DEPARTMENT OF PETROLEUM ENGINEERING

The accurate prediction of the Well Productivity Index (PI) is critical for reservoir management, production optimization, and forecasting. Traditional methods, such as analytical correlations and decline curve analysis, are often limited by simplifying assumptions that fail to capture the complexities of heterogeneous reservoirs like those in the Niger Delta. This research addresses this limitation by developing and evaluating a data-driven framework for PI prediction using machine learning (ML) on historical production data. The study implements and compares three advanced ensemble regression algorithms—Random Forest, XGBoost, and CatBoost—to predict PI from daily records of oil, gas, and water production rates and downhole pressures. A dataset of approximately 7,000 daily records from five Niger Delta wells was utilized, with the PI target variable calculated using a proxy for reservoir pressure drawdown. A clear performance hierarchy was established among the models. Random Forest yielded the weakest performance (R² = 0.18, MAE = 65.62), while XGBoost showed substantial improvement (R² = 0.78, MAE = 34.14). CatBoost emerged as the superior model, achieving exceptional predictive accuracy with an R² of 0.95, a Mean Absolute Error (MAE) of 18.96, and a Root Mean Squared Error (RMSE) of 21.02. Residual and temporal analyses confirmed that CatBoost produced unbiased, homoscedastic errors and effectively tracked the dynamic PI trends of individual wells over time. Interpretability analyses revealed that production rates (oil, gas, and water) were the most influential predictors, a finding consistent with reservoir engineering principles. However, this also highlights a methodological caveat regarding the mathematical coupling between the model's inputs and the PI target. The study concludes that CatBoost provides a robust and highly accurate model for PI prediction from routine field data, offering a significant advantage over
traditional methods for well performance monitoring and screening in the Niger Delta context.

This research focuses on modeling and simulating CO₂ injection and storage in a depleted sandstone reservoir using the CMG-GEM compositional simulator to evaluate the potential of geological carbon sequestration as a sustainable emission reduction strategy. A 3D reservoir model was constructed based on available structural and petrophysical data to replicate the dynamic behavior of CO₂ during and after injection. The simulation was performed under varying pressure and compositional conditions to assess injectivity, storage capacity, and reservoir response over a 69-year period. Results revealed effective CO₂ migration through the formation, with plume dispersion influenced by permeability variations across the ten layers. The estimated total CO₂ storage capacity of the reservoir was approximately 136,863 tonnes, indicating substantial potential for long-term containment. Pressure analysis showed a gradual and controlled buildup within safe limits, confirming caprock stability and the absence of leakage or fracture risk. Additionally, the molality plots demonstrated consistent distribution of CO₂ within the formation, with concentration stabilization after a five-year halt and resumption of injection in 2050, reflecting strong reservoir retention. Overall, the study confirms that the selected depleted reservoir can serve as a viable site for CO₂ sequestration. The findings also highlight the importance of optimizing well placement, incorporating residual and mineral trapping mechanisms, and extending simulation timeframes to improve prediction accuracy and long-term storage performance.

Tight reservoirs contain a large amount of hydrocarbon resources, but producing oil from them is often difficult because of their very low permeability and complex pore structure. Conventional water flooding is commonly used to maintain reservoir pressure and displace oil; however, in tight formations it usually results in low oil recovery due to restricted fluid flow and strong capillary forces. Because of this limitation, there is growing interest in improved water flooding techniques such as smart water injection. This study evaluates the impact of smart water injection on oil recovery in a tight sandstone reservoir using numerical reservoir simulation. A synthetic reservoir model representing a typical tight sandstone formation in the Niger Delta was developed using the Computer Modelling Group (CMG) GEM simulator. Two injection scenarios were considered under the same reservoir conditions: conventional high-salinity water flooding and low-salinity smart water flooding. The smart water case involved reducing the salinity of the injected brine in order to examine its effect on oil displacement and reservoir performance. The performance of both injection strategies was analyzed by comparing cumulative oil production, oil recovery factor, water cut, and reservoir pressure over a production period of 45 years. The simulation results show that smart water injection produced a slightly higher oil recovery compared to conventional high-salinity flooding and also delayed the increase in water cut. However, the overall improvement in recovery was relatively small because fluid movement and ion transport are limited in tight sandstone formations. The results suggest that although smart water injection can improve oil recovery to some extent in tight reservoirs, its effectiveness may be limited when used alone. Combining smart water flooding with other enhanced oil recovery methods may provide better production performance in tight formations.

The necessity of prospect evaluation is undisputed both to translate geology into figures and to expand the uncertainty implicit in hydrocarbon exploration. In this study, the hydrocarbon potential of STOKED field in offshore coastal swamp Niger delta was evaluated to obtain more information about the structures, stratigraphy and hydrocarbon potential of the field from available seismic and a suite of well logs data. The method adopted involves delineation of lithologies from Gamma ray log, Identification of hydrocarbon bearing reservoirs unit from resistivity log, well to well correlation across the field, fault interpretation and horizon mapping, time to depth conversion, Attribute extraction determination of petrophysical parameters and volumetric estimation. One Major and twelve minor faults were interpreted mapped from the well correlation carried out across the four wells in the NE-SW Direction. Two reservoirs were interpreted, and the seed grids generated three top time structure maps. The attribute maps were used to establish the diagnostic ability of 3D seismic attribute analysis in enhancing seismic interpretation and volumetric estimation. Map based volumetrics was calculated and the stock tank oil initially in place estimated is 85 mmstb for reservoir A while reservoir B was inconclusive as the area extended outside the extent of the seismic. The result from the petrophysical analysis and property modelling has shown that the reservoirs have porosity values that range from 17 - 24%, and water saturation ranges from 11 - 56%. Results from this study have shown that, away from currently producing zone at the central part of the field, additional leads and prospects exist, which could be further evaluated for hydrocarbon production.

This project investigates the effect of key drilling parameters on Rate of Penetration (ROP) using real-world field data from a selected well. The parameters analyzed include Weight on Bit (WOB), Rotational Speed (RPM), and mud properties such as Plastic Viscosity, Yield Point, and Gel Strength. The study aims to understand how variations in these parameters influence ROP and to identify combinations that could
enhance drilling efficiency. Microsoft Excel was used for organizing, calculating, and analyzing the data, with additional tools such as Solver applied for basic optimization. By focusing on a practical, data-driven approach, this work contributes to ongoing efforts in optimizing drilling operations, especially in regions where advanced software and models may be inaccessible. The findings provide insight into the practical relationships between operational parameters and ROP, and highlight opportunities for performance improvement in similar field environments

This study was carried out with the aim of delineating and establishing the lithofacies present in the well. From a depth of 4,500 feet (4500 feet) to 11,500 feet (11500 feet) from the Well ‘X’ in the Greater Ughelli Depobelt, Niger Delta Basin, Southern Nigeria, 117 ditch cutting samples were taken from the Shell Petroleum Development Company (SPDC) and processed for sedimentological analysis. We created a lithostratigraphical map for the well that included lithofacies, lithofacies units, formations, and related minerals. Four lithofacies—sandstone, sandy shale, shale, and shale sand—are revealed by the sedimentological examination of the well (4500–11500 feet) in Southern Nigeria's Niger Delta Basin. Twenty-seven (27) lithofacies units were found in the well, including the following: the Sandy Shale lithofacies at units 2, 4, 6, 8, 10, 15, 17, 20, 22, and 26; the Shale lithofacies at units 1, 5, 23, and 25; the Sandstone lithofacies at units 11, 13, 18, and 27; and the Shaly Sand lithofacies at units 3, 7, 9, 12, 14, 16, 19, 21, and 24. In conclusion, the litholog application for studying Well ‘X’ was able to delineate and establishing the lithofacies (Hydrocarbon Play Elements) that are present, as the well under study has hydrocarbon play features such as source rock, reservoir rock, caprock, traps, and seals.

In the present-day oil and gas industry, the chemicals commonly employed as pH controllers in drilling operations are largely imported at exorbitant costs. These high costs contribute significantly to the overall drilling expenditure and create rippleeffects on the national economy. Hence, there is a pressing need to identify and develop locally available substitutes that are both cost-effective and environmentally friendly. This study focuses on investigating the suitability of burnt coconut fibre, a readily available agricultural by-product in Nigeria, as a pH enhancer in drilling mud formulations. The research methodology involved the preparation of water-based mud samples treated with different concentrations of coconut fibre solution, alongside conventional additives such as sodium hydroxide (NaOH) for comparison. Laboratory analyses were conducted to determine pH variation, rheological properties, and mud density under controlled conditions. The performance of the coconut fibre was evaluated based on its ability to increase and stabilize mud alkalinity while maintaining desirable drilling fluid characteristics. The results revealed that burnt coconut fibre imparted a significant pH value of approximately 13.0 in the drilling mud, which is comparable to the 13.8 obtained with sodium hydroxide. Additionally, the coconut fibre showed potential in enhancing rheological properties, such as yield point and gel strength, while also exhibiting ecofriendly and biodegradable characteristics. These findings demonstrate that coconut fibre can serve as a viable supplement to imported chemical additives, thereby reducing dependency on foreign products, lowering drilling costs, and supporting sustainable resource utilization

Water coning and cusping are significant challenges encountered during oil production, particularly in reservoirs with bottom or edge water drives. These phenomena occur due to the unfavorable movement of water into the production wellbore, compromising oil recovery efficiency and resulting in economic and operational losses. Water coning refers to the upward movement of water from an underlying aquifer towards a vertical well, while water cusping describes a similar lateral movement of water towards a horizontal well. These occurrences are typically caused by pressure differentials induced during hydrocarbon production, where the drawdown pressure at the wellbore exceeds the critical rate, causing water to breach the oil-water contact (OWC) and flow into the well. The onset of water coning and cusping can severely reduce the oil-to-water ratio, leading to increased water production, higher separation and disposal costs, reduced oil recovery, and premature well abandonment. Understanding the mechanics of water movement in the reservoir and its interaction with the wellbore is essential for designing effective reservoir management strategies. This project explores the mechanisms, causes, consequences, and predictive models associated with water coning and cusping. It also reviews various solutions and mitigation techniques including critical production rate management, well completion optimization, reservoir simulation, and advanced recovery technologies such as horizontal drilling and inflow control devices (ICDs). The study is motivated by the increasing need to optimize hydrocarbon recovery from mature fields and minimize water handling costs, especially in the face of global energy demands and declining oil reserves. Emphasis is placed on the use of analytical and numerical methods to predict water breakthrough and design production strategies that minimize water influx. Examples of such methods used are change plot comparison, water cut vs time and cumulative oil.

This thesis presents a comprehensive machine learning-based predictive maintenance (PdM) framework for Electric Submersible Pumps (ESPs) in petroleum production, aiming to mitigate unplanned downtime and optimize maintenance scheduling. ESPs are critical for artificial lift in oil wells but are prone to failures—mechanical, electrical, and operational—that disrupt production and incur high intervention costs. Traditional maintenance strategies, whether reactive or calendar-based, fail to leverage real-time sensor data effectively. This study bridges this gap by developing data-driven models to forecast ESP failures using historical and real-time telemetry. The research begins with a detailed exploration of ESP systems, their components, and common failure modes, emphasizing the role of downhole sensors (pressure, temperature, vibration, motor current) in monitoring pump health. A year-long dataset from an onshore oilfield, comprising hourly measurements of motor current, intake/discharge pressures, temperatures, vibration frequency, production rate, and choke settings, is rigorously preprocessed. Key steps data cleaning, imputation of missing values, outlier handling, and feature engineering. Temporal features such as rolling statistics (3-hour mean/standard deviation) and lagged variables (1- and 2-hour delays) are engineered to capture dynamic system behavior. Three ensemble machine learning models—Random Forest, Gradient Boosting, and XGBoost—are benchmarked against a linear regression baseline to predict motor housing temperature one hour ahead. The models are evaluated using time-series cross-validation to prevent data leakage. Results show that Gradient Boosting achieves the lowest mean squared error (MSE: 460.26 °F²) and highest R² (0.926), outperforming the linear baseline (MSE: 477.03 °F², R²: 0.923). Feature importance analysis reveals that lagged intake pressure and motor temperature dominate predictions, accounting for over 80% of explanatory power. However, the models exhibit slight under-prediction during rapid temperature spikes, highlighting a need for further refinement in extreme-event forecasting. Unsupervised clustering identifies five operational modes (startup, steady-state, high-temperature stress, low-pressure events, and current spikes), providing context for model per formance variations during regime transitions. The study underscores the potential of hybrid approaches, combining regime classification with mode-specific regression, to enhance accuracy. The thesis concludes with actionable recommendations: deploying the optimized model in real-time monitoring systems with adaptive alert thresholds, integrating additional sensor modalities (e.g., vibration, acoustics), and establishing feedback loops for continuous model retraining. Economically, the proposed PdM framework could reduce unplanned downtime by 30–50% and lower maintenance costs by 20–40%, as evidenced by industry benchmarks. Future work includes exploring deep learning architectures (e.g., LSTMs) for longer-term dependencies and extending the framework to other artificial-lift systems. This research contributes a scalable, interpretable, and empirically validated PdM pipeline, advancing the transition from reactive to proactive maintenance in petroleum production. By harnessing sensor data and machine learning, operators can anticipate failures, optimize interventions, and maximize ESP run life—translating into significant cost savings and production efficiency gains.

This research aims to assess the performance of drilling bits by employing cost per
foot and breakeven equations as a means of evaluating their cost-effectiveness. The
study focuses on ABC Well X and ABC Well Y, both located onshore in the southern swamp region of OML 11 within the Eastern Niger Delta area. In ABC Well X, the evaluation involved the use of four SEC Bits, three HTC Bits, and one REED Bit. For ABC Well Y, the assessment incorporated three SEC Bits and three HTC Bits. To evaluate the bits, cost per foot and breakeven equations were applied, taking into consideration their respective manufacturers. The study computed the average cost per foot for each type of bit. The results revealed the following average cost per foot values: SEC Bits in ABC Well X: $58.29 per foot, HTC Bits in ABC Well X: $22.24 per foot, REED Bit in ABC Well X: $47.38 per foot, SEC Bits in ABC Well Y: $38.82 per foot and HTC Bits in ABC Well Y: $20.32 per foot. Based on the cost per foot evaluation, it is evident that HTC Bits exhibit superior
performance compared to the other types of bits, as they have the lowest cost per foot. Utilizing HTC Bits for drilling to the total depth in the field would require a total cost of $3,363.6. This investment would enable drilling for 9.108 hours and yield a footage of 78.47 feet, ultimately reaching the breakeven point for cost per foot.