利用神经网络和相渗曲线预测储层的润湿性

Wettability prediction of reservoirs using neural network and relative permeability curves

润湿性是控制储层中流体位置、流动和空间分布的主要因素 ,因此对相对渗透率曲线有重要影响。从研究润湿性在相对渗透率曲线上的特征表现出发 ,利用反向传播神经网络模型预测储层的相对渗透率曲线 ,由得到的相对渗透率曲线上的 6个特征量 ,依据在Craig准则基础上制定的判定储层岩石润湿性的三饱和度准则 ,预测储层的润湿性 ,并与实验室AMOTT方法测量的润湿性结果对比 ,证明该方法是有效的 ,可用于预测储层的润湿性。

Wettability affects relative permeability curves because it is a major factor controlling the location, flow, and spatial distribution of fluids in the reservoirs. Relative wettability curves were calculated using backpropagation neural network model in this paper. From the resulted curves, six characteristic quantities were calculated and used to predict the wettability of reservoirs. Three-saturation rule that is based on Craig's rules of thumbs was adopted in the prediction. The feasibility of the method was confirmed by the comparison between the predicted results and the laboratory measurements obtained from AMOTT method.