Facies-driven, seismic inversion-based property modeling, for flow simulation of Delhi Oil Field, Louisiana

Abstract

Delhi Field is a Cretaceous-age giant oil field on the flank of the Monroe Uplift in northeast Louisiana. Delhi field was acquired by Denbury Resources in 2006 as a candidate for CO2 flooding. Reservoir Characterization Project phase XIII involves 4D and multi-component seismic imaging of the field to monitor the CO2 flood. The goal of this research is to develop a robust structural and stratigraphic framework for property modeling, to be used for flow-simulation of the CO2 flood. The Late Cretaceous volcanic-cored Monroe Uplift is interpreted to influence sedimentation of transgressive Tuscaloosa sandstones. The log signatures of oil and CO2 are defined based on log data. K-means clustering shows improvement in facies delineation using log data after petrophysical correction. Thin-section analysis reveals excellent porosity and permeability in both marine and fluvial reservoirs. Mineralogy from point-counting and XRD analysis are integrated with cluster facies logs, core, and seismic data to interpret the depositional environments of each cluster facies. An investigation of bandwidth-extended seismic data demonstrates the ability to detect thin Tuscaloosa sandstone bodies. A method of time-tying of k-means cluster facies logs to the bandwidth-extended seismic data enables stratigraphic interpretation of the seismic dataset. Three transgressive parasequences are identified in the Tuscaloosa interval in the study area. Sediment preservation is partially controlled by inherited physiography, sediment supply, topographic gradient, and faulting. Shoreline parallel marine sandstone bodies are interpreted in the direction N65E, fluvial sandstone bodies are interpreted in the dip direction S25E. Gassmann fluid substitution evidences a facies-dependent fluid response to oil. The facies dependencies are honored for property modeling of saturation, porosity, and permeability. Seismic inversion for Vp:Vs ratio is used to estimate bulk oil volume. The bulk oil volume model is used to fluid substitute inverted AI to 100% brine to improve the seismic prediction of porosity. Facies-based transforms from porosity to permeability are used to create a permeability model. Property models show agreement with 4D seismic imaging of the CO2 flood - observed CO2 flow corridors are predicted by the permeability model.