Permeability characterises flow in porous rocks/media for upscaling, while steady-state flow fields allow analysis of reactive transport, fines migration, and tight unconventional rocks. Fast calculation of permeability and flow fields obtained from Pore Network Models (PNM) and Laplace Semi-Analytical Solvers (SAS) deviate from computationally demanding simulation of Navier Stokes Equations (NSE) due to flow and geometry simplifications. Coupling PNM/SAS with direct simulation via Lattice Boltzmann Method (LBM) provides 5--$150\times$ speed-up without accuracy loss over 100 samples from 0.7mD to 3.5D. Permeability errors in PNM/SAS show 10-20% (up to 50-70%) error. PNM shows higher variance from geometric simplifications compared to SAS which only makes flow-based assumptions. PNM/SAS errors are eliminated by coupling with LBM at a fraction of LBM-only compute cost. Steady-state conditions with PNM/SAS-LBM are reached in $<2,000$ timesteps, compared to LBM-only which can require $>10$⁵ timesteps in tight domains such as cemented sandstone, carbonate, coal, and shale.

渗透率表征了多孔岩石/介质中的流动以进行放大，而稳态流场则允许分析反应性运移，细粒运移和致密的非常规岩石。由于简化了流动和几何形状，因此从孔隙网络模型（PNM）和拉普拉斯半解析求解器（SAS）获得的渗透率和流场的快速计算与对Navier Stokes方程（NSE）的计算要求较高的仿真有所不同。通过莱迪思·玻耳兹曼方法（LBM）将PNM / SAS与直接仿真耦合，可以提供5--$150\times$从0.7mD到3.5D可以进行100多个采样，而不会造成精度损失。PNM / SAS中的渗透率错误显示10-20％（最多50-70％）的错误。与仅基于流的假设的SAS相比，PNM在几何简化方面显示出更高的方差。通过与LBM耦合，消除了PNM / SAS错误，而仅花费LBM的一部分计算成本。在PNM / SAS-LBM中达到稳态条件$<\mathrm{2个}，000$ 的时间步伐，与可能只需要LBM的情况相比 $>10$在水泥砂岩，碳酸盐，煤炭和页岩等紧密领域中的^5个时间步长。