Optimization of water injection rates in multi-layer oil reservoir under strategies of balanced displacement

Multi-layer heterogeneous oil reservoirs at the ultra-high water-cut stage commonly suffer from severe interlayer conflicts, inefficient circulation of injected water, and uneven reserve mobilization, which together restrict further recovery. To improve zonal water-injection regulation under these conditions, a dynamic control framework was developed by integrating an inter-well connected element model with a balanced-displacement water-injection optimization model. Production history matching was first used to invert the connected volumes between injectors and producers; an IMPES-type connected-element simulator and SPSA-based parameter inversion were used to update the model. An exponential function was then fitted to the oil–water relative permeability ratio and coupled with material balance to derive the critical condition for maximum water-flooding efficiency. The method was applied to a four-layer block in the Daqing Oilfield containing four oil production wells and one water injection wells with an average water cut of 95.2%. The history-matching accuracies for daily oil production and water cut were approximately 98%, and the prediction accuracies were approximately 96%. The fitted exponential relative-permeability relationships produced R² values of 0.992–0.999. After water-injection adjustment over 1.5 years, the average water cut decreased by 0.3%, the average daily oil production increased by 0.3 m³/d, and cumulative oil production increased by 131.1 m³ compared with the unoptimized case. This framework provides a quantitative approach for layer-level water injection optimization in heterogeneous reservoirs. Unlike previous planar or empirically based allocation methods, the proposed model provides a quantitative layer-level relationship for balanced recovery in heterogeneous reservoirs.