Research On Robot Milling Machining Error Compensation Based On Force Feedback Control

Robot machining has the characteristics of high flexibility,strong adaptability and low cost,which plays an important role in the processing of large components.However,due to the low stiffness of the robot itself,the machining accuracy of the robot is generally low.At present,the research at home and abroad mainly includes the position and posture optimization and constant force control for grinding,polishing and deburring robots to improve the machining accuracy;in the aspect of robot milling,it mainly includes the establishment of kinematics and dynamics model to optimize the position and posture to improve the rigidity,the establishment of milling force model to study the vibration suppression,the optimization of the position and posture to improve the rigidity and reduce the cutting force,etc Method to improve the machining accuracy.The above research methods improve the machining accuracy to a certain extent,but the effect is not significant.In order to further improve the accuracy of robot milling,this topic will study the error compensation of robot milling based on force feedback,and the error generated by the data calculation combined with force sensor feedback will be compensated to the trajectory by algorithm,so as to minimize the error.The robot milling in this paper is using FANUC M-10 ia / 12 robot,the homogeneous matrix is chosen to describe the position and pose of the robot,the joint parameter table is obtained by D-H modeling,and the forward and inverse kinematics solutions of the robot are obtained by calculation.The kinematics model can be applied to the robot with similar structure and the same D-H model,which has certain universality.According to the internal structure of the robot joint,the stiffness model of the robot is established and the torsional stiffness of each joint is obtained.By calculating the Jacobian matrix of the robot,the end equivalent stiffness matrix of the robot is obtained.The relationship between the equivalent stiffness of the robot end and each joint angle is analyzed by simulation,and the cloud chart of the equivalent stiffness of the robot end under the adjacent joint linkage is obtained.Through the equivalent stiffness matrix,the offset error of the robot's end position under the force or moment can be obtained.Then,the deflection stiffness model of the robot arm is established,and the necessity of the research on the deflection stiffness of the arm is analyzed.The calculation and analysis show that the deflection stiffness of the arm can be ignored compared with the transmission stiffness model.The milling force model of robot milling is established,and the milling force of ball head milling is calculated and simulated.The rake angle and roll angle parameters corresponding to the minimum milling force in robot milling are obtained,which are used for subsequent tool path planning.The off-line programming code of the robot is obtained by using Power MILL to plan the tool path of the robot milling for the specific workpiece.Based on the feedback of force sensor and the stiffness control system,the trajectory error is obtained by using the stiffness model,and then the RBF neural network compensation algorithm is used to compensate the error of the robot machining trajectory.At the same time,the final error of the two algorithms is analyzed by comparing the BP neural network compensation algorithm.The dynamic characteristics of the robot are analyzed,the FANUC M-10 ia / 12 solid model is established,and the modes of the robot under different postures are analyzed by using the finite element analysis software.At the same time,the harmonic response of one of the postures is analyzed,and the frequency range sensitive to the natural frequency and harmonic response amplitude of the robot is obtained,so as to determine the speed range that the robot milling spindle avoids.Then,ADAMS software is used to simulate the milling process before and after robot compensation,and the milling error before and after robot compensation is obtained,and the compensation effect and practicability of RBF neural network compensation algorithm are analyzed.