Related Experiment Video
Updated: Oct 17, 2026

Image Recognition and Parameter Analysis of Concrete Vibration State Based on Support Vector Machine
Published on: January 5, 2024
Signal-Processing Procedure for Preliminary Concrete Anomaly Screening Based on Continuous Sliding Acoustic
Bochen Sheng1, Shibin Teng1, Qilong Xu1
1School of Smart City Engineering, Qingdao Huanghai University, Qingdao 266427, China.
Abstract:
Internal discontinuities in concrete walls are hidden conditions that can compromise local structural capacity and durability. Rapid and objective on-site anomaly screening is therefore important for structural integrity evaluation. This paper presents a signal-processing procedure for preliminary concrete anomaly screening based on continuous sliding acoustic excitation. Acoustic signals were collected from local reference baseline regions and candidate anomaly regions along defined paths on an in-service reinforced-concrete flue wall. Time-domain waveforms, frequency-domain spectra, and short-time Fourier transform (STFT) distributions were combined to characterize the differences between reference baseline and candidate anomaly regions. Reference signals showed comparatively stable envelopes and concentrated spectral energy, whereas candidate anomaly regions exhibited greater amplitude dispersion, broader spectral distributions, and more pronounced non-stationary time-frequency features. A relative-variation screening workflow based on amplitude, spectral energy, and time-frequency features was formulated. The workflow supports the transition from subjective auditory judgment to quantitative signal analysis and is intended as an efficient Tier-1 preliminary screening approach for concrete structures. Because no destructive coring or paired NDT confirmation was performed on this in-service structure, the detected regions are designated strictly as candidate anomalies rather than confirmed voids.
