Numerical Modelling and Simulation of Seepage in Tight Sandstone Based on Parallel Lattice Boltzmann Method

TANG Mingming,LU Shuangfang,LIANG Hongru,YAN Bihui,MA Huifang,SHEN Shan

DOI: https://doi.org/10.1111/1755-6724.12307_2

2015-01-01

Abstract:Lattice Boltzmann Method (LBM) have a remarkable ability to simulate single, multiphase and multicomponent fluids (Michael & Daniel, 2007). Ludwig Boltzmann’s work is based on a simple idea that a gas is composed of interacting particles which could be analysed by Newton’s laws. So the flowing property of gas could be give out by just considering the particles streaming in space and collision. Although, Boltzmann’s work focused on gas, but with a further simplify, still gave out the behaviour of general real fluids. Usually, lattice gas cellular automation models (CAM) were considered as the harbingers of LBM. CAM is an algorithmic that fluid point can only move along and occupied on a grid or lattice point in space, and interacts with its neighbours following particularly evolution rules. The rules of imposed on the group of cellular automata uniquely, and generally examines the current lattice state and the state of number of its neighbours. Wolfram (Wolfram, 1986) presented a very detailed classification and description of CAM. But CAM only treat the lattice state as 0 or 1, depending on whether there is a particle on the current lattice. Which causes lots simulation instability and errors of real fluids. The LBM (Qian.et al, 1992) describes the lattice state as continues distribution function, instead of discrete binary state. Then streaming and collision of the particles via the distribution function. In this paper, a nono-scale numerical model of tight sandstone is built, then the seepage is calculated out based on LBM. Due to the large simulation cost, we developed a parallel LBM algorithm on high performance cluster.