In the present paper, a new plate-to-plate method is developed for performance prediction of ideal systems and those close to ideal ones. The sum ofmole fractions of heavy components in the top product, ∑HP, as well as mole fraction ratios of heavy components in the upper part of a column and those of light components in the lower part, vij s and μij s, are selected as iterative variables instead of individual concentrations. Two parallel iterations are suggested. The first one is for ∑HP. The mole fractions of various light components are calculated top-down towards the match plate. The mole fraction of various heavy components are calculated bottom-up towards the match plate. ∑HP is then relaxed with the total match tolerance of heavy components. The second one is for vijs and μijs. The subsequent guess of vijs and μijs is calculated from the solution of tridiagnol matrices with the phase equilibrium constants Kijs obtained from the plate-to-plate calculation. Consequently, the related loops consist of a single strong loop accompanied by a number of comparatively insignificant loops. Thus, a multi-variable system can be converted into a single variable system and the convergence can be dramatically accelear-ated. A rapid and stable convergence is finally achieved. This procedure has been successfully used for complex columns with multifeed and multi-side-stream .

In the previous paper, a new plate-to-plate method was successfully used by the proper selection of iterative variables for performance prediction of ideal systems. In the present paper, similar method is used for calculating extractive distillation processes. The mole fraction of solvent in the bottom product XEB, as well as mole fraction ratios of components to be seperated on various plates vijs are selected as iteractive variables. An outer loop iteration and an inner loop iteration are involved in. The outer loop iteration is for the solvent. The mole fraction of solvent is calculated from the bottom upward and is matched at the top; XEB is then relaxed with the match tolerance. The inner loop iteration is for the components to be seperated. The subsequent set of vijs is calculated using the program proposed in the previous paper with ralative volatilities αijs obtained from the stagewise calculation. The inner loop iteration is carried out with fixed αijs and much less computer time is needed. The outer-loop consists of a single strong loop together with a number of comparatively insignificant ones. Therefore, this multivariable system is transferred into a single variable like system as mentioned in the previous paper.

Isothermal effectiveness factors for slab, cylinder and sphere shaped catalysts with uniform or nonuniform intrinsic activity profiles have been investigated. In the case of zero-, first- and second-order kinetics, the effectiveness factors of pellets with increasing activity towards the pellet surface are larger than that of uniform active catalyst, and they are proportional to the square root of the activity at the pellet surface with significant diffusion effect. The effectiveness factor-Thiele modulus curves which are valid for both uniform and nonuniform catalysts have been obtained with the Thiele modulus modified by equivalent thickness of effective layer of the catalyst. Thus, the effectiveness factor for nonuniform active catalyst could be predicted with a maximun deviation of 5% in the case of significant or insignificant diffusion effect, but 10% in general.

In this paper, the superposition rule of the residence time distribution functions for the general system having multiple inlet and outlet streams, Eq. ( 6 ), has been proved rigorously. For the cascade vessels system where processed material in each stage is not ideally mixed in various degrees and the volumes of different stages may not be equal (Fig. 3), the overall residence time distribution function E(t) and each Eij(t) of the flow system, Eq. (10), are derived. If the volumes of different stages are equal, v1 = v2=…=vn, the processed material in each stage is not ideally mixed in the same degree, β1= β2=…=βn and ε1 =ε2=…=εn, the Eq. (10) of this paper can be reduced to Eq. (17) of previous paper[1]. (Please pay attention to some print mistakes in Eq. (17)there, see footnote on page 340 of this paper). The application of Eq. (6) or (10) to the flow system having multiple inlet and outlet streams, the flow system having only one inlet and one outlet stream, as well as the multiple phase system is discussed.

Twelve narrow fractions of uranium ore sand and slime with average diameters dp= 32-460μm(35 to 600 mesh) and densities ρs = 2.58-2.63g/cm3 were fluidized at ambient temperature with water. Bed expansion relation for flui-dization was determined experimentally in the range of voidage ε = 0.530-0.957. The obtained curves for IgRe versus 1 + lgε are all nonlinear. Using two strai-ghtline sections, we approximated the curves, and determined two corresponding terminal velocities vt1 and vt2, and two voidage exponents n1 and n2, as well as bed voidage εc at the intersecting point of the two straight-line sections, the minimum fluidization velocity Umf and the bed voidage εmf at incipient flui-dization for each fraction. By regressive analysis the empirical formulas were obtained for calculating the above fluidization parameters in the range of Archimedes number from 0.6 to 1700. By comparing spherical particles with some non-spherical particles, the fluidization characteristics and relationship for the irregular particles of the uranium ore are discussed.

A new general correlation for predicting average heat transfer coefficient of film condensation has been developed as follows: It is rather simple and convenient in application. The results calculated by this correlation are well in agreement with the experimental data in the literature.

Is this article, a new mathematical procedure for calculating the rate constants(k2/k1, k1)in 2-step series-parallel second-order reactions is presented. Under specific conditions, when k2/k1 is equal to one of 17 specific values and initial molar ratio of the reactants lies between the lower and upper boundary values, a set of exact equations for the determination of k1 are derived, in other cases, approximate equations or approximation method are proposed. All the exact and approximate equations can be covered by a equation. This procedure can be extended to derive multi-step series-parallel reactions. Finally this procedure is applied to the kinetic study of sulfonation of polysulfone and satisfactory results are obtained.

By analysing heat transfer, mass transfer and fluid-dynamics of TLE, a mathematical model for operation of TLE in ethylene plant is proposed. It can be used for process analysis and disign of TLE.

Based on the boundary layer theory and the assumption that the torque by the wall of an agitated vessel is approximately equal to that of the rotating impeller a method to calculate the apparent viscosity of non-Newtonian fluids within the resistant layer of heat transfer in agitated vessel has been proposed. By incorporating this method to the Yuji-Sano s correlation, an unified correlation of heat transfer coefficient for both the Newtonian and non-Newtonian fluids in agitated vessels equipped with different types of impellers is obtained as follows:

The capillary viscometer is often used for viscosity measurement. However,in measuring of viscosity of settling suspensions, the suspensions must he agitated rather vigorously before entering the capillary to get a uniform concentration of slurry. The viscosity thus measured is always in error. A correction term has been introduced in this work to account for the velocity head due to agitation. Viscosities of ammonia solutions and slurries of laterite ore and ammonia solution were determined satisfactorily with correction term mentioned above. The effect of temperature and solid concentration on viscosity was studied. Results indicate that the laterite ore slurry with volume concentration of solid particles less than 8.45% can be considered as Newtonian fluid. The relationship between its viscocity and volume concentration can be expressed by Vand equation: μ/μ0=exp(bcv).