Physics-Informed Neural Network Solution of Thermo–Hydro–Mechanical Processes in Porous Media

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 several coupled PDEs, present unique challenges that hindered the robustness widespread applicability this approach. Here we investigate PINNs forward solution involving thermo-hydro-mechanical (THM) processes in porous media, which exhibit disparate spatial temporal scales thermal conductivity, hydraulic permeability, elasticity. In addition, are faced with multi-objective non-convex nature optimization problem. To address these fundamental issues, we: (1)~rewrite THM governing dimensionless form is best suited deep-learning algorithms; (2)~propose a sequential training strategy circumvents need simultaneous problem facilitates task optimizers search; (3)~leverage adaptive weight strategies overcome stiffness gradient flow Finally, apply framework synthetic 1D and~2D.