基于微观孔隙结构参数的水驱采收率预测方法

A prediction method of waterflood recovery factor based on microscopic pore structure parameters

储层微观孔隙结构参数是影响水驱开发效果的重要因素,二者之间是一种非线性、不确定的复杂关系。用相关性分析优选出对水驱采收率影响较大的参数,分别采用多元回归分析和BP神经网络的方法对水驱采收率进行预测。结果表明:人工神经网络方法具有更好的自适应性,能较好地反映影响水驱效果的各种微观参数与采收率的内在联系,而且预测精度较高。应用BP神经网络方法预测水驱采收率是可行和有效的。

Microscopic pore structure parameters are important factors influencing water flooding response,and are in non-linear,uncertain,complex relationship with the latter.Parameters having big influence on waterflood recovery factor have been optimized through correlation analysis.Waterflood recovery factor is predicted by using multiple regression analysis and BP neural network respectively.The result indicates that the artificial neural network method is more adaptable,can better reflect the intrinsic relationship between various microscopic parameters and waterflood recovery factor,and has high prediction accuracy.BP neural network method is feasible and effective for predicting waterflood recovery factor.