| Curved parts are widely used in the fields of aerospace,automobile,mold manufacturing and medical instruments.In order to meet the quality requirements,the surface of parts often needs to be polished.At present,polishing is mainly done by hand,but the polishing environment is harsh and the efficiency is low.More importantly,the quality and stability of the surface of manual polishing is poor.However,the polishing process based on NC machine tools is limited by the freedom of freedom,which affects the accessibility of the tool and leads to the difficulty of tool path planning.Industrial robots are widely used in free-form surface polishing because of their flexible movement.The related technical problems are becoming the research focus in this field.The trajectory planning of robot tools is one of the key problems in the polishing of small curvature surfaces.This paper studies the surface polishing model based on Hertz contact theory and Preston equation,and optimizes the traditional trajectory planning method.The following work has been completed:(1)The free-form surface polishing model is established based on Hertz contact theory and Preston equation.The effects of material removal mechanism,polishing pressure,distance between polishing tracks and spindle speed of polishing head on surface polishing quality were analyzed.Based on Hertz contact theory,the geometric elements of the contact area between the spherical polishing tool and the surface were calculated,and the expression of pressure distribution on the contact surface was derived.By analyzing the tool linear velocity,the polishing linear velocity distribution model is obtained.Finally,the microscopic material removal model of surface polishing was derived by Preston equation,which laid a theoretical foundation for polishing trajectory planning.(2)As the trajectory generated by the traditional isoparameter method cannot adapt to the curvature change of the surface,which is easy to cause overthrow or underthrow problems,the equivalent parameter method is optimized.In the direction of row spacing,according to the relationship model between residual height and row spacing,the corresponding row spacing values meeting the requirements of residual height can be re-solved for the areas where the curvature of the surface exceeds the preset value.A new tool contact is inserted at the sensitive point of curvature change in the square of step,which improves the problem of overthrow and underthrow.Finally,combined with the material removal model,the polishing trajectory adapted to the curvature changes of the surface was obtained.(3)The prediction of polishing surface roughness is discussed.Aiming at the problem that the roughness prediction error of BP neural network is large and it is difficult to get the global optimal solution,a roughness prediction method based on BP neural network optimized by genetic algorithm is presented.The initial weight and threshold of BP neural network were optimized by genetic algorithm,and different polishing parameters were selected to predict the surface roughness of robot polishing,and the prediction accuracy was verified by simulation.(4)Build simulation experiment platform and complete algorithm verification.Firstly,the inverse kinematics of the six-axis robot was solved,and the relationship model between the end tool displacement and each joint Angle was obtained.Based on the ER＿factory software environment,experimental parameters were configured.For the same typical free-form surface with small curvature,the trajectory data from the traditional parameter method and the trajectory data obtained by this method were respectively imported into the simulation software.The angular displacement data of each joint of the robot was observed,and the actual polishing trajectory was solved by robot kinematics.Experimental data analysis shows that the proposed method has good adaptability to surface curvature,improves the uniformity of residual height,and avoids overcutting and undercutting problems. |
PDF Full Text Request