Bahman Mehdizadeh

Publications

• Solubility prediction of anthracene in binary solvent systems by least square support vector machine

  Bahman Mehdizadeh, Kamyar Movagharnejad

  Computers & Chemical Engineering. 01/2011;

  Solubility prediction of solids is essential for process design involving precipitation, crystallization and many of other processes. Since solubility data are not often available correlation and prediction of solid solubility in different solvents are widely applied. In this paper least square supp... [more] Solubility prediction of solids is essential for process design involving precipitation, crystallization and many of other processes. Since solubility data are not often available correlation and prediction of solid solubility in different solvents are widely applied. In this paper least square support vector machine is applied for solubility modeling of anthracene in two binary solvents (toluene + 2-propanol and toluene + heptane) .Two approaches are applied for solubility modeling. In first approach, one model is optimized for each system. In this model temperature, volume percent of toluene and volume percent of heptane or 2-propanol are used as inputs. Results showed proposed models have an average relative deviation less than 3.88% for two systems. In second approach a generalized model is proposed by using molecular weights of solvents as structural property. Although, accuracy of latter proposed model is high but the most advantage of generalized model is dependency of this model to two adjusting parameter for all binary solvent systems, while in first approach each system require two adjusting parameters.

• A novel semi empirical equation for prediction the solute solubility in supercritical carbon dioxide

  Bahman Mehdizadeh, et al

  Australian Journal Of Basic And Applied Science. 01/2011;

• A Comparison between neural network method and semi empirical equations to predict the solubility of different compounds in supercritical carbon dioxide

  Bahman Mehdizadeh, Kamyar Movagharnejad

  Fluid Phase Equilibria. 01/2011;

  Accuracy of seven semi empirical equations for estimation of solubility of 30 different compounds in supercritical carbon dioxide has been compared with a new neural network method. To base this comparison on a fair basis, a unique set of experimental data was used for both optimization of semi empi... [more] Accuracy of seven semi empirical equations for estimation of solubility of 30 different compounds in supercritical carbon dioxide has been compared with a new neural network method. To base this comparison on a fair basis, a unique set of experimental data was used for both optimization of semi empirical equations’ parameters and training, validation and testing of neural network. Results showed that neural network method with an average relative deviation of about 5.3% was more accurate than the best semi empirical equation with an average relative deviation of about 15.96% for same compounds. It was also found that the average relative deviation of semi empirical equations vary sharply among different compounds, while this quantity is less dependent on material type for neural network method.

  Download

• A Comparative study between genetic algorithm - least square support vector machine method and semi empirical equations to predict the solubility of different solutes in supercritical carbon dioxide

  Bahman Mehhdizadeh, Kamyar Movagharnejad

  Chemical engineering research&design. 01/2011;

  Due to the high solvent ability and diffusivity and low viscosity, supercritical fluids have been used in different processes during the recent decades and many research has been done for characterization of these fluids. In this paper after gathering numerous published solubility data, a genetic al... [more] Due to the high solvent ability and diffusivity and low viscosity, supercritical fluids have been used in different processes during the recent decades and many research has been done for characterization of these fluids. In this paper after gathering numerous published solubility data, a genetic algorithm (GA) based least squares support vector machine (LS-SVM) and 7 semi empirical equations have been used for solubility prediction of 25 compounds in supercritical carbon dioxide. In present work GA-LS-SVM method with three inputs (temperature, pressure and density of supercritical carbon dioxide) has been used for the first time in order to solubility prediction of many solutes in supercritical carbon dioxide. Results showed GA-LS-SVM method present an average relative deviation about 4.92% for 25 solutes while the best semi empirical equations present an average relative deviation about 13.60% for same solutes.

  Download

• Application of least square support vector machine for drying modeling of some agricultural food products

  Bahman Mehdizadeh

  Energy. 01/2011;

• A genetic algorithm based Peng-Robinson equation of state for solubility modeling of cinnamic acid in supercritical carbon dioxide

  Bahman Mehdizadeh et al

  7th International Chemical Engineering Congress & Exhibition-- Kish, Iran, 21-24 November, 2011; 01/2011

• Prediction of second virial coefficient for pure compounds using a genetic algorithm based least square support vector machine

  Bahman Mehdizadeh

  Chemycal Physics Letter. 01/2011;

• Forecasting the differences between various commercial oil prices in Persian Gulf Region by neural network

  Bahman Mehdizadeh, Kamyar Movagharnejad, et al.

  Energy. 01/2011;

  Download

• Application of artificial neural network for solubility modeling of phenanthrene in two binary solvent mixtures

  Bahman Mehdizadeh, Kamyar Movagharnejad

  3rd Technical Conference Of Thermodynamics; 01/2011

  In this paper artificial neural network (ANN) is used for solubility modeling of phenanthrene in two binary solvent systems. In the next step, accuracy of our proposed models is compared to results of a published paper .In this published paper thermodynamic model with different activity coefficient ... [more] In this paper artificial neural network (ANN) is used for solubility modeling of phenanthrene in two binary solvent systems. In the next step, accuracy of our proposed models is compared to results of a published paper .In this published paper thermodynamic model with different activity coefficient equations such as (Wilson, NIBS/Redlich–Kister, UNIQUAC and Dortmund UNIFAC) had been used for solubility estimation. Results showed ANN is the most accurate model.

• Solubility modeling of diamines in supercritical carbon dioxide using artificial neural network

  Bahman Mehdizadeh, et al

  Australian journal Of Basic And Applied Science. 01/2011;

• Microchannel reactor-review

  Bahman Mehdizadeh, Majid Taghizadehand et al

  Fuel cell Seminar-Tehran-Iran. 01/2010;

  Download

• Solubility prediction of pyrene in binary solvent using artificial neural network and genetic algorithm based least square support vector machine

  Bahman Mehdizadeh, Kamyar Movagharnejad

  Fluid phase equilibria. 01/2010;

  Artificial neural network (ANN) and genetic algorithm based least square support vector machine (GA-LS-SVM) were used for solubility modeling of pyrene in six binary solvent systems. In the next step, results of our proposed models were compared to results of a published paper .In this published pap... [more] Artificial neural network (ANN) and genetic algorithm based least square support vector machine (GA-LS-SVM) were used for solubility modeling of pyrene in six binary solvent systems. In the next step, results of our proposed models were compared to results of a published paper .In this published paper thermodynamic model with different activity coefficient equations such as (Wilson, NIBS/Redlich–Kister, UNIQUAC, UNIFAC, UNIFAC-SG, SUPERFAC, SUPERFAC-SG, Dortmund UNIFAC and Dortmund UNIFAC-SG) had been used for solubility estimation. Results showed GA-LS-SVM is the most accurate model. Detailed results are expressed.