激励未知系统模态参数识别研究

System Identification with Unknown Input

研究了输入未知时，根据系统运行实测响应数据识别系统结构，采用迭代逆滤波算法，用高阶ＡＲ模型对输入信号进行估计，以便最终识别系统；在模型识别中采用了正交子空间法，这一方法将正交化引入时间序列，从而实现系统阶次的自动估计及模型识别．针对基本算法的冗余计算提出了改进算法。实验结果表明，本方法对未知输入系统识别良好，具有现实可行性。

System identification using tested vibration data is discussed.The vibration structure is to be identified without input excitation but with the response data under operating condi-tions。An iterative inverse filtering algorithm(ITIF)is used for estimating and updating theunknown input with the orthogonal sub-space method applied later for system identification。When using this method Gram-Schmidt orthogonalization is introduced into time series soas to realize automatic order estimate,An improved algorithm has been developed for redundant computation in the basic algorithm。Experimental resuIts show that the method is goodfor a svstem with unknown input。