Xuyan Tan

Monitoring and early warning systems are essential for ensuring the long-term stability of tunnels. However, existing methods often overlook the impact of multiple external factors, such as water pressure and temperature, on tunnel structural behavior, limiting their intelligence and accuracy. To address this gap, we propose a Dynamic Pre-Warning model (DPWNet), constructed as a Deep Probabilistic Autoregressive Architecture, which integrates deep learning to predict structural responses by considering spatiotemporal correlations and dynamic external loads. DPWNet incorporates Markov Chain Monte Carlo sampling during the decoding phase to simulate a range of potential working conditions, accounting for the complex interactions between environmental and structural factors. This probabilistic framework allows real-time adaptive adjustment of pre-warning thresholds based on scenario likelihoods. DPWNet is applied to an underwater shield tunnel, demonstrating significant improvements in accuracy. Specifically, the model calculates the 90% confidence interval of structural responses under multiple factors, translating probabilistic forecasts into actionable thresholds. Experimental results show that DPWNet reduces mean absolute error by 42.0%, root mean square error by 29.2%, and improves Pearson correlation coefficient by 1.3% over 30 days compared to existing methods. These results highlight the model’s reliability and its potential to advance the intelligent monitoring of underwater shield tunnels.

In tunnel structural health monitoring (SHM) systems, data completeness and accuracy are essential for tasks such as damage detection and early warning. However, environmental disturbances and sensor faults often cause significant missing data, making effective imputation a critical preprocessing step. Traditional statistical methods struggle to capture complex nonlinear temporal and cross‐feature dependencies, while autoregressive models, such as recurrent neural networks, suffer from error accumulation and difficulty adapting to dynamically varying strain distributions in real tunnels. To address these challenges, this work proposes a novel nonautoregressive imputation framework based on diffusion models, which effectively mitigate error accumulation. The model effectively exploits the informative content of observed data to guide the modeling and reconstruction of missing values. A gated temporal‐feature self‐attention fusion module is introduced to accurately capture the complex temporal and spatial dependencies of structural responses. Additionally, external environmental variables such as temperature and water level are integrated to jointly model structural responses and operating conditions, ensuring that the imputation remains robust even under harsh environmental conditions. The method is validated on two real‐world SHM datasets from tunnels in Nanjing and Wuhan with various missing data patterns. Experimental results show consistently robust and superior performance across different missing rates, maintaining high accuracy even under severe data loss, demonstrating its effectiveness and practical value in real SHM applications.

Research on early warning of tunnel anomalies is fundamental for achieving intelligent management. However, most current methods for determining early‐warning values of tunnel mechanics indicators are different to couple the nonlinear variation property and spatial positional difference. Therefore, this research presents a novel approach to tunnel early warning based on deep autoregressive learning method (DL‐AR) that considers the spatiotemporal correlations of structural mechanics responses, specifically tailored to dynamically determine the warning thresholds at different spatial positions. The methodology introduces a framework for predictive modeling and instantiates it on a typical underwater shield tunnel. After thoroughly learning the temporal and spatial correlations of structural mechanical responses, accurate predictions are made for the evolving trends of structural behaviors and the probabilities to the reasonable fluctuation range. Based on these predictions, spatially varying alert thresholds for structural behaviors are proposed. To ensure the reliability of the proposed model, a series of discussions and validation experiments are conducted. Results indicate that the proposed model effectively captured the spatiotemporal characteristics of structural evolution and identified alert ranges, defining permissible variations in structural trends. The prediction results showed near to 99% consistency with actual data, a 5% enhancement compared to classical models. Any deviation beyond this range triggers an early warning, demonstrating the efficacy of model in anticipating and responding to potential structural issues.

The long-term stability of tunnel structures is significantly influenced by the time-dependent behavior of the surrounding rock. Existing constitutive models often deviated from surrounding deformation due to the anisotropic nature of rock mass. In response, this study introduces a novel reinforced learning fusion constitutive model to accurately capture the time-dependent behaviors of soft rock. The framework and methodologies are first outlined, followed by the instantiation of the constitutive model of Burgers and creep parameters using laboratory testing data. To enhance accuracy, an XGBoost model is trained to reinforce the results of the constitutive model. The reliability of the proposed model is then validated against the original constitutive model and other representative machine learning models. Experimental findings demonstrate the superior characterization ability and stability of the presented reinforced model, where the calculation error reduces by 7.2E−06 at least, and [ ] score is improved by at least 1% to others. Consequently, the proposed model is reliable, offering a promising approach to capturing the actual time-dependent behaviors of tunnel surroundings in practical field applications.

High-speed railway suspension bridges (HSRSBs) have been constructed with the new advancements in technology. The deformation prediction for HSRSBs is essential to their safety and maintenance. The conventional prediction methods are developed for bridges without high-speed railway. Different factors, including temperature (TEMP), time delay compensation (TDC), train live load (TLL), are considered in these methods. However, the train side (TS) and train instantaneous position (TIP) have a significant impact on deformation for HSRSBs, and they are not used in the prediction. More importantly, the coupling issue among different factors is so significant that it cannot be neglected. In this study, we propose a deformation prediction model based on a backpropagation (BP) neural network. This model uses different factors as model input, including TEMP, TDC, TLL, TS, and TIP. The coupling issue is addressed by using the new model. The new model was evaluated using a dataset of 10-day field measurements. It achieves a mean absolute error (MAE) of 8.81 mm, a mean relative error (MRE) of 9.82%, and coefficient of determination (R²) of 0.94. The new model will provide high-precision prediction for deformation and will be used in the development of an early warning system.

It is crucial to predict future mechanical behaviors for the prevention of structural disasters. Especially for underground construction, the structural mechanical behaviors are affected by multiple internal and external factors due to the complex conditions. Given that the existing models fail to take into account all the factors and accurate prediction of the multiple time series simultaneously is difficult using these models, this study proposed an improved prediction model through the autoencoder fused long‐ and short‐term time‐series network driven by the mass number of monitoring data. Then, the proposed model was formalized on multiple time series of strain monitoring data. Also, the discussion analysis with a classical baseline and an ablation experiment was conducted to verify the effectiveness of the prediction model. As the results indicate, the proposed model shows obvious superiority in predicting the future mechanical behaviors of structures. As a case study, the presented model was applied to the Nanjing Dinghuaimen tunnel to predict the stain variation on a different time scale in the future.

Monitoring of the mechanical behavior of underwater shield tunnels is vital for ensuring their long‐term structural stability. Typically determined by empirical or semi‐empirical methods, the limited number of monitoring points and coarse monitoring schemes pose huge challenges in terms of capturing the complete mechanical state of the entire structure. Therefore, with the aim of optimizing the monitoring scheme, this study introduces a spatial deduction model for the stress distribution of the overall structure using a machine learning algorithm. Initially, clustering experiments were performed on a numerical data set to determine the typical positions of structural mechanical responses. Subsequently, supervised learning methods were applied to derive the data information across the entire surface by using the data from these typical positions, which allows flexibility in the number and combinations of these points. According to the evaluation results of the model under various conditions, the optimized number of monitoring points and their locations are determined. Experimental findings suggest that an excessive number of monitoring points results in information redundancy, thus diminishing the deduction capability. The primary positions for monitoring points are determined as the spandrel and hance of the tunnel structure, with the arch crown and inch arch serving as additional positions to enhance the monitoring network. Compared with common methods, the proposed model shows significantly improved characterization abilities, establishing its reliability for optimizing the monitoring scheme.