| Large optical systems have high requirements for the accuracy of mirror surface,and the processing of optical mirror belongs to the scope of ultra precision manufacturing.To meet the processing needs of modern large optical mirror,a large optical mirror processing system composed of a five-degree-of-freedom hybrid processing robot and a flexible hydraulic support system is designed.The system can complete the rough grinding,milling,fine grinding,and polishing process of the optical mirror from the mirror blank to the finished mirror.Since the large optical mirror processing system is composed of dual heterogeneous robots,the dynamic characteristics of sub robots are different,and there are different degrees of coupling between the robots itself and the cooperative robots,resulting in a low overall accuracy of the processing system.In addition,the processing equipment is easily disturbed by the behavior of human,the dynamic characteristics of the mechanism and the changes of environmental parameters during the processing process of the optical mirror.This further aggravates the end output error of the large optical mirror processing system,thereby affecting the molding quality of the optical mirror.Therefore,in order to improve the processing accuracy of the large optical mirror processing system,the multiple disturbance analysis and control compensation strategy of the designed large optical mirror processing system are studied,including structural parameters optimization,dynamic characteristic analysis,human-machine-environment disturbance analysis,error prediction and control compensation strategy.The main research contents are as follows:1)To improve the cooperation efficiency between the five-degree-of-freedom hybrid processing robot and the flexible hydraulic support system in the large optical mirror processing system,the structural parameters of the processing system are optimized.Firstly,the structure composition and cooperative relationship of sub robot in the processing system are introduced.Then,taking the optimal effective workspace as the optimization objective,the scale parameters of the parallel module of the processing robot are optimized,and the optimal motion range of the driving branch chain,the moving platform and the fixed platform radius that meet the design requirements of the processing system are obtained.Finally,the parameters of the core components of the flexible hydraulic support system are optimized,combined with the finite element method,the deformation resistance of beryllium bronze reed and rolling diaphragm are analyzed through the control variables,and the optimal structural parameters suitable for the rolling diaphragm and beryllium bronze reed of the flexible hydraulic support system are determined.2)To reveal the modal information of sub robots in the large optical mirror processing system due to their own structural properties and mutual disturbance between robots.Firstly,the vibration differential equations of five-degree-of-freedom hybrid processing robot and flexible hydraulic support system are established,and the vibration differential equations are transformed into modal coordinates.Then,the dynamic analysis of the revolution and rotation of the grinding tool at the end of the processing robot are carried out,the magnitude and direction of the centrifugal force and reaction force of the grinding tool are calculated,which are brought into the simulation model as external disturbances to obtain the natural frequency of the fivedegree-of-freedom hybrid processing robot and the flexible hydraulic support system.Finally,the natural frequencies of the five-degree-of-freedom hybrid processing robot under different postures,the natural frequencies of the flexible hydraulic support system under different support heights and the mirror amplitude during the operation of the large optical mirror processing system are measured through experiments,and the vibration characteristics of the large optical mirror processing system are analyzed.3)To study the specific influence of disturbance factors in the human-machineenvironment on the accuracy of the large optical mirror processing system.Firstly,the disturbance factors existing in human-machine-environment are classified,and the disturbance term of multiple disturbance factors are established.Then,the disturbance model of human-machine-environment corresponding to disturbance factors and the multi-sensor information fusion strategy based on improved BP neural network based on genetic algorithm are established to analyze the mapping relationship between the influencing factors and the affected factors.Finally,through simulation and experimental analysis,the influence of friction and impact of motion pair on joint disturbance force of the five-degree-of-freedom hybrid processing robot,the influence of working medium parameters on the support pressure and support height of the flexible hydraulic support system,and the disturbance weight between the influencing factors of the large optical mirror processing system and the comprehensive output error are studied.4)To predict the error value generated by the large optical mirror processing system in the future,so as to take compensation measures in advance according to the error prediction results.Firstly,an error prediction model for a large optical mirror processing system was established based on the LSTM deep learning algorithm optimized hyperparameters.Then,time stitching,missing data filling,abnormal data processing and data normalization preprocessing are carried out on the data collected during the movement of the large optical mirror processing system,so as to improve the learning efficiency and prediction accuracy of the deep learning network.Finally,through the LSTM error prediction model through the LSTM error prediction model of Bayesian optimization hyperparameters,the motion error of processing robot under different processing trajectories and the output error of flexible hydraulic support system are experimentally analyzed.5)To compensate the comprehensive error that is caused by multiple disturbances on the robot,the force/position output accuracy of the large optical mirror processing system is further improved.Firstly,the adaptive parameters fitting analysis are carried out for the large system composed of multiple nonlinear subsystems,and the stability of the established adaptive decentralized fuzzy compensation control model is analyzed based on Lyapunov function.Then,taking the dynamic characteristic analysis results and error prediction results obtained from the above research and analysis as the feedforward control parameters,and combined with the disturbance correlation term between each robot,the adaptive decentralized fuzzy compensation control model of large optical mirror processing system is designed.Finally,based on the established adaptive decentralized fuzzy compensation control strategy,the experimental analysis of the five-degree-of-freedom hybrid processing robot and the flexible hydraulic support system are carried out,and the processing quality of large optical mirror processing system is evaluated by Preston equation.The research on multiple disturbance error analysis and compensation control strategy of large optical mirror processing system carried out in this subject is of great significance to improve the processing accuracy and efficiency of large optical mirror.At the same time,it also provides a theoretical basis and experimental reference for cooperative work of multiple heterogeneous robots for precision processing.The dissertation has 81 figures,15 tables and 188 references. |
PDF Full Text Request