A rotor support longitudinal weld seam robot automatic polishing control system and method

Abstract

The present application belongs to the technical field of robot control, and particularly relates to a rotor support longitudinal weld seam robot automatic polishing control system and method. The system comprises multiple units such as visual detection, point cloud integration and path planning, and each unit works cooperatively to realize automatic polishing of the weld seam of a curved surface structural member. The system also has functions such as collision analysis, self-checking, process programming and fault alarm, effectively improving polishing efficiency, quality and system safety, solving the deficiencies of the prior art and promoting the development of related industries. The method comprises steps such as workpiece positioning, scanning, data processing, trajectory planning, simulation verification, polishing operation and fault handling. Workpiece information is obtained through remote and near-end scanning, polishing trajectory is accurately planned, and trajectory is retracted considering tool radius. The technical solution solves the problems of existing rotor support weld polishing relying on manual operation, defects of traditional automatic polishing systems and inability to meet the polishing requirements of complex structural members.

Description

Technical Field

[0001] This invention belongs to the field of robot control technology, and in particular relates to an automated grinding control system and method for longitudinal weld seams of rotor supports. Background Technology

[0002] The rotor support, as a crucial component, is constructed by welding a central cylinder, main vertical ribs, and circumferential stiffening plates. During the welding process, the longitudinal weld (referred to as the longitudinal weld) formed between the central cylinder and the main vertical ribs, and the circumferential weld (referred to as the circumferential weld) formed between the circumferential stiffening plates and the central cylinder and main vertical ribs, play a key role in the quality of the rotor support. Currently, the grinding of the rotor support welds mainly relies on manual operation. This method has many drawbacks, such as consuming a large amount of manpower, long grinding time, high worker intensity, and harsh working environment. Furthermore, the quality of the ground longitudinal welds is difficult to guarantee. Therefore, automated transformation is urgently needed to improve production efficiency and grinding quality.

[0003] Traditional automated grinding systems typically consist of a robotic arm, a force control unit, a grinding device, and a control system. In practice, the robot needs to be manually operated to move along the weld seam trajectory, ensuring the grinding tool applies appropriate force to the weld. Multiple points are selected for robot trajectory teaching, and the recorded weld seam trajectory is then stored in the robot program. Finally, the control system calls the program to complete the grinding. While this method achieves a degree of automation, the manual teaching process is cumbersome and requires a high level of expertise from the operator.

[0004] In terms of patented technology, Chinese invention patent application number CN202210506792.4 discloses an automatic flexible grinding system and method for weld seams of plate-shaped structural components. In addition to a robotic arm, force control unit, grinding device, and control system, this system also includes a line-scan laser sensor. Its grinding method encompasses steps such as weld seam input, analysis, on-site contour acquisition, posture-changing grinding, and active floating grinding. This invention imports theoretical weld seam data in parameter form, achieving automatic grinding without manual teaching, reducing the difficulty of use, and allowing ordinary operators to operate the equipment. However, this patented technology has limitations. While it performs well when grinding weld seams of planar plate-shaped structural components, its grinding effect is unsatisfactory when grinding curved structural components, failing to meet the grinding needs of complex structural components.

[0005] In summary, current rotor support weld grinding technology is insufficient in terms of automation and grinding effect. There is an urgent need for a highly efficient and automated weld grinding technology that can be applied to various structural components, especially curved surface components, in order to solve the problems existing in the current technology and promote the development of related industries. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of the existing technology by proposing an automated grinding control system and method for longitudinal weld seams of rotor supports, which realizes automated grinding of weld seams of curved structural components.

[0007] The above objectives are achieved through the following technical solutions: An automated robotic grinding control system for longitudinal weld seams of rotor supports includes: The visual inspection unit includes several laser scanners and corresponding image acquisition devices, which are used to scan the rotor support workpiece, reconstruct the surface model of the rotor support workpiece, and obtain the original image data of the rotor support workpiece surface.

[0008] The point cloud integration unit is used to receive the raw image data collected by the vision inspection unit through the data bus, and to perform preprocessing on the raw image data, including solving, coordinate transformation and filtering, to convert the raw image data into ordered and processable point cloud data, and to perform meshing processing and output it to a file.

[0009] The path planning unit receives near-end point cloud data from the point cloud integration unit, constructs a 3D model of the rotor support workpiece, analyzes the geometry and weld position of the rotor support workpiece, generates environmental data, and calculates the end surface trajectory suitable for grinding, thereby generating trajectory information.

[0010] The execution unit is used to read the trajectory information generated by the path planning unit, plan the specific motion trajectory of the robot according to the robot's kinematic principles, and send the motion trajectory data to the robot for polishing operations.

[0011] The database is used to store the operational data of the robot's automated grinding control system, as well as the intermediate and result data generated by each functional unit.

[0012] The main control unit is used to read the status information of each functional unit in real time and display this status information intuitively on the interface; at the same time, it provides an operation interface for operators to perform operations on the robot automated grinding control system, including parameter setting and start / stop control.

[0013] Preferably, it also includes an external interface unit, which provides a production system interface and a positioner interface; wherein, the production system interface is used for information exchange and coordination with the entire production system; and the positioner interface is used to establish communication and control between the robot automated grinding control system and the positioner.

[0014] Preferably, the execution unit includes a grinding module and a scraping module, which are used to automatically generate the end grinding trajectory and the end scraping trajectory based on the three-dimensional point cloud data and the grinding process requirements, respectively.

[0015] Preferably, it also includes a collision analysis and obstacle avoidance unit, which is used to perform operation simulation verification through simulation function before the robot performs the polishing operation, so as to eliminate the possibility of collision; wherein, if the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory; if the simulation verification passes, the execution unit is triggered to send motion trajectory data to the robot to perform the polishing operation.

[0016] Preferably, it also includes a self-testing and status detection unit for performing system operation self-tests, including monitoring robot status and position parameters, monitoring external axis status and position, and monitoring grinding pressure.

[0017] Preferably, it also includes an integrated process programming unit for editing the grinding process steps and adjusting the grinding process parameters according to process requirements.

[0018] Preferably, it also includes a fault alarm unit for monitoring system faults during the grinding operation; when the robotic automated grinding control system is interrupted by a fault, the audible and visual alarm unit provides an audible and visual alarm reminder, and after the fault is cleared, it triggers the system to continue task execution from the interrupted position.

[0019] An automated grinding control method for longitudinal weld seams of rotor supports, comprising the following steps: S1, the rotor support workpiece is in place, the rotor support workpiece is manually adjusted to the predetermined position and locked, and the robot is moved to the initial position of the weld. S2. Start the vision inspection unit to scan the rotor support workpiece, complete the reconstruction of the surface model of the rotor support workpiece, and obtain the original image data of the surface of the rotor support workpiece. S3. The point cloud integration unit receives the raw image data collected by the visual inspection unit through the data bus, and performs preprocessing on the raw image data, including solving, coordinate transformation and filtering, to convert the raw image data into ordered and processable point cloud data, and performs meshing processing and outputs it to a file. S4. The path planning unit receives near-end point cloud data from the point cloud integration unit, constructs a three-dimensional model of the rotor support workpiece, analyzes the geometry and weld position of the rotor support workpiece, generates environmental data, and calculates the end surface trajectory suitable for grinding, thereby generating trajectory information. S5. The execution unit reads the trajectory information and plans the robot's specific motion trajectory based on the robot's kinematic principles. S6. The motion trajectory is simulated and verified by the collision analysis and obstacle avoidance unit. If the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory. If the simulation verification passes, the execution unit is triggered to send the motion trajectory data to the robot for polishing. S7 monitors system malfunctions in real time during the grinding process; when the robot's automated grinding control system is interrupted, it controls the audible and visual alarm unit to issue an audible and visual alarm reminder, and triggers the system to continue task execution from the interrupted position after the fault is cleared; S8. Use the roller frame automatic positioning device to rotate the rotor support workpiece to the next weld position, and repeat the above steps S1-S7 until all welds of the current rotor support workpiece are ground.

[0020] Preferably, in step S2, scanning the rotor support workpiece includes one far-end scan and one near-end scan; wherein, the far-end scan is to detect the posture of the rotor support workpiece and initially locate the grinding point; the near-end scan is to accurately measure the grinding point.

[0021] Preferably, in step S4, generating trajectory information includes the following steps: S41, based on edge detection of point cloud data, two dividing lines L1 and L2 are obtained, and the point cloud is divided by dividing lines L1 and L2. The area between the two lines is the grinding area, and the area outside the two lines is two planar areas. S42, perform plane segmentation fitting on the point cloud data of the polishing area to obtain two extended intersecting planes P1 and P2, and calculate the normal vectors Normol1 and Normol2 of planes P1 and P2 respectively. S43, calculate the intersection line I of planes P1 and P2, and obtain the direction vector Direction of the intersection line I based on the cross product of normal vectors Normol1 and Normol2; S44, based on the structure of the area to be ground on the rotor support workpiece, preset the grinding arc radius R; take a point on the intersection line I. , command point ,make The point with the smallest radius in the direction of the vector Direction is the area to be polished. S45, based on the junction of the arc radius R. Obtain the center O of the polished arc and the polished arc. ; S46, determine the single-step grinding trajectory, i.e., the point of pass. Construct the perpendicular line of intersection line I to plane P1. ,perpendicular It intersects the dividing line L1 at point Passing point Draw the perpendicular line of the intersection line in the plane P2. ,perpendicular It intersects the dividing line L2 at point Then a complete trajectory along the direction of the first cross section is a straight line. Polishing the arc and straight line ; S47, determine all grinding paths, i.e., the preset extension distance parameter step, and set the current point Extend the vector Direction along the intersection line I by the distance step, and then let... After obtaining the new point Ni, return to step S45 until the grinding trajectory on all cross sections is obtained. S48, the straight trajectories on planes P1 and P2 are reduced by a distance R(t) along their respective normal directions, and the arc trajectories are reduced by a distance R(t) towards the center O. The lines obtained after this reduction are the trajectory information generated by the path planning unit.

[0022] Preferably, in step S45, the center O of the polished arc and the polished arc are obtained. Includes the following steps: S451, with point Draw a circle with center D and virtual radius D. The intersection points of this circle with planes P1 and P2 are respectively... and ; S452, passing Extend a distance of length R along the normal vector Normol1, passing through Extend the normal vector Normol2 by a distance of R, such that the normal vectors Normol1 and Normol2 intersect at point O with a length of R, which is the center O of the polished arc. S453, draw a circle based on the center O and the radius R of the arc to obtain the polished arc. At this point, the angle between the normal vectors Normol1 and Normol2 is θ, and the virtual radius is D = R * tan(θ / 2).

[0023] Preferably, in step S5, the method for planning the robot's specific motion trajectory is as follows: the robot trajectory information is stored in a text file, with each point data recorded on a separate line. Each line records the robot's X, Y, Z, A, B, and C values, separated by spaces. The text file containing this data is then converted into an executable program for the robot, i.e., the specific motion trajectory program, using the SDK provided by the robot. Here, X, Y, and Z represent the axis position coordinates of the robot's end effector in space, and A, B, and C represent the attitude angles of the robot's end effector on the X, Y, and Z axes, to determine the orientation of the robot's end effector.

[0024] Compared with existing technologies, the advantages of this technical solution are: I. Significantly improved level of automation Eliminating the dependence on manual teaching: This invention automatically completes the reconstruction of the rotor support workpiece surface model, point cloud data processing and grinding trajectory generation through a visual inspection unit, a point cloud integration unit and a path planning unit. No manual teaching is required, which reduces the difficulty of operation and makes it easy for ordinary operators to use, greatly improving the degree of automation and production efficiency.

[0025] The entire process is fully automated: from workpiece positioning, scanning, trajectory planning, simulation verification to grinding operations, and then to fault handling and continuous grinding of multiple welds, the whole process is highly automated. The system can automatically identify weld positions, plan grinding trajectories, verify trajectory safety, and automatically alarm and resume operations when faults occur, reducing manual intervention and improving the continuity and stability of production.

[0026] II. Applicable to grinding complex curved surface structures Solving the problem of polishing curved surfaces: This invention uses advanced visual inspection and point cloud processing technology to accurately reconstruct the surface model of complex curved structural components such as rotor supports. Combined with path planning algorithms, it generates end surface trajectories suitable for polishing, effectively solving the polishing problem of curved structural components and broadening the application scope of the system.

[0027] Precise trajectory planning: In step S4, through a series of operations such as precise segmentation of the grinding area, plane fitting, intersection calculation, and arc determination, a precise grinding trajectory can be generated based on the geometry of the curved surface component and the weld position. Simultaneously, the radius of the grinding tool is considered, and the trajectory is adjusted inwards to ensure the consistency and accuracy of the grinding effect.

[0028] III. The system has rich and comprehensive functions. Multi-module collaborative operation: The system includes multiple functional units such as a vision inspection unit, point cloud integration unit, path planning unit, execution unit, database, and interface main control unit. It is also equipped with functional modules such as an external interface unit, collision analysis and obstacle avoidance unit, self-test and status detection unit, integrated process programming unit, and fault alarm unit. The functional modules cooperate with each other to form a complete closed-loop control system, which can realize comprehensive monitoring and precise control of the grinding process.

[0029] Ensuring system safety and reliability: The collision analysis and obstacle avoidance unit performs simulation verification before the robot performs grinding operations to eliminate collision risks. If the verification fails, the trajectory can be replanned or optimized to ensure the robot operates in a safe environment. The self-test and status detection unit monitors parameters such as robot status, external axis status, and grinding pressure in real time to promptly detect potential problems. The fault alarm unit promptly alarms when a system fault occurs and can resume the task from the interrupted position after the fault is cleared, ensuring the system's reliability and stability.

[0030] IV. High flexibility and customizability in process. Integrated Process Programming Unit: The integrated process programming unit allows operators to edit grinding process steps and adjust process parameters according to different process requirements, meeting diverse grinding needs. Operators can flexibly adjust the grinding process based on the rotor support material, weld type, and grinding quality requirements, improving the system's adaptability and flexibility.

[0031] Multi-mode scanning: The vision inspection unit adopts a combination of remote scanning and near-end scanning. Remote scanning initially locates the grinding points, while near-end scanning performs precise measurements. This allows for flexible adjustments based on the actual condition of the workpiece, improving the accuracy and efficiency of grinding. Attached Figure Description

[0032] Figure 1 A schematic diagram of a preferred structure for an automated grinding control system for longitudinal weld seams of a rotor support robot. Figure 2 A preferred implementation flowchart of a robotic automated grinding control method for longitudinal weld seams of rotor supports; Figure 3 A schematic diagram of the connection structure between the central cylinder and the main vertical rib of the rotor support workpiece; Figure 4 This is a schematic diagram of the point cloud structure divided by dividing lines L1 and L2. Figure 5 A schematic diagram illustrating the structure for obtaining extended intersecting planes P1 and P2 by performing planar segmentation fitting on point cloud data; Figure 6 To obtain the center O of the polished arc and the polished arc The schematic diagram.

[0033] In the picture: 1. Longitudinal weld; 2. Main vertical reinforcement; 3. Central cylinder. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should not be construed that the present invention is limited to the following examples. Without departing from the concept of the present invention, all modifications and improvements of the present invention in this field should be included within the protection scope of the claims of the present invention.

[0035] Unless otherwise defined, the technical or scientific terms used in this disclosure shall have the ordinary meaning as understood by one of ordinary skill in the art to which this disclosure pertains. Words such as “or” and “comprising” as used in this disclosure mean that an element or object preceding that word encompasses the elements or objects listed following that word and their equivalents, but do not exclude other elements or objects.

[0036] Example 1 This embodiment discloses an automated grinding control system for the longitudinal weld seam of a rotor support robot. As a preferred implementation of this embodiment, as shown below... Figure 1 As shown, it includes: The vision inspection unit, comprising several laser scanners and corresponding image acquisition devices, is designed to meet the scanning requirements of the rotor support workpiece. The vision inspection unit scans the rotor support workpiece and can be mounted on an adjustable robotic arm or fixed frame to ensure coverage of the entire workpiece surface. Internally, the vision inspection unit is equipped with a high-performance image processing chip to process the acquired image information, supporting multiple scanning modes such as line scanning and area scanning, to more accurately reconstruct the workpiece surface model and obtain the original image data of the rotor support workpiece surface.

[0037] The point cloud integration unit features a powerful data processing chip and large-capacity memory to handle the massive amounts of raw data acquired by the laser scanner. Connected to a data bus, it supports multiple data transmission protocols, such as Ethernet and CAN bus, ensuring stable data transmission. It receives raw image data from the vision inspection unit via the data bus. The unit includes dedicated calculation, coordinate transformation, and filtering modules. These modules utilize hardware acceleration technology to improve data processing speed, performing preprocessing on the raw image data, including calculation, coordinate transformation, and filtering. This converts the raw image data into ordered, processable point cloud data, performs meshing, and outputs it to a file.

[0038] The path planning unit employs advanced 3D modeling software and algorithm libraries to construct a 3D model of the rotor support workpiece based on the received near-endpoint cloud data. The path planning unit embeds a trajectory calculation algorithm that considers factors such as the size, shape, and grinding precision of the grinding tool, as well as the geometry and weld location of the rotor support workpiece, to generate environmental data and calculate a suitable end-surface trajectory for grinding, thereby generating trajectory information.

[0039] The execution unit encompasses the industrial robot controller (hereinafter referred to as the controller) and the robot body (hereinafter referred to as the robot). The robot is typically a multi-axis articulated robot, possessing high precision, high speed, and high load capacity, capable of adapting to rotor support workpieces of varying sizes and weights. The controller integrates trajectory planning and kinematic algorithms to accurately plan the robot's motion trajectory based on the read trajectory information. In summary, the execution unit reads the trajectory information generated by the path planning unit, plans the robot's specific motion trajectory according to the robot's kinematic principles, and sends the motion trajectory data to the robot for grinding operations.

[0040] The database employs a distributed database architecture to ensure the reliability and scalability of data storage. It has a dedicated data storage area for storing the operational data of the robot's automated grinding control system, as well as intermediate and result data generated by various functional units (i.e., vision inspection unit, point cloud integration unit, path planning unit, execution unit, etc.), including but not limited to visual scanning data, system operation logs, and robot motion trajectory data.

[0041] The main control unit (MCU) can include an industrial computer and a high-resolution display screen. The industrial computer possesses powerful computing and graphics processing capabilities. The MCU has an intuitive user interface, facilitating operation and monitoring of various functional units. Specifically, the MCU reads the status information of each functional unit in real time and displays this information intuitively on the interface; it also provides an operator interface for performing operations on the robotic automated grinding control system, including parameter setting and start / stop control.

[0042] Based on the above structure, the working principle of this technical solution is as follows: First, after the system is started, the vision inspection unit scans the rotor support workpiece to obtain the original image data of the workpiece surface.

[0043] Secondly, the raw image data is transmitted to the point cloud integration unit via a data bus. The point cloud integration unit performs preprocessing such as calculation, coordinate transformation, and filtering, converting the disordered raw data into ordered, processable point cloud data, and then performs meshing processing before outputting it to a file and storing it in the database. The calculation process converts the raw sensor data into spatial coordinate information; coordinate transformation unifies data from different coordinate systems into a standard coordinate system, facilitating subsequent processing; filtering removes noise and outliers, improving data quality. The meshing operation transforms the processed point cloud data into a 3D mesh model for subsequent path planning and storage.

[0044] Then, the path planning unit constructs a 3D workpiece model of the rotor support workpiece based on the meshed point cloud data, and analyzes the geometric features and weld locations of the workpiece. Based on the grinding process requirements, such as grinding depth, width, and roughness, the end-effector trajectory of the robot grinding is generated. The trajectory takes into account the robot's workspace, kinematic constraints, and the physical characteristics of the grinding tool to ensure the consistency and reliability of the grinding effect.

[0045] Furthermore, after receiving the trajectory information, the execution unit converts the grinding end surface trajectory into a motion trajectory that the robot can recognize, based on the robot's kinematic model.

[0046] Finally, throughout the system's operation, the database stores a large amount of visual scanning data and system operation data, providing data support for system performance evaluation, fault diagnosis, and process optimization. The interface control unit monitors the system's operating status in real time. Operators can use the interface to set system parameters, schedule tasks, and query status. When abnormal situations occur, timely intervention and adjustments can be made to ensure the stable operation of the system.

[0047] Example 2 This embodiment discloses an automated grinding control system for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiment 1, it further includes an external interface unit, which provides a production system interface and a positioner interface.

[0048] The production system interface is used for information exchange and coordination with the entire production system. It can employ standard industrial communication protocols, such as industrial Ethernet (e.g., PROFINET, Ethernet / IP) or fieldbus (e.g., PROFIBUS, DeviceNet), to connect with other equipment on the production line. The production system interface functions include receiving task instructions from the production system, such as signals to start a grinding task, identification information of the rotor support workpiece, and processing requirements. Simultaneously, it feeds back the status information of the grinding control system (e.g., ready status, running status, and completed status) to the production system for task scheduling and monitoring. Furthermore, the production system interface can synchronize with the material flow information of the production system, ensuring that the rotor support workpiece to be ground is transported to the working position of the grinding control system at the appropriate time, and that the ground workpiece is removed, achieving seamless integration of the entire production process.

[0049] Therefore, the automated grinding control system can be connected to programmable logic controllers (PLCs) or manufacturing execution systems (MES) within the production system via a production system interface. When connected to a PLC, it can receive digital or analog control signals, such as start and stop signals for grinding tasks. When connected to an MES, it can receive more detailed production task information, such as workpiece batches and processing parameters. Simultaneously, it feeds back various status information of the grinding system (such as task completion status, equipment operating status, and fault information) to the production system, enabling the production system to comprehensively understand the status of the grinding system and achieve automated management of the production process.

[0050] A positioner is a device used to change the position and orientation of a workpiece during welding and machining. In this grinding control system, the positioner interface allows the system to communicate and control the positioner. The positioner interface can use specialized control signal protocols, such as analog signals (e.g., 4-20mA current signals or 0-10V voltage signals) or digital signals (e.g., high and low level signals), or more advanced communication protocols, such as CAN bus or Modbus, to achieve precise control of the positioner. The function of the positioner interface is to control the rotation, flipping, and other movements of the positioner to adjust the workpiece on the rotor support to the optimal grinding position. For example, based on the workpiece orientation information provided by the vision inspection unit, control commands are sent to the positioner through the positioner interface to position the longitudinal weld seam of the workpiece in the most favorable position and angle for grinding. Simultaneously, the interface can also receive the positioner's status information (e.g., current position, orientation, motion status, etc.) to ensure the safety and effectiveness of the grinding operation.

[0051] Therefore, the automated grinding control system can communicate with the positioner's controller via the positioner interface. Control commands, such as rotation angle, speed, and direction, are sent to the positioner's controller, causing the positioner to adjust the rotor support workpiece's posture according to the grinding system's requirements. Simultaneously, the system receives current position and status information from the positioner to ensure that the positioner is in the correct position and state during grinding operations, preventing malfunctions that could affect grinding quality and safety.

[0052] Example 3 This embodiment discloses an automated grinding control system for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiment 1 or 2, its execution unit includes a grinding module and a scraping module, which are used to automatically generate end grinding trajectory and end scraping trajectory based on three-dimensional point cloud data and grinding process requirements, respectively.

[0053] The grinding module includes a grinding motor, grinding tools, the drive mechanism for the grinding tools, and related controllers. The grinding motor provides power to the grinding tools, and its power and speed can be adjusted according to the grinding process requirements. Grinding tools are typically grinding wheels or discs with high hardness and wear resistance; different types of tools can be selected based on different grinding needs, such as surface grinding tools and corner grinding tools. The drive mechanism is responsible for precisely controlling the position and orientation of the grinding tools. It generally uses high-precision servo motors and precision transmission mechanisms, such as ball screws and linear guides, to ensure that the tools move accurately along the generated end-grinding trajectory.

[0054] The grinding module generates its trajectory based on the 3D point cloud data obtained from scanning and the grinding process requirements. First, the 3D point cloud data is analyzed to determine the area and surface features to be ground, such as the surface roughness and flatness requirements of the rotor support workpiece. According to the grinding process, parameters such as grinding depth, grinding speed, and grinding direction are considered, and combined with the robot's kinematics and dynamics, the optimal motion trajectory of the grinding tool in space is calculated. For example, for a planar area requiring a specific surface roughness, the feed rate, grinding depth, and number of grinding passes are calculated based on the grinding process requirements, generating a corresponding trajectory to ensure uniform grinding in that area and achieve the required surface quality. When generating the final grinding trajectory, the radius and shape of the grinding tool are also considered to avoid over-cutting or under-cutting due to tool size and shape. Through precise analysis of the 3D point cloud data, it is ensured that the grinding tool accurately conforms to the workpiece surface while avoiding interference with the workpiece.

[0055] The grinding module includes grinding tools, such as scrapers or specialized grinding cutters, as well as their drive and control components. The shape and size of the grinding tools are designed according to the grinding requirements of the longitudinal weld seams of the rotor support workpiece, and typically have special geometries to accommodate weld seams of different shapes and sizes. The drive mechanism of the grinding tools also uses high-precision motors and transmission components to ensure the accuracy and stability of the grinding operation.

[0056] Trajectory Generation in the Grinding Module: The trajectory generation principle of the grinding module is similar to that of the grinding module, but it focuses more on the processing of longitudinal welds. By analyzing the weld portion in the 3D point cloud data, information such as the weld's shape, width, depth, and location is determined. Based on the grinding process, such as the required depth, width, and angle, the motion trajectory of the grinding tool is calculated. Grinding operations typically require precise cutting of the weld; therefore, the entry and exit angles of the grinding tool are precisely controlled during trajectory generation to ensure the grinding effect. For example, for a V-shaped weld, based on its geometry and grinding process requirements, the motion trajectory of the grinding tool within the V-shaped weld is calculated to ensure complete removal of excess weld material while maintaining the flatness of the ground surface and a smooth transition with the surrounding surfaces.

[0057] Thus, the workflow of the execution unit is as follows: First, the execution unit obtains preliminary trajectory information and polishing process requirements from the path planning unit, and acquires the scanned 3D point cloud data from the system.

[0058] Secondly, the grinding module and the scraping module refine and optimize the preliminary trajectory according to their own tool characteristics and process requirements, respectively, to generate accurate end grinding trajectory and end scraping trajectory.

[0059] Then, the generated trajectory is sent to the robot controller, and the grinding and scraping operations are started simultaneously. The grinding and scraping tools move along the trajectory under the control of the drive mechanism to grind the workpiece.

[0060] Furthermore, during the polishing process, the polishing parameters and trajectory are continuously adjusted based on feedback information from the robot controller to ensure polishing quality and safety.

[0061] Finally, after the polishing operation is completed, the final polishing data and results are fed back to other units of the system, such as the interface control unit and the database, for recording and subsequent analysis.

[0062] Example 4 This embodiment discloses an automated grinding control system for a rotor support longitudinal weld seam robot. As a preferred implementation of this embodiment, based on embodiment 1, 2, or 3, it further includes a collision analysis and obstacle avoidance unit. The collision analysis and obstacle avoidance unit includes a simulation software module, a collision detection algorithm module, and an obstacle avoidance strategy module. It is used to perform operation simulation verification through simulation function before the robot performs grinding operations to eliminate the possibility of collisions. If the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory. If the simulation verification passes, the execution unit is triggered to send motion trajectory data to the robot to perform grinding operations.

[0063] The simulation software module includes a 3D physics engine capable of accurately simulating the robot and its surrounding environment. It can create virtual simulation scenes based on the robot's physical parameters (such as dimensions, joint range of motion, mass, and inertia) and the working environment (including 3D models of the workpiece, positioner, and worktable). It supports importing various 3D model file formats, such as STL, STEP, and IGES, to import the geometric models of the robot, workpiece, and other equipment into the simulation environment. Simultaneously, it can receive grinding trajectory information from the path planning unit and apply it to the robot's motion simulation.

[0064] Collision Detection Algorithm Module: Employs efficient collision detection algorithms such as Bounding Volume Hierarchy (BVH) and Spatial Partitioning. These algorithms divide the robot and surrounding objects into different levels or regions, quickly determining whether collisions are possible. It can monitor the distance and positional relationships between the robot's various components (such as the robotic arm, grinding tools, cables, etc.) and surrounding objects in real time while executing the grinding trajectory, with accuracy down to the millimeter level or even higher. When the distance is less than a safety threshold, a collision is considered possible.

[0065] Obstacle Avoidance Strategy Module: Once a potential collision risk is detected, the obstacle avoidance strategy module will formulate corresponding obstacle avoidance plans based on different situations. These strategies may include modifying the robot's motion trajectory, adjusting the robot's posture, pausing the robot's movement, or even replanning the polishing trajectory. Based on the robot's kinematics and dynamics, it can generate optimized obstacle avoidance trajectories, ensuring that collisions are avoided while minimizing impact on the efficiency and quality of the polishing operation.

[0066] Therefore, the working principle of the collision analysis and obstacle avoidance unit is as follows: Scene Setup: First, the collision analysis and obstacle avoidance unit retrieves relevant information from the system. It extracts detailed physical and kinematic parameters of the robot from the database, obtains the grinding trajectory from the path planning unit, and acquires 3D model information of the workpiece and its surrounding environment from the vision inspection unit or other relevant units. This information is then imported into the simulation software module to establish a complete virtual simulation scene, including the robot's initial state, the workpiece's position and orientation, and other equipment and environmental factors that may affect the robot's movement.

[0067] Collision Analysis Process: In the simulated environment, the robot moves according to the planned grinding trajectory. The collision detection algorithm module tracks the robot's movement in real time, performing collision detection on the robot, its components, and the geometry of surrounding objects. For each time step or motion node, the relative position and distance between the robot and surrounding objects are calculated. Algorithms such as hierarchical bounding boxes are used to quickly filter out areas where collisions may occur, followed by precise geometric calculations to determine whether a collision has actually occurred. For example, for a multi-axis robot, it checks whether each joint and its connecting components will collide with the workpiece, positioner, or other equipment during movement.

[0068] Obstacle Avoidance: The obstacle avoidance strategy module is activated when a collision risk is detected. Depending on the specific collision situation, it can employ different obstacle avoidance strategies. For minor interference risks, it may fine-tune the robot's trajectory, such as changing the speed of robot joints, adjusting the direction of movement, or modifying some path points to allow the robot to bypass potential collision areas. In more complex cases, it may be necessary to replan the robot's trajectory, considering changes to the robot's posture to allow it to approach or bypass obstacles in a safer manner. Simultaneously, the requirements of the polishing process will be comprehensively considered to ensure that obstacle avoidance operations do not significantly impact polishing quality.

[0069] Example 5 This embodiment discloses an automated grinding control system for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiments 1, 2, 3 or 4, it further includes a self-testing and status detection unit. The self-testing and status detection unit includes a sensor module (including a robot status sensor, an external axis status sensor and a grinding pressure sensor), a data acquisition and processing module, and a status evaluation and alarm module, which are used to perform system operation self-testing, including monitoring robot status and position parameters, monitoring external axis status and position, and monitoring grinding pressure.

[0070] Robot status sensors, including joint angle sensors, joint torque sensors, and motor encoders, are used to measure the angles of each joint, the torque it experiences, and the rotational position of the motors. These sensors provide accurate feedback on the robot's motion state, offering a basis for subsequent state assessment and control. For example, joint angle sensors typically employ high-precision rotary encoders, capable of accurately measuring the rotational angles of robot joints with an accuracy of ±0.1° or even higher, ensuring the precision and stability of the robot's motion.

[0071] External shaft status sensors: For external shafts (such as positioners), corresponding position sensors and speed sensors are provided. The position sensor can be a linear encoder or a rotary encoder, used to measure the displacement or rotation angle of the external shaft; the speed sensor can measure the movement speed of the external shaft, ensuring that the position and speed of the external shaft are within the set range and accuracy, so as to accurately monitor the working status of the external shaft.

[0072] Grinding pressure sensors: These are typically installed between the grinding tool and the robot's end effector to measure the pressure during the grinding process. Common grinding pressure sensors include strain gauge sensors or piezoresistive sensors. They can monitor the grinding pressure in real time to ensure that the grinding pressure meets process requirements and prevent excessive or insufficient pressure from affecting grinding quality or damaging the workpiece.

[0073] Data Acquisition and Processing Module: This module is responsible for acquiring data from various sensors and preprocessing it. It converts analog signals output by sensors (such as voltage signals from pressure sensors) into digital signals, then performs filtering, amplification, and linearization to ensure data accuracy and usability. Simultaneously, it performs preliminary data analysis and organization, such as averaging multiple data points and removing outliers, to improve data reliability.

[0074] Status Assessment and Alarm Module: Based on the collected and processed data, this module assesses parameters such as robot status, external axis status, and grinding pressure. It has a series of status thresholds, including the maximum and minimum allowable ranges of robot joint torque, the extreme positions of external axes, and the reasonable range of grinding pressure. When monitored parameters exceed or fall below these thresholds, corresponding alarm mechanisms are triggered, such as issuing audible and visual alarm signals, displaying alarm information on the main control unit interface, and prompting the operator to take appropriate action.

[0075] Therefore, the working principle of the self-test and status detection unit is as follows: Data Acquisition: The sensor module continuously measures parameters such as robot status and position, external axis status and position, and grinding pressure. For example, the robot joint angle sensor monitors the angle changes of each joint in real time, the joint torque sensor measures the torque exerted on the joints during robot movement and grinding, the external axis position sensor tracks the position changes of the external axis, and the grinding pressure sensor senses the pressure applied to the workpiece by the grinding tool. These sensors send the measured data to the data acquisition and processing module in different signal forms (such as voltage, current, pulses, etc.).

[0076] Data Processing and Analysis: The data acquisition and processing module processes sensor data. For analog signals, analog-to-digital conversion is first performed to convert them into digital signals that can be processed by a computer. Then, the digital signals are filtered to remove noise and interference, such as using a low-pass filter to remove high-frequency noise to ensure data smoothness. Simultaneously, linearization is performed to establish an accurate correspondence between the sensor output and the actual physical quantity; for example, converting the voltage signal from a pressure sensor into the actual pressure value. The processed data is then analyzed, for example, to calculate the rate of change of robot joint angles over a period of time to evaluate the robot's movement speed; to check whether the external axis position has reached the preset limit position; and to analyze whether the grinding pressure is stable or meets process requirements.

[0077] Status Assessment and Alarm: The status assessment and alarm module evaluates the processed data based on preset status thresholds. Taking robot joint torque as an example, if the measured joint torque exceeds the set safety limit, it may indicate that the robot has encountered abnormal resistance during movement, possibly due to mechanical failure or collision. If the grinding pressure exceeds the specified range, it may affect the grinding quality or damage the workpiece. When an abnormal state is detected, an alarm mechanism is triggered. The alarm can be triggered by a pop-up alarm window on the main control unit interface, displaying specific abnormal information (such as "Robot joint torque is too high, a possible fault"), while simultaneously emitting sound or light signals to remind the operator to handle the situation promptly.

[0078] Example 6 This embodiment discloses an automated grinding control system for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiments 1, 2, 3, 4 or 5, it further includes an integrated process programming unit. The integrated process programming unit includes a process editing module, a process parameter adjustment module and a process storage and recall module, which are used to edit the grinding process steps and adjust the grinding process parameters according to process requirements.

[0079] The process editing module provides a user-friendly graphical or text editing interface, allowing users to input and edit grinding process steps according to specific process requirements. This interface includes various preset grinding operations, such as the starting position, path mode (straight line, curve, circle, etc.), and number of grinding passes. Users can customize the grinding process steps through intuitive operations or by inputting corresponding commands. It features functions for adding, deleting, modifying, and sorting process steps, allowing users to flexibly adjust the grinding operation sequence according to the characteristics of the longitudinal weld seam and grinding requirements of different rotor support workpieces. For example, for complex weld seams, a rough grinding step can be added first, followed by a fine grinding step, to achieve a better grinding effect.

[0080] Process parameter adjustment module: This module allows users to adjust grinding process parameters, which may include grinding depth, grinding speed, grinding tool feed speed, grinding angle, etc. The module performs range checks and validity verification on the input parameters to ensure they are within the limits allowed by the equipment and process. It provides a real-time parameter preview function, allowing users to visually see the impact of parameter changes on the grinding effect. For example, it can display the grinding effect under different parameters through simulated grinding animations or charts, helping users optimize parameter settings.

[0081] Process storage and recall module: Stores user-edited grinding process steps and parameters, which can be stored in a local database or file system. Stored information includes the process name, applicable workpiece types, detailed process steps, and parameter lists, facilitating reuse in subsequent grinding operations. It supports quick process recall; users can select the appropriate process program from the stored process library based on the workpiece type and grinding requirements, avoiding repetitive input and adjustments.

[0082] Therefore, the working principle of the integrated process programming unit is as follows: Process Editing: Users, through the process editing module, formulate the grinding process steps based on the shape, size, material, and final grinding requirements of the longitudinal weld seam of the rotor support. First, a suitable grinding starting position is selected, considering the workpiece geometry and robot accessibility to ensure the grinding operation completely covers the weld seam area to be processed. Then, the grinding path mode is determined; for example, a straight grinding path mode is selected for long, straight weld seams; a circular or curved grinding path mode is selected for circular or arc-shaped weld seams. During the editing process, specific details of each grinding step can be set, such as specifying the grinding length and direction within a straight grinding path.

[0083] Process Parameter Adjustment: Based on the grinding process steps, users can adjust specific parameters using the process parameter adjustment module. For the grinding depth, settings are made according to the weld depth and grinding quality requirements, taking into account material removal and surface roughness to determine a reasonable grinding depth value. When adjusting the grinding speed and the feed speed of the grinding tool, the hardness of the material and the performance of the tool must be considered to ensure a balance between efficiency and quality in the grinding process. Simultaneously, the grinding angle adjustment should be based on the shape of the weld and the geometry of the grinding tool to ensure optimal contact between the tool and the workpiece during grinding, avoiding over-cutting or under-cutting. These parameter adjustments will simulate the grinding process based on preset algorithms and physical models, displaying the possible effects on the interface. Users can continuously optimize the parameters based on the simulation results.

[0084] Process storage and retrieval: After process editing and parameter adjustment are completed, the process storage and retrieval module stores the entire process information. Stored processes can be categorized by process name, workpiece type, and other information for easy subsequent retrieval. In subsequent grinding operations, users can retrieve the corresponding process program from the stored process library based on the information of a new workpiece. Upon retrieval, the stored process steps and parameter information are sent to the execution unit and path planning unit, providing specific guidance for the grinding operation.

[0085] Example 7 This embodiment discloses an automated grinding control system for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiments 1, 2, 3, 4, 5, or 6, it further includes a fault alarm unit. The fault alarm unit includes a fault recording and storage module and a fault recovery and task recovery module, which are used to monitor system fault conditions during the grinding operation. When the automated grinding control system of the robot experiences a fault interruption, the audible and visual alarm unit provides an audible and visual alarm reminder, and after the fault is cleared, it triggers the system to continue task execution from the interruption position.

[0086] The fault monitoring module establishes communication connections with various units in the system (such as the execution unit, vision inspection unit, point cloud integration unit, etc.) to collect their status information and operational data in real time. This information may include, but is not limited to, the robot's motion state, the working status of the grinding tool, sensor measurement data, data processing results, and the communication status of each unit. It possesses fault judgment logic, analyzing the collected data to determine if any abnormalities have occurred. For example, it monitors whether the robot's joint torque exceeds the normal range, whether the grinding tool's rotation speed is abnormal, and whether data transmission is interrupted. Simultaneously, it determines whether preset fault thresholds and conditions have been met to trigger a fault alarm.

[0087] Fault logging and storage module: Once a fault is detected, this module stores fault information (including the time of occurrence, fault type, fault location, and related parameters) on a local storage device or in a database. The stored information serves as a basis for subsequent fault analysis and system maintenance, helping technicians find the root cause of the fault and optimize system performance. Fault information can be categorized for storage, such as mechanical faults, electrical faults, software faults, and communication faults, facilitating subsequent querying and statistical analysis.

[0088] Fault Recovery and Task Resumption Module: After a fault is eliminated, this module coordinates the system to resume task execution from the point of interruption, based on the location of the fault and the system state at the time of the fault. It collaborates with the execution unit, path planning unit, and other components to restore the system's operational state, ensuring that the grinding operation can restart from the point of interruption. Different recovery strategies may be needed depending on the type of fault. For simple faults, such as temporary communication interruptions, the recovery strategy might be to re-establish the communication connection and continue the task; for complex faults, such as grinding tool failure, it may be necessary to wait for tool replacement or repair before replanning subsequent tasks based on the existing trajectory and task progress.

[0089] Therefore, the working principle of the fault alarm unit is as follows: Fault Monitoring and Alarm Triggering: The fault monitoring module continuously monitors the operating status and data of each unit in the system. For example, when the robot is performing a grinding operation, it monitors information such as the robot's joint torque, speed, and position. If a sudden increase in joint torque exceeding a set threshold is detected, it may mean that the robot has encountered abnormal resistance, which could be caused by workpiece position deviation or a malfunction in the grinding tool. When an abnormal situation is detected and reaches the fault alarm standard, the fault monitoring module will trigger a fault alarm operation, sending corresponding control signals to the audible and visual alarm unit to issue an audible and visual alarm. Simultaneously, the fault information is transmitted to the fault recording and storage module for storage.

[0090] Audible and visual alarm operation: After receiving the control signal from the fault alarm unit, the audible and visual alarm unit will activate the alarm according to the preset alarm mode. The audible and visual alarm can use different sound frequencies, sound intensities, and light colors to distinguish different types of faults, allowing operators to quickly identify the severity of the fault. For example, a severe fault may be accompanied by a loud, high-frequency sound and a flashing red light; a minor fault may be accompanied by a soft, low-frequency sound and a yellow light.

[0091] Task recovery after fault elimination: After the operator resolves the fault, the system will feed back the fault elimination information to the fault recovery and task recovery module. This module will first read the fault record, analyze the fault type and the system state at the time of the fault. For task recovery, it will collaborate with the execution unit and path planning unit. If the grinding process is unexpectedly interrupted, it will retrieve the robot position, grinding trajectory, and process progress information before the interruption from the database, coordinate with the execution unit to adjust the robot to the interrupted position, and continue to execute the subsequent grinding task according to the existing grinding trajectory and process requirements.

[0092] Example 8 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports using robots. As a preferred implementation of this embodiment, it employs any one of the automated grinding control systems for longitudinal weld seams of rotor supports described in embodiments 1 to 7 above, and specifically includes the following steps: S1, the rotor support workpiece is positioned manually. The workpiece is then manually adjusted to the predetermined position and locked. The robot is then moved to the initial weld position. This step relies heavily on the operator's experience and skill to precisely place the rotor support workpiece in a position suitable for subsequent automated grinding operations, ensuring its stability and accuracy. The predetermined position is determined based on the equipment layout and grinding process requirements. For example, it ensures that the longitudinal weld seam of the rotor support workpiece is within the working range of the vision inspection unit and the robot, and that there will be no positional shift or instability during subsequent grinding. The locking operation can be achieved using mechanical fixing devices such as clamps and bolts to prevent displacement of the rotor support workpiece due to vibration or other external forces during grinding.

[0093] S2. The vision inspection unit is activated to scan the rotor support workpiece, completing the surface model reconstruction and obtaining the original image data of the rotor support workpiece surface. The vision inspection unit scans the weld seam of the rotor support workpiece twice to complete the surface model reconstruction. The first scan may be a preliminary overall scan of the entire rotor support workpiece, aiming to obtain the approximate outline and basic geometric information of the workpiece, providing a basis for subsequent detailed scanning and processing. The second scan will focus more on the weld seam area of ​​the rotor support workpiece, improving the scanning accuracy and detail, so as to better reconstruct the workpiece surface model.

[0094] S3, the point cloud integration unit receives raw image data acquired by the vision inspection unit via the data bus and performs preprocessing on the raw image data, including calculation, coordinate transformation, and filtering. This converts the raw image data into ordered, processable point cloud data, performs meshing, and outputs it to a file. The calculation process converts the raw image data into point cloud data with spatial coordinates, transforming the raw signal from the laser scanner into meaningful 3D coordinate information. The coordinate transformation operation unifies point cloud data obtained from different scanners or different scanning stages into a standard coordinate system for subsequent processing. The filtering operation removes noise and outliers, improving data quality; algorithms such as Gaussian filtering and median filtering can be used. Meshing transforms the processed point cloud data into a 3D mesh model, connecting discrete point clouds into a mesh, facilitating subsequent calculations and analysis, as well as storage and visualization. The meshed file can be in common 3D file formats such as STL and OBJ.

[0095] S4. The path planning unit receives near-end point cloud data from the point cloud integration unit, constructs a 3D model of the rotor support workpiece, analyzes the geometry and weld location of the rotor support workpiece, generates environmental data, and calculates a suitable end-effector trajectory for grinding, thereby generating trajectory information. The workpiece model constructed using near-end point cloud data is more accurate and can reflect the detailed features of the workpiece. The generation of environmental data considers the environmental information around the workpiece, including the layout of the equipment and the robot's workspace, ensuring that the grinding operation does not interfere with the surrounding environment. The calculation of the grinding end-effector trajectory is based on the grinding process requirements, such as the grinding depth, width, shape and size of the grinding tool, combined with the workpiece model and environmental data, to generate the motion trajectory of the robot's grinding tool end. This trajectory must ensure that the grinding tool can accurately conform to the weld surface while meeting the grinding quality requirements.

[0096] S5. The execution unit reads the trajectory information and plans the robot's specific motion trajectory based on the robot's kinematic principles.

[0097] S6 uses collision analysis and obstacle avoidance units to simulate and verify the motion trajectory. If the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory. If the simulation verification passes, the execution unit is triggered to send the motion trajectory data to the robot for polishing.

[0098] S7 monitors system malfunctions in real time during the polishing process. When the automated robotic polishing control system experiences a malfunction and interruption, the system activates an audible and visual alarm to alert the operator. Once the malfunction is resolved, the system resumes operation from the interrupted position. Malfunctions may include robot failures, sensor failures, and communication failures. Upon malfunction, the alarm unit triggers an audible and visual alarm to notify the operator. After troubleshooting, the system restores itself to its pre-malfunction state based on stored information (such as the robot's interruption position and completed polishing trajectory), ensuring the polishing operation resumes from the interrupted position and maintaining the continuity and integrity of the entire polishing process.

[0099] S8. Use the roller frame automatic positioning device to rotate the rotor support workpiece to the next weld position, and repeat the above steps S1-S7 until all welds of the current rotor support workpiece are ground.

[0100] In this technical solution, the entire grinding control method is a cyclical process. Starting from the positioning of the rotor support workpiece, a series of steps are taken to complete the grinding of one weld seam. Then, the rotor support workpiece is rotated to the position of the next weld seam and grinding continues until all weld seams are completed. Throughout the process, multiple functional units work together, such as a vision inspection unit, a point cloud integration unit, a path planning unit, an execution unit, and a fault alarm unit. These units complete their respective tasks sequentially according to the above steps, while also cooperating and transmitting information to each other to jointly achieve automated grinding operations.

[0101] Example 9 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports. As a preferred implementation of this embodiment, based on embodiment 8, step S2 of which involves scanning the rotor support workpiece includes one far-end scan and one near-end scan.

[0102] When performing remote scanning, the primary purpose is to detect the workpiece's posture. By scanning the overall contour of the workpiece, its approximate shape and posture information can be obtained from a distance, thus roughly locating the grinding area. This scanning method has a large coverage area but relatively low accuracy. It is suitable for initially determining the workpiece's position and posture, providing a rough reference for subsequent operations and avoiding accidents such as collisions between the robot and the workpiece.

[0103] Near-end scanning focuses on precise measurement of the grinding points. Scanning close to the workpiece yields higher resolution and accuracy, allowing for precise measurement of the workpiece's details. Based on these measurements and previously obtained overall contour data, the grinding trajectory can be planned more accurately, ensuring the grinding path closely conforms to the workpiece surface and meets the precision requirements of the grinding process.

[0104] Example 10 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports using robots. As a preferred implementation of this embodiment, based on embodiment 8 or 9, step S2 and step S4 of generating trajectory information include the following steps: S41. Based on the point cloud data, two dividing lines L1 and L2 are obtained, and the point cloud is divided using these lines. The area between the two lines is the grinding area, and the area outside the lines is two planar regions. This divides the complex rotor support workpiece surface into different functional areas, clearly defining the grinding and non-grinding planar regions, providing a clear area range for subsequent precise grinding operations, and helping to improve the targeting and accuracy of grinding.

[0105] S42, perform plane segmentation fitting on the point cloud data of the grinding area to obtain two extended intersecting planes P1 and P2, and calculate the normal vectors Normol1 and Normol2 of planes P1 and P2 respectively. Specifically: First, the Random Sample Consensus (RANSAC) algorithm can be used to fit planes from the point cloud data of the grinding area. This algorithm calculates the plane model and evaluates the number of interior points through multiple random samplings, and finally obtains the best-fitting plane; then, the least squares method is used to fit the point cloud data that satisfies the plane model, and the coefficients of the plane equation are solved to obtain planes P1 and P2; finally, the normal vectors of the planes are calculated through vector operations based on the coefficients of the plane equations. This provides the foundation for subsequent calculation of the intersection line and grinding trajectory. The determination of the two intersecting planes can more accurately describe the geometry of the grinding area, and the normal vectors are used to determine the direction of the planes, which is crucial for calculating the direction of the intersection line, the center of the arc, and trajectory planning. The principle is as follows: the RANSAC algorithm can effectively eliminate noise and outlier interference through random sampling and iteration, and find the plane model that best represents the data; the least squares method obtains the optimal plane fitting parameters by minimizing the sum of squared distances from points to the plane; the normal vector is perpendicular to the plane, reflecting the plane's orientation information.

[0106] S43, calculate the intersection line I of planes P1 and P2, and obtain the direction vector (Direction) of the intersection line I based on the cross product of normal vectors Normol1 and Normol2. Specifically: First, solve the equations of planes P1 and P2 simultaneously to obtain the parametric equation of the intersection line; then, calculate the cross product of normal vectors Normol1 and Normol2 according to the definition of the cross product. This determines the intersection line and its direction of the two intersecting planes. The intersection line I and the direction vector (Direction) are crucial for determining the grinding arc and grinding trajectory, defining the basic direction of the grinding operation in space. The principle is that the intersection line of two planes is the set of points that simultaneously satisfy the equations of both planes; the expression for the intersection line can be obtained by solving the simultaneous equations. The result of the cross product is perpendicular to the two vectors involved in the cross product; therefore, the direction of the cross product of the normal vectors is the direction of the intersection line.

[0107] S44, based on the structure of the area to be ground on the rotor support workpiece, preset the grinding arc radius R; take a point on the intersection line I. , command point ,make The starting point is the point with the smallest coordinate value in the direction of the vector direction within the grinding area. Specifically: First, based on the design requirements of the rotor support workpiece, the weld shape, and the characteristics of the grinding tool, the radius R of the grinding arc is predetermined. Then, the points on the intersection line I are traversed to find the point with the smallest coordinate value in the direction of the vector direction as the starting point. The radius and starting point of the grinding arc are determined in this way. The radius of the grinding arc is predetermined based on the structure of the area to be ground and the grinding process requirements. The selection of the starting point provides a reference for subsequent calculations of the center and trajectory of the grinding arc. The principle is that the setting of the grinding arc radius must comprehensively consider the geometry of the workpiece and the grinding effect to ensure that the grinding achieves the expected quality requirements. Selecting the minimum point in a specific direction on the intersection line as the starting point ensures the orderliness and integrity of the grinding trajectory.

[0108] S45, based on the junction of the arc radius R. Obtain the center O of the polished arc and the polished arc. Specifically: First, the `point_on_line` function is used to calculate a point at a specified location on the line based on the parametric equation of the line, given a point and direction vector. Then, using geometric relationships and the known radius R of the arc, combined with the starting point and the direction of the intersection line, the coordinates of the center O are obtained through mathematical calculations. Finally, with the center O as the center and radius R, and considering the positional relationship between planes P1 and P2, the starting and ending points of the grinding arc are determined, thus defining the entire arc. This precise determination of the center and shape of the grinding arc is crucial, as the center and arc are essential components of the grinding trajectory, directly affecting the accuracy and effect of the grinding. The principle is that the `point_on_line` function, based on the parametric equation of the line, can find points at arbitrary locations on the line by changing the parameters. In determining the center and arc, principles of plane and spatial geometry are utilized; based on the known radius, starting point, and plane information, the positions of the center and arc are calculated through geometric relationships.

[0109] S46, determine the single-step grinding trajectory, i.e., the point of pass. Construct the perpendicular line of intersection line I to plane P1. ,perpendicular It intersects the dividing line L1 at point Passing point Draw the perpendicular line of the intersection line in the plane P2. ,perpendicular It intersects the dividing line L2 at point Then a complete trajectory along the direction of the first cross section is a straight line. Polishing the arc and straight line Specifically: First, based on the normal vectors of the intersection line I and planes P1 and P2, the perpendicular line of the intersection line in the plane is calculated using the property of perpendicular vectors. Then, the equations of the perpendicular line and the dividing line are solved simultaneously to find the coordinates of the intersection point. Finally, the obtained intersection point and the grinding arc are connected to form a complete single-step grinding trajectory. This determines a complete grinding trajectory in a cross-sectional direction. The single-step grinding trajectory is the basic unit of the entire grinding trajectory; multiple single-step trajectories combined can cover the entire grinding area. The principle is that in a plane, the dot product of two perpendicular lines is zero; this property can be used to calculate the perpendicular line of the intersection line in the plane. Solving for the intersection point by solving simultaneous equations is based on the geometric principle of the intersection of two lines. Combining different line segments and arcs into a trajectory is to achieve continuous grinding operations from the planar area to the grinding area and back to the planar area.

[0110] S47, determine all grinding paths, i.e., the preset extension distance parameter step, and set the current point Extend the vector Direction along the intersection line I by the distance step, and then let... After obtaining a new point Ni, the process returns to step S45 until grinding trajectories on all cross-sections are obtained. Specifically: First, vector operations are used to add the direction vector Direction to the current point and multiply it by the extension distance parameter step to obtain a new point Ni; then, using the new point Ni as the starting point, the process returns to step S45, repeating the calculation of the center and arc of the grinding arc, as well as the single-step grinding trajectory, until the extension distance exceeds the maximum range of the grinding area in the intersection direction. Through continuous extension and repeated calculation, all grinding trajectories covering the entire grinding area are obtained, ensuring comprehensive grinding of the longitudinal weld seam of the rotor support workpiece. The principle is: by continuously extending the starting point along the intersection direction, the single-step grinding trajectory is repeatedly generated using the previously determined calculation method, achieving coverage of the entire grinding area. The extension distance parameter step controls the density of the trajectory; a reasonable selection of step can balance grinding accuracy and efficiency.

[0111] S48: The straight lines on planes P1 and P2 are moved inward by a distance R(t) along their respective normal directions, and the arcs are moved inward by a distance R(t) towards the center O. The resulting lines are the trajectory information generated by the path planning unit. Specifically: First, for the straight lines on planes P1 and P2, each point on the line is moved by a distance R(t) along the normal direction of the plane. Then, for the arcs being ground, each point on the arc is moved by a distance R(t) towards the center O. This adjustment considers the radius of the grinding tool and ensures that the grinding tool accurately covers the grinding area, avoiding under- or over-grinding. The principle is that the grinding tool has a certain radius; to ensure the grinding effect, the trajectory needs to be shrunk inward by a distance equal to the tool radius. Through vector operations and geometric transformations, the points on the straight lines and arcs are moved, achieving the inward movement of the trajectory.

[0112] Example 11 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports using robots. As a preferred implementation of this embodiment, based on embodiment 10, step S41, obtaining the dividing lines L1 and L2, is carried out using the following method: Method 1, based on workpiece design information If a precise 3D CAD model of the rotor support workpiece is available beforehand, the geometric information of the weld and surrounding area can be extracted from it. Based on this information, and combined with the requirements of the grinding area, the positions of L1 and L2 can be determined on the model. For example, when the weld is located at the intersection of two planes, two dividing lines can be drawn on the model based on the plane boundaries and the width of the weld. The positional information of these lines can then be converted to the coordinate system corresponding to the point cloud data. Based on the design drawings of the rotor support workpiece, the theoretical dimensional parameters such as the width and position of the weld can be obtained. Using these parameters, the positions of L1 and L2 can be calculated in the point cloud data. For example, if the center position and width of the weld are known, two parallel lines can be determined on both sides of the center position according to the width value as L1 and L2.

[0113] Method 2: Point cloud data processing Cluster analysis is performed on the acquired point cloud data to separate the point clouds belonging to the weld area and the planar area. Clustering algorithms can include density-based DBSCAN or hierarchical clustering algorithms. After clustering, L1 and L2 are determined near the cluster boundaries. For example, using the DBSCAN algorithm to divide the point cloud into weld and planar clusters, straight lines L1 and L2 are drawn at the boundaries of the two clusters, thus defining the grinding area and the planar area. Edge detection algorithms, such as curvature-based edge detection methods, are used to find edge points in the point cloud data. These edge points may represent the intersection of the weld and the plane. Based on the distribution of edge points, L1 and L2 are fitted. For example, the curvature of each point in the point cloud data is calculated, points with larger curvature are considered edge points, and then these edge points are fitted using the least squares method to obtain two straight lines L1 and L2.

[0114] Method 3, based on visual detection and feature recognition If the visual inspection unit acquires image information simultaneously during scanning, image recognition technology can be used to determine L1 and L2. For example, deep learning algorithms can be used to identify and locate weld seams in an image, then segmentation lines can be drawn on the image, and the positional information of these lines can be mapped onto point cloud data. Specific geometric features, such as corner points and line segments, can be extracted from the point cloud data, and the positions of L1 and L2 can be determined through feature matching. For example, some obvious corner features can be extracted near the weld seam area, and two segmentation lines can be fitted based on the relative positional relationship of these corner points.

[0115] Method 4: Determining through manual interaction Operators can observe point cloud data and the 3D model of the workpiece on a visual interface and manually draw L1 and L2 lines based on experience and actual needs. While this method has a degree of subjectivity, it allows for flexible determination of the dividing lines' positions in complex situations. Operators can determine the positions of L1 and L2 by inputting parameters such as the line's start point, end point, and slope. The system then draws the corresponding lines in the point cloud data based on the input parameters.

[0116] This technical solution preferably adopts Method 2, which uses edge detection based on point cloud data to obtain L1 and L2 equations. First, curvature calculation is performed using the `estimate_normals` method to calculate the normals of the point cloud. Then, the curvature of each point is obtained by calculating the trace of the covariance matrix; points with larger curvatures typically represent edge points. Next, edge point selection is performed: a curvature threshold is set, and points with curvature greater than this threshold are selected as edge points. Finally, line fitting is performed: the RANSAC algorithm is used to segment the edge points into planes, obtaining the parametric equations L1 and L2 of two straight lines.

[0117] Example 12 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports using robots. As a preferred implementation of this embodiment, based on embodiment 10 or 11, in step S45, the center O of the grinding arc and the grinding arc are obtained. Includes the following steps: S451, with point Draw a circle with center D and virtual radius D. The intersection points of this circle with planes P1 and P2 are respectively... and This provides the foundation for subsequently determining the endpoints and center of the polished arc. By drawing a circle with a specific center and a specific radius, and finding the intersection of this circle and the plane, the possible endpoints of the polished arc can be located in space. The principle is as follows: a circle is the set of points whose distance to a fixed point (center) is equal to a fixed length (radius). A plane is the set of points that satisfy a linear equation. By solving the equations of both, the points that simultaneously satisfy the conditions of the circle and the plane can be found, which are their intersection points.

[0118] S452, passing Extend a distance of length R along the normal vector Normol1, passing through Extend the normal vector Normol2 by a distance R, so that the normal vectors Normol1 and Normol2 intersect at point O, which is the center O of the polished arc. This precisely determines the center of the polished arc. Using the plane's normal vector and the specific extension length, a center point satisfying the conditions is found based on geometric relationships. This center will be used to construct the polished arc subsequently. The principle is: a normal vector is perpendicular to the plane, and extending it by a specific length defines a straight line in space. The intersection of two straight lines is the point that satisfies the equations of both lines simultaneously. By solving the simultaneous equations, this intersection point, which is the center of the polished arc, can be found.

[0119] S453, draw a circle based on the center O and the radius R of the arc to obtain the polished arc. At this point, the angle between the normal vectors Normol1 and Normol2 is θ, and the virtual radius D = R * tan(θ / 2). The principle is as follows: a circle is the set of points whose distance to the center is equal to the radius, and in three-dimensional space, it is represented as a sphere. By determining the set of points on the sphere that satisfy a specific angle range, the arc can be obtained. The formula for calculating the virtual radius is derived based on the geometric properties of isosceles triangles and trigonometric relationships. In an isosceles triangle, this formula can be obtained using the half-angle formula and the relationship between opposite and adjacent sides.

[0120] Example 13 This embodiment discloses an automated grinding control method for longitudinal weld seams of rotor supports using robots. As a preferred implementation of this embodiment, based on any one of embodiments 8 to 12, step S5, the method for planning the specific motion trajectory of the robot, is as follows: the robot trajectory information is stored in a text file, with each point data recorded on a separate line. Each line records the robot's X, Y, Z, A, B, and C values, separated by spaces. The text file in this data format is converted into an executable program for the robot using the SDK provided by the robot, i.e., the specific motion trajectory program. Here, X, Y, and Z represent the axis position coordinates of the robot's end effector in space, and A, B, and C represent the attitude angles of the robot's end effector on the X, Y, and Z axes, to determine the orientation of the robot's end effector.

[0121] The operation steps are as follows: Obtain the robot's trajectory information from the path planning unit. This information should include the spatial coordinates (X, Y, Z) and attitude angles (A, B, C) of the robot's end effector at each movement point. Store the trajectory information in a text file according to a specified format, with each line recording the information for one point, and the values ​​separated by spaces. Use the software development kit (SDK) provided by the robot manufacturer to convert the text file containing the trajectory information into a program that the robot can recognize and execute. Upload the converted program to the robot controller and start the robot to execute the specific motion trajectory.

[0122] The principle is as follows: robot trajectory information is stored in a text file using a uniform format (one point per line, with values ​​separated by spaces), facilitating subsequent processing and interaction with the robot SDK. The SDK provided by the robot manufacturer has the function of parsing text files in a specific format and converting them into executable programs for the robot. Based on the robot's kinematic model and control protocol, the SDK converts the point information in the text file into motion commands for each joint of the robot. The X, Y, and Z coordinates determine the position of the robot's end effector in space, while the A, B, and C attitude angles describe the orientation of the tool. These six parameters allow for precise control of the robot's end effector's position and attitude in three-dimensional space.

Claims

1. A robotic automated grinding control system for longitudinal weld seams of rotor supports, characterized in that, include: The visual inspection unit includes several laser scanners and corresponding image acquisition devices, which are used to scan the rotor support workpiece, complete the reconstruction of the surface model of the rotor support workpiece, and obtain the original image data of the surface of the rotor support workpiece. The point cloud integration unit is used to receive the raw image data collected by the vision inspection unit through the data bus, and to perform preprocessing on the raw image data, including solving, coordinate transformation and filtering, to convert the raw image data into ordered and processable point cloud data, and to perform meshing processing and output it to a file. The path planning unit receives near-end point cloud data from the point cloud integration unit, constructs a 3D model of the rotor support workpiece, analyzes the geometry and weld position of the rotor support workpiece, generates environmental data, and calculates the end surface trajectory suitable for grinding, thereby generating trajectory information. The generated trajectory information includes: Two dividing lines L1 and L2 are obtained based on edge detection of point cloud data, and the point cloud is divided by dividing lines L1 and L2. The area between the two lines is the grinding area, and the area outside the two lines is two planar areas. Plane segmentation fitting is performed on the point cloud data of the polishing area to obtain two extended intersecting planes P1 and P2, and the normal vectors Normol1 and Normol2 of planes P1 and P2 are calculated respectively. Calculate the intersection line I of planes P1 and P2, and obtain the direction vector Direction of the intersection line I based on the cross product of the normal vectors Normol1 and Normol2. Based on the structure of the area to be ground on the rotor support workpiece, preset the grinding arc radius R; take a point on the intersection line I. , command point ,make The point with the smallest radius in the direction of the vector Direction is the area to be polished. Based on the arc radius R, the junction point Obtain the center O of the polished arc and the polished arc. ; Determine the single-step polishing trajectory, i.e., the point of pass. Construct the perpendicular line of intersection line I to plane P1. ,perpendicular It intersects the dividing line L1 at point Passing point Draw the perpendicular line of the intersection line in the plane P2. ,perpendicular It intersects the dividing line L2 at point Then a complete trajectory along the direction of the first cross section is a straight line. Polishing the arc and straight line ; Determine all grinding paths, i.e., the preset extension distance parameter step, and set the current point Extend the vector Direction along the intersection line I by the distance step, and then let... After obtaining the new point Ni, return to step S45 until the grinding trajectory on all cross sections is obtained. The straight lines on planes P1 and P2 are reduced inward by a distance R(t) along their respective normal directions, and the arcs are reduced inward by a distance R(t) towards the center O. The lines obtained after this reduction are the trajectory information generated by the path planning unit. The execution unit is used to read the trajectory information generated by the path planning unit, plan the specific motion trajectory of the robot according to the kinematic principle of the robot, and send the motion trajectory data to the robot to perform the polishing operation. The database is used to store the operational data of the robot's automated grinding control system, as well as the intermediate and result data generated by each functional unit. The main control unit is used to read the status information of each functional unit in real time and display this status information intuitively on the interface; at the same time, it provides an operation interface for operators to perform operations on the robot automated grinding control system, including parameter setting and start / stop control.

2. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: It also includes an external interface unit, which provides a production system interface and a positioner interface; the production system interface is used to exchange information and coordinate work with the entire production system; the positioner interface is used to establish communication and control between the robot automated grinding control system and the positioner.

3. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: The execution unit includes a grinding module and a scraping module, which are used to automatically generate the end grinding trajectory and the end scraping trajectory based on three-dimensional point cloud data and grinding process requirements, respectively.

4. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: It also includes a collision analysis and obstacle avoidance unit, which is used to perform operation simulation verification through simulation function before the robot performs grinding operations, so as to eliminate the possibility of collisions. If the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory. If the simulation verification passes, the execution unit is triggered to send motion trajectory data to the robot to perform grinding operations.

5. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: It also includes a self-test and status detection unit for performing system operation self-tests, including robot status and position parameter monitoring, external axis status and position monitoring, and grinding pressure monitoring.

6. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: It also includes an integrated process programming unit, which is used to edit the grinding process steps and adjust the grinding process parameters according to process requirements.

7. The automated grinding control system for longitudinal weld seams of rotor supports as described in claim 1, characterized in that: It also includes a fault alarm unit to monitor system faults during the grinding operation; when the robot automated grinding control system is interrupted by a fault, the audible and visual alarm unit will provide an audible and visual alarm reminder, and after the fault is cleared, it will trigger the system to continue the task execution from the interrupted position.

8. A robotic automated grinding control method for longitudinal weld seams of rotor supports, characterized in that, Includes the following steps: S1, the rotor support workpiece is in place, the rotor support workpiece is manually adjusted to the predetermined position and locked, and the robot is moved to the initial position of the weld. S2. Start the vision inspection unit to scan the rotor support workpiece, complete the reconstruction of the surface model of the rotor support workpiece, and obtain the original image data of the surface of the rotor support workpiece. S3. The point cloud integration unit receives the raw image data collected by the visual inspection unit through the data bus, and performs preprocessing on the raw image data, including solving, coordinate transformation and filtering, to convert the raw image data into ordered and processable point cloud data, and performs meshing processing and outputs it to a file. S4. The path planning unit receives near-end point cloud data from the point cloud integration unit, constructs a three-dimensional model of the rotor support workpiece, analyzes the geometry and weld position of the rotor support workpiece, generates environmental data, and calculates the end surface trajectory suitable for grinding, thereby generating trajectory information. Generating trajectory information specifically includes the following steps: S41, based on edge detection of point cloud data, two dividing lines L1 and L2 are obtained, and the point cloud is divided by dividing lines L1 and L2. The area between the two lines is the grinding area, and the area outside the two lines is two planar areas. S42, perform plane segmentation fitting on the point cloud data of the polishing area to obtain two extended intersecting planes P1 and P2, and calculate the normal vectors Normol1 and Normol2 of planes P1 and P2 respectively. S43, calculate the intersection line I of planes P1 and P2, and obtain the direction vector Direction of the intersection line I based on the cross product of normal vectors Normol1 and Normol2; S44, based on the structure of the area to be ground on the rotor support workpiece, preset the grinding arc radius R; take a point on the intersection line I. , command point ,make The point with the smallest radius in the direction of the vector Direction is the area to be polished. S45, based on the junction of the arc radius R. Obtain the center O of the polished arc and the polished arc. ; S46, determine the single-step grinding trajectory, i.e., the point of pass. Construct the perpendicular line of intersection line I to plane P1. ,perpendicular It intersects the dividing line L1 at point Passing point Draw the perpendicular line of the intersection line in the plane P2. ,perpendicular It intersects the dividing line L2 at point Then a complete trajectory along the direction of the first cross section is a straight line. Polishing the arc and straight line ; S47, determine all grinding paths, i.e., the preset extension distance parameter step, and set the current point Extend the vector Direction along the intersection line I by the distance step, and then let... After obtaining the new point Ni, return to step S45 until the grinding trajectory on all cross sections is obtained. S48, the straight lines on planes P1 and P2 are reduced inward by a distance R(t) along their respective normal directions, and the arcs are reduced inward by a distance R(t) towards the center O. The lines obtained after this reduction are the trajectory information generated by the path planning unit. S5. The execution unit reads the trajectory information and plans the robot's specific motion trajectory based on the robot's kinematic principles. S6. The motion trajectory is simulated and verified by the collision analysis and obstacle avoidance unit. If the simulation verification fails, the path planning unit needs to be called to re-plan the path, or the execution unit needs to be called to optimize and adjust the trajectory. If the simulation verification passes, the execution unit is triggered to send the motion trajectory data to the robot for polishing. S7 monitors system malfunctions in real time during the grinding process; when the robot's automated grinding control system is interrupted, it controls the audible and visual alarm unit to issue an audible and visual alarm reminder, and triggers the system to continue task execution from the interrupted position after the fault is cleared; S8. Use the roller frame automatic positioning device to rotate the rotor support workpiece to the next weld position, and repeat the above steps S1-S7 until all welds of the current rotor support workpiece are ground.

9. The automated grinding control method for longitudinal weld seams of rotor supports as described in claim 8, characterized in that, In step S2, scanning the rotor support workpiece includes one far-end scan and one near-end scan; wherein, the far-end scan is to detect the posture of the rotor support workpiece and initially locate the grinding point; the near-end scan is to accurately measure the grinding point.

10. The automated grinding control method for longitudinal weld seams of rotor supports as described in claim 8, characterized in that, In step S45, the center O of the polished arc and the polished arc are obtained. Includes the following steps: S451, with point Draw a circle with center D and virtual radius D. The intersection points of this circle with planes P1 and P2 are respectively... and ; S452, passing Extend a distance of length R along the normal vector Normol1, passing through Extend the normal vector Normol2 by a distance of R, such that the normal vectors Normol1 and Normol2 intersect at point O with a length of R, which is the center O of the polished arc. S453, draw a circle based on the center O and the radius R of the arc to obtain the polished arc. At this point, the angle between the normal vectors Normol1 and Normol2 is θ, and the virtual radius D = R * tan(θ / 2).

11. The automated grinding control method for longitudinal weld seams of rotor supports as described in claim 8, characterized in that, Step S5, the method for planning the robot's specific motion trajectory, is as follows: the robot trajectory information is stored in a text file, with each point data recorded on a separate line. Each line records the robot's X, Y, Z, A, B, and C values, separated by spaces. The text file containing this data is then converted into an executable program for the robot, i.e., the specific motion trajectory program, using the SDK provided by the robot. Here, X, Y, and Z represent the axis position coordinates of the robot's end effector in space, and A, B, and C represent the attitude angles of the robot's end effector on the X, Y, and Z axes, to determine the orientation of the robot's end effector.

Images