Mode shape based damage identification is a global method which diagnoses the structural damage through analyzing the change between structural mode shapes before and after the occurrence of defect at early stage, to prevent the large consecutive damage. Currently, the mode shape is normally obtained by measuring the steady-state vibration of structure via electric measurement techniques such as using accelerometers. This approach is complicated since an electrodynamic vibration system is required to generate steady exciting force and imprecise because of the contact testing which generates additional mass. In this project, a new approach based on knocking method is proposed to solve these problems. A newly reported multi-point laser Doppler vibrometry is introduced to measure the damped free vibration of structure excited by a shock. Vibration signals of multiple points are able to be determined synchronously by the proposed optical vibrometry and the vibratory response can be constructed using signal processing technique. The presented method enjoys the merits of being noncontact, nondestructive, simple and convenient. In most practical applications, the mode shape of intact structure used as baseline is unfeasible to be obtained experimentally, therefore signal processing techniques, such as wavelet transform, is usually adopted to analyze the experimental data on only damaged structure.In this project, to identify the damage, a windowed Fourier transform based processing approach is proposed to treat the mode shape data as a digital signal in spatial domain and detect its local irregularity induced by the damage. The principle is similar with that of wavelet analysis but some recent investigationes show that the windowed Fourier transform may be more suitable for signal type of vibratory mode shape. This project aims to realize a damage identification method based on the non-contact measurement and no baseline needed processing by using the combination of theoretical analysis, numerical simulation and experimental verification on several typical structures.
基于振型的损伤识别技术通过损伤前后结构振型的变化,来分析和识别损伤。目前普遍采用电测法测量物体的稳态振动来获得振型(响应)。稳态激励需要相对复杂的激振系统,而且电测法是接触式测量,精度不高。本项目拟采用一种新的多点激光多普勒测振技术,测量结构在冲击载荷作用下的自由衰减振动,从中提取振型。不同于扫描式激光多普勒测振仪,该技术能实现多点同步测量,从而可建立一种基于敲击法、更简单方便的非接触式损伤检测实验技术。另一方面,实际应用中,作为参考的完整件振型通常难以获取。本项目拟引入加窗傅里叶变换对测得的结构振型或者振型导数进行处理,分析损伤引起的振型局部奇异性,来对损伤进行诊断。该方法仅针对损伤件的振型数据,实现损伤检测。本项目将通过理论推导、数值模拟和典型结构件的实验,对提出的测量技术和算法进行研究和验证。成果将提供一种高精度的非接触式测量技术,简化目前的损伤识别方法。
本项目建立了基于欧拉-伯努利梁结构的损伤参数与固有频率的数学关系,提出了两种算法。在完整梁结构固有频率信息已知的情况下,可通过联立求解给出的公式来计算损伤参数,包括损伤的相对位置和相对程度。在完整梁结构固有频率信息未知的情况下,可采用三交线法来评估损伤。相比以前的频率方法,该理论适用于各种边界条件,纠正了部分错误理论,进行了解的唯一性分析,并且,两种算法都不需要结构的材料参数,有利于在实际中的应用。建立了基于同步多点激光多普勒测振方法的结构振型测量和损伤识别方法。该方法对结构受到冲击载荷作用后的自由振动进行测量,通过加窗傅里叶分析提取结构振型,并通过提出的相关系数调制的模态曲率平方差这一损伤分析指标来进行损伤定位;若针对欧拉-伯努利梁结构,可结合频率方法进行损伤程度分析。该方法基于非接触式测量手段和非稳态振动测量,较大程度地提高了方法的应用范围。
{{i.achievement_title}}
数据更新时间:2023-05-31
坚果破壳取仁与包装生产线控制系统设计
气载放射性碘采样测量方法研究进展
居住环境多维剥夺的地理识别及类型划分——以郑州主城区为例
基于细粒度词表示的命名实体识别研究
宽弦高速跨音风扇颤振特性研究
基于激光多普勒测振技术的梨坚实度在线实时无损检测方法研究
基于相干激光多普勒测振原理的复合材料无损检测能力评价及误差校正方法研究
基于激光多普勒振动分析的采后西瓜成熟度无损检测方法研究
大型复合材料结构快速无损检测与损伤识别方法研究