Dry Submicron Classification by a Small Blow Down Cyclone t

Koichi linoya The Asso. of Powder Process Industry and Eng., japan* Tadashi Fuyuki and Yukiyoshi Yamada Nisshin Flour Milling Co., Ltd., Technical Research Center** Hiroyuki Hisakuni and Eiichi Sue Maruo Calcium Co., Ltd., Powder Laboratory*** At present, dry powders are classified using mainly a vortex centrifuge or sieving equipment, in which the lowest 50% cut size is still greater than one micrometer. Impactor-type classifiers are capable of operating in the submicron range, but have low throughput and high energy consumption. Our new free vortex cyclones are capable of producing excellent sharp classification of dry powders in submicron cut sizes, especially with the use of a 5% to 15% blow-down flow rate from the dust collector chamber. A specified cyclone has been experimentally studied under various operating conditions using several kinds of test powders.


Introduction
Many types of particle size classifiers have been developed for the industrial market, but there is no submicron-cut sizer in practical use except impactors.However, the impactor has practical problems involving very low throughput and high energy consumption.
Our modified cyclone type can classify dry powders of submicron cut size with both large throughput and low energy consumption.The classification performance has been much improved with the use of a blow-down system, which sucks part of the gas flow from the cyclone dust catcher box_ll,Z),3l Experimental results are given under several conditions in this report.

Experimental Apparatus and Operating Procedure
Fig. 1 represents our typical cyclone which has an outer diameter of 70 mm¢, and Fig. 2 is skematic diagram of an experimental set-up.
The two-stage cyclone system (Type II) has been tested in order to obtain improved performances t This report was originally printed in].Soc.Powder Tech- nology, japan, 29, 351-355 (1992) in Japanese, before being translated into English with the permission of the editorial committee of the Soc.Powder Technology, Japan.
KONA No.ll (1993) with the blow-down flow of the 1st cyclone.The one stage cyclone system is called Type I.
The experimental procedure is as follows: The cyclone inlet flow rate is kept constant, and a test powder is fed into the cyclone inlet pipe at a constant rate.The ratio of the blow down flow rate to the inlet flow rate is maintained constant.The test powder is almost completely dispersed upstream the cyclone inlet by means of an air jet disperser.Downstream the cyclone classifier, coarse classified powder is collected into the cyclone• dust box, fine powder is also collected on a paper filter, and blowdown powder is collected on another filter.
The collected yield TJ of the fine powder is calculated using the following equation (1) where W c is the mass of the coarse powder collected at the cyclone, W 8 is the mass of the blow-down powder at the blow-down filter, and WF is the mass of the fine powder at the main filter.In the case of a 2 stage cyclone system, W c is the combined mass caught by both cyclones.
The partial separation efficiency !::,.TJ is also calculated using equation (2), considering both the collected mass fraction and the measured particle size distribution of each collected powder.The mass and size distribution of the feed powder are calculated based on the sum of each collected mass and their mass fraction (ratio), and not directly from the feed powder.where '' /:::, .'' represents the incremental mass of collected powders for every small corresponding size range.
The classification performance is also evaluated using the following equations ( 3) and (4) as the sharpness indexes.
The particle size distributions are measured mainly by a laser diffraction type (Maicrotrac SPA), and occasionally by a X-ray sedimentation type (Sedigraph 5000).

Experimental Results and Comments
The effects of the blow-down flow rate of the partial separation efficiencies of the two stage cyclone system are remarkable as shown in both Fig. 3 and Fig. 4.However, the yield of fine powder decreases with the increase in blow-down follow rate.
Fig. 4 represents an interesting effect of the inlet flow velocity on the classification efficiency.A high velocity results in a deteriorated performance for the case without blow-down and in a good performance for 15% blow-down.
Fig. 5 illustrates the effect of different powders on the efficiency of a single stage cyclone.The powder characteristics affect the performance.However, the blow-down flow improves the classification sharpness in all cases.
Fig. 6 represents the difference between the particle size distributions obtained with the 2 stage cyclone system (Type II) and with the single cyclone system (Type 1).The inlet velocities are kept at 20 m/s for Type II and 30 m/s for Type I, in order to obtain almost similar pressure losses in these two cyclone systems.The 2 stage type produces smaller particle size of the fine powder, with a nearly similar yield.However, the inlet air flow rate of the 2 stage is about 2/3 the flow rate of the single stage in this experiment, therefore the 2 stage energy consumption is also about 2/3 that of the single stage.If the powder concentration of both inlet flows is the same, the powder throughput of the 2 stage is again about 2/3 of the single stage throughput.In these experiments the powder concentration is fairly low (about 20 g/m 3 ), and can be increased up  to 100 g/m 3 , based on our past experiences.Fig. 7 gives similar results for the 2 stage system, and shows a better performance compared to the single stage of Fig. 5. Fig. 8 represents the classification efficiency dif- ference obtained by two types of measuring instruments of particle size distributions for an experiment.The X-ray sedimentation method (Sedigraph) gives usually narrow distributions and better classification results compared to the laser diffraction method (Microtrac).Therefore, these performance evaluations should be done using similar types of instruments for particle size characterizations.
Fig. 9 represents the effect of a bottom cone at the inlet of the cyclone dust chamber or at the bottom of the cyclone body.It is desirable for a cyclone to be equipped with a bottom cone in order to prevent reentrainment of the collected fine powder.
Table 1 gives the entire experimental results including the fine powder yield and the classification sharpness indexes.
Incidentally, the partial separation efficiencies /::,.' YJ involving the blow down flow decrease to zero, instead of the blow down flow rate ratio.The latter case has been obtained in many experimental condi tions of wet type cyclones.The reason for this is not clear, but might be due to strong flow turbulence and reentrainment of fine powder at the cyclone bottom.

Conclusion
The experimental results of a sub-micron particle size classification by a modified blow-down cyclone are as follows: (1) The effects of blow-down flow from a cyclone dust chamber are remarkable, and its performance is as good as that of forced vortex type classifiers for a few micrometers cut size.However, the yield of fine powders drops lower than that with the original type, and a small amount of over one micrometer size particles still remains in the classified fine powder.(2) The effect of blow-down flow is larger at higher inlet velocity than at lower velocity.This may be due to the use of the blow-down flow to prevent reentrainment of coarse particles in a cyclone.(3) The classification performance varies with the powder characteristics for the same operating conditions.(4) The performance of a 2 stage cyclone system (Type II) is substantially higher compared to that of a single stage system with the same pressure loss of the system, though the throughput is about two third that of the single stage.(5) Performance evaluation depends on the type of instrument used for the measurements of particle size distributions.(6) The use of a bottom cone at the inlet of a cyclone dust chamber seems desirable for the maintenance of a good performance.

Fig. 2
Fig. 2 Experimental Cyclone Systems (Type I and Type II)

Table 1
Experimental Results of Cyclone Classifiers