基于支持向量回归的油藏瞬态压力拟合  

Reservoir history matching for transient pressure based on support vector regression

作　　者：许恩华 查文舒[1] 李道伦[1] 陈刚 XU Enhua;ZHA Wenshu;LI Daolun;CHEN Gang(School of Mathematics, Hefei University of Technology, Hefei 230601, China)

机构地区：[1]合肥工业大学数学学院,安徽合肥230601

出　　处：《合肥工业大学学报（自然科学版）》2020年第11期1570-1574,1584,共6页Journal of Hefei University of Technology：Natural Science

基　　金：国家重大科技专项资助项目(2017ZX05009-002)。

摘　　要：试井技术是利用井底压力数据反演地层和井筒参数的方法,是典型的反问题。文章提出了一种基于支持向量回归的反演方法。首先使用随机算法生成试算算例;然后使用不同核函数下的支持向量回归模型对井底压力和压力导数进行拟合,选取误差最小的模型进行试井分析,利用BFGS算法对模型进行优化得到最优不确定参数组;最后比较计算值与观测值之间的压力、压力降落、压力降落导数,从而得到结果。结果表明,该方法能构建带有核函数的回归模型对井底压力以及压力导数进行拟合,拟合效果好、计算效率高,具有很好的应用前景。Well test analysis is an inversion method that uses well bottomhole pressure to evaluate reservoir and wellbore parameters.This paper presents an automatic fitting method based on support vector regression.Firstly,trial examples are generated by randomized algorithms.Then,well bottomhole pressure and derivative are matched by support vector regression models based on different kernel functions.The model with smallest error is used to get the reservoir and wellbore parameters using BFGS algorithm by the comparison of well bottomhole pressure and derivative between the simulated data and observed data.The result shows that the method can match well bottomhole pressure and derivative well and would be useful in the numerical well test interpretation.

关 键 词：数值试井分析 支持向量机(SVM) PEBI网格 反问题 历史拟合 

分 类 号：TE319[石油与天然气工程—油气田开发工程]