基于模糊逻辑的多属性融合油气预测方法

Oil-gas prediction with multi-attributes fusion based on fuzzy logic

地震属性与储层参数不存在严格的一一对应关系,这就导致仅用单一属性解释储层不可避免地存在多解性。在利用属性进行储层预测过程中,单一属性或综合属性能否指示含油气情况,往往没有绝对分明的界限,即存在模糊性。为此,本文提出基于模糊逻辑的多属性融合方法,即将模糊理论应用于属性融合,利用隶属函数将模糊问题进行定量化处理,通过模糊贴近度矩阵求取各自的权重,最终得到一个定量的综合结果。属性融合过程中各种属性权重的求取完全基于数据驱动,降低了人为操控的主观性。数值模拟试验和实际资料处理结果均显示,用本文方法所得融合结果与井资料能较好吻合,表明利用本文提出的基于模糊逻辑的多属性融合方法进行油气预测是有效且实用的。

There is no strictly corresponding relationship between seismic attributes and reservoir parameters,which inevitably gives rise to uncertainty of reservoir characterization when only a single attribute is used.During the process of reservoir prediction,there is no distinctive boundary between whether single or integral attributes are capable of indicating oil and gas or not because of ambiguity. Therefore we propose in this paper a multi-attributes fusion method based on fuzzy logic.We apply fuzzy theory to attributes fusion,quantify the fuzzy problem with the help of membership function, calculate weight factor through fuzzy nearness matrix,and eventually obtain a quantitative comprehensive result.During the process of fusion,the calculation of weight factor of each attribute is completely data driven which could reduce the subjectivity.This method has been tested on synthetic and real data in order to prove its feasibility and effectiveness.Results of the tests demonstrate that the fusion results match very well with well data. Consequently,the proposed multi-attributes fusion method based on fuzzy logic used for oil and gas prediction is effective and available.