.
Groundwater extraction induces pore-pressure variations and land subsidence, posing significant risks to critical infrastructure. However, accurately isolating extraction-induced signals from pore-pressure measurements that capture changes in the groundwater level of specific aquifers remains challenging, as environmental noise from rainfall, humidity, and seasonal fluctuations tends to degrade the signal-to-noise ratio. This study introduces a novel gray-box approach that combines mechanics-based modeling with machine learning techniques for extracting meaningful pore pressure signals induced by groundwater extraction. By leveraging domain-specific knowledge of poromechanics, the framework systematically informs hyperparameter selection. A coupled poromechanics model simulates soil responses under controlled groundwater extraction conditions, providing a reliable reference for denoising field data obtained from pumping experiments conducted in Taiwan. The effectiveness of the proposed framework is evaluated through integration with Autoregressive (AR) models, Gaussian Process Regression (GPR), and Support Vector Regression (SVR), demonstrating substantial improvements in isolating meaningful signals from noisy pore-pressure measurements.
Taiwan’s open-pit mines are frequently exposed to strong seismic shaking, yet rapid post-earthquake loss assessment for mining equipment remains limited. This study develops a Rapid Post-Earthquake Loss Assessment System for open-pit mines in Taiwan, integrating seismic demand estimation, equipment-level fragility modeling, and a pre-simulated earthquake scenario database. Based on field investigations at four representative mines, critical equipment and operational components are identified and evaluated using fragility functions. Once earthquake information is received from the Central Weather Administration, the system automatically matches the event to precomputed scenarios and estimates damage probabilities for monitored equipment. The results are classified into low, moderate, and high risk levels and visualized through an intuitive decision-support interface. By providing rapid, site-specific, and equipment-level risk information, the proposed framework enhances situational awareness, supports inspection prioritization, and facilitates timely emergency response in mining facilities. The system demonstrates a practical approach for applying engineering seismology and fragility-based assessment to industrial seismic risk management.
In this article, we propose a novel real-time terrain recognition and slip estimation method for quadruped robots using proprioceptive sensors and temporal convolutional networks (TCNs). As quadruped robots are increasingly deployed in complex environments, accurate terrain understanding is crucial. External sensors can be affected by lighting variations, occlusion, reflective surfaces, and others. To overcome these challenges, we propose a proprioceptive sensing-based complementary perception module with a TCN, enabling reliable real-time terrain recognition while reducing dependence on external perception. The TCN model effectively captures temporal dependencies in sensor signals, enabling precise and robust detection. The framework is validated through extensive real-world experiments and deployed on an embedded edge computing platform for real-time operation. Results show that the proposed TCN method achieves 98.8% recognition accuracy, outperforming the baseline models compared in this study. In addition, this study analyzes how locomotion speed and environmental conditions affect slip in quadruped robots. These findings confirm that quadruped robots can not only recognize terrain types but also detect surface states, enabling safer and more adaptive locomotion. Therefore, the proposed system is a cost-effective, robust, and low-latency solution for real-time terrain recognition, providing a strong foundation for future deployment across more diverse terrains.
ABSTRACT Flood inundation mapping has become a critical reference for flood risk management, particularly for determining insurance premiums and formulating mitigation strategies in rapidly urbanizing areas with high population density. With recent advancements in mapping technologies and high computing capacity, high‐resolution spatiotemporal datasets can be integrated into flood simulation models. In this study, a hydrological digital elevation model (HyDEM) was implemented at the city scale for Kaohsiung City, Taiwan, to improve urban flood simulations. The HyDEM incorporated specialized datasets, including high‐resolution DEM, building layer, bankline layer, and seadike layer. Using the Delft3D FM 1D–2D modeling platform, 10 basin‐scale models were constructed across Kaohsiung City, and multiresolution computational meshes were employed to achieve a balance between simulation accuracy and computational efficiency. Model performance was validated using two historical flood events, during which it achieved an overall accuracy of ~84% for the June 5 Rainstorm event and 80% for the Typhoon Gaemi (2024) event. These results indicate that flood inundation maps (FIMs) can be considerably improved by preserving critical topographic features, particularly building footprints and hydraulic structures such as river and sea dikes. The proposed method thus addresses limitations of earlier generations of FIMs and provides a framework for advancing next‐generation flood inundation maps.
In response to the rapid deployment of energy storage systems (ESS) to support renewable energy integration and ensure grid stability, this study establishes a project certification framework specifically for outdoor battery energy storage system (BESS) sites. Since 2024, the framework has been applied to newly installed projects, and by June 2025, 30 projects had completed both design and on-site reviews and obtained certification, while 23 projects had completed only the design review. For existing sites, 70 projects had successfully passed project specification and site reviews within the same period. The framework emphasizes safety through mandatory design reviews, requiring electrical and fire safety approvals by licensed engineers. To mitigate project delays and financial burdens on developers, the competent authority allows submission of design approval documents prior to the final acceptance test, enabling earlier project initiation. The results demonstrate that the proposed certification system effectively enhances the safety and reliability of outdoor BESS projects while addressing the practical needs of stakeholders, offering valuable experience for the broader application of ESS project certification systems.
