基于BP神经网络的苏里格气田裂缝强制闭合模型分析

在压裂施工过程中,为了减少压裂液对储层的伤害,压裂液应当被及时排出地层。结合物质平衡原理、渗流力学理论和二维滤失模型,建立了裂缝强制闭合模型,并编制了相应的计算程序,根据井口压力变化规律并结合裂缝闭合条件可得到不同渗透率、裂缝高度和裂缝半长对应的裂缝强制闭合时间。同时,影响裂缝闭合时间的因素权重通过BP神经网络结合理论模型的数据进行了研究。结果显示,渗透率、裂缝高度、返排液粘度、裂缝半长、井筒半径和油嘴半径这六个参数是影响裂缝闭合时间的主要因素,并且在这些因素中,渗透率和油嘴半径的影响权重最大。因此,在压裂液返排过程中,为了实现改善压裂效果的目的,要着重考虑油嘴半径的选择。本文结果为裂缝闭合时间的研究提供了可靠的理论基础。

Model analysis of forced crack closure in the Sulige gas fieldbased on BP neural network

In order to reduce the adverse effect of fracturing on reservoirs, fracturing fluid should be discharged from strata timely. Based on material balance theory, seepage mechanics and a two-dimensional leak-off model, a mathematical model of forced crack closure is established and a corresponding calculation program iscompiled. It permits to determine crack forced close time according to the law of wellhead pressure and fracture close conditionsfor different permeability, crack heights and crack half-lengths. Meanwhile, the weight of the influence of different factors on the forced crack closure time is analyzed using BP neural network coupling with data from the model. The results show that the permeability, crack height, viscosity, crack half-length, wellbore radius and choke radius are the main factors affecting the time of crack closure. Among them, permeability and choke radius have the largest weights of influence.Therefore, during flow back of fracturing fluid, in order to optimize the fracturing, we should consider the choice of choke radius. The results of this workwould be helpful to further studying the time of fracture closure.