U-FNO—An enhanced Fourier neural operator-based deep-learning model for multiphase flow

Numerical simulation of multiphase flow in porous media is essential for many geoscience applications. Machine learning models trained with numerical data can provide a faster alternative to traditional simulators. Here we present U-FNO, novel neural network architecture solving problems superior accuracy, speed, and efficiency. U-FNO designed based on the newly proposed Fourier operator (FNO), which has shown excellent performance single-phase flows. We extend FNO-based highly complex CO2-water problem wide ranges permeability porosity heterogeneity, anisotropy, reservoir conditions, injection configurations, rates, properties. The more accurate gas saturation pressure buildup predictions than original FNO state-of-the-art convolutional (CNN) benchmark. Meanwhile, it utilization efficiency, requiring only third training achieve equivalent accuracy as CNN. provides heterogeneous geological formations critically important applications such “fronts” determination. model serve general-purpose routine simulations 2D-radial CO2 significant speed-ups