Experimental measurement and modeling of saturated reservoir oil viscosity

Abdolhossein Hemmati-Sarapardeh,Amir H. Mohammadi,Ahmad Ramazani S. A.,Seyed-Mohammad-Javad Majidi,Behnam Mahmoudi 한국화학공학회 2014 Korean Journal of Chemical Engineering Vol.31 No.7

A novel mathematical-based approach is proposed to develop reliable models for prediction of saturatedcrude oil viscosity in a wide range of PVT properties. A new soft computing approach, namely least square supportvector machine modeling optimized with coupled simulated annealing optimization technique, is proposed. Six modelshave been developed to predict saturated oil viscosity, which are designed in such a way that could predict saturatedoil viscosity with every available PVT parameter. The constructed models are evaluated by carrying out extensive experimentalsaturated crude oil viscosity data from Iranian oil reservoirs, which were measured using a “Rolling Ballviscometer.” To evaluate the performance and accuracy of these models, statistical and graphical error analyses wereused simultaneously. The obtained results demonstrated that the proposed models are more robust, reliable and efficientthan existing techniques for prediction of saturated crude oil viscosity.

Implementation of soft computing approaches for prediction of physicochemical properties of ionic liquid mixtures

Saeid Atashrouz,Hamed Mirshekar,Abdolhossein Hemmati-Sarapardeh,Mostafa Keshavarz Moraveji,Bahram Nasernejad 한국화학공학회 2017 Korean Journal of Chemical Engineering Vol.34 No.2

The main objective of this study was to develop soft computing approaches for prediction of physicochemical properties of IL mixtures including: density, heat capacity, thermal conductivity, and surface tension. The proposed models in this study are based on support vector machine (SVM), least square support vector machines (LSSVM), and group method of data handling type polynomial neural network (GMDH-PNN) systems. To find the LSSVM and SVM adjustable parameters, genetic algorithm (GA) as a meta-heuristic algorithm was utilized. The results showed that LSSVM is more robust and reliable for prediction of physicochemical properties of IL mixtures. The proposed GALSSVM model provides average absolute relative deviations of 0.38%, 0.18%, 0.77% and 1.18% for density, heat capacity, thermal conductivity, and surface tension, respectively, which demonstrates high accuracy of the model for prediction of physicochemical properties of IL mixtures.

Using an artificial neural network to predict carbon dioxide compressibility factor at high pressure and temperature

Erfan Mohagheghian,Habiballah Zafarian-Rigaki,Yaser Motamedi-Ghahfarrokhi,Abdolhossein Hemmati-Sarapardeh 한국화학공학회 2015 Korean Journal of Chemical Engineering Vol.32 No.10

Carbon dioxide injection, which is widely used as an enhanced oil recovery (EOR) method, has the potential of being coupled with CO2 sequestration and reducing the emission of greenhouse gas. Hence, knowing the compressibility factor of carbon dioxide is of a vital significance. Compressibility factor (Z-factor) is traditionally measured through time consuming, expensive and cumbersome experiments. Hence, developing a fast, robust and accurate model for its estimation is necessary. In this study, a new reliable model on the basis of feed forward artificial neural networks is presented to predict CO2 compressibility factor. Reduced temperature and pressure were selected as the input parameters of the proposed model. To evaluate and compare the results of the developed model with pre-existing models, both statistical and graphical error analyses were employed. The results indicated that the proposed model is more reliable and accurate compared to pre-existing models in a wide range of temperature (up to 1,273.15 K) and pressure (up to 140MPa). Furthermore, by employing the relevancy factor, the effect of pressure and temprature on the Z-factor of CO2 was compared for below and above the critical pressure of CO2, and the physcially expected trends were observed. Finally, to identify the probable outliers and applicability domain of the proposed ANN model, both numerical and graphical techniques based on Leverage approach were performed. The results illustrated that only 1.75% of the experimental data points were located out of the applicability domain of the proposed model. As a result, the developed model is reliable for the prediction of CO2 compressibility factor.

Modeling the permeability of heterogeneous oil reservoirs using a robust method

Arash Kamari,Farzaneh Moeini,Mohammad-Javad Shamsoddini-Moghadam,Seyed-Ali Hosseini,Amir H. Mohammadi,Abdolhossein Hemmati-Sarapardeh 한국지질과학협의회 2016 Geosciences Journal Vol.20 No.2

Permeability as a fundamental reservoir property plays a key role in reserve estimation, numerical reservoir simulation, reservoir engineering calculations, drilling planning, and mapping reservoir quality. In heterogeneous reservoir, due to complexity, natural heterogeneity, non-uniformity, and non-linearity in parameters, prediction of permeability is not straightforward. To ease this problem, a novel mathematical robust model has been proposed to predict the permeability in heterogeneous carbonate reservoirs. To this end, a fairly new soft computing method, namely least square support vector machine (LSSVM) modeling optimized with coupled simulated annealing (CSA) optimization technique was utilized. Statistical and graphical error analyses have been employed separately to evaluate the accuracy and reliability of the proposed model. Furthermore, this model performance has been compared with a newly developed multilayer perceptron artificial neural network (MLP-ANN) model. The obtained results have shown the more robustness, efficiency and reliability of the proposed CSA-LSSVM model in comparison with the developed MLP-ANN model for the prediction of permeability in heterogeneous carbonate reservoirs. Estimations were found to be within acceptable agreement with the actual field data of permeability, with a root mean square error of approximately 0.42 for CSA-LSSVM model in testing phase, and a R-squared value of 0.98. Additionally, these error parameters for MLP-ANN are 0.68 and 0.89 in testing stage, respectively.

Modeling CO2 Loading Capacity of Diethanolamine (DEA) Aqueous Solutions Using Advanced Deep Learning and Machine Learning Algorithms: Application to Carbon Capture

Mahmoudzadeh Atena,Hadavimoghaddam Fahimeh,Atashrouz Saeid,Abedi Ali,Abuswer Meftah Ali,Mohaddespour Ahmad,Hemmati-Sarapardeh Abdolhossein 한국화학공학회 2024 Korean Journal of Chemical Engineering Vol.41 No.5

Several carbon capture techniques have been developed in response to the notable rise of atmospheric carbon dioxide ( CO2 ) levels. The utilization of diethanolamine (DEA) as an absorption method is prevalent in various industries due to its high reactivity and cost-effi ciency. Hence, comprehending the equilibrium solubility of CO2 in DEA solutions is an essential step in developing and optimizing absorption procedures. In order to predict the CO2 loading capacity in the DEA solutions, four advanced deep learning and machine learning models were developed: recurrent neural networks (RNN), deep neural networks (DNN), random forest (RF), and adaBoost-support vector regression (AdaBoost-SVR). The models predict the capacity of CO2 loading as a function of temperature, CO2 partial pressure, and the concentration of DEA in the solution. Intelligent models were developed employing an extensive database which includes new experimental data points published within recent years, which were not considered in the previous studies. The RNN model was found to outperform other models based on graphical and statistical assessments, as evidenced by its lower root mean square error ( RMSE = 0.285 ) and standard deviation ( SD = 0.032 ), and higher determination coeffi cient ( R2 = 0.992 ). While the RNN model resulted in the highest accuracy in predicting CO2 absorption, the DNN, RF, and AdaBoost-SVR models also demonstrated satisfactory accuracy in predicting CO2 solubility, placed in the following ranking. A sensitivity analysis was performed on the four developed models, revealing that the CO2 partial pressure has the strongest eff ect on the CO2 loading capacity. Furthermore, a trend analysis was performed on the RNN model, demonstrating that the developed model has a high degree of accuracy in following physical trends. The binary interaction analysis was conducted with two varying parameters and one constant parameter in the RNN model through 3-D image plots, which illustrated the simultaneous eff ect of two independent parameters on CO2 loading. Finally, outlier detection was conducted by employing the Leverage method to fi nd outlier data points in the data bank, demonstrating the applicability domain of intelligent models.