Statistical Time Features-based Methodology for Fatigue Cracks Detection in a Four-Story Building

Abstract

Fatigue cracks are the most common damages encountered in the civil structures. For this reason, early detection of these types of damages can reduce their accumulation in the elements that conform to a civil structure, preventing its possible collapse. In this sense, this work investigates the usefulness of the statistical time features (STFs) for identifying this type of damages in a four-story building subjected to dynamic forced excitations. The proposed methodology's efficacy is corroborated using three levels of damage (i.e., light-, medium-, and severe-fatigue crack). These conditions are artificially created. Results show that the STFs can determine a building exposed to diverse levels of fatigue cracks with high accuracy using only one sensor placed on the structure.