In order to determine the total surface area of many drops falling from a lower horizontal edge of a vertical wood plate, the sizes of drops were measured on the photographs which were taken with 10^-4 sec lighting.
Experimental results showed that mean diameter of the drops d_m (or d_mv) was independent of the plate thickness (D)and liquid flow rate(L_d or L_d''). From the results obtained by dimensional anaplate thickness (D)and liquid flow rate(L_d or L). From the results obtained by dimensional analysis, the following equations were derived:
[cm] (7)
[-] (16)
where, cm, gr, sec units are used.
The ratio (Z_e) of thc total surface area of all drops existing in the space (1) under a vertical plate to the surface area of a plate of the same height (1) proved to be as follows:
[%] (29)
or, neglecting the effect of viscosity (μ),
[%] (33)
where, units are L>_d[l/m hr], l[m], and τ, ρ, μ, and g with [cm, gr, sec units].

抄録全体を表示

As there had been only a few experimental and theoretical studies made on a disc-type centrifuge in the past, experiments in separating characteristics of centrifuges for solid-liquid system were carried out, using low concentrations of fly-ash particles throughly dispersed in liquid. The particle-size distributions of spherical particles in the feed and in the clarified liquid employed in the test were measured through a microscope and the performance of the centrifuge was evaluated, as follows:
1. In order to describ the perfomance of a centrifuge based on the experimental results, two nondimensional parameters were introduced. One was a fractional penetration, as is commonly used in an aerosol collection, and the other, a non-dimensional particle diameter. In the latter, the particle diameter was determined by the ratio of a particle diameter to a critical particle diameter calculated from the factors of mechanism of separation, such as physical properties of suspension, geometrical dimensions of a disc and operating conditions.
2. The separating characteristics seemed to be determined by two parameters, and the characteristic curves are expressed as: P_f=f(D_pr)
3. As is obviously shown by the characteristic curve, the critical particle diameter proposed by one of the authors might be considered a standard for the operation.
4. The necessaary and sufficient condition for scaling-up is given by:
This condition proved to be more generally applicable than the one presented by C.M. Ambler.

抄録全体を表示

Pneumatic pulverizers have come to be used widely in chemical industries for ultra fine pulverization, but it is very hard to analyze their characteristics quantitatively. We tried, by experiment, to analyze the mechanism of a jet pulverizer, which is expected to be one of the most efficient pneumatic finepulverizers.
Thinking that in the mechanism of jet pulverization, suction of particles by jet was essential, we used a simple jet-suction pulverizer shown in Fig. (2-1), and found the relation between the suction of particles and the nozzle pressure relative to other factors, as given by Eq. (3·1). As to grinding, the relation between the rate of pulverization and the nozzle pressure was discovered as presented by Eq. (4·4), which is closely related to Eq. (3·1).
Constant kj was defined to be grindability of pneumatic pulverization, which is completely independent of sizes of particles fed, and it was found that kj was closely related to the size distribution pattern of the pulverized material and that such characteristic of the material was clearly indicated by coefficient n, which represents the degree of surface grinding. When n=2.0, the mechanism of pulverization was defined as surface grinding, and when n=0, as bulk crushing (See Fig. 4-3). We believe that this analysis of mechanism will be widely applicable to the other sorts of grinding as well.
Further, we experimented in impact jet pulverization by a couple of jets crossing each other (Fig. (7-1)) and obtained the results as shown in Fig. (7-2, 7-3). Another experiment led us to make clear the effects of secondary air on the suction of particles and on pulverization as shown in Fig. (6-2).

抄録全体を表示

There are several methods for estimating process dynamic characteristics by means of various responses, such as frequency response, indicial response, etc. The undamped cycling method which the author proposes here and which has been used by Ziegler and Nichols, in order to decide the optimum control conditions, can also give dynamic properties of a plant by making use of critical control conditions and its cycling period for the closed cycle system.
Process characteristics may generally be assumed to be expressed by Eq. (1), and the transfer function of a controller, by Eq. (2). If the automatic control system is cycling, Eq. (3) may be suggested.
Various cases supposed for each control condition are as follows:
1) Proportional Action Control, (P)
Eqs. (4) & (5), Fig. 2 and Table 1 are obtained. An approximate equation ωPL=1.7 and Fig. 3 are particularly recommended for practical uses. The optimum control conditions of ZieglerNichols are discussed from this standpoint.
2) Integral Action Control, (I)
Eqs. (8) & (9) are obtained and the ratio of critical periods of P to I action is proved to be about 2, provided T/L<1.1)
3) Other Controls
PI action gives Eqs. (10) & (11), and P and PI actions imply a possibility of solving both static and dynamic characteristics of the process by themselves with Eqs. (12) & (13).
A special example of an oil pressure controller with a jet-pipe is given by Eqs. (14), (15) & (16). Some experimental results are also presented in Table 4. This method is convenient for obtaining dynamic properties of pressure or flow rate process without any help of special instruments, becaus their characteristic time is too short to be measured by usual records of the indicial response. The necessary requirement is that the controller should be calibrated strictly and its actions purely theoretical.

抄録全体を表示