This study investigates the application of machine learning (ML) algorithms for seismic damage classification of bridges supported by helical pile foundations in cohesive soils. While ML techniques have shown strong potential in seismic risk modeling, most prior research has focused on regression tasks or damage classification of overall bridge systems. The unique seismic behavior of foundation elements, particularly helical piles, remains unexplored. In this study, numerical data derived from finite element simulations are used to classify damage states for three key metrics: piers' drift, piles' ductility factor, and piles' settlement ratio. Several ML algorithms, including CatBoost, LightGBM, Random Forest, and traditional classifiers, are evaluated under original, oversampled, and undersampled datasets. Results show that CatBoost and LightGBM outperform other methods in accuracy and robustness, particularly under imbalanced data conditions. Oversampling improves classification for specific targets but introduces overfitting risks in others, while undersampling generally degrades model performance. This work addresses a significant gap in bridge risk assessment by combining advanced ML methods with a specialized foundation type, contributing to improved post-earthquake damage evaluation and infrastructure resilience.

The cone penetration tests have been employed extensively in both onshore and offshore site investigations to obtain the strength properties of soils. Interpretation of effective internal friction angle gyp' becomes complicated for cones in silty clays or clayey silts, since the soil around the advancing cone may be under partially drained conditions. Although there exist several robust methods to estimate gyp ' , the pore pressure at the cone shoulder has to be measured to represent the drainage conditions. Many cone penetrometers in practice are not equipped with a pore pressure transducer. Even for a piezocone, the pore pressure recorded in-situ may be unreliable due to the poorly saturated or clogged filter. These limitations prohibit the application of existing methods. Large deformation finite element analyses were carried out within the formula of effective stress to reproduce the cone penetrations under various drainage conditions. The numerical approach was validated against the existing model tests in centrifuge and chamber, with wide ranges of penetration rates and soil types. A backbone curve is proposed to estimate the normalized cone resistance varying with the normalized penetration rate. Based on the backbone curve, a procedure is developed to predict gyp' of cohesive soils under undrained or partially drained conditions, replacing the pore pressure with the normalized penetration rate. The procedure can be used for soils with an overconsolidation ratio no larger than 5.

The water jet trenching technique is widely used in the burial of submarine pipelines. However, its application in cohesive soils often leads to complexity in trench morphology and challenges in predicting trench dimensions due to unclear soil erosion mechanisms. These issues significantly impact pipeline burials. To investigate the soil erosion mechanism of water jet trenching in cohesive soils, two-dimensional physical simulation experiments of submerged vertical water jet erosion were conducted. The influence of jet pressure, impingement height, and nozzle diameter on the shape of the scour hole was analyzed. The erosion damage patterns of water jets on cohesive soils were studied, and a theoretical model for the development of scour holes was established. The study revealed that when the jet velocity reaches 1000 m/s and the nozzle diameter reaches 1 mm, a contracted neck forms at the upper part of the scour hole. The appearance of the contracted neck is due to excessive jet impact energy causing impact shear failure in the soil. The effective height and width of the contracted neck increase with jet pressure and nozzle diameter and decrease with impingement height. Based on Prandtl's bearing capacity model, a model for predicting impact shear failure in cohesive soils was established, and a predictive formula for the effective height and diameter of the neck was proposed. Experimental validation confirmed the accuracy of the predictive formula. These findings provide theoretical support for the application of water jet trenching techniques in cohesive seabed soils.

Most structures supporting solar panels are found on thin-walled metal piles partially driven into the ground, optimizing costs and construction time. These pile foundations are subjected to repetitive lateral loads from various external forces, such as wind, which can compromise the integrity of the pile-soil system. Given that the expected operational lifespan of photovoltaic solar plants is generally 20-30 years, predicting their service life under fatigue loads is crucial. This research analyzes the response of H- piles to lateral fatigue loads in cohesive rigid soils through four field tests, subjected to load cycles of 55%, 72%, and 77% of the static failure load, corresponding to maximum loads of 25 kN, 32 kN, and 35 kN, respectively. Additionally, the effect of load cycles on the degradation of pile-soil adhesion is studied through two pull-out tests following cyclic tests. This study reveals that soil fatigue does not occur under repetitive loads and that soil stiffness remains constant once the cycles causing soil compaction have been overcome. Nevertheless, the accumulated plastic deflection of the soil increases steadily once soil compaction occurs due to cyclic loading. The implications of these results on the fatigue life of photovoltaic solar panel foundations are discussed.