Research of On-Line Parameter Estimation and Vibration Control Based on the ESPRIT

1.在ESPRIT.2的基础上，提出了一种新的子空间方法，可以识别良好激励频率和相应频率的工作偏转形状。然后利用URV分解将该方案转化为时变系统的自适应方案。为了证明该方案对时变信号的适应性，我们演示了它在人工生成响应上的性能。结果表明，该方案可自适应地将结构变形分解为多个谱的底层叠加。然后利用实际结构的响应，用实际数据验证了该方案的有效性。这些结果与分析结果很好地吻合，使我们相信所提出的方案可以成功地用于时变信号的识别。我们提出使用投影的一般描述作为模态分解的工具。该方法用数值模拟频响进行了演示。结果表明，与普通方法相比，该方法能得到更清晰的结果。此外，由于模态向量的准确性在模态分解中非常重要，我们提出了一种新的基于投影一般描述的模型关联技术来验证有限元分析得到的模态向量。这是一个直接的正交性检验，它可以在没有任何系统矩阵的情况下完成。对于所提出的案例研究，我们证明了这种方法的充分性，通常的正交性检验，与普通的。结果表明，该方法对相关分析和正交性检验是有效的。

1.We present a new type of subspace method which can idenitfy well-excited frequencies and the Operating Deflection Shapes corresponding to those frequencies based on the ESPRIT.2.Then we convert the scheme into an adaptive one for time-varying systems, making use of URV decomposition.3.In order to prove the adaptavility of this scheme for time-varying signals, we demonstrate its performance on artifically generated responses. As a result, proposed scheme adaptively decomposed the deformation of a structure into the underlying superposition of several spectrum.4.Then we use the responses of an actual structure and show the effectiveness of this scheme with real data. These results are in good agreement with analytical ones and lead us to believe that the proposed scheme can be successfully used for the identification of time-varying signals.We proposed that the General Description of Projection is used as a tool of the modal decomposition.This approach is demonstrated using numerically simuletad FRFs. As a result, it was shown that this approach gives good clearer results compared to the ordinary method.Moreover, we present a new model correlation technique based on the General Description of Projection to verify the modal vectors obtained by the Finite Element Analysis, because accuracy of the modal vectors is important in the modal decomposition. This is a straight-forwaed orthogonality check to the effect that it can be done wihout any system matrices.For the case study presented, we showed the adequacy of this method for usual orthogonality check, compared with the ordinary one. Results show that this method is effective for correlation analysis and orthogonality check.