With the growing importance and application of Machine Learning in various complex operations in the Oil and Gas Industry, this study focuses on the implementation of data analytics for estimating and/or validating bottom-hole pressure (BHP) of Electrical Submersible Pump (ESP) wells. Depending on the placement of the ESP in the wellbore and fluid gravity of the well fluid, there can be little or no difference between BHP and Pump intake Pressure (PIP); hence these two parameters were used interchangeably. The study focuses majorly on validating PIP when there are concerns with downhole gauge readings. It also has application in estimating PIP when the gauge readings are not available, provided the relevant ESP parameters are obtainable. ESP wells generally have gauges that operate on "Comms-on-Power" principle i.e. downhole communication is via the power cable and loss of signal occurs when there is no good electrical integrity along the electrical path of the ESP system. For proper hydrocarbon accounting and statutory requirements, it is important to have downhole pressure readings on a continuous basis, however this cannot be guaranteed throughout the life cycle of the well. Therefore, an alternative method is essential and had to be sought. In this study, the Response Surface Modelling (RSM) was first used to generate a model relating the ESP parameters acquired real-time to the PIP values. The model was fine-tuned with a Supervised Machine Learning algorithm: Artificial Neural Network (ANN). The performance of the algorithms was then validated using the R-Square and Mean Square Error values. The result proves that Machine Learning can be used to estimate PIP in a well without recourse to incurring additional cost of deploying new downhole gauges for acquisition of well and reservoir data.

Low oil price and increased environmental regulations presents a new frontier for many indigenous oil and gas companies in Nigeria. In mature fields with significant water production, produced water treatment and handling could easily account for up to a third of OPEX. Underground produced water disposal is a tested approach that has been used worldwide with mixed results. Studies have been published on the subject; however, it was observed that there were no Niger Delta case studies. This paper presents SEPLAT's subsurface approach to in-field water disposal, drawing upon geological and petroleum engineering analysis coupled with learnings from over 6 years of produced water re-injection experience. Some of the areas that will be discussed include reservoir selection/screening methodology, water quality impact on permeability, produced water disposal well selection/completion, operating philosophy, general surveillance, and basic separation requirements. Thirteen reservoirs located within 2 proximal fields were screened for suitability and ranked as possible candidates for water disposal based on 8 criteria. The best 2 were then high-graded and detailed studies carried out, spanning detailed geological characterization for reservoir quality and connectivity (including quantitative interpretation), to dynamic simulation, injection well location optimization and performance prediction (for clean water). The results of core flood tests were incorporated. It is recommended that total suspended solids should not exceed 5 mg/L, with a maximum of 5 microns particle size, under matrix injection conditions while oil content should be limited to below 30-50 ppm. Tolerance for TSS can be relaxed to 10ppm – 50ppm at fracturing conditions, depending on the reservoir parameters and process systems. The knowledge of these parameters should drive the technology selection for optimum water treatment and injection.

Accurate prediction of gas compressibility factor is essential for the evaluation of gas reserves, custody transfer and design of surface equipment. Gas compressibility factor (Z) also known as gas deviation factor can be evaluated by experimental measurement, equation of state and empirical correlation. However, these methods have been known to be expensive, complex and of limited accuracy owing to the varying operating conditions and the presence of non-hydrocarbon components in the gas stream. Recently, newer correlations with extensive application over wider range of operating conditions and crude mixtures have been developed. Also, artificial intelligence is now being deployed in the evaluation of gas compressibility factor. There is therefore a need for a holistic understanding of gas compressibility factor vis-a-vis the cause-effect relations of deviation. This paper presents a critical review of current understanding and recent efforts in the estimation of gas deviation factor.

Proper understanding of reservoir producing mechanism forms a backbone for optimal fluid recovery in any reservoir. Such an understanding is usually fostered by a detailed petrophysical evaluation, structural interpretation, geological description and modelling as well as production performance assessment prior to history matching and reservoir simulation. In this study, gravity drainage mechanism was identified as the primary force for production in reservoir X located in Niger Delta province and this required proper model calibration using variation of vertical anisotropic ratio based on identified facies as against a single value method which does not capture heterogeneity properly. Using structural maps generated from interpretation of seismic data, and other petrophysical parameters from available well logs and core data such as porosity, permeability and facies description based on environment of deposition, a geological model capturing the structural dips, facies distribution and well locations was built. Dynamic modeling was conducted on the base case model and also on the low and high case conceptual models to capture different structural dips of the reservoir. The result from history matching of the base case model reveals that variation of vertical anisotropic ratio (i.e. kv/kh) based on identified facies across the system is more effective in capturing heterogeneity than using a deterministic value that is more popular. In addition, gas segregated fastest in the high case model with the steepest dip compared to the base and low case models. An improved dynamic model saturation match was achieved in line with the geological description and the observed reservoir performance. Quick wins scenarios were identified and this led to an additional reserve yield of over 1MMSTB. Therefore, structural control, facies type, reservoir thickness and nature of oil volatility are key forces driving the gravity drainage mechanism.

The production of oil is of great and immense significance as a source of energy worldwide. The major factors affecting the production volume of oil is classified into two groups namely the geological and the human factor. Each group comprises of factors affecting oilfield production volume. The challenge in this project is to find the variable for the crude oil production volume in an oilfield because there are numerous factors affecting the crude oil production volume in an oilfield. The objective of this paper is to provide a more accurate and efficient solution on how to predict the oil production volume. Furthermore, Machine Learning algorithm called Multiple Linear Regression was developed using Python programming Language to predict the production volume of oil in an oilfield. The model was developed and fitted to train and test the factors that affect and influence the oil production volume. After a several studies have been made, the affecting factors were provided from the oilfield which would be trained and tested in order to model the relationship between predictor variable and response variable which are the significant affecting factors and the oil production volume respectively. The predictor variables are the startup number of wells, the recovery percent of previous year, the injected water volume of previous year and the oil moisture content of previous year. The predictor variable is the oil production volume. Moreover, the model was found to possess greater utility in predicting the production volume of oil as it yielded an oil production volume output with an accuracy of 98 percent. The relationship between oil production volume and the affecting factors was observed and drawn to a perfect conclusion. This model can be of immense value in the oil and gas industry if implemented because of its ability to predict oilfield output more accurately. It is an invaluable and very efficient model for the oilfield manager and oil production manager.

There are numerous problems encountered during drilling such as wellbore instability, drilling mud weight estimation, as well as selecting good casing and bit for the drilling operations. It is therefore important to understand and accurately determine the strength of the rock in order to avoid these common drilling problems which are mostly encountered during well operations. It is of paramount importance to determine uniaxial compressive strength (UCS) from core and sonic log data so as to accurately predict rock strength for better well planning. In this work, we were able to obtain a correlation to determine UCS from data obtained from ten (10) wells in different locations in onshore Niger Delta using the regression analysis method. The correlation of UCS versus Poisson's ratio gave R 2 value of 90.0%. The R 2 - value tending towards one (1) indicates that this model can be reliably used to predict ND-UCS and the p<0.05 shows that there is significant relationship between ND-UCS and Poisson's ratio. The model was validated with an entirely different well data and it predicted over 89% rock UCS data when compared to the actual rock UCS data. This study also provides an understanding of the variation in UCS and Poisson's ratio with depth for effective rock property analysis and evaluation. These correlations will help well engineers to make informed decisions on rock strength predictions during well planning and operations as well as manage wellbore stability optimally.

For the past century, optimization of drilling has caught the eyes of many researchers. The main areas center on ROP, fluid treatment, and bit selection. They all share the same goal of maximizing ROP and reducing NPT. In other to develop an optimal control system, ROP must be predicted accurately, unfortunately, it is a complex parameter that is affected by multiple drilling parameters, rock properties, fluid properties, and bit selection. Models used for prediction have developed from empirical models like Bourgoyne and Young's to more intelligent models such as SVM and ANN. With the continuous increase in data obtained from sensors while drilling, there is still much work to be done in this field. In this research, the improvement of an empirical model and the development of an intelligent model are presented. The Bourgoyne and Young's model uses multiple linear regression to estimate coefficients which it then inserts into an empirical formula to predict ROP. This model was modified using non-linear curve-fitting to estimate the coefficients and make it reduce bias to generalize better. Machine learning models such as Gradient Boosting, Random Forest, ANN, and DNN were used in the development of a predictive model for the ROP. These models were easier to develop compared to the empirical model since they rely more on data rather than statistical formulas. The data used in this research include drilling data from 3 wells drilled in 2 fields within the Niger Delta region in Nigeria. The models were developed and trained on one of the wells, while the remaining two were used for testing the performance of the models. The modified empirical model improved the efficiency of the base model by 14% during validation but performs poorly on unseen data from the other two wells. The Machine learning models outperform the empirical models and perform accurately on unseen data from the other wells. DNN was the best performing model achieving an average accuracy of 0.987 for the 3 wells.

The paper aims to estimate the tCO2e from flare stack sites in the Eastern zone of Nigeria and review air monitoring done at the flare sites with the objective of establishing a model for gas tCO2e emission and gaseous pollutants. It focuses on the South-Eastern region of Nigeria where oil and gas production are being carried out (Imo and Abia states). It zero-in on the hydrocarbon processing and handling facilities (flowstation) and the gas flared volumes. The study was carried out using representative data from an onshore flowstation in Eastern Nigeria. The data consist of gas flared volumes from year 2013-2017 and ambient gaseous emission from air quality report done on the same location. Univariate regression and correlation using Excel were carried out on yearly average ambient air quality parameters (VOC, NOx, CO, SOx, CH 4 , SPM, NH 3 , H 2 S) to check the statistical significance of each parameter as an independent variable and calculated tCO2e as the dependent variable. Excel Muti-variate linear regression method was then used to generate a predictive model for tCO2e and gaseous emission parameters. It presented a relationship between the emission from flared gas and air quality index.

An artificial neural network (ANN) was developed to predict skin, a formation damage parameter in oil and gas drilling, well completion and production operations. Four performance metrics: goodness of fit ( R 2 ), mean square error (MSE), root mean square error (RMSE), average absolute percentage relative error (AAPRE), was used to check the performance of the developed model. The results obtained indicate that the model had an overall MSE of 355.343, RMSE of 18.850, AAPRE of 4.090 and an R 2 of 0.9978. All the predictions agreed with the measured result. The generalization capacity of the developed ANN model was assessed using 500 randomly generated datasets that were not part of the model training process. The results obtained indicate that the developed model predicted 97% of these new datasets with an MSE of 375.021, RMSE of 19.370, AAPRE of 6.090 and R 2 of 0.9731, while Standing (1970) equation resulted in R 2 of −0.807, MSE of 9.34×10 16 , AAPRE of 3.10×10 6 and RMSE of 4.10×10 5 . The relative importance analysis of the model input parameters showed that the flow rates (q), permeability (k), porosity (φ) and pressure drop (Δp) had a significant impact on the skin (S) values estimated from the downhole. Thus, the developed model if embedded in a downhole (sensing) tool that capture these basic or required reservoir parameters: pressure, flowrate, permeability, viscosity, and thickness, would eliminate the diagnostic approach of estimating skin factor in the petroleum industry.

Over the years, the industry has been so used to the hard logic of utilizing flare gas meters (notably the ultrasonic flare gas meters) in the measurement of stranded flare gas. This is because it has been a workable solution for years with minimal challenges due to the broader range of accuracy required by regulatory bodies. Usually, companies are either constrained to either utilize the associated gas from the oil and gas facilities as fuel gas to power up the unit or reinject in the reservoir to serve as pressure maintenance agent that pushes the oil towards the reservoir, or stored in the reservoir and/or flare the gas (which in most cases, have been deployed by operators despite the penalties by the regulatory organization). With the recent steer in carbon capture, natural gas utilization, climate change and energy transition, accuracy level demands has been made more stringent with some countries including Nigeria requesting for 2.5 – 3% accuracy level of measurement from operators in a bid to monitor and curb the essence flue gases that are unaccounted for. This can only be for gases flared during routine conditions which does include when process upsets give rise to shut down and blowdown of gases through the flare header to the flare tip. The high demand of measurement accuracy has opened windows for OEM to produce calibrated meters that are bespoke with a longer timeline for recalibration as most of the hardware in critical operations could require a process shutdown to either maintain, repair, calibrate or even replace. With this growing concerns in the industry and the surging growth of digitalization involving AI, data analytics etc in other areas, the software method would be a potential source of synergy to assist the failing hardware which are being impacted by time as calibration issues continue to resurface throughout the life of the meters, giving rise to wider accuracy measurement in the region 5 – 10%, hence attracting the hammer from the regulators. This paper is intended to produce a deep dive of the current regulatory requirements for gas measurement in Nigeria by the regulators (DPR), the impact of the recent 3% accuracy requirements as it impacts both large and medium size operators, the role of gas measurement software for bridging the gaps and shortfalls of the hardware components. A case study of newly developed flare gas measurement software and its impact in assisting operators in gas performance reporting, production allocation and flare penalties where applicable

An evaluation of potential EOR processes applicable in the marginal oil field operation of the Niger Delta region is presented. Technical feasibility, process availability, oil recovery potential, and other uncertainties and risks associated with exploitation of enhanced oil recovery technique in a marginal oil field environment are being assessed. Few Enhanced oil recovery processes, namely polymer flooding, chemical flooding and microbial EOR (MEOR), are considered for possible application in this marginal oil field. The objective of the screening study is to evaluate and rank the EOR options and also select the most attractive method that will have to be further chased to a pilot test stage. Emphasis is strictly on a technical assessment of the incremental oil potential of each of the EOR methods and also identification of critical operational and logistical components of the entire process for their implementation in the offshore operating environment. Recoverable volumes associated with EOR may be significant, but key project development and implementation challenges and extra cost elements must be considered in any EOR forecast for an effective EOR process ranking. Some of these concerns (e.g. Polymer/chemical supply, facilities requirements, and the impact of EOR on reservoir performance and wellbore integrity) may be significant enough to eliminate a method from being considered further and at that point the best EOR option that requires minimal cost exposure for achieving the best recoverable shall be considered. Moreso, there is consideration of the quantity and quality of laboratory data that should support the viability of each EOR process being considered. This paper narrates the state of technical readiness for field implementation of each EOR method and identifies remaining work required to progress EOR process in this marginal oil field.

Faster and more accurate decisions are what the Oil and Gas industry needs with the world's fast-evolving energy needs and economy. The area of Artificial intelligence and Data-driven modelling is relatively new and has not found popular application in the industry. AI is an emerging technology that can be used to predict event outcomes and automate anomaly-detection processes. The various applications of AI in different industries were researched into. This paper highlighted important processes that can be improved with the application of Artificial Intelligence through data-driven modelling. It also highlights areas in the various industries where AI intelligence is already being applied and ways it can be improved. AI and data-driven modelling has the potential to improve exploration accuracy, reduce production down-time, reduce cost of maintenance, and reduce health and safety risks. This body of information can serve as a guideline for adopting AI in the oil and gas industry. A trend of industry-tailored intelligence solutions would be more effective in the evolving energy industry.

Several computer-aided techniques have been developed in recent past to improve interpretational accuracy of subsurface geology. This paradigm shift has provided tremendous success in variety of Machine Learning Application domains and help for better feasibility study in reservoir evaluation using multiple classification techniques. Facies classification is an essential subsurface exploration task as sedimentary facies reflect associated physical, chemical, and biological conditions that formation unit experienced during sedimentation activity. This study however, employed formation samples for facies classification using Machine Learning (ML) techniques and classified different facies from well logs in seven (7) wells of the PORT Field, Offshore Niger Delta. Six wells were concatenated during data preparation and trained using supervised ML algorithms before validating the models by blind testing on one well log to predict discrete facies groups. The analysis started with data preparation and examination where various features of the available well data were conditioned. For the model building and performance, support vector machine, random forest, decision tree, extra tree, neural network (multilayer preceptor), k-nearest neighbor and logistic regression model were built after dividing the data sets into training, test, and blind test well data. Results of metric score for the blind test well estimated for the various models using Jaccard index and F1-score indicated 0.73 and 0.82 for support vector machine, 0.38 and 0.54 for random forest, 0.78 and 0.83 for extra tree, 0.91 and 0.95 for k-nearest neighbor, 0.41 and 0.56 for decision tree, 0.63 and 0.74 for logistic regression, 0.55 and 0.68 for neural network, respectively. The efficiency of ML techniques for enhancing the prediction accuracy and decreasing the procedure time and their approach toward the data, makes it importantly desirable to recommend them in subsurface facies classification analysis.

Crude oil, a base for more than 6000 products that we use on a daily basis, accounts for 33% of global energy consumption. However, the outbreak and transmission of COVID-19 had significant implications for the entire value chain in the oil industry. The price crash and the fluctuations in price is known to have far reaching effect on global economies, with Nigeria hard. It has therefore become imperative to develop a tool for forecasting the price of crude oil in order to minimise the risks associated with volatility in oil prices and also be able to do proper planning. Hence, this article proposed a hybrid forecasting model involving a classical and machine learning techniques – autoregressive neural network, in determining the prices of crude oil. The monthly data used were obtained from the Central Bank of Nigeria website, spanning January 2006 to October 2020. Statistical efficiency was computed for the hybrid, and the models from which the proposed hybrid was built, using the percent relative efficiency. Analyses showed that the efficiency of the hybrid model, at 20 and 100 hidden neurons, was higher than that of the individual models, the latter being the best performing. The study recommends urgent diversification of the economy in order not for the nation to be plunged into a seemingly unending recession.

Machine learning methods have been applied to predict depths of fluid loss in hydrocarbon exploration.During drilling, lost circulation can be described as the unpleasant loss of all or part of drilling mud or fluid into the immediate formations or affected formation by excessive hydrostatic pressure, sufficient to fracture the formation or expand existing fractures encountered during the drilling process. In this study, we deployed Python codes of Support Vector Machine (SVM) and Decision Tree (DT) methodsto categorical data obtained from drilling operations in a producing field to predict lost circulation occurrence. The modelsleveraged the capability of both SVM and DT to achieve binary classification by adopting flow-out percentage of less than 70 percent as the points of lost circulation. That is, < 70% is represented as Loss and > 70% represented asNo Loss. Prediction models were applied to 10 input variables preprocessed with principal component analysis (PCA) to reduce dimensionality and focus on essential variables. The preprocessed SVM model gave an improved result while preprocessing does not affect DT models. Overall, DT models predicted accurate fluid losszones and can be scaled up to field operations with options ofcontinuous sampled variables.

Data analytics has only recently picked the interest of the oil and gas industry as it has made data visualization much simpler, faster, and cost-effective. This is driven by the promising innovative techniques in developing artificial intelligence and machine-learning tools to provide sustainable solutions to ever-increasing problems of the petroleum industry activities. Sand production is one of these real issues faced by the oil and gas industry. Understanding whether a well will produce sand or not is the foundation of every completion job in sandstone formations. The Niger Delta Province is a region characterized by friable and unconsolidated sandstones, therefore it's more prone to sanding. It is economically unattractive in this region to design sand equipment for a well that will not produce sand. This paper is aimed at developing a fast and more accurate machine-learning algorithm to predict sanding in sandstone formations. A two-layered Artificial Neural Network (ANN) with back-propagation algorithm was developed using PYTHON programming language. The algorithm uses 11 geological and reservoir parameters that are associated with the onset of sanding. These parameters include depth, overburden, pore pressure, maximum and minimum horizontal stresses, well azimuth, well inclination, Poisson's ratio, Young's Modulus, friction angle, and shale content. Data typical of the Niger Delta were collected to validate the algorithm. The data was further split into a training set (70%) and a test set (30%). Statistical analyses of the data yielded correlations between the parameters and were plotted for better visualization. The accuracy of the ANN algorithm is found to depend on the number of parameters, number of epochs, and the size of the data set. For a completion engineer, the answer to the question of whether or not a well will require sand production control is binary-either a well will produce sand or it does not. Support vector machines (SVM) are known to be better suited as the machine-learning tools for binary identification. This study also presents a comparative analysis between ANN and SVM models as tools for predicting sand production. Analysis of the Niger Delta data set indicated that SVM outperformed ANN model even when the training data set is sparse. Using the 30% test set, ANN gives an accuracy, precision, recall, and F1 - Score of about 80% while the SVM performance was 100% for the four metrics. It is then concluded that machine learning tools such as ANN with back-propagation and SVM are simple, accurate, and easy-to-use tools for effectively predicting sand production.

This paper presents the research work on using a machine learning algorithm to predict the viscosity of Niger Delta oil reservoirs using formation volume factor and fluid density at bubble point pressure as correlating parameters. Oil Viscosity stands out when considering the amount of oil recoverable from the reservoir hence it is an important input into the recovery factor computation, material balance analysis, reservoir simulation/history match, EOR evaluations and many other applications. Laboratory techniques of obtaining oil viscosity are quite expensive and time consuming, hence the need for various mathematical correlations developed for its estimation. Majority of the correlations make use of empirical and experimental relationships developed from analyzing oil samples to obtain a trend to predict viscosity mostly for a basin. None of these has been developed for oil viscosity for Niger Delta fluids. Viscosity has been globally defined as the resistance to shear stresses within the fluid or the resistance of the fluid molecules to deformation. For a typical reservoir fluid system, where the liquid and gas exist in dynamic equilibrium, reservoir fluid composition along with temperature and pressure has been established to determine reservoir fluid viscosity 1 . Hence for an isothermal system and at a defined pressure in the reservoir the viscosity will be dependent on largely the composition. The reservoir fluid composition is also represented by the reservoir fluid density and the formation volume factor; therefore it is possible to deduce the viscosity of reservoir fluids from the oil density and formation volume factor even though a direct relationship has not been established between these parameters. Therefore, a correlation that can establish a relationship between the specific gravity (density) and FVF with viscosity will have significant value in the oil and industry. The data used for this analysis includes viscosity, formation volume factor, oil density at 2800 sample bubble point pressure. The data was obtained by analyzing over 3500 PVT Analysis reports, extracting the data points using a python work program, cleaning up the data and removing erroneous data, performing preliminary analysis to establish baseline relationships between the data. Supervised learning using a classification tree model was used as the machine learning approach. Seven different machine learning algorithms were reviewed, and the Random Forest Regressor was selected as the most suitable algorithm for the prediction. The model prediction results were quiet encouraging as the model was able to predict viscosity within 10% deviation from the experimental viscosity for over 80% of the cases resulting in about 90% prediction accuracy. The analysis of the results further revealed that the model could better predict viscosity of Medium to Light oil with an R 2 value of between 0.90-0.96 without adjusting some obvious erroneous data points. Future of this research work will involve further in-depth analysis which will merge the preliminary QC plots with the results to evaluate the effect of the outlier sample points on the final predictability of the model. Also explore other machine learning models to further improve predictability and be able to predict viscosity across other pressure values other than the bubble point pressure to capture viscosity along the producing life of the reservoir.

To maximize production from mature fields, it is essential to identify candidate's wells that are not producing up to their potential. Performing periodic interventions or workovers in wells is an established approach for arresting production decline and maximizing production from the fields. However, for mature fields with large well counts, the process of determining the best candidates for well interventions can be complicated and tedious. This can result in less-than-optimal outcomes. Advanced data analytics modeling offers quick and easy access to important information. The main objective of this study is to identify potential candidate wells for workover operation ahead of time so that we can fix them before they become problem. This was achieved via principal component analysis with the aid of XLSTAT in Excel. In this study, we developed a model based on PCA to quickly identify and rank the workover candidate's wells. The dataset used in this project comprises of 66 oil wells and were obtained from a field operating in the Niger Delta. The first step involved data gathering and validation and uploading into XLSTAT software. Data preprocessing procedures were conducted to condition the dataset so as to give optimum performance during model development. A model was built to identify potential wells for workover operation. The results obtained here showed that the wells are separated to areas designated as (A to E). Wells found in area A indicated that they are potential candidates for workover operation. Wells found in area B showed that they are not under immediate danger, but attention would be needed to monitor and prevent increasing water and gas rates in the future. Wells found in area C indicated that they required immediate attention to prevent further decline in oil production. Likewise, wells found in Area D indicated that they also required immediate attention to prevent further decline in oil production. Finally, Wells found in Area E showed that they have highest oil production, lowest water production and moderate gas production, indicating normal condition with no immediate workover operation required. With advanced data analytics modeling, reservoir engineers or geoscientists will now see a bigger picture either field by field or reservoir by reservoir and quicky identify potential candidate wells for workover operation ahead of time before they become a problem. Hence, the results of the analysis can help us to better target wells that are potential candidates for high water cut, high WOR, High gas rates and low oil rates.

The petroleum industry is currently being faced with a growing number of ageing offshore platforms that are no longer in use and require to be decommissioned. Offshore decommissioning is a complex venture, and such projects are expected to cost the industry billions of dollars in the next two decades. Early knowledge of decommissioning cost is important to platform owners who bear the asset retirement obligation. However, obtaining the cost estimate for decommissioning an offshore platform is a challenging task that requires extensive structural and economic studies. This is further complicated by the existence of several decommissioning options such as complete and partial removal. In this paper, project costs for decommissioning 23 offshore platforms under three different scenarios are estimated using information from a publicly available source which only specified the costs of completely removing the platforms. A novel mathematical model for predicting the decommissioning cost for a platform based on its features is developed. The development included curve-fitting with the aid of generalised reduced gradient tool in Excel® Solver and a training dataset. The developed model predicted, with a very high degree of accuracy, platform decommissioning costs for four (4) different options under the Pacific Outer Continental Shelf conditions. Model performance was evaluated by calculating the Mean Absolute Percentage Error of predictions using a test dataset. This yielded a value of about 6%, implying a 94% chance of correctly predicting decommissioning cost.

This paper presents the development of predictive optimization models for autonomous rotary drilling systems where emphasis is placed on the shift from human (manual) operation as the driving force for drill rate performance to Quantitative Real-time Prediction (QRP) using machine learning. The methodology employed in this work uses real-time offset drilling data with machine learning models to accurately predict Rate of Penetration (ROP) and determine optimum operating parameters for improved drilling performance. Two optimization models (physics-based and energy conservation) were tested using Artificial Neutral Network (ANN) algorithm. Results of analysis using the model performance assessment criteria; correlation coefficient (R2) and Root Mean Square Error (RMSE), show that drill rate is non-linear in nature and the machine learning model (ANN) using energy conservation is most accurate for predicting ROP due to its ability in establishing a functional feature vector based on learning from past events.