Dynamic load identification is an inverse problem concerned with finding the load applied on a structure when the dynamic characteristics and the response of the structure are known. In engineering applications, some of the structure parameters such as the mass or the stiffness may be unknown and/or may change in time. In this paper, an online dynamic load identification algorithm based on an extended Kalman filter is proposed. The algorithm not only identifies the load by measuring the structural response but also identifies the unknown structure parameters and tracks their changes. We discuss the proposed algorithm for the cases when the unknown parameters are the stiffness or the mass coefficients. Furthermore, for a system with many degrees of freedom and to achieve online computations, we implement the model reduction theory. Thus, we reduce the number of degrees of freedom in the resulting symmetric system before applying the proposed extended Kalman filter algorithm. The algorithm is used to recover the dynamic loads in three numerical examples. It is also used to identify the dynamic load in a lab experiment for a structure with varying parameters. The simulations and the experimental results show that the proposed algorithm is effective and can simultaneously identify the parameters and any changes in them as well as the applied dynamic load.

中文翻译：

---

基于扩展卡尔曼滤波器的变参数结构在线动态载荷识别

动态载荷识别是一个逆问题，当结构的动态特性和响应已知时，它涉及找到施加在结构上的载荷。在工程应用中，一些结构参数（例如质量或刚度）可能未知和/或可能随时间变化。本文提出了一种基于扩展卡尔曼滤波器的在线动态负荷识别算法。该算法不仅通过测量结构响应来识别载荷，而且还识别未知的结构参数并跟踪它们的变化。我们讨论了在未知参数是刚度或质量系数的情况下所提出的算法。此外，对于具有多个自由度并实现在线计算的系统，我们实现了模型约简理论。因此，在应用所提出的扩展卡尔曼滤波器算法之前，我们减少了所得对称系统中的自由度数。该算法用于恢复三个数值例子中的动态载荷。它还用于在实验室实验中识别具有不同参数的结构的动态载荷。仿真和实验结果表明，该算法是有效的，可以同时识别参数及其任何变化以及施加的动态载荷。