Uncertainty Quantification in Reservoir History Matching Using the Ensemble Smoother

Ensemble-based methods have been widely used in uncertainty quantification, particularly, in reservoir history matching. The search for a more robust method which holds high nonlinear problems is the focus for this area. The Ensemble Kalman Filter (EnKF) is a popular tool for these problems, but studies have noticed uncertainty in the results of the final ensemble, high dependent on the initial ensemble. The Ensemble Smoother (ES) is an alternative, with an easier impletation and low computational cost. However, it presents the same problem as the EnKF. The Ensemble Smoother with Multiple Data Assimilation (ES-MDA) seems to be a good alternative to these ensemble-based methods, once it assimilates tha same data multiple times. In this work, we analyze the efficiency of the Ensemble Smoother and the Ensemble Smoother with multiple data assimilation in a reservoir histoy matching of a turbidite model with 3 layers, considering permeability estimation and data mismatch.