Developing a CRM-IP-Based Digital Twin for the Optimization of Mature Waterflooding Performance: Application to the Haoud Berkaoui Oil Field

Abstract

The Capacitance–Resistance Model with Injection and Production constraints (CRM-IP) is a data-driven reservoir management approach based on analogies between reservoir fluid flow dynamics and RC electrical circuits. In this analogy, the reservoir pore volume behaves similarly to electrical capacitance, whereas fluid transmissibility through porous media is analogous to electrical conductivity. These similarities enable the application of simplified dynamic modeling concepts to characterize fluid flow behavior in heterogeneous petroleum reservoirs. The Haoud Berkaoui oil field is characterized by strong reservoir heterogeneity, permeability variations, premature water breakthrough and complex injector–producer interactions. These challenges reduce the efficiency of conventional reservoir simulation workflows, particularly in operational environments that require rapid decision-making, repeated recalibration and continuous monitoring. The main objective of this study was to develop a CRM-IP-based digital twin framework for robust waterflooding optimization under operational uncertainties in the Haoud Berkaoui field. The proposed framework integrates production and injection data, interwell connectivity analysis, dynamic model updating and closed-loop optimization to improve reservoir management and production performance. The CRM-IP model was calibrated using historical field data from selected injection and production wells and incorporated into a digital twin architecture capable of continuously updating reservoir behavior predictions. The framework also integrates operational and economic constraints through injection rate optimization and Net Present Value (NPV) evaluation. The results demonstrate that the proposed CRM-IP digital twin provides a computationally efficient and operationally adaptive alternative to conventional full-physics reservoir simulation approaches. In addition, the framework improves injector–producer connectivity analysis, supports near-real-time decision-making and enables intelligent and adaptive reservoir management in mature waterflooded reservoirs under uncertainty.