论文部分内容阅读
以伊朗Mansuri油田50口井的常规测井资料为基础,优选人工智能算法,对Mansuri油田白垩系Ilam组4个层的孔隙度和渗透率分布进行模拟。首先利用5口有岩心物性分析资料的井,遴选出常规测井的声波时差、密度和中子孔隙度作为输入参数,采用反向传播人工神经网络(BP神经网络)和支持向量回归方法进行储集层孔隙度和渗透率计算,根据计算结果与岩心实测结果的相关性,选择采用BP神经网络法进行物性计算。然后,利用克里金地质统计算法,对Mansuri油田Ilam组4个层的孔隙度和渗透率分布进行模拟,结果表明,层2.1和层2.2为高孔隙度层,层1、层2.1和和层2.2高渗透层,层3为非储集层;储集层孔隙度和渗透率分布总体呈北部高、南部低的特点。
Based on the conventional logging data from 50 Mansuri oilfields in Iran, the artificial intelligence algorithm was optimized to simulate the porosity and permeability distribution of 4 layers in Cretaceous Ilam Formation in Mansuri Oilfield. First of all, five wells with petrophysical analysis data were selected to select the acoustic logging time-density, density and neutron porosity of conventional logging as the input parameters, back propagation artificial neural network (BPNN) and support vector regression Based on the correlation between calculated results and core measured results, the BP neural network method was chosen to calculate the physical properties. Then, using the Kriging geostatistical algorithm, the porosity and permeability distributions of four layers in the Ilam group of Mansuri field are simulated. The results show that layers 2.1 and 2.2 are high porosity layers, layers 1 and 2.1 and the strata 2.2 High permeability layer, Layer 3 is non-reservoir; The porosity and permeability distribution of reservoir are generally high in the north and low in the south.