Chemical Engineering Research and Design

Develops and analyses a model of a falling film evaporator with mechanical vapour recompression for the purpose of examining the coupling and functional controllability of the two core control loops for effecting temperature and product dry mass fraction. The relative gain array is used to assess coupling and the functional controllability analysis uses the methodology of Skogestad and Postlethwaite (1996). This determines if adequate levels of disturbance rejection can be achieved with allowable levels of manipulation. The aim is to understand the fundamental properties of the process which limit the ability of the control system to reject disturbances. A first principles model of appropriate complexity which captures the essential mechanism is developed and used

In this paper, a detailed study of the drying kinetics for a range of milk products has been conducted. Firstly, an improved glass-filament method for continuous monitoring of the drying process has been established through a series of modifications. The external heat and mass transfer correlation for pure water droplets with large evaporation rates have been obtained. Then, the characteristics of drying of milk droplets have been measured and interpreted. Furthermore, the changes in the diameters of the milk droplets during drying have been recorded using the improved system. This work has yielded some important information on how a dairy particle shrinks during water removal.

In this article we review the concepts and key developments of mixing enhancement through pulsation and oscillation. We focus on more recent research using oscillations in baffled tubes or columns, which generate discrete vortices as distinct from the more random type of flow produced by the well-established pulsed plate and reciprocating plate column devices. The scope of this review ranges from the basic chemical engineering concepts and key findings from recent research projects, including heat/mass transfer and residence time distribution (RTD); to applications and case studies involving specific reactions, e.g. polymerization and process intensification. We hope that this review will enable the reader to identify additional potential and perhaps an unexpectedly wide range of applications for the oscillatory baffled flow technology.

A new long-range predictive control algorithm for non-linear processes is proposed and is based on the popular Generalized Predictive Control (GPC) algorithm using the Takagi-Sugeno Kang (TSK) piece-wise fuzzy modelling approach. The proposed fuzzy modelling approach is integrated with the proposed control algorithm to form an adaptive control scheme based on a Controlled Auto-Regressive Integrated Moving Average (CARIMA) model structure. The performance of the adaptive control scheme is assessed using a series of experiments on the binary distillation column and on the Continuous Stirred Tank Reactor (CSTR) system.

Chemical product design involves a very large search space and evaluation of numerous alternatives. It also requires a number of tools and different sources of data for properties that define the needs of the chemical product. In this paper, a framework for computer-aided chemical product design is presented together with a review of methods and tools that may be useful in chemical product design. The objective of this framework is to provide a set of integrated methods and tools so that some of the chemical product design steps can be carried out faster, over a wider search space and using less resources. The requirements for such a framework and its resident methods and tools are that there exists a reliable set of models. Within this context, the differences between the process model, the property model and the chemical product model are highlighted together with a discussion on the roles of property models with respect to chemical product design. Several types of chemical product design problems are highlighted together with the corresponding methods and tools that can be used. Illustrative examples highlighting the use of the framework are also presented.

Understanding the wetting and evaporation behaviour of volatile droplets on heated surfaces is very important for many industrial applications. In this paper the behaviour of a sessile drop evaporating on a heated surface is investigated both experimentally and numerically. Results are reported for the evaporation of water drops on two different substrates at various temperatures. A numerical model, based on a finite element method, has been developed to describe the hydrodynamics inside the evaporating drop and the effect of the humidity on the evaporation process, assuming the droplet to be a spherical cap. The energy and Navier–Stokes equations are solved within the droplet and the vapour concentration is computed using the diffusion equation. The drop volume and flow and temperature fields within the drop are obtained and the evolution of the volume in time is compared with the experimental results.

A strategy for the integration of a novel CO2 photosynthetic culture and power generation system into a commercial bioethanol plant is presented. Photosynthetic microalgae column photobioreactors, acting as cathodic half cells, are coupled with existing yeast fermentors at a bioethanol plant, acting as anodic half cells, to create coupled microbial fuel cells. The microalgae photobioreactors also sequester CO2 emitted by the yeast fermentors. Incorporating microbial fuel cells into an existing bioethanol plant generates some of the power used in bioethanol production and the microalgae species Chlorella vulgaris contains oil, which acts as a byproduct for the production of biodiesel.The goal of the study is to determine the required design specifications of novel, airlift PBR cathodes to make the integrated system economically feasible at an existing bioethanol plant. Data from previous experimental studies was used to develop the optimum integration strategy. The reported parameters include PBR size, number of integrated MFCs, fuel cell outputs, oil (for biodiesel) production rate, and CO2 consumption rate.

Three types of thermally coupled distillation systems, namely the sequence with side rectifier, the sequence with side stripper and the Petlyuk column, have been shown to provide significant energy savings for the separation of ternary mixtures with respect to the conventional direct and indirect sequences. Although the Petlyuk column in general is more energy efficient than the other options, its structure creates potential operating problems because of the bidirectional flow of the vapor interconnecting streams. Alternative options with potential operational advantages have been recently suggested. In this paper we analyze the energy performance of six alternative schemes to the Petlyuk system with thermal coupling. Two of them make use of two interconnections with unidirectional flow, while the other four show a reduction in the number of interconnections. It is shown that several of the alternate options provide similar energy savings to the Petlyuk column.

Agas-liquid reactor is described, in which a plunging jet issues into a confined downcomer column, entraining headspace gas as it impinges onto the surface of a receiving pool of the same liquid. The two-phase mixture flows down the confining column and then rises through a concentric annulus, before disengaging in a separation vessel. Such a design gives increased gas-liquid contact times compared to jets plunging into open pools. Another feature of the current design is that it is straightforward to recycle both the gas and liquid phases from the separation, using only a single pump, thus ensuring almost complete utilization of the gaseous feed. A hydraulic model of the reactor is presented and tested against experimental data over a range of liquid flow rates and nozzle sizes. The agreement between the predictions and the experimental data indicates that the model provides a good basis for the design of this type of gas-liquid reactor. It is also shown that downcomers with diameter greater than 24 mm appeared to have little effect on the gas entrainment rate, when nozzles of 5 to 12 mm diameter were used. Smaller downcomers significantly reduced the gas volumetric entrainment flow rate.

The single stage polymerization recently proposed for producing micrometer-sized polymer particles in aqueous media was carried out with styrene monomer, potassium persulfate initiator and dimethyldodecylbetaine amphoteric surfactant in ranges of impeller speed (250–500rpm) and addition time of the surfactant to the system (5–30 min) at a monomer concentration of 1.1 mol/dm3H2O and initiator concentrations of 4 and 8mmol/ dm3H2O. Particle size distribution and monomer-to-polymer conversion were measured during the polymerizations. It was shown that the addition of small amount of the surfactant (0.5mmol/dm3H2O) was effective to promote the coagulation of particles, which led to the production of monodisperse, micrometer-sized polymer particles. The particle size distribution significantly depended on impeller speed. At the lowest impeller speed of 250 rpm, secondary particles were generated during the reaction and the size distribution of final particles was bimodal. The generation of secondary particles is explained from an increase in electric surface potential of polymer particles during the reaction. At impeller speeds above 250 rpm, particle size distributions were unimodal throughout the reaction and the average diameter of final particles increased with a decrease in impeller speed. Particles produced at 300 rpm had an average diameter of 2 /an and the highest monodispersity with a coefficient of variation of size distribution of 3.4% that is much smaller than typical monodispersity criterion of 10%.

Interfacial chemistry and particle interactions of aqueous muscovite dispersions have been investigated in the pH range 2–9. Particle zeta potential, reflecting interfacial chemistry, indicated a strong pH-history and solid loading dependency. Pristine particles’ zeta potentials measured from high to low pH indicated an isoelectric point (iep) at ∼pH 4.5. Subsequent measurements from low to high pH showed differing electrokinetic potentials with iep shift to higher pH values, the extent of which depended upon particle volume fraction and aging time. Dispersion shear yield stress analysis revealed similar pH-history and solid loading dependency. Incongruent leaching was observed to be responsible for the interfacial chemistry change and rheological behaviour. Upon decreasing pH from 9 to 2, considerable leaching of the key elements in muscovite, Al(III), Si(IV), K+ and Fe(III), occurred. The species concentrations decreased dramatically upon subsequent pH increase to higher values due to their hydrolysis and specific adsorption. Dispersion shear yield stresses recorded from high to low pH sweep were pH-independent. Similar measurements from low to high pH values, however, showed a strong pH-dependency, with maximum yield values at the ieps. The interfacial chemistry and particle interactions, both displaying bifurcation behaviour, showed good compliance with DLVO theory.

Equilibrium data on the absorption of CO2 in aqueous solutions of single and mixed amine was analysed using the Modified Kent Eisenberg model. The experimental value of the equilibrium constant for the formation of carbamate, instead of the fitted value as usually used by other investigators, was applied in the analysis. Data on CO2 loading in aqueous solutions of DEA and MDEA at various temperature (303–323 K) and CO2 partial pressure (0.09–100kPa) obtained from a stirred reactor was fitted to generate the different parameters in the model. Using these constants, the model was applied to predict the CO2 loading in solutions of DEA and MDEA reported in the literature. Prediction was also made on the loading in solutions of mixed DEA/MDEA obtained from experiments conducted over a range of composition (DEA:MDEA = 0–1) at different temperatures of 303–323 K, as well as those reported in the literature. In all cases, it was found that the model was able to give a relatively good CO2 loading over a wide range of operating conditions both in solutions of single and mixed amine using the constants generated from single amine experiments. It also suggested that the experimental value of the equilibrium constant for the formation of carbamate obtained from an earlier work can be taken as the true value for the reaction.

Preliminary evaluations using a simple but reliable short-cut method indicated that a 15 component aromatics mixture can be separated very efficiently into four fractions according to the given product specifications employing either a single or a multiple partition wall dividing wall column (DWC). The obtained results have been used to initiate rigorous simulations, to determine the number of stages required in different sections, as well as to obtain internal flows of vapour and liquid necessary for dimensioning and adequate cost estimation for two design alternatives. Based on the comparison of total annualised costs it appears that a multi-partition wall configuration that maximizes energy efficiency is a more attractive option for implementation in aromatics processing plants than more practical single partition wall configuration.

Oxyfuel combustion is seen as one of the major options for CO2 capture for future clean coal technologies. The paper provides an overview on research activities and technology development through a fundamental research underpinning the Australia/Japan Oxyfuel Feasibility Project. Studies on oxyfuel combustion on a pilot-scale furnace and a laboratory scale drop tube furnace are presented and compared with computational fluid dynamics (CFD) predictions. The research has made several contributions to current knowledge, including; comprehensive assessment on oxyfuel combustion in a pilot-scale oxyfuel furnace, modifying the design criterion for an oxy retrofit by matching heat transfer, a new 4-grey gas model which accurately predicts emissivity of the gases in oxy-fired furnaces has been developed for furnace modelling, the first measurements of coal reactivity comparisons in air and oxyfuel at laboratory and pilot-scale; and predictions of observed delays in flame ignition in oxy-firing.

The work reported has relevance to the assessment of the response of chemical plant structures to transient pressure loads, which may occur in a range of process operations during normal or failure conditions, and which require consideration at the design stage to contain their effects.Experimental and numerical results are presented for the case of the rupture of a high pressure tube, which is contained within a liquid-filled shell and tube heat exchanger of overall length 3.75 m and diameter 0.75 m. Such a tube failure gives rise to a pressure pulse load on the shell, and the measured pressure transient from this event has been used in a finite element analysis to compare measured and observed strains at the shell surface. Structural failure criteria can be introduced into the analysis and the predictive capability of this has been initially used to predict yield in circular plates exposed to transient pressure loads of different amplitude and duration. Comparison between observed and calculated strains, and the onset of yield, enable a failure limit curve to be constructed as a function of pressure amplitude and duration, and this is compared with the predicted curve based on a prescribed failure criterion.Calculations on the heat exchanger have been extended to explore the region of failure due to yield, and the predicted limit curve demonstrates that pressure amplitudes may be allowed to exceed the yield pressure if pulse widths are short. The engineering benefits arising from these observations are discussed in relation to the likely design implications.

The relay auto-tuning technique for PID controllers is here extended to tunemultivariable controllers. In the case of significant interaction, a fully cross-coupled multivariable controller rather than a decentralized controller should be employed. In this paper, an autotuning method for multivariable controllers from sequential relay feedback is proposed. The frequency response a m×m multivariable process is identified from limit cycles using a FFT-based method, and a multivariable controller is computed with linear least squares frequency response fitting. Various typical examples are included for illustration of the effectiveness of the method.

Protein crystallisation is known to be affected by many factors and inherently difficult to control. Being able to model the crystal growth behaviour, especially at process scale for the population of particles in a crystalliser will no doubt greatly help the formulation and controlled manufacture of protein crystals. In this paper, a morphological population balance model for crystallisation of tetragonal Hen-Egg-White (HEW) lysozyme is presented. Since the population balance model has incorporated crystal shape information, it is able to simulate the dynamic evolution of the shape distribution as well as size distribution. The morphological population balance model requires faceted growth kinetics data, which was obtained from published data in literature for the two identified independent crystallographic faces, {1 0 1} and {1 1 0}, of HEW lysozyme crystals.

Because of the structural change in industry from mass production to the production of small amount of various products with high added value, batch processing has come to the fore in the nineties. In the competitive atmosphere of today, batch chemical reactors are pinpointed as where better performance can be obtained through better process control. Due to their intrinsic characteristics, however, the control of batch reactors involves complex control functionality beyond the well-established norms for continuous processes, and constitutes a challenging area. This paper reviews the progress in the field of batch reactor control. Particular developments in model based control that have been applied to batch reactors, theoretical and experimental experience reported in literature are addressed. Some main directions of current activities are highlighted and suggestions for further research are given.

The effects of different mixing and process parameters on the formation of crystals obtained from semi-batch precipitation are reported. The influence of feed time, feed concentration, feed tube diameter and impeller type on particle morphology of calcium oxalate and calcium carbonate respectively are discussed. In addition, a hybrid mixing-CFD model, the Segregated FeedModel (SFM), is used to predict the influence of process conditions on product particle size via the population balance. Experimental results for the particle size using 1, 5 and 25 l reactors show reasonable agreement with model predictions.

A new process is proposed for CO2 removal from flue gas using the reaction CaO + CO2 ↔ CaCO3. This process consists of two fluidized bed reactors connected by solid transportation lines. In one reactor (absorber), CO2 in the flue gas is captured by CaO at 873 K and the produced CaCO3 is transported to another reactor (regenerator), in which CaCO3 is decomposed to CaO at 1223 K. The produced CaO is transported to the absorber again. The heat of decomposition in the regenerator is supplied by feeding coal and pure oxygen, thus the flue gas from the regenerator is high purity CO2 (>95%, dry base). In this work, a conceptual study is conducted for material balance, heat balance, power generation, and power consumption for O2 production and CO2 liquefaction (compression). Also, a kinetic study of CaO+CO2 → CaCO3 was conducted to design the absorber. The required bed height of the absorber was calculated by use of a bubbling fluidized bed model. The bed height was found to be determined not by the chemical reaction rate but by the arrangement of heat transfer tubes for heat recovery.

A comparison between the flow of air-water and steam-water mixtures through a venturi has established key differences between the two systems. The variations of local void fraction of two-phase flows through a converging-diverging venturi have been studied using gamma ray attenuation. The independent variables are the liquid and gas flow rates, the temperature and the axial position in the flow direction. The difference between the two systems is found to be most evident at, and downstream of the throat. In steam-water, void fraction increases substantially at, and beyond the throat as a result of the low local pressure at the throat. On the other hand, with air-water the void fraction at the throat decreases at low temperature because of local changes in the slip ratio of the two phases. The effect of temperature on flow through a venturi is found to be more significant than it is in a straight pipe. This phenomenon becomes more important as the system approaches its boiling point.

Mathematical modelling based on the continuum assumption and the dense phase kinetic theory is used in the work. It is found that the model captures the main features of solids motion and segregation in the drum and the results agree well with limited experimental data. The results show a clear two-region bed structure in the transverse plane of the drum operated in the rolling mode in terms of both total solids concentration and solids velocity distributions, a surface (active) layer with a low average concentration and a high velocity, and a passive layer close to the drum wall with high solids concentration and low solids velocity. The thickness of the two regions in terms of solids concentration differs greatly from that in terms of solids velocity, suggesting inaccuracy of the thin layer approximation models proposed in the past for predicting solids motion in the surface layer of rotating drums operated in the rolling mode. The results also show a clear core-shell structure in the transverse plane of the drum in terms of concentration of small and large particles. Under the conditions of this work, small particles tend to concentrate in the core region, while large particles tend to occupy the shell region, consistent with most experimental observations. Starting with a homogenous mixture of small and large particles, the above mentioned bed structures are found to be nearly fully developed within ∼2.56 drum revolutions, which is consistent with experimental observations. An initial attempt has also been made to simulate the axial particle segregation. Although the initial results are promising, more work is clearly needed to establish a better picture of axial segregation.

In situ steam reforming of tar from rapid pyrolysis of biomass was investigated experimentally using a two-stage reactor. Pulverized biomass was continuously fed into the rapid pyrolysis zone of the reactor at 773 K, where nascent volatiles and char were isolated from each other. The volatiles were then mixed with steam and introduced into the steam-reforming zone at 873 to 1073 K consisting of a fixed-bed of mesoporous alumina particles. The yield of tar heavier than naphthalene decreased as the amount of volatiles fed into the fixed-bed increased, and finally reached an undetectable level (well below 0.01% on a biomass carbon basis), leaving benzene and naphthalene with yields of 0.5–1% and below 0.05%, respectively. Coke was deposited over the alumina particles with a yield of 10–20%, and it acted as a catalyst active enough to eliminate polyaromatic hydrocarbons as the representative constituents of the heavy tar. Thus the mesoporous alumina played a dual role of eliminating tar and controlling the total yield of char and coke, and would therefore be suitable as the bed material for biomass gasification in two-stage gasifier with the bed material being circulated between the biomass pyrolysis/steam-reforming zone and the char/coke combustion zone.

The profit of a biorefinery is highly affected by the supply of its raw materials and margin from the product. Taking responsive actions to the fluctuating market conditions are of major concern for the efficient management of biorefineries. In this study, an integrated biorefinery process is investigated to tackle this issue by diversifying products as well as raw materials. The optimal operational planning of the integrated process is calculated based on their price scenarios and the product demands. Moreover, the risk involved in the purchase of raw materials due to their varying prices is proposed to be minimized by futures contracts. The proposed planning models will contribute to decreasing the profit variability and escalating the operational flexibility of a biorefinery.

In view of the photoactive properties and potential penetration of fine TiO2 particles to human skin, it is important to seek a compromise to provide the optimum attenuation effects of UV by minimizing the quantity of TiO2 used in developing a sunscreen. In this paper, titanium dioxide powderwas used as fine particles to coat coarse nylon and polyethylene particles by high-speed rotational impact blending. The effects of the coating ratio and size of the coarse particles on the effectiveness of UV attenuation were investigated. The effectiveness of UV protection was found to be increased by decreasing the coating ratio, as this led to a lower level of surface agglomerates being formed. A further advantage of coating large polymer particles with fine titanium dioxide was an improvement in the bulk dispersion, which resulted in optimum UV protection.