基于BP神经网络的碳酸盐岩储层缝洞充填物测井识别方法

缝洞是碳酸盐岩储层重要的储集空间和渗流通道,缝洞充填物的识别对评价油气储集能力和渗流能力具有重要作用。解决缝洞充填物与常规测井响应之间的非线性问题,BP神经网络法具有突出的优势。为此,通过结合成像测井和岩心资料,将碳酸盐岩储层缝洞充填物划分为泥质充填、砂质充填和结晶碳酸盐岩充填3类类型;分析不同充填类型的缝洞测井响应特征,选取敏感性较强的泥质含量、裂缝孔隙度、中子比、密度比和深侧向电阻率等5个参数,利用BP神经网络建立了碳酸盐岩储层缝洞充填物的识别方法。应用所建立的方法对实际井资料进行了处理评价,其预测结果与实际结果有较好的一致性,取得了较好的应用效果。

Fractures and caverns are the mainly part of storage space and seepage channel of carbonate reservoir.The identification on fillings in fractures and caverns plays an important role in the evaluation of storage and seepage capacity of oil reservoir.BP neural network has advantages in solving non-linear problems between fillings and conventional logging responses.Therefore,combining imaging logging with core data,fillings in fractures and caverns in carbonate reservoir are divided into three types,including argillaceous filling,sand filling and crystallization carbonate filling.Five sensitive parameters are used for modeling,such as shale content,fracture porosity,neutron ratio,density ratio and deep lateral resistivity,and so on,which are extracted according to the characteristics of different fillings in the fractures and caverns.Moreover,BP neural network is utilized to establish the identification method for the fillings in fractures and caverns in carbonate reservoirs.The method has been applied to evaluate practical well logging data,and the prediction results is well coinciding with actual data.