摘要
针对巨磁致伸缩系统的自适应精密驱动和微振动控制系统，结合受控自回归滑动平均模型（CARMA）与递推增广最小二乘法（RELS）相结合对巨磁致伸缩驱动器（GMA）实现在线模型辨识；分别用不同类型的信号作为输入，辨识模型能精确描述GMA输出位移，辨识误差达0.23%；将改进的广义预测控制算法（MGPC）应用于GMA的闭环位移控制，与最小方差自适应控制（MVSTR）相比，MGPC具有更好的实时性和更高的控制精度，在0～10μm给定位移下，其驱动控制误差达0.143μm。最后基于上述CARMA模型和MGPC算法对GMA隔振系统进行微振动控制实验，抑制效果达到20 dB。该研究结果对精密工程及航天振动控制应用具有一定的价值。
The adaptive high precision displacement driving and micro vibration control based on giant magnetostrictive actuators （GMA） are studied. The controlled auto regressive moving average （CARMA） model and recursive extended least squares （RELS） algorithm are used to perform the online identification of the GMA system. The identification model can accurately represent the GMA displacement output with different input current signals, and the error is below 0.23%. Then, the modified generalized prediction control （MGPC） algorithm is applied to the high precision displacement control of the actuator. Experiment results show that the MGPC method has a better real-time performance and higher control precision than the minimum variance self-tuning regulator （MVSTR） does. The error of the driving control displacement of the MGPC method is only 0.143 um for the given 0-10μm reference displacement. Finally, the experiment of micro vibration control for a GMA isolation system is done based on the CARMA model and the MGPC algorithm. The results of the experiment show that the vibration attenuation effect can reach 20 dB. This work may have great application significance in precision equipment engineering and vibration control engineering.
出处
《噪声与振动控制》
CSCD
2014年第5期16-21,共6页
Noise and Vibration Control
基金
国家自然科学基金项目(11172169)
上海航天基金项目(USCAST2012-02
USCAST2013-10)
关键词
振动与波
CARMA模型
巨磁致伸缩驱动器
在线辨识
自适应控制
振动控制
vibration and wave
CARMA model
giant magnetostrictive actuator
online identification
adaptive control
vibration control
