C3H8 separation from CH4 and H2 using a synthesized PDMS membrane: Experimental and neural network modeling

A polydimethylsiloxane (PDMS) membrane was synthesized and permeation behavior of ternary gas mixtures including C₃H₈, CH₄ and H₂ through it was studied as a function of operating parameters. Mixed gas permeability values were also compared with pure gas data as well as literature to validate experimental results. The aim was to predict separation factor (SF) of C₃H₈ as a function of feed temperature, pressure, flow rate and C₃H₈ concentration with the aid of artificial neural network (ANN) technique. Multilayer perceptron (MLP), which is the most common type of feedforward neural network (FFNN), was used for prediction. The Levenberg–Marquardt training method was initially employed to train the net. Then, optimum numbers of hidden layers and nodes in each layer were determined. The selected structure (4:4:5:1) was finally used to predict SF of C₃H₈ for different inputs in the domain of training data. The modeling results showed that there is an excellent agreement between the experimental data and the predicted values, with mean absolute errors of less than 1%. Both modeling and experimental results confirmed that increasing feed temperature, feed pressure and C₃H₈ concentration in feed debilitates separation performance; however, SF increases with increasing feed flow rate. As a result, ANN can be recommended for the modeling of mixed gas transport through dense membranes such as PDMS.     

Optimization of artificial neural network for retention modeling in high-performance liquid chromatography

An artificial neural network (ANN) model for the prediction of retention times in high-performance liquid chromatography (HPLC) was developed and optimized. A three-layer feed-forward ANN has been used to model retention behavior of nine phenols as a function of mobile phase composition (methanol-acetic acid mobile phase). The number of hidden layer nodes, number of iteration steps and the number of experimental data points used for training set were optimized. By using a relatively small amount of experimental data (25 experimental data points in the training set), a very accurate prediction of the retention (percentage normalized differences between the predicted and the experimental data less than 0.6%) was obtained. It was shown that the prediction ability of ANN model linearly decreased with the reduction of number of experiments for the training data set. The results obtained demonstrate that ANN offers a straightforward way for retention modeling in isocratic HPLC separation of a complex mixture of compounds widely different in pK_a and log K_ow values.     

Computer simulation for the pervaporation process of wate/ethanol mixture through interpenetrating polymer network (IPN) membranes

In this study, the pervaporation behavior of EtOH-water mixture through interpenetrating polymer network (IPN) membranes was predicted. The pervaporation characteristics of single component membranes were modelled according to the “six coefficients model” proposed by Brun [J. Membrane Sci., 23 (1985) 257]. In the case of the IPN membrane, two models were proposed according to the phase structure of the IPN. For a uniphase membrane with no phase separation, the compositional averages of the single component membrane coefficients were used. In the case of the phase separated IPN, two cases exist. The first is the island and sea model: one phase is continuous and the other is discrete. The second is the cocontinuous model, in which two continuous phases exist. For these cases the permeation rate and separation factor of the IPN membrane were calculated using the experimental sorption data and pure component values for each IPN composition. Comparison with the experimental data indicates that these models could be to predict the performances of IPN membranes depending on the morphology of the IPN.     

Separation of organic acid compounds from biological samples by zinc oxide nanoparticles–chitosan using genetic algorithm based on response surface methodology and artificial neural network

In this study, zinc oxide nanoparticles–chitosan based on solid phase extraction and high performance liquid chromatography was developed for the separation of organic compounds including citric, tartaric and oxalic acids from biological samples. For simulation and optimization of this method, the hybrids of genetic algorithm with response surface methodology (RSM) and artificial neural network (ANN) have been used. The predictive capability and generalization of both predictive models (RSM and ANN) were compared by unseen data. The results have shown the superiority of ANN compared with RSM. At the optimum conditions, the limits of detections of 2.2–2.9 µg L⁻¹ were obtained for the analytes. The developed procedure was then applied to the extraction and determination of organic acid compounds from biological samples. Copyright © 2014 John Wiley & Sons, Ltd.     

Comparison of general rate model with a new model--artificial neural network model in describing chromatographic kinetics of solanesol adsorption in packed column by macroporous resins

Herein, two models, the general rate model taking into account convection, axial dispersion, external and intra-particle mass transfer resistances and particle size distribution (PSD) and the artificial neural network model (ANN) were developed to describe solanesol adsorption process in packed column using macroporous resins. First, Static equilibrium experiments and kinetic experiments in packed column were carried out respectively to obtain experimental data. By fitting static experimental data, Langmuir isotherm and Freundlich isotherm were estimated, and the former one was used in simulation coupled with general rate model considering better correlative coefficients. The simulated results showed that theoretical predictions of general rate model with PSD were well consistent with experimental data. Then, a new model, the ANN model, was developed to describe present adsorption process in packed column. The encouraging simulated results showed that ANN model could describe present system even better than general rate model. At last, by using the predictive ability of ANN model, the influence of each experimental parameter was investigated. Predicted results showed that with the increases of particle porosity and the ratio of bed height to inner column diameter (ROHD), the breakthrough time was delayed. On the contrary, an increase in feed concentration, flow rate, mean particle diameter and bed porosity decreased the breakthrough time.     

Multicomponent adsorption modeling: isotherms for ABE model solutions using activated carbon F-400

Biobutanol has attracted significant interest in recent decades and is seriously considered as a potential biofuel to partly replace gasoline. However, some production challenges must be addressed to make butanol economically viable such as the low product concentration and product toxicity inhibiting the microorganism. To alleviate these limitations, several in situ or ex situ separation techniques have been investigated in view of their integration to the biobutanol production process to enhance its economic viability. One of these techniques is adsorption which is one of the most energy-efficient techniques used for biobutanol separation. Considering the number of chemical species present in the ABE fermentation broth, it is essential to develop multicomponent adsorption isotherms for all components as a first step to design a high performance adsorption process. Few multicomponent isotherm models have been proposed such as multicomponent Langmuir and Freundlich. In this study, these two models as well as artificial neural networks were used to model the isotherms of each component in an ABE fermentation broth as a function of the equilibrium concentrations of all components for activated carbon F-400. Results showed that the multicomponent Langmuir model was not accurate due to the many simplifying assumptions. The multicomponent Freundlich and feedforward neural network (FFNN) isotherm models were able to predict the behavior of multicomponent systems very well. Indeed, the predictive model of the experimental data had a coefficient of determination (R²) of 0.97 and 0.99, for multicomponent Freundlich and FFNN isotherm models, respectively.     

The neural network and multivariate linear regression approach for observing phase transitions of polymers with the differential thermal analysis method

The aim of this study was to correlate the results of experimental data using DTA method and predictions of artificial neural network (ANN) and multivariate linear regression (MLR). Thermal decomposition of polymers was analyzed by simultaneous DTA method, and kinetic parameters (critical points, the change of enthalpy and entropy) of polymers were investigated. A computer model based on multilayer feed forwarding back propagation and multilayer linear regression model were used for the prediction of critical points, phase transitions of low-density polyethylene (LDPE) and mid-density polyethylene. As a result of our study, we concluded that ANN model is more suitable than MLR about prediction of experimental data.     

Concentration polarization in gas permeation

The aim of this article is to show the importance of concentration polarization effects in the separation of gas–vapor mixtures using membranes. In the experimental part of this work, gas mixture measurements are conducted with a specially designed test cell. The experimental data are analyzed using a two-resistance model for the transport through the membrane, which is derived in the theoretical part of this work. The two resistances considered are the transport through the boundary layer on the feed side of the membrane and through the separation layer. For the transport through the separation layer an extended free volume model is derived. This model considers not only the feed side but also the influence of the permeate side on the separation properties of the membrane. The results of the measurements show the influence of concentration polarization effects and their dependencies on feed pressure, membrane thickness, and feed flow rate.     

Artificial neural networks in analysis of indinavir and its degradation products retention

Artificial neural networks (ANN) are biologically inspired computer programs designed to simulate the way in which the human brain processes the information. In the past few years, coupling of experimental design (ED) and ANN became useful tool in the method optimization. This paper presents the application of ED-ANN in analysis of chromatographic behavior of indinavir and its degradation products. According to preliminary study, full factorial design 2⁴ was chosen to set input variables for network training. Experimental data (inputs) and results for retention factors from experiments (outputs) were used to train the ANN with aim to define correlation among variables. For networks training multi-layer perceptron (MLP) with back propagation (BP) algorithm was used. Network with the lowest root mean square (RMS) had 4-8-3 topology. Predicted data were in good agreement with experimental data (correlation was higher than 0.9713 for training set). Regression statistics confirmed good ability of trained network to predict compounds retention.     

Prediction of rheological behavior of MWCNTs–SiO2/EG–water non-Newtonian hybrid nanofluid by designing new correlations and optimal artificial neural networks

In this paper, at the first, new correlations were proposed to predict the rheological behavior of MWCNTs–SiO₂/EG–water non-Newtonian hybrid nanofluid using different sets of experimental data for the viscosity, consistency and power law indices. Then, based on minimum prediction errors, two optimal artificial neural network models (ANNs) were considered to forecast the rheological behavior of the non-Newtonian hybrid nanofluid. One hundred and ninety-eight experimental data were employed for predicting viscosity (Model I). Two sets of forty-two experimental data also were considered to predict the consistency and power law indices (Model II). The data sets were divided to training and test sets which contained respectively 80 and 20% of data points. Comparisons between the correlations and ANN models showed that ANN models were much more accurate than proposed correlations. Moreover, it was found that the neural network is a powerful instrument in establishing the relationship between a large numbers of experimental data. Thus, this paper confirmed that the neural network is a reliable method for predicting the rheological behavior of non-Newtonian nanofluids in different models.

     

Study on growth rate of TiO2 nanostructured thin films: simulation by molecular dynamics approach and modeling by artificial neural network

Effects of the deposition process parameters on the thickness of TiO₂ nanostructured film were simulated using the molecular dynamics (MD) approach and modeled by the artificial neural network (ANN) and regression method. Accordingly, TiO₂ nanostructured film was prepared experimentally with the sol–gel dip‐coating method. Structural instabilities can be expected, due to short‐ and/or long‐range intermolecular forces, leading to the surface inhomogeneities. In the MD simulation, the Morse potential function was used for the inter‐atomic interactions, and equations of motion for atoms were solved by Verlet algorithm. The effect of the withdrawal velocity, drying temperature and number of deposited layers were studied in order to characterize the film thickness. The results of MD simulations are reasonably consistent with atomic force microscopy, scanning electron microscopy and Dektak surface profiler. Finally, the outputs from experimental data were analyzed by using the ANN in order to investigate the effects of deposition process parameters on the film thickness. In this case, various architectures have been checked using 75% of experimental data for training of the ANN. Among the various architectures, feed‐forward back‐propagation network with trainer training algorithm was found as the best architecture. Based on the R‐squared value, the ANN is better than the regression model in predicting the film thickness. The statistical analysis for those results was then used to verify the fitness of the complex process model. Based on the results, this modeling methodology can explain the characteristics of the TiO₂ nanostructured thin film and growth mechanism varying with process conditions. © 2013 The Authors. Surface and Interface Analysis published by John Wiley & Sons Ltd.     

Computer-assisted multifactor optimization of selectivity in gas chromatography

Summary A computer-assisted method is presented for the simultaneous multifactor optimization (stationary phase loading, column temperature and carrier gas flow rate) of the analytical conditions for the optimum separation of multicomponent samples in gas chromatography. The optimization of the separation over the experimental region is based on a special polynomial from twelve preliminary experiments using the resolution as the selection criterion. Computer scanning technique was used for optimum selection in three dimensions. Excellent agreement was obtained between the predicted data and the experimental results.     

Study of the permeability characteristics of porous media with methane hydrate by pore network model

The permeability in the methane hydrate reservoir is one of the key parameters in estimating the gas production performance and the flow behavior of gas and water during dissociation. In this paper, a three-dimensional cubic pore-network model based on invasion percolation is developed to study the effect of hydrate particle formation and growth habit on the permeability. The variation of permeability in porous media with different hydrate saturation is studied by solving the network problem. The simulation results are well consistent with the experimental data. The proposed model predicts that the permeability will reduce exponentially with the increase of hydrate saturation, which is crucial in developing a deeper understanding of the mechanism of hydrate formation and dissociation in porous media.     

Analysis of retrograde behavior and the cross-over effect in supercritical fluids

In this paper, results based upon thermodynamic stability theory are developed which lead to a set of both necessary and sufficient conditions for the existence of retrograde behavior in a multicomponent solute system dissolved in a pure supercritical fluid. While experimental evidence of retrograde behavior in single solute systems has been known for some time, recently data have been obtained showing the retrograde effect in binary solute systems dissolved in a pure supercritical fluid.In such systems, cross-over regions may be defined. These are pressure—temperature regions where the solubility of one solute increases while that of the other decreases with a change in temperature at constant pressure. The existence of cross-over regions in multicomponent mixtures can have implications for a new separation technique using pure supercritical fluids. In conjunction with an equation of state, the results derived here allow cross-over regions to be predicted, thus enabling one to identify candidate systems and thermodynamic conditions for the cross-over process. For this work a variation of a perturbed hard sphere model equation of state was used for the calculations.     

Studies on the applicability of artificial neural network (ANN) in emulsion liquid membranes

An artificial neural network (ANN) model of emulsion liquid membrane (ELM) process is proposed in the present study which is able to predict solute concentration in feed during extraction operation and ultimate % extraction at different initial solute concentration in feed phase, internal reagent concentration, treat ratio, volume fraction of internal aqueous phase in emulsion and time. Because of the complexity in generalization of the phenomenon of ELM process by any mathematical model, the neural network proves to be a very promising method for the purpose of process simulation. The network uses the back-propagation algorithm (BPA) for evaluating the connection strengths representing the correlations between inputs (initial solute concentration in feed phase, internal reagent concentration, treat ratio, volume fraction of internal aqueous phase in emulsion and time) and outputs (solute concentration in feed during extraction operation and % extraction). The network employed in the present study uses five input nodes corresponding to the operating variables and two output nodes corresponding to the measurement of the performance of the network (solute concentration in feed during extraction and % extraction). Batch experiments are performed for separation of nickel(II) from aqueous sulphate solution of initial concentration in the 200–100 mg/l ranges. The network employed in the present study uses two hidden layers of optimum number of nodes being thirty and twenty. A leaning rate of 0.3 and momentum factor of 0.4 is used. The model predicted results in good agreement with the experimental data and the average deviations for all the cases are found to be well within ±10%.     

Computer simulations of thermo-sensitive microgels: quantitative comparison with experimental swelling data

In this work, a quantitative comparison between experimental swelling data of thermo-sensitive microgels and computer simulation results obtained from a coarse-grained model of polyelectrolyte network and the primitive model of electrolyte is carried out. Polymer-polymer hydrophobic forces are considered in the model through a solvent-mediated interaction potential whose depth increases with temperature. The qualitative agreement between simulation and experiment is very good. In particular, our simulations predict a gradual shrinkage with temperature, which is actually observed for the microgels studied in this survey. In addition, the model can explain the swelling behavior for different contents of ionizable groups without requiring changes in the hydrophobic parameters. Our work also reveals that the abruptness of the shrinkage of charged gels is considerably conditioned by the number of monomeric units per chain. The swelling data are also analyzed with the Flory-Rhener theory, confirming some limitations of this classical formalism.     

A quantitative structure–mobility relationship of organic acids using solvation parameters

A quantitative structure–mobility relationship (QSMR) is proposed to estimate the electrophoretic mobility of diverse sets of analyses in capillary zone electrophoresis using Abraham solvation parameters of analyses, such as the excess molar refraction, polarizability, hydrogen bond acidity, basicity, and molar volume. QSMR was developed for prediction the electrophoretic mobility of 231 organic acids using the solvation parameters calculated by Abraham. Multiple linear regression (MLR) as a linear model and artificial neural network (ANN) methods were used to evaluate the nonlinear behavior of the involved parameters. The prediction results are obtained by nonlinear model, ANN, seem to be superior over MLR and were in good agreement with experimental data. In the proposed ANN–QSMR model, the overall mean percentage deviation values were 5.6, 5.4, and 5.3% and the coefficients of determinations (R²) were 0.84, 0.84, and 0.84 for training, test, and verification set, respectively. To investigate the robustness of the model, cross-validation methods have been established, i.e., leave-one-out and leave-N-out (N?=?5 and 10) and model is showed good predictive ability against data variation in cross-validation process. This model is not only able to accurately predict the migration order of a diverse set of organic acids but also model finds that solvation parameters are responsible in separation mechanism.     

The solubility modelling of non-polar gases in polar and non-polar solvent based on scaled particle theory by artificial neural network

In this work, the effective parameters of the scaled particle theory (SPT) are used as the input to the artificial neural network (ANN) to calculate as the output, the solubility (mole fraction of gas in liquid phase) of non-polar gases in polar and non-polar solvents at 298.15?K and 101.325?kPa. It has been found that ANN used in this work should has five neurons in the hidden layer to achieve the least error. The results of ANN have been compared with the experimental values. The results of this comparison are quite satisfactory. The average relative deviations of the simulations in training and testing stages have been calculated 0.92% and 0.89%, respectively. Finally, the results of ANN were compared with the results of SPT. According to this comparison, it is clear that SPT as a thermodynamic model predicts the solubility of the studied gases in the solvents with the same accuracy of ANN which is a purely mathematical model.     

Application of silver nanoparticles and principal component-artificial neural network models for simultaneous determination of levodopa and benserazide hydrochloride by a kinetic spectrophotometric method

A multicomponent analysis method based on principal component analysis-artificial neural network model (PC-ANN) is proposed for the simultaneous determination of levodopa (LD) and benserazide hydrochloride (BH). The method is based on the reaction of levodopa and benserazide hydrochloride with silver nitrate as an oxidizing agent in the presence of PVP and formation of silver nanoparticles. The reaction monitored at analytical wavelength 440 nm related to surface plasmon resonance band of silver nanoparticles. Differences in the kinetic behavior of the levodopa and benserazide hydrochloride were exploited by using principal component analysis, an artificial neural network (PC-ANN) to resolve concentration of analytes in their mixture. After reducing the number of kinetic data using principal component analysis, an artificial neural network consisting of three layers of nodes was trained by applying a back-propagation learning rule. The optimized ANN allows the simultaneous determination of analytes in mixtures with relative standard errors of prediction in the region of 4.5 and 6.3 for levodopa and benserazide hydrochloride respectively. The results show that this method is an efficient method for prediction of these analytes.     

Modification for CAAM approach for simulation of binary gas mixture separation in nanometric tubular membranes

In this study a modified solution for CAAM approach, originally presented in our previous work for simulation of binary gas mixture separation in nanometric tubular membranes, is presented by considering the permeate pressure drop along the permeator length. The accuracy of the modified CAAM method is verified by comparing the predicted results with the exact solution, experimental data and the original CAAM method results. Also, a detailed parametric analysis is done to determine the effects of different parameters (feed operational parameters and membrane physical and dimensional characteristics) on the permeate pressure drop along the permeator length.