Data for: "Synthetic dataset of pore scale multiphase flow from direct numerical simulations"

This dataset provides a comprehensive collection of high-fidelity, pore-scale direct numerical simulations examining multiphase flow in porous media. The simulations systematically investigate the evolution of fluid invasion patterns as functions of pore structure heterogeneity, viscosity ratio between displacing and displaced fluids, and the wettability of the porous matrix. The dataset includes 540 individual simulation cases conducted using the OpenFOAM software. For each case, OpenFOAM input and output files are provided. The output files are provided for every simulated second and just before the invading fluid reaches the outlet, allowing detailed temporal analysis of flow governing variables such as pressure and velocity. In addition, a single visualization snapshot is included for each case, highlighting the invasion profile immediately prior to breakthrough. This dataset enables in-depth analysis of multiphase displacement phenomena, serving as a valuable resource for studying the effects of structural heterogeneity, viscosity contrast, and wettability on fluid invasion at the pore-scale. It is ideally suited for validation of machine-learning surrogate models, and also to benchmark the flows predicted other computational models like pore network models.本数据集收录了一套全面的高保真孔隙尺度（pore-scale）直接数值模拟结果，用于研究多孔介质（porous media）内的多相流动现象。该系列模拟系统地探究了流体驱替模式的演化规律，其影响因素涵盖孔隙结构非均质性、驱替流体与被驱替流体间的粘度比，以及多孔基质的润湿性。本数据集包含540组采用OpenFOAM软件开展的独立模拟算例，每组算例均提供对应的OpenFOAM输入与输出文件。输出文件按每模拟秒间隔生成，并在侵入流体抵达出口前额外提供一组，可支持对压力、流速等流动控制变量开展精细化的时域分析。此外，每组算例均附带一张可视化快照，用以展示突破瞬间前的流体驱替剖面特征。本数据集可支撑多相驱替现象的深入分析，是研究孔隙尺度下结构非均质性、粘度差异与润湿性对流体驱替影响的宝贵资源。其非常适用于机器学习代理模型（machine-learning surrogate models）的验证，同时也可作为孔隙网络模型（pore network models）等其他计算模型所预测流动的基准测试数据集。