To gain an in-depth understanding of the microscopic flow characteristics and residual oil migration during viscoelastic particle flooding， this study established a pore-scale numerical model for viscoelastic particle-water-oil three-phase flow based on digital core. By tracking the dynamic evolution of the oil and water distribution field and analyzing the spatial structure changes of flow paths， the characteristics and evolution laws of flow paths during viscoelastic particle flooding were systematically revealed. Simulation results indicate that compared with conventional water flooding， dynamically branching flow paths are formed in the porous medium due to the plugging effect on the pore channels or increasing the flow resistance of viscoelastic particles in the pore channels. The evolution of these flow paths exhibits three distinct stages： path reconstruction， where the number of flow paths rapidly decreases from 60 to 25； path optimization， where the standard deviation of the weight increases from 5.00 to 10.00， accompanied by a 22% reduction in residual oil volume fraction； and dynamic equilibrium， where the number of flow paths remains around 15， and the residual oil volume fraction further decreases from 0.26 to 0.13. Quantitative analysis further shows that the resistance regulation and pressure fluctuations during viscoelastic particle flooding prompt fluids to enter previously unmobilized regions， improving displacement efficiency. The volume fraction of mobilized zones increases to 0.50-0.70， and the overall oil recovery is increased by 35% compared to the end of water flooding. The research reveals the dynamic evolution characteristics of the flow path during the viscoelastic particle flooding and clarifies the relationship between its microscopic regulatory mechanism and the macroscopic oil recovery efficiency.

In this article, an advanced nine-point (9P) scheme is proposed for solving multiphase flow problems in 2D heterogeneous porous media. The proposed 9P scheme is an extension of the five-point (5P) scheme constructed using the finite analytical method (FAM). As the heterogeneity of porous media increases, the error of the traditional algorithms becomes uncontrollable due to their significant underestimation of nodal transmissibility. However, the transmissibility in the FAM is calculated based on the local analytical solution, and its accuracy is high and does not depend on the strength of heterogeneity. The proposed FAM-9P scheme offers two distinct advantages. Compared to the traditional 9P scheme, it provides much more accurate simulation results, especially for strongly heterogeneous porous media. Additionally, compared to the FAM-5P scheme, it can the alleviate grid orientation effect (GOE) under adverse mobility ratios. In practice, 2 x 2 or 3 x 3 subdivisions for one original grid are recommended for the proposed FAM-9P method, regardless of the strength of heterogeneity.

Since the beginning of the 21st century, the rapid advancements in micro-mechanical systems and bioengineering, along with the extensive exploration of unconventional oil and gas resources, have made fluid flow in micro-nano spaces a prominent research focus. This study investigates the flow characteristics of oil, gas, and water at the micro-nano scale using quartz capillaries and alumina channels, with a particular emphasis on unconventional oil and gas reservoirs. The research led to the development of corresponding models, including a nonlinear seepage model, a non-Newtonian fluid mechanics model, and a neural network prediction model. These models not only elucidate the complex mechanisms of fluid flow at the micro-nano scale but also provide theoretical support for practical applications.

In fractured reservoirs, the capillary pressure is matrix and is usually much greater than in the fracture. This significant difference of capillary pressure induces an interface effect, where the saturation is discontinuous at the matrix–fracture interface. The conventional transfer function between matrix and fracture in the dual porosity model dose not account for this matrix–fracture interface effect and sometimes leads to results violating the physical principle. To improve the calculation of transfer function, an advanced form is proposed with the consideration of the matrix–fracture interface effect. The proposed form is quite simple and can be directly applied in the engineering application of numerical reservoir simulation and also in commercial software such as ECLIPSE and CMG with little modification. The numerical tests indicate a noticeable improvement of the proposed DP model compared to the conventional one.

Jiyang shale oil is a terrestrial self-generated and self-accumulated reservoir with deep burial， low maturity， and high pressure coefficient. Artificial reconstruction spaces with complex flow mechanisms form after large-scale volumetric fracturing， and it isn't easy to obtain accurate physical properties of the reservoir in the stimulated spaces through conventional experiments， which presents new challenges to numerical simulation methods. In view of the unclear internal flow law of shale oil， the reservoir around the wellbore was regarded as a grey box system， and the numerical simulation model of the horizontal well of shale oil was established based on the understanding of physical property and fracturing effect of the reservoir from the previous exploration work. The production law of the horizontal well of shale oil in Jiyang Depression was studied to investigate the internal relationship between oil production and influencing factors. The reservoir around the horizontal well was divided into an artificial main fracture zone， micro-fracture reconstruction zone， and unstimulated matrix zone in the model. By considering fracturing fluid injection into the reservoir， the production dynamics of typical horizontal shale wells of shale oil were fitted， the physical properties of the reservoir were deduced， and the final oil production was predicted. The influence law of the working system and fractures on development indexes were analyzed， and the method for determining the rational well spacing for well group development was further explored， which provided support for the development and design optimization of shale oil in Jiyang Depression.

To address the needs of reservoir development with complex geological conditions in Shengli Oilfield， this paper systematically elaborates on the independent R&D process， core technologies， and application achievements of reservoir numerical simulation software. Through years of dedicated research， Shengli Oilfield has developed a series of proprietary software， including a 3D three-phase black oil model， as well as compositional numerical simulation software for chemical flooding and CO2 flooding. These innovations have effectively resolved key technical challenges such as characterizing flow laws at ultra-high water-cut stages and describing the oil displacement mechanisms of novel chemical flooding systems. The software demonstrates significant technical advantages in simulating differentiated oil-water flow at ultra-high water-cut stages and nonlinear flow in low-permeability reservoirs. Furthermore， by establishing efficient mathematical model discretization and solution algorithms， such as decoupling sequencing， alternating direction， adaptive grids， and algebraic multigrid methods， the software achieves rapid solutions for reservoir models with millions to tens of millions of grid cells， largely meeting the needs of field development planning. However， compared to international commercial reservoir simulation software， Shengli Oilfield’s self-developed numerical reservoir simulation software still exhibits gaps in functional completeness， large-scale parallel computing efficiency， and the integration of emerging technologies. To bridge these gaps， further advancements are required in core reservoir simulation technologies， including new development paradigms for integrating modeling and simulation as well as combining multiple recovery methods； CPU + GPU large-scale parallel computing； the integration and application of emerging technologies based on big data and artificial intelligence. Through continuous technological innovation and functional optimization， Shengli Oilfield’s reservoir numerical simulation software will strengthen its competitiveness， providing robust support for the efficient development of complex reservoirs.

Reservoir numerical simulation technology is an important means of reservoir analysis,and it is a powerful tool for oil-field developers to design development schemes,proceed with dynamic tracking adjustment,and enhance oil recovery of reser-voirs.In order to clarify the future development directions of reservoir numerical simulation technology in Shengli Oilfield,this pa-per reviewed the development process of reservoir numerical simulation application technology and independent intellectual prop-erty software in Shengli Oilfield over the past 60 years and summarized the innovative progress in numerical simulation application technology in many fields since the"13th Five-Year Plan in 2011-2015"period,which is applicable to current development and geo-logical characteristics of Shengli Oilfield.The fields covered fine reservoir description in ultra-high water-cut stage,pressure flood-ing development in low-permeability reservoirs,heterogeneous combination flooding,multiple thermal combination flooding in heavy oil reservoirs,high-pressure CO2 miscible flooding,and integrated development of shale oil by large-scale fracturing.Then,this paper introduced the development and application of several reservoir numerical simulation software with Shengli characteris-tics,involving water flooding,chemical flooding,and micro and intelligent simulation.Finally,it was pointed out that under the current development situation of Shengli Oilfield,the reservoir numerical simulation is facing great challenges in terms of refine-ment,scale,efficiency,and collaboration approach,and the numerical simulation of reservoirs should be further developed to-wards integration,parallelization,and intelligence,so as to provide technical support for improving the quality and efficiency of Shengli Oilfield development.

In this work, we improve the traditional adaptive mesh refinement (AMR) by combining it with the finite analytical method (FAM) to solve the multiphase flow in heterogeneous porous media numerically. The FAM can provide rather accurate internodal transmissibility, and it is employed to improve the accuracy of coarsening and refining processes in AMR. The high performance of the proposed AMR-FAM is indicated through numerical tests for solving the two-phase flow in 2D heterogeneous media. The numerical simulation results indicate that the proposed AMR-FAM is more accurate than the traditional AMR-FAM. Compared with the simulation in the original fine grids, the proposed AMR-FAM can provide nearly the same results. Moreover, the computational cost in the AMR grids is only approximately one-third of the cost in the original fine grids according to our numerical tests.

Pore-scale direct numerical simulations were conducted to explore the changes of flow paths and dynamic distribution of water-oil phases during two-phase displacement in porous media. Three contact angles, i.e., 45 degrees, 90 degrees, and 135 degrees, correspond to water-wet, intermediate-wet, and oil-wet rock surfaces, respectively. The medial axis within complex pore channels is obtained using a grid-based method. Subsequently, a novel particle tracking-based approach is proposed to extract the flow paths during two-phase displacement. The results show that when water enters a porous medium filled with oil, the distribution of oil phase flow paths is uniform, indicating that all oil phases are movable. As water continues to be injected, some water-oil interfaces stop at pores or throats due to capillary force, causing the flow paths through these locations to disappear. In oil-wet and intermediate porous media, capillary forces always act as resistance; therefore, the water-oil interfaces stop at the throat channels. Interestingly, under water-wet conditions, the water-oil interface can reverse, and the capillary force presents resistance, resulting in the interfaces being blocked at enlarging areas, usually from throats to pores. Once breakthrough occurs, preferential flow paths for pure water are established, resulting in a rapid decrease of flow paths with moving water-oil interfaces under oil-wet and intermediate-wet conditions, while movement continues in a water-wet medium. Due to the typically lower viscosity of water than oil, the viscous force decreases as water saturation increases. As water injection continues, the water in some channels gradually displaces oil, resulting in differences in flow rates between oil-filled and water-filled channels. This study provides a method for identifying the changes in flow paths during two-phase displacement in porous media and gives potential benefits for designing strategies to improve oil recovery.

Strong heterogeneity, complicated lithology, and chaotic seismic reflection characteristics are all common features of glutenite reservoirs. It is challenging to pinpoint the interior lithology and quantitatively describe the heterogeneity of the single-stage glutenite. To explore the distribution and superimposition features of subaqueous fans, the upper Es4 in the Y229 region of Dongying sag was used as an example. We have developed a multitrend fusion constraint modeling method based on deposition patterns. First, the truncated Gaussian simulation method is used to establish the sedimentary subfacies model of the subaqueous fan. Second, one lithologic probability volume is generated according to the proportion of various lithology of different sedimentary subfacies. Then, the lobes model is constructed using the object-based simulation method and the quantitative parameters from outcrop and flume sedimentation simulations. In addition, a different lithology probability volume is determined based on how far a particular lobe is from the centerline. The two probability volumes are combined to create the integrated lithology probability volume, which reflects the planar trend and the internal differences of different lobes. To constrain lithologic modeling, the integrated lithology probability volume is used. By comparing the model results with a single-trend constraint, the findings indicate that the multitrend integration constraint modeling method may more accurately depict the internal variability of the glutenite reservoir. In addition, the lithologic model built on this foundation is consistent with the depositional model.

Porous carbon nanofibers doped with nickel (Ni) were successfully fabricated through electrospinning, carbonization, and CO2 activation techniques using polyacrylonitrile (PAN) and petroleum pitch as carbon sources and nickel acetate as the dopant. During the activation process, Ni was reduced and dispersed in situ on the carbon matrix. The effects of Ni doping content on the morphology and structure of the carbon nanofibers were systematically investigated using SEM, TEM, XPS, XRD, Raman, and BET analyses. The experimental results revealed that the prepared materials had a hierarchically porous structure and that Ni nanoparticles played multiple roles in the preparation process, including catalyzing pore expansion and catalytic graphitization. However, particle agglomeration and fiber fracture occurred when the Ni content was high. In the adsorption/desorption experiments, the sample with 10 wt% Ni doping exhibited the highest specific surface area and micropore volume of 750.7 m(2)/g and 0.258 cm(3)/g, respectively, and had the maximum hydrogen storage capacity of 1.39 wt% at 298 K and 10 MPa. The analyses suggested that the hydrogen adsorption mechanism contributed to enhanced H-2 adsorption by the spillover effect in addition to physisorption.

Development mode adjustment is an important measure to further enhance oil recovery after primary waterflooding. Investigating the oil-water flow behaviors in the pores is significant to deepening the understanding of the macrobehavior of waterflooding and the designation of the reservoir development plan. Previous studies mainly focused on the change in recovery factor and macroflow characteristics, while less attention was paid to the causes of the change in recovery factor and the difference in macrocharacteristics. In this paper, the numerical simulation technology by coupling the Navier-Stokes equation with the method of volume of fluid is employed to investigate the dynamic formation mechanism of the remaining oil at pore-scale in the primary waterflooding, and then the pore-scale oil-water two-phase behaviors under three waterflooding development adjustment schemes: changing flooding direction, turning extraction well to injection well, and increasing injection rate are studied. The research shows that when the viscous resistance of the water-bearing channel is less than the sum of the capillary barrier and drainage capillary resistance (the resistance in the displacement of wetting phase displacement using the nonwetting phase), the remaining oil is formed. Water ➔oil ➔water ➔oil displacement mode is formed in the process of changing flow direction, which makes the force of the phase interface tend to balance, reduces the capillary effect in waterflooding, and improves oil recovery. In the process of developing the scheme of turning extraction into an injection well, through multipoint injection, it advances from the central water-bearing area to the oil-bearing area on both sides in multiple paths, forming a larger spatial spread range than that in the change of flooding direction. Under the influence of the capillary barrier effect and drainage capillary resistance, when the injection rate is increased, the remaining oil can restart to move only when the flowing rate exceeds a certain value. The small viscous resistance in the water-bearing channel and the lateral resistance from the capillary barrier limiting the lateral sweep of water are the primary reasons for the insignificant improvement of oil recovery under the condition of a low liquid injection rate. The findings of this study can help for better understanding of pore-scale flow mechanism behaviors and their influences on the macroscopic development features in the waterflooding process.

Reservoir stochastic simulation usually can be divided into two categories: object-based modeling method and pixel-based modeling method. For fluvial reservoirs, the complex geometry of sedimentary microfacies, such as the morphology of channels and natural levees, can be reproduced better by using the object-based methods. But these object-based approaches also have difficulty in conditioning models to dense well data. Most of previous conditional methods are iterative optimization algorithms, which take a long time to reach convergence when there are many conditional data, and they generally obtain a low degree of conditionalization. Therefore, we develop a new conditional methodology. The process of conditioning to well data is as follows. First, the classical object-based method is used to establish the channel reservoir model. Based on this, the distance parameter field (D1) related to the center lines of channels is established. Next, we convert the information of channel and nonchannel wells into the distance from the center lines and then set up the distance parameter field (D2) by kriging and other methods. Finally, D1 is modified by using the logical relationship between D2 and D1, and then D1 is truncated to obtain the river channel model. The analysis of the results of three examples indicates that this new approach can improve the conditional level greatly.

A modified sequential fully implicit method (MSFI) is proposed for the compositional reservoir model in this paper. An advanced discrete pressure equation is derived through Taylor's expansion of the discrete transport equations of water phase and the components. Its Newtonian iterative form can be obtained through Gaussian elimination of local cell equations with negligible computational cost. The constructed pressure equation is solved first, then the transport equations of water phase and components are solved sequentially by the topological sorting algorithm. In the conventional SFI algorithm, an outer loop iteration is imposed for the coupled system of solving the pressure and transport equations sequentially, resulting in the same solution as the fully implicit method after convergence. It is noticed that the convergence of the outer iteration is not necessary, at least in our proposed MSFI. For the strongly coupled compositional reservoir model, several outer iterations are still needed to ensure the accuracy and stability of the whole algorithm. Three times of outer iterations is recommended according to our numerical tests. By fixing the number of outer iterations instead of reaching convergence, the computational efficiency can be greatly improved. To further increase the calculation efficiency, the proposed MSFI is performed in the adaptive mesh refinement (AMR) grids. Numerical tests indicate that the proposed MSFI, even performed in AMR grids, can provide accurate simulation results. Meanwhile, the computational efficiency can be greatly improved compared to the conventional SFI.

由于储层非均质性及开发措施的影响,易造成水驱过程驱替不均衡.基于统计学方法,提出了采用克里斯琴森均匀系数定量评价多层水驱油藏开发过程中均衡驱替程度,以最大化均衡驱替为目标建立了分层注采参数优化数学模型,采用协方差矩阵自适应智能进化算法进行求解,形成了多层水驱油藏注采参数协同优化方法,并通过实例验证了方法的可靠性.将建立的新方法应用于胜利油区胜坨油田坨142油藏典型井组,通过对比智能协同分层注采参数优化前后结果,相比于原始方案克里斯琴森均匀系数提高了3.49％,累积产油量增加1.024×104m3,同时含水率降低1.34％.表明该方法可以精准地控制各层间的配产和配注,有效地改善水驱油藏层间和井间的均衡驱替状况,实现注采参数和剩余油分布的精准匹配.

Tight reservoirs have poor physical properties: low permeability and strong heterogeneity, which makes it difficult to predict productivity. Accurate prediction of oil well production plays a very important role in the exploration and development of oil and gas reservoirs, and improving the accuracy of production prediction has always been a key issue in reservoir characterization. With the development of artificial intelligence, high-performance algorithms make reliable production prediction possible from the perspective of data. Due to the high cost and large error of traditional seepage theory formulas in predicting oil well production, this paper establishes a horizontal well productivity prediction model based on a hybrid neural network method (CNN-LSTM), which solves the limitations of traditional methods and produces accurate predictions of horizontal wells’ daily oil production. In order to prove the effectiveness of the model, compared with the prediction results of BPNN, RBF, RNN and LSTM, it is concluded that the error results of the CNN-LSTM prediction model are 67%, 60%, 51.3% and 28% less than those of the four models, respectively, and the determination coefficient exceeds 0.95. The results show that the prediction model based on a hybrid neural network can accurately reflect the dynamic change law of production, which marks this study as a preliminary attempt of the application of this neural network method in petroleum engineering, and also provides a new method for the application of artificial intelligence in oil and gas field development.

沉积相建模是储层建模中的一个重要环节,有多种方法可以用来建立沉积相模型.传统的建模方法需要利用各种参数对变量的空间结构信息进行刻画,如变差函数、数据样式等,在模拟中再现这种空间结构.利用生成对抗神经网络方法(GAN,Generative Adversarial Nets)建模采用了不同的策略,通过对大量图像(模型)的学习,生成与学习样本具有高度相似特征的模型.基于单一图像生成对抗神经网络方法(SinGAN,Generative Adversarial Nets based on single image)对传统的GAN方法进行改进,仅需一张图像进行训练就能够生成高度相似的图像.以N气田2个小层的沉积微相图为例,建立了相应的沉积相模型,并与经典的基于样式的多点地质统计学建模方法(Simpat)对比可以看出,SinGAN方法与训练图像刻画的沉积微相空间结构更相似,具有良好的应用前景.

The presence of natural fractures can significantly affect the quality of hydraulic fracturing operations in tightsand and shale or oil/gas formations. This paper describes the procedure used to model natural fractures with continuum damage tensor and the resulting orthotropic permeability tensor. A damage model that uses damage variable in tensor form is presented. In the procedure presented, a nonnegligible angle is assumed to exist between directions of principal stresses in the formation and in the natural-fractures-related damage tensor, and this difference in orientation is modeled by introducing local directions in the model. A damage-dependent permeability tensor in tabular form is then proposed. A second case scenario when the directions of principal stresses and natural fractures align is also analyzed. Numerical results of fracture distribution are presented for both cases, and differences can be seen from the computed contour of the damage variable. The results indicate that the model can effectively simulate the fracture propagation phenomena during hydraulic fracturing.

目前油藏、采油依托各自专业数据和信息系统进行异常问题的分析,对于两个系统间的复杂关联关系考虑不够,导致生产异常的诊断仍较局限,治理措施针对性不强.基于随机森林算法和卷积神经网络算法集成学习构造了油藏井筒一体化智能诊断模型,根据注水失效、泵漏失等不同油藏、井筒问题,以基于随机森林的决策树分析油藏异常工况,卷积神经网络诊断井筒异常故障,通过集成学习方法将两类分类器结合起来,形成一体化诊断.现场验证结果表明,所建立的方法通过集成学习提升了单分类器性能与范化能力,应用准确率达到90％以上,实现了油藏和井筒问题的一体化诊断,为油田智能化管控提供了有力支撑.

油气田开发生产系统包含油藏、井筒、管网三大生产环节.传统数值模拟方法主要模拟油藏中流体的流动,而油藏-井筒-管网一体化耦合模拟同时反映油藏供液能力、井筒举升能力和管网集输能力,对准确预测生产态势、实现生产系统全流程优化具有重要意义.通过研究井筒-管网模型建立方法及流体多相管流相关式适应性,实现了井筒、管网内流动模拟和流动保障模拟;通过节点链接建立一体化模型,耦合关键节点的流动关系,形成了油藏-井筒-管网一体化耦合模拟方法,实现了油气田开发全流程模拟.基于埕岛油田西A区块,根据海上油田实际生产需求和约束条件,优化生产制度、配产配注等,基于一体化模型进行全流程优化,最大程度提高了油气田开发系统产量效益.