摘要
水平井体积压裂是致密油藏高效开采的主要技术手段,准确预测产能对油田施工方案编制具有重要的指导意义,开采效果受地层因素、原油物性因素、压裂施工因素等影响,基于机理计算公式的传统预测方法存在一定局限性,提出一种基于K-means聚类与支持向量回归的产能预测组合模型,采用主成分分析算法解决K-means中欧氏距离对所有特征贡献程度一致性问题,K-means聚类结果与压裂施工参数结合作为SVR预测样本,有效解决不同区域间差异较大等问题。通过实验对比SVR、BP神经网络,预测准确性和稳定性优于单一模型,具有较高的合理性,可为致密油田高效开发提供指导性建议。
Volumetric fracturing of horizontal wells is the main technical method of efficient production in tight reservoirs,Accurate prediction of productivity has important guiding significance for oilfield construction plan.The mining effect is affected by formation factors,crude oil property factors,fracturing construction factors and so on.The traditional prediction method based on mechanism calculation formula has some limitations.A combined model of capacity prediction based on K-means clustering and support vector regression is proposed.Principal component analysis algorithm is adopted to solve the consistency problem of K-means central European distance's contribution to all features.The combination of K-means clustering results and fracturing operation parameters as SVR prediction samples can effectively solve the problems such as large differences between different regions.Compared with SVR and BP neural network prediction model,the prediction accuracy and stability are better than the single model,and it has a high rationality,which can provide guidance for the efficient development of tight oil field.
作者
刘新平
邓杰
杨鹏磊
LIU Xinping;DENG Jie;YANG Penglei(College of Computer Science and Technology,China University of Petroleum(East China),Qingdao 266580)
出处
《计算机与数字工程》
2023年第9期1949-1953,共5页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61801517)
中央高校基本科研业务专项经费(编号:19CX02029A,19CX02027A)资助。
关键词
致密油藏
产能预测
K-MEANS
支持向量回归
主成分分析
tight reservoir
productivity prediction
K-means
support vector regression
principal component analysis
