02. Master's Thesis
Permanent URI for this collection
Browse
Recent Submissions
Item Embargo ASSESSMENT OF THE PERFORMANCE OF NATURAL POLYMER ASSISTED BY NANOPARTICLES(Nazarbayev University School of Mining and Geosciences, 2025-04-21) Zhuniskenov, YermukhanWith the increase in energy demand the natural resources like oil and gas are depleting steadily. Maximizing the potential of the old reservoir is an important, future oriented task. This task goes alone with the preservation of the environment. One of the EOR methods utilizes chemicals to inject into reservoir that helps to displace oil. Natural materials can very much be the material for chemical injections. Some byproducts of food or other organic plants have polysaccharides in their composition. These polysaccharides are considered as polymers, that in fact can increase the viscosity of the injective fluids and replace the synthetic polymers. However, biopolymers have drawbacks related to the vitality within harsh conditions of the reservoir and need assistance via other components. One of the new, promising methods is to combinate polymers with nanoparticles. Nanoparticles are particles with the scale of 1-100 nm that can alter wettability, reduce interfacial tension (IFT) and increase stability of other materials. Despite having evaluations of the nanoparticle effect on polymers, the literature lacks evaluation of nanoparticles in combination with biopolymers. Therefore, the aim of this thesis is to evaluate the effect of nanoparticles modification on biopolymer. Where nanoparticles were chosen to be SiO2 and ZnO metal oxide nanoparticles because they can improve rheological behavior under salinity and temperature. Natural polymer of Flaxseed gum previously evaluated under reservoir conditions in Nazarbayev University and Xanthan gum to compare results with established polymer. First, modified polymers were characterized to proof the adsorption and interaction with the polymer. FTIR, SEM and TEM were used to characterize modification. Samples undergone rheological tests such as analysis of concentration, temperature, salinity and stability. Then efficiency checked in core flooding experiments. Six experiments were done with Flaxseed and Xanthan gum modifications at ambient temperature and 35,000 ppm. Three core flooding experiments were conducted with Flaxseed gum and its modifications with silica and zinc oxide nanoparticles under close to reservoir conditions of 55oC and 35,000 ppm. Core flooding under harsh reservoir environment help to evaluate nanoparticles effect under worst salinity concentration and common Kazakhstan oil reservoir where the salinity and temperature were elevated. Silica nanoparticle modification worked best for both polymers especially in salinity. ZnO nanoparticle possibly was not dispersed in the medium well and showed aggregation. Moreover, ZnO nanoparticles showed negative impact on rheology at elevated temperature and salinity. Silica modified polymer showed the best results in the core flooding as incremental oil recovery of 4% and 7% was for, flaxseed and xanthan modifications. Zinc oxide showed less recovery at high temperature than Flaxseed.Item Open Access APPLICATION OF FOAM FLOODING FOR RESIDUAL OIL PRODUCTION IN THE “X” OILFIELD(Nazarbayev University School of Mining and Geosciences, 2025-04-25) Verendeyev, StepanFoam is a known diverting agent capable of improving sweep efficiency. It is speculated that foam flooding can also improve microscopic displacement efficiency in the reservoir core by the same mechanism. Recovery of trapped residual hydrocarbon fluid in the oilfield “X” core is investigated in this thesis. To quantify the recovery of bypassed resources on the core level, the application of foam flooding was investigated in the reservoir core and contrasted with the same recovery process sequence on homogeneous sandstone and carbonate quarried outcrop core. The study was the first to identify promising foam formulations using atmospheric bulk foam stability tests and use such foaming formulations with and without nanoparticle addition in immiscible core flood tests with nitrogen (N2) and carbon dioxide (CO2) carrier fluids. Screening experiments of bulk foam stability were conducted as a precursor to core flooding studies. Foam half-life tests substantiated a negative influence of increasing salinity (NaCl) on foam stability. Appropriate baseline production from sequential water flooding and gas flooding preceded foam flooding assessments with both nitrogen and carbon dioxide gas options. Such tests were conducted in both reservoir and outcrop cores. Oil recovery and resistance factor observations were compared between reservoir and outcrop core, N2 and CO2 gas, and foam formulations with and without nanoparticle addition. The X oilfield experiments confirmed significant EOR potential with foam processes, especially by adding nanoparticles. Baseline studies on gas flooding prior to foam introduction showed that CO2 gas flooding is more efficient than N2 flooding, even at low pressure, presumably due to the presence of additional interaction mechanisms with CO2. Whether nanoparticles were present or not, the data showed that N2 foam was more efficient than CO2 foam, as evidenced by higher Recovery-8.15 % of Original Oil in Place (OOIP) with N2 foam in “X” core samples compared to 2.45 % with CO2 foam. Moreover, recovery was improved with the addition of NP to the foam formulation, resulting in an average additional recovery factor of 5.05%. However, the higher apparent viscosity of N₂ foam with nanoparticles (about 9 cP) must be carefully considered before field usage due to the associated mobility reduction, despite the dramatic increase in the recovery efficiency. In general, from the findings of this study, it appears that foam flooding is a feasible approach for enhancing sweep in the "X" oilfield.Item Open Access APPLICATION OF POLYMER FLOODING IN HIGH SALINITY CONDITIONS IN UZEN FIELD(Nazarbayev University School of Mining and Geosciences, 2025-04-23) Tursymat, OmirzhanThis research investigates improved oil production techniques for high-salinity oil reservesfocusing on Uzen Field in Kazakhstan. According to standard primary and secondary recoverypractices current operations in mature reservoirs recover only limited quantities of the initial oilbefore EOR techniques become essential. Polymer flooding represents an advanced EOR methodwhich offers successful viscosity increase for water injection which leads to superior sweepefficiency together with reduced water output. The extreme salinity of formation water in the Uzen Field (TDS reaches 60,000 ppm) createsconditions that degrade polymers and decreases their viscosity. Through lab experimentsdeveloped for real reservoir conditions this research thoroughly screened polymer materialsutilizing hydrolyzed polyacrylamide (HPAM) as the main focus. The study performed laboratorywork by conducting tests for rheology and thermal stability and static adsorption and injectivitytesting. Additionally, ion modification experiments were performed to determine the effects ofselectively adding Na2SO4, CaCl2, and MgCl2 on polymer performance without altering the overallsalinity of the formation water. As evidenced through experimental testing appropriate control of ions results in majorimprovements of polymer stability alongside maintaining viscosity levels. The addition of 4× Mg²⁺ions produced better injectability by generating elevated resistance factors and decreased viscosityloss during core flooding assessments. A high Ca²⁺ content affected the repeated configurations ofpolymer chains which caused viscosity levels to deteriorate. The addition of measured Na2SO4amounts that rose sulfate concentration moderately strengthened polymer performance yet raisingthe amount too high decreased viscosity by damaging the polymer chain structure. Multiple studiesincluding Jouenne (2020) and Saha et al.(2021) and Hu et al.(2024) confirm that achieving theright balance of ions is essential for optimizing polymer flooding operations under high-salinityconditions. Field research conducted at Daqing, Shengli, Mangala, Pelican Lake, and Burgan locations aroundthe world validates how polymer flooding solves high water cut conditions in heterogeneousreservoirs. The examples show that proper polymer choice together with customized ion controlstrategies result in significant enhancement of oil recovery when dealing with challenging extremereservoir conditions. The study introduces an effective method to enhance polymer performance which mergeslaboratory testing methods with ion modification techniques. The obtained results help understandpolymer flooding techniques while providing essential discoveries for enhancing Uzen Field oilextraction. Follow-up pilot testing combined with sustained monitoring should be conducted toimprove polymer formulation methods and injection practices leading to efficient sustainable EOR operations.Item Embargo ADVANCED MACHINE LEARNING FOR PREDICTING CO₂ SOLUBILITY AND DIFFUSION IN RESERVOIR FLUIDS: INTEGRATING DATA-DRIVEN AND PHYSICS-BASED APPROACHES FOR CARBON SEQUESTRATION AND ENHANCED OIL RECOVERY(Nazarbayev University School of Mining and Geosciences, 2025-04-24) Hassan, SuleimanAccurate prediction of carbon dioxide (CO₂) solubility and diffusion in reservoir fluids is essential for designing effective carbon sequestration and enhanced oil recovery (EOR) strategies. This thesis presents a unified machine learning framework developed to predict CO₂ solubility and diffusion coefficients across two distinct reservoir fluid systems: (1) crude oils and liquid hydrocarbons, and (2) pure water and saline brines. Four data-driven models were applied, including Artificial Neural Networks (ANN), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Regression (SVR). Each model was trained, optimized, and evaluated independently for each fluid-property combination. In addition, a Physics-Informed Neural Network (PINN) was implemented to model CO₂ diffusion in aqueous systems by incorporating governing equations, including the Stokes–Einstein and Arrhenius relations, directly into the loss function during training. The compiled datasets include over 3,000 experimental records from more than 80 published references, covering a wide range of conditions. These span temperatures from 252 to 513 K, pressures from 0.1 to 200 MPa, salinities up to 6.8 mol/kg, viscosities up to 224,500 cP, and densities from 443 to 1026 kg/m³. The datasets contain limited measurements at high salinity and pressure ranges, which limits the model’s ability to generalize accurately in those conditions. A structured preprocessing pipeline was applied, including outlier handling, salinity and viscosity grouping, engineered features, and categorical encodings tailored to each fluid system. Among the data-driven models, ANN and XGBoost showed stronger performance than RF and SVR, which can be attributed to their ability to capture nonlinear dependencies and generalize well across diverse input conditions., with R² values exceeding 0.98 for most tasks. The PINN achieved a test R² of 0.992 on the aqueous diffusion dataset, along with an 80% reduction in RMSE and a 75% reduction in MAE compared to the ANN model. When compared to XGBoost, RMSE and MAE dropped by approximately 89% and 84%, respectively. These improvements were achieved while preserving physical consistency through integration of the Stokes-Einstein and Arrhenius relations. The PINN also showed strong generalization in regions with limited experimental data, offering a distinct advantage over conventional models in low-data settings. This research shows that modern machine learning methods, when combined with structured data and guided by physical principles, can provide accurate, generalizable, and interpretable tools for CO₂ transport prediction. The developed models, curated datasets, and the PINN framework offer practical value for fluid screening, digital reservoir simulation, and the design of CO₂ storage and EOR operations.Item Embargo MITIGATING CLAY SWELLING AND PERMEABILITY LOSS IN THERMAL EOR WITH A QUATERNARY AMMONIUM CLAY STABILIZER UNDER HIGH-TEMPERATURE LOW-SALINITY CONDITIONS(Nazarbayev University School of Mining and Geosciences, 2025-04-21) Labak, AishaClay swelling and fines migration pose major challenges to thermal enhanced oil recovery (EOR), especially in clay-rich reservoirs like the East Moldabek Field in Kazakhstan. While previous studies offer limited solutions under high-temperature, low-salinity (HTLS) conditions, this study addresses this gap by evaluating three different clay inhibitors to mitigate formation damage. Through static swelling tests, a quaternary ammonium-based inhibitor demonstrated superior performance via ion exchange and surface charge stabilization. Among the inhibitors tested, it achieved the highest permeability retention maintaining 48.89% in distilled water at 100°C, compared to only 17.05% in untreated samples. It also reduced the Critical Salt Concentration (CSC) from 7380 ppm to 4920 ppm. Scanning Electron Microscopy (SEM) confirmed structural integrity in treated clays, contrasting with severe delamination in untreated samples. Dynamic core-flooding tests further validated its effectiveness by reducing pressure drops, minimizing fines migration, and enhancing permeability retention under HTLS conditions. Performance decreased in high-salinity environments due to natural ionic stabilization. While field-scale studies are needed to assess long-term performance and economic viability, the quaternary ammonium inhibitor offers a promising solution for controlling clay swelling, preserving permeability, and improving thermal EOR efficiency in challenging reservoir conditions.Item Open Access GEOSPATIAL ANALYSIS AND OVERVIEW OF TAILINGS STORAGE FACILITIES IN KAZAKHSTAN: A CASE STUDY OF VARVARA TAILINGS STORAGE FACILITY(Nazarbayev University School of Mining and Geosciences, 2025-04-22) Kairat, AkbotaTailings Storage Facilities (TSFs) are essential to mining operations but pose ongoing environmental and structural risks. This study investigates the Varvara TSF in northern Kazakhstan using geospatial analysis of Sentinel-2 and Landsat 8 satellite imagery from 2019 to 2024. Spectral indices including NDWI, mNDWI, and NDVI were applied to detect changes in the supernatant water pond and wet tailings, both of which are critical indicators of facility stability. The results reveal a consistent expansion of the water pond, with its proximity to the dam infrastructure raising concerns, especially during spring flood events. The spatial extent of water and tailings was highly responsive to fluctuations in precipitation and temperature, highlighting the vulnerability of the facility to climatic variability. The study further critiques the limitations of traditional slurry-based tailings management under such environmental stressors. Despite improvements introduced after the 2016 dam incident, the findings emphasize the need for adaptive strategies, including real-time satellite monitoring, adoption of dewatered tailings technologies, and stronger containment measures. Additionally, the absence of a comprehensive national TSF inventory hampers effective risk governance in Kazakhstan. This research advocates for improved regulatory frameworks and sustained monitoring to ensure safer and more resilient tailings management.Item Embargo INTEGRATION OF MACHINE LEARNING AND GEOSTATISTICS FOR DOMAINING: A DATA AUGMENTATION PRACTICE IN A TAILING STORAGE FACILITY(Nazarbayev University School of Mining and Geosciences, 2025-04-23) Karakozhayeva, AyanaTailings Storage Facilities (TSFs) pose a multifaceted problem in mining owing to their environmental ramifications. One of the challenges in managing sulfidic TSFs is the presence of elevated sulfur (S) and iron (Fe) levels, which can lead to environmental contamination. This occurs through the generation of acid mine drainage (AMD), impacting surrounding soils, water, and vegetation. Traditional geostatistical techniques, however, struggle to accurately delineate compact and contiguous areas of these zones, often resulting in patchy and irregular clusters that are challenging to interpret and manage. This thesis introduces a novel approach that integrates machine learning (ML) and data augmentation with geostatistical simulations, incorporating variogram component filtering to delineate compact hazardous zones within the studied domains more effectively.Item Open Access IMPLEMENTATION OF ARTIFICIAL INTELLIGENCE IN SHORT-TERM MINE PLANNING OF OPEN PIT MINE. A CASE STUDY OF BOZSHAKOL MINE.(Nazarbayev University School of Mining and Geosciences, 2025-04-28) Kairatuly, NurassylUsing Bozshakol Mine as a case study, this thesis investigates how Artificial Intelligence (AI) is applied in short-term mine planning. Open-pit mining operations, like those at Bozshakol, struggle to develop flexible, responsive mine plans that fit with the changing dynamics of the mining environment. Often relying on static, rule-based systems, traditional approaches find it difficult to handle real-time data and unanticipated operational interruptions. This study offers an AI-driven solution to mine planning that combines machine learning and optimization techniques to increase planning accuracy, lower human error, and optimize Net Present Value (NPV) in order to meet these difficulties. The project creates a web-based artificial intelligence model that analyzes data from the Bozshakol Mine, including block models, mining face polygons, and operational factors such as ore grade, mining costs, and haulage distances. The AI system aims to produce best mining scenarios, hence maximizing NPV and reducing running costs. The performance of the AI model is compared with conventional human-generated plans, showing notable gains in NPV, time efficiency, and operational flexibility. By means of dynamic, data-driven decision-making, this thesis emphasizes how artificial intelligence might transform short-term mining planning. The results imply that artificial intelligence can simplify planning procedures, cut scenario evaluation time, and boost general profitability. The study also points out important issues including data quality and system integration and offers advice for more progress in AI-augmented mine planning.Item Open Access NUMERICAL STUDY OF GEOMECHANICAL FAULT REACTIVATION AND CAPROCK SYSTEM INTEGRITY IN CO₂ STORAGE: A SENSITIVITY ANALYSIS USING FULLY COUPLED FLOW-DEFORMATION THREE-DIMENSIONAL DISTINCT ELEMENT MODELLING(Nazarbayev University School of Mining and Geosciences, 2025-05-15) Maratov, TorekeldiTo examine fault reactivation and caprock seal integrity during CO₂ injection, a fully coupled flow-deformation model is used to evaluate geomechanical performance. The In Salah greenhouse gas storage project in Algeria is used as a field example to investigate fault slip and potential for induced seismicity across geomechanical and operational conditions. A sensitivity analysis is conducted with variations in fault geomorphology, in-situ stress ratio, injection pressure, and rate. Results indicate fault slip is sensitive to fault dip angles, where faults with 15° fault dips experience the largest shear displacement (1.07×10⁻² m) relative to higher dipping fault angles (60° faults experience maximum slip of 2.24×10⁻³ m). In addition, variations in the in-situ horizontal-to-vertical stress ratio revealed that raising this ratio from 1.0 to 2.0 led to a 50% decrease in vertical displacement, meaning horizontal stresses are beneficial for caprock containment. The evaluated parameters of fluid injection revealed that the injection rate affects fault stability; increasing the rate from 5 l/s to 6 l/s increased fault slip by 55%, with the resultant maximum displacements of 1.34 mm and 2.08 mm, respectively. Yet, increasing fluid viscosity by 20% increased fault slip from 1.34 mm to 1.26 mm, meaning increased fluid viscosity equals reduced fluid migration and pressure stabilization, thus decreasing fault reactivation likelihood. Furthermore, the results showed that caprock strength controls deformation; deformation occurred in the form of 0.0004 m vertical displacement up for weak caprocks, but no vertical displacement occurred for strong caprocks. Therefore, this shows caprock failure is dependent upon stress regime and fluid properties with a possibility of failure via leakage. Ultimately, this study contributes to the body of knowledge about the complex geomechanical and fluid interactions associated with injecting/sequestering CO₂ into/from subsurface geological formations. It highlights fault geometry as well as the relative location of injection, injection volume and pressure, and caprock strength as important factors of stress manipulation for successful sequestration safety over the long termItem Open Access EVALUATING BOREABILITY OF METAMORPHIC ROCKS: INTEGRATING TEXTURE COEFFICIENT, ROCK ABRASIVITY, AND OPERATIONAL PARAMETERS FOR ENHANCED TBM PERFORMANCE PREDICTION(Nazarbayev University School of Mining and Geosciences, 2025-05-03) Smirnov, GlebThe efficiency and cost-effectiveness of mechanized tunneling in hard rock conditions are closely tied to the geological characteristics of the rock mass, particularly its texture and mechanical properties. This study investigates the potential of assessing rock boreability using the Texture Coefficient (TC), which quantifies rock texture based on grain shape, orientation, and interlocking. By examining the relationship between TC, other physical rock properties, and Tunnel Boring Machine (TBM) performance, this research aims to improve the prediction of Rate of Penetration (ROP) and support effective tunneling strategies. The methodology is based on the analysis of rock samples collected from the Queens Water Tunnel excavation project in New York City, 1996-1999. Thin sections of the rock samples were examined using image analysis software (ImageJ), following the TC identification procedure outlined by Howarth and Rowlands (1987). Additionally, mechanical and physical properties, including Uniaxial Compressive Strength (UCS), Brazilian Tensile Strength (BTS), Cerchar Abrasivity Index (CAI), and others, were obtained from previously performed laboratory tests. TBM performance parameters such as penetration rate, thrust, torque, and power were gathered from field data recorded during excavation. The linear and nonlinear regression models were developed to estimate the TBM penetration rate using TC, UCS, CAI, Cutterhead Power (CP), and Alpha Angle as input variables. The analysis indicated that CP and CAI strongly influenced ROP prediction, while the contribution of other variables was less significant. The most optimal models achieved a considerable degree of fit with R² values between 0.82 and 0.85, indicating a strong correlation between the input parameters and actual TBM performance. The results show that the application of TC for TBM performance prediction models is limited in its current form. Despite the overall high predictive power of the presented models, the TC plays an insignificant role if included in the model, occasionally increasing the prediction error. Future work may explore the potential modifications of TC depending on the rock type and the automation of the TC calculation process, improving its practicality for research and engineering applications.Item Open Access MACHINE LEARNING-BASED PREDICTION OF INTERFACIAL TENSION IN CO₂-WATER AND CO₂-OIL SYSTEMS(Nazarbayev University School of Mining and Geosciences, 2025-04-22) Zhanabilev, TemirkhanAccurate IFT prediction in CO₂–water and CO₂–oil systems play an important role in enhancing the performance of carbon sequestration, enhanced oil recovery (EOR) and subsurface fluid modeling. The accurate experimental procedures require extended periods of time along with restricted system investigation parameters (Zhang et al., 2023). This research utilizes seven machine learning models to predict IFT at different thermodynamic and compositional conditions: Artificial Neural Networks (ANN), Symbolic Regression–Genetic Programming (SR-GP), Decision Trees (DT), Random Forest (RF), Gradient Boosting Regression (GBR), Support Vector Regression (SVR) and XGBoost. The methodology highlighted the development of predictive models that can accommodate the complexity of interfacial phenomena. Carefully curated datasets were used to train and test machine learning algorithms sensitive to the intricate relationship that governs CO₂–water and CO₂–oil systems. The ability of each model to handle non-linear and complex interactions was exhaustively examined to identify the best IFT prediction approach. Model training and testing involved extensive datasets consisting of CO₂–water and CO₂–oil systems defused by various parameters including pressure, temperature, salinity, salt types, oil API gravity and impurities. Standard performance metrics consisting of RMSE, MSE, R², MAPE, AIC and BIC determined the assessment of model accuracy. GBR together with XGBoost and RF proved to be the most accurate ensemble models since they achieved R² (coefficient of determination) values greater than 0.97 while delivering superior predictive capabilities in the evaluation of both systems. The predictive capabilities of ANN models increased remarkably when optimized parameters were applied to the hidden layer structure. The experimental data was verified through KDE plots alongside scatter diagrams which showed excellent correlation between predicted results. The results of sensitivity analyses demonstrated that pressure and temperature highly affect the IFT while salt types and impurities had notable effects. The presented work demonstrates that ML techniques can substitute experimental IFT measurement by providing dependable high-speed scalable options that benefit reservoir modeling and process optimization.Item Open Access MACHINE LEARNING MODELING OF WETTABILITY AND CONTACT ANGLE BEHAVIOR IN CO₂-WATER-ROCK SYSTEM(Nazarbayev University School of Mining and Geosciences, 2025-04-24) Tiyntayev, YernarThe focus of this research project is to understand how machine learning algorithms work with regarding wettability and contact angles in CO₂-water-rock combinations. A broad dataset containing static contact angles along with advancing and receding contact angle measurements originated from published experimental studies that studied different types of rocks across a range of pressures, temperatures, and salinities. The analysis implemented six machine learning algorithms, including Decision Trees, Random Forests, XGBoost, Gradient Boosting Regressor, Support Vector Machines, and Artificial Neural Networks. Both training and testing datasets received evaluation through different error metrics, which assessed the performance of the models. The GBR model delivered optimum performance in static contact angle prediction for CO₂-water-rock systems by reaching R² =0.99 value during training and R² = 0.92 value during testing. The high accuracy value shows that GBR effectively identifies intricate non linear patterns in wettability patterns. The GBR model produced the most accurate results for testing dataset advancing and receding contact angles with a calculation error rate of R² =0.93 training and R² =0.88 testing. GBR model demonstrate excellent proficiency for understanding the changing behavior of liquid wetting in such systems. Different input conditions such as pressures and temperatures and rock types and salinities were subject to sensitivity analysis to check model prediction accuracy and to investigate the impact on the advancing, receding and static contact angles. The research shows that pressure produces the largest impact on contact angle measurement whereas temperature and salinity impacts differ based on the rock type. The predictive capabilities of developed models worked effectively in predicting the contact angles of silicate, clay, carbonate and basalt and effectively captured their complex non-linear relationships.Item Open Access A HIERARCHICAL SEQUENTIAL GAUSSIAN CO-SIMULATION ALGORITHM WITH ACCEPTANCE-REJECTION SAMPLING TECHNIQUE FOR MINE TAILINGS EVALUATION(Nazarbayev University School of Mining and Geosciences, 2025-04-25) Ibraimov, AlikhanDespite the environmental impacts, it can be acknowledged that mine tailings may contain large amount of valuable and critical minerals for re-valorization or re-processing. However, evaluation of the same tailings poses a major challenge while dealing with multiple elements in the composition. Conventional methods may not appropriately capture the essence of the bivariate relationship between components of interest. To address this challenge, our research investigates the hierarchical sequential Gaussian co-simulation approach with acceptance-rejection sampling methods. This algorithm is designed towards the fulfillment of the linearity constraints dealing with two major elements among which are copper and gold. Both hierarchical and conventional co-simulation approaches were employed in the study, highlighting differences in data reproduction based on the obtained results. Notably, our findings showed that the effectiveness of the proposed algorithm is superior to those of conventional methods due to the more accurate reproduction of the linearity constraints. By using an acceptance-rejection sampling technique, the proposed technique guarantees the replication of values based on the identified linearity requirements. To apply this algorithm initially, regression analysis was performed to check the validity of linearity between copper and gold in the dataset and to obtain the coefficients of a formula which represents the required linearity constraint. Thereafter, the obtained formula with the linearity constraint was used to co-simulate the values of the copper and gold conditions to their bivariate linearity relations by either accepting or rejecting the simulated values based on whether they meet the predefined constraint. Thus, the proposed algorithm provides more accurate and reliable results in the bivariate relationships of copper and gold than the conventional co-simulation approach that does not take into account the linearity constraint. Nonetheless, the use of hierarchical Gaussian co-simulation is restricted to the chemical elements, which show a poor correlation between them.Item Embargo UTILIZATION OF THE RMR SYSTEM AND MACHINE PARAMETERS FOR PERFORMANCE ESTIMATION OF HARD-ROCK TBMS: A STATISTICAL AND MACHINE LEARNING APPROACH(Nazarbayev University School of Mining and Geosciences, 2025-04-18) Olaiya, Toluwase DanielAccurate estimation of TBM performance is crucial for project planners, as it helps predict the expected excavation time and overall project costs. The rock mass rating (RMR) system has been used widely for this purpose because it is easy to apply and accounts for the effect of rock mass discontinuities. However, TBM penetration models based on the RMR system are limited, with many existing ones developed using relatively small datasets and relying solely on RMR as the input variable, often overlooking the critical influence of machine parameters. This thesis aims to develop models for estimation of TBM penetration per revolution (Prev) using RMR and machine parameters as inputs. To achieve the study aim, a database containing 908 data points was compiled from seven hard rock tunnels in Italy and Iran. This database was further developed into 137 datasets, each containing RMR, TBM operational parameters, and performance metrics, categorized based on tunnel lithologies. After establishing the datasets, simple linear regression (SLR) was initially performed to examine the relationships between individual input and the Prev. The SLR results showed that single input variable could not provide an accurate estimation of Prev. However, all the input parameters, except the number of cutters (Nc), were statistically significant at p-values less than 0.005, leading to the exclusion of Nc from the input variables used for developing the models. After selecting the input parameters, linear and non-linear multivariable regressions were conducted using SPSS (V.28) to establish the empirical models for estimating Prev. Five non-linear models using RMR, thrust, cutterhead rotational speed, and cutterhead diameter as input variables, were introduced. The introduced non-linear models achieved an average determination coefficient (R²) of 0.75 and 0.72 on training and testing datasets, respectively. Random Forest (RF) and Extreme Gradient Boosting (XGBoost) algorithms were employed to estimate Prev using the same datasets and input variables employed for the non-linear models. The results showed that the RF models achieved an average R² of 0.91 and 0.89, respectively on the training and testing datasets, while the XGBoost models achieved an average R² of 0.93 and 0.89, respectively on the training and testing datasets. It is necessary to understand the impact and feature importance of individual input in the machine learning-based models. Therefore, a sensitivity analysis was performed using SHAP. The analysis showed that cutterhead diameter has the highest impact on Prev, followed by RPM, thrust, and RMR. Conclusion from the study is that introduced non-linear models may be used for estimating Prev under conditions similar to those observed in this study.Item Open Access STUCK PIPE PREDICTION IN DEEP WELLBORES DRILLED IN COMPLEX EVAPORITE FORMATIONS USING MLBASED INTELLIGENT CLASSIFIERS(Nazarbayev University School of Mining and Geosciences, 2025-04-21) Amanzhol, ElmiraStuck pipe events represent a very challenging and high-cost problem in drilling industry. Hence, accurate stuck pipe prediction is crucial for successful field development. Various stuck pipe prediction models using Machine Learning (ML) approaches have been developed in the past. In this study, an attempt was made to address critical problem of stuck pipe incidents prediction while drilling through deep and complex evaporite formations by ML-based models for early hazard detection. Development and testing different intelligent models performed using real field data that included actual drilling parameters along with geological information and drilling mud properties. Actual field data from 10 wells were used to train and test the models using acquired 61 data sets that consist of 610 datapoints. Five supervised classification ML algorithms were used: Logistic Regression (LR), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Trees (DT) and CatBoost. Before proceeding with training phase, an essential step of data pre-processing was conducted to identify any missing data and detecting outliers. Each model was trained and tested using the 60-40 data splitting strategy to identify stuck pipe condition at both individual hole section level and the entire field level. The developed intelligent models can identify patterns of different stuck pipe conditions including stuck pipe, non-stuck, pack-off, overpull, and tight spot. Error metrices were used to ascertain accuracy, precision, recall, and F1-score of the developed intelligent models in conjunction with ROC analysis to assess performance of the developed intelligent models. Two intelligent models were identified as the most effective classifiers for specified stuck pipe issues: DT model attained 99.59% accuracy at the field level; while, CatBoost reached 100% accuracy in hole section-based assessments. Feature importance analysis together with SHAP (SHapley Additive exPlanations) analysis showed that formation lithology is the leading factor affecting stuck pipe occurrences, as salt and clay contents remain as a primary contributors to this issue. Two key drilling parameters that control stuck pipe occurrence were identified as drilling rotational speed and flow rate. Additionally, mud weight combined with gel strength demonstrated major effects of rheological properties of the mud. SHAP analysis showed that some hole sections outcomes are locally affected by depth and clay content levels. The results of this research demonstrated that successful field development operations require interpretable intelligent models to enhance operational decision-making process to prevent different drilling hazards.Item Embargo AN ANALYTIC REPRESENTATIVE ELEMENT RATE DECLINE MODEL WITH FRACTURE AND MATRIX COUPLING IN NATURALLY FRACTURED RESERVOIR DEPLETION(Nazarbayev University School of Mining and Geosciences, 2025-04-23) Mynbayeva, TogzhanNaturally Fractured Reservoirs (NFRs) host significant global hydrocarbon reserves but pose considerable modeling challenges due to their inherent heterogeneity and complex flow dynamics. While dual-porosity concepts are widely used, many analytical models, particularly those employing representative element volumes (REVs) for rate decline analysis, simplify the problem by decoupling the fracture and matrix pressure systems. This often involves assuming instantaneous fracture pressure depletion, which may be misleading in flow regime timing in reservoirs with substantial matrix storage or moderate fracture-to-matrix permeability contrasts (kf/kx). An improved analytical REV model that explicitly couples the transient drainage of the matrix with the transient depletion of the fracture system is developed and validated in this thesis to overcome this limitation. The core novelty lies in incorporating a time-dependent average fracture pressure, Pf(t), as the dynamic boundary condition governing 1D linear matrix flow, relaxing the assumption of instantaneous fracture pressure drop. The model is formulated for a 2D REV consisting of a matrix block penetrated by two parallel symmetric fractures and explicitly includes parameters for matrix block geometry (aspect ratio a/b) and size variations (scaling factor S). Analytical solutions for average system pressure, instantaneous flow rate, and their Bourdet derivatives were derived and implemented numerically using Python programming language. Model validation confirmed that the new coupled solution converges to previous decoupled results (Hazlett et al., 2024) under limiting conditions of high permeability contrast (kf/kx > 104). However, significant and theoretically expected departures were observed at lower kf/kx ratios (≤104) and low fracture volume fractions (vf < 0.01), demonstrating the importance of coupling in these regimes. Sensitivity analyses quantified the influence of key parameters: kf/kx primarily controls the transition duration between fracture and matrix linear flow regimes; vf influences the intercepts of these regimes and transition zone magnitude; matrix aspect ratio (a/b) affects the separation and timing of depletion signatures; and the scaling factor (S) introduces a predictable time shift related to block size. Empirical correlations relating transient diagnostic features to kf/kx and vf were developed. Analysis of heterogeneous mixtures (varying a/b or S) revealed that the presence of even small fractions of slower-draining blocks disproportionately impacts the overall system response, thus indicating that a simple volume-weighted averaging can be misleading, especially for derivative analysis. According to this study's findings, the coupled analytical model that was created offers a more physically sound framework for examining the pressure and rate decline in NFRs, particularly in situations where the timescales for matrix depletion and fracture are not widely separated. The explicit inclusion of coupling and geometric factors enhances the understanding of transient behavior and highlights the complex impact of heterogeneity on production signatures.Item Open Access APPLICATION OF MACHINE LEARNING ON PREDICTING OIL RECOVERY DURING SPONTANEOUS IMBIBITION BY LOW SALINITY WATER(Nazarbayev University School of Mining and Geosciences, 2025-04-13) Bukayev, AzamatThis thesis explores how machine learning can be used to predict spontaneous imbibition recovery, the process where oil is naturally displaced by water in porous rock, such as in fractured reservoirs. Traditionally, running these lab experiments can take a lot of time, sometimes even months. Instead of waiting that long, this research gathers data from real laboratory results and applies different machine learning models to predict how much oil can be recovered. The focus is on using input parameters like core size, porosity, salinity, temperature, and more to estimate recovery performance without having to physically run the test every time. Six different models were tested: Artificial Neural Networks, Decision Tree, Gradient Boosting, Random Forest, Support Vector Machine, and Extreme Gradient Boosting. The performance of each model was evaluated based on how accurately it could predict real experimental outcomes. The Gradient Boosting model stood out as the most accurate, especially when trained on a combination of secondary and tertiary imbibition data. While the predictions were promising, the study also discusses the limitations of the current dataset and suggests that including more physical parameters like interfacial tension and contact angle could make future predictions even better. Overall, this work shows that machine learning has real potential to speed up and improve decision-making in enhanced oil recovery research.Item Open Access WAX DISAPPEARANCE TEMPERATURE MODELING WITH MACHINE LEARNING TECHNIQUES(Nazarbayev University School of Mining and Geosciences, 2025-04-23) Bayanova, AinurWax precipitation along with asphaltene and scale deposition present major problem for flow assurance in the oil production. Mitigation and prevention measures for wax precipitation requires understanding of wax forming conditions. Key parameter to define wax forming borderline is Wax Disappearance Temperature (WDT). The available methods to determine WDT include laboratory measurements, empirical correlations, thermodynamic modeling, and data-driven approaches. Laboratory measurements are costly and may logistically be challenging, thermodynamic approaches require detailed and accurate characterization of production fluid, and some may experience convergence issues. Data driven approaches provide a good alternative to earlier methods with almost no loss to accuracy of data. Use of machine learning (ML) techniques to determine WDT is reported in more recent literature, though limited studies are conducted. Use of more recent developments in ML for WDT determination are not well documented. Hence, an attempt was made to develop intelligent models using decision tree (DT) with boosting algorithms for WDT prediction: AdaBoost, Gradient Boosting Machines, XGBoost and CatBoost. Conventional Linear Regression (LR) and K-Nearest Neighbor (KNN) methods were also used for comparison with DT approaches. A detailed analysis of the input data from published WDT experimental studies in literature was performed that includes selection of input features, building dataset, validating data sources, and the input data analysis with statistical tools and graphical analysis. This research work resulted in building a database with 380 data points of experimental WDT. Overall, all DT boosting algorithms have performed better than the conventional LR and KNN techniques, with XGBoost and CatBoost being the top performers (R2 = 0.996 and RMSE = 0.791 and 0.7916, respectively). AdaBoost could be a model of choice if simplicity is preferred with negligible difference in performance (R2 = 0.9945 and RMSE = 0.9272) compared to the top performer. Model performance evaluation was based on both statistical assessment and graphical analysis such as parity plots and error distribution plots. Further trend analysis was assessed where all models indicate WDT increase with MW increase that validate models’ generalization and predicting capacity. The developed XGBoost and CatBoost models are superior to the existing models reported in literature, both thermodynamic and data-driven methods. The developed XGBoost and CatBoost models offer higher accuracy and better generalization with possible implications in flow assurance schemes involving wax.Item Open Access ANALYSIS OF PARTICULATE MATTER (PM) AND LUNG DEPOSITED SURFACE AREA (LDSA) CONCENTRATIONS IN OPERATIONAL AREAS OF A ROOM-AND-PILLAR OIL SHALE MINE(Nazarbayev University School of Mining and Geosciences, 2025-04-14) Korshunova, RuslanaParticulate matter (PM) in the context of underground mining results from various operations such as rock drilling and blasting, ore loading, hauling, crushing, dumping, and from diesel exhaust gases as well. These operations result in the formation of fine particles that can accumulate in the lungs of mineworkers. The lung deposited surface area (LDSA) concentration is a variant solution to evaluate potential health impacts. The aim of this study is to analyse PM and LDSA concentrations in operational areas of a room-and-pillar oil shale mine. Measurements were carried out by a direct-reading real-time PM monitor, Dusttrak DRX, and a multimetric fine particle detector, Naneous Partector 2, during the loading and dumping processes using the diesel engine loader. Consequently, the analysis was conducted on PM, LDSA, particle surface area concentration (SA), average particle diameter (d), particle number concentration (PNC), and particle mass (PM₀.₃), producing a few valuable correlation factors. Averaged LDSA was around 1433 μm²/cm³ and reached maximum peaks of 2140 μm²/cm³ during the loading, which was mostly related to diesel exhaust emissions, and within the dumping, 730 μm²/cm³ and 1840 μm²/cm³, respectively. At the same time, average PM₁ was about 300 μg/m³ during the loading, but within the dumping peaks, it reached up to 10,900 μg/m³. During the loading phase, particle diameter ranged from 30 to 90 nm, while during the dumping phase peaks, it varied from 90 to 160 nm. On this basis, a relationship between PNC and particle diameter has been produced to demonstrate an approximate split between diesel particulate matter (DPM) and oil shale dust diameters. The Ventsim simulation demonstrated a potential concentration of diesel and dust particles in a remote area of the mine where no direct measurements had been conducted. This study offers important data on PM and LDSA concentration that can be used for estimating potential exposure to miners at various working operations in the room and pillar oil shale mine, and will be used for air quality control in accordance with establishing toxic aerosol health effects.Item Open Access ADVANCING ROCK-TYPE SPECIFIC BRITTLE HOEK-BROWN PARAMETER “S”(Nazarbayev University School of Mining and Geosciences, 2025-05-15) Nesterov, RomanThe work analyzes sandstone mechanics through laboratory tests consisting of UCS measurements using strain gauges, acoustic emission measurements, and microscopic observations. The first goal of this study was to classify sandstone types through petrographic analysis, followed by the measurement of their compressive stress responses in order to establish rock-type-specific parameters for the Hoek-Brown failure criterion. The study used ten sandstone samples: five fine-grained sandstones and five coarse-grained sandstones. The objective of the study is to determine the relationship between grain size, matrix type, and the brittle Hoek-Brown constant “s” (i.e., cohesion constant). Macroscopic studies combined with microscopic studies provide important results about the structure of the sample, but require more precise measurements between petrographic characteristics such as grain sizes and the Hook-Brown parameter. Petrographic observation showed thin-grained samples had higher matrix material that likely included clay-rich argillaceous material based on their compact, dull appearance. The fine-grained samples studied display arkosic wacke characteristics as a result of high matrix content and poor sorting and angular grain textures, which on thin sections resemble plagioclase feldspar. The thin-section analysis showed that the two sandstone types possessed poor sorting characteristics and included polymictic grain compositions of quartz and plagioclase feldspar, together with iron oxide cement. The fine-grained sandstones contained grains measuring between 60-150 microns, whereas the coarser sandstones held grains from 100 to 300 microns, accompanied by reduced matrix occurrence. The restricted transport patterns, along with quick deposition of angular to sub-angular grain shapes, probably took place in a setting of alluvial fans. Results from UCS tests showed noticeable differences between the strength of fine- and coarse-grained sandstones. The UCS values of coarse-grained sandstone (103.7-152.6 MPa) exceed the UCS values of fine-grained sandstone (73.9-124.1 MPa) because of different microstructural elements. The better grain interlocking and stronger grain-to-grain contacts within coarse-grained sandstones improve their load-bearing capability. The distribution of stress becomes less effective in fine-grained sandstones because their matrix material includes clay-rich components, which diminishes overall cohesion. The uniaxial compressive strength decreases because poorly sorted and less angular grains found in fine-grained sandstones lead to early stress-induced crack formation. The failure strain of fine-grained sandstones proved lower based on strain gauge measurements, thus indicating these rocks would fail in a more brittle manner than the coarse-grained rocks. The damage initiation stress values are lower in the fine-grained samples. Under stress conditions, the material experiences early microcrack formation since it contains higher amounts of matrix and grain bonding because of these factors. The research demonstrates that grain size and matrix content directly affect rock strengths as well as the processes through which they fail. The research studied how to improve the brittle Hoek-Brown failure criterion through proper adjustment of its 's' parameter that represents rock material cohesion attributes. Sample analyses indicate that fine-grained sandstone has s average parameter value of 0.142, while coarse-grained sandstone has s parameter average value of 0.34. The use of s = 0.11 as a generalized cohesion constant lead to forecasting inaccuracies due to significant deviations from the actual values measured during this study, which indicates the need to calibrate the parameter "s" for the type of rock. The study shows that sandstone exhibits varying mechanical responses according to grain size variation and matrix content. The strength values for fine-grained sandstones remain lower than those of coarse-grained sandstones due to their elevated matrix fraction, along with inferior grain bonding strength. This leads to earlier material deterioration. The advanced Hoek-Brown failure criterion gives geotechnical assessments better reliability, which leads to improved underground excavation planning efficiency and safety.