A dual-signal feature fusion and rebar debonding detection framework based on unified multiscale features for RC structures

The reliable detection of hidden rebar debonding in reinforced concrete structures remains a critical challenge. Addressing this issue is essential for advancing structural health monitoring. The prevailing data-driven rebar debonding detection methods predominantly rely on single-dimensional physical features, which severely restricts their reliability in complex environments. To overcome this fundamental limitation, this article presents a novel application and integration framework for dual-signal feature fusion and debonding detection (DSFD) using unified multiscale features. The proposed methodology adopts multiscale dispersion Lempel–Ziv complexity to extract dynamic characteristics from dual signals of nonlinear harmonics (NLH) and linear sine sweep (SP). A gated mechanism is utilized to integrate these heterogeneous features, leveraging the high micro-damage sensitivity of NLH signals and the macroscopic robustness of SP signals. The state classification process involves implementing a multilayer perceptron. To ensure structural safety, a damage detection method that strictly adheres to the principle of ‘avoiding missed detection’ is adopted. The experimental validations demonstrate that the proposed gated fusion strategy achieves the mean test accuracies of 98.33 and 96.09%, respectively, under simple and complex damage scenarios. The spatial localization and quantification explicitly demonstrate that the framework can accurately evaluate the physical debonding length and boundaries achieving a remarkable dice similarity coefficient metric ranging from 0.70 to 1.00, thereby successfully fulfilling the strict structural safety redundancy requirements. The proposed framework shows notable resilience to noise interference, achieving the accuracy and F1_score that consistently exceed 80% and an area under the receiver operating characteristic curve that maintains a highly stable level exceeding 0.90 even under the signal-to-noise ratio of 15 dB. The proposed DSFD provides a highly robust and resilient technical approach for identifying hidden internal damage, enabling the transition to proactive early-warning maintenance of civil infrastructure.