用于油田产量预测的多元线性回归和自回归模型

Multivariate Linear Regression and Automatic Regression Models Applied to Output Prediction of Oilfield

为了预测油田的动态产量,通过对多个自变量进行综合回归分析,确定了影响油田产量的重要因素和自变量的数目,并建立了相应的多元线性回归预测模型。基于多元线性回归模型和计算统计量,分析了回归参数求解过程中的重要信息,并建立了年产油量自回归预测模型。将两种预测模型应用于某油区,获得了较为满意的预测结果。

In order to predict the dynamic output of an oilfield, many factors were analyzed synthetically. The prominent factors and the independent variables influencing oilfield output were confirmed. A prediction model of multivariate linear regression is proposed in this paper. Meanwhile, the automatic regression model for predicting annual output of an oilfield is built up by analyzing the statistic and the important information obtained from the solving process of regression parameters. The applied results of the two models in some oilfields are satisfactory. 19