Fractal dimension-based methodology for detecting and quantifying the severity of fatigue cracks in a four-story structure

摘要

Civil structures (CS) frequently suffer damage during their service lives. The accumulation of these damages can cause a weakness of the structure and if these damages are not early detected, they can produce a collapse of it. Cracks produced by fatigue in the material are common deterioration situations that can be found in CS. In recent years, a concept named Structural Health Monitoring (SHM) is used for assessing the condition of the structures, with the purpose of generating a monitoring system capable of identifying damages in an early stage in order to avoid economical and human losses. One of steps more important in SHM scheme is the signal processing technique, which has to be capable of estimating features that allow assessing the condition of the civil structure. Hence, the purpose of this paper is to introduce a Fractal Dimension based methodology in order to identify and quantify cracks in a four-story structure subjected to forced excitations. The fractal value of a of time-series signals are calculated for three different FD methods: Katz's FD (KFD), Box Dimension (BD), and Higuchi's FD (HFD). The proposed FD methodology demonstrate to be effective for monitoring the structure condition.