Milorad Dudukovic - Academia.edu (original) (raw)

Papers by Milorad Dudukovic

Industrial & Engineering Chemistry Process Design and Development, 1986

... z{ = component k of the composition vector d zm = composition vector of the point which satis... more ... z{ = component k of the composition vector d zm = composition vector of the point which satisfies eq 3 Greek Letters A = net flow Literature Cited Chimowitz, EH; Anderson, T. F.; Macchieto, S.; Stutzman, L. F. Ind. Eng. ... Ind. Eng. Chem. Process Des. Dev., Vol. 25, No. 3, 1986 635 ...

Industrial & Engineering Chemistry Research, 1995

A reproducible procedure for packing small-diameter packed beds with a mixture of fines and catal... more A reproducible procedure for packing small-diameter packed beds with a mixture of fines and catalyst is reported. Such beds are essential for process catalyst testing in trickle-flow reactors.

Chemical Engineering Science, 2001

Radioactive gas tracer measurements conducted during liquid-phase methanol synthesis from syngas ... more Radioactive gas tracer measurements conducted during liquid-phase methanol synthesis from syngas in a pilot-scale slurry bubble column at the alternate fuels development unit (AFDU), La Porte have been compared with simulations from two mechanistic reactor models — ...

Chemical Engineering Science, 1999

The knowledge of gas holdup, liquid recirculation and turbulent parameters is important for desig... more The knowledge of gas holdup, liquid recirculation and turbulent parameters is important for design and performance calculation of bubble column reactors. Although numerous experimental studies have been reported on this subject, most are point measurements limited to columns without internals operated at low gas velocities. In this study, we present the results obtained for the gas holdup pro"les, time-averaged liquid velocity pro"les, turbulent stresses and eddy di!usivities (radial and axial) obtained in a 18 (44 cm) diameter column without and with internals similar to those used in industrial scale units (e.g., heat exchanger tubes) using both air/water and air/drakeoil 10 (viscosity&30 cP) at gas super"cial velocities of 2, 5 and 10 cm/s. The scale-up procedure suggested by is critically examined in light of these results. (H. Al-Dahhan).

Chemical Engineering Science, 1996

The isothermal decomposition of hydrogen peroxide on a CuCr catalyst in a laboratory scale trickl... more The isothermal decomposition of hydrogen peroxide on a CuCr catalyst in a laboratory scale trickle bed reactor was used to test model predictions of the dependence of liquid reactant conversion on space time for different operating conditions. It is assured that the decomposition of ...

Chemical Engineering Science, 1997

Simple models, based on liquid-solid and gas-liquid interaction are proposed in order to model a ... more Simple models, based on liquid-solid and gas-liquid interaction are proposed in order to model a gas-liquid flow in trickle-bed reactors operating in high interaction regimes. Good results are obtained, when the liquid-solid and the gas-liquid interactions are modeled separately. Besides the classical Ergun equation, a simple model of the liquid-solid interactions, based on a boundary layer flow is proposed. This approach results in a relation which is very close to the well-known liquid saturation correlation proposed by Specchia and Baldi (1977, Chem. Engng Sci.32, 515–523), and might be its theoretical justification. The gas-liquid interaction is modeled using a drift-flux approach. The different models are tested against the data bank of about 1500 saturation and pressure drop measurements, established mainly by the Nancy research group. The importance of the geometrical characteristics of the porous media and their incomplete description are emphasized. Finally simple predictive correlations based on the models are proposed, tested against the data bank and compared to literature correlations.

Chemical Engineering Science, 1990

... 1275 10.0 0.0 5.0 15.0 Radial Position, em Fig. 5. Stagnant region map in a bubble column. ..... more ... 1275 10.0 0.0 5.0 15.0 Radial Position, em Fig. 5. Stagnant region map in a bubble column. ... The entry region is determined to be approximately ome column diameter. ... Engra., 52, 1 9. Joshi. .3.8. and NM Sharma, (1979), A Circulation Cell Model for Bobble Columns, Trans. lust. ...

Catalysis Today, 2003

ABSTRACT A multicomponent one-dimensional dynamic mathematical model for the reacting slurry syst... more ABSTRACT A multicomponent one-dimensional dynamic mathematical model for the reacting slurry systems with a change in gas flow rate due to the chemical reaction is developed. A change in gas flow rate caused by the chemical reaction is modeled using the overall gas mass balance. Thus, all relevant chemical species are included in the model. Linear first-order reaction kinetics is considered. The gas phase is modeled using the two-bubble class hydrodynamic model. The interaction between small and large bubbles is included as the cross-flow. Suspension of liquid and solids is assumed to form a pseudo slurry phase. Back-mixing in all of the three considered phases, small bubbles, large bubbles and slurry, is accounted for using the axial dispersion model (ADM). Energy balance of the slurry phase is also included in the model. The developed general reacting slurry system model is used to simulate the performance of the Fischer–Tropsch (FT) slurry bubble column (SBC) reactor. Performance of the developed ADM based model is compared with the reactor scale models in which the reactor back-mixing is represented using some combination of the two limiting ideal reactor models of, complete stirred or plug flow.

Catalysis Today, 2001

A phenomenological (mechanistic) model has been developed for describing the gas and liquid/slurr... more A phenomenological (mechanistic) model has been developed for describing the gas and liquid/slurry phase mixing in churn turbulent bubble columns. The gas and liquid phase recirculation rates in the reactor, which are needed as inputs to the mechanistic reactor model are estimated via a sub-model which uses the two-fluid approach in solving the Navier-Stokes equations. This sub-model estimates the effective bubble diameter in the reactor cross-section and provides a consistent basis for the estimation of the volumetric mass transfer coefficients. The strategy for the numerical solution of the sub-model equations is presented along with the simulation results for a few cases. The overall reactor model has been tested against experimental data from radioactive gas tracer experiments conducted at the Alternate Fuels Development Unit (AFDU), La Porte, TX under conditions of methanol synthesis.

A laboratory fixed-bed gasification reactor was designed and built with the objective of collecti... more A laboratory fixed-bed gasification reactor was designed and built with the objective of collecting operational data for model validation and parameter estimation. The reactor consists of a 4 inch stainless steel tube filled with coal or char. Air and steam is fed at one end of the reactor and the dynamic progress of gasification in the coal or char bed is observed through thermocouples mounted at various radial and axial locations. Product gas compositions are also monitored as a function of time. Results of gasification runs using Wyoming coal are included in this report. In parallel with the experimental study, a two-dimensional model of moving bed gasifiers was developed, coded into a computer program and tested. This model was used to study the laboratory gasifier by setting the coal feed rate equal to zero. The model is based on prior work on steady state and dynamic modeling done at Washington University and published elsewhere in the literature. Comparisons are made between model predictions and experimental results. These are also included in this report. 23 references, 18 figures, 6 tables.

The fabrication of composite laminates having a thermosetting resin matrix is a complex process. ... more The fabrication of composite laminates having a thermosetting resin matrix is a complex process. It involves simultaneous heat, mass, and momentum transfer along with chemical reaction in a multiphase system with time-dependent material properties and boundary conditions. Two critical problems, which arise during production of thick structural laminates, are the occurrence of severely detrimental voids and gradients in resin concentration.

Industrial & Engineering Chemistry Process Design and Development, 1986

1979, 78(2), 343. Des. Dev. 1982, 27(1), 79. Knopf, F. C.; Okos, MR; G. V. Reklattls, 0. V. Ind. ... more 1979, 78(2), 343. Des. Dev. 1982, 27(1), 79. Knopf, F. C.; Okos, MR; G. V. Reklattls, 0. V. Ind. Eng. Chem. Process Maudefli, A.; Rlppin, DWT Chem. Eng. Rug. 1980, 76(4). 37. Ohtake, Y.; NisMda, N. Oper. Res. Lett. 1985, 4(1), 41. Overturf. E. W.: Reklaltk. GV: Woods. J. M Ind, ...

Pure and Applied Chemistry, 2000

Powder Technology, 2001

We propose a simple method for accurately estimating the solids circulation rate in a closed-loop... more We propose a simple method for accurately estimating the solids circulation rate in a closed-loop liquid–solid circulating fluidized bed. The technique is based on the use of radioactive isotopes for estimating the solids velocity and volume fraction distribution in a chosen section of the circulating fluidized bed loop. The method is demonstrated for a liquid–solid circulating fluidized bed, and is

Powder Technology, 2004

ABSTRACT It is important to know the overall solids flux and solids Cycle Time Distribution (CTD)... more ABSTRACT It is important to know the overall solids flux and solids Cycle Time Distribution (CTD) to properly characterize Circulating Fluidized Bed (CFB) systems. Although conceptually simple, estimating the solids mass flux in these closed loop systems, either via experiments or by using a model, is a non-trivial problem. In the present work, a non-invasive in situ technique is applied, which is sensitive, and can be used on a large-scale circulating flow systems. The technique is based on estimating the solids velocity and volume fraction distribution in a section (downcomer) of the CFB loop. Solids velocity is obtained by tracking a single radioactive particle as tracer, using a two detector setup, and the volume fraction distribution is obtained by γ-ray densitometry measurements. In the downcomer section of a 2-in.-diameter column, the solids flow was found to be close to a moving packed bed condition, with minimal random fluctuations in the solids volume fraction. Hence, only a negligible error is incurred by using the cross-sectionally averaged solids velocity and holdup values for the estimation of the solids mass flow. A calibration curve was established for the overall solids mass flux as a function of superficial gas velocity, and the mass flux values compared with those obtained from a timing and weighing method were within 4%.

Polymer Composites, 1987

... solution to one-dimensional consolidation and three-dimensional resin flow in a composite lam... more ... solution to one-dimensional consolidation and three-dimensional resin flow in a composite laminate during processing is obtained by simultaneously solving the above ... An experimentally obtained relationship be-tween the void ratio and effective stress for the skeletal ...

Polymer Composites, 1994

ABSTRACT A one-dimensional devolatilization model is developed for autoclave processing of themro... more ABSTRACT A one-dimensional devolatilization model is developed for autoclave processing of themroplastic polyimide, graphite-reinforced, composites. The model rests on the assumption that continous gas passages are formed during processing due to partical crystallization of the polymer in the course of polymerization. Both DSC and optica microscopy indicate that crystallinity develops in the Avimid-III polyimide resin during polymerization, which in turn facilitates formation of cracks. These cracks heal upon removal of volatiles and polymer melting. Model-Calculated temperature profiles in the laminate during polymerization, and the predicted removal rates of volatile species (water, ethanol, NMP) are compared to experimental miniautoclave data for the DuPont Avimid K-III/IM6 system. Good agreement is obtained when the volumetric mass transfer coefficient required by the model is assumed to vary proportionally to the inverse of the changing resin viscosity. The model allows the investigation of desired cure cycle parameters for different laminate thicknesses.

Industrial & Engineering Chemistry Research, 2009

ABSTRACT The effect of gas sparger design on the gas holdup radial profile in a bubble column (wi... more ABSTRACT The effect of gas sparger design on the gas holdup radial profile in a bubble column (with a diameter of 0.162 m) has been studied using γ-ray computed tomography (CT). Six different configurations of gas spargers were examined, using an air−water system for selected superficial gas velocities from 2 cm/s to 30 cm/s, covering the homogeneous and heterogeneous (churn-turbulent) flow regimes. Two operating pressures were used: 1 and 4 atm. Differences were found between the gas holdup distributions produced by different spargers at dimensionless radii of r/R < 0.8 in the central region of the column. The cross and single nozzle spargers produced closely similar gas holdup distributions, while the perforated plate sparger produced a higher gas holdup when compared to other spargers with the same percentage open area (POA). At 4 atm, the sparger design did not have a significant effect on the gas holdup profiles, compared to atmospheric pressure, except for the case of the single-hole sparger, which was found to yield a higher gas holdup.

Industrial & Engineering Chemistry Research, 2011

ABSTRACT In field tomography, the total number of scan lines collected during imaging is usually ... more ABSTRACT In field tomography, the total number of scan lines collected during imaging is usually limited. At the same time, the demands on the technique regarding resolution are low, corresponding to diagnostic purposes (e.g., presence or absence of maldistribution). In this study, we investigate the level of information needed for purpose of field diagnostics when using a fan beam tomography scanner. We utilize the theoretical framework of Fisher information content to examine dependence between projection data collection methodology and image quality. The findings are substantiated by experimental results obtained on suitable phantom designed to mimic maldistribution in packed bed. The performance of expectation–maximization and alternating-minimization image reconstruction algorithms for varying number of scan lines used is compared.

Industrial & Engineering Chemistry Process Design and Development, 1986

... z{ = component k of the composition vector d zm = composition vector of the point which satis... more ... z{ = component k of the composition vector d zm = composition vector of the point which satisfies eq 3 Greek Letters A = net flow Literature Cited Chimowitz, EH; Anderson, T. F.; Macchieto, S.; Stutzman, L. F. Ind. Eng. ... Ind. Eng. Chem. Process Des. Dev., Vol. 25, No. 3, 1986 635 ...

Industrial & Engineering Chemistry Research, 1995

A reproducible procedure for packing small-diameter packed beds with a mixture of fines and catal... more A reproducible procedure for packing small-diameter packed beds with a mixture of fines and catalyst is reported. Such beds are essential for process catalyst testing in trickle-flow reactors.

Chemical Engineering Science, 2001

Radioactive gas tracer measurements conducted during liquid-phase methanol synthesis from syngas ... more Radioactive gas tracer measurements conducted during liquid-phase methanol synthesis from syngas in a pilot-scale slurry bubble column at the alternate fuels development unit (AFDU), La Porte have been compared with simulations from two mechanistic reactor models — ...

Chemical Engineering Science, 1999

The knowledge of gas holdup, liquid recirculation and turbulent parameters is important for desig... more The knowledge of gas holdup, liquid recirculation and turbulent parameters is important for design and performance calculation of bubble column reactors. Although numerous experimental studies have been reported on this subject, most are point measurements limited to columns without internals operated at low gas velocities. In this study, we present the results obtained for the gas holdup pro"les, time-averaged liquid velocity pro"les, turbulent stresses and eddy di!usivities (radial and axial) obtained in a 18 (44 cm) diameter column without and with internals similar to those used in industrial scale units (e.g., heat exchanger tubes) using both air/water and air/drakeoil 10 (viscosity&30 cP) at gas super"cial velocities of 2, 5 and 10 cm/s. The scale-up procedure suggested by is critically examined in light of these results. (H. Al-Dahhan).

Chemical Engineering Science, 1996

The isothermal decomposition of hydrogen peroxide on a CuCr catalyst in a laboratory scale trickl... more The isothermal decomposition of hydrogen peroxide on a CuCr catalyst in a laboratory scale trickle bed reactor was used to test model predictions of the dependence of liquid reactant conversion on space time for different operating conditions. It is assured that the decomposition of ...

Chemical Engineering Science, 1997

Simple models, based on liquid-solid and gas-liquid interaction are proposed in order to model a ... more Simple models, based on liquid-solid and gas-liquid interaction are proposed in order to model a gas-liquid flow in trickle-bed reactors operating in high interaction regimes. Good results are obtained, when the liquid-solid and the gas-liquid interactions are modeled separately. Besides the classical Ergun equation, a simple model of the liquid-solid interactions, based on a boundary layer flow is proposed. This approach results in a relation which is very close to the well-known liquid saturation correlation proposed by Specchia and Baldi (1977, Chem. Engng Sci.32, 515–523), and might be its theoretical justification. The gas-liquid interaction is modeled using a drift-flux approach. The different models are tested against the data bank of about 1500 saturation and pressure drop measurements, established mainly by the Nancy research group. The importance of the geometrical characteristics of the porous media and their incomplete description are emphasized. Finally simple predictive correlations based on the models are proposed, tested against the data bank and compared to literature correlations.

Chemical Engineering Science, 1990

... 1275 10.0 0.0 5.0 15.0 Radial Position, em Fig. 5. Stagnant region map in a bubble column. ..... more ... 1275 10.0 0.0 5.0 15.0 Radial Position, em Fig. 5. Stagnant region map in a bubble column. ... The entry region is determined to be approximately ome column diameter. ... Engra., 52, 1 9. Joshi. .3.8. and NM Sharma, (1979), A Circulation Cell Model for Bobble Columns, Trans. lust. ...

Catalysis Today, 2003

ABSTRACT A multicomponent one-dimensional dynamic mathematical model for the reacting slurry syst... more ABSTRACT A multicomponent one-dimensional dynamic mathematical model for the reacting slurry systems with a change in gas flow rate due to the chemical reaction is developed. A change in gas flow rate caused by the chemical reaction is modeled using the overall gas mass balance. Thus, all relevant chemical species are included in the model. Linear first-order reaction kinetics is considered. The gas phase is modeled using the two-bubble class hydrodynamic model. The interaction between small and large bubbles is included as the cross-flow. Suspension of liquid and solids is assumed to form a pseudo slurry phase. Back-mixing in all of the three considered phases, small bubbles, large bubbles and slurry, is accounted for using the axial dispersion model (ADM). Energy balance of the slurry phase is also included in the model. The developed general reacting slurry system model is used to simulate the performance of the Fischer–Tropsch (FT) slurry bubble column (SBC) reactor. Performance of the developed ADM based model is compared with the reactor scale models in which the reactor back-mixing is represented using some combination of the two limiting ideal reactor models of, complete stirred or plug flow.

Catalysis Today, 2001

A phenomenological (mechanistic) model has been developed for describing the gas and liquid/slurr... more A phenomenological (mechanistic) model has been developed for describing the gas and liquid/slurry phase mixing in churn turbulent bubble columns. The gas and liquid phase recirculation rates in the reactor, which are needed as inputs to the mechanistic reactor model are estimated via a sub-model which uses the two-fluid approach in solving the Navier-Stokes equations. This sub-model estimates the effective bubble diameter in the reactor cross-section and provides a consistent basis for the estimation of the volumetric mass transfer coefficients. The strategy for the numerical solution of the sub-model equations is presented along with the simulation results for a few cases. The overall reactor model has been tested against experimental data from radioactive gas tracer experiments conducted at the Alternate Fuels Development Unit (AFDU), La Porte, TX under conditions of methanol synthesis.

A laboratory fixed-bed gasification reactor was designed and built with the objective of collecti... more A laboratory fixed-bed gasification reactor was designed and built with the objective of collecting operational data for model validation and parameter estimation. The reactor consists of a 4 inch stainless steel tube filled with coal or char. Air and steam is fed at one end of the reactor and the dynamic progress of gasification in the coal or char bed is observed through thermocouples mounted at various radial and axial locations. Product gas compositions are also monitored as a function of time. Results of gasification runs using Wyoming coal are included in this report. In parallel with the experimental study, a two-dimensional model of moving bed gasifiers was developed, coded into a computer program and tested. This model was used to study the laboratory gasifier by setting the coal feed rate equal to zero. The model is based on prior work on steady state and dynamic modeling done at Washington University and published elsewhere in the literature. Comparisons are made between model predictions and experimental results. These are also included in this report. 23 references, 18 figures, 6 tables.

The fabrication of composite laminates having a thermosetting resin matrix is a complex process. ... more The fabrication of composite laminates having a thermosetting resin matrix is a complex process. It involves simultaneous heat, mass, and momentum transfer along with chemical reaction in a multiphase system with time-dependent material properties and boundary conditions. Two critical problems, which arise during production of thick structural laminates, are the occurrence of severely detrimental voids and gradients in resin concentration.

Industrial & Engineering Chemistry Process Design and Development, 1986

1979, 78(2), 343. Des. Dev. 1982, 27(1), 79. Knopf, F. C.; Okos, MR; G. V. Reklattls, 0. V. Ind. ... more 1979, 78(2), 343. Des. Dev. 1982, 27(1), 79. Knopf, F. C.; Okos, MR; G. V. Reklattls, 0. V. Ind. Eng. Chem. Process Maudefli, A.; Rlppin, DWT Chem. Eng. Rug. 1980, 76(4). 37. Ohtake, Y.; NisMda, N. Oper. Res. Lett. 1985, 4(1), 41. Overturf. E. W.: Reklaltk. GV: Woods. J. M Ind, ...

Pure and Applied Chemistry, 2000

Powder Technology, 2001

We propose a simple method for accurately estimating the solids circulation rate in a closed-loop... more We propose a simple method for accurately estimating the solids circulation rate in a closed-loop liquid–solid circulating fluidized bed. The technique is based on the use of radioactive isotopes for estimating the solids velocity and volume fraction distribution in a chosen section of the circulating fluidized bed loop. The method is demonstrated for a liquid–solid circulating fluidized bed, and is

Powder Technology, 2004

ABSTRACT It is important to know the overall solids flux and solids Cycle Time Distribution (CTD)... more ABSTRACT It is important to know the overall solids flux and solids Cycle Time Distribution (CTD) to properly characterize Circulating Fluidized Bed (CFB) systems. Although conceptually simple, estimating the solids mass flux in these closed loop systems, either via experiments or by using a model, is a non-trivial problem. In the present work, a non-invasive in situ technique is applied, which is sensitive, and can be used on a large-scale circulating flow systems. The technique is based on estimating the solids velocity and volume fraction distribution in a section (downcomer) of the CFB loop. Solids velocity is obtained by tracking a single radioactive particle as tracer, using a two detector setup, and the volume fraction distribution is obtained by γ-ray densitometry measurements. In the downcomer section of a 2-in.-diameter column, the solids flow was found to be close to a moving packed bed condition, with minimal random fluctuations in the solids volume fraction. Hence, only a negligible error is incurred by using the cross-sectionally averaged solids velocity and holdup values for the estimation of the solids mass flow. A calibration curve was established for the overall solids mass flux as a function of superficial gas velocity, and the mass flux values compared with those obtained from a timing and weighing method were within 4%.

Polymer Composites, 1987

... solution to one-dimensional consolidation and three-dimensional resin flow in a composite lam... more ... solution to one-dimensional consolidation and three-dimensional resin flow in a composite laminate during processing is obtained by simultaneously solving the above ... An experimentally obtained relationship be-tween the void ratio and effective stress for the skeletal ...

Polymer Composites, 1994

ABSTRACT A one-dimensional devolatilization model is developed for autoclave processing of themro... more ABSTRACT A one-dimensional devolatilization model is developed for autoclave processing of themroplastic polyimide, graphite-reinforced, composites. The model rests on the assumption that continous gas passages are formed during processing due to partical crystallization of the polymer in the course of polymerization. Both DSC and optica microscopy indicate that crystallinity develops in the Avimid-III polyimide resin during polymerization, which in turn facilitates formation of cracks. These cracks heal upon removal of volatiles and polymer melting. Model-Calculated temperature profiles in the laminate during polymerization, and the predicted removal rates of volatile species (water, ethanol, NMP) are compared to experimental miniautoclave data for the DuPont Avimid K-III/IM6 system. Good agreement is obtained when the volumetric mass transfer coefficient required by the model is assumed to vary proportionally to the inverse of the changing resin viscosity. The model allows the investigation of desired cure cycle parameters for different laminate thicknesses.

Industrial & Engineering Chemistry Research, 2009

ABSTRACT The effect of gas sparger design on the gas holdup radial profile in a bubble column (wi... more ABSTRACT The effect of gas sparger design on the gas holdup radial profile in a bubble column (with a diameter of 0.162 m) has been studied using γ-ray computed tomography (CT). Six different configurations of gas spargers were examined, using an air−water system for selected superficial gas velocities from 2 cm/s to 30 cm/s, covering the homogeneous and heterogeneous (churn-turbulent) flow regimes. Two operating pressures were used: 1 and 4 atm. Differences were found between the gas holdup distributions produced by different spargers at dimensionless radii of r/R < 0.8 in the central region of the column. The cross and single nozzle spargers produced closely similar gas holdup distributions, while the perforated plate sparger produced a higher gas holdup when compared to other spargers with the same percentage open area (POA). At 4 atm, the sparger design did not have a significant effect on the gas holdup profiles, compared to atmospheric pressure, except for the case of the single-hole sparger, which was found to yield a higher gas holdup.

Industrial & Engineering Chemistry Research, 2011

ABSTRACT In field tomography, the total number of scan lines collected during imaging is usually ... more ABSTRACT In field tomography, the total number of scan lines collected during imaging is usually limited. At the same time, the demands on the technique regarding resolution are low, corresponding to diagnostic purposes (e.g., presence or absence of maldistribution). In this study, we investigate the level of information needed for purpose of field diagnostics when using a fan beam tomography scanner. We utilize the theoretical framework of Fisher information content to examine dependence between projection data collection methodology and image quality. The findings are substantiated by experimental results obtained on suitable phantom designed to mimic maldistribution in packed bed. The performance of expectation–maximization and alternating-minimization image reconstruction algorithms for varying number of scan lines used is compared.