Hyper-velocity impact (HVI) caused by a collision between orbital debris and spacecraft exists widely in outer space, and it poses a threat to spacecraft. This paper proposes a probabilistic hyperbola method based on Lamb waves analysis to detect and locate the impact in stiffened aluminum (Al) plates. A hybrid model using finite element analysis (FEA) and smoothed particle hydrodynamics (SPH) was developed to gain an insight into characteristics of HVI-induced acoustic emission (AE) and shock wave propagation. In addition, an experimental validation was carried out with a two-stage light gas gun, giving an aluminum projectile a velocity of several kilometers per second. Then a quantitative agreement is obtained between numerical and experimental results, demonstrating the correctness of the hybrid model and facilitating the explanation of received AE signals in experiments. Signal analysis shows that the shock wave quickly converts to a Lamb wave as it propagates from the HVI spot, and the zeroth-order symmetric wave mode (S0) dominates wave signal energy. The S0 wave is dispersive and shows a wide frequency range, with dominant magnitudes below 500 kHz. Finally, the HVI experiment results obtained with a light gas gun showed that the average location error could be less than 1 cm with only four sensors for a 1-square-meter stiffened metal plate.

This study proposes an improved two-dimensional finite-difference time-domain method (2D-FDTD) to simulate the propagation of guided waves in three-dimensional (3D) stiffened plates, which can greatly reduce the calculation amount and accelerate the simulation process. This method can be used for damage-detection imaging based on time-reversal with wavefield reconstruction (TR-RW) as a quick numerical simulation of wavefield is available. First, the piezoelectric wafer active sensors excite and collect Lamb wave signals in the damaged stiffening plate, after which these signals are reversed in the time domain. Second, the TR signals are again triggered by each sensor in turn on the reconstructed numerical model of the tested structure with 2D-FDTD. Third, the wavefield animation is obtained through simulation. Finally, the energy of the wavefield can be focused in certain areas, indicating the location of the damage. As an example, TR-RW, combined with 2D-FDTD method, is adopted to detect damage in an integral grid-stiffened plate, showing that the proposed 2D method can decrease the calculation time from dozens of minutes in the 3D model to dozens of seconds, which is of great significance for damage-detection applications. Furthermore, compared with the virtual TR method, the combined TR-RW and 2D-FDTD method can overcome the multipath effect and improve the damage location accuracy of stiffened structures.

Accurate impact localization on spacecraft structural panels subjected to contact loading by on-orbit servicing robots is critical for real-time structural health monitoring (SHM), yet remains challenging due to heterogeneous elastic wave propagation in complex aluminum structures with stiffener ribs and bonded joints. Conventional Received Signal Strength Indicator (RSSI)-based weighted centroid methods rely on fixed path-loss exponents that cannot accommodate spatially varying wave attenuation, resulting in position-dependent localization errors that worsen significantly near structural discontinuities. This paper proposes a Crow Search Algorithm (CSA)-optimized adaptive weighted centroid algorithm using distributed Fiber Bragg Grating (FBG) sensors, featuring three principal innovations: (i) a novel FBG wavelength-shift-to-RSSI amplitude mapping derived from elastic wave attenuation theory, bridging optical fiber sensing with centroid localization; (ii) per-event online weight optimization via CSA that adapts sensor contributions to each individual impact’s strain-wave signature; and (iii) a multi-objective fitness function simultaneously optimizing localization accuracy, noise robustness, and temporal consistency. The proposed method is validated across 200 impact events distributed over five representative positions on a 1 m³ Al6061 satellite-like structure with 64 FBG sensors (8 × 8 grid, 125 mm pitch), under three Gaussian noise levels (σ = 1%, 3%, 5% of signal RMS), and benchmarked against classical weighted centroid (WC), PSO-WC, GA-WC, DE-WC, and GWO-WC using paired t-tests (p < 0.01). CSA-WC achieves a mean localization error of 4.63 mm—an 83.29% improvement over classical WC and the lowest error among all five compared algorithms—with an average computation time of 0.14 s per event, satisfying real-time monitoring requirements.

Accurate detection, quantitative assessment, and localization of delamination damage in composite materials remain major challenges in structural health monitoring (SHM). This paper proposes Adaptive Graph Transformer (AGFormer), a framework integrating a graph convolutional neural network with Transformer encoders, for size quantification, severity grading, and localization of circular delamination damage in carbon-fiber-reinforced polymer (CFRP) laminates. First, ten-dimensional edge features, including root-mean-square deviation, correlation coefficient, energy ratio, time-of-flight shift, and spectral variation, are extracted from differences between multipath Lamb-wave baseline and damaged-state signals. These features are aggregated into 42-dimensional enhanced node features to fuse multipath damage information at the sensor nodes. Second, a multimodal adjacency-matrix construction strategy combining sensor spatial distance with signal time–frequency distance is proposed to provide an adaptive topology for the graph convolutional network. On this basis, the AGFormer model employs three residual graph-convolution modules to extract local topological features, a multilayer perceptron for dimensional mapping, and two multi-head self-attention Transformer encoder layers to capture long-range dependencies among sensors. Attention-gated pooling and parallel output heads then simultaneously perform damage-radius regression and four-class severity classification. In addition, damage-centroid localization is achieved by combining multipath damage indices with the RAPID (Reconstruction Algorithm for Probabilistic Inspection of Damage) elliptical imaging algorithm. Experimental validation was conducted on a 450 mm × 450 mm six-ply CFRP laminate instrumented with a 12-piezoelectric-sensor network, under progressively increasing simulated delamination damage. The model was trained and evaluated using five-fold cross-validation grouped by damage size, while Gaussian-noise injection and random masking were introduced for data augmentation to improve generalization. The results demonstrate that the proposed method can effectively quantify and grade delamination damage of different sizes, achieving a mean absolute error of 4.38 mm, a root-mean-square error of 5.37 mm, and an R² of 0.920 for damage-radius regression, an accuracy of 81.3% for four-class severity classification, and a mean localization error of 6.50 mm, thereby offering a new technical approach to intelligent health monitoring of composite structures.