Identification of breathing crack induced nonlinearity in dynamic responses of bridge structures using shape metrics in phase space topology

Abstract

The traditional system identification techniques using modal analysis assume the system as linear which may not be valid in the case of structures exhibiting breathing crack behavior during the initial damage stage. Detection of nonlinearity in the case of weakly nonlinear structures is essential for the detection of incipient damages as well as the initiation- and propagation- of cracks. Time series analysis techniques are found to be effective in extracting the nonlinear features for the characterization of system dynamics for timely fore-warning of failures. One of the promising features is the changes in phase space topology (derived from time-series information) for the detection of nonlinearity in dynamic responses of a damaged structure. In this paper, phase space topology is constructed using the acceleration-velocity-deflection responses obtained from controlled force excitation to a reinforced concrete bridge girder-deck structure at different stages of damage. Phase space trajectories from the intact and damaged structure (at different levels) are used to evaluate damage progression from the formulated dissimilarity methods. The integrity of the structure including the initial stage of damage is assessed based on dissimilarity features formulated by the change in the dynamics of the nonlinear system and determined through time-series based statistical shape analysis. The distortion in phase space topology is found to be an efficient nonlinear damage index feature and is very sensitive to the damages of low magnitude, distinguishing the structural nonlinearity from material one. It can thus be an effective tool in identifying the damage initiation through nonlinear features which form the base for structural health monitoring of weakly nonlinear civil infrastructures.