Use of Wavelets for Mixed-Mode Damage Diagnostics in Warren Truss Structures

The use of the wavelet transform has been gaining widespread acceptance over the last decade as a valuable tool for damage detection. This paper investigates the use of wavelets for detecting mixed-mode, also known as combined mode, cracks in truss structures. The propagation of an open, mixed-mode crack is simulated by using a macroscopic model of damage that is combined with a finite element model of the Warren truss. The natural modes of the truss with varying levels of damage are then used to determine crack location on a specific member of the truss. A damage detection algorithm is developed and the influence of multiple parameters such as truss geometry, crack geometry, number of truss members, etc. is investigated. A direct correlation between damage severity and the magnitude of wavelet coefficients is found for a predefined damage location. It is observed that the proposed damage detection algorithm can be used to successfully detect mixed-mode cracks even in the presence of noise, and even when a relatively coarse sampling of natural modes is used. Multiple simulations are presented and some shortcomings of the proposed algorithm are also discussed in detail.