神经网络在评价塔北低阻油气储层参数中的应用

本文阐述了塔里木盆地北部地区（简称塔北）低阻油气储层的成因及评价北低阻油气储层 方法一神经网络法。该方法利用神经网络原理中前馈网络中的逆传播（ＢＰ）学习算法求取孔隙度、渗透率和饱和度等储怪参数。在选择解释模型时，考虑了各参数的不同选取了不同的网络结构。通过实际资料的处理结果与化验资料、测试结果相比较，认为神经网络法在塔北低阻油气储层测井解释中提高了储层参数的评价精度，具有良好的地质效果。

Application of Neural Network to Lowresistivity Reservoir Evaluation in the North of Talimu Basin

The origin of low-resistivity reservoir in the north of Talimu Basin and itsevaluation method-neural network method are given in this paper.The backpropagation algorithm of feed-forward neural network is adopted to estimatereservoir parameters(porosity, permeability and water saturation). Whileselecting the interpretation model,different network structure is applied to eachkind of parameters.Comparing the processing results with lab test and well testresults proves this method greatly increases the precision of reservoir parametersevaluation in low-resistivity reservoir in the north of Talimu Basin, and has a bettergeological application effect.