Analysis of stationary roving mass effect for damage detection in beams using wavelet analysis of mode shapes

11th International Conference on Damage Assessment of Structures

物理学;材料科学

Solís, Mario^1 ; Benjumea, Antonio J.^1 ; Algaba, Mario^1 ; Galvín, Pedro^1

Escuela Técnica Superior de Ingeniería, Universidad de Sevilla, Sevilla

41092, Spain^1

关键词: Continuous Wavelet Transform;    Damage detection technique;    Mathematical tools;    Measuring points;    Natural frequencies and modes;    Practical issues;    Singular behavior;    Wavelet coefficients;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/628/1/012014/pdf
DOI  :  10.1088/1742-6596/628/1/012014

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

One of the main challenges in damage detection techniques is sensitivity to damage. During the last years, a large number of papers have used wavelet analysis as a sensitive mathematical tool for identifying changes in mode shapes induced by damage. This paper analyzes the effect of adding a mass to the structure at different positions. Depending on the location and severity of damage, the presence of the mass affects the natural frequencies and mode shapes in a different way. The paper applies a damage detection methodology proposed by the authors, although it has been modified in order to consider the addition of the mas. This methodology is based on a wavelet analysis of the difference of mode shapes of a damaged and a reference state. The singular behavior of a normalized weighted addition of wavelet coefficients is used as an indicator of damage. The presence of damage is detected by combining all the information provided by mode shapes and natural frequencies for different positions of the roving mass. A continuous wavelet transform is used to detect the difference between the response of a healthy state and a damaged one. The paper shows the results obtained for a beam with different cracks. The paper analyzes the sensitivity to damage of the proposed methodology by considering some practical issues such as the size of the crack, the number of measuring points and the effect of experimental noise.