The Chemical Engineering Journal and the Biochemical Engineering Journal最新文献

A realistic model has bene developed for the hydrodynamics and mass transfer in a cocurrent packed column. It postulates that the liquid phase is divided into stagnant and dynamic regions. Unlike in other models, the stagnant phase is not completely mixed and its concentration profile is given by a Fickian-type model equation. The model considers axial dispersion in the dynamic phase and mass transfer between the stagnant and dynamic regions of the liquid phase. The resulting partial differential equations are solved numerically by the method of orthogonal collocation. Simulation results highlighting the effects of various parameters for a step input and a step decrease in tracer concentration are also presented. The model predictions for both the downflow and upflow modes of operation are compared with available experimental data and are found to agree well.

This paper deals with the application of a neural controller for temperature control of a batch reactor. The term “neural controller” is used to refer to a multilayer neural network which computes the control values to be applied to the process.

We present the design and the development of the neural network: architecture, learning database and learning procedure. In a first step, the learning phase consists in teaching the neural network to map the dynamics of a classical adaptive controller (generalized predictive control with double model reference) implemented on the process. Although the neural controller performance is good for operating conditions included in the learning set (interpolation), it exhibits limitations on extrapolation. In this work, two methods for the on-line adaptation of the network's weights are developed: one of them is the “specialized” learning technique, whereas the other uses another neural network in order to model the reactor dynamics. Several results are shown and prove the good capacities of neural networks for controlling batch processes.

Mathematical modelling of a non-catalytic gas-solid reaction in a porous pellet, where both diffusion and chemical reaction are important, is represented by a pair of coupled partial differential equations. When the kinetics are half-order with respect to solid concentration, no analytical solution exists. Therefore numerical solutions or approximate methods have been used.

In this work a new incremental analytical method of solution has been developed. The results of this method in comparison with the existing numerical and approximate solutions show high accuracy and good fitting. The method is very simple to handle with small calculators and also gives some analytical expressions for gas and solid concentrations.

An experimental study was carried out to evaluate ϵ around an oscillating grid producing zero-mean shear flow. The spatial average $\text{ϵ}$ throughout the mixing vessel was determined by the mesurement of the net power input of the grid, this complying with a relationship of the form $\text{ϵ}$αf³_DS³ Re^{−$\text{1}\text{2}$} h⁻¹, in which f_D is the driving frequency, S is the stroke length, Re the grid Reynolds number and h the depth of the mixing column. From mesurements of the power spectrum using laser Doppler anemometry, spectral collapse of the power spectra in the domain of the energy containing eddies at different distances from the grid was demonstrated using u² (u as the turbulence r.m.s. velocity) and τ_E (Eulerian time integral scale) as the scaling parameters. From a balance of the power input and the energy losses, together with the feature of spectral collapse, it was shown that ϵ could be estimated by ϵ = γ₁u²/τ_E in this domain of the power spectrum with γ₁ as a multiplying coefficient independent of distance from the grid.

From the spatial dependence of the turbulence parameters it was evident that within about 1.5 mesh lengths of the grid there was a transition region beyond which it was found that spatial variations were consistent with the dependences uαz⁻¹, τ_Eαz², and ϵαz⁻⁴, where z is the distance from the grid.

Attrition experiments lasting several hours were carried out for low concentration crystal suspensions (3 kg m⁻³) at two different stirring rates (960 and 1154 rev min⁻¹). The size distributions were measured at fixed time intervals by continuously circulating the suspension through the cell of a laser particle sizer. Monosized sodium chloride crystals (500−560 μm), after attrition, gave rise to a bimodal size distribution with a higher peak in the parent crystal region and a much lower one below 32 μm: the amount of fine fragments increased continuously throughout the run.

The size distributions of the particles present in the crystallizer at each time were successfully modelled by superimposing the effects of abrasion and breakage fracture mechanisms. At the beginning of the run, abrasion largely prevails, being responsible for more than 98% of the fracture, but its importance rapidly declines with time and at the end of the run (8–12 h) the contributions of abrasion and breakage mechanisms to attrition appear comparable.

A mathematical model was proposed to describe the behaviour of liquid emulsion membranes for the extraction of penicillin G in a continuous countercurrent mixing column. A polyamine-type surfactant acts non only as carrier but also as surface-stabilizing agent; thus the influence of surfactant on extraction should be considered in mathematical modeling when its effect is significant. The proposed model takes into account the influence of surfactant on mass transfer. The advancing front model was employed for deriving the overall mass transfer coefficient in the emulsion globule, and the axial dispersion model was applied to the external feed phase. The experimental data were compared with the proposed model, the calculations without considering the contribution of the surfactant to extraction, and the calculations without considering diffusion in the emulsion phase.

The reduction of anthraquinone (AQ) with sodium sulphide in alkaline medium has been studied by measuring conversion as a function of time at various temperatures, sodium sulphide concentrations, initial radii of anthraquinone particles and sodium hydroxide concentrations. The reduced product is the disodium salt of 9,10-dihydroxyanthracene. The kinetic data were fitted to the isothermal shrinking core model (SCM) for cylindrical particles without porous solid product layer formation. The results indicate that the surface chemical reaction is the controlling step of the overall process rate. AQ reduction is of first order with respect to sodium sulphide concentration. The activation energy was obtained from the Arrhenius law and was found to be 68.81 kJ mol⁻¹ for the investigated temperature range.

Blending of viscous newtonian liquids was studied in dual- and triple-turbine agitated tanks. The dimensionless power, mixing and interstage flow numbers were all found to be dependent on the Reynolds number as long as the full development of turbulence was inhibited by viscosity (Re < 2 × 10⁴) and they were found to be constant above that limit. The complexity of the transitional regime makes it adverse for simple correlation, but, surprisingly, modelling of mixing in the upper transition region (Re > 400) could still be achieved using a compartment model developed elsewhere for turbulent mixing simulation in this geometry. The model corresponds to a simplified physical representation of the hydrodynamics of the tank

The cathodic reduction of ferricyanide ions is used to characterize liquid—solid mass transfer in porous electrodes in the creeping flow regime. Our study deals with packed beds of spheres, long cylinders and plates of low height-to-side ratio. Correlations are proposed and compared with predictive equations based on the association of the capillary representation of the porous medium with the analytical solution for mass transfer at a pipe wall, for fully developed laminar flow in short tubes.

The model leads to satisfying values of the mass transfer coefficient in the case of beds packed with spheres and parallelepipedal particles.