Bond damage evolution and spatial micro-mechanism analysis of corroded reinforced concrete and failure mode assessment by acoustic emission

Bond performance is an important guarantee for the structural integrity and long-term service performance of reinforced concrete, which is seriously threatened by the reinforcing bar corrosion in corrosive environments. This study employs an experimental framework combining macroscopic pull-out test, microscopic scanning electron microscopy (SEM), microscopic X-ray photoelectron spectroscopy (XPS), and acoustic emission (AE) monitoring to investigate bond behavior evolution in corroded reinforced concrete with HRB400 ribbed reinforcing bar. Pull-out tests were conducted on specimens with varying corrosion extents (CE: 0%, 3%, 5%, 8%), complemented by SEM-XPS to elucidate spatial micro-mechanisms. AE technology was employed to monitor real-time damage progression, with Mel-frequency cepstral coefficients (MFCC) enabling precise segmentation of bond damage stages. Based on MFCC segmentation, the bond failure stages and modes are quantitatively classified by leveraging RA-AF values and Gaussian mixture model (GMM) clustering. Results revealed a 24.2% decline in bond strength at CE=8% compared to uncorroded specimens, accompanied by a transition in failure modes from splitting to hybrid splitting-pull-out debonding. Low corrosion (≤3%) enhances initial friction with ductile failure via rust compaction, while higher corrosion (≥5%) causes abrupt brittle failure due to corrosion-induced cracking that degrades bond integrity. Microscopic characterization analysis showed that corrosion products (e.g., porous FeO and brittle Fe(OH) 3 ) reduced mechanical interlock, while microcrack propagation and stress redistribution accelerated interfacial debonding. The bond-slip process was quantitatively divided into four stages using MFCC analysis, in which higher CE values led to the disappearance of some stages. GMM clustering of RA-AF values successfully differentiated shear-dominated (low CE) and tensile-shear mixed (high CE) failure modes. Based on Pearson correlation coefficient, an established predictive model for bond behavior demonstrated high accuracy (R 2 >0.9) in predicting bond strength and slip. This study bridges multi-scale mechanisms, providing some references for durability assessment and maintenance strategies of corrosion-affected reinforced concrete infrastructure.