Research and Application of Injection-Production Linkage Coupling Control Technology

The main oilfields in China have generally entered the ultra-high water cut period. The number of wells using separate-layer injection-production strings is increasing, and it has become an important means to alleviate horizontal and vertical contradictions of water flooding. Traditionally, the control strategies of layered water injection wells and layered oil production wells are designed independently, without considering the linkage relationship between injection and production wells, resulting in insufficient technological potential. For this reason, the injection-production linkage coupling control technology is proposed, which fully considers the dynamic response relationship between injection-production wells in different layers, and realizes fine optimization of injection-production parameters in each layer. Divide the injection-production wells into different injection-production units on the plane, use the Buckley-Leverett oil-water two-phase flooding theory to calculate the water driving state in the unit, and build a proxy model. Combined with the intelligent optimization algorithm, the proxy model is corrected, and the corrected model is further invoked to predict and screen multiple sets of injection-production control schemes until the optimal stratified injection-production parameters are output. Compared with the reservoir numerical simulation, the data-driven model formed based on reservoir engineering has an index calculation coincidence rate of 95.3%, and the calculation speed can be increased by more than 5 times. Through the integration with the intelligent optimization algorithm, it can make full use of massive historical production data quickly and use the actual block data for trial calculation. The fitting degree of key indicators can reach more than 85%, and the distribution of remaining saturation can be output in time. With oil accumulation, NPV, balanced displacement, etc. as the target, layered injection-production parameters are the adjustment objects, and the revised data-driven model is used to quickly predict, analyze and compare massive measure schemes, further automatic iterative optimization, and the final output is the best plan. The research results have been applied to the injection-production control of several water flooding reservoirs in Shengli oilfield. After the application of the typical block B2, the daily oil increase reached 14.7%, the water cut decreased by 0.2%, the oil recovery rate increased by 0.25%, and the recovery rate increased by 0.7%. The development effect is improved significantly.