摘要
讨论了以模糊理论为基础的合试层产能分权法。由SOM自组织映射与TFI模糊聚类协同确定分类数及变量组合 ,以隶属度分权、动态地剖分合试层产能为预测单层样本数据体 ,以遗传神经网络建立预测模型来预测储层自然产能 ,形成独特、灵活的泛合试层测井产能预测方法。经对鄂尔多斯盆地中部气田 16口井 5 2个测试层的应用 ,取得较好的应用效果。
This paper discusses weighted division of multi layer combined test production based on fuzzy theory.A method for logging prediction of production capacity of combined test zones was developed. Classifications and clustering features were determined by self organizing Mapping(SOM) and total fuzzy iteration(TFI). The production capacity of combined test zones was divided into single layer data sets by weights on membership to fuzzy classes. The natural production capacity of reservoirs was established by genetic neural network.Satisfactory results have been achieved by applying the method to 52 combined test zones of 16 wells in gas fields of middle Ordos basin.
出处
《勘探地球物理进展》
2003年第2期109-113,共5页
Progress in Exploration Geophysics
