Predicting the productivity of multistage fractured horizontal wells plays an important role in exploiting unconventional resources. In recent years, machine learning (ML) models have emerged as a new approach for such studies. However, the scarcity of sufficient real data for model training often leads to imprecise predictions, even though the models trained with real data better characterize geological and engineering features. To tackle this issue, we propose an ML model that can obtain reliable results even with a small amount of data samples. Our model integrates the synthetic minority oversampling technique (SMOTE) to expand the data volume, the support vector machine (SVM) for model training, and the particle swarm optimization (PSO) algorithm for optimizing hyperparameters. To enhance the model performance, we conduct feature fusion and dimensionality reduction. Additionally, we examine the influences of different sample sizes and ML models for training. The proposed model demonstrates higher prediction accuracy and generalization ability, achieving a predicted R² value of up to 0.9 for the test set, compared to the traditional ML techniques with an R² of 0.13. This model accurately predicts the production of fractured horizontal wells even with limited samples, supplying an efficient tool for optimizing the production of unconventional resources. Importantly, the model holds the potential applicability to address similar challenges in other fields constrained by scarce data samples.

With the global energy consumption on the rise and the gradual decline in conventional oil production, unconventional reservoirs have received considerable attention in the last decade. However, due to the unique physical properties and a large number of micro/nanopores in unconventional reservoirs, fluid flow in these reservoirs is considerably different from conventional ones. Therefore, it is highly important to conduct research on elucidating these fluid flow mechanisms. Furthermore, to avoid problems associated with the rapid production decline and low recovery efficiency in such reservoirs, an enhanced oil recovery technology that can efficiently and economically develop unconventional reservoirs is urgently required. This paper systematically summarizes the current research on flow mechanisms, including capillary imbibition, molecular-scale fluid flow and productivity prediction in unconventional reservoirs, and introduces the enhanced oil recovery and application status of hydraulic fracturing assisted oil displacement technology, along with a brief analysis of their advantages and disadvantages. This study is intended to serve a reference for the efficient development of unconventional reservoirs.

Water flooding is a commonly used technique to improve oil recovery, although the amount of oil left in reservoirs after the procedure is still significant. Gas displacement after water flooding is an effective way to recover residual oil, but the occurrence state and flow principles of multiphase fluid after gas injection are still ambiguous. Therefore, the gas displacement process after water flooding should be studied on the pore scale to provide a basis for formulating a reasonable gas injection program. Most of the current pore-scale studies focus on two-phase flow, while simulations that account for the influence of oil-gas miscibility and injected water are seldom reported. In this work, the multi-component multi-phase Shan-Chen lattice Boltzmann model is used to simulate the gas displacement after water flooding in a porous medium, and the effects of injected water, viscosity ratio, pore structure, and miscibility are analyzed. It is established that the injected water will cause gas flow path variations and lead to premature gas channeling. Under the impact of capillary pressure, the water retained in the porous medium during the water flooding stage further imbibes into the tiny pores during gas injection and displaces the remaining oil. When miscibility is considered, the oil-gas interface disappears, eliminating the influence of the capillary effect on the fluid flow and enabling the recovery of remaining oil at the corner. This study sheds light on the gas displacement mechanisms after water flooding from the pore-scale perspective and provides a potential avenue for improving oil recovery.

Different from the conventional oil reservoirs, the primary storage space of shale is micro/nano pore networks. Moreover, the multiscale and multi-minerals characteristics of shale also attract increasing attentions from researchers. In this work, the advances and challenges in the development of shale oil are summarized from following aspects: phase behavior, flow mechanisms, reservoir numerical simulation and production optimization. The phase behavior of fluids confined in shale nanopores are discussed on the basis of theoretical calculations, experiments, and molecular simulations. The fluid transport mechanisms through shale matrix are analyzed in terms of molecular dynamics, pore scale simulations, and experimental studies. The methods employed in fracture propagation simulation and production optimization of shale oil are also introduced. Clarifying the problems of current research and the need for future studies are conducive to promoting the scientific and effective development of shale oil resources.