Our SFB 1313 doctoral researcher Hanchuan Wu (research project A02) successfully defended his doctoral thesis "Dimensionally reduced model concepts for the simulation of multi-phase flow and transport processes in porous media - advanced pore-network and embedded tubular network models" on 13 March 2026. Congratulations!
Hanchuan Wu completed his doctoral thesis in the framework of the SFB 1313 Intrgrated Research Training Group "Interface Driven Multi-Field Processes in Porous Media" (IRTG-IMPM).
Abstract
For many applications involving flow and transport in porous media, fully resolved three-dimensional simulations are computationally expensive. Model reduction techniques therefore play a crucial role by enabling efficient lower-dimensional representations, which form the focus of this work. While such reduced models significantly decrease computational costs, the associated physical and geometrical simplifications introduce new numerical and modeling challenges. In this work, two representative scenarios are investigated.
The first scenario considers embedded network models for describing exchange processes between lower-dimensional structures and a surrounding bulk domain. Typical applications include blood vessel networks in biological tissue, root systems in soil, and geothermal wells in the subsurface. By representing these structures as one-dimensional networks, the computational efficiency is greatly improved; however, modeling mass exchange via line sources introduces mathematical singularities. This work addresses these challenges and develops robust formulations for accurately capturing exchange processes in embedded tubular network systems.
The second scenario focuses on pore-network models as an efficient representation of the void space in porous media. Pore-network models resolve multiphase flow and transport on a simplified yet physically representative geometry, enabling monolithic coupling across domain interfaces. However, discontinuities in pore-throat conductivity relations introduce significant numerical challenges for nonlinear solvers under implicit time discretization. To address these issues, a generalized flux formulation is proposed to enhance the convergence and robustness. Within this framework, two solution strategies are developed: a regularization-based approach and an additional-variable formulation. The proposed methods demonstrate improved accuracy and convergence behavior, providing a reliable framework for multiphase pore-network simulations.