Evaluating nonlinear Kalman filters for parameter estimation in reservoirs during petroleum well drilling

Abstract

When drilling into a petroleum reservoir, the geological properties of the reservoir might require that the well pressure is kept slightly below the reservoir pore pressure. This leads to a migration of reservoir fluids from the reservoir into the oil well. The amount of reservoir fluids flowing into the well is dependent of the reservoir parameter named production index. This paper evaluates the performance of the extended Kalman filter, the ensemble Kalman filter and the unscented Kalman filter to estimate the production index. The comparison is based on a nonlinear two-phase fluid flow model. The results show that all three filters are capable of identifying the reservoir production index parameter, but that the unscented Kalman filter gives the best performance both when evaluating the least squares deviation from the true value and calculation resource requirements.