油田剩余可采储量预测算法设计及性能测试  

Design and Performance Test of Prediction Algorithm of Residual Recoverable Reserves

作　　者：郭瑾 矫欣雨 孙肖 GUO Jin;JIAO Xin-yu;SUN Xiao(College of Oil and Gas Engineering,Shandong Institute of Petroleum and Chemical Technology,Dongying Shandong 257061,China;College of Petroleum Engineering,China University of Petroleum,Beijing(CUPB),Beijing 102249,China)

机构地区：[1]山东石油化工学院油气工程学院,山东东营257061 [2]中国石油大学(北京)石油工程学院,北京102249

出　　处：《计算机仿真》2022年第8期144-147,170,共5页Computer Simulation

摘　　要：由于已有算法未能对油藏数据进行预处理,导致预测结果不准确,预测时间较长,提出一种基于蒙特卡洛法的油田剩余可采储量预测算法。选取密度高于数据集平均密度的数据对象作为初始聚类中心,引入信息熵计算数据点和簇聚类中心的加权欧式距离,借助簇中数据的加权欧式距离和簇中全部数据点的平均加权欧式距离对油藏数据进行预处理。根据油藏数据的特性,采用蒙特卡洛法建立油田剩余可采储量预测模型,估算油田剩余可采储量,并借助粒子群算法优化预测模型,完成油田剩余可采储量预测。实验分析结果表明,所提算法能够有效提升预测结果的准确性,同时还能够减少预测时间。Currently,some algorithms fail to preprocess the oil reservoir data,leading to inaccurate prediction results and a long prediction time.Based on the Monte Carlo method,an algorithm to predict the remaining recoverable reserves of an oilfield was proposed.Firstly,the data object whose density was higher than the average density of the data set was selected as the initial clustering center.Secondly,the information entropy was introduced to calculate the weighted Euclidean distance between data points and cluster centers.Thirdly,the oil reservoir data were preprocessed with the help of the weighted Euclidean distance of the data in the cluster and the mean weighted Euclidean distance of all data points in the cluster.According to the characteristics of reservoir data,the Monte Carlo method was adopted to build a model to predict remaining recoverable reserves,and thus to estimate the remaining recoverable reserves of the oilfield.Finally,the particle swarm optimization algorithm was used to optimize the prediction model,thus completing the prediction of the remaining recoverable reserves of the oilfield.Experimental results show that the proposed algorithm can effectively improve the predictive accuracy and reduce the prediction time.

关 键 词：蒙特卡洛法 油田 剩余可采储量 加权欧式距离 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]