多维ARX模型在高层建筑模态识别中的应用

Action of multiple ARX model for modal identification of high - rise buildings

针对高层建筑模态识别具有激励难以实现、噪声影响大等特点,从状态空间理论出发,建立了多维ARX模型,运用极大似然法估计ARX模型参数,推得结构的模态参数。采用线性滤波器法对白噪声进行滤波,模拟脉动风的作用,分析了不同噪信比和不同模型阶次对固有频率、阻尼比和振型的影响。结果表明,多维ARX模型适合于高层结构在线识别,特别适用干结构振型的识别,具有很强的抗噪声能力。

A multiple auto regressive model (ARX) model has been set up based on the state space theory. By using the maximum likelihood method, the parameters of ARX model are estimated and the modal parameters of the structure are obtained. The action of pulse wind was imitated with linear filter to filter white noise. The effect of different signal to noise ratio and different model orders on free frequency, damping ratio and vibration curve are analyzed. Imitation analysis results show that the multiple ARX model with the characteristics of anti-noise is suitable for the high-rise building identification on line, especially for the structure vibration curves.