基于VMD的建筑结构模态参数识别

在建筑结构的健康监测、控制和状态评估中经常遇到的一个关键性问题是如何根据实测响应信号准确估计结构阻尼比及自振频率等模态参数。基于变分模态分解(VMD)提出一种新的结构模态参数识别方法。该方法首先对实测振动信号进行VMD分解,获得单模态信号,然后采用自然环境激励技术(NEXT)得到单模态信号的自由衰减响应,最后利用直接插值法(DI)和曲线拟合获得结构的自振频率和阻尼比。通过三层框架结构的数值模拟验证了该方法的准确性及可靠性。利用该技术对台风“达维”作用下广州中信广场的实测加速度数据进行分析,并将估计的结构模态参数和其他识别方法的分析结果进行对比,进一步验证了该方法的准确性和有效性。

Modal parametric identification of building structures based on VMD

A key issue in health monitoring, control and condition assessment of building structures is correctly estimating their modal parameters including structural damping ratios, natural frequencies, etc. based on measured response signals. Here, a new structural modal parametric identification approach was proposed based on variational mode decomposition (VMD).With the proposed method, a measured vibration signal was decomposed using VMD to obtain various modal signals. Then, free decay response of each mode was obtained using NEXT technique. Finally, the direct interpolation (DI) method and the curve fitting one were used to estimate various modes’ natural frequencies and damping ratios. The numerical simulation for a 3-story frame structure was employed to verify the correctness and reliability of the proposed method. The measured acceleration response data of CITIC Plaza in Guangzhou under Typhoon Damrey were analyzed using the proposed approach. The estimated structural modal parameters were compared with those obtained using other recognition methods to further verify the correctness and reliability of the proposed method.