Search results

Wavelets are being increasingly used for damage diagnostics. The purpose of this paper is to present an algorithm that uses the wavelet transform for detecting mixed-mode, also known as combined mode, cracks in large truss structures.

The mixed-mode crack is modeled by superposing two damage modes, and this model is combined with a finite element model of the truss. The natural modes of the truss are processed through the wavelet transform and then used to determine the damage location. The influence of multiple parameters such as truss geometry, crack geometry, number of truss members, orientation of truss members, etc. is investigated as part of the study.

The proposed damage detection algorithm is found to be successful in detecting single mode as well as mixed-mode cracks even in the presence of significant end effects, and even when a relatively coarse sampling of natural modes is used. Results from multiple simulations that involve three commonly used truss structures are presented. A correlation between damage severity and the magnitude of wavelet coefficients is observed.

The proposed algorithm is found to be successful in accurately detecting damage, but direct determination of damage severity is found to be challenging.