Proceedings · 2013 · 35 citations · 12 references
EngineeringEarth SciencesPetroleum EngineeringPetrophysicsReservoir EngineeringReservoir SimulationData AssimilationHistory MatchingPetrophysical AnalysisPetrophysical EvaluationReservoir ArchitectureReservoir ForecastingChemical EorChemical Eor ScreeningData Integration For Reservoir ModelsEnsemble Kalman FilterGeological ModellingIntegration Of Core DataPetrophysical ModelingReservoir Parameter Estimation
The paper focuses on a Chemical EOR study for a mature field. The field was selected due to its volume in place and good petrophysical properties. Indeed, the preliminary screening gave indication that polymer injection could be a promising EOR technique. New core data, SCAL and PLT were acquired and a high resolution model of the pilot area was built to integrate such new data and to properly capture the behaviour of the chemicals. The sector modelling was challenging due to the complexity of the history match and polymer injection optimization. The field has been producing for 60 years. Moreover, due to the complex structural settings, the sector model is not completely isolated from the full field model and dummy wells were introduced to mimic the flow interaction with the rest of the reservoir. A Computer Assisted History Matching (CAHM) was carried out by the means of the Ensemble Kalman Filter (EnKF). The EnKF is a Monte-Carlo method that automatically updates an ensemble of reservoir models by production data integration. The EnKF is capable of providing a set of matched models that preserve the geological coherence which can be used to quantify uncertainty in forecast production. In this paper, we present the application of the EnKF to history match the sector model and the consequent optimization for polymer injection. EnKF was used to calibrate petrophysical properties, relative permeability and faults transmissibility integrating measurements, shut-in pressures and rates, of 14 wells including the dummy wells. The final output is a set of 100 alternative models that properly match production data which were used to set up and optimize the forecast development strategy through polymer injection. This application provides evidence that the EnKF is effective and efficient for history matching. Moreover, dealing with multiple models put the basis for a conscious estimation of future production and a more realiable risk evaluation on EOR strategy.
12
The Ensemble Kalman Filter in Reservoir Engineering--a Review
S. I. Aanonsen, Geir Nævdal, Dean S. Oliver et al. · SPE Journal · 2009 · 800 citations