Abstract
In this work, we applied a Bayesian time-lapse (4D) seismic inversion to a deep-water heavy oil field located in the Campos Basin (Brazil). In this procedure, the differences of seismic amplitudes of baseline and repeated surveys are the input to calculate the changes in acoustic impedances. To ensure the quality of the results, the parametrization of the inversion was performed by analyzing the anomalies resolution and signal-to-noise ratio with different values of input parameters, and the estimates are crosschecked with the results of a petroelastic modeling applied to the dynamic model. Results show a relevant improvement in the vertical resolution of estimated acoustic impedances in comparison to the amplitude data in a thin reservoir (with thickness ranging from 15 to 45 m). This improvement was possible because of the high-quality of acquired seismic amplitudes (permanent reservoir monitoring) and because of the data variability that is controlled in the Bayesian 4D inversion algorithm. The high resolution of Bayesian inverted results enabled a more detailed vertical zoning of the reservoir, revealing important geological features, such as channels in the main interval and below the reservoir's base, and improved the understanding of the fluid movement over time, enabling the analysis of the injection efficiency and by-passed oil zones as well as the aquifer movement. This information is important to update the static model, by improving the definition of geological features, and to calibrate dynamic models, in the future, by adding the inverted acoustic impedance maps into the data assimilation procedure.
NSTL
Elsevier