Predicting the frost depth of soils in pavement design is critical to the sustainability of the pavement because of its mechanical vulnerability to frozen-thawed soil. The reliable prediction of frost depth can be challenging due to the high uncertainty of frost depth and the unavailability of geotechnical properties needed to use the available empirical- and analytical-based equations in literature. Therefore, this study proposed a new framework to predict the frost depth of soil below the pavement using eight machine learning (ML) algorithms (five single ML algorithms and three ensemble learning algorithms) without geotechnical properties. Among eight ML models, the hyperparameter-tuned gradient boosting model showed the best performance with the coefficient of determination (R²) = 0.919. Furthermore, it was also shown that the developed ML model can be utilized in the prediction of several levels of frost depth and assessing the sensitivity of pavement-related predictors for predicting the frost depth of soils.

The artificial ground freezing (AGF) method is an environmentally friendly ground improvement technique for numerous geotechnical applications. It can be used in fine-grained soils, which may not be efficiently improved via conventional cement-based ground improvement techniques. However, some of the issues hindering the application of the AGF method to fine-grained soils include inefficiency in achieving the target volume of frozen soil and degradation in mechanical properties of the soil after the freezing-thawing process. In this paper, the freezing rate and degradation in strength and stiffness of a clayey silt in South Korea were investigated using field experiments. At two different outlet temperatures ($-180°\mathrm{C}$ and $-120°\mathrm{C}$ ), liquid nitrogen was injected into the freezing pipe to evaluate the freezing rate. A simple equation to estimate the theoretical radial freezing rate was proposed and compared with the experimental results. In addition, a piezocone penetration test (CPTu) and pressuremeter test (PMT) were performed to assess the degradation in strength and stiffness of the soil after the freezing-thawing process. Results of the CPTu, PMT, and laboratory experiments revealed that the degradation in mechanical properties of Korean clayey silt could be attributed to the rearrangement of soil particles.

The artificial ground freezing (AGF) method has been used in many geotechnical engineering applications such as temporary excavation support, underpinning and groundwater cutoff. The AGF method utilizes a refrigerant such as liquid nitrogen or brine, circulating through embedded freezing pipes in order to freeze the ground. In this paper, two in-situ cryogenic freezing experiments (i.e., single freezing-pipe test and frozen-wall formation test) were performed using liquid nitrogen to simulate the AGF in a Korean marine clay deposit, in which the freezing rate was evaluated. The thermal conductivity of frozen and unfrozen marine clay was evaluated by performing typical laboratory experiments. In addition, the strength and stiffness of frozen-thawed deposits were comparatively measured by sounding tests (i.e., piezocone penetration test and lateral loading test). The freezing rate of the frozen-wall formation test in the Korean marine clay deposit was approximately twice as high as that of the single freezing pipe test. Compared to the original marine clay deposit, the frozen-thawed marine clay showed a significant reduction in strength and stiffness.

Rotating compressors are foundational in various industrial processes, particularly in the oil-and-gas sector, where reliable fault detection is crucial for maintaining operational continuity. While Graph Attention Network (GAT) frameworks are widely available, this study advances the state of the art by introducing a Bayesian GAT method specifically tailored for vibration-based compressor fault diagnosis. The approach integrates domain-specific digital-twin simulations built with Rotordynamic software (1.3.0), and constructs dual adjacency matrices to encode both physically informed and data-driven sensor relationships. Additionally, a hybrid forecasting-and-reconstruction objective enables the model to capture short-term deviations as well as long-term waveform fidelity. Monte Carlo dropout further decomposes prediction uncertainty into aleatoric and epistemic components, providing a more robust and interpretable model. Comparative evaluations against conventional Long Short-Term Memory (LSTM)-based autoencoder and forecasting methods demonstrate that the proposed framework achieves superior fault-detection performance across multiple fault types, including misalignment, bearing failure, and unbalance. Moreover, uncertainty analyses confirm that fault severity correlates with increasing levels of both aleatoric and epistemic uncertainty, reflecting heightened noise and reduced model confidence under more severe conditions. By enhancing GAT fundamentals with a domain-tailored dual-graph strategy, specialized Bayesian inference, and digital-twin data generation, this research delivers a comprehensive and interpretable solution for compressor fault diagnosis, paving the way for more reliable and risk-aware predictive maintenance in complex rotating machinery.

Amid the ongoing global warming crisis, there has been growing interest in hydrogen energy as an environmentally friendly energy source to achieve carbon neutrality. A stable and large-scale hydrogen storage infrastructure is essential to satisfy the increasing demand for hydrogen energy. Particularly for hydrogen refueling stations located in urban areas, technological solutions are required to ensure the stability of adjacent civil structures in the event of hydrogen storage tank explosions. In this study, a numerical analysis using equivalent trinitrotoluene (TNT) and Concrete Damage Plasticity (CDP) models was employed to analyze the dynamic behavior of the ground in response to hydrogen gas explosions in shallow underground hydrogen storage facilities and to assess the stability of nearby structures against explosion effects. According to the simulation results, it was possible to ensure the structural stability of nearby buildings and tunnel structures by maintaining a minimum separation distance. In the case of nearby building structures, a distance of at least 6 to 7 m is needed to be maintained from the underground hydrogen storage facility to prevent explosion damage from a hydrogen gas explosion. For nearby tunnel structures, a distance of at least 10 m is required to ensure structural stability.

Ensuring the operational safety and reliability of rotary machinery systems, especially in oil plants, has become a focal point in both academic and industry arenas. Specifically, in terms of key rotary machinery components such as shafts, the diagnosis of these systems is paramount for achieving enhanced generalization capabilities in fault diagnosis, encompassing multiple sensor-derived variables with their respective fault patterns. This study introduces a multi-stage approach to generalize capabilities for fault diagnosis that considers multiple sensor-derived variables and their fault patterns. This method combines the Convolutional Triplet Network for feature extraction with an ensemble model for fault classification. Initially, vibration signals are processed to yield the most representative temporal and spatial features. Then, an ensemble approach is used to maximize both diversity and accuracy by balancing the contributions of the individual classifiers. The approach can detect three representative types of shaft faults more accurately than traditional single-stage machine learning models. Comprehensive experiments, detailed within, showcase the method’s efficacy in diagnosing rotary machine faults across diverse operational scenarios.