基于子带自适应滤波的振动台功率谱复现

PSD replication of a shaking table based on subband adaptive filtering algorithm

为了提高液压振动台功率谱密度复现的精度和减少功率谱均衡时间,研究了子带自适应滤波系统辨识技术。在分析振动台功率谱均衡原理的基础上,指出液压振动控制系统辨识的精度直接影响到功率谱复现的精度及均衡时间。针对系统辨识技术研究算法的结构,推导出算法迭代公式和收敛条件。同时,为了将信号更好地分解到各个子频带,探讨了一种近似完全重构滤波器组设计方法,近于消除滤波器组的幅值失真和混叠失真。基于液压振动台功率谱复现流程,采用Matlab/Simulink仿真软件进行仿真试验,仿真结果证明了子带自适应滤波算法在液压振动台功率谱复现过程中的有效性。

To improve the accuracy of power spectral density （ PSD ） replication of a hydraulic shaking table and to reduce the PSD equalization time, a new system identification （ SID） technology based on the subband adaptive filtering algorithm was studied here. The accuracy of SID of the shaking system was one of key factors influencing the PSD replication precision and equalization time through analyzing the principle of PSD equalization of the hydraulic shaking table. The structure of the subband adaptive filtering was investigated, the iteration equation was derived and the convergence conditions of this SID technology were discussed. Meanwhile, a new method for the design of almost full reconstruction cosine modulated filter banks was introduced to decompose a signal into different subbands smoothly and decrease distortions of amplitude and aliasing of the filter banks. Finaly, the efficiency of the subband adaptive filtering SID technology was verified using the simulation of PSD replication of a shaking table with Matlab/Simulink.