Non-Parametric Bayesian Networks for Parameter Estimation in Reservoir Engineering

The ultimate goal in reservoir engineering is to optimize hydrocarbon recovery from a reservoir. To achieve the goal, good knowledge of the subsurface properties is crucial. One of these properties is the permeability. Ensemble Kalman Filter (EnKF) is the most common tool used to deal with this situation. However, it is not the only way. Recently, a research on a more general approach based on a dynamic Bayesian network using the Non Parametric Bayesian Networks (NPBN) has been initiated. This research, which uses a twin experiment, indicates the NPBN approach to be a promising alternative to EnKF as a tool to tackle history matching problem. Analysis of the spatial correlation of the permeability estimates from both methods reveals puzzling behavior. For the same pair of cell, the EnKF method tends to have higher correlation than the NPBN method. Two pairs of cells from the NPBN estimate even have completely negative correlation. Nevertheless, the NPBN method is still in its infancy and further investigations and improvements still need to be performed. However, based on the obtained results, there are even more reasons to believe it as a promising approach in tackling history matching problem in reservoir engineering.