摘要
影响重复压裂选井的因素很多,且各个因素之间具有复杂的非线性关系,利用传统的预测方法受到诸多限制。针对此问题,应用现代数学理论,确定了影响重复压裂选井选层的主要因素,建立了选井选层模型。研究结果表明,应用灰色系统理论方法可综合考虑多种影响因素,并对其进行量化,确定出这些因素对压后效果的权重影响,增加了决策的科学性;建立的BP神经网络模型能够自组织、自适应地解决复杂的非线性问题,增加了重复压裂选井选层预测结果的可信度;应用遗传算法及自适应学习效率法对模型进行了改进,提高了网络的学习效率和精确度。本模型对现场实际施工有一定的指导作用。
There is complicated nonlinear relationship between many factors affecting well and layer selection for re-fracturing,so traditional forecasting methods are always useless.For this problem,modern mathematical theory is applied to determine the impact of main factors on well and layer selection for re-fracturing,and a well and layer selection model is established in this paper.The results show that: a variety of factors are considered and quantified by gray system theory approach,and the impact weight of these factors is determined,which improve science of decision-making;the complex nonlinear problems were solved by the BP neural network model self-organizationally and adaptively,which increase the credibility of well and layer selection for re-fracturing prediction.The network learning efficiency and accuracy has been improved greatly by the application of genetic algorithms and efficient method for adaptive learning method.This model is simple and convenient and of certain application value on site.
出处
《油气井测试》
2013年第2期4-6,12,共4页
Well Testing
基金
国家科技重大专项子课题"油田开采后期提高采收率新技术"(2011ZX05009-004)资助
关键词
重复压裂
选井选层
影响因素
灰色关联度
BP神经网络
re-fracturing,well and layer selection,factor
grey relational grade,BP nerve network
