Band 3, Heft 1

The multizone circulating reactor (MZCR) technology is a newly announced reactor design for solid-catalyzed gas-phase polymerization. The process is characterized by two hydrodynamically distinct but interconnecting polymerization zones, thus forming a continuous loop of polymer flow between the said zones. In the present study, a dynamic model for the MZCR reactor is developed to illustrate the basic dynamic behavior of the new reactor design; the model is used to study the copolymerization of ethylene with butene. Several parameter sensitivity analyses are performed to show the computer-simulated time responses for reactor temperature, number-average molecular weight, weight-average molecular weight, catalyst feed rate and the monomer concentration along the reactor length.

To investigate the factors affecting splashing in Peirce-Smith copper converters of Sarcheshmeh copper complex, South-East Iran, a laboratory scale model with dimensions one eighth of the actual converter was designed and assembled. The model was filled with water. Air was impinged onto the water. The sequence of events was subsequently recorded using high speed cameras. Factors affecting splashing such as blowing angle, volumetric flow rate of air, and the distance of the blowers from water surface were studied.Using a value of 2.65×10-11 for Morton dimensionless number, it was found that splashing may be reduced by 75% if the air speed in the tuyeres is reduced from 155 m/s to 30 m/s while the nozzle distance from the melt free surface is between 2 to 0.5 meters.

Expandable Polystyrene (EPS), a material of the insulation productions and packages, is commonly produced in a batch process. The control of the batch process is based on predefined process recipes and process parameters such as mixing properties. The EPS production has to be able to satisfy the aims and quality requirements of the markets, which causes additional demands on the process control. In this paper we demonstrate the optimization and modeling application for the EPS-batch process. The application consists of a production optimization tool and a simulation tool for the process parameters based on the Multi-Layer Perceptron network (MLP) with retrain properties. The software can be used at the operational level as well as at the process management level. The features are programmed into standalone software built up in the Matlab environment and it is tailored for the needs of the EPS production company StyroChem Ltd. The results show that the application can offer an efficient tool to economize production and storage costs.

The outputs of statistical process control (SPC) tools developed for fault detection are comparatively examined while applied to actual data collected in an industrial plant. The influence of added information gathered from the plant operation under different strategies is analyzed. Particularly, standard principal component analysis (PCA), kernel PCA and the Hotelling's T2 charts are inspected for a reported problem. The effect of training the tools either with an extended historic databank obtained under standard operation, or including also non-conventional conditions, is studied. The ability of the tools to provide a specific alarm and identify the responsible variable is examined by analyzing the contributions per variable to the SPE and the T2 statistics. In addition, the capacity of the tested tools to adapt to a new operation strategy is compared.

Several types of flow patterns in cyclones were found in recent publications. As a means for structuring the findings the publications were split up into two classes. Each class represented a certain feature of the axial velocity profile of the gas flow inside the cyclone. Class V represents cyclones with a maximum of the axial velocity at the vortex core of the cyclone, whereas Class W represents cyclones with an axial velocity profile that resembles an upside down W. This class has a local minimum of axial velocity at the vortex core or even displays backflow.The geometries of the cyclones belonging either to Class V or to Class W showed only small relative differences. It was assumed that the diameter of the vortex finder had a great influence on the class membership. This influence was simulated and measured for different basic geometries. The goal of this work was to develop the basics for reliable simulation results using the preprocessor GAMBIT and the finite volume CFD code FLUENT (which also serves as a postprocessor). Laser Doppler Anemometry (LDA) experiments were carried out on a laboratory scale cyclone to verify simulation results.In addition the results of cyclones which were investigated in previously published papers were used as a basis for developing a method of simulation, to validate numerical results and class membership. The prediction of pressure drop over the cyclone is overestimated by FLUENT (regarding the test cases). It was proven experimentally that downstream geometry (pipes etc.) has a great influence on flow profiles inside the cyclone. The axial flow profile can change dramatically if the pipes downstream the vortex finder are removed. A Class V cyclone displaying a maximum of axial velocity at the vortex core can change to a Class W cyclone with a minimum of axial velocity at the vortex core.

Purified lactic acid is used for production of biodegradable polymer. Esterification with methanol and subsequent hydrolysis in distillation column using cation exchange resin as catalyst to produce purified lactic acid has obvious advantages. In this paper, we present a theoretical study on optimization of industrial production (1000 t/year) of pure lactic acid using differential evolution (DE), a robust global optimization technique. The objective function is to minimize the overall cost with a view to obtain the optimized values of the decision variables: esterification reaction temperature, number of stages, feed location, reflux ratio, catalyst weight, number of reactive stages and position of reactive stages for hydrolysis.

This paper discusses the importance of phase equilibrium modeling in heterogeneous azeotropic distillation. Seven ternary systems are used to examine the behavior of phase equilibria and its representation, through activity coefficient models. This work used vapor-liquid equilibrium, liquid-liquid equilibrium, vapor-liquid-liquid equilibrium, azeotropic points and residue curve data to pick up a model and its parameters with the goal that the election correlate the systems in a way that it is useful for conceptual design of heterogeneous azeotropic distillation. The paper presents results done with UNIQUAC. A procedure to obtain binary parameters is proposed. The procedure is validated through an azeotropic distillation conceptual design and its following optimization using a rigorous simulation.

Thermally coupled distillation sequences can provide significant energy savings with respect to the operation of conventional distillation columns. Coupled sequences exhibit a complex structure, with recycle streams, that appear to affect their controllability properties. One potential solution to this problem has been suggested through the operation of complex schemes under conditions that do not provide minimum energy consumption. The basic idea is that if one changes the value of the interconnection flowrate, the control properties might change as well. In this work, we analyze the dynamic behavior of two coupled structures under different operating points, including the one with minimum energy consumption. The control analysis properties are analyzed with the application of the singular value decomposition technique in all frequency domains. The results show that the controllability properties of distillation sequences may change significantly depending on the selected value of interconnection flowrate.

The aim of this research is to simplify an existing model (model-1) of solar flux radiation containing many parameters as compared to a new mathematical model (model-2) using just two independent variables, time and roof's angle of inclination. The impact of the solar radiation on the surface of the roof in Sabah (state of Malaysia) has been examined under different orientations and slopes by means of model-1 at different time of the day. Furthermore, the data of solar flux radiation computed using model-1 has been explored to develop the model-2 from which the correlation coefficient of 0.87 was achieved using regression analysis. It has been proven statistically using the t-test that the difference between the existing model-1 and the proposed model-2 is not significant. The new mathematical model-2 developed can be applied for a double-pitched roof at any given orientation in the area of Sabah only, where the slope of the roof and the local time have to be taken from 0 to 75 degrees and from 7 hrs to 17 hrs, respectively. Additionally, the method used in the derivation of this new model can also be employed for other states of Malaysia as well as other countries. From a technical point of view, the architects or designers can use the proposed model for a quick estimation of solar flux radiation, to identify the type of insulations to be applied under the roof or above the ceiling, as well as the size area of the natural ventilation and the decision-making on the thermal system to be employed, either as an active or a passive cooling system.

A 3-phase, one dimensional, heat transfer model for the fluid catalytic cracking riser reactor is developed. This model incorporates a new detailed kinetic model based on pseudo components' cracking. The model also includes the effect of local temperature on the reaction rate constants, variation of endothermic heat of cracking all along the riser height, and the heat of adsorption of the coke on the catalyst. Also, the developed model is compared with a 2-phase, one dimensional heat transfer model.The proposed model is capable of predicting the products' yields, phase temperatures, and phase velocities all along the riser height. Model predictions are very close to the data reported in literature.

Bodymaker lubricants are oil-in-water emulsions used as coolants in the ironing process during production of aluminium cans. Lubricants are formulated to diminish friction during ironing and to cool down the mechanical contact interface. In the present work, a methodology based on algebraic modelling of the process is proposed to select among bodymaker lubricants, featuring their tribological performance. Experimental results are obtained by "pin-and-vee" test protocol, mathematical model exhibits good agreement with experimental observations. In general, tests carried out indicate that different bodymaker lubricants exhibit different wear rates, therefore evaluation of this rate can be used as tool to discriminate among lubricants performance. At present, this approach is being successfully used by a lubricant making company to evaluate and compare the performance of commercial bodymaker lubricants.

Calculation of the pressure and flow profiles of a simulation has a major effect on the fidelity, reliability, robustness and performance of the dynamic simulator. A pressure-flow (P-F) network consists of several unit operations, connected by streams, where pressure and flow relations must be calculated. The resistance and volume balance equations produce these P-F relations within the flowsheet.This study compared two different solution methods for solving the resultant nonlinear simultaneous P-F equations, namely the Referred Derivatives method introduced by Thomas in 1997 and the Newton-Raphson method. In this comparison the advantages and disadvantages of the Referred Derivative method are provided. This paper also discusses the use of the Referred Derivatives method in solving flowsheets that include unit operations with holdup.

The objective of the research presented in this paper is to investigate the effectiveness of a new stochastic algorithm to obtain maximum hydrocarbon production rate in petroleum reservoirs. The hydrocarbon production rate is a function of several variables. Some of these variables are production well location, production well flow pressure, injection well location and fluid injection rate. Changing any of the variables will change the hydrocarbon production rate value. By using stochastic optimization techniques, there is no limit to the number of decision variables that can be optimized simultaneously. The presented method in this paper is a new metaheuristic algorithm which enables us to analyze a system of mathematical equations containing a large number of decision variables and determine the optimum values of them that should give the most economic result. This is referred to as the Ant Colony Optimization Algorithm. Ant Colony Optimization technique leads us to lower computational cost to optimize complicated problems too. Although applied to discrete domains, this algorithm with some modifications has been applied to continuous optimization. Here, a numerical simulation is initially applied for modeling the hydrocarbon reservoir and then the results are analyzed and furthermore optimized to enhance profitability using Ant Colony Optimization algorithm. The developed code in MATLAB environment, based on Ant Colony Optimization, is able to estimate production and injection wells location, fluid injection rate and well flow pressure which would result in improved hydrocarbon production. This algorithm is simple to implement and the results of case studies show its ability to provide fast and accurate solutions. Results prove the reliability of algorithm for solving a large class of optimization problems in petroleum engineering.

Nanofiltration (NF) is considered to be an intermediate process between reverse osmosis (RO) and ultrafiltration (UF). In the present work, separation data obtained for nickel sulfate-water and nickel chloride-water systems, using an N, O-carboxymethyl chitosan (NOCC)-polyether sulfone (PES) composite NF membrane, are analyzed with the help of two nonlinear membrane transport models, viz., combined film theory-solution-diffusion (CFSD) model and combined film theory-Spiegler-Kedem (CFSK) model. The membrane transport parameters and mass transfer coefficient (k) are estimated, simultaneously, from both the models using the Levenberg-Marquardt method. The k values estimated from two models show a variation that may be attributed to the presence of reflection coefficient in the Spiegler-Kedem model. Though both the models predict the membrane performance reasonably, the values of CFSK model are more accurate in the present case. Hence, it can be suggested that CFSK model may be the better method for estimating the membrane transport parameters and k. Various parameters estimated from the above models are used to find the enrichment factors, concentration polarization modulus and Peclet numbers for both the systems.

The aim of this work is to develop and compare kinetic models for the oxidative coupling of methane (OCM) based on the gas hourly space velocity (GHSV) value and CH4/O2 ratio in two scales: laboratory and bench. The experiments were carried out in tubular fixed bed reactors at 1023 K, using 0.7-1.5 g and 30 g of perovskite titanate as the reaction catalyst for laboratory and bench scales, respectively. The various GHSVs (8000, 12000, 17000 h-1) and (3400, 4300, 5200 h-1) and methane to oxygen ratios (1, 2, 3, 4, 7.5) and (2, 2.5, 3) were selected for laboratory and bench scales, respectively. We have proposed a mechanism in which the consumption rate of methane is always twice of production rate of C2. A power law model was assumed for rate of reaction in terms of partial pressure of oxygen and methane. Using a linear regression analysis, the kinetic models were determined. Comparison of the calculated rate of reaction with the experimentally measured data confirmed the accuracy and applicability of the developed model for both scales.

Numerous unit operations in chemical and process engineering can be represented as a system of non-linear algebraic equations, when modeled for steady state operation ,e.g. isothermal and non-isothermal operations of a series of CSTRs, batteries of evaporators, networks of various separation operations (flash drum, mixers), distillation, extraction and absorption columns, and pumps and piping networks etc. These governing equations are sometimes very difficult to solve due to the nonlinear and uneven nature associated with them. The difficulty level increases when the resulting set of equations become flat near their zeros and thus derivatives based schemes mostly diverge or give poor results. Many a times, even a good initial guess in conventional numerical techniques do not guarantee to have a true solution and problem specific methods have to be designed. This whole scenario can also be viewed as an optimization problem having equality constraints only and casting the equations in the form of norm of function vector, which formulates an objective function to be minimized. The true minimum thus found gives us the correct solution vector. Recently, Genetic Algorithms have been quite effectively used to solve many complex engineering optimization problems. In continuation of our earlier research work where an elitist genetic algorithm was developed for the solutions of various difficult MINLP problems (Danish et al. 2006a and b), this research work extends its application for the solution of difficult non-linear algebraic equations. A novel scheme of dynamic mutation parameter as a function of fitness along with dynamic penalty has been proposed. The small value of mutation parameter in initial stages enables the algorithm to search globally, and the solution thus found is refined by keeping its value higher in later generations. This new scheme is found to be very effective in the sense that the algorithm requires very small population size and comparatively lesser number of generations to give reasonably good solutions. To test the efficacy of algorithm we have solved five sets of difficult non-linear algebraic equations (Dennis and Schnabel, 1983). It is worthwhile to mention that one of these equations was having both its Jacobian and Hessian as zero, at its true solution. Applicability of the developed GA was also demonstrated by simulating an industrial case study of triple effect evaporator used for concentrating the caustic soda solution (Zain and Kumar, 1996), which also poses difficulty during numerical simulation by Newton-Raphson method.

Flows in highly heterogeneous porous media are found in many practical fields, such as hydrology, petroleum engineering, and chemical engineering. The case of two-region heterogeneous media (fractured media, catalytic beds, etc.) plays a fundamental role. The different questions associated with this specific case are illustrated in this paper for two different kinds of transport: (i) flow of a slightly compressible fluid, (ii) dispersion of a tracer. Many different theoretical models are implemented using COMSOL Multiphysics™. These models correspond to direct simulation and macro-scale or large-scale models such as fully averaged models or mixed models.

This paper discusses the design and application of a filter based on an Artificial Neural Network (ANN) in a chemical engineering process. The design of a filter consists of adapting the algorithms that make up the filter to the process to be filtered. Taking into account that the ANN is able to model almost every type of chemical process, the design and application of a filter based on ANN was studied. In this work, every ANN used was based on Multilayer Perceptron (MLP). Bearing in mind that ANN should reproduce the process as accurately as possible, an optimisation of the ANN (training function and parameters) was carried out. A mathematical model of a reflux in the upper part of a distillation column was used to test the ANN filter. The ANN is able to filter noisy signals with a mean prediction error less than 2.5•10-3 %.

Today, steam engines are used for special purposes only, for example to reduce steam pressure in pressure reduction stations, where they replace the traditional and inefficient throttling process. Throttling is the most used way to control the pressure in steam reduction stations. This way is unsatisfactory from the economical point of view, because the exergy is lost uselessly. It is a part of heat energy that can perform a work. The better way of a pressure reduction is an expansion in a backpressure turbine or in a steam engine by simultaneous transformation of the heat energy into electricity (cogeneration). This article describes the design and implementation of the mathematical model of a steam engine used as pressure regulator in a pressure reduction station. The present model is a part of a comprehensive mathematical model of a cogeneration unit and also a part of the author's doctoral thesis. The model assumes detailed mathematical description of physical processes in a steam engine and implementation in an MATLAB-SIMULINK software environment.

Flowshop scheduling via mixed integer linear programming (MILP) has received considerable attention in the past four decades. However, alternate models are limited; most numerical studies have used small problem sizes. A need for good model evaluation methodology exists; and limited work exists on flowshops with no intermediate storage. This paper presents a classification of flowshops and MILP scheduling models, and addresses some of these issues. It develops a host of new MILP formulations for minimizing makespan in a permutation flowshop with no storage and with or without unit setups. It presents some useful insights into model building by employing a variety of new and old binary variables and coupling them creatively. In contrast to previous work, it evaluates a range of new and existing MILP models using many larger test problems with no or unlimited intermediate storage, and presents a reliable procedure to rank various models based on problems with varying data and sizes. It shows that the top models for the two flowshops indeed show slightly different computational performance.

Computational heat transfer flow modeling is one of the great challenges in the classical sciences. As with most problems in Engineering, the interest in the heat transfer augmentation is due to its extreme importance in various industrial applications. CFD simulation for the heat transfer augmentation in a circular tube fitted with right-left helical tape inserts in laminar flow conditions has been explained in this paper using Fluent version 6.2.16. The data obtained by simulation are matching with the literature value for plain tube with the discrepancy of less than plus or minus 5% for Nusselt number and friction factor. Enhanced heat transfer with decreasing twist ratio was observed. Heat flux is more uniform all along the tube and decreases uniformly towards the center. Right-left movement of the flow as per direction twist was also observed from the velocity contour.

This paper presents a convective dominated reliable diffusion process in an axi-symmetric tube with a local constriction simulating a stenos artery considering the porosity effects. The investigations demonstrate the effects of wall shear stress and recirculation flow on the concentration distribution in the vessels lumen and on wall mass transfer keeping the porosity in view. The flow is governed by the incompressible Navier-Stokes equations for Newtonian fluid in porous medium. The convection diffusion equation has been used for the mass transport. The effect of porosity is examined on the velocity field and wall stress. The numerical solutions of the flow equations and the coupled mass transport equations have been obtained using a finite difference method. This paper explains the reliable effects of flow porosity on the mass transport.

This paper presents a comparative closed-loop performance study on a multicomponent batch distillation column that produces high-purity distillate product. Initially, the distillation model representing a ternary system of cyclohexane-toluene-chlorobenzene has been simulated to observe the open-loop dynamics at the start-up as well as production phase. The next three control schemes, namely classical proportional integral (PI) controller, nonlinear proportional integral (NLPI) control strategy and gain-scheduled proportional integral (GSPI) law, have been designed and applied on the simulated distillation column. Numerical experiments based on servo and regulatory tests show that the GSPI algorithm provides relatively better results than the other two PIs.

Process systems engineering is rapidly moving from steady state simulation towards operator training simulation, based on dynamic models, automated procedures, and predictive systems. Conventional operating conditions are well-known and easily controlled by field operators as well as control-room operators, whereas other situations are not yet. This is the case of more unusual circumstances, such as plant start-ups as well as planned and emergency shutdowns.The implementation of detailed dynamic mathematical models allows simulating the behavior of single process units as well as industrial plants, with the possibility to study unusual scenarios, especially for two reasons. First of all, it is necessary to train the operator and prepare him/her to face several events. Secondly, there is the need for evaluating the best automatic procedure to manage either unconventional or critical situations, without waiting for them to occur. In this sense, the safety approach is changing from reacting to predicting, thanks to the spread of tools that offer the possibility to simulate accidental events, unit and/or valve malfunctions, and significant process transients.The paper investigates and compares different procedures to control process transients, described by a sequence of actions, in order to coordinate synergistically the actions dictated by control-room operators and the operations to be accomplished manually by field operators. The main objective is to improve process start-up and shutdown reliability as well as the corresponding plant safety of an existing plant-wide control structure, without requiring any structural modification. A propane/butane splitter is adopted as a case study.

The Artificial Neural Network (ANN) has been used as a predictive tool for the estimation of certain parameters of Subernarekha, an important river in the Jharkhand state in India. The network used two algorithms for this purpose and was sufficiently accurate in predicting the most economically heavy and time-consuming set of data. The Levenberg-Marquardt Backpropagation Network (trainlm) and the Resilient Backpropagation Network (trainrp) were the two algorithms used for the estimation of metallic species and physicochemical parameters of the river. The MAPE for metallic species were found to be 0.71 for cadmium, 0.182 for copper and 0.771 for chromium, while physicochemical parameters were 16.645 for alkalinity, 5.883 for dissolved oxygen (DO) and 23.28 for chemical oxygen demand (COD). Both the algorithms were used with different sets of hidden layers i.e., one for trainlm with 5 neurons and three for trainrp with 5, 4 and 5 neurons, and these were determined using the trial and error method. This method is not only economically favorable but time-adaptive as well, as it depends on the amount and span of data available and it can predict values to a very high degree of accuracy. This method can successfully be employed for the prediction of parameters of any river system with confidence.

Dead streptomyces rimosus was found to be an effective biosorbent for the removal of chromium from industrial tanning effluents. A sorption level of 65 mg/g was observed at pH 4.8 while the precipitation effect augmented this value at a higher pH range. Chromium desorption increased with decreasing desorption agents pH (including HCl and H2SO4) to a maximum value of 95% at approximately zero pH. The biosorption data of trivalent chromium by streptomyces rimosus has been used for kinetic studies based on fractional power, Elovich, pseudo-first order and pseudo-second order rate expressions. The time-dependent Cr (III) biosorption data were well-described by a pseudo-second-order kinetic model. The intraparticle diffusion is not the rate-limiting step for the whole reaction. It was found that the biosorption equilibrium data fit well with the Langmuir model.

This article presents the design of an exponential state estimator for a distillation column. The proposed algorithm mainly estimates the imprecisely known parameters based on the available measurements. For the example distillation column, the nonlinear state estimator is formed with two component material balance equations; one is around the condenser-reflux drum system and the other one is around the reboiler-column base system. As a result, there is an excessive process/model mismatch. In this situation, the proposed scheme estimates the augmented states with sufficient accuracy. The simulation experiments have been performed to investigate the convergence ability of the estimator. The less computational requirements and simple design make the estimator attractive for online use.

Structure-activity relationships were studied for a series of 46 2.6-dimethyl-3.5-dicabomethoxy-4-phenyl-1.4-dihydropyridine derivatives by means of multiple linear regression (MLR) and artificial neural network (ANN) techniques. The values of log (1/EC50), which represents the 50% effective concentration for blocking the Ca2+ channel of the studied compounds were correlated with the descriptors encoding the chemical structures. Using the pertinent descriptors revealed by the regression analysis, a correlation coefficient of 0.99 (s = 0.23) for the training set (n = 46) was obtained for the ANN using the Levenberg-Marquardt algorithm with a 3-10-1 configuration. The results obtained from this study indicate that the activity of 2.6-dimethyl-3.5-dicabomethoxy-4-phenyl-1.4-dihydropyridine derivatives is strongly dependent on molar refractivity (MR), electronic factors (especially on the connectivity indices (IC0)) and hydrogen-bond donor's (HBD) of the molecule. Comparison of the descriptor's contribution obtained with MLR and ANN models indicates the presence of non-linearity in the data and the interaction effect between them since the efficiency of these descriptors was increased by the ANN model.On the other hand, we have used a new, robust structure-activity mapping technique, a Bayesian-regularized neural network, to develop a quantitative structure-activity relationship (QSAR) model, and the ability of the model was tested by using the cross validation technique. The results show that the method is robust and reliable and gives good results. Comparisons of Bayesian neural net models with those derived by classical MLR model analysis showed its superiority in generalization.

In this work, a model for particle formation and growth is used to simulate the aluminum particle synthesis by an evaporation-condensation process. The effect of commonly employed process parameters (vaporization temperature, cooling rate, system pressure) is investigated. The model to be solved is the general dynamic equation (GDE) that accounts for particle nucleation, condensation and coagulation at non-isothermal conditions. The GDE was solved using the nodal method. The methodology approximates the particle size distribution to a few nodes by introducing size-splitting operators. The simulation results show that particle formation and growth take place in a short temperature range. The coagulation increases the particle size while maintaining the number concentration. On the other hand, the surface condensation tends to shift the particles size distribution towards a larger size. It was shown that production of nanoparticles with a more uniform size distribution and smaller particles are favored using low pressure and low vaporization temperature. The particle size distribution is nearly influenced by the cooling rate in the range study here.

A continuous intensified reactor requires precise information to adapt the design to the reaction to be implemented. In this article we present a two step approach to pre-designing such devices. The first step consists of the optimisation of the operating conditions based only on stoichio-chemical schemes and kinetic laws. This is carried out by adjusting the temperature profile and feeding rate strategy in a batch or pseudo-batch operation. The second step departs from the ideal thermo-kinetic path by taking into account heat removal and reagent injection. A simulator is used to evaluate the impact of these constraints on the selectivity and productivity. The approach is applied to the linear alkylbenzene sulfonation which characteristics present most of the challenges encountered in the intensification. Final results give a pre-design in terms of reactor characteristics (size, injections positions) and operating conditions.

The objective of this work is to introduce and demonstrate the technical feasibility of the continuous flash fermentation for the production of butanol. The evaluation was carried out through mathematical modeling and computer simulation which is a good approach in such a process development stage. The process consists of three interconnected units, as follows: the fermentor, the cell retention system (tangential microfiltration) and the vacuum flash vessel (responsible for the continuous recovery of butanol from the broth). The efficiency of this process was experimentally validated for the ethanol fermentation, whose main results are also shown. With the proposed design the concentration of butanol in the fermentor was lowered from 11.3 to 7.8 g/l, which represented a significant reduction in the inhibitory effect. As a result, the final concentration of butanol was 28.2 g/l for a broth with 140 g/l of glucose. Solvents productivity and yield were, respectively, 11.7 g/l.h and 33.5 % for a sugar conversion of 95.6 %. Positive aspects about the flash fermentation process are the solvents productivity, the use of concentrated sugar solution and the final butanol concentration. The last two features can be responsible for a meaningful reduction in the distillation costs and result in environmental benefits due to lower quantities of wastewater generated by the process.

Knowledge of the heat transfer coefficient is an essential prerequisite for any industrial gas/liquid bubble column reactor design. In this paper, experiments were carried out in a laboratory scale bubble column using air–water and 10wt% glycerin–air solution systems. Experimental values of the heat transfer coefficient were compared with some of the previous published correlations and were also compared with the model predicted by ANN in our previous work (Ahmedzeki, 2007). The ANN model was found to be better than all previous published correlations. %AARE was 8.2 and %standard deviation was 7.85. A mechanistic model was also developed to analyze the system. The resulting correlation was a %AARE of 17.5 and %standard deviation of 15. Both the above models and the experimental results compared well with the previous published correlations.

In this study, the (catalysed) esterification process of acetic acid with ethanol producing ethyl acetate and water is modelled using detailed mass and energy balances and thermodynamic properties within gPROMS modelling software (version 3.0.3, 2004). The basic model was taken from Mujtaba and Macchietto (1997) assuming no azeotrope formation in the system. Kinetic (hydrochloric acid as a catalyst at 100 °C) and vapour-liquid equilibrium models are taken from Bogacki et al. (1989) and Suzuki et al. (1970) respectively. Physical properties are calculated using the Ideal Physical Properties Foreign Objective (IPPFO) package interfaced to gPROMS. The performance of the batch reactive distillation is evaluated in terms of the maximum conversion of ethanol to ethyl acetate. Five cases with varying amounts of reactants (including the cases with the reduced amount of water in the feed and keeping the amount of acetic acid and ethanol fixed) are utilized to improve the conversion of ethanol to ethyl acetate. Both piecewise constant and linear reflux ratio profiles (single time interval) are considered as a control variable for the base case while the piecewise constant single reflux ratio strategy is adopted for other optimized cases. The results show that both the maximum conversion and distillate product (base case) are improved using the linear reflux strategy as a control variable compared to those with a constant reflux ratio profile. Increasing the amount of water in the feed leads to a reduction in the conversion and distillate product; moreover, the column was operated at a higher reflux ratio compared to the case without any water in the feed.

A novel macrokinetic model is proposed for the hydrogenation purifying of terephthalic acid (HPTA), which is catalyzed by Pd/C in an industrial trickle-bed reactor (TBR). First, the reaction orders of HPTA are estimated by the mass transfer-free experiment data in the laboratory stirred batch reactor (SBR). The hydrogenation reaction kinetics is assumed to be first-order with respect to 4-carboxybenzaldehyde (4-CBA), which is the most important contamination of terephthalic acid and hydrogenated to produce p-toluic acid. Second, due to the fact that the reaction factors of HPTA have a high-nonlinear and complex effect on the reaction rate constant of HPTA, the nonlinear partial least squares regression method was used to model the influence of the reaction factors on the rate constant based on the transfer-free experiment data of HPTA in the laboratory SBR. Thus, the macrokinetic model of HPTA in the laboratory SBR is developed. Third, to indicate the difference between HPTA in the industrial TBR and that in the laboratory SBR, the correction coefficients were proposed and introduced into the obtained rate constant model to describe the influence of the reaction factors on HPTA in the industrial TBR. Based on the macrokinetic model with the unknown correction coefficients and the data obtained in the industrial TBR, differential evolution (DE) algorithm was used to obtain the optimum correction coefficients according to the fitting performance of the macrokinetic model, and the macrokinetic model for HPTA in the industrial TBR was developed. Further, the reliability of the model was investigated and satisfactory results were obtained.

In this study, several local composition models and stochastic optimization methods have been used and compared in data fitting of activity coefficients in aqueous electrolytes. We have utilized the electrolyte-NRTL model of Chen et al. and the modified Wilson models proposed by Xu and Macedo, and Zhao et al. to fit the activity coefficients of several quaternary ammonium salts in water at 25°C. These electrolytes have interesting properties for their application as ionic liquids. However, the modeling of their thermodynamic behavior using local composition models is a global optimization problem. In this study, several stochastic optimization methods have been used to solve this optimization problem and their numerical performances have been compared. Specifically, we have tested the classical Simulated Annealing and the hybrid methods: Direct Search Simulated Annealing, Simplex Coding Genetic Algorithm, Simulated Annealing Heuristic Pattern Search, and Directed Tabu Search. Our results show that Simulated Annealing is a suitable tool for data fitting of the activity coefficients of aqueous electrolytes. Finally, the tested models can satisfactorily correlate the mean activity coefficients of electrolytes treated in this study, and are suitable for process design.

Carbon-11 radiopharmaceuticals are gaining an increasing importance in positron emission tomography due to their importance in diagnostic medicine. The most wide spread method of production of these radiopharmaceuticals is by methylation of an appropriate precursor with the highly reactive [11C]methyl iodide. Conventional synthesis of this intermediate involves liquid phase synthesis of [11C]methanol, which is the step that limits the specific activity of the final product. To avoid the loss of specific activity, a catalytic gas phase methanol synthesis process was evaluated. In this procedure, [11C]carbon monoxide would be reduced to [11C]methanol using a copper zinc oxide catalyst in the presence of hydrogen.In this study, a reactor to catalytically convert 50 ppm carbon monoxide to methanol was developed. A copper zinc oxide catalyst was prepared by a co-precipitation method. The catalyst was activated by reduction with hydrogen and passivated with oxygen prior to methanol synthesis. The effects of temperature, pressure and flowrate on the conversion of carbon monoxide to methanol were studied. The experimental results were used in conjunction with a commercially available process simulator to fit a kinetic model for methanol synthesis from carbon monoxide. This model was used to determine optimal operating conditions for this reactor and predicts a 60% conversion of [11C]carbon monoxide to [11C]methanol. These findings suggest that gas phase [11C]methanol synthesis is a viable alternative to the conventional liquid phase method, meriting further studies with carbon-11.

This work is a simulation of a countercurrent moving bed reactor in which a reduction of porous iron ore pellets to sponge iron is performed. Simultaneous mass and energy balances of the reactor will lead to a set of ordinary differential equations. The iron ore reduction kinetics was modeled with a three-stage shrinking core model. Solving the ODE system allows prediction of the concentration and temperature profiles of all species within the reactor.The model is validated with experimental data obtained during operation of the bench-scale and full-scale processes and allows optimization of the process. The model is used to explore the behavior of the reactor when residence time of pellets is changed in order to increase or decrease the solid conversion. Also, it is used to explore the performance of the reactor under different conditions.

As a successive work to clarify the film growth kinetics of GaAs, a new lumped model for computer simulation of film growth during MOCVD process is proposed. The film growth process in a horizontal reactor using trimethylgallium (TMGa) and tertiarybutylarsine (TBAs) is considered. Analyzing methodologies at two different scales have been used to revise our previous simulation model. One is the micro-scale analysis using a selective-area growth method and the other is the reactor-scale macro analysis using numerical simulation. Species concentration profiles are obtained from the macro-scale computational fluid dynamics study incorporated with chemical reactions. Surface reaction rate constants are extracted from the selective-area growth experiments. From the multi-scale analysis that combined the two-scale analyses, surface formation from the source gas, TMGa, is included to the previous model. The kinetic data for the surface reaction of TMGa are deduced from the quantitative analysis for the species concentrations from both of the analyses. Simulation results using the new model show that monomethylgallium is the major source of film deposition, but deposition from TMGa is important in the upstream region of the reactor.

The development of alternative technologies for the removal of gas pollutants is an important aspect for the environmental friendliness of energy production. During coal gasification, N2 contained in coal is converted to NH3 and, as much as 50% of the ammonia in the fuel gas can be converted to nitrogen oxides (??x). At these conditions, decomposition seems to be the only applicable solution for the removal of NH3. The application of a high temperature catalytic membrane reactor process appears to offer an efficient and cost effective method of removing the NH3 from coal gasification gas streams.The present work examines the operation of such a selective membrane, used for the decomposition of NH3, under a 2-D axissymetric CFD approach where the flow field, the chemical reactions and the selective porous membrane behavior are being modeled and computed. The main target of this effort was to obtain a more detailed view of the flow field and to investigate the decomposition of ammonia in comparison with a simpler 1-D modeling approach and, thus, to evaluate the advantages and disadvantages of each method.

In order to improve air quality, catalytic converters have been extensively employed throughout the world. These monolithic catalytic converters are very effective in reducing pollution only after they have heated up. The catalyst warming-up process takes some time and during this period a considerable amount of pollutants, especially hydrocarbons, is released into the atmosphere. In this paper, a one-dimensional model for the hydrocarbon propylene was developed, taking into account heat and mass transfer between the exhaust gas and the catalyst surface and the catalytic reaction. The equations so formed are a set of ordinary differential equations (ODEs) and a partial differential equation (PDE). The ODEs are solved using the Runge-Kutta method of fourth order and the PDE by backward implicit scheme. The effect of varying the inlet gas temperature and initial catalyst solid temperatures on conversion of propylene was analysed.

An efficient Monte Carlo (MC) algorithm and a two-dimensional fixed pivot technique (2-D FPT) are described for the calculation of the average and distributed molecular properties of polymers in batch free-radical polymerization and copolymerization reactors. Simulations are carried out, under different reactor conditions, to calculate the individual monomer conversions, the leading moments of the 'live' and 'dead' polymer chain length distributions as well as the dynamic evolution of the distributed molecular properties (i.e., molecular weight distribution (MWD), long chain branching distribution (LCBD), copolymer composition distribution (CCD), joint LCB-MW and CC-MW distributions, etc.). The validity of the numerical calculations is examined via a direct comparison of the simulation results, obtained by the two numerical methods, with experimental data on the styrene-methyl methacrylate and vinyl acetate batch free-radical polymerization systems. Additional comparisons between the MC and the 2-D FPT methods are carried out for both polymerization systems, under different polymerization conditions. It is clearly shown that both numerical methods are capable of predicting the distributed molecular, branching and copolymer properties, with high accuracy, up to very high monomer conversions.

Low frequency and large scale flow variations, often called macro-instability (MI) phenomena, were studied by means of a combination of large eddy simulations (LES) and sliding mesh (SM) models. Numerical predictions of MI characterictic in a six-bladed Rushton turbine stirred tank were performed by varying the off-bottom clearance at constant impeller rotational speed. The occurrence of MI in this study was identified by two methods: observing visually the flow velocity vector field and analyzing the time-series data of both velocity in the bulk flow and dynamic pressure on the vessel wall. The transient flow field visualization revealed the secondary circulation flow and the asymmetrical flow pattern beside of the mean flow pattern. The high variation intensity of flow pattern variation was clearly identified in the region with smaller space. The flow pattern variation was corroborated with the presence of high amplitude peaks in the low frequency part of the frequency spectrum of flow velocity. The distinct peak that was usually designated as the frequency of MI revealed the characteristic of MI. The physical meaning of the frequency of MI was the periodical appearance of the pronounced flow pattern. The numerical predictions of the frequency of MI in the region below and above impeller studied in this paper were confirmed well with the experimental results reported by Matsuda et al. (2004). The numerical prediction of the dynamic pressure monitored on the tank wall was also in agreement with the flow velocity monitored in the bulk flow for indicating the MI phenomena. This way of monitoring will be useful for the study of MI in multiphase agitated tank in the near future and in the practical application.

Retrograde vaporization calculation is a hard problem, which possesses several solutions and demands a robust algorithm. In the present work, we solved the double retrograde vaporization problem by hybrid global-local optimization. In fact, we propose a new methodology that involves fuzzy clustering means together with Luus-Jaakola and Nelder-Mead algorithms. We applied the proposed methodology to solve the double retrograde vaporization problem for binary systems of methane + n-butane. For this mixture, the dew point curve exhibits an S-shape at temperatures slightly below the critical temperature of the more volatile component. This binary mixture was modeled using the original Peng-Robinson equation of state and the classical one-fluid van der Waals mixing rule.

In this work, the dynamic behavior of two chemical reactors (Fixed Bed Catalytic Reactor and the Three-Phase Reactor), in which different reactions of great industrial interest take place, is presented. Rigorous and reduced models were developed and their capabilities to adequately predict the dynamic behavior of these reactors were compared, indicating which model is more suitable for a specific application (design, optimization and control). Different approaches [Classic, Hermite, Finlayson, Pirkle, Dixion, Generic and Orthogonal collocation (one point)] were used to obtain the reduced models. With these techniques, a mathematical order reduction was carried out, which eliminates the spatial co-ordinate (of the catalyst particle or of the reactor) and promotes radially lumped-differential formulations. The proposed rigorous models were able to predict the main characteristics of the dynamic behavior of the fixed bed catalytic reactor as well as of the three-phase reactor, including the inverse response phenomena and the hot spot present in the former. This knowledge is essential to design and control these reactors. The computational time demanded for the solution of the rigorous models is high in comparison to the reduced ones, which restricts the use of the rigorous models to cases where time is not a limiting factor. Otherwise, when on-line applications are required, the reduced models are more adequate. The models based on reduction techniques overcame computational burden with a faster and easier numerical solution, as well as other difficulties found in rigorous heterogeneous models, especially related to the large number of parameters and sophisticated numerical procedures required by the solution.

The time to measure and maintain distillate quality in the presence of disturbances is the main objective of the distillation process. To maintain the product quality, a suitable tray temperature is kept constant with the help of a control strategy near its set point. As the controlled variables cannot be measured online easily due to measurement lags and sampling delays, the secondary measurement techniques are used to estimate the distillate quality by adjusting the values of manipulated variables (inferential control). This paper presents an inferential control scheme for a distillation column using an Ethernet-based data acquisition system to provide a standard online estimator. In the present scheme the distillation parameters are acquired from a client PC connected to the network and the distillate composition is estimated online by a developed ANN-based estimator. The above estimator is used for an inferential control strategy to control distillate composition. The inferential control scheme is exercised for various disturbances and the effect of the disturbances is observed on distillate quality.

In the near future of chemical industry, communication between design, manufacturing, marketing and management should be centered on modeling and simulation, which could integrate the whole product and process development chains, process units and subdivisions of the company. Solutions to this topic often set aside one or more components from product, process and control models, hence, as a novel know-how, an information system methodology was developed. Its structure integrates models of these components with a process data warehouse where integration includes information, location, application and time integrity. It supports complex engineering tasks related to analysis of system performance, process optimization, operator training systems (OTS), decision support systems (DSS), reverse engineering or software sensors (soft-sensors). The case study in this article presents the application of the proposed methodology for product quality soft-sensor application by on-line melt index prediction of an operating polymerization technology.

The heat transfer cooling of a hot liquid in a stirred vessel has been studied experimentally with coolant flowing through a half-coil around the vessel. Correlations have been developed for the heat transfer coefficient of the half coil jacket. A mathematical model for the half coil jacket liquid temperature dynamics and its analytical solution is used to find the shell side temperature profile as a function of time. It is found that the model predictions are in satisfactory agreement with the experimental data and that the developed correlation is superior to previously published correlations for similar systems.

This work addresses the optimum design of a phosphoric acid (PA) production process under uncertainty, both from an economic and environmental perspective. A design and optimization framework is employed that utilizes generic process modules to enable systematic process flowsheet development and to facilitate interconnectivity among the utilized processing components. This is coupled with a systematic method of incorporating uncertainty into the process design and optimization, hence, enabling a realistic representation of industrial conditions during the PA process design stages. The integrated design of PA production processes and downstream phosphogypsum utilization processes is further proposed, in the form of the Ammonium Sulfate (AS) production process, as a means of mitigating the environmental impact of the PA production process associated with phosphogypsum. The case studies, presented to illustrate the proposed developments, highlight the significant impact of considering uncertainties in the design of optimum PA process schemes as well as generate useful insights regarding the synergies developed in the integrated PA and AS process design.

The processes that take place in industrial reactors are mainly highly nonlinear and, due to this, their controls have become a major factor in determining the aspect of process safety, quality, and productivity of these systems. For this reason, we have designed an inexpensive, practical reactor on a pilot plant scale to conduct such online control studies where the heat of reaction in the reactor is simulated by injecting steam into the reactor system. This paper highlights the development of the pilot plant with its software development to implement advanced control strategies using artificial intelligence approaches such as fuzzy logic and genetic algorithm. The online implementation results conclude that the Genetic Algorithm Model Based Controller (GAMBC) gave similar performance with the Fuzzy Logic Controller (FLC) for the set point tracking studies but in the load disturbance rejection studies, it was found that the FLC performed better. The results obtained show the usefulness and importance of these experimental studies to further understand advanced control applications for a typical industrial chemical process such as a reactor.

This paper describes a robust hybrid artificial neural network (ANN) methodology, which can offer superior performance for important process engineering problems. The method incorporates hybrid artificial neural network and differential evolution technique (ANN-DE) for efficient tuning of ANN meta parameters. The algorithm has been applied for the prediction of pressure drop of solid liquid slurry flow. A comparison with selected correlations in the literature showed that the developed ANN correlation noticeably improved the prediction of pressure drops over a wide range of operating conditions, physical properties, and pipe diameters.

Process design and simulation rely heavily on the accuracy and availability of transport property correlations. General models that combine the properties of pure components often lack the necessary accuracy. In this investigation, neural networks were used to model some important transport properties for the ethanol-water binary system. Specifically, a three-layer feed-forward neural network with six neurons in the hidden layer was used to model viscosity, thermal conductivity, surface tension and the Fick diffusion coefficient based on an array of experimental data. These neural network models were then compared to some conventional models that are commonly used to predict the aforementioned transport properties. The results showed that the neural network models were able to represent the experimental data very well for the system studied. One advantage in using neural network models to represent these properties is their ability to predict complex and interrelated behaviors without a priori information about the underlying model structure. Further, since all the models retain the same simple matrix structure, their integration into computer codes becomes straightforward and non-repetitive.

This paper presents artificial intelligence-based process modeling and optimization strategies, namely, support vector regression – differential evolution (SVR-DE) for modeling and optimization of catalytic industrial ethylene oxide (EO) reactor. In the SVR-DE approach, a support vector regression model is constructed for correlating process data comprising values of operating and performance variables. Next, model inputs describing process operating variables are optimized using Differential Evolution (DE) with a view to maximize the process performance. DE possesses certain unique advantages over the commonly used gradient-based deterministic optimization algorithms. The SVR-DE is a new strategy for chemical process modeling and optimization. The major advantage of the strategy is that modeling and optimization can be conducted exclusively from the historic process data wherein the detailed knowledge of process phenomenology (reaction mechanism, kinetics, etc.) is not required. Using SVR-DE strategy, a number of sets of optimized operating conditions leading to maximized EO production and catalyst selectivity were obtained. The optimized solutions, when verified in an actual plant, resulted in a significant improvement in the EO production rate and catalyst selectivity.

A comparison between linear least-squares method and non-linear regression method of the widely used equilibrium isotherms (Langmuir, Freundlich and Redlich-Peterson) for the sorption of Cr(VI) onto Cystoseira indica, which was chemically-modified by cross-linking with epichlorohydrin (CB1, CB2), or oxidized by potassium permanganate (CB3), or only washed by distilled water (RB) was examined. The biosorption equilibriums were established in about 2 h and the maximum removal was observed at pH 3.0 with solid to liquid ratio of 0.05 g/l. The four different linearized forms of Langmuir isotherms are also discussed. Langmuir isotherm parameters obtained from the four Langmuir linear equations are different but they are the same by using non-linear Langmuir equation. The best-fitting isotherms are Langmuir and Redlich-Peterson. The present investigation showed that the non-linear method is the more appropriate method to determine the isotherm parameters. A design procedure was proposed using the Langmuir isotherm to design a two stage sorption system to minimize the amount of biomass required for the treatment of Cr(VI) solution using Cystoseira indica. A two stage sorption system reduced the amount of biomass required by 51.2, 50.7, 51.1 and 51.3%, for CB1, CB2, CB3 and RB, respectively, to achieve the required amount of Cr(VI) removal for any solution volume.

A comprehensive triple-site-model was studied. This model was based on the new molecular dynamics outcomes obtained from the researchers on the different active sites that exist on the heterogeneous Ziegler-Natta catalyst. The model was composed of high-isospecific site (IS), low-isospecific site (IS), and aspecific site (AS). A transformation occurred on the high isotactic site and a second isotactic site was generated, as the low-isospecific site. The model is able to predict experimental results of bulk propylene polymerization with great accuracy. Moreover, the kinetic parameters and constants for each individual site has an Arrhenius type behavior. These parameters were determined using genetic algorithm adaptation method. The concentration of each type of sites, the isotactic and atactic fraction were determined during the polymerization, per unit mass of catalyst. A very good agreement was found in between the model and reported experimental results.

This study was conducted to present a mathematical model with numerical solutions to predict the mass distributions inside a cross sectional area of an individual Fajer variety of rough rice (Oryza Sativa L.) kernel as a function of drying time considering the effects of coupled heat and mass transfer processes. The modified Luikov's equation was used for the simulation of drying kinetics of a single rough rice grain. The applied drying models with simultaneous heat and mass diffusion in biological materials were developed by finite element method. A finite element formulation and solution of a set of nonlinear coupled conductive heat and diffusive moisture transfer equation to improve grain drying simulation of axisymmetric bodies is presented. Axisymmetric linear triangular elements with two degrees of freedom per node were used to discretize the rice grain. Moisture distribution inside the individual rough rice kernel was produced by the model. Good agreement was observed when the output of nonlinear model was compared to experimental data.

The aim of this work is to simulate the dynamic behavior of a phenol biodegradation process in a fluidized bed bioreactor (FBR). Pseudomonas putida is used for the biodegradation of phenol. A mathematical model was developed to describe the dynamic behavior of the biodegradation process. The model equations describing the process have been solved, and the rate of biodegradation and the biofilm thickness at different points of time have been determined. The mathematical model has been directly mapped onto the network architecture. The network is used to find an error function. Minimization of error function with respect to the network parameters (weights and biases) has been considered as training of the network. A real-coded genetic algorithm has been used for training the network in an unsupervised manner. The system is tested for two different inlet concentrations of feed. The results obtained are then compared with the experimental results. It is found that there is a good agreement between the experimental results and the results obtained from the model.

In this paper the historical perspective of the use of simulation in pulp and paper industry is presented and different applications discussed. Scientific papers as well as research and development work made by suppliers and software vendors is covered. The review covers a number of applications. First data reconciliation, root cause analysis and decision support is covered. This forms the basis for more advanced optimization and control. Physical models are then covered as they are especially good for engineering and design of new processes or the rebuilding of existing processes. As a complement for on-line applications quality property predictions with soft sensors like MVDA, ANN and grey box models are covered. When the process signals have been checked and soft sensors developed it is possible to perform process optimization and model based control and also to develop expert systems for decision support. Simulators also are used for operator training to increase the understanding of the processes as well as the interaction with the DCS system. Finally the future of process simulation in pulp and paper industry is discussed.

In this paper a brief, commented introduction to thermodynamics highlighting its connections with the simulation of chemical processes is presented. With these connections established I proceed to comment on current challenges faced by process and product modelers and conclude with a peek into the future of process and product design and the role of thermodynamics as the unifying discipline for engineers.

The objective of this paper is to review the application of the finite element-based CFD methods in mixing engineering. It provides a good opportunity to summarize the thirty five years of finite element achievements in the field of fluid flow started in the 70's, with the development of the first 2D Navier-Stokes solver for Newtonian and non-Newtonian viscous fluids, and extended in the 80's, with new solution algorithms to tackle 3D problems. For mixing simulations, a corner stone was the introduction of the "virtual finite element method" in the mid 90's for the simulation of flow systems with internal moving parts. In the on-going quest to improve the characterization of mixing systems for industrial needs, further developments are required, and progress could come from the next generation of computationally efficient multi-physics solvers.