Time–Frequency Signal Processing for Integrity Assessment and Damage Localization of Concrete Piles

Abstract

This paper presents a new integrity assessment and damage localization method for piles based on one-dimensional wave propagation theory by integrating the analytical mode decomposition (AMD), recursive Hilbert transform (RHT) and complex continuous wavelet transform (CCWT) into a single assessment tool. The AMD is first used as a band pass filter to extract the mono-component over a frequency band of interest from the response of a pile head, aimed at attenuating the interference from various noisy signals. Then, the mono-component signal is demodulated into a purely frequency-modulated signal by means of RHT, which greatly reduces the interferences from the amplitude-modulated function. Finally, the CCWT is utilized to process the frequency-modulated signal and to calculate phase angles; the latter are subsequently mapped into the time–frequency domain to localize pile damage. The methodology is verified by a numerical example, in which a concrete pile is modeled by the finite element method considering the soil-pile interaction, and by an experimental case study on an actual pile. The results from the numerical and experimental examples demonstrate that the proposed method improves the efficiency of damage identification when compared with other three methods ([Formula: see text], [Formula: see text] and CCWT). In addition, the proposed method enables the localization of damage in full-scale piles situated in soil with an acceptable engineering accuracy by mutual validation with other pile integrity assessment methods, e.g. the ultrasonic emission method.