Research On Construction Of Grinding System For Complex Surface Robot Grinding And Grinding Force Control

At present,the machining of complex curved surface parts mostly adopts the forming technology of no allowance precision casting or precision forging at home and abroad,which makes the parts basically formed at one time,and then the manual polishing and finishing processing are carried out,so as to obtain the parts that meet the requirements.However,the manual polishing process depends heavily on the processing experience of workers,and the workpiece yield and processing efficiency are relatively low.However,due to the complexity of the structure of complex curved surface parts,there are some problems in the automatic polishing of industrial robots,such as the blind area of parts processing and the lack of processing.It is difficult to realize the one-time and integrated automatic polishing,and the polishing force in the process of polishing is difficult to control in real time,which makes the effect of automatic polishing of industrial robots difficult to achieve the expected.In order to solve the above problems,the design of the industrial robot variable polishing head polishing system and the research of the polishing force advance control algorithm are carried out.(1)Based on the industrial robot,a variable head polishing system for complex curved parts is designed and built,including the design of force sensor module,power supply module,structure design of variable head polishing platform and the design of polishing platform control scheme.Based on ANSYSWorkbench software,the static and dynamic characteristics of the polishing platform with variable polishing head are analyzed by using the finite element method,including the static analysis and modal analysis.The structural strength of the polishing platform is verified.The communication scheme among the working threads of PC,industrial robot controller and PLC host in the variable head polishing system of complex curved surface parts was established.(2)The mathematical and physical basis of the industrial robot's polishing force control process is studied in detail: the industrial robot coordinate conversion and the calculation method of the polishing position and attitude.The control principle of the polishing force during the polishing process is analyzed in detail,and the normal calculation of the polishing force was carried out on this basis.Based on the impedance control algorithm,the polishing force control strategy of the industrial robot under the condition that the workpiece and the polishing tool are not in contact or in contact is established,and the normal polishing position and attitude correction amount at the processing point of the workpiece and the polishing tool is calculated.(3)The principle,advantages and disadvantages of traditional recurrent neural network(RNN)and long-short-term memory neural network(LSTM)are analyzed.Based on the LSTM neural network,a polishing force prediction model is established,and the influencing factors and data processing methods of polishing force are studied.The training methods of the polishing force prediction model are studied in detail,including: the research on the factors affecting the model training performance,the model training process based on the back propagation method and the model training results evaluation method.(4)The position error test of the variable polishing head polishing platform is carried out,including the positioning accuracy and rotation error detection of the polishing platform,which verified the working performance of the variable polishing head polishing platform.By using the designed polishing system of variable polishing head,the typical complex curved surface parts aeroengine blades are selected to carry out the polishing experiment of industrial robot,which verifies the effectiveness of the system to solve the problem of machining blind area and insufficient machining when the complex curved surface parts are polished by industrial robot.The simulation experiment of the polishing force prediction is carried out to verify the effectiveness of the algorithm to solve the real-time control problem of the polishing force of industrial robot.