of the target wind power tower drum based on the master-slave control architecture.

[0039] In the embodiments of the present application, the cooperation control module 15 connects multiple mechanical arms for collaborative work, combines the first error compensation data and the second error compensation data, and configures a master-slave control architecture. Specifically, based on the preset outer wall polishing path, the first error compensation data for rust area trajectory compensation and the second error compensation data for weld area attitude correction are combined to dynamically optimize the path structure, set multiple groups of variable pitch spiral polishing paths adapted to complex outer wall structures, and record key parameters related to polishing quality on path nodes, including surface roughness Ra value, rust residual rate and weld flatness. After the polishing path and its surface feature parameters are determined, the task division mechanism among multiple mechanical arms is established by comprehensively analyzing the path density and processing difficulty of each area based on the requirements of work division, synchronization efficiency and processing accuracy in the wind power tower drum outer wall polishing task. Based on this, a master-slave control architecture is configured, one mechanical arm is set as a master unit responsible for overall path beat, task scheduling and synchronization signal issuing, and the remaining mechanical arms are set as slave units responsible for executing tasks in different areas, and based on the error compensation and surface state parameters recorded in their respective paths, trajectory correction and synchronous response are performed.

[0040] The final cooperative control module 15 performs multi-end cooperative control of the outer wall polishing of the target wind tower through the master-slave control architecture configured above, so that the multiple mechanical arms realize dynamic task allocation, error adaptive adjustment and motion behavior linkage in different polishing areas, thereby ensuring high efficiency, high precision and high consistency of the entire wind tower outer wall polishing operation under complex surface conditions.

[0041] Further, the system provided by the application embodiment is further used for:

[0042] Based on the preset outer wall polishing path, the first error compensation data and the second error compensation data are combined to set a variable pitch spiral polishing path, and the surface roughness, the rust residual rate and the weld flatness are recorded; based on the wind tower outer wall polishing task, the surface roughness, the rust residual rate and the weld flatness are combined to establish a master-slave control architecture corresponding to the cooperative operation of the multiple mechanical arms.

[0043] In the application embodiment, in the process of performing the multi-mechanical arm cooperative polishing control of the wind tower outer wall, first, based on the preset outer wall polishing path, the first error compensation data and the second error compensation data are combined to perform multi-region adaptive adjustment on the path trajectory, and a variable pitch spiral polishing path is set. The path is developed in a spiral shape along the surface of the cylinder body with the axial direction of the wind tower as the reference, and the path interval dynamically changes according to the complexity of different processing regions and the error correction requirements. In the rusted region, the first error compensation data is combined with the rust depth offset information and the feed control parameters in the first error compensation data to automatically reduce the pitch to improve the polishing density per unit area; in the weld segment, the second error compensation data is combined with the weld position offset and the attitude angle adjustment value in the second error compensation data to finely adjust the path direction, so that the trajectory tangent and the weld main shaft direction remain consistent, and the feed depth of the grinding head is controlled to realize the conformal polishing. The entire path generation process adopts a path segment interpolation and attitude constraint control algorithm to ensure the continuity, conformality and stability of the variable pitch path under complex surface topography.

[0044] In the path execution preparation stage, the laser roughness sensor, the vision unit and the displacement measuring device integrated at the end of the polishing head are used to sample the effect of each path execution and record the three types of key quality parameters in real time. The surface roughness is evaluated by taking the Ra value as the standard to reflect the smoothness of the polished surface; the rust residual rate is calculated by identifying the rust stain area that is not removed through the image recognition algorithm to reflect the removal effect; and the weld flatness is calculated from the difference between the front and rear height profiles to indicate whether the polished weld meets the requirements of structural continuity and flatness. The above three types of indexes are bound with the path segment number to form a path-quality mapping structure, which constitutes a key basis for subsequent task scheduling.

[0045] After the path setting and state parameter recording are completed, according to the overall structure division, path priority and processing complexity requirement of the wind power tower cylinder outer wall polishing task, combined with the surface quality distribution reflected by the surface roughness, rust residual rate and weld flatness, the polishing area is spatially divided and task classified, and a master-slave control architecture corresponding to the collaborative work of multiple mechanical arms is established. In the control architecture, one mechanical arm is set as the master control arm, responsible for task beat management, path segment distribution and state coordination, and the remaining mechanical arms are slave control arms, which complete the collaborative polishing work of the corresponding path segment according to the execution path and control rhythm signal issued by the master control arm, and perform trajectory position adjustment and grinding head attitude fine adjustment according to the error compensation data corresponding to each path segment. Through the master-slave coordination mechanism, task synchronization, quality alignment and trajectory continuous control among multiple mechanical arms under the complex variable pitch spiral polishing path are realized.

[0046] Further, the system provided by the application embodiment is also used for:

[0047] The mechanical arm pose data includes joint rotation angle, telescopic translation displacement and end effector attitude angle; a master mechanical arm performing coarse polishing operation and a slave mechanical arm performing fine polishing operation are defined; based on the mechanical arm pose data, the master-slave control architecture is used to cooperatively control the master mechanical arm and the slave mechanical arm in the multiple mechanical arms.

[0048] In the application embodiment, first, the mechanical arm pose data of each mechanical arm is collected, which includes the joint rotation angle obtained by the joint encoder, the telescopic translation displacement obtained by the linear displacement sensor of the telescopic slide rail, and the end effector attitude angle measured by the inertial measurement unit (IMU) on the end effector. These three types of data reflect the current form of each mechanical arm in space in real time, which is used for subsequent action control and attitude matching.

[0049] After the collection is completed, according to the processing requirements of different regions for the polishing task, the task division of the master mechanical arm and the slave mechanical arm is set in the multiple mechanical arms. The master mechanical arm is distributed to the region with heavier rust and larger surface fluctuation, and performs coarse polishing operation, focusing on fast removal of large-area materials. A larger path pitch and a higher feed rate are set in the path segment of the master mechanical arm, while allowing the grinding head to apply a larger contact pressure. Correspondingly, the slave mechanical arm is distributed to the weld or the rough processed region, and performs fine polishing operation, which is encrypted in path density, slows down the feed speed, and reduces the contact pressure, so as to ensure the realization of the lower surface roughness target. Path division is realized through a simple space region surrounding frame and surface roughness threshold screening algorithm.

[0050] After the task division is completed, based on the collected mechanical arm pose data, a master-slave control strategy is started. First, the current joint angle and end pose of the master mechanical arm are taken as a reference, and a reference trajectory is generated in real time through a path tracking control algorithm. Then, the real-time pose data of the slave mechanical arm is read, and a servo error compensation control method is called to compare the reference trajectory of the master mechanical arm, calculate the deviation between the two, and adjust the joint angle and end pose of the slave arm in the control period to ensure that it is synchronized with the master arm in the spatial motion direction and trajectory rhythm.

[0051] When the master mechanical arm changes the path speed or the trajectory pose deflects (for example, the angle is adjusted due to the mutation of the machined surface), a time synchronization mechanism is used to fine-tune the timing of the path of the slave mechanical arm, so that the slave arm continues to follow the master arm in angle, displacement and execution rhythm, avoiding cooperation rupture caused by trajectory deviation or processing delay.

[0052] Further, the system provided by the application embodiment is also used for:

[0053] Based on the master mechanical arm, pose data packets are broadcast in real time to the slave mechanical arm, and the mechanical arm pose data is stored in the pose data packet; after receiving the pose data packet, the slave mechanical arm determines the lag compensation amount.

[0054] In the application embodiment, in order to realize the multi-arm action synchronization in path execution, based on the master-slave control strategy, the master mechanical arm is responsible for generating and broadcasting its own motion state in real time. First, the master mechanical arm obtains the current joint rotation angle of each joint through the joint angle encoder, obtains the telescopic translation displacement of each segment of the arm body through the linear displacement sensor, and calculates the current end effector pose angle by using the inertial measurement unit (IMU) integrated on the end tool.

[0055] The above three types of mechanical arm pose data are packaged to generate a standard structure pose data packet, and a timestamp and a path segment number are attached in the data packet. The master mechanical arm broadcasts the pose data packet to each slave mechanical arm in the network in real time through the EtherCAT bus at a fixed frequency (such as every 10 ms). This broadcast process uses a cyclic master communication mechanism to ensure that the data packet can reach all receiving ends within milliseconds.

[0056] From the mechanical arm after receiving the pose data packet, the local communication decoding module is called to analyze the data content, extract the pose information of the master mechanical arm, and compare it with the current sensor data read by itself item by item. The pose difference calculation method is adopted to calculate the errors of the corresponding joint angles, the end space position and the attitude angle of the master mechanical arm and the slave mechanical arm respectively. For example, the 4th joint angle of the current master mechanical arm is 75.3°, and the slave mechanical arm is 72.8°, so the error is 2.5°; the difference of the end Z position is +1.8mm; the attitude yaw angle deviation is +3.1°.

[0057] In order to correct these differences, the slave mechanical arm determines the lag compensation amount according to the above difference, that is, according to the pose difference and mechanical force difference between the master mechanical arm and the slave mechanical arm, the correction amount to be executed is calculated, which is used to adjust the position, angle and force state of the slave mechanical arm, so as to realize the synchronization and stability between the master and slave mechanical arms. The compensation amount is input into the trajectory controller of the slave mechanical arm as a feedforward correction parameter, and the slave mechanical arm is driven to perform fine adjustment operation. The position loop control and attitude closed loop control are used to adjust the linear position and direction of the end respectively, so that it quickly approaches the reference state of the master mechanical arm, and keeps the dynamic consistency of the overall trajectory.

[0058] Further, the system provided by the application embodiment is further used for:

[0059] Based on the mechanical arm pose data, the force conditions of each joint of the master mechanical arm at different positions and attitudes are analyzed, and joint torque distribution data is obtained; based on the mechanical arm pose data and the lag compensation amount, the contact state of the end effector of the slave mechanical arm with the outer wall of the tower drum at different attitudes and positions is analyzed, and end effector force distribution data is obtained; based on the master mechanical arm and joint torque distribution data, and the slave mechanical arm and end effector force distribution data, a master-slave force-position hybrid control architecture is set.

[0060] In the application embodiment, based on the obtained mechanical arm pose data, first, the force conditions of each joint of the master mechanical arm at different attitudes are calculated, and joint torque distribution data is obtained. Specifically, the torque sensor in the joint motor is used to read the motor output current of each joint in real time when executing the preset trajectory. Since the motor output current is proportional to the load of the joint, the corresponding joint torque value is calculated according to the current value and the motor parameters (such as current-torque conversion coefficient) by table lookup. During the entire path execution process, the joint torque value in each control period is sampled and recorded, and is segmented and counted according to the path number, and finally the joint torque distribution data according to the joint number and path segment index is formed, which is used to represent the actual force distribution of each joint of the master mechanical arm in the entire trajectory.

[0061] In processing the state of the mechanical arm, according to the obtained mechanical arm pose data and the lag compensation amount between the master mechanical arm, the contact state of the end effector with the outer wall of the wind turbine tower drum is analyzed, and the end effector force distribution data is obtained. Specifically, the three-axis force sensor output installed on the end effector is read, including the contact force values in X, Y and Z directions. Combined with the attitude angle of the end (such as Pitch, Roll, Yaw) and the angle between the end and the normal of the tower drum surface, the three-axis contact force is decomposed into normal force and tangential force using the spatial force decomposition method. For example, the measured force of the end in the Z-axis direction is 12N in a certain section of the path, and the attitude angle is-15°, then through the trigonometric function conversion, the normal pressure under this attitude is 11.6N, and the tangential pressure is 3.1N. This process is repeated in each path segment cycle, and the calculation results are associated with the path segment number to form the end effector force distribution data, which is used to judge the fitting state, polishing stability and load balance in the polishing process.

[0062] After obtaining the joint torque distribution data of the master mechanical arm and the end effector force distribution data of the slave mechanical arm, the master-slave control architecture under the master-slave force-position hybrid control is configured. In the control mode, the master mechanical arm uses the position control method to perform trajectory execution to maintain the accuracy and structural rigidity of the path. According to the path segment where the high-load joint (such as more than 80% of the rated torque) identified in the joint torque distribution data, the running speed of the master arm is appropriately reduced, and the gain parameters of the joint are adjusted to reduce the risk of overload. The slave mechanical arm adopts force control method, compares the actual contact force of the end with the expected polishing pressure (such as the target normal force is 10N), if the deviation exceeds the set range (±1N), adjusts the Z-axis feed amount and the end attitude angle for fine tuning, to ensure the contact state with the tower drum surface is stable and continuous.

[0063] Further, the system provided by the application embodiment is further used for:

[0064] Through the master mechanical arm and the joint torque distribution data, the master mechanical arm adopts the feedforward impedance control under the joint torque distribution data, and a first master-slave force-position hybrid control network layer is configured; through the master mechanical arm and the joint torque distribution data, the slave mechanical arm adopts the active compliant control of the six-dimensional force sensor, and a second master-slave force-position hybrid control network layer is configured; the first master-slave force-position hybrid control network layer and the second master-slave force-position hybrid control network layer are connected to obtain the master-slave control architecture.

[0065] In the application embodiment, based on the motion state of the master mechanical arm and the force response of the slave mechanical arm, two functionally complementary control network layers are established, which correspond to the feedforward position stability control of the master arm and the force compliance adjustment of the slave arm respectively, and are connected through unified coordination logic to form the master-slave control architecture.

[0066] Specifically, by joint torque distribution data of the master robot arm, a first master-slave force-position hybrid control network layer is constructed by using a feedforward impedance control method. Specifically, before executing a path, the actual output torque of each joint is read from the torque sensor of each joint, for example, the current torque of the 3rd joint is 2.8 Nm, and the expected stiffness (such as K = 100 Nm / rad) and damping (such as D = 10 Nm·s / rad) of the joint are set in combination with the expected joint displacement, velocity and acceleration corresponding to the target path segment. In each control cycle, the feedforward torque compensation of the joint is calculated using the impedance control formula τ = K·Δθ + D·Δω, where Δθ is the difference between the current joint angle and the target angle, and Δω is the angular velocity error. The feedforward torque is directly superimposed on the position controller input, so that the master robot arm can output stable torque in advance when entering the high load segment or the path curvature mutation zone, avoiding trajectory fluctuations caused by sudden load on the joint, and ensuring that the master arm path execution is more rigid and stable.

[0067] Subsequently, based on the six-dimensional force sensor at the end of the slave robot arm, the coupling relationship between the tangential friction and the normal pressure at the current execution point is extracted, which is used to reflect the action state between the end posture and the contact direction. In combination with the spatial information about the weld direction and width in the surface topological graph, the orthogonal direction of the weld is determined, and a posture adjustment strategy is used in the direction for compliant compensation control. The compensation process extracts the normal pressure oscillation component in the stable contact state of the weld by calculating the normal force fluctuation trend of the end force, which is used to limit the normal micro-motion range of the slave arm in the weld segment, and ensures that the pressure is balanced and the posture is stable when the grinding head switches on both sides of the weld. Thus, the configuration of the second master-slave force-position hybrid control network layer is completed.

[0068] Finally, the two control network layers are connected by a time synchronization and path segment alignment method to establish a unified master-slave control architecture. The master robot arm serves as the master control node, which broadcasts its current path number and starting time stamp (such as T = 200 ms) to the bus network before executing a path segment (such as path segment ID = P22). After receiving the data packet, the slave robot arm aligns its current task segment execution clock to T = 200 ms, and retrieves the corresponding end force target parameters according to the path number, and enters the preparation state. If a path segment execution offset (such as a time sequence deviation greater than ± 5 ms) is detected, the slave arm automatically adjusts the execution delay or starts in advance to ensure that the slave arm synchronously adjusts its end posture and grinding pressure when the master arm executes the path segment P22. In this way, the master-slave robot arms are consistent in path segment control rhythm, parameter calling, and posture adjustment in three dimensions, preventing problems such as path discontinuity or posture disconnection.

[0069] Further, the system provided by the application embodiment is also used for:

[0070] According to the variable-pitch spiral polishing path, a track overlap coefficient is determined; a belt wear rate of the plurality of mechanical arms is introduced, and path fine adjustment is performed in combination with the track overlap coefficient.

[0071] In the embodiments of the present application, first, based on the variable-pitch spiral polishing path, the spatial projection of each path segment is analyzed, and the track overlap coefficient is calculated. Specifically, the polishing path is unfolded on the tower cylinder unfolding plane, the three-dimensional path unfolded along the surface of the cylinder is projected into a spiral track in a two-dimensional coordinate system, and according to the pitch distance d between each circle path and the belt polishing width w, the formula is used to calculate the track overlap coefficient η, where η represents the overlap ratio between unit polishing belts. For example, when the belt width is 20 mm and the path interval is 15 mm, the track overlap coefficient is 25%, which means that there is a 25% coverage area between each polishing path and the adjacent path, preventing unpolished blanks between paths.

[0072] Subsequently, the belt wear rate of each mechanical arm currently in use, that is, the relative decline of the actual material removal efficiency of the belt per unit area compared with that of new belts, is introduced. This data is derived from historical data statistics of the system during operation. Specifically, the belt use time, the polished surface area, the polishing pressure and temperature change during operation of each mechanical arm are recorded, and the wear rate is calculated by comparing with the reference polishing efficiency of new belts under the same working conditions. For example, by recording that a certain mechanical arm belt can polish 50 m 2 of the surface in the first 30 minutes, but only completes 42.5 m 2 of polishing in the same time period, the wear rate of this period is calculated to be 15%. This data is refreshed regularly and is synchronized with the path task timeline to ensure its dynamic nature and availability.

[0073] Finally, the above track overlap coefficient and the belt wear rate of each mechanical arm are considered together to perform path fine adjustment. The adjustment strategy is that when the belt wear rate of a certain mechanical arm is high (for example, more than 15%), the path interval in the polishing area it is responsible for is reduced, and the overlap coefficient is appropriately increased, for example, from 30% to 40%, so as to increase the polishing coverage density per unit area and make up for the insufficient polishing depth due to the decrease in efficiency; for the mechanical arm with low wear rate, the path interval remains unchanged to maximize the efficiency. For example, the mechanical arm with a wear rate of 5% still performs polishing according to the original path interval, while the mechanical arm with a wear rate of 18% reduces the path interval from 14 mm to 12 mm in the next path segment, ensuring polishing consistency.

[0074] Through this process, the coverage control method of the track overlap coefficient and the belt wear rate analysis method based on historical data statistics are combined to realize dynamic adaptive adjustment of the path of the plurality of mechanical arms.

[0075] Further, the system provided by the application embodiment is further used for:

[0076] Based on the six-dimensional force sensor, a coupling relationship between the tangential friction force and the normal pressure is extracted; through the coupling relationship between the tangential friction force and the normal pressure, the tower drum outer wall weld in the surface topography topological graph is combined to perform orthogonal direction compliance compensation, and a normal pressure oscillation component conforming to the weld stable constraint is determined.

[0077] In the application embodiment, first, the system uses a sliding covariance analysis method to extract the coupling relationship between the tangential friction force and the normal pressure in the six-dimensional force sensor. Specifically, the system collects a group of F x , F y (tangential) and F z (normal) data every 20 ms during the execution of the path segment, and takes 50 groups as a sliding window, and calculates the covariance values of F x -F z and F y -F z for each group of data. If the covariance value is positive, it means that F x or F y increases with the increase of F z , and there is a positive coupling trend. If the covariance values of continuous multiple windows (such as 5) are greater than a set threshold (for example, 0.8), it is judged that the pressing state of the current end and the surface friction trend do not match, and there may be unreasonable attitude or surface high point mutation, resulting in abnormal friction.

[0078] Secondly, after confirming that there is a strong coupling relationship, the polishing point of the current path segment is mapped to the surface topography topological graph, and whether the position belongs to the identified weld structure is read. The topological graph is constructed by superimposing the laser radar and the structured light scanning data, and the weld area appears as a continuous convex belt area in the Z-axis direction, with a clear weld main direction vector. The weld orthogonal direction, that is, the unit vector perpendicular to the weld main direction, is calculated based on the angle between the path direction and the weld trend, and is used as a compensation direction reference.

[0079] Then, a posture compliance compensation method based on Z-axis fine adjustment is used for control adjustment. In the weld orthogonal direction, whether the F z value of the current end is fluctuated sharply (such as deviation exceeding ±1.5N) from the average value is analyzed, and F x or F yIf the force is continuously higher than the task setting value (e.g. 3N), the pose fine-tuning of the end effector is triggered. The specific operation includes controlling the end effector to lift 0.3-0.5mm in the Z direction, and at the same time, through the end angle control module of the robot arm, the pitch angle or roll angle is fine-tuned by 0.8° to 1.5°. This compensation is realized through a servo controller and is completed within 2-3 control cycles to ensure that the pose is flexibly attached to the weld undulating surface.

[0080] Finally, a fast Fourier transform (FFT) method is called to obtain the frequency domain data of the force F z The data fluctuation in the weld path segment is analyzed in the frequency domain. The maximum oscillation amplitude in the main frequency interval (e.g. 1-5Hz) is extracted, and whether it belongs to stable fluctuation is judged in combination with the total time length of the path segment. If the maximum oscillation amplitude identified is less than 2N, the main frequency of the fluctuation is stable and the continuous time length reaches more than 60% of the total time length of the path segment, it is determined that the pressure fluctuation is within the reasonable compensation range, and the normal pressure oscillation component conforming to the weld stability constraint is formed. The component will be used for subsequent force and position parameter reuse and pose correction in similar weld structures.

[0081] Further, the system provided by the application embodiment is also used for:

[0082] If the strong wind disturbance tolerance jumps into the strong wind disturbance tolerance risk controllable interval, three-dimensional wind load vector data is obtained; the eddy current damping unit is activated, and a dynamic time warping mechanism is started to realign the motion trajectory, and the trajectory overlap coefficient is updated in combination with the three-dimensional wind load vector data.

[0083] In the application embodiment, when it is detected that the current environmental wind speed reaches the preset strong wind disturbance tolerance risk controllable interval, that is, the wind speed is not higher than the safe shutdown threshold but is sufficient to affect the polishing trajectory stability (for example, between 8-10m / s), the wind disturbance coping mode is immediately entered to ensure the accuracy of the robot arm running trajectory and the stability of the multi-arm collaborative control.

[0084] Firstly, three-dimensional wind load sensing nodes arranged at different heights in the polishing area are called. Each node is composed of a group of ultrasonic anemometers and can measure the wind speed components in X, Y and Z directions in real time. According to the spatial coordinates of the current robot working position, a plurality of wind speed nodes closest to the working position are selected, and the three-dimensional wind load vector data of the working point is extracted by using the multi-point average sampling method.

[0085] After the wind load data collection is completed, the eddy current damping unit is immediately activated. The unit is integrated in the middle and end joints of the robot arm and is composed of a magnetoresistance structure and an aluminum disc. The angular velocity change is detected in real time through the built-in rotary encoder, and when it is found that the rotation speed of a joint fluctuates sharply in a short time (for example, the angular velocity fluctuation is more than 25° / s 2The magnetic field strength is increased to generate a larger eddy current reaction force in the aluminum disk, thereby reducing the attitude deviation caused by the joint's overly rapid response, improving the robotic arm's anti-sway ability under strong wind disturbances, and enhancing the stability of the grinding head's fit.

[0086] Subsequently, Dynamic Time Warping (DTW) is activated to counteract trajectory execution timing offsets caused by wind load interference. By comparing the timestamps of the preset trajectory points within the current path segment with the actual trajectory points at the end of the robotic arm, if multiple consecutive trajectory points are found to be ahead or behind (e.g., the deviation exceeds 200ms), the DTW algorithm is used to perform non-linear remapping of the path time axis, adjusting the path execution rate so that the trajectory point sequence is re-aligned with the task-planned trajectory on the time axis.

[0087] Finally, the trajectory overlap coefficient η of the current path segment is updated based on the three-dimensional wind load vector data. The original trajectory overlap coefficient is defined as follows: Where w is the width of the abrasive belt and d is the spacing between adjacent paths. Under wind disturbance conditions, firstly, based on the three-dimensional wind load vector... The angle θ between the lateral component (usually the X direction) and the current grinding path direction vector is used to calculate the wind offset Δx at the end in the lateral direction using the projection offset estimation method. |×sinθ×Δt, where Δt is the execution time of the path segment. Then, adjust the effective grinding width based on the offset. The updated trajectory overlap coefficient is .

[0088] In summary, the embodiments of this application have at least the following technical effects:

[0089] The application maps the surface morphology topological graph of the wind turbine tower outer wall corrosion depth data and the tower outer wall weld by the surface morphology characteristics of the target wind turbine tower; according to the surface morphology topological graph, a preset outer wall polishing path associated with the wind turbine tower outer wall polishing task is generated, and the matching parameters of the grinding head pressure and speed are determined synchronously in combination with the tower material hardness and the grinding tool wear coefficient; the first error compensation data is determined by introducing the mechanical arm pose data dynamically adjusted by the fuzzy PID controller and compensating the tower outer wall corrosion depth data in the surface morphology topological graph in combination with the preset outer wall polishing path; the second error compensation data is determined by compensating the tower outer wall weld in the surface morphology topological graph in combination with the preset outer wall polishing path through the wind turbine tower outer wall polishing task; a plurality of mechanical arms in collaborative work are connected, the master-slave control architecture is configured in combination with the first error compensation data and the second error compensation data, and the outer wall polishing multi-end collaborative control of the target wind turbine tower is performed by the master-slave control architecture. The application solves the technical problem that the existing technology cannot realize high-precision collaborative polishing of the wind turbine tower outer wall under complex topography, and achieves the technical effect of improving the polishing precision and work consistency of the wind turbine tower outer wall by introducing the topological graph fitting and multi-arm error compensation collaborative control mechanism.

[0090] In the embodiment two, based on the same inventive concept as the intelligent wind turbine tower outer wall polishing system in the foregoing embodiment, as shown in the figure, the embodiment of the application provides an intelligent wind turbine tower outer wall polishing method, which comprises the following steps: Figure 2

[0091] The surface morphology topological graph of the wind turbine tower outer wall corrosion depth data and the tower outer wall weld is mapped by the surface morphology characteristics of the target wind turbine tower; according to the surface morphology topological graph, a preset outer wall polishing path associated with the wind turbine tower outer wall polishing task is generated, and the matching parameters of the grinding head pressure and speed are determined synchronously in combination with the tower material hardness and the grinding tool wear coefficient; the first error compensation data is determined by introducing the mechanical arm pose data dynamically adjusted by the fuzzy PID controller and compensating the tower outer wall corrosion depth data in the surface morphology topological graph in combination with the preset outer wall polishing path; the second error compensation data is determined by compensating the tower outer wall weld in the surface morphology topological graph in combination with the preset outer wall polishing path through the wind turbine tower outer wall polishing task; a plurality of mechanical arms in collaborative work are connected, the master-slave control architecture is configured in combination with the first error compensation data and the second error compensation data, and the outer wall polishing multi-end collaborative control of the target wind turbine tower is performed by the master-slave control architecture.

[0092] Further, a plurality of mechanical arms in collaborative work are connected, the master-slave control architecture is configured in combination with the first error compensation data and the second error compensation data, and the method further comprises:

[0093] ​Based on the preset outer wall polishing path, combined with the first error compensation data and the second error compensation data, a variable pitch spiral polishing path is set, and surface roughness, rust residual rate and weld flatness are recorded; based on the wind power tower outer wall polishing task, combined with the surface roughness, rust residual rate and weld flatness, a master-slave control architecture corresponding to the collaborative work of multiple mechanical arms is established.

[0094] Further, the method further comprises:

[0095] The mechanical arm pose data includes joint rotation angle, telescopic translation displacement and end effector attitude angle; a master mechanical arm performing rough polishing operation and a slave mechanical arm performing fine polishing operation are defined; based on the mechanical arm pose data, the master-slave control architecture is used to cooperatively control the master mechanical arm and the slave mechanical arm in the multiple mechanical arms.

[0096] Further, the method further comprises:

[0097] Based on the master mechanical arm, a pose data packet is broadcasted to the slave mechanical arm in real time, and the mechanical arm pose data is stored in the pose data packet; after receiving the pose data packet, the slave mechanical arm determines the lag compensation amount.

[0098] Further, based on the mechanical arm pose data, the master-slave control architecture is used to cooperatively control the master mechanical arm and the slave mechanical arm in the multiple mechanical arms, and the method further comprises:

[0099] Based on the mechanical arm pose data, the force conditions of each joint of the master mechanical arm at different positions and attitudes are analyzed to obtain joint torque distribution data; based on the mechanical arm pose data and the lag compensation amount, the contact state of the end effector of the slave mechanical arm with the tower outer wall at different attitudes and positions is analyzed to obtain end effector force distribution data; according to the master mechanical arm and the joint torque distribution data, and the slave mechanical arm and the end effector force distribution data, a master-slave control architecture under master-slave force-position hybrid control is set.

[0100] Further, the master-slave control architecture corresponding to the collaborative work of multiple mechanical arms is established, and the method further comprises:

[0101] Through the master mechanical arm and the joint torque distribution data, feedforward impedance control under joint torque distribution data is adopted for the master mechanical arm, and a first master-slave force-position hybrid control network layer is configured; through the master mechanical arm and the joint torque distribution data, active compliant control of a six-dimensional force sensor is adopted for the slave mechanical arm, and a second master-slave force-position hybrid control network layer is configured; the first master-slave force-position hybrid control network layer and the second master-slave force-position hybrid control network layer are connected to obtain the master-slave control architecture.

[0102] Further, the method further comprises:

[0103] According to the variable-pitch spiral polishing path, a trajectory overlap coefficient is determined; a belt wear rate of the plurality of mechanical arms is introduced, and path fine adjustment is performed in combination with the trajectory overlap coefficient.

[0104] Further, active compliant control of the mechanical arms using a six-dimensional force sensor is adopted, a second master-slave force-position hybrid control network layer is configured, and the method further comprises:

[0105] Based on the six-dimensional force sensor, a coupling relationship between tangential friction and normal pressure is extracted; through the coupling relationship between the tangential friction and the normal pressure, in combination with the tower outer wall weld in the surface topological graph, orthogonal direction compliant compensation is performed to determine a normal pressure oscillation component conforming to the weld stable constraint.

[0106] Further, the method further comprises:

[0107] If the strong wind disturbance tolerance jumps into the strong wind disturbance tolerance risk controllable interval, three-dimensional wind load vector data are acquired; the eddy current damping unit is activated, and a dynamic time warping mechanism is started to realign the motion trajectory, in combination with the three-dimensional wind load vector data, the trajectory overlap coefficient is updated.

[0108] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0109] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0110] The present application and the drawings are only exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and changes.

Claims

1. An intelligent wind turbine tower outer wall grinding system, characterized in that, include: The mapping module is used to generate a topological map of the surface morphology of the target wind turbine tower by mapping the corrosion depth data of the outer wall of the tower and the weld seam of the outer wall of the tower. The parameter determination module is used to generate a preset outer wall grinding path associated with the wind turbine tower outer wall grinding task based on the surface topology map, and simultaneously determine the grinding head pressure-speed matching parameters by combining the tower material hardness and the grinding tool wear coefficient. The first error compensation data determination module is used to introduce the robot arm pose data dynamically adjusted by the fuzzy PID controller, and combine it with the tower outer wall corrosion depth data in the surface topology diagram of the preset outer wall grinding path to determine the first error compensation data. The second error compensation data determination module is used to determine the second error compensation data by combining the wind turbine tower outer wall grinding task with the preset outer wall grinding path to compensate for the tower outer wall weld in the surface morphology topology diagram. The collaborative control module is used to connect multiple robotic arms working collaboratively, and to configure a master-slave control architecture by combining the first error compensation data and the second error compensation data, and to perform multi-terminal collaborative control of the outer wall grinding of the target wind turbine tower using the master-slave control architecture. The collaborative control module includes: Based on the preset outer wall grinding path, combined with the first error compensation data and the second error compensation data, a variable pitch spiral grinding path is set, and the surface roughness, rust residue rate, and weld flatness are recorded. Based on the aforementioned wind turbine tower outer wall grinding task, and combined with the surface roughness, rust residue rate, and weld flatness, a master-slave control architecture corresponding to the collaborative operation of multiple robotic arms is established. The robotic arm pose data includes joint rotation angles, telescopic translational displacements, and end effector attitude angles. Define the master robotic arm that performs the coarse grinding operation and the slave robotic arm that performs the fine grinding operation; Based on the robotic arm pose data, the master-slave control architecture is used to coordinate the control of the master robotic arm and slave robotic arms among the multiple robotic arms.

2. The intelligent wind turbine tower outer wall grinding system as described in claim 1, characterized in that, Also includes: Based on the master robotic arm, a pose data packet is broadcast to the slave robotic arm in real time, and the pose data of the robotic arm is stored in the pose data packet. After receiving the pose data packet from the robotic arm, the hysteresis compensation amount is determined.

3. The intelligent wind turbine tower outer wall grinding system as described in claim 2, characterized in that, Based on the robotic arm pose data, the master-slave control architecture is used to coordinate the control of the master and slave robotic arms among the multiple robotic arms, including: Based on the robotic arm pose data, the force conditions of each joint of the main robotic arm under different positions and postures are analyzed to obtain joint torque distribution data; Based on the robotic arm pose data and hysteresis compensation, the contact state between the end effector of the robotic arm and the outer wall of the tower at different postures and positions is analyzed to obtain the force distribution data of the end effector. Based on the torque distribution data of the master robotic arm and joints, and the force distribution data of the slave robotic arm and end effector, a master-slave control architecture under master-slave force-position hybrid control is set up.

4. The intelligent wind turbine tower outer wall grinding system as described in claim 3, characterized in that, Establish a master-slave control architecture for collaborative operations of multiple robotic arms, including: Based on the main robotic arm and joint torque distribution data, the main robotic arm is subjected to feedforward impedance control under the joint torque distribution data, and a first master-slave force position hybrid control network layer is configured. Based on the master robotic arm and joint torque distribution data, the slave robotic arm is subjected to active compliant control using a six-dimensional force sensor, and a second master-slave force position hybrid control network layer is configured. The master-slave hybrid control network layer and the second master-slave hybrid control network layer are connected to obtain the master-slave control architecture.

5. The intelligent wind turbine tower outer wall grinding system as described in claim 4, characterized in that, Also includes: The trajectory overlap coefficient is determined based on the variable pitch spiral grinding path; The belt wear rate of the multiple robotic arms is introduced, and the path is fine-tuned by combining it with the trajectory overlap coefficient.

6. The intelligent wind turbine tower outer wall grinding system as described in claim 5, characterized in that, Active compliant control of the robotic arm using a six-dimensional force sensor is configured with a second master-slave force-position hybrid control network layer, including: Based on the aforementioned six-dimensional force sensor, the coupling relationship between tangential friction force and normal pressure is extracted; By leveraging the coupling relationship between tangential friction and normal pressure, and combining the tower outer wall weld in the surface topology diagram with orthogonal compliance compensation, the normal pressure oscillation component that conforms to weld stability constraints is determined.

7. The intelligent wind turbine tower outer wall grinding system as described in claim 6, characterized in that, Also includes: If the strong wind disturbance tolerance falls into the controllable range of strong wind disturbance tolerance risk, obtain three-dimensional wind load vector data; The eddy current damping unit is activated, and the dynamic time warping mechanism is initiated to realign the motion trajectory. The trajectory overlap coefficient is then updated in conjunction with the three-dimensional wind load vector data.

8. A method for intelligent grinding of the outer wall of a wind turbine tower, characterized in that, The method is performed by an intelligent wind turbine tower outer wall grinding system according to any one of claims 1 to 7, comprising: Based on the surface morphology characteristics of the target wind turbine tower, a surface topology map mapping the corrosion depth data of the outer wall of the tower and the weld seam of the outer wall of the tower is proposed. Based on the surface topology map, a preset outer wall grinding path associated with the wind turbine tower outer wall grinding task is generated, and the grinding head pressure-speed matching parameters are determined simultaneously by combining the tower material hardness and the grinding tool wear coefficient. The robot arm pose data dynamically adjusted by a fuzzy PID controller is introduced, and the tower outer wall corrosion depth data in the surface topology diagram is compensated by the preset outer wall grinding path to determine the first error compensation data. By using the wind turbine tower outer wall grinding task, combined with the preset outer wall grinding path to compensate for the tower outer wall weld in the surface morphology topology diagram, the second error compensation data is determined. Multiple robotic arms are connected to work together. By combining the first error compensation data and the second error compensation data, a master-slave control architecture is configured, and the target wind turbine tower is subjected to multi-terminal collaborative control for external wall grinding using the master-slave control architecture.

Images