Ensemble-Based Reservoir Simulation with 4D Seismic for Assessment of Infill Opportunities: Case Study in a Deepwater Gulf of Mexico Field

Abstract We present the ensemble-based reservoir modeling and uncertainty study in Miocene reservoirs of a deepwater Gulf of Mexico (GoM) field with extensive production history. The purpose of the simulation study was to identify and characterize infill opportunities to arrest the field decline and maximize asset value. Static and dynamic uncertainty parameters were identified and built into the integrated geological and dynamic modeling workflow, where the full range of geologically consistent reservoir models are history matched and used as a basis for production forecasts. The ensemble modeling workflow (commercial software) includes: i) the creation of an initial ensemble of cases (e.g. 100 simulation cases) based on the initial set and range of uncertainty parameters; ii) defining the objective functions and post-processing/visualizing the results of initial ensemble modeling to ensure coverage of the objective functions; iii) defining the uncertainty variables to be optimized in the history matching process; and iv) running the iterations of the history matching ensembles. The ensembles with the best history match quality were then selected for forecasting and testing infill locations. In addition to reservoir pressure and production data, 4D seismic was also used as a history match objective function. Matching the 4D seismic with hydrocarbon pore thickness (HPT) difference was essential to constrain the range of aquifer size and oil in place as sources of reservoir energy. To achieve a good history match quality, first an approximate match was obtained for 4D seismic HPT difference by testing sensitivity to aquifer size range using a small number of ensembles (10-20 ensembles). After narrowing down the aquifer size range, the ensemble modeling history matching workflow was used to match production and pressure data while fine tuning the 4D - HPT difference match. Initial productivity index (PI) and PI degradation assumptions were the most impactful subsurface parameters on incremental Estimated Ultimate Recovery (EUR) for the tested infill well locations. The integrated geo-dynamic modeling workflow enabled a consistent assessment of the infill bottomhole locations (selected based on seismic attributes) using the simulator calculated initial PI for infills. This was achieved by: i) conditioning the geostatistical facies and property modeling workflow to seismic amplitudes, and ii) ensuring the simulator calculated well PI matched well test PI for the existing wells. The compelling results of the uncertainty evaluation influenced the decision to progress the maturation of a new infill well by the asset co-owners. We demonstrate the application of a fully integrated workflow for geological and dynamic modeling using ensemble-based reservoir simulation which uses 4D seismic to calibrate HPT difference as one of the objective functions to constrain the aquifer size uncertainty. The integrated reservoir modeling approach results in a consistent assessment of infill bottomhole locations where estimates of the initial well PI is driven by seismic attributes.