Specific Reservoir Research Articles

Comprehensive Investigation of Factors Affecting Acid Fracture Propagation with Natural Fracture

Acid fracturing is a crucial stimulation technique to enhance hydrocarbon recovery in carbonate reservoirs. However, the interaction between acid fractures and natural fractures remains complex due to the combined effects of mechanical, chemical, and fluid flow processes. This study extends a previously developed hydro-mechano-reactive flow coupled model to analyze these interactions, focusing on the influence of acid dissolution. The model incorporates reservoir heterogeneity and simulates various scenarios, including different stress differences, approaching angles, injection rates, and acid concentrations. Numerical simulations reveal distinct propagation modes for acid and hydraulic fractures, highlighting the significant influence of acid dissolution on fracture behavior. Results show that hydraulic fractures are more likely to cross natural fractures, whereas acid fractures tend to be arrested due to wormhole formation. Increasing stress differences and approaching angles promote fracture crossing, while lower angles favor diversion into natural fractures. Higher injection rates facilitate fracture crossing by increasing pressure accumulation, but excessive acid concentrations hinder fracture initiation due to enhanced wormhole formation. The study demonstrates the importance of tailoring fracturing treatments to specific reservoir conditions, optimizing parameters to enhance fracture propagation and reservoir stimulation. These findings contribute to a deeper understanding of fracture mechanics in heterogeneous reservoirs and offer practical implications for improving the efficiency of hydraulic fracturing operations in unconventional reservoirs.

Optimizing the heat extraction performance in the Qingfeng Karst geothermal reservoir

Faults are typically associated with hydrothermal reservoirs and may control the heat extraction of geothermal reservoirs. However, in Karst geothermal reservoirs, the dissolution of soluble rocks results in pronounced spatial porosity and permeability anisotropy (PPA), potentially weakening the dominant role of faults. The question arises: which plays a dominant role, PPA or faults? Should geothermal exploitation strategies be adjusted under the interaction of both factors? In this study, we proposed a coupled Thermal-Hydraulic-Geothermal Wellbore (THG) numerical simulation model to streamline the integration of non-isothermal flow and heat transfer processes within a 3D wellbore. Utilizing this model, we analyzed the influence of faults compared to PPA and assessed the geothermal well doublet's heat extraction performance, including thermal breakthrough time, temperature losses along the wellbore wall, and recovery factor (Rg). There are specific geothermal reservoir development conditions leading to the latest thermal breakthrough time and the maximum Rg. The results of contribution rates highlight the influence of injection rate in determining Rg. It is crucial to note that thermal breakthrough time does not necessarily increase with injection temperatures. A more suitable definition for thermal breakthrough time in the Karst geothermal reservoir has also been identified. This study provides an essential reference value for the efficient development of similar Karst geothermal reservoirs.

Machine learning approaches identify immunologic signatures of total and intact HIV DNA during long-term antiretroviral therapy.

Understanding the interplay between the HIV reservoir and the host immune system may yield insights into HIV persistence during antiretroviral therapy (ART) and inform strategies for a cure. Here, we applied machine learning (ML) approaches to cross-sectional high-parameter HIV reservoir and immunology data in order to characterize host-reservoir associations and generate new hypotheses about HIV reservoir biology. High-dimensional immunophenotyping, quantification of HIV-specific T cell responses, and measurement of genetically intact and total HIV proviral DNA frequencies were performed on peripheral blood samples from 115 people with HIV (PWH) on long-term ART. Analysis demonstrated that both intact and total proviral DNA frequencies were positively correlated with T cell activation and exhaustion. Years of ART and select bifunctional HIV-specific CD4 T cell responses were negatively correlated with the percentage of intact proviruses. A leave-one-covariate-out inference approach identified specific HIV reservoir and clinical-demographic parameters, such as age and biological sex, that were particularly important in predicting immunophenotypes. Overall, immune parameters were more strongly associated with total HIV proviral frequencies than intact proviral frequencies. Uniquely, however, expression of the IL-7 receptor alpha chain (CD127) on CD4 T cells was more strongly correlated with the intact reservoir. Unsupervised dimension reduction analysis identified two main clusters of PWH with distinct immune and reservoir characteristics. Using reservoir correlates identified in these initial analyses, decision tree methods were employed to visualize relationships among multiple immune and clinical-demographic parameters and the HIV reservoir. Finally, using random splits of our data as training-test sets, ML algorithms predicted with approximately 70% accuracy whether a given participant had qualitatively high or low levels of total or intact HIV DNA . The techniques described here may be useful for assessing global patterns within the increasingly high-dimensional data used in HIV reservoir and other studies of complex biology.

Machine learning approaches identify immunologic signatures of total and intact HIV DNA during long-term antiretroviral therapy

Understanding the interplay between the HIV reservoir and the host immune system may yield insights into HIV persistence during antiretroviral therapy (ART) and inform strategies for a cure. Here, we applied machine learning (ML) approaches to cross-sectional high-parameter HIV reservoir and immunology data in order to characterize host–reservoir associations and generate new hypotheses about HIV reservoir biology. High-dimensional immunophenotyping, quantification of HIV-specific T cell responses, and measurement of genetically intact and total HIV proviral DNA frequencies were performed on peripheral blood samples from 115 people with HIV (PWH) on long-term ART. Analysis demonstrated that both intact and total proviral DNA frequencies were positively correlated with T cell activation and exhaustion. Years of ART and select bifunctional HIV-specific CD4 T cell responses were negatively correlated with the percentage of intact proviruses. A leave-one-covariate-out inference approach identified specific HIV reservoir and clinical–demographic parameters, such as age and biological sex, that were particularly important in predicting immunophenotypes. Overall, immune parameters were more strongly associated with total HIV proviral frequencies than intact proviral frequencies. Uniquely, however, expression of the IL-7 receptor alpha chain (CD127) on CD4 T cells was more strongly correlated with the intact reservoir. Unsupervised dimension reduction analysis identified two main clusters of PWH with distinct immune and reservoir characteristics. Using reservoir correlates identified in these initial analyses, decision tree methods were employed to visualize relationships among multiple immune and clinical–demographic parameters and the HIV reservoir. Finally, using random splits of our data as training-test sets, ML algorithms predicted with approximately 70% accuracy whether a given participant had qualitatively high or low levels of total or intact HIV DNA . The techniques described here may be useful for assessing global patterns within the increasingly high-dimensional data used in HIV reservoir and other studies of complex biology.

Productivity Prediction Model of Tight Oil Reservoir Based on Particle Swarm Optimization–Back Propagation Neural Network

Single-well productivity is a crucial metric for assessing the effectiveness of petroleum reservoir development. The accurate prediction of productivity is essential for achieving the efficient and economical development of oil–gas reservoirs. Traditional productivity prediction methods (empirical formulae and numerical simulation) are limited to specific reservoir types. There are few influencing factors, and a large number of ideal assumptions are made for the assumed conditions when predicting productivity. The application scenario is ideal. However, in tight oil reservoirs, numerous factors affect productivity, and their interactions exhibit significant complexity. Continuing to use traditional reservoir productivity prediction methods may result in significant calculation errors and lead to economic losses in oilfield development. To enhance the accuracy of tight reservoir productivity predictions and achieve economical and efficient development, this paper investigates the tight reservoir in the WZ block of the Beibuwan area, considering the impact of geological and engineering factors on productivity; the random forest tree and Spearman correlation coefficient are used to analyze the main influencing factors of productivity. The back propagation neural network optimized by particle swarm optimization was employed to develop a productivity prediction model (PSO-BP model) for offshore deep and ultra-deep tight reservoirs. The actual test well data of the oilfield are substituted into this model. The analysis results of the example application yielded an RMSE of 0.032, an MAE of 1.209, and an R2 value of 0.919. Compared with traditional productivity prediction methods, this study concludes that the model is both reasonable and practical. The calculation speed is faster and the calculation result is more accurate, which can greatly reduce productivity errors. The model constructed in this paper is well suited for predicting the productivity of tight reservoirs within the WZ block. It offers substantial guidance for predicting the productivity of similar reservoirs and supports the economical and efficient development of petroleum reservoirs.

Saturation Prediction of Specific Tight Sandstone Reservoirs Based on Experimental Analysis of Oil-Water Relative Permeability

Saturation Prediction of Specific Tight Sandstone Reservoirs Based on Experimental Analysis of Oil-Water Relative Permeability

Research on compound amidohydroxysulfobetaine foamer for high-temperature and high-salinity clastic reservoirs

A high-quality foaming agent should exhibit excellent solubility in formation water and strong foaming ability, while also meeting the requirements of stability under high-temperature and high-salinity conditions, as well as low adsorption capacity. This study explores the potential of amidohydroxysulfobetaine surfactant as a foaming agent in sandstone reservoirs under specific reservoir conditions of 110 ℃ and 25.0×104 mg/L salinity. The findings indicate that compounding EHSB with a small amount of LHSB to form LA13 significantly reduces the Krafft point of EHSB. Furthermore, LA13 demonstrates outstanding solubility in 25.0×104 mg/L brine and exhibits an overall adsorption capacity on quartz sand surfaces lower than 0.6 mg/g when its concentration is below 1.0 wt%. Performance evaluation of bulk foam for LA31 at concentrations ranging from 0.05 wt% to 0.4 wt% suggests that using a foaming agent concentration of 0.05 wt% results in relatively low foaming volume and foam decay half-life; however, consistent foaming performance is observed across other concentrations, leading to highly stable foam due to the compound system's higher surface dilational modulus being the primary factor contributing to foam stability. Flow experiments reveal that the foam generated in cores with permeabilities of 46.5 mD, 185.5 mD, 632.0 mD, and 2005.3 mD exhibits a high resistance factor when the concentration of LA13 exceeds 0.1 wt%, while also demonstrating commendable gas-channeling regulation capability. Therefore, it is recommended to use LA13 as a foaming agent for field applications when its concentration surpasses 0.1 wt%. Additionally, comparing the mobility of foam formed by concentrations of 0.1 wt% and 0.2 wt% LA13 in cores with varying permeabilities shows that LA13 exhibits selective mobility regulation specifically in cores with permeabilities lower than 632.0 mD.

Permeability prediction method of unconsolidated sandstone reservoirs using CT scanning technology and random forest model

Developing unconsolidated sandstone reservoirs presents a formidable challenge due to their loose formation, which often triggers alterations in pore parameters and seepage characteristics during water injection processes. This study focuses on a specific reservoir, utilizing micro-CT scanning to examine the intricate relationship between permeability and pore throat structure. Leveraging a random forest model, we establish an empirical formula tailored for high permeability reservoirs. Furthermore, we conduct in-situ CT scanning experiments across various displacement multiples to analyze the pore structure of unconsolidated sandstone cores. The derived relationship curves elucidate the positive correlations between porosity and average pore throat radius with displacement multiples, while revealing a negative correlation with tortuosity. These findings enable the formulation of quantitative formulas for permeability and displacement multiples within the studied block. Such insights prove instrumental in devising effective water injection development strategies, predicting dynamic reserves, and projecting water drive development for analogous unconsolidated sandstone reservoirs undergoing high water cut phases.

Carbonate reservoir characterization and permeability modeling using Machine Learning ـــ a study from Ras Fanar field, Gulf of Suez, Egypt

Predicting facies and petrophysical properties along and between wells is challenging in carbonate reservoir modeling. In the Nullipore carbonate reservoir, Ras Fanar field, depositional and long-term diagenetic processes result in a high degree of heterogeneity and complex distribution of facies, which in turn affect the reservoir quality. This provides a significant obstacle to building accurate geological models. This study integrates thin sections, routine core analyses, and well logging data to overcome such difficulties and model the Nullipore carbonate facies and permeability. The detailed petrographic analysis revealed the existence of seven microfacies in the reservoir, which are summed up into three facies associations (FAs), each of which represents a specific reservoir rock type (RRT): (1) supratidal FA, (2) intertidal FA, and (3) shallow subtidal FA. The three FAs were correlated with the gamma-ray logs to create facies logs for the studied wells, which were further populated via the Truncated Gaussian Simulation method. Cross-validation was used to evaluate the model's accuracy. The analysis of the available core data infers that the three RRTs are prospective and have a wide permeability distribution. However, RRT3 constitutes the best reservoir quality. The sedimentological analysis revealed that the long-term diagenetic events, involving the dolomitization of limestone and the dissolution of allochems have a major role in improving the pore connectivity and permeability of the reservoir. Fracture characterization discloses that fractures play a significant role in fluid storage and migration. Three Machine Learning (ML) models, including Adaptive boosting (AdaBoost), Gradient Boosting (GB), and Extreme Gradient Boosting (XGB), were developed to integrate the RRTs, porosity, and permeability to improve permeability prediction. Statistical analysis revealed that the XGB model outperforms other models and exhibits the highest prediction performance. The present study provides further insights into the characterization and modeling of facies and permeability of complex carbonate reservoirs. It can be applied in similar geological settings to better interpretation of depositional and diagenetic controls on reservoir quality assessment and aid in the field development plan.

Enhancing carbon sequestration efficiency and safety through albumin-optimized CO2 hydrate distribution and geological layer stability: An innovative approach

The persistently high global CO2 emissions are exacerbating the greenhouse effect issue. Scientists have explored methods such as carbon sequestration via subsea hydrate formation in response to global warming. However, this method faces a series of challenges when injecting CO2 on a large scale, including the slow formation rate of CO2 hydrates and potential leakage risks. Specially, the direct injection of promoters could lead to excessively high concentrations of additives close to the injection wells. This would consequently accelerate the local hydrate conversion, potentially causing reservoir blockage problems. The current study proposes a novel strategy involving additives that exhibit varying effects with concentrations. The high-concentration areas close to the wellbore can maintain the permeability by inhibiting hydrate formation; while the reduced concentration further away upon gas migration facilitates the rapid formation of CO2 hydrates. We successfully demonstrate the spatial and temporal differences in hydrate formation when using such additives. This can also facilitate the rapid formation of a dense hydrate cap layer over the reservoir, potentially preventing the leakage of CO2 gas. A new type of additive based on egg albumin was developed, which possesses the dual function and exhibits good biocompatibility and specific reservoir stability.

A New Method for the Evaluation of Productivity of Water-Producing Gas Wells in Specific Tight Sandstone Reservoirs

A New Method for the Evaluation of Productivity of Water-Producing Gas Wells in Specific Tight Sandstone Reservoirs

Specific quantification of inducible HIV-1 reservoir by RT-LAMP

BackgroundStrategies toward HIV-1 cure aim to clear, inactivate, reduce, or immunologically control the virus from a pool of latently infected cells such that combination antiretroviral therapy (cART) can be safely interrupted. In order to assess the impact of any putative curative interventions on the size and inducibility of the latent HIV-1 reservoir, robust and scalable assays are needed to precisely quantify the frequency of infected cells containing inducible HIV-1.MethodsWe developed Specific Quantification of Inducible HIV−1 by RT-LAMP (SQuHIVLa), leveraging the high sensitivity and specificity of RT-LAMP, performed in a single reaction, to detect and quantify cells expressing tat/rev HIV-1 multiply spliced RNA (msRNA) upon activation. The LAMP primer/probe used in SQuHIVLa was designed to exclusively detect HIV-1 tat/rev msRNA and adapted for different HIV-1 subtypes.ResultsUsing SQuHIVLa, we successfully quantify the inducible viral reservoir in CD4+ T cells from people living with HIV-1 subtypes B and C on cART. The assay demonstrates high sensitivity, specificity, and reproducibility.ConclusionsSQuHIVLa offers a high throughput, scalable, and specific HIV-1 reservoir quantification tool that is amenable to resource-limited settings. This assay poses remarkable potential in facilitating the evaluation of potential interventional strategies toward achieving HIV-1 cure.

Machine learning approaches identify immunologic signatures of total and intact HIV DNA during long-term antiretroviral therapy.

Understanding the interplay between the HIV reservoir and the host immune system may yield insights into HIV persistence during antiretroviral therapy (ART) and inform strategies for a cure. Here, we applied machine learning approaches to cross-sectional high-parameter HIV reservoir and immunology data in order to characterize host-reservoir associations and generate new hypotheses about HIV reservoir biology. High-dimensional immunophenotyping, quantification of HIV-specific T cell responses, and measurement of genetically intact and total HIV proviral DNA frequencies were performed on peripheral blood samples from 115 people with HIV (PWH) on long-term ART. Analysis demonstrated that both intact and total proviral DNA frequencies were positively correlated with T cell activation and exhaustion. Years of ART and select bifunctional HIV-specific CD4 T cell responses were negatively correlated with the percentage of intact proviruses. A Leave-One-Covariate-Out (LOCO) inference approach identified specific HIV reservoir and clinical-demographic parameters, such as age and biological sex, that were particularly important in predicting immunophenotypes. Overall, immune parameters were more strongly associated with total HIV proviral frequencies than intact proviral frequencies. Uniquely, however, expression of the IL-7 receptor alpha chain (CD127) on CD4 T cells was more strongly correlated with the intact reservoir. Unsupervised dimension reduction analysis identified two main clusters of PWH with distinct immune and reservoir characteristics. Using reservoir correlates identified in these initial analyses, decision tree methods were employed to visualize relationships among multiple immune and clinical-demographic parameters and the HIV reservoir. Finally, using random splits of our data as training-test sets, machine learning algorithms predicted with approximately 70% accuracy whether a given participant had qualitatively high or low levels of total or intact HIV DNA. The techniques described here may be useful for assessing global patterns within the increasingly high-dimensional data used in HIV reservoir and other studies of complex biology.

Study on the Damage and Failure Process of Prefabricated Hole-Fracture Combination Defects in Shale

The rock formations in oil and gas reservoirs are predominantly composed of laminated shale, which contains various defects such as pores and fractures. These defects have a significant impact on the stability of wellbore walls. As a result, this study utilized rock samples from a specific oil and gas reservoir and introduced pre-existing pore and fracture defects within them. Uniaxial compression tests were conducted on rock specimens with varying angles between the fractures and the bedding planes. The study involved measuring and analyzing the impact of pre-existing defect morphology on shale’s mechanical properties. Additionally, through the use of digital image correlation (DIC) technology, a comprehensive strain field map was obtained, depicting the initiation, propagation, and ultimate failure of surface cracks in shale under loading conditions. This allowed for both qualitative and quantitative analysis of the connection between shale’s damage and failure processes and the evolution of strain fields. Ultimately, this research provides a theoretical basis for wellbore stability and the development of shale oil and gas fields. Conclusions drawn from the study are as follows. With an increase in fracture angle, the rock’s elastic modulus gradually increases, and both compressive strength and strain exhibit a pattern of higher values on both ends and lower values in the middle. When the fracture angle is less than 30°, significant stress concentration occurs at the tip of the fracture. When the angle exceeds 60°, stress concentration around pores dominates. In the range of 30° to 60°, there is a combined stress concentration around both pores and fractures, significantly reducing rock stability. Crack propagation is influenced by bedding planes and exhibits varying degrees of ductility, ultimately resulting in a tension-shear mixed failure. When fractures are parallel to bedding planes (angle = 0°), defects are symmetrically distributed, and stress concentration at the fracture tip dominates, leading to crack initiation from the fracture tip and eventually forming an “H”-shaped tensile failure. When fractures are perpendicular to bedding planes (angle = 90°), stress concentration around pores is greater than at the fracture tip, causing cracks to initiate around pores and eventually collapsing on one side of the fracture.

A comparative study of brine solutions as completion fluids for oil and gas fields

Completion fluids play a vital role in well-related processes within the oil extraction industry. This article presents a comprehensive study of the properties and performance of various brine solutions as completion fluids for different well and reservoir conditions. Attributes examined include density, corrosion resistance, temperature stability, compatibility with formation fluids, clay swelling potential and influence on wettability. The research highlights the significance of selecting appropriate completion fluids to optimize well and reservoir operations. Zinc chloride emerges as an excellent option for high density applications, while sodium chloride and potassium formate solutions are ideal for extreme cold conditions. Potassium acetate outperforms calcium chloride and potassium chloride and has excellent pH stability. The compatibility of completion fluids with formation water has been observed to be excellent, with no sedimentation or emulsion formation. Potassium acetate also experiences minimal clay swelling, making it suitable for clay-rich formations. On the other hand, calcium chloride has a higher clay swelling than most of the brines tested, making it less suitable for sandstone formations with a higher clay content than these brines. The research evaluates the water-wetting abilities of completion fluids in carbonate and sandstone formations. Potassium chloride and zinc chloride have the most significant impact in carbonate formations, while potassium acetate and potassium formate excel in sandstone formations. This study provides a comprehensive understanding of completion fluids, facilitating informed decisions that maximize operational efficiency, protect reservoir integrity, and enhance hydrocarbon recovery. The appropriate selection of completion fluids should align with specific well and reservoir conditions, considering the priorities of the application.

Post-Fracture Production Prediction with Production Segmentation and Well Logging: Harnessing Pipelines and Hyperparameter Tuning with GridSearchCV

As the petroleum industry increasingly exploits unconventional reservoirs with low permeability and porosity, accurate predictions of post-fracture production are becoming critical for investment decisions, energy policy development, and environmental impact assessments. However, despite extensive research, accurately forecasting post-fracture production using well-log data continues to be a complex challenge. This study introduces a new method of data volume expansion, which is to subdivide the gas production of each well on the first day according to the depth of logging data, and to rely on the correlation model between petrophysical parameters and gas production to accurately combine the gas production data while matching the accuracy of the well-log data. Twelve pipelines were constructed utilizing a range of techniques to fit the regression relationship between logging parameters and post-fracture gas production These included data preprocessing methods (StandardScaler and RobustScaler), feature extraction approaches (PCA and PolynomialFeatures), and advanced machine learning models (XGBoost, Random Forest, and neural networks). Hyperparameter optimization was executed via GridSearchCV. To assess the efficacy of diverse models, metrics including the coefficient of determination (R2), standard deviation (SD), Pearson correlation coefficient (PCC), mean absolute error (MAE), mean squared error (MSE), and root-mean-square error (RMSE) were invoked. Among the several pipelines explored, the PFS-NN exhibited excellent predictive capability in specific reservoir contexts. In essence, integrating machine learning with logging parameters can be used to effectively assess reservoir productivity at multi-meter formation scales. This strategy not only mitigates uncertainties endemic to reservoir exploration but also equips petroleum engineers with the ability to monitor reservoir dynamics, thereby facilitating reservoir development. Additionally, this approach provides reservoir engineers with an efficient means of reservoir performance oversight.

Study on the Division of Remaining Recoverable Reserves Abundance in Specific Reservoirs

Study on the Division of Remaining Recoverable Reserves Abundance in Specific Reservoirs

Evaluating the Impact of Biochemical Reactions on H2 Storage in Depleted Gas Fields

Summary Underground hydrogen storage (UHS) is an emerging technology to store energy, produced by renewable sources, in subsurface porous formations. UHS efficiency in depleted gas reservoirs can be affected by H2 biochemical degradation due to interactions with rock, brine, and gas. In the reservoir, subsurface microorganisms can metabolize H2 with possible hydrogen losses, H2S production, clogging, and formation damage. In this work, we investigate the impact of hydrogen losses due to microbial activities on UHS operations in depleted gas reservoirs lying in sandstone formations. We developed a workflow to exploit the chemical reactive transport functionalities of a commercial reservoir simulator, to model biochemical processes occurring in UHS. Kinetic chemical reaction formulation was used to replicate a Monod’s type microorganism growth, using PHREEQC to tune reaction parameters by matching a 0D process in an ideal reactor. Then, we applied the methodology to evaluate the impact of biotic reactions on UHS operations in depleted gas fields. Eventually, various sensitivities were carried out considering injection/production cycle lengths, cushion gas volumes, and microbial model parameters. Benchmark against PHREEQC demonstrated that, by properly tuning the kinetic reaction model coefficients, we are capable of adequately reproducing Monod-like growth and competition of different microbial community species. Field-scale results showed that hydrogen losses due to biochemistry are limited, even though this may depend on the availability of reactants in the specific environment: In this work, we focus on gas reservoirs where the molar fraction of the key nutrient, CO2, is small (≤2%) and the formation is a typical sandstone. Operational parameters (e.g., storage cycle length) have an impact on the biochemical dynamics and, then, on the hydrogen degradation and generation of undesired byproducts. Similar considerations hold for the model microbial growth kinetic parameters: In this study, they were established using available literature data for calibration, but we envisage tuning them using experimental results on specific reservoirs. The current model setup does not account for rock-fluid geochemical interactions, which may result in mineral precipitation/dissolution affecting the concentration of substrates available for biotic reactions. Nonetheless, it can provide an estimate of hydrogen consumption during storage in depleted gas reservoirs due to microbial activities. This study is among the first attempts to evaluate the impact of hydrogen losses by the presence of in-situ microbial populations during hydrogen storage in a realistic depleted gas field. The assessment was performed by implementing a novel workflow to encapsulate biochemical reactions and bacterial dynamic growth in commercial reservoir simulators, which may be applied to estimate the efficiency and associated risks of future UHS projects.

Forecasting Gas Well Classification Based on a Two-Dimensional Convolutional Neural Network Deep Learning Model

In response to the limitations of existing evaluation methods for gas well types in tight sandstone gas reservoirs, characterized by low indicator dimensions and a reliance on traditional methods with low prediction accuracy, therefore, a novel approach based on a two-dimensional convolutional neural network (2D-CNN) is proposed for predicting gas well types. First, gas well features are hierarchically selected using variance filtering, correlation coefficients, and the XGBoost algorithm. Then, gas well types are determined via spectral clustering, with each gas well labeled accordingly. Finally, the selected features are inputted, and classification labels are outputted into the 2D-CNN, where convolutional layers extract features of gas well indicators, and the pooling layer, which, trained by the backpropagation of CNN, performs secondary dimensionality reduction. A 2D-CNN gas well classification prediction model is constructed, and the softmax function is employed to determine well classifications. This methodology is applied to a specific tight gas reservoir. The study findings indicate the following: (1) Via two rounds of feature selection using the new algorithm, the number of gas well indicator dimensions is reduced from 29 to 15, thereby reducing the computational complexity of the model. (2) Gas wells are categorized into high, medium, and low types, addressing a deep learning multi-class prediction problem. (3) The new method achieves an accuracy of 0.99 and a loss value of 0.03, outperforming BP neural networks, XGBoost, LightGBM, long short-term memory networks (LSTMs), and one-dimensional convolutional neural networks (1D-CNNs). Overall, this innovative approach demonstrates superior efficacy in predicting gas well types, which is particularly valuable for tight sandstone gas reservoirs.

Optimal operation of the dam reservoir in real time based on generalized structure of group method of data handling and optimization technique

The historical data on water intake into the reservoir is collected and used within the framework of a deterministic optimization method to determine the best operating parameters for the dam. The principles that have been used to extract the best values of the flow release from the dam may no longer be accurate in the coming years when the inflow to dams will be changing, and the results will differ greatly from what was predicted. This represents this method’s main drawback. The objective of this study is to provide a framework that can be used to guarantee that the dam is running as efficiently as possible in real time. Because of the way this structure is created, if the dam’s inflows change in the future, the optimization process does not need to be repeated. In this case, deep learning techniques may be used to restore the ideal values of the dam’s outflow in the shortest amount of time. This is achieved by accounting for the environment’s changing conditions. The water evaluation and planning system simulator model and the MOPSO multi-objective algorithm are combined in this study to derive the reservoir’s optimal flow release parameters. The most effective flow discharge will be made feasible as a result. The generalized structure of the group method of data handling (GSGMDH), which is predicated on the results of the MOPSO algorithm, is then used to build a new model. This model determines the downstream needs and ideal release values from the reservoir in real time by accounting for specific reservoir water budget factors, such as inflows and storage changes in the reservoir. Next, a comparison is drawn between this model’s performance and other machine learning techniques, such as ORELM and SAELM, among others. The results indicate that, when compared to the ORELM and SAELM models, the GSGMDH model performs best in the test stage when the RMSE, NRMSE, NASH, and R evaluation indices are taken into account. These indices have values of 1.08, 0.088, 0.969, and 0.972, in that order. It is therefore offered as the best model for figuring out the largest dam rule curve pattern in real time. The structure developed in this study can quickly provide the best operating rules in accordance with the new inflows to the dam by using the GSGMDH model. This is done in a way that makes it possible to manage the system optimally in real time.