桥梁结构动力测试信号的自适应形态学滤波研究

Research on Adaptive Morphological Filter of Dynamic Test Signals for Bridge Structure

桥梁健康监测数据中不可避免地会掺杂系统噪声和测试噪声,噪声的存在将严重影响桥梁状态评估的准确性。为了抑制噪声对桥梁状态评估的影响,获得精确的桥梁状态评估结果,本文提出了一种自适应的形态学滤波器(Adaptive Morphological Filter,AMF)。首先比选了AMF的各类结构元素类型,并依据AMF对信号傅里叶谱相对幅值的影响程度确定适宜的结构元素尺寸,据此构建滤波器。实现原始信号的自适应滤波后,用确定-随机子空间识别方法对原始数据和滤波后数据的模态参数进行识别,并使用稳定图法对滤波效果进行对比。以大比尺斜拉桥模型试验的测试数据为支撑对所提算法进行了验证。结果表明,AMF可显著提高桥梁模态参数识别结果的稳定性,并能挖掘出被噪声淹没的高阶桥梁模态参数。

The measured data of bridge health monitoring is inevitably interfered by systematic noise and test noise,the noise will seriously reduce the accuracy of evaluation of bridge state.In order to suppress the inf luence of noise on the evaluation of bridge state and obtain accurate evaluation results,an adaptive morphological f ilter(AMF)was proposed in this paper.First of all,the structural element types of AMF were compared and the appropriate structural element size was determined on the condition that the degree of inf luence of AMF on the relative amplitude of the signal Fourier spectrum.Then,the AMF was established to be applied to adaptive f ilter for original signal.Finally,the combined deterministic-stochastic subspace identif ication(CDSI)method was respectively applied on the original data and the f iltered data to identify the modal parameters.The stabilization diagram method was adopted to compare the f iltering effects.The proposed algorithm was verif ied with the data of a large scale cable-stayed bridge model test.The results show that the AMF can signif icantly improve the stability of the modal parameter identif ication results of bridge,and dig out the high order modal parameter submerged by noise.