A Gaussian Process Regression‐Based Multi‐Path Guided Wave Fusion Strategy for Delamination Evolution Tracking in Composite Laminates

ABSTRACT Accurate detection and evolution tracking of delamination damage in composite laminates pose significant challenges for structural health monitoring. This study proposes an integrated framework combining a novel damage index (DI) based on two‐dimensional continuous wavelet features of guided wave signals with a multi‐path fusion strategy employing gaussian process regression (GPR) for damage detection and evolution tracking. Initially, the received guided wave is transformed into a time‐frequency representation using the continuous wavelet transform (CWT). Singular value decomposition (SVD) is then applied to extract the principal characteristics from the acquired time‐frequency matrix, followed by the use of dynamic time warping (DTW) distance to characterize the dissimilarity between signals. Subsequently, to mitigate the influence of different path locations and enhance the robustness in damage monitoring, the quantification curves derived from multiple sensing paths are fused with GPR. This process yields damage evolution curves with associated confidence intervals, enabling more accurate and reliable tracking of delamination evolution. The proposed approach is experimentally validated through delamination tests on a carbon fiber reinforced polymer composite plate. Results demonstrate that the proposed index exhibits higher sensitivity to damage detection compared to other DI techniques, and the GPR‐based multipath fusion strategy provides more reliable damage evolution tracking under small‐sample conditions.