应用量子神经网络快速预测储层敏感性  

FAST PREDICTING THE SENSIBILITY OF RESERVOIRS BY USING QUANTUM NEURAL NETWORKS

作　　者：孙玉学[1] 谢建波[1] 赵景原[1] 历艳明 王娇[3] 

机构地区：[1]东北石油大学石油工程学院 [2]大庆油田钻探集团第三公司 [3]东北石油大学化学化工学院

出　　处：《钻采工艺》2012年第5期72-74,11,共3页Drilling & Production Technology

基　　金：中国石油大庆油田公司"龙西地区储层损害机理及保护措施研究"(Dq-1204003201-js-318)项目内容

摘　　要：针对室内评价储层敏感性预测的不足和传统神经网络模型的缺陷,在收集岩心分析资料、分析生物神经元的信息处理方式和量子特性的基础上,提出一种具有量子特性权值和活性值的量子神经网络预测模型[1]。对吉林油田伊通地区储层敏感性进行快速预测,其结果表明,量子神经网络模型的敏感性伤害程度预测结果明显优于传统BP神经网络,与岩心流动实验结果的符合率达到88%。该方法能快速、准确地预测储层的敏感性指数,为保护油气层提供可靠的理论依据。Aiming at the shortcomings of the indoor evaluation to the reservoir sensitivity prediction and the defects of traditional neural network model,the quantum neural network model was put forward,which has quantum properties metric and activity values based on collecting core analysis data,analysis of biological neuron information processing and quantum characteristics.Quick to predict the reservoir sensitivity of Yitong area,Jilin Oilfield,the results showed that the quantum neural network model was much better than traditional BP neural network to the degree of sensitivity damage,and the accuracy with experimental results of core flowed up to 88%.This method could quickly and accurately forecast reservoir sensitivity index,it provided a reliable theoretical basis for the reservoir protection.

关 键 词：量子神经元 量子神经网络 储层保护 预测 敏感性 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] P618.13[自动化与计算机技术—控制科学与工程]