A review of mechanisms and ML-based research on factors affecting spontaneous imbibition of surfactant

Spontaneous imbibition using surfactant has become an effective method for enhanced oil recovery (EOR) in tight reservoirs. In order to guide EOR in the oilfield, it is of great significance to clarify the main controlling factors of spontaneous imbibition and its influence mechanism. This paper focuses on governing factors selecting of spontaneous imbibition of surfactant. First, the common experimental methods of spontaneous imbibition were introduced. Second, the main governing factors of spontaneous imbibition of surfactants and their relative mechanism were summarized. Subsequently, basing on pre-existing experimental data, machine learning (ML) models were devised for a quantitative analysis aimed at elucidating the significance of factors influencing the spontaneous imbibition of surfactants. These models encompassed several algorithmic frameworks of ensemble learning, including Random forest (RF), Adaptive boosting (AdaBoost), Gradient boosting decision tree (GBDT), and Extreme gradient boosting (XGBoost). Besides, the importance analysis of the factors affecting spontaneous imbibition and the prediction of oil recovery by imbibition were realized. Finally, conclusions and outlooks are suggested. In the review, the main governing factors of imbibition. Spontaneous of surfactant were obtained, including permeability, viscosity of crude oil, wettability and length of the core. Oil-water interfacial tension (IFT) has less influence on imbibition oil recovery than other factors. Our research proposes that more quantitative indicators of the factors affecting spontaneous imbibition, such as overburden pressure on rocks and emulsification capacity, should be taken into account by ML model. Additionally, the model should be built based on rocks with specific physical properties.