应用神经网络计算储层参数

Application of Neural Network to the Calculation of Reservoir Parameters

以岩心分析资料及多种测井信息为依据，首先利用样本信息的神经元模型（ＣＵＳＩ）解决了储层参数的计算问题，并利用改进后的自适应神经元模型（ＡＣＵＳＩ）提高了分析精度。最后利用前馈神经网络的误差反向传播模型（ＢＰ）网络的外延和信息表达能力解决了非储层的定量识别。应用上述方法对辽河油田四口井进行了逐点参数分析，分析结果与实际情况吻合很好。

Based on data from core analysis and various log information, CUSI model is used for the calculation of reservoir parameters. The adaptive capacity of the improved ACUSI model enhances the accuracy of the analysis result. And the extrapolation and information expressivity of BP network is a great help in quantitative identification for non-reservoir. The above-mentioned models have been applied to parameters calculation point by point for four wells in l.iaohe oil Field. The interpretation results which are given in the paper agrees well with the actual situation.