压裂油气藏产能岭回归预测分析方法

RIDGE REGRESSION PREDICTION AND ANALYSIS FOR PRODUCTIVITY OF FRACTURING RESERVOIRS

采用压裂后产能的岭回归预测分析方法,与目前广泛采用的单因素回归预测产能的方法进行了对比研究,并以大牛地气田盒三气层为例,对该气层的26口井的实际数据进行样本统计。通过岭回归分析的岭迹曲线确定独立变量的表现形式,建立了该气层的无阻流量的预测模型。预测结果表明,该模型拟合精度高,克服了目前广泛采用的单因素分析方法的不足,并通过预测模型分析了地层参数和压裂施工参数的影响,讨论了提高压裂效果的途径。

The ridge regression prediction and analysis method for productivity after fracturing is studied and compared with the single-factor regression prediction. Samples statistics have been carried on for actual data from 26 wells of Hesan Formation in Daniudi Gasfield. Through the independent variable obtained from the ridge trace curve of ridge regression analysis, a prediction model for open-flow capacity is established. Prediction results show that the model has a high matching accuracy and overcomes a shortcoming of single-factor analysis method which is widely adopted currently. Moreover, the paper analyzes the influence of formation and fracturing operation parameters by the model, and discusses how to improve fracturing results.