An algorithm for multistage multicomponent separation calculations was presented for the extractive distillation columns including both the first column for separating C5 alkanes and alkenes from dienes and the second column for separating alkynes and cyclopentadiene from isoprene. Based on the equilibrium data of C5 hydrocarbons in the presence of solvents, the simplest linear equation was recommended for the vapour-liquid equilibrium relations log(yi-/yip) =log(vi/viP).xs where V represents activity coefficients, Y represents y at infinite dilution, x represents the molar fractions, and the subscripts i, ip, s represent i-compo-nent, isoprene and solvent respectively. Using this equation together with the T(temperature)-xs and the Yip-xs relation data, a calculation procedure proposed by Lu Huan-zbang[1] was employed with some modifications. The computations were carried out on TQ-16 computer for the above mentioned extractive distillation columns, and the results were compared with those obtained from a model column equipped with 92 sieve-plates and 9 sampling sites along the column. The above results - indicated that the calculated and experimental values agreed quite well with each other both for the first column with as many as about 20 components and the second column with components in ppm amounts. The average deviation was 0.36% for all components of the first column, 0.20% for main components of the second column, 27 ppm for components less than 1000ppm, and 0.9ppm for those less than 10ppm. The average deviation in temperature was 2.5℃ for the first column and l.3℃for the second. It may be concluded therefore, that the vapour-liquid equilibrium data and the proposed vapour-liquid equilibrium equation as well as the computational procedure employed were reliable for the calculation of C5 hydro-carbons-DMF extractive distillation.

Based on the partial differential equation of mass transfer, one dimensional (radical) and two dimensional (radical and axial) diffusion models of the falling film molecular still are developed and solved. The results of both analytical solutions and digital computations of these two models are on the whole in agreement with the experimental data taken from literatures. Factors affecting the design and operation of a falling film molecular still are discussed.

Reactive disperse mono-azo dyes containing a β-hydroxy-ethyl sulfonyl sulfuric ester group in diazo component have been synthesized. The preparation of anthraquinone derivatives with the same reactive group are also studied. These dyes reacted chemically in weak alkali medium with synthetic polyamide fibres, such as nylon 66 and 6. The dyeing properties of both azo- and anthraquinone reactive disperse dyes on nylon 66 and 6 are being studied. The fastness properties of the dyed substrate are also examined. As expected, these dyes showed bright colours on various substrates, and the dyed fabrics have good fastness to light and wet treatment. The formation of the covalent bond between the reactive group and nylon fibre is proved and discussed. The reactive disperse dyes containing β-hydroxy-ethyl-sulfonyl-sul-furic ester group behaves the important characterictics of either acid reactive or disperse dyes for dyeing polyamide fibres. These dyes are suitable for dyeing various kinds of polyamide fibres as well as natural silk, tussah silk and crepe fabrics of silk with cellulose acetate and others.

The dawncomer is the key component of a multi-stage fluidized bed. A pneumatically controlled downcomer has been devised, consisting of a fully fluidized dipleg followed by a conical spout for upward discharge in dense phase. Such a design has been found to insure against bridging of solids in descent and against spouting of solids from a lower layer caused by gas by passing. In experimental investigations, flow characteristics are measured including the pressure drops of the different fluidized beds and of the downcomers for various gas and solids flow rates as well as the solids residence time distribution, operating either with a multiplicity of gas sources for the various downcomers or with only two gas sources. In a particular design, a multi-stage fluidized bed with dual gas supply, in which one half is incorporated centrally in the other, has been tested on pilot scale for recovering solvents fromlow-concentration C2H4Cl2 streams of a plant waste gas. Main test results obtained were as follows: adsorption rate >97% concentration of waste gas <25mg/m3 active carbon loss <0.1% This kind of conical downcomer in which the solids are either transported from a high-to a low-pressure region, or from a low-to a high-pressure region, is also used to control the solids discharge of cyclone diplegs or the feeding and discharging in other solids circulation systems.

Hydrodynamic behavior of flow-guided sieve plate (a Linde type sieve plate) has been studied. Studies were carried out using air and water in a rectangular column. To obtain information on the effects of surface tension and specific gravity of liquid on plate hydraulics, surfactants were added to alter the surface tension of water and saturated aqueous solution of common salt was used instead of water for some experiments. The experimental results about dry plate pressure drop, pressure drop caused by vapor through liquid and formation of bubbles, entrainment and weeping point are reported in detail and correlation equations and graphes for design purposes are presented. Application of the results to commercial column design has been verified by checking with plant data. Based on adequate experimental work the aeration factor β against calculated clear-liquid depth hL with hole F factor as parameter was plotted. It must be emphasized that appreciable error may be resulted if the β plot based on air-water system is applied to those systems with low interfacial tension. Thus, a correction formula for aeration factor β is specially proposed as. This formula may be used for conventional sieve plate and other type plates as well. After examining various reports about weeping point it has been found that the bases used by different authors are quite different. Considering the combined effects of the action of weeping on plate efficiency and the dependence of weeping rate on vapor velocity, it is concluded that choosing a relative weeping rate q=10% as weeping point may be suitable. Experimental results on clear-liquid height and froth height and corresponding correlation equations are also reported.

In this paper the -reasonable design on the uniform fluid distribution device for radial flow reactors is explored. The flow of fluid in the main chan nel may be described by the modified momentum equation where p, Y/g, w, , DB and x denote fluid hydrostatic pressure, density of fluid, velocity through the main channel, tube friction coefficient, equivalent diameter of main channel and length of main channel respectively. The momentum exchange coefficient k has been measured. In divided flow channel k ranges between 0.65 and 0.72, but in combined flow channel it is a little, more than unity. The hydrostatic pressure distributions in the main channels have been analyzed. According to the magnitude of friction resistance item relative to the momentum exchange item, four flow models in the divided flow channel have been presented Condition Model Momentum exchange is in control Friction resistance is in control Momentum exchange is dominant Friction resistance is dominant In the divided flow channel the tendency of variance of the hydrostatic pressure varies with the model, because the momentum exchange item and the friction resistance item have contrary effects on hydrostatic pressure. But in the combined flow channel their effects are consistant and they always cause the hydrostatic pressure to decrease. In order to reduce the difference of variance of the hydrostatic pressure between the divided flow channel and the combined flow channel in the radial reactors, it is reasonable to adopt "IT" type distribution as the. divided flow channel is controlled or dominated bythe momentum exchange. On the other hand, the "Z" type distribution would be more suitable when the divided flow channel is controlled or dominated by the friction resistance, A reasonable ratio of the cross section of the divided flow channel to that of the combined flow channel has been proposed in accordance with the various flow models. The resepective resistance coeffi-cients of the side flow through the hole for the divided flow channel and the combined flow channel have been determined. The resistance coefficients depend essentially on the ratio of the velocity through the hole to the velocity along the axis direction in the main channel and the ratio of the thickness of the wall to the diameter of the hole. The holes along the axis direction may be arranged equally or unequally according to the degree of difference in the variances of the hydrostatic pressure between the divided flow channel and the combined flow channel. Several problems such as the selection of the initial pressure drop through the hole, the catalyst seal and the axis flow inside the radial bed have been discussed. The design method has been used in various situations and the predicted results have been obtained.

Based on Tinkers flow model and the concepts of equivalent piping diagram suggested by HTRI (Heat Transfer Research Inc.), the authors, in this paper, proposed a "row by row" calculating method for solving flow ratesof WA and WE in streams A and E respectively. For practical uses, however, this "row by row" method is somewhat tedious. After some simplified treatment, and can be obtained. These results are quite satisfactory for most situations. Furthermore, a complete mathematical model for fluid flow and pressure drop calculation is given in this paper as follow: Since the K values are functions of the stream Reynolds numbers, experimental data from 14 models have been correlated to give resulting equations for resistance coefficient K of individual stream. For K&, however, the situation is much more complicated and a purely empirical correlation of K缺 has the following: form. By means of the mathematical model and experimental result, and flows rates of five streams can be solved simultaneously by trial and error.