A structural health monitoring approach using ERA, real numbered mode shape transformation, and basis mode screening

To improve structural health monitoring (SHM) and damage detection, the Real Numbered Mode Shape Transformation and Basis Mode Screening are developed. These methods are used with the Natural Excitation Technique, the Eigensystem Realization Algorithm (ERA), and a least squares stiffness solution method for a complete system identification and damage detection process. The Real Numbered Mode Shape Transformation converts complex mode shapes identified by ERA to real ones, such that their shapes are still close to the original complex mode shapes, to satisfy the undamped eigenvalue problem of the physical system. The Basis Mode Screening determines the best modal parameters by choosing modes that have shapes close to the theoretical mode shapes and show significantly high consistency between different parameter sets identified by ERA with different numbers of assumed mode. Using simulated and experimental cases of the IASC-ASCE SHM benchmark problem and the simulation case of the USC ceiling truss structure as testbeds, the proposed algorithms show favorably 90-100% accuracy in damage detection.