部分输入未知条件下结构参数及激励识别

Structural parameters and dynamic loading identification with partially unknown excitations

针对部分输入未知条件下的结构参数和荷载识别问题,提出一种改进的基于最小二乘准则的自适应加权迭代算法。该方法通过引入自适应学习因子和加权正定矩阵,以任意假定的未知外激励作为初始迭代条件,以相邻两次迭代后荷载的识别值的误差作为收敛判断准则,有效地改进了迭代收敛速率、稳定性和识别精度。同时,针对比例阻尼,对现有非线性参数识别的松弛法进行改进,提出一种转换算法。通过一个具有15个自由度的高层数值模型的模拟数据和一个4层结构模型的动力试验实测数据分别验证了该方法有效性,同时,分别探讨了噪声水平、权重系数、学习因子等对算法收敛性的影响。数值算例和基于模型动力测试数据的识别结果表明,该算法具有稳定的收敛特性,参数和荷载识别精度高以及对测量噪声的鲁棒性强的特点。

The identification of structural parameters and dynamic loadings when partial inputs are unknown is important for damage detection and remaining service life forecasting of the structures. In this paper, an updated iterative identification approach, referred to as weighted adaptive iterative least-squares estimation with partially unknown inputs （WAILSE-PUI）, was proposed to identify the structural parameters as well as the unknown external loadings simultaneously. In the proposed iteration approach, a learning coefficient using the information of the previous two iterations and a weighted positive definite matrix were employed for the purpose of improving the convergence of the identification approach. The initial values of the unknown external excitations were arbitrarily assumed and the iteration process would continue until the errors of the identifed external dynamic loadings of two adjacent iterations were within a pre-defined tolerance. Moreover, an approach to handling the Rayleigh damping which leads to nonlinear problems in the identification was proposed. The efficiency, accuracy, and robustness of the proposed method were validated via numerical simulation of a 15-degree-of-freedom high-rise building model and dynamic test of a four-story frame structure. The effect of noise level, weight, and learning coeficients on the convergence was discussed. Results show that the proposed approach can simulstanously identifiy structural parameters and unkown excitations with high accuracy, and is robust even with noise-polluted dynamic response measurements.