基于优化算法的六自由度气浮台垂向控制  

Vertical Control of 6-DOF Pneumatic Platform Based on Optimized Neural Network Algorithm

作　　者：陈保秀 王岩[1] CHEN Bao-xiu;WANG Yan(School of Astronautics,Harbin Institute of Technology,Harbin 150001,Heilongjiang Province,China)

机构地区：[1]哈尔滨工业大学航天学院,黑龙哈尔滨150001

出　　处：《计算机仿真》2018年第9期315-319,共5页Computer Simulation

基　　金：国家自然科学基金(61174037);国家自然科学基金创新群体项目(61021002)

摘　　要：针对六自由度气浮台垂向控制系统,考虑垂向气浮轴承气腔体积变化、气膜漏气不均衡和气源输入时变性等因素,可以看作内环含有滞后的非线性串级控制系统。由于被控对象参数未知且时变,仅用双闭环PID控制不能有效补偿参数、滞后对系统的不利影响。故在对垂向控制系统进行垂向位移和气压控制的基础上,采用基于遗传算法优化的BP神经网络算法对参数、滞后时间进行在线辨识,利用辨识模型结合两级Smith预估控制进行实时闭环控制。仿真结果与理论一致,证明了辨识算法的正确性、有效性。上述算法与两级Smith预估控制共同应用于双闭环控制时,极大改善系统的抗干扰性、实时性,并提高了垂向控制的稳定性及鲁棒性。Aiming at the vertical control system of six-degree-of-freedom air-floating platform, considering the variation of air volume, the leakage of gas film and the time-varying of gas source input of vertical air-filled beating, it can be regarded as the nonlinear string level control system. As the controlled object parameters are unknown and time-varying, only double closed-loop PID control cannot effectively compensate the adverse effects of parameters and time-delays in the system. Based on the vertical displacement and pressure control of the vertical control system, BP neural network algorithm based on genetic algorithm optimization is used to identify the parameters and delay time. A real-time closed-loop control is carried out by using the identification model combined with two-stage Smith predictive control. The simulation results prove the correctness and validity of the identification algorithm. The algorithm and the two-stage Smith predictive control are applied to double closed-loop control, which greatly improves the anti -jamming and real-time of the system and improves the stability and robustness of the vertical control.

关 键 词：滞后 串级时变 遗传算法 神经网络 预估 

分 类 号：V411.8[航空宇航科学与技术—航空宇航推进理论与工程]