Improving the Calculation of Petrophysical Properties in Vugular Carbonates Using Logs and Rock Samples: A Case Study in a Brazilian Pre-salt Well

Abstract

Calculating log-scale petrophysical properties is essential in formation evaluation, since it directly impacts the estimates of hydrocarbon reserves and reservoir flow capacity. These calculations traditionally use the nuclear magnetic resonance (NMR) log through cutoffs in the transverse relaxation time (T2) in the Brazilian pre-salt carbonates. T2 values above and below 100 ms are associated with free and irreducible fluids, respectively. Above the oil-water contact, free fluid is interpreted as oil when calculating saturation and permeability, assuming the rock has only water and oil. As the NMR readings are influenced by oil-based mud invasion, the free fluid is also related to the oil in the mud filtrate. The filtrate and formation oil present T2 bulk values of 500 ms, and the traditional interpretation calculates reliable properties in intervals with matrix porosity. Nevertheless, this interpretation does not hold in formations with vugular porosity. The wide pore throats prevent mudcake formation, and therefore mud invades the vugs with the solid particulates as well. The mud particles decrease the mud bulk T2 to 50 ms, being misinterpreted as irreducible fluid by the traditional approach, and therefore artificially increasing the water saturation and decreasing permeability estimates. Additionally, mud has hydrogen index below one, decreasing total porosity estimated from the NMR measurements. In this work we present a workflow for calculating petrophysical properties on the Brazilian pre-salt carbonates, integrating laboratory analyses and well logs. The chosen well was drilled with oil-based mud, has an extensive core interval, and a complete wireline suite. First, NMR and Dean-Stark distillation were performed on preserved core plugs with matrix and vugular porosities. These analyses allowed the mud T2 signal identification in vugs by decomposing the T2 spectrum into base functions. Then, the samples' porosity and permeability were measured. The mud signal was propagated to the NMR logs in intervals dominated by vugs. These intervals were defined using photoelectric factor and acoustic image. The T2 spectrum was decomposed into base functions, which allowed identifying and quantifying the irreducible fluids, mud-filled vugs, and free fluid. The last two were interpreted as the pore volume initially filled by formation oil, redefining the saturation and permeability calculations. Finally, the NMR total porosity was corrected using the mud hydrogen index. Our workflow showed a better agreement with laboratory analyses. Estimated average water saturation decreased by 0.18 v/v, average permeability increased by 408 mD, and average porosity increased by 0.02 v/v compared to the traditional interpretation. The importance of correctly interpreting the different types of fluid and pores when analyzing the T2 spectrum becomes evident, especially in complex formations like the Brazilian pre-salt carbonates.