Structural damage detections based on a general vibration model identification approach

Abstract

This paper presents a novel vibration-based damage detection method using a general vibration model identification (GVMI) approach. A damage index based on the identified general vibration model is constructed for damage detection and localization, with the damage being regarded as a virtual excitation on the structure. The proposed GVMI approach utilizes a general form of high order derivative equation that to be identified and further used to detect changes in the vibration characteristics of the structure. Therefore, the proposed damage detection method requires neither baseline signals nor prior knowledge on the structural parameters, thus offering great application potentials for complex structures with unknown parameters. As a proof-of-concept example, a honeycomb sandwich cantilever beam is investigated for validating the proposed approach. The influences of the key parameters on the detection resolution, such as the measurement interval, the order of the displacement derivative and the selection of the excitation frequency, are investigated. Furthermore, an enhanced version of the GVMI method with an excitation frequency extension is developed by using a data fusion scheme. Taking advantages of the broadband excitations, the blind inspection area can be completely eliminated, whilst improving the effectiveness and the accuracy of the detection. Experimental results with both single and multiple structural damage show the validity and the accuracy of the proposed approach.