Abstract:The upper Submember of the fourth Member of the Eocene Shahejie Formation （Es ₄^U）and the Lower Submember of the third Member of the Eocene Shahejie Formation （Es₃ ^L） shales in Dongying Sag are the main horizons of shale oil in Shengli Oilfield.There are three reservoir types：carbonate-rich， mixed and felsic-rich， among which the carbonate-rich type is the most developed.The sections of carbonate-rich shale where lamellar sparry calcites are well-developed exhibit favorable characteristics in reservoir properties， oil bearing， permeability， and compressibility， and the peak daily oil production of many horizontal well exceeds 100 tons in Niuzhuang， Minfeng， and Lijin sags. In this paper， the genetic mechanism of lamellar sparry calcite and its influence on reservoirs were further analyzed through the thin section， cathodoluminescence， inclusion， isotope geochemistry， and argon ion polishing scanning electron microscope. The results showed that the lamellar sparry calcites could be divided into recrystallized granular calcite laminae and lamellar sparry calcite veins crystallized by hydrocarbon generating fluids. The recrystallized granular calcite laminae were formed by the in-situ recrystallization of the lacustrine authigenic micritic calcite laminae during the early diagenesis stage， and the formation process was closely related to the reduction of sulfate bacteria during the shallow burial period. The calcite crystals are granular with irregular shapes， the laminae are equally thick and the distribution is continuous. The formation of lamellar sparry calcite veins was closely related to the hydrocarbon generation and expulsion during the thermal evolution maturity of the organic matter and the organic acids dissolved the micrite calcite in shale to form carbonate-rich fluids. These fluids migrated along the lamellar fractures under the action of hydrocarbon generation pressures， then crystallized and precipitated again to form calcite veins. The laminae are thicker， lenticular， and distributed intermittently on the plane. In sparry calcite lamellae， the interlayer fractures （bedding fractures）， intergranular， intragranular pores， and dissolved pores are well-developede， the proportion of macropores is high and fractures and pores have good connectivity. Sparry calcite laminae and organic-rich laminae are interbedded in“ couplets”， which constitute two lithofacies： lamellar organic-rich sparry caly-lime shale and lamellar organic-rich sparry limycaly shale. The "source-reservoir" shale has high-quality which is the most favorable lithofacies in the carbonate-rich shale of Jiyang shale oil.

Abstract:To clarify the formation mechanism and evolution characteristics of the deep Permian overpressure in western Well Pen-1 Sag and its periphery in Junggar Basin， based on the data of drilling， logging， and measured formation pressure， the formation mechanism and evolution characteristics of overpressure were comprehensively studied， and the contribution rate of different types of overpressure genesis in the target layer was quantitatively characterized by using the methods of analysis of well logging curve combination， cross plots， and basin simulation. The results show that ① the overpressure genesises of different lithologies of the deep Permian in the study area have obvious differences， among which the overpressure genesises of source rock of Fengcheng Formation and the Lower Wuerhe Formation are mainly hydrocarbon generation and disequilibrium compaction， with hydrocarbon generation dominating； the overpressure genesis of mudstone caprock is mainly disequilibrium compaction， while the overpressure genesises of reservoirs of the Fengcheng Formation and Lower Wuerhe Formation are mainly overpressure transfer and disequilibrium compaction. ②The overpressure induced by hydrocarbon generation of Fengcheng Formation in the study area has appeared since the Early Permian and reached the maximum at present， and there are little differences in the size of the pressurization among different tectonic positions； the overpressure of the source rock is mainly distributed in the range of 39.43-49.16 MPa， but its contribution rate has obvious differences， with larger salients and smaller sags， and the contribution rates of pressurization induced by hydrocarbon generation in salients and sags to the total overpressure are 84.49%-94.41% and 65%-67.3%， respectively. The pressurization induced by hydrocarbon generation in the source rock of the Lower Wuerhe Formation has a similar size to that of Fengcheng Formation. The contribution rate of disequilibrium compaction to the overpressure in the mudstone caprock is basically 100%. ③ Overpressure transfer pressurizations in the Permian reservoir of the study area are mainly formed from the Late Jurassic to Early Cretaceous and since the Paleoproterozoic to now， and their contribution rates vary significantly in different salient areas， with 21.86%-23.35% and 100% in Dabasong Salient and Shixi Salient， respectively. Clarifying the overpressure distribution law of deep and ultra-deep formations in the study area can provide a basis for the further development of the new basin.

Abstract:The reservoir lithology of the fourth Member of Dengying Formation in the north-central Sichuan Basin is complex with strong heterogeneity，which limits the gas reservoir exploration in the fourth Member of Dengying Formation. To further deepen the understanding of sedimentary facies and reservoir characteristics in Well DB1 area in north-central Sichuan， the sedimentary facies and reservoir characteristics of the fourth Member of Dengying Formation in Well DB1l area in the north-central Sichuan were studied based on the data of drilling core，thin section observation，well logging，log，physical property test，and drilling results. The results show that two sedimentary subfacies of platform margin and restricted platform are mainly developed in the fourth Member of Dengying Formation in the study area， and sedimentary microfacies such as algal mound，grain beach，and inter-beach sea were further identified. Among them， algal mound microfacies mainly develop algal clot dolomite and algal stromatolite dolomite， grain beach microfacies mainly develop calcarenite dolomite， and inter-beach sea microfacies mainly develop micritic dolomite and powder crystal dolomite. The algal mounds and grain beaches in the fourth Member of Dengying Formation in the study area are mainly developed on the side of the platform margin， and the deposition of algal mounds and grain beaches in the platform gradually decreases mainly inter-beach sea deposition. The development of algal mounds in the fourth Member of the Dengying Formation in the study area is characterized by longitudinal superposition，and the horizontal extension is relatively close. Meanwhile， the algal mounds are concentrated in the Upper Submember of the fourth Member of Dengying Formation with large monomer thickness，while the Lower Submember of the fourth Member of Dengying Formation is less developed and thinner. The reservoir rock types of the fourth Member of Dengying Formation in the study area are mainly algal clot dolomite，algal stromatolite dolomite，and calcarenite dolomite. The reservoir space is mainly algal binding framework pore，intergranular dissolved pore，and intergranular dissolved pore. The core porosity of the fourth Member of Dengying Formation in the study area is distributed between 2.02% and 6.03% with an overall average porosity of 3.11%，the core permeability is distributed between 0.007 04 and 9.78 mD，and the overall average permeability is 0.692 64 mD. The overall physical properties are characterized by low porosity and ultra-low permeability，and some samples show low porosity and high permeability due to the influence of fractures.

Abstract:Accurate characterization of interlayer heterogeneity is the key to solving the problem of severe interlayer disturbances caused by large well sections in commingling production reservoir. The numerical applicability of the Lorenz method in development field is investigated by experiments and reservoir engineering theories because it is the most commonly employed method for describing interlayer heterogeneity， but its applicability remains uncertain. The study reveals that the Lorenz method’s characterization of the heterogeneity has a weak correlation with development effects， exerting significant influence on development effects only within a localized sensitive range. Additionally，curve shape also affects development effects when Lorenz coefficients are equal，but traditional methods are ineffective in capturing this phenomenon. Consequently， a modification to the Lorenz method is introduced， followed by physical simulation experiments and validation with oilfield data. Experimental results demonstrate that the modification significantly broadens the sensitive ranges of Lorenz coefficients，leading to a substantial enhancement in the correlation between waterflooding development effects and Lorenz coefficients. Example applications show that the correlation coefficient between the Lorenz coefficients after the modification and recovery factors increases from 0.48 to 0.87 in the KL-A area and from 0.74 to 0.87 in the KL-B area. This highlights a significant enhancement in the correlation between interlayer heterogeneity and development indicators.

Abstract:The tight sandstone reservoirs are deeply buried and strongly influenced by diagenetic processes such as mechanical compaction，dissolution，and metasomatism. Therefore，their lithological characteristics are complex. As the exploration and development of oil and gas resources go towards deeper reservoirs， the high ground stress and high confining pressure in deep layers have a significant impact on the brittleness and plasticity of reservoirs， which affects the hydraulic fracturing effect and oil and gas production.The deep tight sandstone reservoir in central Junggar Basin has abundant reserves. However， traditional oil and gas extraction methods fail to meet production needs due to the deep burial depth，poor physical properties，and abnormally high confining pressure of the reservoir， as well as limited understanding of the brittleness variation law of its in-situ geomechanical core. To this end，through physical experiments and numerical simulation techniques， this paper studies the differences in rock mechanics characteristics， brittleness and plasticity transformation， fracture modes， and acoustic emission characteristics of deep rock cores in the central Junggar Basin. The results show that as the confining pressure increases， the rock transitions from brittle failure to plastic failure，and the fracture mode changes from splitting failure to shear failure， weakening the brittleness of sandstone. The overall degree of sandstone brittleness reflected by the six brittleness indices based on the energy balance method and normalization treatment decreases，but the numerical ranges of different brittleness indices and their correlation with brittleness vary. The triaxial compression test experiments of core models with different brittle values show that when the content of brittle minerals is low，the rock exhibits single oblique shear failure. As the content of brittle minerals increases，rock exhibits a composite failure mode with multiple fracture surfaces. There is a correlation between the acoustic emission characteristics during rock fracture and its brittleness，involving three modes：swarm type，foreshock–main shock–aftershock type，and main shock type.

Abstract:In response to the contradiction between the deep mobility control capability and near-well-bore area injectability of oil displacement agents，this paper proposed that an ideal oil displacement agent for mobility control should fulfill the technical requirements of low viscosity preparation/transportation， low adsorption in the near-well-bore area for long-term stable injection，and produced liquid with low concentrations of the component. Meanwhile，multi-level flow resistance peaks （Δpmax） at different positions in the deep reservoir should be established during the flow process，and flow resistance of fluid after displacing front should be maintained at a lower level. Based on the synergistic effect of the adsorption at the solid/liquid interface and the inter-molecular interaction，a hydrophobically associating water-soluble polymer/anionic surfactant binary system was designed with the characteristics of dynamically changing the system constituent and microscopic solution structure， thereby changing flow resistance. Compared with HP-1 （1 500 mg/L）， the binary systems HP-1 （1 500 mg/L）/SDSB （150 mg/L） and HP-1 （1 500 mg/L）/SDSB （200 mg/L）， with similar apparent viscosity and different constituents can construct dynamic flow resistance with higher values and better spatial distribution in the middle-rear position of the porous medium flow during the slug injection and subsequent water flooding processes. It verified that the binary system had the characteristics of viscosity increasing and delayed breakthrough during migration，and the mobility control capability in the displacing front was more powerful. In addition，the spatial distribution of flow resistance in the reservoir becomes more reasonable，which prolonged the overall breakthrough time of the slugs，thus expanding the swept volume and enhancing the oil displacement efficiency. Oil displacement experiments were conducted in a Bohai Oilfield with heavy oil，strong heterogeneity， and 80% water cut during water flooding， with chemical industrial products with basically the same dosage and similar cost adopted. The results show that the spatial dynamic distribution characteristics of the binary system before subsequent water flooding breakthrough are a crucial factor affecting oil displacement efficiency， and the binary system （APP4 （1 400 mg/L） + ZX-27（300 mg/L）， viscosity of 6.4 mPa·s） can improve the oil recovery by more than 10% compared with hydrophobically associating water-soluble polymer AP-P4 （1 750 mg/L， viscosity of 62.9 mPa·s）.

Abstract:Surfactants are widely used in the tertiary recovery of oilfields， and it is of great significance to study the mechanism of surfactant flooding from the microscopic point of view. In recent years， molecular dynamics （MD） simulation has become essential for oil and gas field development research. The use of MD simulation to study surfactants has become a hot spot. Through MD simulation，the numerical solutions of Newton’s equations are calculated for all the moving particles in the system， and the change in atomic positions with time is analyzed， from which some laws are found. MD simulation can be used to study the microscopic behavior of surfactant molecules to explore the properties of surfactant molecules and the microscopic flooding mechanism. MD simulation of the movement and aggregation state of surfactants at the interface and the analysis of the influence of surfactant molecules at the interface on the interfacial system are of some guiding significance for the field application of surfactants. To this end，based on the principle overview of MD simulation， firstly， the force field， boundary conditions， system types， and numerical algorithms in MD simulation are summarized；secondly，the microscopic mechanisms of MD simulation of surfactant flooding are highlighted，including the reduction of oil-water interfacial tension， the change in surface wettability， the increase in interfacial charge and emulsification，etc.；then，examples of the application of MD simulation in surfactant flooding are presented. Finally，the development directions of MD simulation technology in surfactant flooding are proposed， including the flooding environment similar to production， the design of new surfactants，the relationship between theory and experiment，the selection of molecular force field and potential energy model，and the research and development of composite and multifunctional integrated surfactants.

Abstract:Ultradeep carbonate gas reservoirs have typical characteristics such as diverse storage media，strong heterogeneity，high temperature，and high pressure. They are generally developed using deviated wells，which complicates the evaluation of gas well productivity. Considering factors such as multiple storage media，stress sensitivity，nonlinear percolation，threshold pressure effect，and well deviation angle，a trinomial productivity prediction model was established based on the Forchheimer’s gas-phase differential equation，stress sensitivity and physical simulation experimental data of gas-phase flow. The gas well productivity of different types of carbonate reservoirs in different regions was predicted，and the effects of the above factors on gas productivity were analyzed.The results indicate that pore-type and low-permeability cavity-type reservoirs are affected by threshold pressure effects，while fracture-cavity type and high-permeability cavity-type reservoirs are influenced by nonlinear flow. Therefore，the trinomial productivity prediction model is more suitable for ultradeep carbonate gas reservoirs with diverse reservoir types. Different types of reservoirs exhibit different degrees of stress sensitivity and nonlinear flow characteristics. Physical simulation experiments on the cores from various reservoirs are necessary to determine the parameters required for the productivity equation. The well deviation angle creates a negative skin factor affecting gas well productivity. When the well deviation angle is greater than 55°，the gas well productivity of all types of reservoirs starts to increase rapidly， especially for fracture-cavity type reservoirs. The threshold pressure effect has a significant negative impact on productivity at low pressure differentials， whereas stress sensitivity and nonlinear flow mainly impact at high pressure differentials. Starting pressure gradient and nonlinear flow significantly inhibit gas well productivity when their coefficients are 0.01-0.048 MPa/m and 109-1012 m?1， respectively. In contrast， the productivity loss caused by stress sensitivity is relatively stable. The dominant factors affecting well productivity are the skin coefficient and formation coefficient. During the development stage， the impact of the nonlinear flow effect on high-permeability reservoirs and the threshold pressure effect on low-permeability reservoirs is also significant.

Abstract:Tight sandstone gas reservoirs have the characteristics of poor physical properties of the reservoir， complex pore throat structure， strong anisotropy， and extensive coverage of initial water saturation， which lead to the complex flow mechanism of fluids in reservoirs and bring significant challenges to the accurate prediction and evaluation of gas well productivity. Therefore， based on the theory of gas-water two-phase non-Darcy flow， this study comprehensively considers the complex productivity factors， such as reservoir anisotropy， different degrees of stress sensitivity between reservoir matrix and fracture， two-phase starting pressure gradient，high-speed non-Darcy flow effect， gas slippage effect， finite conductivity of fracture， and inter-fracture interference. A new gas-water two-phase productivity prediction model for multistage fractured horizontal wells in anisotropic tight sandstone gas reservoirs is established by using coordinate transformation， perturbation ellipse theory， equivalent development rectangle theory，equivalent well diameter principle， pressure superposition principle， and hydropower similarity principle. Field examples verify the accuracy and practicability of the model. The gas-water two-phase productivity prediction curve is drawn， and the influence of sensitive parameters on productivity is evaluated. The results show that the open flow capacity of a fractured gas well increases with the increase in slippage effect， fracture conductivity， fracture half-length， fracture number， and the angles between the wellbore of the horizontal well and the main permeability of the formation. The open flow capacity of the fractured horizontal well decreases with the increase in reservoir anisotropy， reservoir stress sensitivity， fracture stress sensitivity， two-phase starting pressure gradient， and water-gas volume ratio. Water has an inhibitory effect on the flow of gas， and greater displacement pressure difference indicates a more significant inhibitory effect， so it is necessary to take waterproof and water control measures in advance. The research results further narrow the gap between the productivity prediction results of tight sandstone gas reservoirs and the actual production of the field， which is helpful for parameter evaluation， dynamic prediction， productivity evaluation， and exploration and development decision-making of tight sandstone gas reservoirs.

Abstract:Long horizontal wells and large-scale volumetric fracturing are key technologies for shale oil development. Due to reservoir characteristics and the fracturing scale， the optimal length of horizontal wells varies in different fields. It is necessary to consider both reservoir flow and wellbore flow of long horizontal wells when the optimal length of horizontal wells is determined.North American marine shale has good continuity and is relatively homogeneous， with thick oil layers， where long horizontal wells play a decisive role in improving single-well oil production. In contrast， Chinese continental shale is highly heterogeneous， with rapid lateral changes and developed thin interbeds， and it requires the determination of reasonable lengths of horizontal wells to increase single-well oil production and economic benefits. Therefore， accurately characterizing shale oil reservoir flow and wellbore flow of long horizontal wells is crucial for the production capacity evaluation of shale oil and the determination of appropriate lengths of horizontal wells. This study， based on the five-zone linear flow， comprehensively considered the starting pressure gradient of the shale matrix， complex heterogeneous fracture networks， and stress sensitivity. It also established a multi-stage hydraulic fracture-wellbore coupled flow model of shale oil horizontal wells and solved the model using perturbation transformation and Laplace transformation. Based on this model， production capacity was evaluated and predicted， and complex fracture network parameter inversion was achieved. The influence of reservoir thickness and the number of fractures on the optimal length of shale oil horizontal wells was demonstrated. On this basis， a length optimization chart for shale oil horizontal wells was developed. The results show that the number of fractures is the main controlling factor affecting the optimal length of horizontal wells for thick shale. By applying the above model to actual fields， the study determined the optimal lengths of horizontal wells in different fields under current reservoir conditions. This research provides a theoretical basis for the determination of the optimal lengths of horizontal wells in different shale oil reservoirs.

Abstract:During the hydraulic fracturing construction of unconventional reservoirs， hydraulic fractures easily intersect with natural fractures to form complex fractures. The migration and the laying form of the proppant directly determine the stimulation effect of the reservoirs. In order to study the laying rules of proppants in complex fractures， a proppant-fracturing fluid two-phase flow mathematical model was established based on the Euler-Euler method， and the dune laying shapes in complex fractures of the indoor test device and the numerical simulation were compared to verify the model. The results show that the selected Euler two-fluid numerical model can be used to study the migration and laying rules of proppants carried by slick water in complex fractures. Based on the similarity criterion， a simplified complex fracture plate model was established， and the morphological parameters and area of the dune were normalized to obtain the influence of the fracturing fluid displacement，viscosity，proppant particle size， fracture width， and fracture morphology on proppant laying in complex fractures. The results show that increasing the displacement and viscosity and reducing the particle size is beneficial to proppant migration to the depth of the fractures， and the displacement has the most obvious impact on the shape of the dune in the branch fractures，but increasing the displacement is not conducive to proppant to fill the near-well zone. The influence of the fracture width is reflected in the wall effect. Under the same injection time， when the viscosity of the fracturing fluid increases from 1 mPa·s to 5 mPa·s， the normalized length of the main fracture dune increases by 37.9% on average，and the normalized height reduces by 61.4%. A more complex fracture structure indicates a higher proportion of proppant laying area in all branch fractures and a more significant diversion effect. As the number，series，and extension length of the branch fractures of complex fractures increase， the height and length of the dune in the branch fractures decrease； compared with the T1 type fracture， the dune in the T3 and T5 type fractures decreases the most in terms of length and height.

Abstract:In recent years， tight oil horizontal wells in Daqing oilfield are mainly fractured by close cutting. However， the multicluster fracture propagation generates induced stress， which leads to uneven fracture extension in each cluster. To this end， the finite element software ABAQUS was used to establish the numerical model of multi-cluster fracture propagation in horizontal wells in Daqing tight-oil reservoir. The effects of two fracture parameters （cluster spacing， clusters within one fracturing stage） and three construction parameters （fracturing processes， construction displacement， and liquid viscosity） on fracture propagation were studied by the extended finite element （XFEM） simulation method. At the same time， it is found that there is a good consistency between the simulated fracture length of each cluster and the oil production contribution of each cluster when the simulated result is compared with the capacity contribution result of each cluster tested by actual optical fiber in the field. This confirms the correctness of the simulation results. The simulation results show that the stress interference of each cluster is obvious when the cluster spacing is reduced to 5 m， which is not conducive to fracture propagation. The mean fracture half-length of a single cluster does not change significantly when the cluster spacing reaches 10 m or more， and the reasonable cluster distance is about 10 m. The fracture halflength increases with increased clusters within one fracturing stage， but the non-uniform extension also becomes obvious. Compared with the simultaneous stimulation of multiple clusters in the bridge plug stage， the coiled tubing single-cluster stimulation mode is more favorable to fracture propagation. The average half-length of single cluster fractures increases with the rise of the construction displacement， and the uniform extension of fractures can be promoted by increasing the construction displacement. The average fracture half-length of a single cluster rises with the increase of liquid viscosity， and the proportion of high viscous liquid can be increased appropriately.

Abstract:Carbonate reservoirs are often stimulated by acidizing and acid fracturing techniques， but the microstructure of the rock surface is corroded and destroyed after acid etching， and the mechanical properties of the rock are changed significantly， which further affects the final stimulation effect. The experiment on the mechanical properties， the study on surface morphological characteristics by scanning electron microscopic （SEM）， and the microscopic pore structure characterization by nuclear magnetic resonance （NMR） are carried out before and after limestone acid etching to clarify the influence of acid etching on the mechanical properties and microscopic pore structure changes of limestone. The results show that acid etching softening layers with a certain thickness of about 11.61 μm appear on the rock surface after acid etching. Remarkably， the compressive strength， elastic modulus， and surface hardness of limestone are all reduced to about 40% before acid etching due to the acid etching softening layers. Dissolved pores of different sizes are formed inside the rock surface after acid etching and stress concentration is induced around the dissolved pores，changing the stress distribution of the acid etching softening layers. Under high closure stress， the dissolved pores are easier to crush deformation or collapse failure， which is also the main reason for the change of mechanical properties of acid etching softening layers.According to the T₂ spectrogram of NMR， it is found that there is the largest change in the number of pores with pore sizes between 1 nm and 1 μm， and the acid dissolution alters the internal pore structures of the acid etching softening layers， hence further affecting the mechanical properties of rock after acid etching.

Abstract:Emulsification is one of the main mechanisms of chemical flooding to enhance oil recovery in heavy oil reservoirs. The emulsification of heavy oil in porous media is closely related to reservoir temperature and oil-displacing agent concentration. The relationship between the critical concentration and temperature of the oil-displacing agent required to form oil in water （O/W） emulsions in heavy oil was studied by emulsification experiment. The emulsification and migration rules of crude oil during chemical flooding of heavy oil were studied by using high-temperature and high-pressure microscopic visual oil displacement devices. The results show that a higher temperature in the emulsification experiment indicates that the crude oil is easier to emulsify， and the critical emulsification concentration of the required oil-displacing agent is lower. The critical emulsification concentration decreases by 90% when the temperature increases from 30 °C to 90 °C during the experiment. There are three modes of emulsification of heavy oil in porous media during microscopic oil displacement experiments for chemical flooding of heavy oil： the emulsification mode of heavy oil shows dispersion， migration， and coalescence at the throat in porous media when the concentration of the oil-displacing agent is less than the critical emulsification concentration of the oil-displacing agent required for heavy oil emulsification at this temperature；the emulsification mode of heavy oil indicates emulsification， migration， re-emulsification， and re-migration when the concentration of the oil-displacing agent is greater than the critical emulsification concentration of the oil-displacing agent required for heavy oil emulsification at this temperature； the emulsification mode of heavy oil is manifested as contact， stripping， and migration when the concentration of the oil-displacing agent is much larger than the critical emulsification concentration of the oildisplacing agent required for heavy oil emulsification at this temperature. Therefore， the concentration of the oil-displacing agent should be increased as much as possible to achieve efficient emulsification of heavy oil and improve oil recovery under the premise of controlling the cost when chemical flooding of heavy oil is carried out.

Abstract:The conventional reservoir automatic history matching method requires multiple computationally time-consuming reservoir numerical simulations. Deep learning agent models can perform alternative reservoir numerical simulation calculations with approximate accuracy and greater computational efficiency. In the reservoir automatic history matching method based on the deep learning agent model， the reservoir uncertainty parameters adjusted by the reservoir automatic history matching method are usually used as the input parameters of the deep learning agent model. Existing deep learning agent models are often single input-output neural network model architectures. They do not consider that reservoir automatic history matching methods require the adjustment of multiple reservoir uncertainty parameters. Multiple deep learning agent models need to be trained to predict water saturation field distribution and pressure field distribution in reservoirs. For solving this problem， a reservoir automatic history matching method based on a dual input-output convolutional neural network agent model was proposed to simultaneously predict the water saturation field distribution and pressure field distribution in a reservoir by using a dual input-output convolutional neural network， with the reservoir permeability field distribution and phase permeability parameters as input. The production was calculated with the help of the Peaceman equation. It was coupled to the ensemble smoother with multiple data assimilation （ES-MDA） methods to invert the reservoir permeability field distribution and phase permeability parameters to achieve a more efficient reservoir automatic history matching solution. The results of the study show that the prediction accuracy of the reservoir water saturation field distribution and pressure field distribution is above 93% at the specified time step based on the dual input-output convolutional neural network agent model. Compared with the traditional reservoir automatic history matching method， the proposed reservoir automatic history matching method based on a dual input-output convolutional neural network agent model avoids the time-consuming computation of multiple calls to the reservoir numerical simulator and improves the efficiency of the matching.

Abstract:The existing non-linear flow model does not consider the influence of temperature change on the non-linear flow law.Therefore， based on the capillary flow model， the Hagen-Poiseuille equation， and the boundary layer theory， as well as the influence of the temperature on the thickness of the boundary layer and the yield stress of the fluid， a dynamic non-linear flow model in low-permeability porous media considering the influence of temperature was derived. Then， the fitting verification was carried out according to the experimental data in published literature， and the calculated flow velocity curve and minimum start-up pressure gradient were highly consistent with the experimental data， which revealed the correctness of the dynamic non-linear flow model.Meanwhile， the comparative analysis with the static non-linear flow model and the Darcy flow model illustrated the rationality of the dynamic non-linear flow model. The sensitivity analysis of the flow velocity curve revealed that the increase in coefficients A and B made the flow velocity curve right-shift； the curvature degree of the non-linear flow section was weakened， and the minimum start-up pressure gradient was increased. The increase in the temperature had a much more significant effect on the flow velocity curve， which made the velocity curve move up obviously. The curvature degree of the non-linear flow section weakened， and the minimum start-up pressure gradient increased. The proposal of this dynamic non-linear flow model further enriches the development of the non-linear flow theory and can be extended to the numerical simulation of two-phase flow and high-temperature and highpressure oil and gas reservoirs.

Abstract:In order to achieve the multiple purposes of the highest CO² flooding recovery and the largest CO² storage in high water cut reservoirs，an objective function of collaborative CO² flooding and storage optimization for high water cut reservoirs based on economic net present value （NPV ） was established by considering CO² injection cost， CO² flooding and storage benefits. Combined with the reservoir numerical simulation method，we used simultaneous perturbation stochastic approximation （SPSA） to solve the objective function iteratively. Taking a high water cut reservoir in Shengli Oilfield as an example，the different injection timings of water transferring to CO² were set according to the different characteristics of water cut in the reservoir at different times. The optimization of water injection and CO² injection schemes was carried out， respectively， and the differences in production behavoir and NPV under the two injection methods were compared. The influence of gas injection timing on cumulative oil production and CO² storage was analyzed， and the optimal gas injection timing was determined. In addition， the effects of different gas injection prices and CO² storage subsidy schemes on their economic benefits were analyzed. The results show that the water cut should be lower than 0.97 during waterf transferring CO² injection to make the CO² flooding project profitable in high water cut reservoirs. The high project profit can be guaranteed as CO² is injected earlier when the CO² injection price is 1.2 yuan/m³， and the carbon subsidy is 0.114 yuan/m³.