In the case of the flame in which gas and air flow upward through separate tubes, it was confirmed that: (a) The greater the port velocity of air and the width of air port with a constant gas velocity, the shorter the flame. (b) The burning velocity of gas is also governed by the rate of mixing of gas and air. Therefore, from the same consideration as before we can easily obtain the equation expressing the mean concentration of combustible gases.
By using the experimental formula representing the relation between the length of the flame and the gas port velocity, air port velocity, diameter of gas port and the width of air port, the expression for the mean concentration of gases in the flame was simplifyed. Thus the heat of liberation in the flame can, easily be determined. Combining the value so obtained and the heat transmitted from the flame, the flame temperature is calculated. A comparison of the observed temperature indicates a fairly good agreement with the calculated value.
In order to determine the effect of the direction of flame on the burning velocity, the study of the horizontal flame was carried out. It was confirmed that the rate of combustion of gas is also governed by the same basic principles as before. Hence, on the base of the same procedure as described in the vertical flame the concentration of combustible gases in the flame and the flame temperature were calculated.
In burning coal gas in a small furnace, the flame length decreases with the increasing amount of air and constant gas velocity. With the constant aeration, however, it is almost indepent of the amount of gas, that is, even if the amount of gas increases with the constant aeration, the length of flame does not change.

View full abstract

Experimental work on liquid-liquid spray columns is presented, which includes 56 runs in 2.92cm. and 5.64cm. I.D. glass towers, where the mass transfer does not occur. The behavior of the dispersed phase projected from the various nozzle tips is observed over a wide range of velocity at the tip (5-200cm/sec), and the hold-up is measured under the following conditions up to the flooding.
Systems: Four kinds of mineral oils and water
Dispersed phase: Both oils and water, alternatively
Tip dimensions: 0.50, 0.90, 1.50, 3.05 and 5.00mm. I.D., 1, 2, 3, 7, and 13 points. (No-tip plate type distributor is also used.)
Flow rates (Superficial velocity based on the tower cross-section):
Dispersed phase (U_D): 2-60m³/m²hr
Continuous phase (U_C): 0-110m³/m²hr U_C/U_D: 0.3-36
Viscosity of both phases (μ_C and μ_D): 0.8-300c.p.
Density difference (Δρ): 0.075-0.117gr/cub.cm.
Interfacial tension (σ): 34.0-52.7dyne/cm
Temperature: Room temperature (7-31°C)
The results obtained here are summarized as follows: 1) Distributor nozzles and projecting velocities.
The observed drop formation is classified into three stages according to the projecting velocities of the dispersed phase. The first stage is in the range of We=au_D²ρ_D/σ<2, where the drop forms one by one at the tip. The second stage, where the uniform drops form from a stream projected, is in the range of u_D<100cm/sec. In the third stage, u_D>100cm/sec, the size of drops become non-uniform probably due to the turbulence in the nozzle tube. It is also observed that no-tip plate type distributor is not satisfactory, because the dispersed phase tends to wet the surface of the plate, and forms large irregular drops. The large tip diameter (a=0.50cm) is also unsatisfactory, since the flooding occurs at low velocity.
2) Hold-up
The percentage hold-up, ε, is generally correlated with some demensionless groups as shown in Fig. 4-Fig. 6. The ordinate of Fig. 6 is (1+U_C/U_D)^0.4(1-a/D)²ε, and the abscissa is (1+UC/U_D)^0.5(U_D/√ga) for √β>25, and is (2.3/β^0.13)(1+U_C/U_D)^0.5(U_D/√ga) for √β<25. The curve AA represents average values of all data obtained here, which include the range of U_C/U_D=0-36, √β=0.3-400, u_D<100cm/sec., and U_C<110m/hr. For the large tip diameter (a=0.50cm), the data of au_Dρ_D/μ_D>1500 are omitted in Fig. 6, because they are higher than the curve as shown in Fig. 4 and Fig. 5. For comparison the data of Johnson and Bliss are also plotted in Fig. 6.
3) Flooding
Although all of the flooding data are plotted near the right end of the curve AA in Fig. 6, the better correlation will be obtained if the exponents of (1+U_C/U_D) in the coordinates, 0.5 and 0.4, be replaced by 0.6 and 0.3, respectively, so far as the flooding phenomenon concerned. Thus, for the safe operation the value of (1+U_C/U_D)^0.3(1-a/D)²ε should not be greater than 20%, and the value of (1+U_C/U_D)0.6(U_D/√ga) or (2.3/β^0.13)(1+U_C/U_D)^0.6 (U_D/√ga) should not be greater than 0.08.

View full abstract

Thermo-compression system (Fig. 3) is, of course, much more economical than single or multiple effect (Fig. 1 and 2) according to the theory of thermo-dynamics. But when evaporation take place at low temperature, high vacuum, or large temperature difference, it is very difficult to use the thermo-compression system practically. In this case when using medium by indirect thermo-compression system (Fig. 4), we can easily accomplish the purpose under these bad conditions.
For example, using trichloroethylene (C₂HCl₃) or dichloromethane (CH₂Cl₂) as medium, the pressure at compressor comes near the atmosphere (Fig. 5), the head and suction volume become very small (Fig 6 and 7), and the compressor is simlified, so that we can use this method practically. Fig. 8, 9 and 10 are examples of this system.

View full abstract

In the previous paper [Chem. Eng. Japan, 14, 148 (1950)] a method was presented for obtaining path curves on psychrometric chart in dehumidification and water cooling, and calculating, the transfer units by the graphical integration based on absolute humidity, temperature and enthalpy under the assumption that the water rate change is negligible thronghout the apparatus. This paper presents the more precise method for determining path curves on psychrometric chart in these cases, without the above approximate assumption. A calibration method is presented for this graphical procedure, and a method is proposed to determine in dividual coefficients in these cases by means of a single experiment of humidification, dehumidification or water cooling.

View full abstract