摘要
针对斜拉索-磁流变阻尼器系统,提出一种数据驱动的分布参数系统建模方法.利用K-L时空分解方法得到表征系统空间结构的空间基函数,将其与系统输出作内积求出表征系统状态幅值的时间序列,根据时间序列和外载荷等数据信息,利用递推最小二乘算法辨识时域ARMAX模型,将空间基函数与时域ARMAX模型进行空间重构得到空间ARMAX模型,最后对该时域ARMAX模型进行预测控制.实验结果表明:建立的分布参数系统模型计算量小、精度高,可以满足实际工程要求;采用的预测控制方法能够对斜拉索-磁流变阻尼器分布参数系统模型进行有效的减振控制.
A data-driven distributed parameter system modeling method is proposed to analyze the stay cable-magnetorheological damper system.The space basis function that characterizes the space structure of the system is obtained by using the KL spatio-temporal decomposition method,and the inner product of such function and the system output is calculated to obtain the time series representing the system state amplitude.Based on the time series and other data information like the external load,the time-domain ARMAX model is identified by recursive least squares algorithm.The spatial ARMAX model is established by the spatial reconstruction of the spatial basis function and the time-domain ARMAX model.Finally,the time-domain ARMAX model is studied for predictive control.The analysis of experimental results shows that the established distributed parameter system model exhibits less calculation and higher accuracy,which can meet the engineering requirement.The proposed predictive control method can effectively control the cable-MR damper distributed parameter system model for vibration reduction.
作者
王振
张刚
陈跃华
冯志敏
胡洋洋
WANG Zhen;ZHANG Gang;CHEN Yuehua;FENG Zhimin;HU Yangyang(Faculty of Maritime and Transportation,Ningbo University,Ningbo 315832,China)
出处
《宁波大学学报（理工版）》
CAS
2020年第3期91-98,共8页
Journal of Ningbo University：Natural Science and Engineering Edition
基金
国家自然科学基金(51675286)
国家自然科学基金青年科学基金(51505237)
宁波大学王宽诚幸福基金.
关键词
K-L时空分解
递推最小二乘辨识
分布参数系统
ARMAX模型
广义预测控制
K-L decomposition
recursive least squares identification
distributed parameter system
ARMAX model
generalized predictive control
