论文部分内容阅读
将所统计数据分别在普通坐标系、半对数坐标系和双对数坐标系中进行绘制,并分析其间的关系,选择回归预测技术进行预测,建立一元线性回归模型与多元线性回归模型。回归预测技术依据最小二乘法原理,进行数据拟合,确定经验公式的系数,求解经验公式。后继水驱阶段的累积产油量在半对数与双对数坐标系中与时间呈近似线性关系,但二者预测值有不同方向的偏差,将预测值求和取平均,预测结果更准确。应用该方法对实际聚合物驱块进行预测,拟合误差和后验误差很小,预测精度较高,因此预测结果是可靠的。
The statistical data are respectively plotted in ordinary coordinate system, semi-logarithmic coordinate system and double logarithmic coordinate system, and the relationship between them is analyzed. The regression prediction technique is selected for prediction, and a linear regression model and a multiple linear regression model are established. Regression prediction techniques based on the principle of least squares, data fitting, empirical formula to determine the coefficient to solve the empirical formula. The cumulative oil production in subsequent waterflooding stages is approximately linear with time in semi-logarithms and double-logarithmic co-ordinates, but the predicted values of the two have different deviations in different directions. The predicted values are averaged and the prediction results are more accurate . This method is used to predict the actual polymer flooding block. The fitting error and the posterior error are very small, and the prediction accuracy is high. Therefore, the prediction result is reliable.