Multiphysics uncertainty decoupling methods for stochastic poroelastic analysis

• Efficient multiphysics uncertainty decoupling methods are proposed. • Multiphysics stochastic solutions are decoupled into solutions for each physical field. • Stochastic solutions are approximated and solved using dedicated algorithms. • Linear and nonlinear stochastic poroelastic problems are analyzed. • High-dimensional stochastic multiphysics analysis can be performed efficiently. This article presents an efficient multiphysics framework for linear and nonlinear stochastic poroelastic analysis. Specifically, an uncertainty decoupling method is developed to decouple the overall multiphysics stochastic solution into stochastic solutions of each physical field. Stochastic solutions of each physical field are approximated by a unified representation and expressed as the summation of a series of products of triplets of spatial functions, temporal functions and random variables. Each triplet is then solved in a greedy way using a dedicated iteration, which transforms the original stochastic problems into deterministic problems and very low-dimensional stochastic problems that can be solved almost independently of stochastic dimensions. Compared with existing coupled stochastic solution approximations, the proposed uncertainty decoupling method has better convergence without loss of accuracy, thus saving computational efforts. Numerical examples of linear and nonlinear stochastic poroelasticity with high-dimensional random inputs demonstrate the promising performance of the proposed framework.