Enhanced oil recovery (EOR) is a proven method to increase oil production from the brown fields. One of the efficient EOR methods is injecting surfactants to release the trapped oil. However, few unconsolidated behaviours were observed in both field and laboratory practice. In this study, a new framework was adapted to evaluate the continuous surfactant flooding (CSF) in Bentiu reservoir. The study aims to quantify the expected range of the oil production, recovery factor and residual oil saturation (S_or). The motivation came from the oil demand in Sudan and the insufficient cores. The framework adopted in the study includes numerical simulation modelling and proxy modelling. Thirty-six cores obtained from the field were revised and grouped into five main groups. The interfacial tension (IFT) data were obtained experimentally. The CSF sensitivity study was developed by combining different experimental design sets to generate the proxy model. The CSF numerical simulation results showed around 30% additional oil recovery compared to waterflooding and approximately oil production between (20–30) cm³. The generated proxy model extrapolated the results with concerning lower ranges of the input and showed an average P50 of oil production and recovery of 74% and 17 cm³, respectively. Overall, the performance of CSF remained beneficial in vast range of input. Moreover, the generated proxy model gave an insight on the complexity of the interrelationship between the input factors and the observants with a qualitative prospective factors. Yet, the results confirmed the applicability of CSF in core scale with an insight for field scale application.

提高采收率（EOR）是提高棕地采油量的一种行之有效的方法。一种有效的EOR方法是注入表面活性剂以释放被困的油。但是，在野外和实验室实践中都很少观察到松散的行为。在这项研究中，采用了一个新的框架来评估本提乌水库的连续表面活性剂驱（CSF）。该研究旨在量化预期的采油范围，采收率和剩余油饱和度（S_或）。动机来自苏丹的石油需求和核心不足。研究中采用的框架包括数值模拟建模和代理建模。修订了从该领域获得的36个核心，并将其分为五个主要组。界面张力（IFT）数据是通过实验获得的。通过结合不同的实验设计集来生成CSF敏感性研究，以生成代理模型。CSF数值模拟结果表明，与注水相比，采油量提高了约30％，产油量在（20-30）cm ^3之间。生成的代理模型以较低的输入范围外推结果，结果表明采油和采收率的平均P50为74％和17 cm ³， 分别。总体而言，CSF的性能在广泛的投入范围内仍然是有益的。此外，生成的代理模型提供了对输入因子与具有定性预期因子的观察者之间相互关系的复杂性的见解。然而，结果证实了CSF在核心规模上的适用性，并具有对现场规模应用的见识。