As the core function of dual-porosity model in fluids flow simulation of fractured reservoirs, matrix-fracture transfer function is affected by several key parameters, such as shape factor. However, modeling the shape factor based on Euclidean geometry theory is hard to characterize the complexity of pore structures. Microscopic pore structures could be well characterized by fractal geometry theory. In this study, the separation variable method and Bessel function are applied to solve the single-phase fractal pressure diffusion equation, and then the obtained analytical solution is used to deduce one-dimensional, two-dimensional and three-dimensional fractal shape factors. The proposed fractal shape factor can be used to explain the influence of microstructure of matrix on the fluid exchange rate between matrix and fracture, and is verified by numerical simulation. Results of sensitivity analysis indicate that shape factor decreases with tortuosity fractal dimension and characteristic length of matrix, increases with maximum pore diameter of matrix. Furthermore, the proposed fractal shape factor is effective in the condition that tortuosity fractal dimension of matrix is roughly between 1 and 1.25. This study shows that microscopic pore structures have an important effect on fluid transfer between matrix and fracture, which further improves the study on flow characteristics in fractured systems.

Lithofacies identification is essential in reservoir evaluation, especially in ultradeep carbonate reservoirs. In general, coring samples are the best sources to identify carbonate lithofacies because they are taken directly from reservoirs. However, the core is expensive to obtain, and generally its availability is greatly limited. In recent years, deep learning has attracted enormous attention because of its robust nonlinear regression and classification ability. This study applies a deep-learning algorithm to identify the lithofacies using geophysical well-log data. Six types of well-log data, such as natural gamma ray, density (DEN), neutron porosity (CNL), acoustic (AC), and shallow and deep lateral resistivity well logs (RT/RXO), are smoothed by the average sliding method and converted to 2D data. Then, the 2D data are treated as inputs to predict the carbonate lithofacies through the convolutional neural network (CNN). The results indicate that the prediction accuracy rate is 90.2%. This indicates that the CNN can identify different carbonate lithofacies well.

Carbonates commonly present a stacking-pattern architecture determined by the depositional processes despite the diagenetic modifications. This study integrates spectral cyclostratigraphy and petrophysical analysis to locate prolific reservoir beds to optimize reservoir development. This involves subdividing the reservoir succession into different hierarchical units to understand the variation of microfacies and pore-structure types. The variation trends and discontinuity surfaces (chronostratigraphic and lithostratigraphic boundaries) are determined by conjointly interpreting the prediction error filter analysis (PEFA) and integrated PEFA (INPEFA) and synthetic seismic trace and seismogram. We define the rock texture types (microfacies) through self-consistent approximation based on the INPEFA log while considering the variation of depositional sequence and grain size, core/thin-section petrography, and borehole electrical image interpretation. We establish lithology-depth and petrophysical classes-depth profiles using largely well-log data and present new petrophysical classes for naturally fractured vuggy reservoirs. We predict porosity and permeability from the distribution of high-frequency cycles and lithofacies while considering the implications of diagenesis. The results indicate that the depositional processes and diagenesis largely control the quality of the reservoir beds in response to the relative sea-level variations. The reservoir beds comprise the reservoir units with connected-vugs- and fractures-dominated pore systems and the horizons with microporosity-dominated pore systems that are candidates for secondary recovery processes. The study proposes a total porosity variation model that describes an external cardiac compression-like depth profile, revealing alternation between low/tight and high-porosity horizons, with the highly porous intervals associated with the solution-enlarged porosity zones. Permeability varies with the pore type mixing and rock texture type rather than the total porosity. The method well applies to distinguish flow conduits from baffles and barriers in complex carbonate reservoirs.