No abstract

No abstract

In this article, a computational model is presented for the analysis of coupled thermo-hydro-mechanical process with phase change (evaporation/condensation) in fractured porous media in order to model multiphase fluid flows, heat transfer, and discontinuous deformation by employing the extended finite element method. The ideal gas law and Dalton's law are employed to consider vapor and dry air as miscible gases. To take into account the phase change, latent heat and specific vapor enthalpy are incorporated into the physical model. The set of governing equations consists of linear momentum for the solid-phase, energy balance equation and mass conservation equations of water species (liquid and vapor) and dry air, which are derived within the framework of the generalized Biot's theory. The weak forms are presented in terms of the displacement, water pressure, capillary pressure, and temperature as the primary variables. The spatial and temporal discretizations are carried out applying the extended finite element method and the generalized Newmark scheme, respectively, resulting in the final system of fully coupled nonlinear equations. Finally, several numerical examples are solved to demonstrate not only the robustness of the proposed computational model but also the effect of the crack orientation, intrinsic permeability, and elastic modulus on the fluid flow patterns, relative humidity, and temperature distribution. Moreover, an edge-crack propagation is investigated under mode I fracture due to drying conditions.

Physics-Informed Neural Networks (PINNs) have received increased interest for forward, inverse, and surrogate modeling of problems described by partial differential equations (PDE). However, their application to multiphysics problem, governed by several coupled PDEs, present unique challenges that have hindered the robustness and widespread applicability of this approach. Here we investigate the application of PINNs to the forward solution of problems involving thermohydro-mechanical (THM) processes in porous media, which exhibit disparate spatial and temporal scales in thermal conductivity, hydraulic permeability, and elasticity. In addition, PINNs are faced with the challenges of the multi-objective and non-convex nature of the optimization problem. To address these fundamental issues, we: (1) rewrite the THM governing equations in dimensionless form that is best suited for deep-learning algorithms; (2) propose a sequential training strategy that circumvents the need for a simultaneous solution of the multiphysics problem and facilitates the task of optimizers in the solution search; and (3) leverage adaptive weight strategies to overcome the stiffness in the gradient flow of the multi-objective optimization problem. Finally, we apply this framework to the solution of several synthetic problems in 1D and 2D.