Fuzzy Neural Network Synchronous Control Of H-type Platform Based On Iterative Super Twisting Sliding Mode

In recent years,with the improvement of processing quality and accuracy,high-speed and high-precision motion system has become an indispensable part of high-end equipment manufacturing industry.The direct drive H-type platform composed of three permanent magnet linear synchronous motors（PMLSM）has been extensively used in precision machining and other high-end equipment manufacturing field due to its advantages of large thrust,high reliability and high positioning accuracy.However,PMLSM is susceptible to external disturbances,friction and periodic thrust fluctuations,leading to tracking errors in the single-axis PMLSM of the direct drive H-type platform.In addition,factors such as mechanical coupling will reduce the synchronization accuracy between the two-axis.Therefore,a fuzzy neural network synchronous control system of H-type platform based on iterative super twisting sliding mode was designed to reduce the single-axis tracking error and improve the precision of double-axis synchronous control.Firstly,the development situation and control strategy of direct drive H-type platform at domestic and overseas were introduced.The basic structure and operating principle of PMLSM were described,and the mathematical model of PMLSM was built.Based on the mathematical model of PMLSM,the mathematical model of direct drive H-type platform including mechanical coupling,external disturbance and other uncertain factors was established.Secondly,an iterative super twisting sliding mode controller was designed for the PMLSM which was susceptible to external disturbances,friction and periodic thrust fluctuations.Super twisting sliding mode control can suppress the influence of external disturbance and improve the robustness of the system.A novel iterative learning law based on super twisting sliding mode control was designed to restrain the influence of periodic disturbances such as friction and thrust fluctuation.The stability of the system was proved by Lyapunov stability theory,and the valid of the designed controller was demonstrated by simulation.Finally,in order to reduce the influence of mechanical coupling between the double-axis on the direct drive H-type platform,it can obtain high quality synchronous control performance.In this thesis,a recurrent Chebyshev fuzzy neural network（RCFNN）compensation controller was designed,with Chebyshev orthogonal polynomial as the activation function of hidden layer of fuzzy neural network.The controller can compensate for synchronization error and single axis tracking error,and regulate the parameters of the neural network in real time by using BP algorithm,so that the system error gradually tends to the minimum.The proposed controller was modeled and simulated by MATLAB/Simulink,and the simulation results were compared with those of the cross-coupling controller,which proves the superiority of the RCFNN compensation controller.