Average Absolute Percent Relative Error Research Articles

Modeling of Brine/CO2/Mineral Wettability Using Gene Expression Programming (GEP): Application to Carbon Geo-Sequestration

Carbon geo-sequestration (CGS), as a well-known procedure, is employed to reduce/store greenhouse gases. Wettability behavior is one of the important parameters in the geological CO2 sequestration process. Few models have been reported for characterizing the contact angle of the brine/CO2/mineral system at different environmental conditions. In this study, a smart machine learning model, namely Gene Expression Programming (GEP), was implemented to model the wettability behavior in a ternary system of CO2, brine, and mineral under different operating conditions, including salinity, pressure, and temperature. The presented models provided an accurate estimation for the receding, static, and advancing contact angles of brine/CO2 on various minerals, such as calcite, feldspar, mica, and quartz. A total of 630 experimental data points were utilized for establishing the correlations. Both statistical evaluation and graphical analyses were performed to show the reliability and performance of the developed models. The results showed that the implemented GEP model accurately predicted the wettability behavior under various operating conditions and a few data points were detected as probably doubtful. The average absolute percent relative error (AAPRE) of the models proposed for calcite, feldspar, mica, and quartz were obtained as 5.66%, 1.56%, 14.44%, and 13.93%, respectively, which confirm the accurate performance of the GEP algorithm. Finally, the investigation of sensitivity analysis indicated that salinity and pressure had the utmost influence on contact angles of brine/CO2 on a range of different minerals. In addition, the effect of the accurate estimation of wettability on CO2 column height for CO2 sequestration was illustrated. According to the impact of wettability on the residual and structural trapping mechanisms during the geo-sequestration of the carbon process, the outcomes of the GEP model can be beneficial for the precise prediction of the capacity of these mechanisms.

Determination of the Gas-Oil Ratio below the Bubble Point Pressure Using the Adaptive Neuro-Fuzzy Inference System (ANFIS).

Determining the solution gas–oil ratio (Rs) below the bubble point is a vital requirement that aids in multiple production engineering and reservoir analysis issues. Currently, there are some models available for the determination of the solution gas–oil ratio under the bubble point. However, they still may prove unreliable due to the applied assumptions and their specification to operate only under a particular range of data. In this study, the neuro-fuzzy, i.e., the adaptive neuro-fuzzy inference system (ANFIS) approach, is utilized to develop an accurate and dependable model for determining the Rs below the bubble point pressure. A total of 376 pressure–volume–temperature datasets from Sudanese oil fields were used to establish the proposed ANFIS model. The trend analysis was applied to affirm the proper relationships between the inputs and outputs. Furthermore, using different statistical error analyses, the developed model was benchmarked against widely used empirical methods to evaluate the proposed method’s performance in predicting the Rs at pressures below the bubble point. The proposed ANFIS model performs with an average absolute percent relative error of 10.60% and a correlation coefficient of 99.04%, surpassing the previously studied correlations.

Estimation of bubble point pressure and solution gas oil ratio using artificial neural network

Bubble point pressure and gas solubility are considered vital pressure-volume-temperature properties of crude oil samples as both play a significant role in reservoir and production engineering calculations. These two fluid oil properties can be measured experimentally or using mathematical correlations, the experimental methods are time consuming and expensive, although they give a reliable result. Therefore, using mathematical correlations can be as an alternative way, however, existing correlation having number of limitations. Hence, this study presents a novel numerical model based on artificial neural network (ANN) model and universal mathematical correlation to forecast bubble point pressure and gas solubility. The artificial neural network model based on back-propagation learning algorithm and sigmoid function with numerous inputs include API gravity, reservoir temperature, and gas specific gravity. The ANN model is developed and validated by utilizing more than 450 data points collected from numerous fields located worldwide with various reservoir characterization. In this study, the bubble point pressure is estimated first and then used to forecast gas solubility based on the calculated correlation coefficient. The results of this study concluded that in order to optimize the ANN model, collected data must to be distributed as follows; 65% for training, 20% for testing and 15% for the validation processes with R2 of 0.988 and average absolute percent relative error (AAPRE) =5.56% for the bubble point pressure and 4.5% for the gas solubility. A comparison study is performed between the new ANN extracted correlation and other well-known models (Standing, Petrosky and Frashad, McCain, Vasquez and Beggs, Marhoun, Glaso, Dokla, and Lasater correlations) to show the robustness of the presented correlation. The results show that the presented correlation has the ability to predict bubble point pressure and gas solubility precisely with lowest AAPRE compared to other models.

Integrating advanced soft computing techniques with experimental studies for pore structure analysis of Qingshankou shale in Southern Songliao Basin, NE China

Evaluating pore structure of unconventional shale reservoirs enables us to determine their productivity, allowing for better operational decisions. Despite extensive studies in this field, considering the complexity of shale plays, pore structure analysis of such formations still requires novelties and further research. In this study, 10 samples from the Qingshankou Formation (from 5 wells) were analyzed with X-ray diffraction (XRD), programmed pyrolysis, N2 adsorption, and mercury intrusion capillary pressure (MICP). In the next step, several modern intelligent smart models including multilayer perceptron (MLP), radial basis function (RBF), generalized regression neural network (GRNN), cascaded forward neural network (CFNN), and least squares support vector machine (LSSVM), that were optimized by levenberg-marquardt (LM), Bayesian regularization (BR), genetic algorithm (GA), ant colony optimization (ACO), particle swarm optimization (PSO), and differential evolution (DE) algorithms were employed, to estimate the volumes of N2 adsorbed and desorbed based on the mineralogy and geochemical properties of the samples. Results show that samples are mainly composed of clay (up to 42.3 wt%) and quartz (up to 34.6 wt%), low in total organic carbon (TOC) (up to 2.89%) and in the oil generation window. Complexity of smaller pores was found higher compared to medium and larger ones. In addition, deconvolution of N2 adsorption pore size distribution (PSD) curve revealed that samples are composed of up to three families in the range of macropore size and different families in mesopore size. We found that LSSVM with applicability to the entire input dataset, outperformed all other models in predicting the amount of nitrogen adsorption and desorption with an average absolute percent relative error (AAPRE) value of 1.94%. Ultimately, clay minerals and potash feldspar had the greatest effect on increasing and decreasing the amount of nitrogen adsorbed and desorbed, respectively. The Leverage technique's findings demonstrate that more than 97% of total data points in the LSSVM model are in the valid domain. This study proves that smart methods if used properly, would enable us to study a large group of samples independent from exhaustive, time consuming and expensive experimental methods.

An advanced computational intelligent framework to predict shear sonic velocity with application to mechanical rock classification

Shear sonic wave velocity (Vs) has a wide variety of implications, from reservoir management and development to geomechanical and geophysical studies. In the current study, two approaches were adopted to predict shear sonic wave velocities (Vs) from several petrophysical well logs, including gamma ray (GR), density (RHOB), neutron (NPHI), and compressional sonic wave velocity (Vp). For this purpose, five intelligent models of random forest (RF), extra tree (ET), Gaussian process regression (GPR), and the integration of adaptive neuro fuzzy inference system (ANFIS) with differential evolution (DE) and imperialist competitive algorithm (ICA) optimizers were implemented. In the first approach, the target was estimated based only on Vp, and the second scenario predicted Vs from the integration of Vp, GR, RHOB, and NPHI inputs. In each scenario, 8061 data points belonging to an oilfield located in the southwest of Iran were investigated. The ET model showed a lower average absolute percent relative error (AAPRE) compared to other models for both approaches. Considering the first approach in which the Vp was the only input, the obtained AAPRE values for RF, ET, GPR, ANFIS + DE, and ANFIS + ICA models are 1.54%, 1.34%, 1.54%, 1.56%, and 1.57%, respectively. In the second scenario, the achieved AAPRE values for RF, ET, GPR, ANFIS + DE, and ANFIS + ICA models are 1.25%, 1.03%, 1.16%, 1.63%, and 1.49%, respectively. The Williams plot proved the validity of both one-input and four-inputs ET model. Regarding the ET model constructed based on only one variable,Williams plot interestingly showed that all 8061 data points are valid data. Also, the outcome of the Leverage approach for the ET model designed with four inputs highlighted that there are only 240 “out of leverage” data sets. In addition, only 169 data are suspected. Also, the sensitivity analysis results typified that the Vp has a higher effect on the target parameter (Vs) than other implemented inputs. Overall, the second scenario demonstrated more satisfactory Vs predictions due to the lower obtained errors of its developed models. Finally, the two ET models with the linear regression model, which is of high interest to the industry, were applied to diagnose candidate layers along the formation for hydraulic fracturing. While the linear regression model fails to accurately trace variations of rock properties, the intelligent models successfully detect brittle intervals consistent with field measurements.

Application of robust intelligent schemes for accurate modelling interfacial tension of CO2 brine systems: Implications for structural CO2 trapping

Given the current global climate change, renewable energy sources, carbon capture, utilization, and storage (CCUS) are being considered as a potential solutions to this critical global issue. Structural and capillary processes can be used to store carbon dioxide (CO2) in deep saline aquifers in a way that is safe and does not harm the environment. Due to this fact, the interfacial tension (IFT) of the CO2-brine system is an important factor influencing the capacity of storage formations to sequester CO2. As a result, IFT is essential for conducting a thorough and accurate evaluation in order to determine the optimal strategy for CO2 storage projects. This paper used intelligent models such as Gaussian Process Regression (GPR), Radial Basis Function (RBF), and Random Forest (RF) to forecast IFT in the CO2-brine system with high precision and substantial time saving. The results reveal that the constructed RF model could deliver excellent performance in predicting IFT with the lowest average absolute percent relative error (AAPRE = 1.9156 percent), highest coefficient of determination (R2 = 0.9907), and lowest root mean squared error (RMSE = 0.7279). Furthermore, a sensitivity analysis was performed to ascertain the most critical parameters in the RF model to be considered. The parametric analysis found that both pure and non-pure CO2 systems had a significant impact on IFT prediction.Also, the RF model was used to assess the structural trapping capacity of a storage location in the Cuu Long Basin. The estimation results obtained in this study agreed perfectly with the previous ones. The findings of this study can aid in a better understanding of how machine learning models can be applied to predict IFT values for the evaluation of the structural CO2 storage capacity.

Modeling solubility of CO2\u2013N2 gas mixtures in aqueous electrolyte systems using artificial intelligence techniques and equations of state

Determining the solubility of non-hydrocarbon gases such as carbon dioxide (CO2) and nitrogen (N2) in water and brine is one of the most controversial challenges in the oil and chemical industries. Although many researches have been conducted on solubility of gases in brine and water, very few researches investigated the solubility of power plant flue gases (CO2–N2 mixtures) in aqueous solutions. In this study, using six intelligent models, including Random Forest, Decision Tree (DT), Gradient Boosting-Decision Tree (GB-DT), Adaptive Boosting-Decision Tree (AdaBoost-DT), Adaptive Boosting-Support Vector Regression (AdaBoost-SVR), and Gradient Boosting-Support Vector Regression (GB-SVR), the solubility of CO2–N2 mixtures in water and brine solutions was predicted, and the results were compared with four equations of state (EOSs), including Peng–Robinson (PR), Soave–Redlich–Kwong (SRK), Valderrama–Patel–Teja (VPT), and Perturbed-Chain Statistical Associating Fluid Theory (PC-SAFT). The results indicate that the Random Forest model with an average absolute percent relative error (AAPRE) value of 2.8% has the best predictions. The GB-SVR and DT models also have good precision with AAPRE values of 6.43% and 7.41%, respectively. For solubility of CO2 present in gaseous mixtures in aqueous systems, the PC-SAFT model, and for solubility of N2, the VPT EOS had the best results among the EOSs. Also, the sensitivity analysis of input parameters showed that increasing the mole percent of CO2 in gaseous phase, temperature, pressure, and decreasing the ionic strength increase the solubility of CO2–N2 mixture in water and brine solutions. Another significant issue is that increasing the salinity of brine also has a subtractive effect on the solubility of CO2–N2 mixture. Finally, the Leverage method proved that the actual data are of excellent quality and the Random Forest approach is quite reliable for determining the solubility of the CO2–N2 gas mixtures in aqueous systems.

Toward predicting SO2 solubility in ionic liquids utilizing soft computing approaches and equations of state

BackgroundThe use of novel and green solvents like ionic liquids (ILs) for the capture of air pollutant gases has gained extensive attention in recent years. However, getting reliable and fast predictions of gases solubility in ILs is complex. MethodsFour soft computing methods including deep belief network (DBN), group method of data handling (GMDH), genetic programming (GP), and K-nearest neighbor (KNN) were utilized for estimating the solubility of sulfur dioxide (SO2) in ILs. A total of 374 experimental data points of SO2 solubility in 15 types of ILs were collected and used for model development. Moreover, Valderrama-Patel-Teja (VPT), Zudkevitch-Joffe (ZJ), Peng-Robinson (PR), Redlich-Kwong (RK), and Soave-Redlich-Kwong (SRK) equations of state (EOSs) were applied for the solubility predictions in the SO2 + ILs systems. Significant findingsThe results illustrated that DBN model is the most reliable predictive tool for the SO2 solubility in ILs by having an average absolute percent relative error (AAPRE) of 3.56%. Furthermore, the proposed simple to use GMDH mathematical correlation also provides good estimations with an AAPRE of 8.05%. Despite the weaker performance of the EOSs than the intelligent models, the PR EOS presented better estimations among other EOSs for the SO2 solubility in ILs.

Experimental measurement and compositional modeling of bubble point pressure in crude oil systems: Soft computing approaches, correlations, and equations of state

No one can deny the ever-increasing importance of oil since it has influenced every aspect of humans' life. One of the most important pressure-volume-temperature (PVT) properties of crude oil, which is needed in a majority of production and reservoir engineering calculations, is the saturation pressure (bubble point pressure (Pb)). Having accurate knowledge about Pb is significant for both academia and industry. This communication concentrates on providing reliable experimental data from constant composition expansion (CCE) test as well as rigorous compositional models to predict saturation pressure of crude oils based on oil composition (H2S, N2, CO2, C1 to C7+), reservoir temperature, C7+ specifications (molecular weight and specific gravity). Seven advanced machine learning approaches, namely, decision trees (DTs), random forest (RF), extra trees (ETs), cascade-forward back-propagation network (CFBPN), and generalized regression neural networks (GRNN) as well as multilayer perceptron (MLP) and radial basis function (RBF) neural networks were used for modeling. The CFBPN and MLP models were trained by three different training algorithms, namely scaled conjugate gradient (SCG), Bayesian regularization (BR), and Levenberg-Marquardt (LM). The modeling was done based on a databank consisting of 206 data points (130 points were previously published in the literature plus 76 points determined experimentally in this study). The results show that the DT model could provide the most reliable prediction with an average absolute percent relative error (AAPRE) of 4.43%. The efficiency of various equations of state (EoS) and empirical correlations were checked. According to the results, Peng-Robinson (PR) and the correlation developed by Elsharkawy were the most efficient EoS and empirical correlation with AAPRE values of 8.46% and 12.05%, respectively. Then, the sensitivity analysis revealed that the Pb was extremely affected by methane and C7+ mole percent. Finally, the Leverage approach confirmed the validity of the employed data, detecting only 7 points as outliers.

Estimating reservoir properties using downhole temperature and pressure data

Predicting the behavior of an underground reservoir requires precise reservoir properties determination. One of the most effective tools for this purpose is well test analysis. Conventional well test analysis employed bottomhole pressure data to understand the reservoir performance better. As the heterogeneity of the system increases, the complexity of well test analysis increases and nonunique solutions can be obtained. Using another source of data, like bottomhole temperature and wellbore temperature profile, can significantly minimize the uncertainty in reservoir model identification, particularly in more heterogeneous reservoirs like multi-layered systems.In the current study, the permeability of various reservoir models has been estimated via optimization approaches. For this purpose, uncertain reservoir parameters have been considered as optimization variables. The particle swarm optimization (PSO) and differential evolution (DE) algorithms have been utilized to find the optimal value of these parameters. The temperature and pressure data have been assumed as the input parameters and were simultaneously analyzed to determine the properties of reservoir layers more accurately. Some statistical parameters were employed to verify the accuracy of the developed approaches. Using simultaneous analysis of the bottomhole data and wellbore temperature profile, the average absolute percent relative error (AAPRE) values for the multi-layered heterogeneous reservoir model was 1.57%. This approach, compared to the bottomhole pressure analysis technique, in which AAPRE of 19.91% was obtained, led to significantly more reliable results.Furthermore, the best history-matched models created through various analysis approaches were selected to perform the prediction scenarios. Absolute relative error (ARE) values for the multi-layered heterogeneous reservoir model for bottomhole pressure analysis and simultaneous analysis were 12.00% and 0.26%, respectively. The results demonstrated that employing the simultaneous bottomhole temperature and pressure analysis approach has the lowest estimations error, reducing the uncertainties in the reservoir model predictions.

Modeling of wax disappearance temperature (WDT) using soft computing approaches: Tree-based models and hybrid models

Solid scales can cause significant problems in oil production and transmission systems such as oil flow rate reduction. Wax is one of the most critical substances that are highly prone to precipitate and deposit in oil pipelines. This occurrence typically happens in low temperatures, and can extremely affect the flow rate. In the wax deposition process, wax disappearance temperature (WDT) is a crucial factor since it denotes the lowest temperature, at which wax deposits melt and the consequences of wax deposition vanish. In this research, smart models were utilized to predict WDT as a function of molar mass and pressure according to a comprehensive database. The proposed intelligent models are radial basis function (RBF) and multilayer perceptron (MLP) neural networks, adaptive neuro-fuzzy inference system (ANFIS), random forest (RF), extra tree (ET), and decision tree (DT). The MLP networks were trained with Bayesian Regularization (BR) and Levenberg-Marquardt (LM) algorithms, and the ANFIS models were coupled with three optimization algorithms, namely Cultural Algorithm (CA), Biogeography-based Optimization (BBO), and Teaching-Learning-Based Optimization (TLBO). Results demonstrate that the developed RF model could present an outstanding performance and provide predictions for WDT with the lowest average absolute percent relative error (AAPRE = 0.246%), the maximum coefficient of determination (for predicting = 0.993), and minimum root mean squared error (RMSE = 1.01). Trend analysis showed that increasing pressure and molar mass leads to wax disappearance temperature increase. Lastly, outlier discovery was performed using the Leverage approach to recognize the suspected data points. The outlier detection found that only 6 points (out of 346) are located in the upper and lower suspected data zones. Wax disappearance temperature is modeled as a function of pressure and crude oil molar mass. RBF, MLP, ANFIS, DT, RF, and ET are used for modeling. LM, BR, CA, BBO, and TLBO are used to optimize the developed models. RF as the best model could predict WDT with the lowest average absolute percent relative error (AAPRE = 0.246%) The Leverage approach demonstrate the reliability of the developed RF model.

Modelling density of pure and binary mixtures of normal alkanes: Comparison of hybrid soft computing techniques, gene expression programming, and equations of state

Determining the properties of hydrocarbons, especially density, is one of the most important measures in the oil and gas industry. In this study, robust artificial intelligence techniques have been investigated to predict the density of pure and binary mixtures of normal alkanes (between C 1 and C 44 ). Since there are various conditions in oil and gas reservoirs, it has been tried to study the properties of different hydrocarbons in a wide range of temperature (up to 522 K) and pressure (up to 275 MPa). An extensive databank, including 2143 and 985 data points for pure and binary mixtures of normal alkanes, respectively, was extracted from the literature. The crow search algorithm (CSA), firefly algorithm (FFA), grey wolf optimization (GWO), and wind-driven optimization (WDO) algorithms were utilized to improve the learning process of least square support vector machine (LSSVM) and radial basis function neural network (RBFNN) models, which were developed to predict the density of hydrocarbons. Also, gene expression programming (GEP) correlations were presented using complex mathematical calculations to estimate the density of hydrocarbons. The results obtained from the models were also compared with seven equations of state (EOSs). The obtained results showed that the predictions of the proposed techniques are in a great match with the experimental data. By performing a comparison on the models’ outcomes, LSSVM-GWO and RBFNN-CSA were found to be the most accurate models for pure and binary mixtures of normal alkanes with overall average absolute percent relative error (AAPRE) values of 0.0622% and 0.0098%, respectively. It is noteworthy that the GEP correlations with the AAPRE values of 0.1955% and 0.3525% for pure and binary mixtures of normal alkanes, respectively, have a high accuracy compared to the equations of state and are suitable practical correlations for estimating the density of hydrocarbons. • Density of pure and binary mixtures of normal alkanes is modeled. • 2143 and 985 data points for pure and binary mixtures of normal alkanes, respectively, are used. • LSSVM and RBFNN models are optimized by CSA, FFA, GWO, and WDO. • Simple-to-use correlations are developed by GEP. • The proposed models outperform equations of state.

Predicting formation damage of oil fields due to mineral scaling during water-flooding operations: Gradient boosting decision tree and cascade-forward back-propagation network

Water-flooding is one of the main options employed by the oil industry to meet the world's ever-increasing demand for oil, as the primary source of energy. This approach is highly prone to cause formation damage if the injected water is not compatible with the formation brine. In this study, decision tree optimized with gradient boosting (GBDT), cascade-forward back-propagation network (CFBPN), and generalized regression neural networks (GRNN) were employed, as relatively novel intelligent models, for the first time to develop accurate models to estimate the formation damage during a waterflooding operation in terms of damaged permeability. To compare the performance of these models, radial basis function (RBF) and multilayer perceptron (MLP) neural networks were also developed. The Levenberg-Marquardt algorithm (LMA), scaled conjugate gradient (SCG), and Bayesian regularization (BR) were used for training the MLP and CFBPN models. The results of this study showed the outperformance of the proposed GBDT model compared to the other developed models as well as previously proposed ones with an average absolute percent relative error (AAPRE) of 0.1465 % and correlation coefficient (R 2 ) of 0.9991 for the whole dataset. According to the results, the accuracy of the developed models could be ranked as follows: GBDT > CFBPN-LM > CFBPN-BR > RBF > MLP-LM > GRNN > MLP-BR > CFBPN-SCG > MLP-SCG. Moreover, it was shown that the GBDT mode could estimate more than 90 % of points with an absolute relative error of lower than 0.5 %. The trend analysis showed the high capability of the developed models in detecting the physical trend of the formation damage with variation of inputs. Then, the variable impact analysis was performed for this model, and the results reflect the high dependency of the model's predictions on the volume of injected water (Vinj), initial permeability (Ki), and ionic concentration of sulfate. Lastly, the Leverage approach was employed to determine suspected points as well as the applicability realm of the GBDT model. The results of the outlier detection indicated that only 4 points (0.93 % of the dataset) were detected as outliers, and the applicability realm of the proposed GBDT was verified. The findings of this communication shed light on the application of intelligent models and their power in predicting the formation damage caused during water-flooding operations before their occurrence. • Formation damage during water-flooding is modelled using GBDT, RBF, MLP, CFBPN, and GRNN. • LMA, SCG, and BR are used for optimizing MLP and CFBPN. • GBDT outperforms all models with AAPRE of 0.14 %. • The Leverage approach was used to determine suspected points and as the applicability realm of the GBDT model.

Modeling of H2S solubility in ionic liquids using deep learning: A chemical structure-based approach

Hydrogen sulfide (H2S) is a toxic, flammable, corrosive, and acidic gas that can have harmful impacts on the environment. Ionic liquids (ILs) are relatively new solutions that have desirable characteristics such as high thermal and electrochemical stabilities, negligible vapor pressure, non-combustibility, and high solubility, allowing them to be an appropriate choice of liquid solvents in gas removal processes. In this study, deep learning (DL) approaches were utilized to predict the H2S solubility in ionic liquids (ILs). The proposed models include convolutional neural network (CNN), recurrent neural networks (RNNs), deep belief networks (DBN), and deep neural network initialization with decision trees (DJINN). To this end, a large databank in a broad range of pressure and temperature encompassing 1516 data points of H2S solubility in 37 various ILs was used for developing models. In these models, the chemical structure of ILs, temperature, and pressure were considered as the model’s inputs, while H2S solubility was the model’s output. The results reveal that the CNN model provides more accurate, rapid, flexible, and inexpensive estimations for H2S solubility in ILs with an average absolute percent relative error (AAPRE) of 2.92% and a determination coefficient (R2) of 0.99. The results were then compared to those of previously proposed techniques in the literature. According to the results of the sensitivity analysis, it was found that pressure and OH (as substructure) impose the highest positive and the highest negative impacts on the solubility of H2S in ILs, respectively. Also, based on sensitivity analysis, temperature has a reverse effect on the solubility of H2S and with increasing temperature, the H2S solubility decreases. The Taylor diagram illustrated that the CNN approach is very efficient, accurate, and realistic for predicting the solubility of H2S in diverse ILs based on the chemical structure, pressure, and temperature. Finally, trend analysis revealed that the trends of CNN predictions are in great agreement with the measured data of H2S solubility in ILs as a function of pressure. The findings of this work may be used not only to overcome the difficulties of calculating H2S solubility in ILs, but also to develop novel and accurate forecasting algorithms for a large dataset that cover a broad range of temperatures and pressures.

Experimental measurement and modeling of water-based drilling mud density using adaptive boosting decision tree, support vector machine, and K-nearest neighbors: A case study from the South Pars gas field

Exact determination of drilling mud weight in order to prevent fracture pressure and at the same time overcoming pore pressure is a key parameter in the drilling of oil and gas wells. Accurate design of drilling mud weight increases drilling safety, reduces mud loss, and also the duration of drilling operations. In this study, five efficient artificial intelligent models including Bayesian ridge regression (BRR), K-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT), and Adaptive Boosting Regressor with Decision Tree (ABR-DT) were proposed for estimating mud weight based on a databank of 817 data points from five wells in the South Pars gas field. Variables influencing mud weight include true vertical depth (TVD), hole size, inclination, funnel viscosity, yield point (YP), plastic viscosity (PV), gel strength (measured at 10th second, 10th minute, and 30th minute), API fluid loss (FL), mud type (seawater, KCl-polymer, and salt-polymer), and formation lithology. According to the statistical analysis, the proposed ABR-DT and DT were the most accurate approaches that could predict mud weight with average absolute percent relative errors (AARE) of 0.5159% and 0.7560%, respectively. The accuracy of the models was ranked as: ABR-DT > DT > SVM > KNN > BRR. The sensitivity analysis showed that the predicted mud density is highly influenced by the values of plastic viscosity and true vertical depth. Lastly, the Leverage approach confirmed the validity of the employed data and the applicability area of the proposed ABR-DT model. • BRR, KNN, SVM, DT, and ABR-DT intelligent models were proposed for estimating water-base mud density. • A databank of 817 data points from five wells in the South Pars gas field was used. • Actual field data were used for the first time as inputs of the models. • The accuracy of the models was ranked as: ABR-DT > DT > SVM > KNN > BRR. • The predicted mud density is highly influenced by the values of plastic viscosity and true vertical depth.

Modeling hydrogen solubility in hydrocarbons using extreme gradient boosting and equations of state

Due to industrial development, designing and optimal operation of processes in chemical and petroleum processing plants require accurate estimation of the hydrogen solubility in various hydrocarbons. Equations of state (EOSs) are limited in accurately predicting hydrogen solubility, especially at high-pressure or/and high-temperature conditions, which may lead to energy waste and a potential safety hazard in plants. In this paper, five robust machine learning models including extreme gradient boosting (XGBoost), adaptive boosting support vector regression (AdaBoost-SVR), gradient boosting with categorical features support (CatBoost), light gradient boosting machine (LightGBM), and multi-layer perceptron (MLP) optimized by Levenberg–Marquardt (LM) algorithm were implemented for estimating the hydrogen solubility in hydrocarbons. To this end, a databank including 919 experimental data points of hydrogen solubility in 26 various hydrocarbons was gathered from 48 different systems in a broad range of operating temperatures (213–623 K) and pressures (0.1–25.5 MPa). The hydrocarbons are from six different families including alkane, alkene, cycloalkane, aromatic, polycyclic aromatic, and terpene. The carbon number of hydrocarbons is ranging from 4 to 46 corresponding to a molecular weight range of 58.12–647.2 g/mol. Molecular weight, critical pressure, and critical temperature of solvents along with pressure and temperature operating conditions were selected as input parameters to the models. The XGBoost model best fits all the experimental solubility data with a root mean square error (RMSE) of 0.0007 and an average absolute percent relative error (AAPRE) of 1.81%. Also, the proposed models for estimating the solubility of hydrogen in hydrocarbons were compared with five EOSs including Soave–Redlich–Kwong (SRK), Peng–Robinson (PR), Redlich–Kwong (RK), Zudkevitch–Joffe (ZJ), and perturbed-chain statistical associating fluid theory (PC-SAFT). The XGBoost model introduced in this study is a promising model that can be applied as an efficient estimator for hydrogen solubility in various hydrocarbons and is capable of being utilized in the chemical and petroleum industries.

Estimation of tetracycline antibiotic photodegradation from wastewater by heterogeneous metal-organic frameworks photocatalysts

In this work, the potential ability of various modern and powerful machine learning methods such as Categorical Boosting (CatBoost), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), Adaptive Boosting (AdaBoost), Gradient-Boosted Decision Trees (GBDT), Extra Tree (ET), Decision Trees (DT), and Random Forest (RF) were investigated to estimate tetracycline (TC) photodegradation from wastewater by 10 different metal-organic frameworks (MOFs). A comprehensive databank was gathered, including 374 data points from the photodegradation percentage of MOFs in various practical conditions. The inputs of the employed models were chosen as catalyst dosage, antibiotic concentration, Illumination time, solution pH, and specific surface area and pore volume of the investigated MOFs, and the output was TC degradation efficiency. Different statistical criteria were calculated for the validation of the developed models. Average absolute percent relative error (AAPRE) and standard deviation error (STD) values of 1.19% and 0.0431, 3.07% and 0.0628, 2.88% and 0.0751, 2.86% and 0.1304, 8.73% and 0.2751, 4.24% and 0.1024, 2.83% and 0.0934, and 11.56% and 0.4459 were obtained for CatBoost, LightGBM, XGBoost, AdaBoost, GBDT, ET, DT, and RF approaches, respectively. Among all implemented models, the CatBoost was found to be the most trustable model. Moreover, this model followed the expected trends of the TC degradation process with variation of catalyst dosage, initial TC concentration, and reaction pH. The developed CatBoost model predicted the removal of TC by MOFs accurately, which proved the capability of this approach in solving complex problems with numerous data points and its straightforwardness and cost-effectiveness for environmental applications.

Deep Learning Approach for Robust Prediction of Reservoir Bubble Point Pressure.

The bubble point pressure (Pb) is a crucial pressure–volume–temperature (PVT) property and a primary input needed for performing many petroleum engineering calculations, such as reservoir simulation. The industrial practice of determining Pb is by direct measurement from PVT tests or prediction using empirical correlations. The main problems encountered with the published empirical correlations are their lack of accuracy and the noncomprehensive data set used to develop the model. In addition, most of the published correlations have not proven the relationships between the inputs and outputs as part of the validation process (i.e., no trend analysis was conducted). Nowadays, deep learning techniques such as long short-term memory (LSTM) networks have begun to replace the empirical correlations as they generate high accuracy. This study, therefore, presents a robust LSTM-based model for predicting Pb using a global data set of 760 collected data points from different fields worldwide to build the model. The developed model was then validated by applying trend analysis to ensure that the model follows the correct relationships between the inputs and outputs and performing statistical analysis after comparing the most published correlations. The robustness and accuracy of the model have been verified by performing various statistical analyses and using additional data that was not part of the data set used to develop the model. The trend analysis results have proven that the proposed LSTM-based model follows the correct relationships, indicating the model’s reliability. Furthermore, the statistical analysis results have shown that the lowest average absolute percent relative error (AAPRE) is 8.422% and the highest correlation coefficient is 0.99. These values are much better than those given by the most accurate models in the literature.

An Artificial Intelligent Approach for Black Oil PVT Properties

In absence of experimental oil fluid samples, it is usually difficult to select the suitable correlation to estimate oil properties. However, the accuracy of these empirical correlations has become insufficient for the best calculation. The main objective of this work is to test the capability of Particle Swarm Optimization with Neural Network (PSONN) and Neuro-Fuzzy (NFuzzy) approaches to predict oil properties with simply and accurately. The proposed approaches are developed based on clustering the oil data into the three groups (light, medium and heavy light oil). Over five hundred of black oil samples were collected from Middle East Field to train the hybrid models whereas additional oil data samples were selected to validate. The developed models used to estimate bubble points pressure (Pb), solution gas oil ratio at and below Pb, undersaturated oil compressibility, saturated formation volume factor, saturated and undersaturated density. The recommended guidelines and optimal configuration of the PSONN and NFuzzy models are developed to estimate any oil property in future. Statistical error analyses show that the proposed models exhibit a robust predictive capability for estimating oil properties. The validation results show the PSONN model achieve the lowest average absolute percent relative error (0.04, 2.89 and 1.0) for estimating formation volume factor, gas oil ratio and oil compressibility respectively whereas the NFuzzy model obtains the best approximation in oil density and bubble point pressure with average absolute percent relative error (0.18 and 0.97) respectively.

On the Evaluation of Interfacial Tension (IFT) of CO2–Paraffin System for Enhanced Oil Recovery Process: Comparison of Empirical Correlations, Soft Computing Approaches, and Parachor Model

Carbon dioxide-based enhanced oil-recovery (CO2-EOR) processes have gained considerable interest among other EOR methods. In this paper, based on the molecular weight of paraffins (n-alkanes), pressure, and temperature, the magnitude of CO2–n-alkanes interfacial tension (IFT) was determined by utilizing soft computing and mathematical modeling approaches, namely: (i) radial basis function (RBF) neural network (optimized by genetic algorithm (GA), gravitational search algorithm (GSA), imperialist competitive algorithm (ICA), particle swarm optimization (PSO), and ant colony optimization (ACO)), (ii) multilayer perception (MLP) neural network (optimized by Levenberg-Marquardt (LM)), and (iii) group method of data handling (GMDH). To do so, a broad range of laboratory data consisting of 879 data points collected from the literature was employed to develop the models. The proposed RBF-ICA model, with an average absolute percent relative error (AAPRE) of 4.42%, led to the most reliable predictions. Furthermore, the Parachor approach with different scaling exponents (n) in combination with seven equations of state (EOSs) was applied for IFT predictions of the CO2–n-heptane and CO2–n-decane systems. It was found that n = 4 was the optimum value to obtain precise IFT estimations; and combinations of the Parachor model with three-parameter Peng–Robinson and Soave–Redlich–Kwong EOSs could better estimate the IFT of the CO2–n-alkane systems, compared to other used EOSs.