Study On Structural Optimization And Control System Of Legged Parallel Robot For Manufacturing

In the field of automobile and aviation,the large thin-walled partsaremuch adopted to improve the comprehensive performance.At present,during assembly or manufacturing process,fixtures layout remain unchanged once planned.However,these thin-walled parts often have multiple processing characteristics,so the cutting load distribution is complex,and the unchanged layout is difficult to adapt to complex conditions.Therefore,there is a necessary to develop self-reconfigurable flexible fixture.The flexible fixture based on legged parallel self-reconfigurablerobot with high flexibility and mobility advantages,and the supporting head can swing with high magnitude,by these,the deformation and vibration is effectively suppressed,then the machining precision can be greatly improved.In this thesis,based on dynamic self-reconfigurable fixture system,the basic unit which is a 8-SPU legged parallel robot is analyzed,from workspace,motion performance,motion control system and the fault diagnosis system.The main contents are as follows:(1)The CAD model and mathematical model of this mechanism are established,and the freedom is analyzed also.Inverse kinematics model and kinematic constraints are analyzed,by which,the workspace is solved.And the effects of the mechanism parameters are analyzed also.(2)By the robot Jacobian matrix,kinematics mapping model is established.Against to rotation and motion freedom,a global condition number index is proposed,combined with workspace volume index,a comprehensiveobjective function is present.Integrated with particle swarm optimization,three mechanism parameters are optimized.What's more,the effect of optimization variables is analyzed also,andthe compared results indicate the feasibility of this procedure.(3)The trajectory planning of the linear and circular motion of the robot is carried out.Based on the kinematics model,the motion control strategy is designed.By using the control mode of "industrial computer + motion control card",the control interface of is designed.(4)Combining the YU norm,biological lateral inhibition theory and Fuzzy ART network,the soft competition learning ART algorithm which allows multiple nodes to be learned simultaneously is proposed.In order to improve the diagnostic efficiency,the feature selection algorithm based on YU norm similarity criterion is present.Based on the combination of these two methods,a real-time monitoring and fault diagnosis system for the robot drive motor is designed.This research has important academic significance to improve the flexibility and intelligence of fixture,has important guiding significance to promote the design theory applied to engineering practice.What's more,it also has great engineering significance to improve the machining accuracy,reduce the cost of design,and incentivize manufacturingautomation.