Identification of local debonding in bolted panels using nonlinear pseudo-forces

Abstract

Bolt looseness can occur subject to long-term structural service. With this concern of structural integrity and safety, there is a huge demand to identify bolt-looseness-caused local debonding in connected structural components such as bolted panels. With the aid of non-contact laser scanning, transverse Operating Deflection Shapes (ODSs) of a bolted panel can be measured with high spatial resolutions. Perturbation to the linear transverse dynamic equilibrium of the panel can be regarded as the Linear Pseudo-Force (LPF), which is applied to the debonding region only and vanishes at undamaged locations. However, nonlinearities caused by contact of debonding interfaces during vibrations are not taken into consideration in the LPF model. As a consequence, only linear damage features can be contained in the LPFs which are established on linear ODSs, leading to incompleteness of damage features. Addressing this problem, this study establishes a Nonlinear Pseudo-Force (NPF) model from the nonlinear transverse motion of equation of a beam-type bi-layer panel model with local debonding. Superior to LPFs, NPFs can extract linear and nonlinear damage features from linear and nonlinear ODSs, respectively. Similar to LPFs, NPFs concentrate in the debonding regions to form local peaks. Therefore, the NPF can be utilized as an ideal nonlinear indicator for the identification of local debonding in bolted panels. The applicability of the NPF is experimentally validated by identifying width-through debonding in a steel panel connected by bolts, whose ODSs at linear and nonlinear (higher) harmonics are acquired through non-contact laser scanning measurement. Experimental results reveal that the NPF can extract complete linear and nonlinear damage features, and hence has a higher-dimensional capacity for identifying local debonding in connected structural components such as bolted panels, whose occurrence, location, and size can be graphically characterized.