Search

In this paper, we attempted to analyze the experimental results about different CO2 injection rates affect the oil recovery and pressure gradient in low permeability dip reservoirs. For these exper...

As a recent noticeable topic, domain generalization aims to learn a generalizable model on multiple source domains, which is expected to perform well on unseen test domains. Great efforts have been made to learn domain-invariant features by aligning distributions across domains. However, existing works are often designed based on some relaxed conditions which are generally hard to satisfy and fail to realize the desired joint distribution alignment. In this paper, we propose a novel domain generalization method, which originates from an intuitive idea that a domain-invariant classifier can be learned by minimizing the KL-divergence between posterior distributions from different domains. To enhance the generalizability of the learned classifier, we formalize the optimization objective as an expectation computed on the ground-truth marginal distribution. Nevertheless, it also presents two obvious deficiencies, one of which is the side-effect of entropy increase in KL-divergence and the other is the unavailability of ground-truth marginal distributions. For the former, we introduce a term named maximum in-domain likelihood to maintain the discrimination of the learned domain-invariant representation space. For the latter, we approximate the ground-truth marginal distribution with source domains under a reasonable convex hull assumption. Finally, a Constrained Maximum Cross-domain Likelihood (CMCL) optimization problem is deduced, by solving which the joint distributions are naturally aligned. An alternating optimization strategy is carefully designed to approximately solve this optimization problem. Extensive experiments on four standard benchmark datasets, i.e., Digits-DG, PACS, Office-Home and miniDomainNet, highlight the superior performance of our method.

CO2 storage mechanisms in an EOR process in mature reservoirs are measured to determine three types of storage factors, which are introduced into compositional numerical simulation. The hybrid objective function coupli ng the oil recovery factor and the CO2 storage ratio is proposed to optimize the injection and production parameters in CO2 flooding. Three storage factors of the oil and water partition coefficient, the permeability change coefficient and the CO2 retention factor are measured in a laboratory, which is utilized to modify the grid properties of oil, brine, rock in compositional numerical simulation. The restart procedure is automatically adopted to consider these storage mechanisms in CO2 EOR. The bi-objective function of the oil recovery factor and the CO2 storage ratio is used to optimize the injection and production parameters for CO2 EOR, which concludes the design principles on CO2 EOR and storage. The oil and water partition coefficient defined as the ratio of the CO2 solubility in the oil phase and the brine phase is a constant for a specific reservoir condition. The permeability change coefficient caused by the mineral dissolution effect of carbonate water decreases slightly in the early stage and increases gradually with the long term injection. The CO2 retention factor that is induced by the relative permeability hysteresis decreases with the pressure and the permeability. These equivalent treated methods that modify fluids and rock in the real-time are inserted into the procedure of compositional numerical simulation to take into account the storage mechanisms in CO2 EOR. The results show that the effect of the storage mechanisms on EOR is evident. Furthermore, the bi - objective optimization indicates that the injection rate should be reduced largely in the medium and the later stages to control gas channeling as the EOR scenario is focused. And the bottom wellhole pressure of producers should be decreased to the lower level to maximize oil recovery. As the storage scenario is concentrated, the injection rate is required to be slightly controlled. As the producers are shut off, the injection rate must be increased significantly to maximize CO2 storage. The storage mechanisms in the CO2 EOR process have not been understood thoroughly. The methodology of numerical simulation coupling CO2 EOR and storage is not mature, which is still not taken into account in commercial software. The results above provide a way to optimize CO2 EOR and storage simultaneously, which is significant for the large scale storage after CO2 EOR in mature oilfield.

During lean gas reinjection after depletion in high-pressure and low-permeability reservoirs, the fluid composition in the formation changes greatly, the microscopic interaction process among fluid...

This paper focuses on the CO2-EOR in fractured tight oil reservoirs after water-flooding treatment. In previous works, few studies were presented about the spontaneous imbibition experiments of CO2-rich brine at formation pressure. We investigated the influence of CO2 injection on spontaneous imbibition, which is an essential mechanism to improve oil recovery in tight reservoir. In this paper, a laboratory equipment was set up to conduct spontaneous imbibition experiments at formation temperature of 65 °C and pressures of 10–22 MPa on different low-permeability core samples from Nugget, Kentucky, Colton, and Crab-Orchard in the United States. Moreover, we proposed a saturation-based dimensionless time model to scale the spontaneous imbibition and a modified Ma model to fit the oil recovery curves of spontaneous imbibition of CO2-rich brine with double peaks of imbibition rate. The results of quantitative imbibition experiments confirm that both the oil production per unit area and the oil recovery have a ...

Density-driven convection, which can accelerate the dissolution rate of carbon dioxide (CO2) in resident brine, is critical for the long-term fate of the injected CO2 permanence and security of CO2 geological storage. Visualization experiments and pressure–volume–temperature (PVT) testing were conducted to investigate the influence from gravitational convection. For investigate gravitational instabilities and convective diffusion, we designed a Hele-Shaw cell rated to 70 MPa and Rayleigh number of 346 to conduct visualization experiments with the micro-schlieren technique. The average diffusion coefficient and time-dependent values were measured in the PVT experiments. We also calculated the convection parameters, including Rayleigh number and critical onset time, with a series of PVT testing at the temperature ranging 293.15–423.15 K and pressure ranging 14–24 MPa by using the constant-pressure method. Through visualization experiments, we observed convective currents triggered by the density gradient in the gas–liquid interface, which noticeably enhanced the CO2 dissolution rate. The PVT testing confirmed that the diffusion coefficient increased sharply under the influence of the gravitational convection at the early stage and then decreased toward the average diffusion coefficient with time. The PVT testing also demonstrated the Rayleigh number increasing with temperature or pressure increasing under the reservoir conditions. The gravitational convection will be more likely to occur and more rapid with a greater pressure or higher temperature.

An oilfield reservoir over long-term operation may have different petrophysical information, which has a significant impact on oilfield maintenance and finance. Successful oilfield enhanced oil recovery benefits a lot from identifying and analyzing the variations of the critical properties after long-term water-flooding treatments. Since the inspection wells drilled within different development periods contain the core samples that have the petrophysical information at that period, it is necessary to collect and test the samples from different periods to investigate the overall tendency of the petrophysical properties. The samples from four inspection wells, which were drilled in four stages since the very beginning of development, were subjected to in-laboratory core analysis methods to illustrate the variation of some critical parameters in the reservoir. The permeability and porosity variation are revealed clearly by the experimental results. The migration and dissolution of clay minerals play a crucial role in the variation of petrophysical information and pore structure. To quantify the variations above, we applied the multiple linear regression model into our investigation. The dependent variable and all of the predictors in the model come from the experimental results. The quantitative results show the closed correlation between different parameters in the formation. With the development stage moving forward, the weight coefficients for different predictors have multiple trends. The experimental and statistical approach provides a novel understanding of the reservoir properties with the effect of waterflooding treatment. National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [41972098]; Virginia Tech Open Access Subvention Fund The authors thank the National Natural Science Foundation of China (No. 41972098) for their financial support to carry out this research. The authors also thank the support from the Virginia Tech Open Access Subvention Fund. The collaborators from the Statistical Applications & Innovations Group (SAIG) at Virginia Tech will also be thanked for their assistance and comments on the regression model. The insightful and constructive comments of the anonymous reviewers are also gratefully acknowledged.

In China, CO 2 EOR technology is mostly applied in small scale field test after water flooding. Microscopic mechanism of CO 2 EOR at extra-high water cut stage is not clear to us. Based on the characteristics of residual oil after water flooding, three types of ideal core models were designed including dead-end model, shaped island model and cluster model. The microscopic experiments were carried out in high temperature high pressure micro-visualization experiment apparatus of independent research and development. The microscopic mathematic model of CO 2 -water-oil interaction was established. The process of CO 2 dissolution and extraction during CO 2 injection at extra-high water cut stage (water cut>98%) were monitored in the experiments for the first time. The microscopic mechanism of CO 2 EOR on trapped oil droplets and the influence of water barrier on miscible displacement were discussed. The results from experiments and mathematic calculation conclude that water barrier should postpone and complicate miscible process between CO 2 and trapped oil. No matter how thick the water film, CO 2 can diffuse through the water film into the trapped oil. The higher injection pressure is, the faster CO 2 diffuses into oil, the more easily miscibility is achieved. CO 2 miscible displacement contributes to oil recovery largely after water flooding. CO 2 EOR might benefit from the increasing of injection pressure and the full contact time between CO 2 and fluid (water and trapped oil) underground in extra-high water cut reservoir.

Alkaline-Surfactant-Polymer (ASP) flooding is an emerging chemical Enhanced Oil Recovery (EOR) technology which has significantly enhanced oil recovery of Daqing Oilfield. ASP flooding benefits from the synergy effects of alkali, surfactant and polymer to improve both volumetric and displacement efficiencies and meanwhile lower surfactant adsorption. However, ASP flooding also induces some negative formation damage effects such as scaling, adsorption, and mineral dissolution. In this paper, we investigated the formation damage caused during ASP flooding in Block Sanan-5 in Songliao Basin – one of the most productive blocks of Daqing Oilfield in China. It was found that the distribution of formation damage caused by ASP flooding followed flow paths of chemical solutions and was dependent on well locations. The severity of damage varies as distance increases from the near-injection-well area to the near-production-well area. Understanding the effects of well locations on formation damage during ASP flooding could provide more accurate evaluation of formation damage and helped to guide reservoir development strategies. To analyze the well location factor, we collected scaling samples and more than 970 m of core samples from Block Sanan-5 of Daqing Oilfield covering different wells on various flow paths before and after ASP flooding. The changes of some key petrophysical parameters such as porosity and permeability before and after ASP flooding were investigated. A series of experiments, including Scanning Electron Microscopy (SEM), Casting Thin Sections (CTS), X-Ray Diffraction (XRD) and ion analysis of produced water were performed to test properties of core samples. In addition, absorption of different components in the ASP solutions was also measured. Experimental results indicate that the ASP flooding has considerably different influences on different parts of flow paths. After ASP flooding, permeability distribution of core samples exhibits different variability trends from the near-injection-well areas to near-production-well areas. Due to absorption of alkali and polymer, grains migration and scaling of calcium and magnesium, permeability decreases at the near-injection-well area, then increases at an intermediate distance and decreases again at the near-production-well. Moreover, porosity of samples shows a similar tendency with variability of permeability, which is interpreted by the strong mineral corrosion due to high concentration of alkali in the near-wellbore area, while its extent of variation is smaller than permeability.

The Pucheng Oilfield, located in the northeast of Dongpu Depression, is one of the oldest and largest oil fields in the Zhongyuan Oilfield Company of SINOPEC. The CO 2 captured from the Chinese coal industry tail gas has been injected into the Es1 reservoir of Pucheng Oilfield for geological CO 2 storage and enhancing oil recovery. The Es1 formation is a high water-flooded reservoir and consists of sandstones with heterogeneous permeabilities. The Es1 reservoir has a formation temperature of 85 °C and a formation pressures of 15–24 MPa. After injecting CO 2 into reservoir formation, complicated interactions of CO 2 -brine-rock induces the changes of reservoir properties, which have an impact on the performance of geological storage and enhancing oil recovery. The reactions of brine, CO 2 and natural debris samples were studied by conducting laboratory batch experiments, which were conducted under reservoir conditions. During the batch experiments, Scanning Electron Microscopy (SEM) and X-Ray Diffraction (XRD) were used to evaluate the formation damage caused by mineral dissolution. Natural rock batch experiments demonstrated that calcite crystals were obviously dissolved and formed secondary minerals. Calcite content and feldspar content reduced, clay minerals content increased, and quartz content changed only slightly. Based on chemical equilibrium theory, mineral dissolution and deposition were discussed. Flooding experiments under various reservoir conditions were conducted to dynamically evaluate the influence of mineral dissolution and mobilization on formation permeability. The result demonstrated that minerals can migrate in sandstone with any permeability. For low permeability cores, the mineral secondary migration will cause the blocking of a porous medium, even though the flow rate is notably low. For medium permeability cores, the permeability firstly changes little due to the blocking, which is caused by mineral migration, can be canceled via minerals dissolution, carrying the plugging minerals out of the sandstone. Sufficient dissolution effect will weaken the blocking to enhance the permeability. For high permeability cores, it is easier to carry the mobilized minerals out of sandstone, and the mineral mobilization can first decrease the permeability and then increase it. Considering all of the permeabilities, mineral dissolution and mobilization can enhance reservoir heterogeneity.

Liquid repellent surfaces are being promisingly applied in industry and our daily lives. Herein we report a facile and effective sol–gel method for fabricating hybrid coatings with highly omniphobic and self-cleaning properties. The fluorinated hybrid nanocomposite was synthesized via one-step hydrolytic condensation of a nanosilica sol, methyltriethoxysilane (MTES) and 3-[(perfluorohexyl sulfonyl) amino] propyltriethoxysilane (HFTES). The solvent mixture of water and 2-propanol surrounding the hydrophobic nanosilica is a key factor in the control of nanoparticle aggregation, which leads to the formation of a multi-scale roughness surface with different wettabilities. The fluorinated nano-sol can be easily coated on various hard and soft substrates by spraying or dipping methods, endowing the substrate with omniphobicity to different organic liquids and biofoulings especially solidified egg white. Furthermore, the designed coating shows excellent self-cleaning and anti-adhesion properties in various harsh environments such as high temperature, acid and alkaline treatment and oil contamination. Owing to the facile method and its remarkable omniphobic abilities, the fluorinated hybrid coatings can be expected to have potential industry applications in a material system requiring robust antifouling, protein resistance and self-cleaning functions.

A kind of starch graft copolymer (SGC) water shut off agent, a novel, efficient and economic crude macromolecular plugging agent, has been applied to some oilfields. In this paper, a three dimensional physical model of horizontal well was designed for the water shutoff with an edge water simulating the edge water driving reservoir and with a higher permeability band imitating the preferential paths. Using the model and X-ray computed tomography (CT), the effect of blocking the water channeling paths on the model taken by starch strong gels blocking agent was tested. The result showed that the agent was selective and got a good effect. Besides, a simple Computer Modeling Group (CMG) model was used to deduce the development process of the water breakthrough, channeling, and flooding, so as to explain previously what had happened in the physical model. Finally, compared to field experiment, the advance of the model lies in that it can be used in performance appreciation of water shut off agent well.

A superhydrophobic surface with anti-icing performance has been the focus of research, but few studies have reported the effective and low cost strategy that met the requirements under overcooled conditions. In this article, the fluorinated sol-gel colloid coatings were simply prepared via hydrolytic condensation of nanosilica sol, methyltriethoxysilane (MTES) and 3-[(perfluorohexylsulfonyl)amino]propyltriethoxysilane (HFTES). The multi scale morphology and chemical composition of the artificial surfaces were characterized by scanning electron microscopy (SEM), atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS). The influence of the surface roughness structure and fluorinated groups on the wettability and freezing delay time of the colloid surface under overcooled conditions were explored. As the HFTES content was higher than 6 wt%, the prepared colloid surface showed excellent superhydropobicity with a contact angle (CA) of about 166° at room temperature. The CA gradually reduced with the decrease of the temperature. Only the samples with high HFTES contents (above 30 wt%) exhibited special superhydrophobic and anti-icing properties under freeze temperature. Besides the surface roughness structure, the high fluoride enrichment on the surface plays a major role in the superhydrophobic and anti-icing properties under overcooled conditions.