Iranian Journal of Oil and Gas Science and Technology

Abstract Unlike traditional approaches, Support Vector Regression (SVR), Multilayer Perceptron Neural Network (MLPNN), Probabilistic Neural Network (PNN), Random Forest (RF), Decision Tree (DT), and eXtreme Gradient Boosting (XGBoost) are utilized as predictive algorithms to simulate the cementation exponent based on various pore descriptions and total porosity. To optimize the parameters of the MLPNN approach, Levenberg-Marquardt (LM) is coupled with MLPNN, leading to the development of a hybrid approach named MLPNN-LM. This hybrid model efficiently optimizes neural network parameters, significantly improving accuracy and convergence speed. The necessary databank for constructing, validating, and predicting with the proposed models is derived from Ragland's work and classified into test, validation, and train subsets. The results reveal high precision for the hybrid MLPNN-LM and RF techniques, with key statistical measures such as the Average Absolute Percentage Relative Deviations (AAPRD%) (i.e., the percentage of relative error) of 4.3781% for MLPNN-LM and 4.8690% for RF, and determination coefficients (R²) (i.e., the fitness magnitude of estimated and measured values around Y=X line) of 0.8654 for MLPNN-LM and 0.8731 for RF. The highly accurate estimates of the cementation exponent provided by MLPNN-LM and RF surpass those from commonly applied literature correlations. Sensitivity analysis shows the significant impact of interparticle, moldic, and connected vuggy pore types on the modeling output. The trustworthiness of the databank and the accuracy of the proposed MLPNN-LM and RF approaches are verified by Williams’ plot, with approximately 93.75% and 91.96% of the databank within the applicability domain, respectively. Trend analysis demonstrates a good match of predicted cementation exponent with actual data, especially for deep formations (i.e., depth greater than 1200 m) and tight carbonate reservoirs (i.e., total porosities less than 10%), where traditional correlations face challenges due to the noticeable complexity in the behavior of cementation exponent.

Abstract Optimization of transportation for organizational projects is paramount, as identifying and evaluating transportation barriers constitutes key strategic decisions for managers and decision-makers. This study focuses on three main objectives: identifying the factors influencing lean logistics and transportation barriers, clarifying the interrelationships among the research criteria, and determining their relative importance and ranking in petroleum product distribution. A hybrid model was employed to achieve these objectives. First, the Delphi method was utilized to facilitate the examination of interrelationships among the research criteria. Additionally, fuzzy multi-criteria decision-making techniques, including the fuzzy Analytic Hierarchy Process, were applied to assign specific weights to the criteria. Subsequently, the VIKOR method enabled the prioritization of these criteria. The findings of this study confirm the existence of significant relationships between lean logistics, lean criteria, and transportation barriers, highlighting that managerial barriers and lean managerial logistics are the most critical obstacles and factors, respectively, in petroleum product distribution. These findings can assist managers in improving distribution processes and mitigating key transportation barriers.

Abstract Cryogenic nitrogen rejection from methane-rich natural gas is a critical operation for meeting LNG and pipeline gas specifications. Yet, it is highly susceptible to operational disturbances caused by solid formation in heat exchangers and distillation columns. This study proposes an integrated simulation framework that couples steady-state process modeling in Aspen HYSYS with rigorous solid–liquid–vapor equilibrium (SLVE) analysis in ThermoFAST to evaluate freezing risks in CH4 systematically–N2 mixtures under nitrogen-rejection unit (NRU) conditions. The NRU process, including multi-stream heat exchange, Joule–Thomson expansion, and cryogenic distillation, is modeled using the Peng–Robinson equation of state. ThermoFAST employs a Helmholtz-energy-based PC-SAFT equation of state, together with Lennard-Jones Weeks–Chandler–Andersen (LJ-WCA) pure-solid references, to generate freezing envelopes over a wide pressure range (0.1–10 MPa) and across methane-rich to nitrogen-rich compositions relevant to industrial operations.

The framework is validated against available experimental solid–liquid equilibrium data, yielding mean absolute and relative deviations of 3.45 K and 5%, respectively, demonstrating its suitability for hazard screening applications. Simulation results reveal that increasing pressure elevates the eutectic temperature and expands the stability region of the solid phase. In contrast, increasing nitrogen concentration depresses the eutectic point and narrows the solid stability range. Risk maps indicate that solid formation is most probable in methane-rich streams (CH4 > 0.75) at pressures of 10 MPa or higher, as well as in nitrogen-rich streams (CH4 < 0.55) at extremely low temperatures, particularly after expansion and in the upper trays of the distillation column. The proposed integrated approach provides a predictive tool for identifying vulnerable operating zones, defining safe temperature–pressure margins, and enhancing the safety, operability, and efficiency of cryogenic nitrogen rejection processes.

Abstract Accurate pore pressure estimation is a critical component of geomechanical modeling, essential for maintaining wellbore stability and optimizing drilling fluid density. This study proposes a Grey Wolf Optimizer-Supported Vector Machine (GWO-SVM) workflow to predict pore pressure in a complex carbonate reservoir in southwestern Iran. While traditional empirical correlations like Eaton and Bowers are widely used, their reliance on continuous calibration can be challenging. To address this, discrete Repeating Formation Test (RFT) pressure points were utilized to calibrate baseline empirical trends, creating continuous reference profiles for the entire depth. The GWO-SVM model was then deployed to automate the replication of these calibrated baselines from standard petrophysical logs (sonic, density, resistivity, porosity, and shale volume). Using a dataset from five wells (four for training, one for blind validation), the GWO optimally tuned the Support Vector Regression (SVR) hyperparameters. The model demonstrated exceptional fidelity in replicating the calibrated baseline on the blind test well, achieving an RMSE=37.96psi and an R^2=0.998. Finally, the 1D well-based predictions were integrated into a 3D geostatistical model using co-kriging to visualize pressure compartmentalization influenced by the local tectonic stress regime. This workflow offers a replicable, high-precision alternative for pre-drill pore pressure modeling in data-limited reservoir settings.