Real-Time Interpretation Model of Reservoir Characteristics While Underbalanced Drilling Based on UKF

Abstract

This study presents a novel interpretation model for reservoir characteristics while underbalanced drilling (UBD), by incorporating an unscented Kalman filter (UKF) algorithm in a three-phase variable mass flow model of oil, gas, and liquid. In the model, the measurement parameters are simplified to bottomhole pressure and liquid outlet flow, for decreasing the amount of the computation and time. By taking into account real-time measurements, the permeability and reservoir pressure along the well can be continuously updated. Three cases including single-parameter and double-parameter estimations have been simulated, and the performance is tested against the extended Kalman filter (EKF). The results show that single-parameter estimation of reservoir permeability or pressure achieves superior performance. The filtered values of bottomhole pressure and outlet flow trace the measured values in real time. When a new section of a reservoir is opened, the estimated reservoir permeability or pressure can always be quickly and accurately returned to its true value. However, it is not possible for the double-parameter estimation to obtain good results; its interpretation accuracy is low. UKF is superior to EKF in both estimation accuracy and convergence speed, which further illustrates the superiority and accuracy of the novel interpretation model based on UKF. Benefits from this model are seen in accurate bottomhole pressure and reservoir characteristic predictions, which are of major importance for safety and economic reasons during UBD and follow-up completion operations.

An Improved Knothe Time Function Model in the Prediction of Ground Mining Subsidence by Using the Kalman Filter Method

“…These factors can also be used in predicting the mining progress. Moreover, there are other Kalman filter methods, such as extended Kalman filter [32,33], unscented Kalman filter [34][35][36], and total Kalman filter [37,38]. These methods can be widely used in prediction models.…”