致密油藏水平井压裂后产能预测方法PRODUCTIVITY PREDICTING METHOD OF THE FRACTURED HORIZONTAL WELLS IN THE TIGHT OIL RESERVOIRS
纪天亮,卢双舫,唐明明,王民,赵宏宇,梁宏儒
JI Tianliang,LU Shuangfang,TANG Mingming,WANG Min,ZHAO Hongyu,LIANG Hongru
摘要(Abstract):
以松辽盆地南部红岗北水平井开发目标区扶余油层水平井压裂后各项参数为基础,首先采用灰色关联分析方法对扶余致密油藏压裂水平井产能参数的影响程度进行排序,优选出主要影响参数;然后对研究区压裂水平井进行模糊聚类分析,将研究区压裂水平井分成4类;最后将水平井分类结果与影响参数一起加入到BP神经网络输入端,以日产能指标为硬数据,建立预测压裂后水平井产能的神经网络模型,并利用检验数据对模型进行验证。结果表明,基于水平井类型控制的神经网络模型产能预测效果要优于传统神经网络模型,可以作为有效的水平井压裂后产能预测方法。
Based on all parameters of the fractured horizontal wells in Fuyu oil layers in the horizontal well development target block of North Honggang in South Songliao Basin,the method of grey relational analysis is adopted first to sequence the influence degrees of the productivity parameters of the fractured horizontal wells in the layers,and then the major impacting parameters are selected; secondly,the fuzzy clustering analysis is applied to categorize and research the wells in the study area into four kinds; finally the classified results and influencing parameters are input into BP neural network to built the neural network model to predict the horizontal wells' productivity after being fractured,and moreover the testing data are adopted to verify the model. The results show that the predicted effects of the neural network model based on the control of the types of the horizontal well are well better than those of the traditional neural network model,so the productivity predicting method for the fractured horizontal wells is proven to be effective.
关键词(KeyWords):
致密油藏;水平井;压裂产能;灰色关联;模糊聚类;神经网络
tight oil reservoir;horizontal well;fractured well's productivity;grey relation;fuzzy clustering;neural network
基金项目(Foundation): 国家自然科学基金项目(41172134,41402108);; 中石油股份公司专项(2012E-0501)
作者(Author):
纪天亮,卢双舫,唐明明,王民,赵宏宇,梁宏儒
JI Tianliang,LU Shuangfang,TANG Mingming,WANG Min,ZHAO Hongyu,LIANG Hongru
参考文献(References):
- [1]刘新,张玉纬,张威,等.全球致密油的概念、特征、分布及潜力预测[J].大庆石油地质与开发,2013,32(4):168-174.
- [2]张应安.松辽盆地致密砂岩气藏水平井多级压裂现场实践[J].天然气工业,2011,31(6):46-48.
- [3]Joshi S D.Augmentation of well productivity using slant and horizontal wells[R].SPE 15735,1986.
- [4]傅荟璇,赵红.Matlab神经网络应用设计[M].北京:机械工业出版社,2010:94-95.
- [5]叶双江,姜汉桥,陈民峰.基于灰色关联与神经网络技术的水平井产能预测[J].大庆石油学院学报,2009,33(3):53-55.
- [6]梁淑贤,高建,周体尧,等.吉林红岗北致密油特征、成因及开采技术[J].大庆石油地质与开发,2014,33(2):165-170.
- [7]Michelevichius D,Zolotukhin A B.Evaluating productivity of a horizontal well[R].SPE 79000,2002.
- [8]傅立.灰色系统理论及其应用[M].北京:科学技术文献出版社,1992:185-211.
- [9]李巧云,张吉群,邓宝荣,等.高含水油田层系重组方案的灰色决策优选法[J].石油勘探与开发,2011,38(4):463-468.
- [10]高建,侯加根.多油层弱非均质油藏流动单元划分及控制因素分析[J].大庆石油地质与开发,2008,27(3):72-75.
- [11]谢季坚,刘承平.模糊数学方法及其应用[M].4版.武汉:华中科技大学出版社,2013:68-71.
- [12]崔传智,赵晓燕.模糊聚类分析方法在水平井开发指标预测中的应用[J].数学的实践与认识,2007,372(4):78-82.
- [13]陈仁保,师俊峰.神经网络系统在稠油区块优先动用排序中的应用[J].大庆石油地质与开发,2005,24(6):84-86.
- [14]连承波,李汉林,渠芳,等.基于测井资料的BP神经网络模型在孔隙度定量预测中的应用[J].天然气地球科学,2006,17(3):382-384.
- [15]廖明燕.基于神经网络多参数融合的钻井过程状态监测与故障诊断[J].中国石油大学学报(自然科学版),2007,31(4):149-152.
- [16]周金应,桂碧雯,李茂,等.基于岩控的人工神经网络在渗透率预测中的应用[J].石油学报,2010,31(6):311-318.
- [17]韩立群.人工神经网络教程[M].北京:北京邮电大学出版社,2006:58-80.
- 致密油藏
- 水平井
- 压裂产能
- 灰色关联
- 模糊聚类
- 神经网络
tight oil reservoir - horizontal well
- fractured well's productivity
- grey relation
- fuzzy clustering
- neural network