Ground subsidence resulting from underground coal mining poses significant challenges to urban safety, infrastructure stability, and environmental protection, particularly in regions extending beneath water bodies. This study investigates subsidence patterns in the Kozlu coal basin by integrating Interferometric Synthetic Aperture Radar (InSAR), numerical modelling, and machine learning techniques. The Kozlu coal basin, located in Zonguldak, Turkey, serves as a critical example, where extensive mining activities have led to complex deformation patterns. InSAR effectively captures terrestrial subsidence but is limited in underwater regions. Numerical modelling provides insights into geological behaviour but requires extensive input data. Machine learning, specifically Gaussian Process Regression (GPR), bridges this gap by predicting subsidence in unobservable underwater zones with high accuracy. The integrated approach reveals consistent deformation trends across terrestrial and marine environments, offering practical tools for risk mitigation and resource management. These findings underscore the importance of interdisciplinary methods in addressing complex geological challenges and pave the way for future advancements in subsidence monitoring and prediction.

As urbanization gathers pace, projects involving adjacent subway tunnels are increasing, thereby amplifying the need for robust tunnel protection measures. Currently, there exists a notable lack of precise analyses on the three-dimensional (3D) deformation laws and mechanisms of tunnels affected by adjacent deep excavation. Moreover, the influence patterns of retaining wall stiffness and deep excavation depth on the 3D deformation of pit-side tunnels remain unclear. The purpose of this research is to bridge the existing disparity by adopting the hypoplastic model, which effectively captures soil stiffness that is dependent on soil state, strain, and stress path, even at small strains, as well as soil strength, based on reported centrifuge model tests. This approach facilitates a comprehensive, precise numerical analysis of the interaction between deep excavation and preexisting tunnels located outside the retaining wall. The analysis sheds light on the deformation mechanisms and trends of pit-side tunnels not solely confined to the longitudinal axis but extending to the transverse plane as well, while systematically examining the influence of varying excavation depths and retaining wall stiffness on key tunnel parameters, including longitudinal deformation, diameter changes, bending strains, and soil pressure distributions around the tunnels. The study reveals that if the tunnel situated outside the retaining structure lies beneath the foundation pit's base, deep excavation only slightly deforms the tunnel. However, when the tunnel outside the retaining structure is positioned above the pit's base, its deformation progressively intensifies with deeper excavation, but the growth rate has a decreasing trend. An enhancement in the stiffness of the retaining wall results in a notable decrease in the deformation exhibited by the adjacent tunnels. The findings contribute to a deeper understanding of the complex responses of pit-side tunnels to excavation activities, ultimately facilitating the design and construction of safer and more resilient urban subway infrastructure.

Site response analyses at large strains are routinely carried out neglecting the shear strength of soil and the stiffness degradation due to the increase in pore pressures, leading to unrealistic predictions of the seismic response of soil deposits. The study investigates the performance of a simplified nonlinear (NL) approach, implemented in the Deepsoil code, constituted by coupling a hyperbolic model incorporating shear strength with a strain-based semi-empirical pore pressure generation model. The first part of the study, based on a large one-dimensional parametric study, shows that above a shear strain of 0.1%, it is necessary to include shear strength in the site response modelling to get more realistic results. Then, the approach has been evaluated with reference to the well-known downhole Large-Scale Seismic Test array located in Lotung (Taiwan): numerical results have been compared with recordings in terms of acceleration response spectra and pore water pressure time histories at different depths along the soil profiles. The comparison shows that the NL simplified model is characterized by an accuracy comparable with more sophisticated advanced elasto-plastic NL analyses adopting essentially the same input data of the traditional equivalent linear approaches(shear modulus and damping curves) and simple physical-mechanical properties routinely determined during geotechnical surveys (i.e., shear strength, relative density, fine content). This approach is therefore recommended for site response analyses reaching large strains (i.e., soft soil deposits and moderate-to-high input motions).

Proper evaluation of seismic-induced excess pore water pressures in saturated sandy soils is still an open issue, which can be tackled with fully coupled to uncoupled approaches. The former are more accurate but computationally onerous, while the latter require seismic demand and pore pressure build-up to be computed in two successive steps, typically employing simple constitutive assumptions. Starting from the work by Seed et al. (1975), this paper presents a novel uncoupled procedure to compute excess water pressures developing in a 1D soil column under partially drained conditions, when subjected to horizontal seismic excitations. Fundamental modifications are introduced to account for: non-uniform distribution of equivalent loading cycles; soil stiffness degradation; and modification of the frequency content of ground motion due to pore pressure build-up. The approach was implemented in Matlab via the Finite Difference Method and validated against both fullycoupled Finite Element analyses and one centrifuge test. An extensive parametric study was also performed for a two-layer soil column, by varying the thickness and hydraulic conductivity of the shallow layer, as well as the seismic input. The good agreement with both numerical and experimental data demonstrates that key features of liquefaction are well-captured by the proposed uncoupled approach.

The paper presents the results of 3D coupled cyclic time history numerical analyses of a monopile supporting a 12 MW Offshore Wind Turbine, installed in dense cohesionless soils and subjected to a 600-s load history corresponding to the high phase of a 35-h design storm. The goal of the study is to investigate the governing mechanisms and gauge potential conservatisms or uncertainties in approaches for monopile analysis used in practice. The Ta-Ger constitutive model, implemented in FLAC3D and calibrated against site-specific cyclic tests, is used to model the complex soil response. Emphasis is placed on the effect of drainage conditions, an aspect typically overlooked in practice, although often stated as critical. Analyses show that the drainage of the system can substantially affect the response. In low-permeability soils (e.g., cohesionless soils with low-plasticity fines) widespread liquefaction may occur inducing high rotations above allowable limits. On the contrary, systems that can drain effectively within each cycle, develop moderate excess pore pressures which do not jeopardize performance. Current design procedures are often unable to accurately capture these effects possibly leading to either conservative or unconservative outcomes. Suitably validated advanced numerical analyses can be used as complementary tools to standard methods to assess these uncertainties.