An Artificial Filter Bank (AFB) for Structural Health Monitoring of Civil Structures - Part1: A Concept and Design of the AFB

Abstract

This study developed an artificial filter bank (AFB) that emulates the auditory system in order to efficiently obtain the significant dynamic responses. The developed AFB is composed of digitally coded band-pass filter optimizing algorithm (BOA) and peak-picking algorithm (PPA). The developed AFB was then optimized using the random El-Centro earthquake wave, which is often used in structural dynamic studies, and the reconstructed signal and compressed signal were compared to the original signal using the reconstruction error (RE) and compressive ratio (CR). The AFB developed, designed, and evaluated in this study adequately captures the time and frequency information of the original signal while capturing the significant dynamic response compressed about the frequency range of interest. Also the digitally developed AFB overcomes the shortcomings of conventional analog filter banks, supporting for quick and easy software-based implementation of new logics, optimizations, and designs. Ultimately the AFB developed in this study is expected to provide new paradigm in the structural health monitoring of construction structures as well as other structures.