Abstract: Origin mechanism of multiscale percolation phenomena of porous flow in reservoir is discussed,and the growth rule of displacing front is uncovered in the view of percolation.Percolation rule is determined by pressure field calculated by implicit method according to dynamic equation of macroscopic and microscopic porous flow,then in this paper,macroscopic and microscopic percolation models for reservoir numerical simulation are constructed separately.Development of fluid fingering and trap of remaining oil in the models are not only be affected by heterogeneity of porous media,but also correlated with displacing history variation of pressure field,which is the modification of percolation simulation models for porous flow constructed by Stenby (1990) and Barrefet (1994).The models in this paper can reflect the dynamic universality and irregularity of porous flow.The model for calculation of pressure drop in grids adjacent to outlet after water breakthrough is constructed also,which makes the reservoir percolation simulation models can not only simulate the water displacing oil process before water breakthrough,but also can predict the dynamics after water breakthrough.Then one of the technical problems in application of percolation network to simulation of reservoir dynamics has been solved.Testified by flooging experiment in real cores and plane sand packing experiment,the multiscale percolation models for reservoir numerical simulation constructed in this paper are reliable.Microscopic percolation simulation can be applied to evaluate the influence of porous structure and pressure difference on recovery,fingering development and distribution of remaining oil,while macroscopic percolation network simulation can simulate porous flow in reservoirs with heterogeneous distribution of permeability and porosity,which can make more accurate prediction of distribution of fingering and remaining oil.Characteristics curves by macroscopic and microscopic percolation methods can be correlated with the corresponding results from physical and classical simulation methods,which deepens and broadens reservoir numerical simulation technique.

Key words: percolation, fingering, reservoir, numerical simularion, recovery

摘要： 指出油藏渗流中不同尺度的渗滤现象形成机制,从渗滤理论角度揭示驱替前缘的生长规律.根据宏观与微观渗流的动力学方程,以隐式计算的压力场确定渗滤准则,分别建立了微观与宏观油藏数值模拟的渗滤模型.该模型中指进的发育、剩余油的圈闭不仅受介质非均质性的影响,而且与驱替历史(压力场变化)相关,是对Stenby(1990)及Barefet(1994)的渗滤模拟模型理论上的修正与完善,更能体现渗流的动力学统一性及不规则性.此外,建立了排液道见水后与排液道相临节点的压力降计算模型,使油藏渗滤模型不仅可模拟见水前的水驱油过程,还可对见水后的动态进行预测,突破了网络模拟向油藏动态模拟应用转化的一个难题.通过实际岩心驱替实验及平面填砂模型实验的验证,所建立的多尺度油藏数值渗滤模拟模型是可靠的.微观渗滤模拟可以评价孔隙结构对采收率、指进发育、剩余油分布的影响.宏观渗滤网络模拟方法,可对渗透率、孔隙度非均质分布的油藏在不同驱替条件下的渗流进行渗滤模拟,使指进、剩余油分布的预测更为客观.宏观与微观渗滤模拟特征曲线可与物理模拟及经典数值模拟方法的结果进行对比研究,使油藏数值模拟的深度与广度都得以延伸.

关键词: 渗滤, 指进, 油藏, 数值模拟, 采收率