No abstract

No abstract

No abstract

Carbonate rocks constitute a large number of petroleum reservoirs worldwide. Notwithstanding, the characterization of these rocks is still a challenge due to their high complexity and pore space variability, indicating the importance of further studies to reduce uncertainty in reservoir interpretation and characterization. This work was performed for coquina samples from Morro do Chaves Formation (Sergipe-Alagoas Basin), analogous to important Brazilian reservoirs. Computed tomography (CT) was used for three-dimensional characterization of rock structure. The neural network named Self-Organizing Maps (SOM) was used for CT images segmentation. According to our tests, CT demonstrated to be a consistent tool for quantitative and qualitative analysis of heterogeneous pore space, by the evaluation of porosity, connectivity and the representative elementary volume.

No abstract