摘要
偏最小二乘回归分析通过从自变量和因变量数据表中提取包含原数据变异信息的成分来建立回归模型,能够解决回归建模过程中由于自变量之间的高度相关关系而引起的多重共线性问题。以油田操作成本为研究对象,以操作成本为因变量,选取产液量、产油量、注水量、含水率、措施工作量、工业品购进价格指数、电力价格等7个因素为自变量,以DX油田2010年至2016年的实际数据为基础,对DX油田各自变量指标进行偏最小二乘回归分析,建立回归预测模型,并对该模型进行验证。结果表明,自变量指标对操作成本的解释能力达到了0.99902,模型具有较高的可靠性。这一情况说明,将偏最小二乘回归分析应用于油田操作成本预测具有可行性。
Partial least squares regression analysis establishes a regression model by extracting the components containing the original data variation information from the independent variable and dependent variable data table,which can solve the problem of multiple collinearity due to the high correlation between the independent variables in the regression modeling process.With oil field operation cost as the research object,and operating cost as the dependent variable,we analyze the partial data of the DX Oilfield's variables through partial least squares regression analysis by using SIMCA-P software.The regression prediction model is established and tested.The results show that the independent variable index has an explanatory power of 0.99902,and the model has very high reliability.This research shows that partial least squares regression method is applicable to the prediction of oilfield operation cost,and can be used for reference in other research objects.
引文
[1]张继成,梁文福,赵玲,宋考平,甘晓飞.喇嘛甸油田特高含水期开发形势分析[J].东北石油大学学报,2005,29(3):23-25.
[2]李斌,吴晨洪,邵玉明.油气生产成本的关联分析[J].国际石油经济,2001,8(6):37-40.
[3]李丰,张晓辉,曲德斌,曲海旭.基于主成分回归模型的水驱油田操作成本预测[J].石油天然气学报,2012,34(9):136-139.
[4]Balch G R.Electricity Cost Reductions in the Oil Field[C]//Permian Basin Oil and Gas Recovery Conference.Midland:Society of Petroleum Engineers,1990.
[5]Robert F,Stiles M,Steven Slezak.Strategies for reducing oilfield electric power costs in a deregulated market[J].SPE Production&Facilities,2002,17(3):171-178.
[6]王惠文.偏最小二乘回归方法及其应用[M].北京:国防工业出版社,1999:200-234.
[7]王惠文.偏最小二乘回归的线性与非线性方法[M].北京:国防工业出版社,2006:97-141.
[8]吴茜茜,侯春华,陈武,赵小军,余晓钟.特高含水水驱油田操作成本组合预测方法研究[J].石油化工技术与经济,2014,30(6):5-9.
[9]Wold S,Albano C,Dunn W J.Pattern regression finding and using regularities in multivariate data[J].Analysis Applied Science,1983:147-188.
[10]王惠文.用PLS回归方法对中国沿海与内陆城市经济的比较分析[J].数理统计与管理,1998(5):10-15.
[11]李萍,毛琼,王新颖,王旭东,王晓芸.基于多元线性回归分析的操作成本预测模型建立与应用[J].石油规划设计,2018,29(3):33-37,52.