Carbonated water injection (CWI) is a modified CO2 flooding technique for enhanced oil recovery, which takes advantage of both CO2 flooding and water injection and has attracted much attention recently. However, the dedicated research so far has focused heavily on conventional reservoirs. The objective of this research is to investigate the performance of CWI in a tight oil reservoir. First, a set of well-designed multiple contact tests were conducted to simulate the dynamic mass transfer process of fresh carbonated water (CW) to live crude oil. In each test, CW was brought into contact with live crude oil in a high-temperature and high-pressure PVT cell. Pressure changes during the test were observed and recorded. After equilibrium, all the transferred CW was taken out of the cell and the swelled oil proceeded to the next contact. The volumes of water and liberated gas were measured. Then, the oil swelling factor could be calculated, which would verify the existence of the moving interface between CW and live crude oil. It was observed that the system pressure built up immediately after CW was mixed with live crude oil in the PVT cell. For the first contact, the equilibrium pressure increased by 6.46 MPa, and the equilibrium pressure increased by 2.16 MPa at the last contact. This result indicates a strong interaction between CW and live crude oil, which is beneficial to maintain reservoir pressure. Because a large amount of CO2 from CW was transferred to the live crude oil, the swelling factor of 1.26 was obtained at the end of the tests. The diffusion of CO2 into the live crude oil also leads to subsequent oil viscosity reduction. In addition, a series of coreflood experiments under real reservoir conditions were carried out to evaluate the performance of CWI for improving oil recovery in core samples from a tight sandstone reservoir. Coreflood results showed that both secondary and tertiary CWI recovered additional oil compared to water flooding. Finally, a significant amount of CO2 was stored in the cores. Our experimental results clearly indicate the potential of CWI for improving oil recovery and CO2 storage capacity in tight oil reservoirs.

This paper presents an analytical solution of Buckley-Leverett equation for gas flooding with constant-pressure boundary including the effect of miscibility on the viscosity and relative permeability. First, a relative permeability model and a viscosity model with consideration of miscibility are used to describe the variations of relative permeability and viscosity of oil and gas. Then, based on the fractional-flow theory, the Buckley-Leverett equation for gas flooding with constantpressure boundary including the effect of miscibility is constructed and solved analytically. From the analytical solution, the saturation and pressure profiles, the total volumetric flux and the breakthrough time are determined. To verify the theory, the analytical solution is compared with the numerical solution. The comparison shows that the analytical solution is in reasonable agreement with numerical solution. Through the study on the influential factors, it can be concluded that total volumetric flux is increasing with the increases of permeability and pressure and decrease of gas viscosity. The increase of total volumetric flux accelerates breakthrough of the injected gas. Furthermore, with the pressure increase, there are remarkable reduction in residual oil saturation and improvement of relative permeability, resulting in higher gas saturation and oil displacement efficiency. The analytical solution presented in this paper provides guidance on analyzing the distribution of saturation and pressure profiles, predicting the gas production and oil recovery efficiency of oil well.

The precondition of well testing interpretation is to determine the appropriate well testing model. In numerous attempts in the past, automatic classification and identification of well testing plots have been limited to fully connected neural networks (FCNN). Compared with FCNN, the convolutional neural network (CNN) has a better performance in the domain of image recognition. Utilizing the newly proposed CNN, we develop a new automatic identification approach to evaluate the type of well testing curves. The field data in tight reservoirs such as the Ordos Basin exhibit various well test models. With those models, the corresponding well test curves are chosen as training samples. One-hot encoding, Xavier normal initialization, regularization technique, and Adam algorithm are combined to optimize the established model. The evaluation results show that the CNN has a better result when the ReLU function is used. For the learning rate and dropout rate, the optimized values respectively are 0.005 and 0.4. Meanwhile, when the number of training samples was greater than 2000, the performance of the established CNN tended to be stable. Compared with the FCNN of similar structure, the CNN is more suitable for classification of well testing plots. What is more, the practical application shows that the CNN can successfully classify 21 of the 25 cases.

No abstract

No abstract