An improved method is presented for describing the apparent intrinsic pore properties (mean pore size and pore size distribution) of ultrafiltration membrane from solute rejection curves. To obtain the apparent intrinsic pore properties, firstly the mean pore diameter is determined from the solute rejection curve by correcting the steric and hydrodynamic hindrance effect of pores on molecules, and secondly the effects of operating conditions (pressure, feed concentration and recirculation velocity) on the pore properties are examined by ultrafiltration of seven solutes using SEPA-PS-0 membrane and are eliminated by extrapolating to “zero” operating conditions. The pore property parameters (μ, Dp and θ) obtained from the solute rejection curve are affected by two factors: 1) the plugging of membrane pores by the solute and 2) the deformation of solute due to the variation of pressure and feed concentration. But the recirculation velocity does not affect the pore property parameters. the mean pore diameter is about 60 Å and the pore diameter distribution range is from 30 Å to 103 Å for SEPA-PS-0 membrane at the “zero” operating condition state.
Spherical particles made of cellulose powder were pyrolyzed in a 4.3 cm i.d. fluidized bed over a temperature range of 350 to 550°C in a stream of nitrogen gas with the primary aim of producing tar and levoglucosan in the tar. Higher yields of tar and levoglucosan were achieved under operating conditions in which the cellulose particles tended not to sink into the bed, and secondary decomposition of tar was reduced accordingly. The highest yields of tar and levoglucosan respectively at 400°C were 63% and 32% of the original weight of cellulose. These values are much higher than those reported in previous studies of fluidized beds and fixed-bed reactors in atmospheres of inert gases. The variation with time of the concentration of gas containing carbon dioxide and carbon monoxide could be simulated satisfactorily by a kinetic model in which secondary decomposition of the tar was taken into account.
A process has been developed to recover gallium and vanadium from coal fly ash in which they exist at very low concentration in a mixture of high concentration of less desirable species. Aqueous solution containing 2.8 mg/l gallium and 35 mg/l vanadium was obtained by leaching with 1.5 mol/l sulfuric acid. These metals were then concentrated by use of a chelating resin column with a functional group of iminodiacetic type after reduction of ferric ion to ferrous ion and pH adjustment. the eluate was conditioned and passed again through the column. Gallium was concentrated to 4.0 g/l and vanadium to 26.0 g/l in the resulting eluate. This liquor was further treated to remove impurities by a solvent extraction technique, employing TOMAC and D2EHPA as extractants. Stripping solutions of 96.6% gallium purity and 94.8% vanadium purity were finally obtained.
The extraction equilibria of various amino acids with tri-n-octylmethylammonium chloride (TOMAC) were studied in the high pH range. TOMAC reacted with the anion of amino acid and OH– according to the stoichiometric ratio determined by ionic valency. The extraction equilibrium constants KA of various amino acids differed considerably and could be correlated well with each hydrophobicity scale. The highest KA value obtained for tryptophan was larger than the lowest value for glycine by a factor of about 260. These findings suggested that separation of amino acids is possible by utilizing the differences in KA values. The extraction equilibrium constants of inorganic anions which were present in the fermented broth was comparable to those of amino acids exhibiting a low degree of extraction.
Numerical simulations of RF argon-oxygen and argon-nitrogen thermal plasmas under atmospheric pressure were performed. Two-dimensional continuity, momentum, energy and species equations were solved simultaneously with the electromagnetic equations by using the SIMPLER algorithm. Dissociation and recombination rates of oxygen or nitrogen in the plasmas were taken into account in the numerical model. Distributions of the plasma enthalpy and concentration were measured with a water-cooled probe. The numerical results were in good agreement with the experimental ones. The argon-oxygen plasma has a strong recirculating eddy. O2 is dissociated completely downstream from the eddy except near the tube wall. In the argon-nitrogen plasma, the mass fraction of N2 is significant even in the high-temperature region.
In many previous investigations, the elutriation rate of multi component gas-solid fluidized beds has been reported to be in proportion to the concentration of particles in the bed. We, however, found the very interesting phenomenon that the ratio of elutriation rate to particle concentration gradually decreases as the concentration becomes larger. A simple model, in which the obstructive effect on elutriation due to suspended particles in the freeboard is taken into consideration, is proposed and applied to elutriation from a fluidized bed which consisting of two kinds of glass beads of different sizes. The virtual elutriation rate in the proposed model, which is the imaginary elutriation rate when no obstructive effect exists, agrees well with the common elutriation rate predicted by the equations proposed previously, such as that by Wen and Hashinger. The other parameter in the model, kp, which relates directly to the magnitude of the obstructive effect, decreases as the superficial gas velocity becomes high.
The filtration characteristics of crossflow ultrafiltration are studied for two types of solutes, a protein (BSA) and a colloid (silica sol), under constant-pressure conditions. It is shown that the time variation of the filtration rate coincides with that of dead-end ultrafiltration until the filtration rate drops to a certain value, and that the gel-cake is easily swept away at a relatively small crossfiow velocity. Effects of the crossflow velocity, the filtration pressure and the solute concentration on the dynamically balanced filtration rate are explained by considering the balance between the “particulate” solutes accumulating on the gel-cake surface and those to be swept away by the shear force due to crossflow at steady state. It was also found that the dynamically balanced filtration rate in crossfiow upward ultrafiltration coincides with that in dead-end upward ultrafiltration under conditions below the critical shear stress τw·c, whereas it is in accord with that in crossflow downward ultrafiltration above τw·c.
The electroosmotic dewatering of a homogeneous semisolid material is investigated. The basic differential equation for electroosmosis through compressible materials is derived by use of Kobayashi’s equation, which considers the tortuosity of a flow path and the effect of the hydraulic pressure profile in the materials. It is recognized from Kobayashi’s equation that at the impermeable upper wall the electroosmotic flow and the pressure flow compensate each other. The relation among local hydraulic pressure, solid compressive pressure and external pressure is also derived, and is the same as those in mechanical filtration and expression. It is assumed in the analysis that no electroosmotic flow occurs in the layer where the local void ratio is smaller than a certain critical value. Based on numerical calculation, the pressure flow opposed to the net flow propagates from the upper electrode to the lower electrode. The electroosmotic dewatering rates of a homogeneous bentonite clay material calculated by the present method compare favorably with experimental observations.
An on-line dynamic optimizing control procedure for the operation of a binary distillation column is proposed, and its performance was examined experimentally. A linear input-output model for a multi-input/multi-output system was employed to represent the column, and its parameters were recursively updated using the instrumental variable method. The model was used in the prediction of output process values dynamically. The optimization of profit function was carried out using the golden section search method in every sampling step while the feed flow rate and the set point of top-tray temperature were varied. For the control of column operation, an adaptive control technique using quadratic programming was implemented. A satisfactory application of the optimizing control procedure was obtained, and it was found that the procedure gives the optimal solution in a short time and is applicable to every sampling time of advanced control to maximize profit while the system is changing dynamically. Also, good convergence in the procedures of identification, optimization and adaptive control was observed.
The extraction of cobalt and hydrochloric acid from acid solution by tri-n-octylmethylammoniuin chloride (TOMAC) was studied, employing diluents with differing solvating characteristics to the quaternary salt such as benzene and chloroform. The association of the extractant in each diluent is taken into account and only the monomeric species of the extractant is considered to take part in the extraction of acid and cobalt in both diluents. A simple scheme for extraction of cobalt occurring in both diluents is presented which assumes that the extraction occurs via replacement of water in the salt hydration sphere by cobalt chloride. This scheme can express the extraction of cobalt over the whole range of cobalt loading of the extractant and the liberation of water and acid with the progress of cobalt loading from the organic solution in both diluent systems.
A mathematical model suitable for a continuous electrophoretic concentration of protein is presented. The model equation was solved by using a finite-difference numerical method with an upwind-difference scheme. A stable and satisfactory numerical result was obtained. The experimental results quantitatively, or at least qualitatively, agree with the theoretical predication, indicating that the mathematical model and numerical method are valid. The effect of various operating factors including the buffers added to the feed solution and the composition of the electrode solution has been analyzed on the basis of the experimental results and theoretical predications. The result indicates that the degree of concentration of protein depends on the pH distribution in the electrophoretic chamber. The pH distribution is affected not only by the kinds of buffers added to the feed solution but also by the composition of the electrode solutions.
Foam separation experiments were carried out by use of sodium dodecylbenzene sulphonate. In the presence of various coexisting salts, the dynamic surface excess of dodecylbenzene sulphonate ion (DES–), ΓDBS–, at the dynamic gas–liquid interface of bubble was experimentally and theoretically examined. The Davies adsorption isotherm could be well applied to ΓDBS– in the range of surfactant concentrations of CDBS–/Cm ≤ 0.15: CDBS– and Cm mean bulk concentration of DBS– and critical micelle concentration respectively. It was also found that the values of the ratio of adsorption constant to desorption constant in the Davies adsorption isotherm, B1/B2, have values in the two regions of electrical potential, φ0: in the region of –φ0 < 4.0 × 10–4 esu, B1/B2 value is 6.85 × 10–7 cm and in that of –φ0 > 4.0 × 10–4 esu, the value is 6.07 × 10–6 cm. Moreover, the effect of various coexisting salts on the surface excess of DBS– in static equilibrium determined by surface tension measurements could be thermodynamically interpreted by a two-dimensional state equation by use of reduced concentration of DBS– based on CMC. Saturated surface excess of DBS– in static equilibrium could also be correlated to the ionic strength of coexisting metal ions. Using the equation derived from both the Davies adsorption isotherm and the Gouy–Chapman theory of the diffuse double layer (G.C. model), ΓDBS– under various coexisting salts could be estimated within an error of about ±10%. Estimations of ΓDBS– in various solution systems were also carried out.
Pressure fluctuations in the plenum chamber of a solid-gas fluidized bed were measured and analyzed in order to develop a method for the diagnosis of fluidization conditions. Power spectral density functions obtained by FFT analysis of the pressure fluctuation data had three principal frequencies. These frequencies were compared with the bubble eruption frequency observed at the bed surface and the bubble generation frequency, which was obtained by an optic fiber probe installed above the distributor opening, and two of the three frequencies were found to agree well respectively with the obtained frequencies of bubble generation and of eruption. The third principal frequency was judged to be a natural frequency of the fluidized bed, since the frequency depended on the particle content in the fluidized bed and on the plenum volume. The frequency corresponding to bubble generation was observed to be influenced significantly by the natural frequency of the fluidized bed when the natural frequency had a low value.
Ionic transport in a continuous Donnan dialyzer with a parallel-plate channel and an agitated tank was studied for a bi-ionic exchange system. A theoretical model of the system was formulated on the basis of diffusion equations in terms of diffusion, migration and convection of each ion. Theoretical solutions were obtained by numerical calculations using a finite-difference technique. Simultaneously, continuous Donnan dialytic experiments were conducted for the K+–H+ exchange system with a cation-exchange membrane. The validity of this model was confirmed by comparison with experimental results for mean dialytic rates. The numerical calculation also provided distributions of ionic concentrations and the electric potential in the parallel-plate channel. An effectiveness factor was introduced to discuss the proportion of the mass transfer resistance in the membrane phase to the overall mass transfer resistance. The influence of Reynolds number and channel height on the mean dialytic rates and the effectiveness factor was also examined.
The drop size distribution in a vessel is decided by the mutual relation between energy dissipation rate, residence time of drops at a certain location, break-up rate of drops and coalescence rate of drops. Experimental results so far, however, leave us far from an understanding of the precise mutual relation. Accordingly, the authors have made correlative equations representing the drop-size distribution under various mixing conditions. It is observed in this study that a combination of three normal distribution curves gives a good fit for the volumetric drop-size distribution and that the number distribution of drops is also given by a combination of two normal distribution curves.
A fuzzy control system for a fed-batch culture was constructed and applied to coenzyme Q10 fermentation. For inputs, cell concentration ([OD] and [OD2]), specific growth rate ([SGR]) and fermentation time ([BTIM]) were selected. A base value of aeration rate ([AIRB]) and an adjustment term for [AIRB] ([DAIR]) were defined as outputs. In this system, cell concentration was measured on-line by a turbidimeter and specific growth rate was then calculated. After cell concentration, specific growth rate and fermentation time were transferred into four inputs using membership functions, and these were further transferred into two outputs by 70 fuzzy rules. After defuzzification of [AIRB] and [DAIR], a set value of the aeration rate was determined as the sum of defuzzied [AIRB] and [DAIR]. In this fuzzy control system, the aeration rate was set at two-hour intervals according to its state and the productivity was higher and had less deviation than model fermentation (predetermined stepwise shift up each 20 hours in the aeration condition). Such result was also obtained even in the case of low initial cell concentration. These data indicated the usability and stability of this fuzzy control system.
A study of soot oxidation in silent discharge aiming for the development of an efficient soot removal technique, is presented. The soot collected on the filter decreased in silent discharge by the addition of H2O and/or O2 to Ar carrier gas. In the case of a flow system also, the soot concentration was reduced with an increase of residence time in the discharge reactor. These facts support the conclusion that the soot particles were oxidized by OH radicals generated by the destruction of water vapor in the exhaust gas. To explain the soot reduction by OH radicals quantitatively, simple kinetic simulations were carried out. In this model, the electron energy in the discharge was adjusted to obtain agreement with experimental measurements of OH radical concentration as measured by resonance absorption spectroscopy. The calculated results well explained the observed trends of soot reduction when the collision efficiency of OH radicals with the soot surface was assumed to be 0.1.
An analytical investigation is made on the recovery of hydrogen isotopes to a level of very low concentration by means of a metallic particle bed. Our attention is focused on the effect of plateau pressures in equilibrium isotherm on the shapes of absorption breakthrough curves of hydrogen. A dimensionless material balance equation is numerically calculated, taking account of the effects of axial dispersion and the resistances of external diffusion, internal diffusion and surface reaction. If the equilibrium isotherm intersects the line of x = ym on the x–ym diagram, a plateau zone is found on breakthrough curves. The width of the plateau zone increases with increase of bed height, and two independent mass transfer zones go through the bed with different propagation velocities. On the other hand, the plateau zone occasionally emerges in the developing region close to the inlet of the bed even when the two lines do not intersect.