INSIM-BHP: A physics-based data-driven reservoir model for history matching and forecasting with bottomhole pressure and production rate data under waterflooding

Abstract

In this work, we present a new physics-based data-driven interwell three-dimensional (3D) numerical simulation model with front tracking. It provides accurate and computationally fast calculation of bottomhole pressures (BHPs) for history matching, forecasting, and well-control optimization for a 3D reservoir with slanted wells with sectionally perforated intervals under waterflooding. The new model is referred to as INSIM-BHP. The previous versions of INSIM-based models cannot accurately generate BHPs since they use the upstream saturation (or upstream weighting) to obtain the transmissibility of each connective unit and neglected the changes of total mobility and the gravitational effect while water front propagates. INSIM-BHP implements harmonic average transmissibility and fine-scale gravity term calculations to accurately represents flow between a pair of connected nodes by making use of fine-scale saturation distribution calculated by front tracking. Besides, unlike the previous versions of INSIM-based models, it approximates oil reservoirs associated with limited volume aquifer drive systems. The accuracy of the BHPs generated with INSIM-BHP is validated by comparison with corresponding results from a high-fidelity commercial simulator (Eclipse 100™1) considering two synthetic T-shaped reservoir examples with complex gravitation effects with fine grids and small-time steps. Results show that INSIM-BHP model provides accurate BHPs that agree quite well with the corresponding BHPs computed from Eclipse 100 (less than or equal to 2 psi for these two synthetic examples considered in this study). We test INSIM-BHP by history matching three different sets of observed data types for a synthetic 3D channelized reservoir; (1) well oil production rate only; (2) BHP only; (3) oil production rate and BHP simultaneously. Results show that INSIM-BHP with harmonic average transmissibility and fine-scale gravitational term provides a better history matching and prediction ability for well oil production rate and BHPs performance. We have validated our aquifer modeling approach used in INSIM-BHP with an Eclipse's Fetkovich-based aquifer model and presented an application of INSIM-BHP to history match and predict oil production rate and BHP from Brugge model, a field scale synthetic reservoir model.