Study the elastic properties and the anisotropy of rocks using different machine learning methods

This paper aims to demonstrate that the elastic stiffnesses and the anisotropic parameters of rocks can be accurately predicted from geophysical features such as the porosity, the density, the compression stress, the pore pressure and the burial depth using relevant machine learning methods. It also suggests that the extreme gradient boosting method is the best method for this purpose. It is more accurate, extremely faster to train and more robust than the artificial neural networks and the support vector machine methods. Very high R-squared scores was obtained for the predicted elastic stiffnesses of a relevant dataset that is available in the literature. This dataset contains different types of rocks, and the values of the features are in large ranges. An optimal set of parameters was obtained by considering an appropriate sensitivity analysis. The optimized model is very easy to implement in Python for practical applications.     

Semi-Analytical Solution for Stresses and Displacements in a Tunnel Excavated in Transversely Isotropic Formation with Non-Linear Behavior

A semi-analytical solution based on the transfer matrix technique is proposed to analyze the stresses and displacements in a two-dimensional circular opening excavated in transversely isotropic formation with non-linear behavior. A non-isotropic far field can be accounted for and the process of excavation is simulated by progressive reduction of the internal radial stress. A hyperbolic stress–strain law is proposed to take into account the non-linear behavior of the rock. The model contains seven independent parameters corresponding to the five elastic constants of an elastic material with transverse isotropy and to the friction coefficient and cohesion along the parallel joints (weakness planes). This approach is based on the discretization of the space into concentric rings. It requires the establishment of elementary solutions corresponding to the stress and displacement fields inside each ring for given conditions at its boundaries. These solutions, based on complex variable theory, are obtained in the form of infinite series. The appropriate number of terms to be kept for acceptable approximation is discussed. This non-linear model is applied to back analyze the convergence measurements of Saint-Martin-la-Porte access gallery. Short-term and long-term ground parameters are evaluated.     

Effect of saline intrusion on the properties of cohesive soils in the Red River Delta,Vietnam

Abstract

This study conducted a series of laboratory experiments and established numerical models on selected undisturbed soil samples in the Red River Delta (RRD) to determine the effect of change in soils intruded by saline water. The variation in the technical parameters of soils was verified in soils fully saturated by solution of four salt concentrations, that is, 0.0, 9.9, 19.8, and 33.0?g/L. Results show that the content and composition of clay minerals in cohesive soils before and after saline intrusion are unchanged. The same finding is obtained for clay after removing absorbed water layer by using a centrifuge apparatus. The zeta potential and settlement velocity of soils in the RRD increase when salt is added to the saturated solution. Similarly, the deformation of soils increases proportionally with the salt concentrations of that solution. This result is attributed to the linear decrease in deformation modulus. The decrease in modulus versus salinities is nearly consistent for pressure stages from 100 to 400?kPa. The safety factor of bearing capacity also decreases linearly with salinities. The decrease reaches 12.5–16.3% when soils are in the maximum saline solution. All these changes are considered as the degradation of soils in saline media.