AbstractThe capability of the wave method for structural health monitoring, based on fitting a layered Timoshenko beam by waveform inversion of impulse responses, to localize damage is investigated using shake table experiment data. The specimen is a full-scale slice of a seven-story reinforced concrete (RC) shear wall building, which was tested on the University of California San Diego–Network for Earthquake Engineering Simulation (UCSD-NEES) outdoor shake table under consecutive exposure to four earthquake motions with increasing intensity, which progressively damaged the structure. The damage is localized by partitioning the beam into horizontal layers, comprising a single floor in the parts with the most severe damage and a group of floors in the lightly damaged parts, and detecting changes in the wave velocities in each layer with respect to a reference state. Small amplitude, ambient vibration response data were used as input for the inversion. The paper describes a number of improvements in the methodology, which were necessary to improve the quality of the inversion, the success of the method to detect the existence and the intensity of the damage in different parts of the structure, and how it fared in comparison to other methods applied to the same test structure. Also, recommendations are proposed for future improvements of the method.