Basic Research for On-line Identification of Physical Parameters Involved in Flexible Structures

本基础研究的主要目的是建立在线算法来确定悬臂梁模型中涉及的物理参数，如杨氏模量、开尔文-沃伊特结构和空气阻尼系数。无论采用何种仪器或传感器，这些未知参数都无法直接测量。只对梁的位移进行测量。因此，问题是如何利用间接测量的动态数据构建计算机识别算法。为了逐步进行研究，首先建立了均匀悬臂梁两个未知物理参数（内阻尼系数和杨氏模量）的识别算法；其次，通过发展上述算法，提出了一类均匀力学梁的两个未知参数的在线计算算法。然后考虑了非均匀悬臂梁的空间分布（即空间变化）模型两个参数的反问题。不用说，在处理地面柔性结构时，空气阻尼的影响是不可忽视的。考虑空气阻尼效应，提出了一种确定Kelvin-Voigt结构、空气阻尼系数和杨氏模量的同时识别算法。针对三种未知参数的辨识问题，构造了三种未知参数的在线辨识问题。此外，我们还开发了三种空间变化参数的（离线）识别算法。本研究提出的识别方法是间接测量法，即所谓的“软传感器”，通过动态测量数据确定不可测量的物理参数。

The principal purpose of this basic research is to establish the on-line algorithms for determining physical parameters involved in the model of cantilevered beam such as the Young's modulus, Kelvin-Voigt structural and air damping coefficients. These unknown parameters can not be measured directly no matter what instruments or sensors are employed. The measurement is performed only for the displacement of the beam. So the problem is to construct any computer algorithm for identification by using indirectly measured dynamic data.In order to research step by step, at first the identification algorithm for two unknown physical parameters(the internal damping coefficient and the Young's modulus) of the uniform cantilevered beam is established. Next, by developing the above-mentioned algorithm an on-line computational algorithm for determining the two unknown parameters is proposed for a class of uniform mechanically beams. Then the inverse problem for the spatially distributed (i.e., spatially-varying)model two parameters are considered for the non-uniform cantilevered beam.Needless to say, when we deal with flexible structures on the ground, the effect of the air damping cannot be neglected. Taking the air damping effect into account, a simultaneous identification algorithm is proposed for determining Kelvin-Voigt structural and the air damping coefficient and Young's modulus. For the identification problem about three kinds of unknown parameters, we also construct an on-line identification problem for three parameters. Furthermore we develop the (off-line) identification algorithm for three spatially-varying parameters.The identification method proposed in this research are the indirectly measurement method, so-called 'soft sensor', by which we decide unmeasurable physical parameters using the dynamic measurement data.