BP神经网络在剩余油分布预测中的应用研究

BP Artificial Neural Networks for the Remaining Oil Prediction

我国多数油田经过一次、二次采油后,仅能采出地下总储量的30%左右,这意味着有60%～70%的剩余石油仍然残留在地下成为剩余油。加强剩余油分布规律研究、提高石油采收率不仅有着可观的经济效应,而且关系到国家石油战略的安全。本研究应用神经网络的原理,基于BP网络使用MATLAB语言建立一个剩余油分布的预测系统。该系统通过学习在地理坐标和孔隙度之间建立一个非线性函数关系,以此来预测任何区域的孔隙度,再通过孔隙度与剩余油饱和度之间的关系达到剩余油分布预测的目的。

After a majority of our oilfield and secondary recovery,we just get 30% of the total reserves,which means that 60% to 70% of the remaining residual oil in the ground become remaining oil.China's 2007 crude oil output grew only by 1.6 percent to 12.872 million tons.Clarifying the law of distribution of the remaining oil and improving oil recovery is not only an economic effect,but also a National oil strategic issue.In this study,we used MATLAB language to establish a forecast system of distribution of remaining oil based on BP network.After studying,this system can get a nonlinear function between the geographical coordinates and porosity.We can get the regional porosity us-ing this system.Then using the relationship the porosity and the remaining oil,we can known the distribution of the remaining oil.