Article
Thermodynamics
Viet Nguyen-Le, Hyundon Shin
Summary: This paper proposes three artificial neural network architectures for predicting the production and decline parameters of shale gas wells. The testing results confirm the efficiency and reliability of the proposed method. The predicted production profiles can be used to update development plans and calculate the project's net present value.
Article
Engineering, Multidisciplinary
Sarmad Dashti Latif, Ali Najah Ahmed, Mohsen Sherif, Ahmed Sefelnasr, Ahmed El-Shafie
Summary: Developing accurate forecast for water losses and reservoir final storage using Artificial Neural Networks (ANN) models with radial basis function (RBF) can lead to better monitoring of water quality and more efficient reservoir operation, thus potentially reducing flash flood and water crisis problems in Malaysia.
ALEXANDRIA ENGINEERING JOURNAL
(2021)
Article
Geochemistry & Geophysics
Kai Zhang, Niantian Lin, Jiuqiang Yang, Dong Zhang, Yan Cui, Zhiwei Jin
Summary: Gas reservoir identification using seismic data is a significant focus in geophysical exploration. This study introduces an approach using artificial neural networks (ANNs) and the Viterbi algorithm to identify gas reservoirs and evaluate their structural characteristics, improving processing efficiency. The initial identification utilizes deep neural networks (DNNs) and obtains composite seismic attributes sensitive to gas reservoir responses. The ANN-based gas reservoir identification results are evaluated comprehensively based on structural characteristics, reducing the uncertainties predicted by mathematical methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Multidisciplinary Sciences
Kiana Peiro Ahmady Langeroudy, Parsa Kharazi Esfahani, Mohammad Reza Khorsand Movaghar
Summary: Oil viscosity is crucial in petroleum engineering, and experimental methods and compositional methods can accurately estimate it. However, experimental data is difficult to obtain, so there is a need for convenient and fast methods to predict viscosity. This study uses machine learning methods (XGBoost, CatBoost, and GradientBoosting) based on gradient boosting decision tree to reduce the prediction error of viscosity by considering dissolved gas content, temperature, pressure, and API gravity. XGBoost outperforms other methods with higher precision and lower error, showing the effectiveness of the approach.
SCIENTIFIC REPORTS
(2023)
Article
Energy & Fuels
Niloofar Salmani, Rouhollah Fatehi, Reza Azin
Summary: A new method of overlapping grids is introduced in this paper for simulating flow around a single well in reservoirs. By combining Cartesian and radial grids, connectivity is established between the two grids to transfer pressure and saturation data effectively. The proposed Double-Scale method using a 21 x 21 coarse Cartesian grid shows comparable accuracy to a 1001 x 1001 simple Cartesian fine grid model, but is almost 2 times faster in specific problems.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Energy & Fuels
Xuliang Liu, Wenshu Zha, Zhankui Qi, Daolun Li, Yan Xing, Lei He
Summary: Well test analysis is crucial in monitoring reservoir performance. This paper proposes an intelligent reservoir model identification method using convolutional neural network (CNN), which improves classification accuracy and achieves good results.
JOURNAL OF ENERGY RESOURCES TECHNOLOGY-TRANSACTIONS OF THE ASME
(2022)
Article
Energy & Fuels
S. M. Alizadeh, A. Khodabakhshi, P. Abaei Hassani, B. Vaferi
Summary: The technique of using GoogleNet to analyze transient signals in petroleum engineering is able to decrease uncertainty and accurately classify different reservoir interpretation classes with an overall accuracy of 98.36%.
JOURNAL OF ENERGY RESOURCES TECHNOLOGY-TRANSACTIONS OF THE ASME
(2021)
Article
Geochemistry & Geophysics
Lei Song, Xingyao Yin, Linjie Yin
Summary: Reservoir lithology identification is crucial for reservoir characterization, reserves calculation, and geological modeling. This study proposes an improved adversarial learning (IAL) method for reservoir lithology identification, addressing the overfitting and multisolution problems caused by inadequate labeled data and massive learnable parameters. A probabilistic lithology classification neural network (PLCNN) is constructed to predict lithology from density, P-velocity, and S-velocity, and an IAL lithology identification workflow is designed to train the PLCNN with limited labeled data and large-scale unlabeled data. The proposed method is successfully applied to the Book Cliffs model, achieving a classification accuracy of 92.71% and efficiently relieving the misclassification of sand and sandy shale compared to the conventional supervised learning workflow.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Chemistry, Multidisciplinary
Daniel Asante Otchere, Mohammed Abdalla Ayoub Mohammed, Tarek Omar Arbi Ganat, Raoof Gholami, Zulkifli Merican Aljunid Merican
Summary: Accurately measuring wettability is crucial for understanding reservoir parameters and optimizing reservoir potential, recovery, and management. This study introduces a new mathematical model based on Amott-USBM wettability measurement and field NMR T2LM log, which shows promising results in characterizing wettability at both lab and field scales.
APPLIED SCIENCES-BASEL
(2022)
Article
Energy & Fuels
Abdelrahman Gouda, Sayed Gomaa, Attia Attia, Ramadan Emara, S. M. Desouky, A. N. El-hoshoudy
Summary: Optimal management and development plans for gas condensate reservoirs require accurate prediction of dew point pressure (DPP). Traditionally, DPP is measured through a constant mass expansion test (CME), but this method is time-consuming. This study developed a new intelligent predictive model based on 453 gas condensate samples, showing excellent performance compared to widely used empirical correlations.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Optics
Mengyao Han, Muguang Wang, Yuchuan Fan, Shiyi Cai, Yuxiao Guo, Naihan Zhang, Richard Schatz, Sergei Popov, Oskars Ozolins, Xiaodan Pang
Summary: This paper proposes an approach for simultaneous modulation format identification (MFI) and optical signal-to-noise ratio (OSNR) monitoring based on optoelectronic reservoir computing (RC) and signal's amplitude histograms (AHs). The proposed method achieves 100% MFI accuracy and accurate OSNR estimation for different modulation formats. The approach also exhibits good robustness in the presence of noise interference.
Article
Energy & Fuels
Mehrafarin Moghimihanjani, Behzad Vaferi
Summary: A novel approach combining wavelet transform and recurrent neural networks is proposed for analyzing long-term well testing signals to improve the classification accuracy of hydrocarbon reservoir systems. By extracting features and identifying reservoir types, the method demonstrates superior performance compared to traditional techniques.
JOURNAL OF ENERGY RESOURCES TECHNOLOGY-TRANSACTIONS OF THE ASME
(2021)
Article
Energy & Fuels
Thales de Oliveira Souza, Kyung Jae Lee
Summary: This study aims to reduce the risk and impact of pollutant releases from offshore wells by using machine learning technology to predict potential breaches and their environmental consequences. By conducting numerical simulations and generating data from different scenarios, multiple Artificial Neural Networks were trained to predict various outputs, resulting in excellent correlations between the input and output features. This is the first study to combine machine learning technology with advanced reservoir simulation to mitigate the hazard of gas escaping from offshore production wells.
PETROLEUM SCIENCE AND TECHNOLOGY
(2023)
Article
Engineering, Civil
Ningpeng Dong, Wenhai Guan, Jixue Cao, Yibo Zou, Mingxiang Yang, Jianhui Wei, Liang Chen, Hao Wang
Summary: This study develops data-driven reservoir operation schemes based on extreme gradient boosting (XGBoost) and artificial neural network (ANN) to predict reservoir release and storage. A hybrid modeling framework is proposed by coupling a high-resolution hydrologic model with the developed data-driven reservoir operation schemes and a calibration-free conceptual reservoir operation scheme for data-scarce reservoirs. The framework quantitatively assesses the cumulative impacts of dam operation on the hydrologic regime under different reservoir data availability. The results show the effectiveness of the framework in reconstructing reservoir releases and storage, and improving daily streamflow in the Upper Yangtze River Basin (UYRB).
JOURNAL OF HYDROLOGY
(2023)
Article
Geochemistry & Geophysics
Cao Song, Wenkai Lu, Yuqing Wang, Songbai Jin, Jinliang Tang, Lei Chen
Summary: Reservoir prediction is a significant issue in seismic interpretation. This study proposes a semisupervised deep-learning framework using a closed-loop CNN and virtual well-logging labels to improve the accuracy and spatial continuity of the predicted reservoir.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Biotechnology & Applied Microbiology
Bahador Daryayehsalameh, Miralireza Nabavi, Behzad Vaferi
Summary: The study estimates the solubility of CO2 in ionic liquid using various artificial intelligence techniques, with the cascade feed-forward neural network identified as the best model. This model accurately predicts the experimental data, showing that the maximum mole fraction of CO2 can be obtained at the highest pressure and the lowest temperature.
ENVIRONMENTAL TECHNOLOGY & INNOVATION
(2021)
Article
Biotechnology & Applied Microbiology
Zahra Keshtkar, Sajad Tamjidi, Behzad Vaferi
Summary: Wastewater pollution by heavy metals, especially nickel, has negative impacts on human health and the environment. The synthesized gamma-alumina nano-adsorbents are effective in removing nickel ions from wastewater, with different surface and pore characteristics between low and high specific surface area samples. Optimum conditions for nickel uptake include temperature of 40 degrees C, adsorbent dosage of 2 g, pH of 4, nickel ions initial concentration of 25 mg/L, and contact time of 60 min, resulting in high removal rates by both types of nano-adsorbents.
ENVIRONMENTAL TECHNOLOGY & INNOVATION
(2021)
Article
Polymer Science
Jing Wang, Mohamed Arselene Ayari, Amith Khandakar, Muhammad E. H. Chowdhury, Sm Ashfaq Uz Zaman, Tawsifur Rahman, Behzad Vaferi
Summary: This research utilizes machine learning methods to estimate the relative crystallinity of biodegradable PLLA/PGA composites. A model with a single hidden layer CFFNN is found to be the most accurate, predicting an experimental database with high accuracy. The study shows that relative crystallinity increases with PGA content and crystallization time, while the effect of temperature is more complex.
Article
Chemistry, Physical
Seyed Mehdi Seyed Alizadeh, Zahra Parhizi, Ali Hosin Alibak, Behzad Vaferi, Saleh Hosseini
Summary: This study predicts the hydrogen uptake ability of 28 zeolites using artificial neural networks. The most accurate model is determined, and the leverage method is used to verify the reliability of the data.
INTERNATIONAL JOURNAL OF HYDROGEN ENERGY
(2022)
Article
Green & Sustainable Science & Technology
Xiaolei Zhu, Marzieh Khosravi, Behzad Vaferi, Menad Nait Amar, Mohammed Abdelfetah Ghriga, Adil Hussein Mohammed
Summary: This study uses a machine learning model to predict the absorption capacity of deep eutectic solvents (DESs) for sulfur dioxide (SO2), providing a reliable model based on comprehensive experimental data. The model is highly accurate and can be used to screen and find the best DESs candidates.
JOURNAL OF CLEANER PRODUCTION
(2022)
Article
Environmental Sciences
Hadi Adloo, Saeed Foshat, Behzad Vaferi, Falah Alobaid, Babak Aghel
Summary: This study investigates the critical factors causing non-Darcian flow in porous media and explores the distinct roles of pores and throats in total dissipation using numerical simulation. The Forchheimer model is used to analyze the non-Darcian flow. The results show that pores are more likely to deviate from the Darcy model than throats, and increasing the pore-to-throat ratio leads to earlier onset of non-Darcian flow in the pores.
Article
Energy & Fuels
Mohsen Karimi, Marzieh Khosravi, Reza Fathollahi, Amith Khandakar, Behzad Vaferi
Summary: This study develops a machine learning model to estimate the heat capacity of cellulosic biomass samples with different origins, and the results show that the least-squares support vector regression model with the Gaussian kernel function is the best estimator. The model's accuracy is validated using laboratory heat capacity data, and it is found to have a 62% higher prediction accuracy compared to the empirical correlation method.
ENERGY SCIENCE & ENGINEERING
(2022)
Article
Pharmacology & Pharmacy
Maryam Najmi, Mohamed Arselene Ayari, Hamidreza Sadeghsalehi, Behzad Vaferi, Amith Khandakar, Muhammad E. H. Chowdhury, Tawsifur Rahman, Zanko Hassan Jawhar
Summary: This study constructs a stacked model using machine learning tools to predict the solubility of anticancer drugs in supercritical CO2. The model demonstrates excellent performance according to experimental validation.
Article
Energy & Fuels
Saleh Hosseini, Amith Khandakar, Muhammad E. H. Chowdhury, Mohamed Arselene Ayari, Tawsifur Rahman, Moajjem Hossain Chowdhury, Behzad Vaferi
Summary: The fouling factor is an index measuring the undesirable effect of solids' deposition on heat transfer ability. This study uses machine-learning algorithms and traditional models to accurately predict the fouling factor, with Gaussian Process Regression achieving the most accurate predictions.
Article
Energy & Fuels
Lan Xu, Aboozar Khalifeh, Amith Khandakar, Behzad Vaferi
Summary: Nanofluids have been used in experimental studies to improve the performance of flat plate solar collectors (FPSC). However, the results regarding the effect of nanofluids on FPSC are often ambiguous and contradictory. This research develops a straightforward approach to predict the thermal efficiency of nanofluid-based FPSC and compares different machine learning models to determine the most accurate tool for this task, finding that LS-SVR performs the best.
Article
Energy & Fuels
Yan Cao, Elham Kamrani, Saeid Mirzaei, Amith Khandakar, Behzad Vaferi
Summary: This study utilized machine-learning approaches to simulate the electrical performance of PV/T systems cooled by water-based nanofluids, finding the ANFIS as the most effective method. The optimized condition with 30 lit/hr of water-silica nano-coolant at a radiation intensity of 788.285 W/m(2) maximized electrical efficiency by 27.7%. The ANFIS model successfully predicted a large amount of experimental data and an external database with a low average relative deviation and high R-squared value.
Article
Energy & Fuels
Vishal Singh, Nabindra Ruwali, Rakesh Kumar Pandey, Behzad Vaferi, David A. Wood
Summary: This study proposes a surrogate model based on deep learning to predict cumulative oil production under water flooding conditions. Compared to traditional machine learning methods, this model achieves higher accuracy.
PETROLEUM SCIENCE AND TECHNOLOGY
(2023)
Article
Engineering, Environmental
Hamed Mohammaddoost, Maryam Asemani, Ahmad Azari, Behzad Vaferi
Summary: New polymeric pipes have been used to overcome the challenges posed by metal pipes in gas transmission, such as corrosion and leakage. This study investigates the diffusivity and potential leakage of methane and ethane gases in glass-reinforced epoxy (GRE) composite using diffusion and solubility cells. The results demonstrate that the diffusivity of both gases increases with temperature, while gas solubility decreases. Furthermore, novel correlations for estimating gas diffusion coefficient and mass flow rate have been developed based on experimental data.
PROCESS SAFETY AND ENVIRONMENTAL PROTECTION
(2023)
Article
Polymer Science
Yaser Ahmadi, Mohamed Arselene Ayari, Meysam Olfati, Seyyed Hossein Hosseini, Amith Khandakar, Behzad Vaferi, Martin Olazar
Summary: This study investigates the effect of green polymeric nanoparticles on the interfacial tension (IFT) and wettability of carbonate reservoirs for enhanced oil recovery (EOR). The performance of xanthan/magnetite/SiO2 nanocomposites and green materials, such as eucalyptus plant nanocomposites and walnut shell nanocomposites, were compared through spontaneous imbibition tests. The results showed that eucalyptus plant nanocomposites performed better than walnut shell nanocomposites in reducing contact angle (CA) and IFT under different salinities. The EOR of carbonate rocks was improved with eucalyptus plant nanocomposites at specific salinity concentrations.
Article
Multidisciplinary Sciences
Saleh Hosseini, Behzad Vaferi
Summary: The study accurately determines methanol loss in a three-phase separator using intelligent connectionist approaches like least-squares support vector machines (LS-SVM). The LS-SVM model shows excellent consistency with real-field datasets and the economic impact on gas processing plants is also evaluated.
ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING
(2022)
