Yuequan Bao

Although the guided wave testing is a common method for the non-destructive testing of plates, continuous wave excitation is rarely employed in practical applications. Therefore, this study proposes a damage localization method that integrates compressive sensing and continuous wave excitation, which achieves damage location prediction via the two-dimensional fast Fourier transform (2D-FFT) technique. First, sparse laser-acquired signals are reconstructed into a full wavefield by using a group sparsity-aware convolutional neural network, thereby ensuring compliance with the sampling requirements for 2D-FFT. When continuous waves are excited, primarily one mode of waves is excited in the plate, which manifests as a single peak in the frequency-wavenumber spectrum. Notably, the amplitude of this peak correlates directly with the damage location. Leveraging this characteristic, the damage location can be calculated by identifying the lines with the maximum amplitude from randomly selected reference points and determining the centroid of their intersection. However, for damage outside the measurement region, deviations arise between the actual and predicted locations. A deep forest model thus adjusts localization using the relationship between the amplitude of slices and the tangent of the line passing through the damage. Owing to the steady-state nature of continuous wave excitation, the proposed method extracts frequency-wavenumber features that are inherently more robust to measurement noise than conventional transient-excitation approaches. Both simulated and experimental results of steel plates demonstrate that the proposed method can precisely locate damage localizations both within and outside the measurement region.

Long-term bridge monitoring systems inevitably contain missing or unusable data segments after sensor faults, equipment failures, transmission interruptions, and other non-structural data-quality problems are identified and removed. Traditional reconstruction methods struggle to preserve the non-stationary, multi-scale, and sensor-dependent characteristics of bridge responses. This article proposes a mechanics-prior latent diffusion model for high-fidelity bridge monitoring data reconstruction. The method converts one-dimensional monitoring sequences into image representations, maps them into a regularized latent space through a pre-trained variational autoencoder, and trains a conditional diffusion U-Net to recover missing regions through reverse denoising. Mechanics-prior condition encoding is introduced as a global prior constraint, incorporating frequency-domain energy distribution and quasi-static morphological features tailored to different sensor types. Validation using numerical simulations and real bridge data, including acceleration, Global Positioning System (GPS) displacement, and strain, shows robust reconstruction under both long continuous gaps and fragmented missing scenarios. In representative 40% long-gap cases, the proposed method achieves NRMSE values of 0.0921, 0.0620, and 0.0523 for acceleration, GPS displacement, and strain, respectively, outperforming LSTM, GAN, and cubic spline interpolation baselines.

Urban bridge systems underpin sustainable mobility and metropolitan productivity, yet assets exposed to similar geology, climate, and traffic loads can exhibit correlated displacement responses at the group level. Detecting such correlations is essential for resilience-oriented asset management because it enables prioritization of inspection and coordinated maintenance planning across a city. Persistent scatterer interferometric synthetic aperture radar (PS-InSAR) provides dense, city-scale displacement time series, but converting high-dimensional monitoring data into actionable and interpretable bridge cohorts remains challenging. This paper develops uniform manifold approximation and projection with kernel density–based pattern identification (UMAP-KDPI), an unsupervised algorithmic framework that transforms PS-InSAR time series into bridge-group maintenance decision units via interpretable manifold learning and complex network analysis. Displacement histories are embedded using UMAP-KDPI to preserve temporal shape similarity and to extract four characteristic displacement patterns. Each bridge is then represented by a pattern-distribution signature, from which a weighted bridge-to-bridge similarity network is constructed and partitioned using the Louvain community detection algorithm to identify cohesive deformation subnetworks. Using 35 real satellite acquisitions on eight major cross-river bridges in a major metropolis in Central China, UMAP-KDPI reveals two stable subnetworks with distinct displacement-pattern compositions. Spatial organization, displacement-field consistency, and sensitivity analyses support the validity and robustness of the identified subnetworks, providing an interpretable system-level basis for group-wise inspection, coordinated interventions, and early risk mitigation.

Bridge and tunnel groups are integral components of urban infrastructure, and evaluating their service state is essential for ensuring operational safety and resilience. Current research often analyzes structures in isolation, neglecting the systemic interdependencies and spatiotemporal correlations inherent in a network. To address this gap, this paper proposes a novel, data-driven framework for the intelligent evaluation of the service state of urban bridgetunnel groups. The methodology includes traffic flow modelling of the spatiotemporal prediction and vehicle load estimation, deformation field analysis method based on PS-InSAR, and vulnerability analysis based on complex network theory. An actual bridge-tunnel groups with 681 bridges and tunnels are employed to illustrate the ability of the proposed method. The results demonstrate that the proposed method can accurately model network behavior, predict operational loads, identify systemic deformation patterns, and quantify vulnerability to disruptions. This research provides a scientific and holistic basis for enhancing the operational management, maintenance planning, and resilience of urban bridge-tunnel systems.

Performing Bayesian inference of structural damage detection is always computationally intractable, as the corresponding posterior probability density function (PDF) is commonly configured with hundreds/thousands of parameters. For Markov Chain Monte Carlo (MCMC) methods, thousands of likelihood calls are required for stochastic exploration of the parameter space, while evaluation of the likelihood function requires complex forward modeling for each set of structural parameters explored during the sampling process. This study proposes an accelerated Bayesian inference (ABI) algorithm for structural damage identification based on a convolutional neural network (CNN)‐based surrogate model. As an emulator for the likelihood function, the CNN is constructed using samples from the prior distribution of damage parameters, and the weighted sampling is employed to select uniformly distributed samples in the parameter space for model training. The Metropolis–Hasting (MH) algorithm with this surrogate model is then integrated into the delayed acceptance framework, in which the surrogate model behaves as an approximate model to generate easily computed proposals transferred to the exact model, thus reducing the computational burden imposed by repeated evaluations of the computationally expensive likelihood. Considering the non‐negligible bias of the surrogate model, the enhanced error model is used to adaptively correct the surrogate model using online posterior samples. An experimental three‐story frame and a numerical cable‐stayed bridge are utilized to verify the effectiveness of the proposed method. Compared to traditional sampling methods, the computational cost of damage detection can be reduced by up to 75%, and at the same time the damage identification accuracy has been improved, especially for large‐scale complex structures. Although the proposed algorithm is developed for a particular structural damage identification problem, it can be extended to all other models with a costly likelihood for accelerating uncertainty quantification. Moreover, the proposed CNN surrogate model can be incorporated into other sampling‐based methods.