欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
镇泾油田长9致密砂岩储层基质孔隙度小,渗透率低,非均质性强,储集结构复杂,依据常规测井交会图和基于不同参数“距离”差异等方法识别流体效果不理想,难以有效应用.提出基于离散Hopfield网络（DHNN）的稳态收敛原理实现流体识别.该方法利用研究区评价井试油资料建立流体信息编码方式,分别对油层、油水同层、含油水层、水层和干层进行信息编码,通过DHNN网络对流体信息编码进行聚类分析实现流体识别.从研究区其他多口井的应用看,流体预测结果与试油结果吻合,从而验证了该方法的可靠性.
英文摘要：
Tight sandstone reservoirs of Chang9 in Zhenjing oilfield have the characteristics of low porosity, low permeability and great heterogeneity, some fluid identification methods are not effective, such as well logging cross-plot and cluster analysis aimed at ＂distances＂ in attributes. A new approach based on Discrete Hopfield Neural Network（DHNN） is proposed to solve this problem. The new approach sets up fluid identifying code rules using appraisal wells data and fluid types （oil, oil-water, water-with-oil and water）, which are compiled as a unique matrix codes before these codes are used for fluid prediction through DHNN cluster analysis. According to the application in other wells of Chang formation, the predicted results coincide with oil testing, the results of which validate that the method is effective.
同期刊论文项目
同项目期刊论文
期刊信息
位置：