Permeability represents the flow conductivity of a porous media. Since permeability is one of the most vital as well as the complex properties of a hydrocarbon reservoir, it is necessary to measure/estimate accurately, rapidly and inexpensively. Routine methods of permeability calculation are through core analysis and well tests, but due to problems and weaknesses of the aforementioned methods such as excessive costs and time, these are not necessarily applied on neither in all wells of a field nor in all reservoir intervals. Therefore, log-based approaches have been recently developed. The goal of this research is to provide a flowchart to estimate permeability using well logs in one of Iranian south oil fields and finally to introduce a new algorithm to estimate the permeability more accurately. Permeability is firstly estimated using artificial neural network (ANN) employing routine well logs and core data. Subsequently, it is estimated using Stoneley-Flow Zone Index (ST-FZI) and is compared with the results of core analysis. Correlation coefficients in permeability estimation by artificial neural network and Stoneley-FZI are R² = 0.75 and R² = 0.85, respectively. On the next step, an improved algorithm for permeability prediction (improved ST-FZI) is presented that includes the impact of lithology and porosity type. To improve the permeability estimation by ST-FZI method, electro-facies clustering based on MRGC method is employed. For this purpose, rock pore typing utilizing VDL and NDS synthetic logs is employed that considers the porosity types and texture. The VDL log separates interparticle porosity from moldic and intra-fossil porosities and washes out and weak rock-type zones. Employing MRGC method, three main facies are considered: good-quality reservoir rock, medium-quality reservoir rock and bad-quality (non-reservoir) rocks. Permeability is then estimated for each group employing ST-FZI method. The estimated permeability log by improved ST-FZI method shows better match with the measured permeability (R² = 0.93). The average error between estimated and measured permeability for ANN, ST-FZI method and improved ST-FZI method is 1.83, 1.18 and 0.796, respectively. The increased correlation is mainly due to involving the impact of porosity types on improved ST-FZI method. Therefore, it is recommended to apply this algorithm on variety of complicated reservoir to analyze its accuracy on different environments.

渗透率表示多孔介质的流导率。由于渗透率是碳氢化合物储层最重要和最复杂的特性之一，因此有必要进行准确，快速和廉价的测量/估算。渗透率计算的常规方法是通过岩心分析和试井，但是由于上述方法的问题和缺点（例如，成本和时间过长），未必将这些方法既不适用于油田的所有井，也不适用于所有储层间隔。因此，最近开发了基于日志的方法。这项研究的目的是提供一个流程图，以使用伊朗南部某油田的测井曲线估算渗透率，并最终引入一种新算法，以更准确地估算渗透率。首先使用人工神经网络（ANN）估算渗透率，该人工神经网络采用常规测井和核心数据。随后，使用斯通利流动区指数（ST-FZI）进行估算，并将其与岩心分析结果进行比较。人工神经网络和Stoneley-FZI估算渗透率的相关系数为R ²  = 0.75和R ² 分别为0.85。下一步，提出了一种用于渗透率预测的改进算法（改进的ST-FZI），其中包括岩性和孔隙度类型的影响。为了提高基于ST-FZI方法的渗透率估计，采用了基于MRGC方法的电相聚类。为此，采用了利用VDL和NDS合成测井曲线进行岩石孔隙分型的方法，该方法考虑了孔隙类型和质地。VDL测井将颗粒间孔隙度与发霉和化石内部孔隙度分开，并冲刷出来并形成弱岩石型区域。采用MRGC方法，考虑了三个主要相：优质储层岩石，中等质量储层岩石和劣质（非储层）岩石。然后使用ST-FZI方法估算每个组的渗透率。R ²  ＝ 0.93）。ANN，ST-FZI方法和改进的ST-FZI方法的估计渗透率和测量渗透率之间的平均误差分别为1.83、1.18和0.796。相关性的增加主要是由于孔隙类型对改进的ST-FZI方法的影响。因此，建议将该算法应用于各种复杂油藏，以分析其在不同环境下的精度。