支持向量机建模及游梁式抽油机抽汲参数优化  

Support Vector Machine Model and Beam Pumping Unit Pumping Parameters Optimization

作　　者：徐向前[1] 周好斌[1] 徐生辉[2] 

机构地区：[1]西安石油大学,西安710065 [2]长庆油田分公司机械制造总厂,西安710201

出　　处：《石油矿场机械》2016年第9期1-5,共5页Oil Field Equipment

基　　金：陕西省教育厅项目"基于支持向量机的游梁式抽油机平衡判定方法"(2013JK1024);西安石油大学<材料科学与工程>省级优势学科资助

摘　　要：为了提高抽油机的运行效率,需要在采油过程中对抽汲参数进行优化。提出了采用支持向量回归机在线预测油井产量及调整抽汲参数的方法。根据游梁式抽油机的工作原理,找出与油井产量密切相关的数据,构造支持向量机的训练样本。建立了支持向量回归机抽油机在线抽汲参数优化模型,主要包括以KKT条件为新增样本条件,根据样本的信息熵调整新增样本的数量,用粒子群算法调整、优化了支持向量机参数。根据油井生产数据进行了仿真,结果表明预测准确度较高,抽汲参数调整准确。To improve pumping efficiency of the pumping unit machine,the swabbing parameters are optimized in the process of oil production.Therefore,oil well output would be predicted by support vector machine online modeling method in this paper,in order to adjust swabbing parameter of pumping unit.According to the working principle of the beam pumping unit,the data that is closely related to the oil well production are found,and the training samples of the support vector machine are constructed with the data.The support vector regression machine oil pumping machine online swabbing parameter optimization model was built.The condition that the new sample is added to the training sample is to meet the KKT condition.The number of new samples is adjusted according to the information entropy of the sample,the parameters of support vector machine is modified by using particle swarm algorithm,and then the model can be refreshed online as time passes by.The simulation with the production data of oil well results show that the prediction accuracy is high,and swabbing parameters can be optimized.

关 键 词：游梁式抽油机 支持向量机 参数优化 

分 类 号：E933.107[军事—军事装备学]