The dynamic impact of the extra-matrix pore systems in Brazilian pre-salt carbonate reservoirs has driven the development of new techniques and workflows for their characterization. The presence of large-scale and well-connected pore systems can significantly influence various processes, from drilling and completion efficiency to flow modeling and history matching. Characterizing these systems requires a specific type of modeling that is based both on measurements and assumptions, as the pore network geometry extends beyond the well and resides in an intermediate scale between well logs and seismic data. Borehole image logs, although not fully capturing this pore scale, provide the most representative detection of these systems. Consequently, quantitative analysis of these structures based on borehole image logs became mandatory for pre-salt carbonate projects. In these reservoirs, porosity, and permeability models from the well to the field scale have been built by integrating attributes extracted from borehole image logs, conventional petro-physical log suits, routine core analysis and permeability estimations based on drill stem test measurements. However, the traditional petrophysical logs used for modeling the matrix porosity partially detect the extra-matrix pores system, even after proper environmental corrections. As a result, the logs typically used to characterize the matrix porosity are consistently contaminated by a mixture of pore volumes that actually belong to different pore types and scales, each with distinct flow behavior.
In this study, a comparative analysis of the different porosity logs acquired in pre-salt karstified carbonate sections is conducted to isolate the extra-matrix pore systems and characterize the matrix porosity more precisely. The decomposition of the T2 spectrum into log-normal base functions is considered the most accurate method for representing matrix porosity, as it separates the mud signal detected in large cavities from the relaxation of the matrix pore volumes. Conversely, the quantitative analysis of the borehole image logs must be used to represent the extra-matrix porosity. By separating the effects of mega and giga pores on porosity logs, it becomes possible to model parts of the reservoir that are expected to behave as standard porous media, while treating the large-scale pore system with different mathematical representations.
A sensitivity analysis was performed on a 3D model to demonstrate the impact on STOOIP under three scenarios: one with the ideal matrix porosity log, another one with the NMR porosity log contaminated by extra-matrix pore volume, and another considering the sum of ideal matrix porosity and extra-matrix porosity. The extra-matrix pores in karstified reservoirs do not flow like traditional porous media and can significantly impact project scope and field productivity as they may channel fluids and cause early breakthrough of gas or water, due to their high permeability. Hence, it is essential to represent these systems as accurately as possible. This approach aims to determine the best practices for representing porosity in karstified reservoirs while separating contaminated and non-representative measurements generated by large-scale pores in direct porosity tool readings, leading to more predictive models.