摘要
针对我国陆相沉积油层组在多层合采井的单层产能和贡献率配分中精度不高的现状,进行了5层产能配比实验.通过气相色谱指纹技术定量了指纹化合物绝对浓度,并结合建立的BP人工神经网络模型,计算多层配比实验的单层产能贡献率,计算结果精度较高,其中最大绝对误差为0.34%,最小绝对误差为0,表明气相色谱指纹技术结合人工神经网络在多层合采井的单层产能贡献率计算中占有优势.
Aimed at the reservoir exploration characteristics of continental sedimentary and the accuracy of the productivity contribution in the commingled producing wells,allocation experiments of five layers were accomplished.Then absolute concentrations of finger compounds were obtained through gas chromatography technology,using the BP-ANN model to calculate the productivity contribution,the result shows that the accuracy is high,the maximum error is 0.34% among them.Therefore,the gas chromatography technology combined with ANN model probable has a good future in the calculation of productivity contribution.
出处
《大庆石油学院学报》
CAS
北大核心
2007年第6期8-10,21,共4页
Journal of Daqing Petroleum Institute
基金
黑龙江省杰出青年基金项目(200003)
关键词
合采井
产能贡献率
BP人工神经网络
气相色谱
配比实验
commingled producing well
productivity contribution
BP-ANN
gas chromatography
allocation experiment
