In "Anchored Physics-Informed Neural Network for Two-Phase Flow Simulation in Heterogeneous Porous Media" By Jingqi Lin, Xia Yan, Kai Zhang, Zhao Zhang, and Jun Yao, the authors introduce Anchored-EPINN, a new approach that integrates an adaptive tensorization mechanism with a novel anchoring strategy. This “tensorize-then-anchor” design expands learning capacity while reducing computational costs, achieving ~30% faster runtime for NN-based two-phase flow simulation in heterogeneous porous media — all while preserving accuracy. The method demonstrates a promising path for faster and more efficient simulations.

👉 Read the full article in InterPore Journal: https://doi.org/10.69631/ipj.v2i3nr67