Collaborative optimization design method for water injection layer division and injection-production fluid allocation in high water cut reservoirs

Fine layered water injection is a crucial method to enhance oil recovery in high water cut reservoirs. However， the design of water injection layer division and injection-production fluid allocation often lacks coordination， limiting the full realization of process synergy. To address this， under the condition of an unknown optimal number of water injection layers， a mathematical model of fine layered water injection optimization was developed. This model aimed to maximize economic net present value while considering multiple engineering and geological constraints， integrating water injection layer division and injection-production fluid allocation into a unified framework. By leveraging the Dung Beetle Optimizer， a collaborative optimization design method for water injection layer division and injection-production fluid allocation in high water cut reservoirs was proposed， enabling the integrated optimization design of key parameters for fine layered water injection. Using the Egg model as an example， this study compared the effects of three approaches： standalone optimization of injection-production fluid allocation， standalone optimization of layered water injection parameters， and integrated fine optimization. The results indicate that for high water cut reservoirs with severe interlayer physical property interference， the integrated fine optimization approach significantly outperforms the other standalone methods by optimizing layer division of layered water injection wells， fluid allocation， and injection-production fluid allocation of other wells. Through automatic layer division and reorganization of water injection layers， coupled with intelligent matching and adjustment of injection-production parameters， the approach effectively alleviates interlayer conflicts during the high water cut stage， optimizes the flow fields， and achieves balanced displacement across layers. Over a 5-year production forecast， this approach improves the recovery factor by 1.08 percentage points compared to the pre-optimization scenario.