Liquefaction hazard analysis is crucial in earthquake-prone regions as it magnifies structural damage. In this study, standard penetration test (SPT) and shear wave velocity (Vs) data of Chittagong City have been used to assess the liquefaction resistance of soils using artificial neural network (ANN). For a scenario of 7.5 magnitude (Mw) earthquake in Chittagong City, estimating the liquefaction-resistance involves utilizing peak horizontal ground acceleration (PGA) values of 0.15 and 0.28 g. Then, liquefaction potential index (LPI) is determined to assess the severity of liquefaction. In most boreholes, the LPI values are generally higher, with slightly elevated values in SPT data compared to Vs data. The current study suggests that the Valley Alluvium, Beach and Dune Sand may experience extreme liquefaction with LPI values ranges from 9.55 to 55.03 and 0 to 37.17 for SPT and Vs respectively, under a PGA of 0.15 g. Furthermore, LPI values ranges from 25.55 to 71.45 and 9.55 to 54.39 for SPT and Vs correspondingly. The liquefaction hazard map can be utilized to protect public safety, infrastructure, and to create a more resilient Chittagong City.

Buried pipelines are essential for the safe and efficient transportation of energy products such as oil, gas, and various chemical fluids. However, these pipelines are highly vulnerable to ground movements caused by geohazards such as seismic faults, landslide, liquefaction-induced lateral spreading, and soil creep, which can result in potential pipeline failures such as leaks or explosions. Response prediction of buried pipelines under such movements is critical for ensuring structural integrity, mitigating environmental risks, and avoiding costly disruptions. As such, this study adopts a Physics-Informed Neural Networks (PINNs) approach, integrated with a transfer learning technique, to predict structural response (e.g., strain) of both unreinforced and reinforced steel pipes subjected to Permanent Ground Displacement (PGD). The PINN method offers a meshless, simulation-free alternative to traditional numerical methods such as Finite Element Method (FEM) and Finite Difference Method (FDM), while eliminating the need for training data, unlike conventional machine learning approaches. The analyses can provide useful information for in-service pipe integrity assessment and reinforcement, if needed. The accuracy of the predicted results is verified against Finite Element (FE) and Finite Difference (FD) methods, showcasing the capability of PINNs in accurately predicting displacement and strain fields in pipelines under geohazard-induced ground movement.

The ability to predict the soil mechanical parameters swiftly is critical for off-road vehicle mobility. This paper introduces a novel interpretation methodology for determining critical soil mechanical parameters by impact penetration tests, enabling rapid and remote assessment of terramechanics properties. Initially, the method employs the Mohr-Coulomb constitutive model and the Coupled Eulerian-Lagrangian (CEL) finite element method to generate a dataset of soil impact penetration resistance and acceleration responses. Subsequently, a Radial Basis Function (RBF) neural network is employed as a surrogate model and integrated with the Nondominated Sorting Genetic Algorithm II (NSGA-II) to accurately interpret parameters such as density, cohesion, internal friction angle, elastic modulus, and Poisson's ratio. Experimental validation using sand and silty clay from Yangbaijing, Tibet, confirmed the accuracy and robustness of the method. The results indicate that the mean absolute percentage error for interpreted values was below 25%, with relative errors for some key parameters even below 10%. Furthermore, each single-condition calculation was completed on a standard computer in less than one minute. Comparative analyses with other algorithms, including MIGA and POS, demonstrated the superior performance of NSGA-II in avoiding local optima. The proposed interpretation framework offers a rapid, reliable, and remote solution for identifying the soil mechanical properties. Its potential applications range from disaster mitigation and emergency response operations to extraterrestrial soil exploration and other scenarios where in-situ investigations are challenging.

The hilly and mountainous regions of China are characterized by unique features such as small plots of land, steep slopes, fragmented fields, and high soil viscosity, which result in a decline in the efficiency of conventional agricultural machinery, or even render its use impractical. To address this issue, this study developed a micro universal chassis adapted to hilly terrains. First, a four-wheel-drive multifunctional electric micro chassis was designed, considering the terrain characteristics of hilly regions and the agronomic requirements of maizesoybean strip intercropping. Second, the kinematics of the chassis were modeled and analyzed to determine optimal posture control strategies, and a fuzzy RBF neural network-based PID control algorithm was designed to enable dynamic adjustment of the chassis. Then, extensive testing was conducted on the prototype chassis, including straight-line driving tests, steering tests, climbing tests, and passability tests, which demonstrated its excellent operational performance. The straight-line driving tests showed an average lateral deviation of 30 mm and a maximum deviation of 60 mm, while the in-situ steering tests recorded a deviation of 20 mm. Finally, the prototype was applied to field weeding operations, where results indicated that its performance, including travel speed, weeding efficiency, and seedling damage rate, significantly outperformed existing traditional models. The findings suggest that the designed multifunctional micro universal chassis is highly effective for use in hilly and mountainous regions, with superior performance particularly under intercropping systems.

The main problem in expansive soil treatment with steel slag (SS) is the relatively slow hydration reaction that occurs during the initial period. To circumvent this, SS-treated expansive soil activated by metakaolin (MK) under an alkaline environment was investigated in this study. Based on a series of tests on the engineering properties of the treated soil, it can be reported that SS could enhance the strength and compressibility of expansive soil, with strength increasing by approximately 108 % for SS contents exceeding 10 % compared to 3 % lime-treated soil, and the compression index reducing by 20 %. Further addition of MK plays a dual role, enhancing strength for higher SS content while excessive MK leads to strength reduction due to insufficient pozzolanic reactions and hydration product transformation. Expansive and shrinkage behaviors are notably improved, with a 5 % increase in SS content reducing the free swelling ratio by 0.66 %-5.9 %, and the combination of 15 % SS and 6 % MK achieving a nearly 300 % reduction in the linear shrinkage ratio. Microstructural analysis confirms the formation of hydration gels, densification of the soil structure, and reduced macropores, validating the enhanced mechanical and shrinkage resistance properties of the SS-MK-treated expansive soil. Additionally, to develop predictive models for mechanical and the content of hardening agents (SS and MK), the experimental data are processed utilizing a backpropagation neural network (BPNN). The results of BPNN modeling predict the mechanical properties perfectly, and the correlation coefficient (R) approaches up to 0.98.

Understanding slope stability is crucial for effective risk management and prevention of slides. Some deterministic approaches based on limit-equilibrium and numerical methods have been proposed for the assessment of the safety factor (SF) for a given soil slope. However, for risk analyses of slides of earth dams, a range of SFs is required due to uncertainties associated with soil strength properties as well as slope geometry. Recently, several studies have demonstrated the efficiency of artificial neural network (ANN) models in predicting the SF of natural and artificial slopes. Nevertheless, such techniques operate as black-box models, prioritizing predictive accuracy without suitable interpretability. Alternatively, multivariate polynomial regression (MVR) models offer a pragmatic interpretability strategy by combining the analysis of variance with a response surface methodology. This approach overcomes the difficulties associated with the interpretability of the black-box models, but results in limited accuracy when the relationship between independent and dependent variables is highly nonlinear. In this study, two models for a quick assessment of slope SF in earth dams are proposed considering the MVR and the ANN models. Initially, a synthetic dataset was generated considering different soil properties and slope geometries. Then, both models were evaluated and compared using unseen data. The results are also discussed from a geotechnical point of view, showing the impact of each input parameter on the assessment of the SF. Finally, the accuracy of both models was measured and compared using a real-case database. The obtained accuracy was 78% for the ANN model and 72% for the MVR one, demonstrating a great performance for both proposed models. The efficacy of the ANN model was also observed through its capacity to reduce false negatives (a stable prediction when it is not), resulting in a model more favorable to safety assessment.

This study utilizes a combined approach of Finite Element Method (FEM) simulation and Artificial Neural Network (ANN) modeling to analyze and predict the load-displacement relationship of bored piles in clayey sand. FEM is applied to simulate the nonlinear relationship between load and vertical displacement, with input parameters including load and the mechanical properties of the soil. The results obtained from FEM are used as input data for the ANN model, enabling accurate predictions of vertical displacement based on these parameters. The findings of this study show that the predicted ultimate bearing capacity of the bored piles is highly accurate, with negligible error when compared to field experiments. The ANN model achieved a high level of accuracy, as reflected by an R2 value of 0.9992, demonstrating the feasibility of applying machine learning in pile load analysis. This research provides a novel, efficient, and feasible approach for analyzing and predicting the bearing capacity of bored piles, while also paving the way for the application of machine learning in geotechnical engineering and foundation design. The integration of FEM and ANN not only minimizes errors compared to traditional methods but also significantly reduces time and costs when compared to field experiments.

Shield tunnels in operation are often affected by complex geological conditions, environmental factors, and structural aging, leading to cumulative damage in the segments and, consequently, increased deformation that compromises structural safety. To investigate the deformation behavior of tunnel linings under random damage conditions, this study integrates finite element numerical simulation with deep learning techniques to analyze and predict the deformation of shield tunnel segments. First, a refined three-dimensional finite element model was established, and a random damage modeling method was developed to simulate the deformation evolution of tunnel segments under different damage ratios. Additionally, a statistical analysis was conducted to assess the uncertainty in deformation caused by random damage. Furthermore, this study introduces a convolutional neural network (CNN) surrogate model to enable the rapid prediction of shield tunnel deformation under random damage conditions. The results indicate that as the damage ratio increases, both the mean deformation and its variability progressively rise, leading to increased deformation instability, demonstrating the cumulative effect of damage on segment deformation. Moreover, the 1D-CNN surrogate model was trained using finite element computation results, and predictions on the test dataset showed excellent agreement with FEM calculations. The surrogate model achieved a correlation coefficient (R2) exceeding 0.95 and an RMSE below 0.016 mm, confirming its ability to accurately predict the deformation of tunnel segments across different damage conditions. To the best of our knowledge, the finite-element-deep-learning hybrid approach proposed in this study provides a valuable theoretical foundation for predicting the deformation of in-service shield tunnels and assessing structural safety, offering scientific guidance for tunnel safety evaluation and damage repair strategies.

Determining the burial depth for offshore pipelines to resist impact load is challenging owing to the spatial variability of soil strengths, which proves to significantly affect failure behaviours of soils and pipelines. To facilitate the design, accurate and fast evaluation on pipeline damage is required. Here, an integrated surrogate model was developed to forecast impact damage of pipelines buried in spatially varied soils. Through coupling the random field and numerical simulation, a stochastic finite element analysis framework was derived and verified to yield the datasets; Based on the scheme of feature extraction - integration from convolution neural network, the surrogate model was established, which mapped the three-dimensional soil spatial field to the structural response. Prediction mechanism of the developed model was explored, where correlations among soil spatial distribution patterns, failure mechanisms and feature recognitions were discussed. The models enabled to capture the key features representing the failure mechanisms under random soil conditions, including the local failure mode of soil and pipe-soil interactions, which theoretically explained its feasibility in damage estimation. Further, model performance was comprehensively evaluated with regard to prediction accuracy, uncertainty quantification, and transfer learning, and the corresponding causes were investigated. Satisfactory performance and high computation efficiency were demonstrated.

The joint roughness coefficient (JRC) is a key parameter in the assessment of mechanical properties and the stability of rock masses. This paper presents a novel approach to JRC evaluation using a genetic algorithm-optimized backpropagation (GA-BP) neural network. Conventional JRC evaluations have typically depended on two-dimensional (2D) and three-dimensional (3D) parameter calculation methods, which fail to fully capture the nonlinear relationship between the complex surface morphology of joints and their roughness. Our analysis from shear tests on eight different joint types revealed that the strength and failure characteristics of the joints not only exhibit directional dependence but also positively correlate with surface dip angles, heights, and back slope morphological features. Subsequently, five simple statistical parameters, i.e. average dip angle, median dip angle, average height, height coefficient of variation, and back slope feature value (K), were utilized to quantify these characteristics. For the prediction of JRC, we compiled and analyzed 105 datasets, each containing these five statistical parameters and their corresponding JRC values. A GA-BP neural network model was then constructed using this dataset, with the five morphological characteristic statistics serving as inputs and the JRC values as outputs. A comparative analysis was performed between the GA-BP neural network model, the statistical parameter method, and the fractal parameter method. This analysis confirmed that our proposed method offers higher accuracy in evaluating the roughness coefficient and shear strength of joints. (c) 2025 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/).