摘要
由于油藏非均质性的影响,单个岩心的油水相渗曲线并不能代表整个油藏或区块的水驱渗流特征。为了获得具有代表性的油水相渗曲线,通常的做法是将渗透率相近的岩心的油水相渗曲线进行归一化处理。针对现有归一化处理方法的不足,本文提出一种油水相渗曲线归一化的新方法。首先将油(水)相对渗透率和含水饱度标准化;然后采用单调保形插值拟合;在插值拟合的基础上,利用BP神经网络计算归一化的油水相渗曲线。该方法得到的归一化油水相渗曲线更接近平均化的相渗曲线(实测曲线束中部),因此更能代表整个油藏或区块的水驱渗流特征,同时为后续的油藏数值模拟、动态分析、参数计算等提供更可靠的相渗资料。
In order to obtain the typical oil- water relative permeability curves,the relative permeability curves of cores with similar permeability are generally normalized. Aiming at the defects of the existing normalization method,this paper presents a new method for normalizing oil- water relative permeability curves. First,the water saturation and relative permeability are standardized. Secondly,the relative permeability and water saturation are fitted by shape- preserving interpolation. Finally,on the basis of the interpolation fitting,the normalized relative permeability curves are obtained by BP neural network calculation. The result of this method was closer to the average relative permeability curve and more representative of the water flooding seepage characteristics of the whole reservoir or block. Meanwhile,it can provide reliable data for the further reservoir numerical simulation,dynamic analysis and parameter calculation.
出处
《复杂油气藏》
2015年第1期38-40,51,共4页
Complex Hydrocarbon Reservoirs
基金
十二五国家科技重大专项<复杂裂缝性碳酸盐岩油田开发关键技术>(2011ZX05014-004)子课题<裂缝性复杂介质油藏相渗曲线测试>项目的支助
关键词
相渗曲线
单调保形插值
神经网络
岩心
relative permeability curve
shape-preserving interpolation
neural network
core
