Abstract
Evaluation of filtration-capacitive properties of fractured carbonate section in most cases is associated with difficulties due to the high lithological heterogeneity, the complex structure of the pore space, the presence of a system of fractures that causes fluid filtration in reservoirs of this type. For a more homologous modeling of filtration behavior and achieving the maximum production level, it is important to carry out a differentiated study of the filtration-capacitive properties of the carbonate reservoir.
The identification of rock types by well log data is the first step to their differentiated study. The typification of reservoirs of carbonate deposits of Jurassic age, represented by limestones and dolomites, was carried out by a comprehensive analysis of the results of interpretation of borehole images of electrical resistivity, a conventional wireline logs and the results of core studies by using clustering method, such as MRGC (Multi Resolution Graph Based Analysis). The main purpose of the analysis was to identify areas of increased fracturing and divide reservoirs into pore-fracture and fracture pore types and refinement of the permeability model.
The resulting rock types, considering the structure of the pore space, were used with seismic attribute data to calculate volumes of lithofacies and the facies probability using the DNNA (Democratic Neural Network Association) seismic data classification algorithm. In this work, 5 forecasts were performed, each of which used 6 wells with a different number of training and test wells. A single tracing of the main anomalies associated with the target lithofacies is demonstrated for all predictions. As a statistical assessment of the results of all iterations, it was proposed to use the so-called cube of "repeatability" of the presence of the facies of interest, which is the sum of how many realizations the facies of interest occur at a certain point in space. The results have been used in the fracturing model as trends for the propagation of fractures intensity for productive horizons and for a conceptual understanding of the development of fracturing in the work area. The option to forecast the spatial distribution of high-fractured rock facies and clarify the filtration properties is key in the search for promising areas and the successful placement of new wells.
Rock typing reservoirs using the described methodology is a crucial aspect of hydrodynamic studies. The presence of fractured reservoirs significantly impacts deposit drainage and development approaches, so it is necessary to consider this feature in the hydrodynamic model to reproduce the characteristic effects of a fracture system. This approach increases the informativeness of the geological model and allows us to flexibly adjust the characteristics of fractured rocks in the model for further hydrodynamic modeling.
