| With the continuous and rapid development of national economy,the quality and precision of finished steel products are also put forward higher requirements.As is known to all,the special steel bar in the rolling finishing area has some problems,such as frequent storage operation,bad working environment,heavy repetitive labor,high security risks,etc.As an important part of the finishing process,many grinding procedures still stay in the traditional grinding stage of manual experience,there are low grinding precision,poor adaptability and other problems,it is urgently needed to free people from the heavy,repetitive,monotonous and harsh environment,and realize the robot operation of grinding processing procedures.Based on this,in order to effectively and quickly realize the grinding of bars after rolling,to meet the needs of high safety,strong reliability,fatigue life resistance and high economy in the process of social development,this thesis focuses on further research on the grinding process,to realize the transformation of the process from traditional manual operation to robotic operation.The specific research contents are as follows:(1)In order to realize the robotization of grinding process in the finishing process of special steel bar,a robot operating system for finishing grinding of special steel bar was established.The overall layout of the operating system was planned,and the working process of the operating system was introduced.Aiming at the selected grinding robot,the kinematics model of the grinding robot was established by D-H method,and the forward and inverse kinematics equations of the grinding robot were solved.On this basis,considering the uncertainty of diameter change caused by grinding wheel wear during bar grinding,a deep neural network method based on multi-layer forward propagation was proposed,the prediction model of robot kinematics was established by using deep neural network.The curve profile error caused by grinding wheel wear was compensated and corrected by this model,so as to realize the online grinding of the robot quickly and efficiently.(2)In order to realize the continuous and smooth motion trajectory of the robot during the bar grinding,effectively ensure the stability and smoothness of grinding,the robot motion trajectory was planned and optimized during bar grinding process.Based on the statistical classification of a large number of defective bars in the rolling field,various locations and appearance distribution of defects on the surface of the bars were given.In order to support the follow-up research,the existing laser ultrasonic testing platform in the laboratory was used to detect the defects of the sample bar,and the parameter information such as the location and depth of the defect on the surface of the sample bar was determined.On this basis,according to the distribution of defects on the surface of the bar,the trajectory planning of the grinding robot was classified,and the shortest time spent on the completion of the grinding of the bar was selected as the objective function,an improved particle swarm optimization trajectory planning algorithm for grinding robot was proposed.(3)In order to directly reflect the working state of the bar during grinding,the contact force between the grinding wheel and the bar was analyzed,and the dynamics of the contact was discussed.Based on Hertz rolling contact friction theory,a contact model was established to obtain the stress distribution and boundary curve in the contact area.ABAQUS software was used to simulate and verify the bar grinding process,and the influence of different parameters on the grinding contact force was analyzed.According to the results of simulation and experimental verification,the formula of contact force was obtained.On this basis,the dynamics equation of the grinding robot was established based on Lagrange method.Aiming at the dynamic problem caused by the contact between grinding wheel and bar,the nonlinear fractional calculus theory was introduced to analyze and discuss the vibration phenomenon caused by the contact between grinding wheel and bar.(4)In order to ensure the completion of the bar grinding task and realize the robust stability control of the robot,the dynamic performance of the robot was analyzed.Aiming at the problem of superposition of contact force at the end of robot,gravity compensation technology was used to solve the problem of contact force perception during bar grinding.According to the impedance characteristics of the robot,the impedance control system model was established,and the system parameter performance and steady-state error were analyzed.Based on the complex nonlinear characteristics of robot grinding,a fractional force/position hybrid control strategy for robot grinding was proposed by introducing the fractional-order control theory.Based on the robot dynamics model,the 3D robot model was simulated and analyzed by ADAMS software,and the joint torque curve of the robot was obtained when it was loaded.On this basis,a fractional PI~λD~μ adaptive impedance control strategy of the robot was proposed,the dynamic control performance of the grinding robot was analyzed. |
PDF Full Text Request