Application of Discrete Wavelet Transformation to Defect Detection in Truss Structures with Rigidly Connected Bars

The structures examined in this paper are bridge-type trusses that were previously used as railway viaduct support structures. The considered trusses are modelled as 2D and 3D structures. The lower chord bar of the considered structure can be loaded by external forces located outside the rigid nodes (the points where truss bars are connected). Hence, in the numerical experiment in terms of 2D approach, the truss structure consists of the set of two-node beam finite elements with three degrees of freedom per node and exact shape functions. According to 3D approach, the truss is described as the set of two-node beam elements with six degrees of freedom per node. Axial and twisting displacements of the element are described by linear shape functions and the bending is described by polynomials of the third order corresponding to Euler-Bernoulli beam fields of deformation. The defect (damage) in truss structure is modelled as the local stiffness reduction of one or two lower chord bars. The analysis of a structural response is carried out using the discrete wavelet transformation (DWT). The aim if this work is to detect the localization of damage provided that it exists in the considered structure and to examine whether the DWT will prove to be the effective tool to defect detection. It is expected that the disturbance of the response signal will appear in the vicinity of the point where the defect exists. The family of Daubechies 4 wavelet is implemented. Numerical investigation is executed based on signal analysis of structural static response. Some numerical examples are presented.