更新时间：05-12 （julu1004）提供原创文章
摘要：在测量试验中，采用的是新疆油田公司的低渗岩心。该研究首先利用粒度实验、压汞实验、离心实验、相渗实验、敏感实验、润湿性等实验得出测量的岩心的数据，然后采用典型参数研究方法，尽可能利用多种参数进行反演目标参数，从而填补未取得实验参数部位的目标参数，得出一个区块各个层位上的参数。利用多元线性回归和BP神经网络方法，通过单参数、多参数分析研究，建立数据与其他一个或多个数据的关系的参数模型，得出了渗透率、孔隙度、油水饱和度、驱油效率、敏感性等参数与其他实验参数的关系，形成了毛管力、相对渗透率、粒度等典型曲线的建立方法，并根据图形拟合出一定的数量关系，得出低渗砂岩岩心的实验数据的关联性。
关键词：多元线性回归方程；BP神经网络；典型参数研究方法
Abstract : Among the measurement test, the low permeability core comes from the Xinjiang Oilfield Company. The study firstly use the particle size experiment, mercury intrusion experiments, centrifugation, and relative permeability experiments, sensitive experiments, the wettability of the experimental data obtained core, Then use the method of typical parameters, using a variety of parameters inversion of target parameters as much as possible, Thereby filling the parts of the target parameters of the experimental parameters which not obtained. Draw the layer parameters on each block . We can use the method of multiple linear regression and BP neural network, by a single parameter, multi-parameter analysis, then we can establish a model which can show the relationship between the data and one or more data. Derived the relationship between permeability, porosity, oil and water saturation, flooding the oil efficiency, sensitivity and other parameters and other experimental parameters, we can find the establishment of the typical curve of the formation of capillary pressure, relative permeability, particle size, according to the graphics ,we can fit a certain number of relations, and obtained the correlation of the experimental data in low permeability sandstone cores.
Key words : Multiple linear regression equation; BP neural network; Typical parameters of research methods