基于运动放大的振动结构的模态识别

Modal identification of Vibration structure based on motion magnification

摘要:

快速准确地获取结构的振动信息是确定结构模态参数的关键。随着计算机视觉和高速相机技术的发展,视觉测量振动逐渐受到人们的重视。对一种可检测结构微小运动的计算机视觉技术进行了研究,并将最新的实时运动放大算法用于振动结构的模态识别。此算法可在不提取位移的情况下直接展示振动结构的模态特征。为了验证提出方法的可行性,以悬臂梁为例搭建了加速度计和高速相机的模态识别实验系统,并依据理论模态对其实验数据进行模态置信准则(MAC)检验。分析实验结果表明：提出的方法可以识别振动结构的模态参数。

Abstract:

Modal analysis of structures depends on the accurate and swift collection of data from a vibrating structure so the data can be analyzed to determine the modal characteristics. With the development of computer vision techniques and high-speed camera techniques, vision measurement vibration is gradually attended by people. This paper researches a motion magnification technique based on computer vision, which can detect the small sub-pixel motion in vibration structure. And we use the new real-time motion magnification algorithm to identify mode of vibration structure, which can determine directly the modal characteristics without extracting displacements. Taking the cantilever beam as a model, and setting the accelerometer measurements and high-speed camera modal identification experimental systems for verification on the method we proposed is viable, and verifying the test results based on modal assurance criterion(MAC). The results demonstrate the method we researched can be applied to modal identification of vibration structure.