Error compensation control method for heavy load forming robot based on strain feedback

Abstract

This invention relates to an error compensation control method for a heavy-duty forming robot based on strain feedback, comprising the following steps: establishing a dynamic model of the heavy-duty forming robot; obtaining the actual strain of the connecting rods in each branch and the actual motor angle of each servo motor; calculating the actual displacement of each slider; solving for the motion state, deformation state, and stress state of the heavy-duty forming robot; calculating the theoretical motor torque of each servo motor based on the motion state, deformation state, and stress state; calculating the compensation value of each slider displacement based on the motion error between the theoretical and actual motion of the enveloping mold; and performing closed-loop control on each servo motor based on the theoretical motor torque, the compensation value of the slider displacement, the actual motor angle, and the actual motor speed to compensate for the motion error of the enveloping mold in real time. This invention does not significantly reduce the structural stiffness of the robot and can reflect the stress and deformation state of the robot in real time, significantly improving the control performance of the heavy-duty forming robot.

Description

Technical Field

[0001] This invention relates to the field of forming equipment and robot control technology, and more specifically, to an error compensation control method for heavy-duty forming robots based on strain feedback. Background Technology

[0002] Thin-walled, highly ribbed components combine lightweight design with high load-bearing capacity, and are widely used in aerospace, rail transportation, and high-end equipment manufacturing. These components typically feature thin skins, high ribs, and complex local features, placing high demands on the load-bearing capacity, motion accuracy, and dynamic control performance of forming equipment.

[0003] To achieve high-performance manufacturing of thin-walled, high-rigidity components, heavy-duty forming robots with multi-degree-of-freedom enveloping motion capabilities can be employed. These robots typically bear significant forming loads, and during the forming process, the moving platform, connecting rods, sliders, and support structures experience non-negligible elastic deformation. This elastic deformation leads to deviations between the actual motion of the enveloping mold and the theoretically planned motion, thus affecting the forming accuracy of the component.

[0004] Traditional force feedback control methods utilize force sensors with relatively low stiffness compared to heavy-duty forming robots. Installing these sensors on the mold for direct force feedback reduces the overall robot stiffness, leading to decreased control precision and hindering the high-precision forming of thin-walled, highly ribbed components. Due to significant differences in precision control between heavy-duty and light-duty robots, traditional control methods result in reduced robot control accuracy, consequently impacting the performance of thin-walled, highly ribbed components. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a strain feedback-based error compensation control method for heavy-duty forming robots that does not significantly reduce the structural stiffness of the robot and can reflect the stress and deformation state of the robot in real time, thereby significantly improving the control performance of heavy-duty forming robots.

[0006] The technical solution adopted by this invention to solve its technical problem is as follows: Constructing an error compensation control method for a heavy-duty forming robot based on strain feedback. The heavy-duty forming robot includes an upper base, a moving platform, and six branches connecting the upper base and the moving platform. Each branch includes a servo motor, a reducer, a ball screw, a slider, and a connecting rod. The two ends of the connecting rod are respectively connected to the slider and the moving platform via ball joints. The control method includes the following steps: Establish a dynamic model of the heavy-duty forming robot; Obtain the actual strain of the connecting rods in each branch and the actual motor angle of each servo motor; The actual deformation of each link is calculated based on the actual strain, and the actual displacement of each slider is calculated based on the actual motor angle. The actual deformation of each link and the actual displacement of each slider are used as input variables of the dynamic model to solve the motion state, deformation state and force state of the heavy-duty forming robot. The theoretical motor torque of each servo motor is calculated based on the motion state, deformation state, and force state. Based on the motion error between the theoretical and actual motion of the envelope mold, calculate the compensation value for the displacement of each slider. Based on the theoretical motor torque, the compensation value of the slider displacement, the actual motor angle, and the actual motor speed, closed-loop control is performed on each servo motor to compensate for the motion error of the envelope mold in real time.

[0007] According to the above scheme, establishing the dynamic model of the heavy-duty forming robot includes: The slider and the connecting rod are respectively equivalent to two mass components, upper and lower, and an equivalent spring and an equivalent damper are provided between the upper and lower mass components; The moving platform is equivalent to a central mass component and multiple side mass components, and an equivalent spring and an equivalent damper are provided between the central mass component and the multiple side mass components; Based on the mass components, elastic deformation, and damping characteristics of the slider, the connecting rod, and the moving platform, a set of rigid-flexible coupling dynamic equations is established.

[0008] According to the above scheme, the actual strain is obtained by strain gauges installed on each connecting rod, and the strain gauges are installed in the middle region of the connecting rod.

[0009] According to the above scheme, before the strain gauge is installed on the connecting rod, the deformation sensitivity and installation error sensitivity of the slider, connecting rod and moving platform as strain gauge installation positions are analyzed respectively, and the connecting rod with high sensitivity to forming load and low sensitivity to strain gauge installation error is selected as the strain gauge installation position.

[0010] According to the above scheme, the calculation of the actual deformation of each link based on the actual strain includes: Obtain the actual strain of each link in the six branches; Based on the correspondence between the actual strain and the rated length or equivalent length of the connecting rod, the deformation of each connecting rod along its force direction is calculated.

[0011] According to the above scheme, calculating the actual displacement of each slider based on the actual motor angle includes: Obtain the actual motor angles of the six servo motors; Based on the actual motor angle and the transmission relationship between the reducer and the ball screw, the actual displacement of the six sliders is calculated.

[0012] According to the above scheme, the actual deformation of the six links and the actual displacement of the six sliders are used as known input variables to input into the dynamic model to solve for robot state variables, including slider motion variables, link motion variables, moving platform motion variables, slider deformation variables, link deformation variables, moving platform deformation variables, constraint force variables, slider driving force variables, and load variables applied to the moving platform.

[0013] According to the above scheme, the theoretical motor torque is calculated based on the actual motor angle, the first derivative of the actual motor angle, the second derivative of the actual motor angle, the actual strain of the connecting rod, the first derivative of the actual strain of the connecting rod, and the second derivative of the actual strain of the connecting rod.

[0014] According to the above scheme, the calculation of the compensation value for each slider displacement based on the motion error between the theoretical and actual motion of the envelope mold includes: Determine the theoretical and actual poses of the envelope mold; Calculate the motion error rotation of the envelope mold based on the theoretical and actual poses; With the goal of making the motion error spinor approach zero, the compensation value of each slider displacement is solved; The compensation value of the slider displacement is superimposed on the theoretical displacement command of the corresponding slider.

[0015] According to the above scheme, closed-loop control of each servo motor includes position loop control, speed loop control, and current loop control, wherein: In the position loop control, the position error is calculated based on the theoretical motor angle, the compensation value of the slider displacement, and the actual motor angle, and the position loop adjustment amount is generated based on the position error. In the speed loop control, the speed error is calculated based on the theoretical motor speed and the actual motor speed, and the speed loop adjustment amount is generated based on the speed error; In the current loop control, the current error is calculated based on the theoretical current corresponding to the theoretical motor torque and the actual motor current, and the current loop adjustment amount is generated based on the current error. The real-time output torque of each servo motor is calculated based on the theoretical motor torque, the position loop adjustment, the speed loop adjustment, and the current loop adjustment.

[0016] According to the above scheme, the position loop control adopts proportional control, the speed loop control adopts proportional-integral control, and the current loop control adopts proportional-integral-derivative control.

[0017] According to the above scheme, a motion planning step is also included before closed-loop control: Receive user-inputted envelope mold motion parameters; Based on the motion parameters of the envelope mold, kinematic calculations are performed to obtain the theoretical displacement of each slider; The theoretical motor angle and theoretical motor speed of each servo motor are calculated based on the theoretical displacement of each slider.

[0018] The error compensation control method for heavy-duty forming robots based on strain feedback of the present invention has the following beneficial effects: 1. This invention utilizes linkage strain feedback to replace the direct series connection of low-stiffness force sensors for force measurement, which can avoid reducing the overall stiffness of the heavy-duty forming robot due to the installation of force sensors.

[0019] 2. This invention uses a dynamic model to convert the strain of the connecting rod and the angle of the motor into the motion, deformation and force state of the robot, thereby realizing real-time prediction of the state of the heavy-duty forming robot.

[0020] 3. This invention uses slider displacement compensation to make the actual movement of the envelope mold approach the theoretical movement, which can significantly reduce the position and angle errors of the envelope mold.

[0021] 4. This invention combines theoretical motor torque, slider displacement compensation value and three-ring servo control, which can improve the multi-degree-of-freedom envelope forming accuracy of thin-walled high-rib components. Attached Figure Description

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a schematic diagram of the configuration of a heavy-duty forming robot; Figure 2 This is a schematic diagram outlining the dynamic equations and variables of a heavy-duty forming robot. Figure 3 This is a schematic diagram of potential mounting points for strain gauges; Figure 4 This is a diagram analyzing the impact of strain gauge installation errors on the deformation of strain gauge installation points on the slider, connecting rod, and moving platform. Figure 5 This is a schematic diagram of an experimental setup for detecting the strain of a connecting rod and the angle of a motor. Figure 6 This is a schematic diagram illustrating the prediction of motion, deformation, and stress state of a heavy-duty forming robot. Figure 7 This is a schematic diagram of motion error compensation for the envelope tool; Figure 8 This is a schematic diagram of an error compensation control method for heavy-duty forming robots based on strain feedback; Figure 9 This is a schematic diagram of the hardware structure of the control system; Figure 10This is a diagram of the software's control and monitoring interface; Figure 11 These are schematic diagrams of the blank, the enveloping mold, and the constraint mold; Figure 12 The error between slider displacement and envelope mold motion under traditional force feedback method; Figure 13 The error compensation control method based on strain feedback is used to control the slider displacement and the motion error of the envelope mold. Figure 14 This is a schematic diagram of the motion of feature points on the envelope tool under different control methods; Figure 15 It is a comparison diagram of the geometric parameters of two typical thin-walled high-ribbed components, experimental envelope mold, lower mold, sheet blank and final formed thin-walled high-ribbed component; Figure 16 These are graphs showing the measurement results of the forming accuracy of thin-walled, highly ribbed components under different control methods. Detailed Implementation

[0023] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] The error compensation control method for heavy-duty forming robot based on strain feedback of the present invention includes the following steps: S1. Establish a dynamic model of the heavy-duty forming robot.

[0025] Figure 1 The structure of the heavy-duty forming robot is demonstrated. The robot is driven by six servo motors mounted on an upper base. The robot employs a parallel structure with six identical branches. Each branch includes a servo motor, a reducer, a ball screw, a slider, and a connecting rod. The connecting rods have S-joints at both ends, connecting to the slider and the moving platform respectively. The servo motors drive the slider to perform vertical translational motion, which in turn drives the moving platform to perform six degrees of freedom motion through the parallel structure.

[0026] In the dynamics modeling of the heavy-duty forming robot, the slider and the connecting rod are defined as two upper and lower mass components connected by springs and dampers. The moving platform is defined as a central mass component and six side mass components connected by springs and dampers. Figure 2This diagram illustrates the general dynamics equations and variables of a heavy-duty industrial robot. It reveals a total of 516 equations: 72 for slider force balance, 72 for link force balance, 72 for moving platform force balance, and 300 for boundary conditions. The dynamics system has 528 variables: 18 for the slider, 54 for the link, and 54 for the upper mass of the moving platform; 108 for the slider, 108 for the link, and 108 for the moving platform; 30 for the slider, 18 for the upper S-joint, and 18 for the lower S-joint; 6 for the slider's driving force; and 6 for the load applied to the platform. Generally, for forward dynamics, the driving force is known, and the robot load needs to be solved. For inverse dynamics, the robot load is known, and the driving force needs to be solved. In fact, in this dynamic model, since the rotation of the link along its axis can be ignored, the dynamic model can be theoretically solved once any six variables are known.

[0027] S2. Establish a state prediction model for heavy-duty forming robots.

[0028] Indirect and direct force feedback control methods are not suitable for heavy-duty forming robots. Therefore, detecting forces using strain gauges mounted on the machine's moving elements is a good option. Based on the established dynamic model, when the servo motor encoder detects the displacement of six sliders, the machine's stress state can be theoretically solved if the other six deformation variables are known. This means that at least six independent deformation variables should be detected by strain gauges. To determine which moving element is more suitable for mounting strain gauges, three potential mounting points for the strain gauges are considered: sliders, links, and platforms (e.g.,...). Figure 3 As shown, the sensitivity of forming load and installation error is analyzed.

[0029] Calculate the slider based on the corresponding deformation screw. ,link Hedong Platform The deformation at the detection point, i.e.: (1) in The installation phase angle of the strain gauge can be set to... .

[0030] Based on the above formula, such as Figure 4 As shown in (a), the dynamic platform load was obtained. That is, the dynamic platform force and dynamic platform torque on , and The effect of the slider. The results show that, with the slider Hedong platform Compared to the connecting rod Maximum load on the moving platform Most sensitive.

[0031] Considering the installation error of the strain gauge, calculate the deformation of the slider, connecting rod, and moving platform detection points, i.e.: (2) in The installation error of the strain gauge along the Z-axis is considered. The installation error of the strain gauge along the Z-axis.

[0032] Based on the above formula, there are strain gauge installation errors. , , With no strain gauge installation error , , The difference, that is , , We can obtain the following: (3) Strain gauge installation error , , The impact such as Figure 4 As shown in (b) of the diagram. The results indicate that the linkage... Almost zero and minimal, indicating the relationship with the slider. Hedong platform Compared to the connecting rod It is least sensitive to strain gauge installation errors.

[0033] In summary, the deformation of a two-force bar is not only the largest among all moving elements, but it is also the most sensitive to the two main motion errors of the mold envelope. Simultaneously, the deformation of a two-force bar is most sensitive to large forming loads, but insensitive to strain gauge installation errors. Therefore, the link is most suitable for mounting strain gauges, and link strain feedback is suitable for heavy-duty forming robots.

[0034] Based on the above discussion, the strain of the connecting rod is detected by a strain gauge connected in the middle of the connecting rod (e.g., Figure 5 As shown in (a)), the motor angle is detected by an encoder connected to the tail of the servo motor (as shown in [reference]). Figure 5 As shown in (b)). The actual link strain is The actual motor angle is set to Therefore, the connecting rod deforms as follows: (4) The slider displacement is: (5) in It is the transmission ratio of the speed reducer.

[0035] like Figure 6 As shown, when the connecting rod strain and motor angle are detected, the connecting rod deformation and slider displacement can be obtained, meaning 12 variables are known. These 12 known variables serve as input variables for the dynamic model, and the other variables can be solved (the total number of unknown variables is 12+54+54+108+102+108+30+18+18+6+6 = 516, exactly equal to the number of equations), including the kinetic spin of the mass component on the slider. (2 (6 = 12 variables), kinetic spin of the mass component on the connecting rod (9 (6 = 12 variables), the kinematic spinor of the platform-side mass component (9 (6 = 54 variables), the deformation spin of the slider (18 (6 = 108 variables), the deformation spin of the connecting rod (5 6+6 6+6 (6 = 102 variables), the deformation spindle of the moving platform (18 (6 = 108 variables), the constraint force of the slider (5 (6 = 30 variables), constraint forces of the link (3 (6 = 18 variables), constraint force of the lower S-joint (3 (6 = 18 variables), the driving force of the slider (1 6 = 6 variables), loads applied to the moving platform (6 variables)

[0036] Therefore, it was redefined as The theoretical motor torque can be expressed as: and The function, that is: (6) in, For motor angle The first and second derivatives, For the strain of the connecting rod The first and second derivatives.

[0037] In summary, by detecting the strain of the connecting rod using strain gauges and the motor angle using a motor encoder, 516 variables can be solved based on a dynamic model, thereby predicting the machine's motion, deformation, and stress state for machine control.

[0038] S3. Establish an error compensation control model for heavy-duty forming robot based on strain feedback.

[0039] The basic idea for motion error compensation of an enveloping mold is to dynamically adjust the slider displacement to ensure that the actual motion of the enveloping mold matches the theoretical motion. Therefore, the compensation value of the slider displacement can be solved by making the motion error screw of the enveloping mold equal to zero. For example... Figure 7 As shown, the theoretical positions of the connecting rod and the envelope mold are drawn with black lines, while their actual positions are drawn with blue lines. The coordinate system of the envelope mold with motion errors is... The center of the ball on the connecting rod is , , No. The position vectors of the links are The deviation vector of the sphere's center Defined as The calculation is as follows: (7) When the compensation value of the slider displacement is set to The displacement vector of the slider is .therefore, The following can be calculated: (8) Based on equations (50)-(51), we can obtain: (9) Therefore, the compensation value of the slider displacement The above formula can be used to solve for real-time motion error compensation of the envelope mold, i.e. .

[0040] Based on the machine state prediction and motion error compensation discussed above, it is possible to achieve linkage strain feedback and motion error compensation control of the machine. The machine's control system is as follows: Figure 8 As shown, it consists of three subsystems.

[0041] In the motion planning subsystem, the enveloping mold motion parameters are first input via the user interface, and the enveloping tool motion is based on kinematic calculations. Then, the theoretical slider displacement is calculated. Finally, the theoretical motor angle was calculated ( ) and motor speed ( ), which is the output of the motion planning subsystem.

[0042] In the feedback calculation subsystem, the actual motor angle and the actual link strain ( ) is used as input. Motor torque ( Based on the machine condition prediction method discussed above, the compensation value for the slider displacement is calculated as follows: The calculation is based on the motion error compensation method discussed above.

[0043] The six servo control subsystems contain three motor control loops, including position, speed, and current loops. In the position loop, the theoretical motor angles are based on those in the motion planning subsystem. ), slider displacement compensation value in the feedback calculation subsystem ( ) and the actual motor angle from the encoder Using a proportional parameter A proportional (P) controller is used to adjust the position error between the theoretical and actual motor angles. In the velocity loop, the theoretical motor speed is based on the motion planning subsystem. ) and the actual motor speed from the encoder Using a proportional parameter and integral parameters A proportional-integral (PI) controller is used to adjust the speed error between the theoretical and actual motor angles. In the current loop, the theoretical force is calculated based on the feedback calculation subsystem. ) and the actual current from the motor coils ( ), using a proportional parameter Integral parameters and differential parameters A proportional-integral-derivative (PID) controller is used to adjust the current error between the theoretical current and the actual current. ).

[0044] The error for each loop is: (10) in and These are the number of pole pairs and the flux linkage of the motor, respectively.

[0045] Figure 8 middle, This refers to the dynamic torque of six servo motors. Control parameters. It is a diagonal matrix, represented as follows: (11) Based on the above formula, by controlling the instantaneous torque of the six servo motors, the displacement of the six sliders can be controlled in real time, thereby achieving precise control of the enveloping mold motion. The instantaneous torque of each servo motor can be calculated from four parts: the first part is the theoretical torque ( ), based on real-time motor motion ( ) and connecting rod strain ( The dynamic model calculation is performed. The second part is the adjustment of the position loop. Due to motor position error ( ), Slider error compensation value ( and control parameters Calculations. The third part is the adjustment of the speed loop. Due to motor speed error ( ), the integral of motor speed error and corresponding control parameters Calculations. Part Four is the adjustment of the current loop. Due to motor current error ( ), the integral r of the motor current error ), the derivative of motor current error ( and the corresponding control parameters Calculation. Thus, by calculating the real-time torque of the servo motor... The real-time displacement of the six sliders and the real-time movement of the enveloping mold can be controlled according to the proposed control method.

[0046] Based on the above discussion, the core of the linkage strain feedback and motion error compensation control method proposed in this invention is a multi-axis three-loop PID servo control system. The three-loop PID servo control system is a second-order linear system. The stability of the control system can be verified using the Routh-Herwitz criterion. In the Routh-Herwitz criterion, the coefficients of a second-order system should be greater than zero. Specifically, if the PID control parameters... If the value is greater than zero and sufficiently large, the steady-state error will tend to zero.

[0047] According to the method provided by this invention, the solution to the dynamic model of the heavy-duty forming robot can be obtained. According to the method provided by this invention, an error compensation control method for the heavy-duty forming robot based on strain feedback can be implemented. In such cases… Figure 9The experiment was conducted within the hardware framework shown. Linkage strain was first detected by strain gauges, then converted into a current signal by an isolated strain gauge input module (ADAM3016), and finally converted into a digital signal by a multi-function data acquisition card (PCI 1716). Each slider was driven by a servo control system with a motor, controller, and encoder. The actual slider displacement was detected by the encoder and updated to a 6-axis servo motor control card (PCI 1265). The industrial PC (IPC 610) calculated the dynamic model, linkage strain feedback, and motion error compensation control parameters, and sent the calculated data to the servo motor control card (PCI 1265) to drive the servo motor. Figure 10 shows the software control and monitoring interface. The motion planning interface can be used to plan slider displacement, such as... Figure 10 As shown in (a) above, the deformation monitoring interface can be used to monitor the deformation of the connecting rod in real time and predict the motion error of the envelope mold, such as... Figure 10 As shown in (b) of the diagram. During the experiment, the processed blank, the envelope mold, and the constraint mold were as follows: Figure 11 As shown, the blank is a simple cylindrical shape, the enclosing mold is a basic conical shape, and the constraining mold is a ring shape.

[0048] like Figure 12 As shown in (a) of the figure, under the traditional indirect force control method with motion error compensation, the compensation value of the slider displacement ranges from approximately 51 to 112 μm. Figure 13 As shown in (a), under the linkage strain feedback control method with motion error compensation, the compensation value of the slider displacement ranges from approximately 130 to 208 μm. Under the traditional indirect force control method with motion error compensation, the Z-axis angle error of the envelope tool is the largest, ranging from approximately -2000 to 0 μrad. Figure 12 As shown in (b) above, the Z-axis position error of the envelope mold is the largest, ranging from approximately -94 to 7 μm. Figure 12 As shown in (c) above. Under the linkage strain feedback control method with motion error compensation, the Z-axis angular error of the envelope mold is still the largest, but its range is reduced to approximately -600 to 600 μrad, as... Figure 13 As shown in (b) of the diagram, the Z-axis position error of the envelope mold becomes minimal, with its range reduced to approximately -20 to 22 μm, as... Figure 13 As shown in (c) in the figure.

[0049] Figure 14 This demonstrates the motion of feature points on the envelope tool under different control methods. From Figure 14 It can be seen that, compared with traditional control methods, the motion of feature points on the envelope mold under the linkage strain feedback and motion error compensation control method is more consistent with the theoretical control method.

[0050] Using the heavy-duty forming robot and the proposed control method of this invention, multi-degree-of-freedom envelopes of two typical 6061 aluminum alloy thin-walled high-rib components with different geometries were experimentally completed at room temperature. Figure 15 The table shows the geometric parameters of two typical thin-walled high-rib components, the experimental envelope mold, the constraint mold, the blank, and the final formed thin-walled high-rib component. Table 3-1 shows that the two typical thin-walled high-rib components have good appearances. For the first typical thin-walled high-rib component with ribs on one side, the skin thickness can be reduced to 1 mm, and the ribs can grow well to a final height of 10 mm and a thickness of 0.8 mm, i.e., the ratio of rib height to rib thickness can reach 12.5, and the ratio of rib height to skin thickness can reach 10. For the other typical thin-walled high-rib component with ribs on both sides, the skin is thinned to the target thickness, and the ribs grow well. Furthermore, the multi-degree-of-freedom envelope formation time for each of the two typical thin-walled high-rib components is within 10 seconds, which is highly efficient.

[0051] The forming accuracy of the first typical thin-walled, highly ribbed component was measured using a coordinate measuring machine (CMM). Figure 16 The results of forming accuracy measurement of thin-walled, highly ribbed components under different control methods are presented. Figure 15 It can be seen that the geometric deviation of the thin-walled, highly ribbed component formed under the error compensation control method based on strain feedback is significantly reduced from -64-54 μm under the traditional force feedback method to -18-32 μm, meaning that the forming accuracy of the thin-walled, highly ribbed component is significantly improved under the control method proposed in this invention. In summary, under the control method proposed in this invention, the motion error of the envelope tool can be significantly reduced, thereby significantly improving the multi-degree-of-freedom envelope forming accuracy of the thin-walled, highly ribbed component.

[0052] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for error compensation control of a heavy-duty forming robot based on strain feedback, characterized in that, The heavy-duty forming robot includes an upper base, a moving platform, and six branches connecting the upper base and the moving platform. Each branch includes a servo motor, a reducer, a ball screw, a slider, and a connecting rod. The two ends of each connecting rod are connected to the slider and the moving platform via ball joints. The control method includes the following steps: Establish a dynamic model of the heavy-duty forming robot; Obtain the actual strain of the connecting rods in each branch and the actual motor angle of each servo motor; The actual deformation of each link is calculated based on the actual strain, and the actual displacement of each slider is calculated based on the actual motor angle. The actual deformation of each link and the actual displacement of each slider are used as input variables of the dynamic model to solve the motion state, deformation state and force state of the heavy-duty forming robot. The theoretical motor torque of each servo motor is calculated based on the motion state, deformation state, and force state. Based on the motion error between the theoretical and actual motion of the envelope mold, calculate the compensation value for the displacement of each slider. Based on the theoretical motor torque, the compensation value of the slider displacement, the actual motor angle, and the actual motor speed, closed-loop control is performed on each servo motor to compensate for the motion error of the envelope mold in real time.

2. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The establishment of the dynamic model of the heavy-duty forming robot includes: The slider and the connecting rod are respectively equivalent to two mass components, upper and lower, and an equivalent spring and an equivalent damper are provided between the upper and lower mass components; The moving platform is equivalent to a central mass component and multiple side mass components, and an equivalent spring and an equivalent damper are provided between the central mass component and the multiple side mass components; Based on the mass components, elastic deformation, and damping characteristics of the slider, the connecting rod, and the moving platform, a set of rigid-flexible coupling dynamic equations is established.

3. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The actual strain is obtained by strain gauges installed on each connecting rod, with the strain gauges located in the middle region of the connecting rod.

4. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 3, characterized in that, Before placing the strain gauge on the connecting rod, the deformation sensitivity and installation error sensitivity of the slider, connecting rod and moving platform as strain gauge installation positions are analyzed respectively. The connecting rod with high sensitivity to forming load and low sensitivity to strain gauge installation error is selected as the strain gauge installation position.

5. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The calculation of the actual deformation of each link based on the actual strain includes: Obtain the actual strain of each link in the six branches; Based on the correspondence between the actual strain and the rated length or equivalent length of the connecting rod, the deformation of each connecting rod along its force direction is calculated.

6. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The calculation of the actual displacement of each slider based on the actual motor angle includes: Obtain the actual motor angles of the six servo motors; Based on the actual motor angle and the transmission relationship between the reducer and the ball screw, the actual displacement of the six sliders is calculated.

7. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The actual deformation of the six links and the actual displacement of the six sliders are input into the dynamic model as known input variables. The robot state variables, including slider motion variables, link motion variables, moving platform motion variables, slider deformation variables, link deformation variables, moving platform deformation variables, constraint force variables, slider driving force variables, and load variables applied to the moving platform, are solved.

8. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The theoretical motor torque is calculated based on the actual motor angle, the first derivative of the actual motor angle, the second derivative of the actual motor angle, the actual strain of the connecting rod, the first derivative of the actual strain of the connecting rod, and the second derivative of the actual strain of the connecting rod.

9. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, The calculation of compensation values ​​for each slider displacement based on the motion error between the theoretical and actual motion of the envelope mold includes: Determine the theoretical and actual poses of the envelope mold; Calculate the motion error rotation of the envelope mold based on the theoretical and actual poses; With the goal of making the motion error spinor approach zero, the compensation value of each slider displacement is solved; The compensation value of the slider displacement is superimposed on the theoretical displacement command of the corresponding slider.

10. The error compensation control method for heavy-duty forming robot based on strain feedback according to claim 1, characterized in that, Closed-loop control of each servo motor includes position loop control, speed loop control, and current loop control, among which: In the position loop control, the position error is calculated based on the theoretical motor angle, the compensation value of the slider displacement, and the actual motor angle, and the position loop adjustment amount is generated based on the position error. In the speed loop control, the speed error is calculated based on the theoretical motor speed and the actual motor speed, and the speed loop adjustment amount is generated based on the speed error; In the current loop control, the current error is calculated based on the theoretical current corresponding to the theoretical motor torque and the actual motor current, and the current loop adjustment amount is generated based on the current error. The real-time output torque of each servo motor is calculated based on the theoretical motor torque, the position loop adjustment, the speed loop adjustment, and the current loop adjustment.

Images