In this paper we introduce for the first time an innovative approach for deriving Oil Formation Volume Factor (Bo) by mean of artificial intelligence method. In a new proposed application Self-Organizing Map (SOM) technology has been merged with statistical prediction methods integrating in a single step dimensionality reduction, extraction of input data structure pattern and prediction of formation volume factor Bo. The SOM neural network method applies an unsupervised training algorithm combined with back propagation neural network BPNN to subdivide the entire set of PVT input into different patterns identifying a set of data that have something in common and run individual MLFF ANN models for each specific PVT cluster and computing Bo. PVT data for more than two hundred oil samples (total of 804 data points) were collected from the north African region representing different basin and covering a greater geographical area were used in this study. To establish clear Bound on the accuracy of Bo determination several statistical parameters and terminology included in the presentation of the result from SOM-Neural Network solution. the main outcome is the reduction of error obtained by the new proposed competitive Learning Structure integration of SOM and MLFF ANN to less than 1 % compared to other method. however also investigated in this work five independents means of model driven and data driven approach for estimating Bo theses are 1) Optimal Transformations for Multiple Regression as introduced by (McCain, 1998) using alternating conditional expectations (ACE) for selecting multiple regression transformations 2), Genetic programing and heuristic modeling using Symbolic Regression (SR) and cross validation for model automatic tuning 3) Optimization by successive alteration of the empirical constant of the correlation parametric form of Standing correlation or any other, 4) Correlation refinement using Rafa Labedi approach tuning existing working correlation of Standing and others for calculating Bo by adjusting constant in the linear trend component of the formula structure of Standing correlation using a subset of the original correlation called correlation variable of formation volume factor this type of correlation refinement was extended and applied to modify any other available correlation to predict Bo 5) Machine learning predictive model (Nearest Neighbor Regression, Kernel Ridge regression, Gaussian Process Regression (GPR), Random Forest Regression (RF), Support Vector Regression (SVM), Decision Tree Regression (DT), Gradient Boosting Machine Regression (GBM), Group modeling data handling (GMDH). Regression Model Accuracy Metrics (Average absolute relative error, R-square), diagnostic plot was used to address the more adequate techniques and model for predicting Bo.

The relative permeability concept has been used extensively in reservoir engineering. As numerical reservoir simulation has become more popular as a tool for reservoir development, the role of relative permeability data became even more evident and important. Its key use is to control the advancement and mobility of different fluids simultaneously coexisting in the porous media, and hence controlling the recovery of the fluids. However, deriving a reliable relative permeability data set remained a major challenge. In reservoir engineering, this challenge has been present for many decades and might be so in the foreseeable future. Another challenge is to have a data set which is internally consistent and does not hinder the simulation performance. Optimistically, with the significant literature accumulated over the years in deriving and using relative permeability, some techniques can be extracted for data quality check, control and assurance. This paper covers the limitations of the conventional methods used for calculating relative permeability from displacement experiments. It also compiles all contemporary techniques in a systematic workflow for quality assessment and consistency evaluation. The workflow has been demonstrated with different synthetic and field examples. This paper will provide a reference for reservoir engineers who have an interest in investigating, checking the quality, and preparing relative permeability data set usable for reservoir simulation process. Introduction Special core analysis (SCAL) is the hub of the evaluation and management of hydrocarbon reservoirs. Relative permeability is one the main constituent of the SCAL which importance is widely recognized for the prediction of oil recovery during displacement by water. As any other piece of data, high quality and reliable relative permeability data set can reduce uncertainty in dynamic reservoir modelling and provide a sound foundation for reservoir engineering studies. Conversely, ppoor quality data can result in lost time due to rework and additional studies, inadequate development plans, and inefficient investment. Relative permeability curves can be generated from different sources such as mathematical models and experimental methods. However, experimental methods are more desirable for two reasons. First, they produce specific relative permeability relationships for specific reservoirs. Second, it is best available approach to resemble the flooding process in the field provided that the experiments performed on representative core samples and fluids from the reservoir under study. Therefore, our discussion will be restricted to relative permeability data are derived from laboratory experiments. The main challenge in the derivation process is how to obtain a reliable relative permeability data set. The term reliable will often be used in referring to relative permeability data set with is a good probability that the defined relationships are representative of the reservoir and inherently repeatable. Judgment regarding reliability will be made according to analysis of results that are judged to have been obtained using valid laboratory procedures. When the term valid is used in referring to laboratory measurements it will mean that none of the procedures used during the test are inconsistent with obtaining reliable results for the sample tested. For instance, the use of an extracted core plug with altered wettability other than the reservoir one would generally mean that the results are invalid. The source of unreliability may be attributed to the following reasons:The derived relative permeability data set are usually prone to excremental artefacts.The state of the experiment does not fit the model's assumptions used to derive the relative permeability data sets. In another word, the methods of calculating relative permeabilities from data obtained from displacement experiments don't describe all physical effects encountered in the experiment.

With the advances made in drilling long horizontal wells over the past decades it has become economically attractive to produce oil from thin oil rims. However, the production from these types of reservoirs presents several challenges. Gas coning is one of the most important ones. Horizontal well drilling traditionally helps to improve the oil recovery and avoid problems of premature gas/water breakthrough. In Bouri field, offshore Libya, the main concern of the operator was to establish an advanced method of controlling gas and water encroachment in a fractured carbonate reservoir characterized by high vertical permeability. This paper describes the first Inflow Control Device (ICD) installation for Mellitah Oil & Gas, and the first such application in Libya Offshore field. It was an integral part of a well completion aimed at evenly distributing inflow in a horizontal well, and at limiting the negative effects after occurrence of expected gas breakthrough. Due to small clearances involved, the ICD deployment presented a significant operational challenge. Despite the higher initial completion costs associated with ICDs, they can provide a cost-effective way to reduce long-term operating costs and increase yield. Production targets are achieved with longer, but fewer wells, maintenance and overhead. From a reservoir management point of view, ICDs can improve the productivity index (PI) by maximizing reservoir contact, minimizing gas coning by operating at lower drawdown, and increasing overall efficiency .Swellingpackers were used to compartmentalize the horizontal and build sections, allowing better drawdown control and eliminating crossflow issues. The completion required re-thinking of the established acid-wash treatment procedures, ultimately improving the overall well clean-up. Integrated analysis methods using steady-state wellbore hydraulic and 3D dynamic simulators were performed to generate flow profiles and calculate ICD pressure drop along the horizontal section. The models were updated using results from logging-while-drilling (LWD) and with real-time modifications to the initial design. To verify the inflow profile along the length of the ICD completion, production logging (PLT) was conducted. The inflow profiles compared favorably with those predicted by the models.

Horizontal well drilling traditionally helps to improve the oil recovery and avoid problems of premature gas/water breakthrough. In Bouri field, offshore Libya, the main concern of the operator was to establish an advanced method of controlling gas and water encroachment in a fractured carbonate reservoir characterized by high vertical permeability. This paper describes the first Inflow Control Device (ICD) installation for Mellitah Oil & Gas, and the first such application in Libya. It was an integral part of a well completion aimed at evenly distributing inflow in a horizontal well, and at limiting the negative effects after occurrence of expected gas breakthrough. Due to small clearances involved, the ICD deployment presented a significant operational challenge. Swelling-packers were used to compartmentalize the horizontal and build sections, allowing better drawdown control and eliminating cross-flow issues. The completion required re-thinking of the established acid-wash treatment procedures, ultimately improving the overall well clean-up. Integrated analysis methods using steady-state wellbore hydraulic and 3D dynamic simulators were performed to generate flow profiles and calculate ICD pressure drop along the horizontal section. The models were updated using results from logging-while-drilling (LWD) and with real-time modifications to the initial design. To verify the inflow profile along the length of the ICD completion, production logging (PLT) was conducted. The inflow profiles compared favorably with those predicted by the models.

No abstract