FLUID MIXING II A symposium organised by the Yorkshire Branch and the Fluid Mixing Processes Subject Group of the Institution of Chemical Engineers and held at Bradford University, 3-5 April 1984 Organising Committee: Prof. M.F. Edwards Dr. N. Harnby Dr. J.C. Middleton THE INSTITUTION OF CHEMICAL ENGINEERS SYMPOSIUM SERIES No. 89 (i) PUBLISHED BY THE INSTITUTION OF CHEMICAL ENGINEERS Copyright ö 1984 The Institution of Chemical Engineers All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without permission in writing from the copyright owner. First edition 1984 — ISBN 0 85295 171 X MEMBERS OF THE INSTITUTION OF CHEMICAL ENGINEERS (Worldwide) SHOULD ORDER DIRECT FROM THE INSTITUTION Geo. E. Davis Building, 165-171 Railway Terrace, Rugby, Warks CV21 3HQ. Australian orders to: R. M. Wood, School of Chemical Engineering and Industrial Chemistry, University of New South Wales, PO Box 1, Kensington, NSW, Australia 2033. Distributed throughout the world (excluding Australia) by Pergamon Press Ltd, except to IChemE members. U.K. Pergamon Press Ltd., Headington Hill Hall, Oxford 0X3 OBW, England U.S.A. Pergamon Press Inc., Maxwell House, Fairview Park, Elmsford, New York 10523, U.S.A. CANADA Pergamon Press Canada Ltd., Suite 104, 150 Consumers Rd., Willowdale, Ontario, M2J 1P9, Canada FRANCE Pergamon Press SARL, 24 rue des Ecoles, 75240 Paris, Cedex 05, France FEDERAL REPUBLIC Pergamon Press GmbH, 6242 Kronberg- OF GERMANY Taunus, Hammerweg 6, Federal Republic of Germany British Library Cataloguing in Publication Data Fluid mixing II. - (Institution of Chemical Engineers symposium series: 89) 1 Mixinq 2. Fluid dynamics ' institution of Chemical Engineers, Yorkshire Branch ii. Institution of Chemical Engineers, Fluid Mixing Processes Subject Group III. Series 660.2'84292 TP156.M5 Pergamon Press ISBN 0-08-031416-3 Library & Congress No. 84-14849 (ii) PREFACE Fluid Mixing II continues the series of conferences which began at Bradford in March 1981. It is hoped to hold a third in the series in 1987. The conferences are intended to cover all aspects of mixing including the assessment of mixture quality, experimental and theoretical studies of mixing, chemical reaction and mass transfer, heat transfer, novel experimental techniques, scale-up and optimisation. The authors represented include most of the leading names in the field in the UK and Continental Europe. Symposium Series No. 89 THE DRAWDOWN OF FLOATING SOLIDS INTO MECHANICALLY AGITATED VESSELS M. F. Edwards and D. I. Ellis Schools of Chemical Engineering, University of Bradford. SYNOPSIS Data are reported of the critical impeller speed required to draw float­ ing solids down into the bulk of the liquid in a cylindrical vessel of diameter 1.22 m. A variety of impeller types and baffle configurations was studied. In each case water was the test liquid and various concen­ trations of polyethylene chips were used. Measurements were taken of the power input at the minimum drawdown speed and these are used to assess the optimum impeller/baffle arrangement. INTRODUCTION The use of agitated vessels to contact solids and liquids is encountered in a number of industrially important processes. The basic objective may be to provide intimate contact between the two phases to facilitate mass transfer or chemical reaction, or simply to produce a general dis­ persion of the solid particles in the liquid prior to hydraulic transport through pipelines and processing equipment. In either case a major de­ sign consideration is the appropriate impeller speed to achieve good con­ tacting conditions using a minimum input of power. In the case of solids which are denser than the liquid, much work has been done to establish relationships between the minimum speed required for complete suspension of the solids, Njs, and the potentially large number of system parameters involved. An early, comprehensive and fre­ quently quoted study is that of Zwietering (1), who proposed the follow­ ing empirical correlation:- ϊθ.45 0.1 N. = s v d — D-°·85 X0·13 (1) P where s is a constant for a given type of impeller and system geometry. Here Njs is the impeller speed which is just sufficient to prevent the particles from remaining "stationary" on the base of the tank and various criteria have been put forward to define this suspension state in which all of the surface area of the particles becomes effective for processing. More recent works (e.g. (2,3,4)) have considered a wider range of geomet­ rical configurations than those used by Zwietering (1). In particular impeller types, the height of the impeller above the base, the size and location of draught tubes have all been studied in an attempt to achieve adequate suspension conditions with a minimum of power input. In certain effluent treatment, polymerisation and fermentation processes, it is often necessary to contact solids which are less dense than the 1 Symposium Series No. 89 liquid. This problem of the drawdown of floating solids has received very little attention, the only major investigation being that reported by Joosten et al (5). These authors found that for vessel sizes from 0.27 to 1.8 m the baffle arrangement was critical. It was observed that the optimal flow pattern for drawdown was achieved with a single baffle (one fifth of the tank diameter in width) leading to an eccentric vortex. Using this baffle configuration, a four bladed angled paddle of diameter 0.6 times the tank diameter required the least power input compared to other 2, 4 and 6 bladed angled paddles and 3 bladed marine propellers. The minimum speed for drawdown was correlated by:- = 3 b xl {2> -i~ - ° ITJ [T] for the preferred stirrer/baffie arrangement (one baffle 0.2 T/four bladed inclined paddle). Ohiaeri (6) carried out some experiments with floating solids and conclu­ ded that propellers were more efficient than pitched paddles and that the best geometry is achieved with the impellers located near tne liquid sur­ face and pumping upwards. These comments were applicable to tanks of diameter 0.3 to 0.91 m fitted with four baffles. The present study reports information on a series of tests in which the minimum drawdown speed and the associated power requirement have been measured in a 1.22 m diameter flat bottomed vessel with a variety of impel­ ler and baffle configurations. EXPERIMENTAL WORK Equipment Measurements of the minimum drawdown speed and the associated power requi­ rement were carried out in a 1.22 m diameter flat bottomed, cylindrical vessel. Up to four vertical strip baffles could be inserted at various positions around the periphery of the vessel, with essentially no clear­ ance between the outer edge of the baffle and the vesse lwall. Each baf­ fle had a width equal to one-tenth of the tank diameter and could be arran­ ged to rest on the bottom of the vessel or be raised to give a clearance of 0.305 m between the lower edge of the baffle and the base of the vessel. When the baffles were in the lowered position their wetted length was equal to the depth of liquid in the vessel, which was maintained through­ out at 1.15 m. The range of baffle configurations investigated is detail­ ed in Table 1. Agitation was provided by three impellers; a downward pumping, 3-bladed marine propeller, a six-bladed disc turbine and a simple flat paddle. The principle dimensions are shown in Figure 1. The impellers were mounted on a 4.5 cm diameter mild steel shaft which was located on the axis of the vessel and was driven by a 2.54 kW motor through a variable speed drive and a 3.98:1 reduction gear, giving a range of rotational speeds of 1.06 to 12 rev/s. The whole drive unit was vertically mounted on a thrust bear­ ing comprising a fixed base plate and a well-greased ball race. A parall­ elogram mechanical linkage transmitted the torque produced on the motor assembly to a load cell. Assuming that the friction loss in the linkage 2 Symposium Series No. 89 TABLE 1 Baffle Configurations Used The configurations used are identified by the coding m/n. m is the total number of baffles, n the number in the raised position. Number of Arrangement around Codes baffles vesselo perip hery 1 o I/O, 1/1 2 o 2/0, 2/1, 2/2 2A/0, 2A/1, 2A/2 o 2B/0, 2B/1, 2C/2 o 3 3/0, 3/1, 3/2, 3/3 4 G 4/0, 4/1, 4/2a, 4/2b, 4/3, 4/4 (4/2a - two opposed baffles rai­ sed 4/2b - two adjacent baffles rai­ sed) and the ball race is negligible, the load cell reading gives a direct mea­ sure of the torque on the shaft and hence the power consumption of the impeller. The speed of rotation of the impeller was taken from the motor shaft coupling using a mechanical tachometer. Throughout the experimental work the liquid used was water and the parti­ cles were high density polyethylene chips having a size range of 1.4 - 3.35 mm and a density of 925 kg/m3. For all three impellers used, measure­ ments of the minimum suspension speed, N, and the corresponding power s requirement, p i were made for each of the 22 different baffles config­ m n/ urations shown in Table 1 and for solid concentrations of 0.75, 1.5, 2.25, 3.0 and 3.75% w/w. All these experiments were performed with the impel­ lers located to give a clearance between each impeller and the bottom of 3 Symposium Series No. 89 the vessel of V rdof the liquid depth. In addition, N and pj_n were 3 s m measured for the propeller and the turbine using a concentration of 3.75% w/w but with an impeller clearance of 2/3rd of the liquid depth. The sim­ ple paddle could not readily be raised into this position. Procedure In each series of runs, the speed of the impeller was progressively increased from its minimum value until the minimum speed for complete drawdown was adjudged to have been reached. The determination of the minimum suspension speed was based on a criterion similar to that used by Zwietering (1). Thus, when a stagnant zone, even one consisting of only a few floating polyethylene chips, stayed on the surface for less than 3-5 seconds, the drawdown was considered to be just complete. On this basis, it was possible for the same observer to determine the minimum impeller speed for complete suspension with a reproducibility of 2 - 7%. RESULTS Effect of Concentration For each impeller the general effect of concentration on minimum suspen­ sion speed was essentially insignificant. This may be illustrated, for example, by the data obtained for the turbine, as seen in Figure 2 for a range of different baffle configurations. Very similar results were obta­ ined for the paddle and for the propeller, although in the latter case a considerable scatter of the experimental values of N was observed when a s single baffle was used (configurations 1/0 and 1/1). This resulted from the difficulty experienced in applying the basic criterion for complete suspension under these conditions, particularly at the higher concentra­ tion levels. Thus with a single baffle the solids were largely suspended by the gross vortex produced. However, to reach the centre of the vortex the majority of the floating particles had to travel around the tan kperi­ phery until the rising swell produced immediately behind the baffle direc­ ted them into the vortex. In attempting to correlate the observed power consumption at minimu msus­ pension speed, Pmin* it was found that the data were widely scattered when represented as a function of concentration for all baffle configurations. This is illustrated by the results obtained for the turbine, as shown in Figure 3. Whilst large variations in Pmin a*"e apparent, no consistent pattern may be discerned. These large variations are, however, very much reduced when the power consumption is represented in terms of the power number, Po = Pmin/P Ns D · Thus reference to Figure 4, which includes data for all three impellers and baffle configurations 1/0, 2/0, 3/0 and 4/0, shows that Po is essentially independent of concentration. Bearing in mind the relative low concentration levels involved, this result is not surprising. It does, however, suggest that the small variations in N observed at different concentration levels (see Figure 2) are suffi­ s cient, when incorporated into Po, to compensate for the relatively large variations observed for Pmin (see Figure 3). Effect of Impeller Type Of the parameters investigated, the impeller type used had a major effect on both N and Pmin· Tne influence on N is shown very clearly in Table 2 s s 4 Symposium Series No. 89 which includes the average values of N for all baffle configurations and s particle concentrations at the lower impeller clearance. TABLE 2 Mean Values of N and Associated Standard Deviation s (impeller clearance, c = V3 H) Ns r.p.m. Turbine Propeller Paddle Unbaffled (mean 129* (-) 180 (+_ 5) 93 (+_ 11) for all concns.) Baffled (mean for all concns. 126 ( + 6) 223 (+12) 120 (+ 9 ) and baffle con­ figurations) * based on single value for X = 3.75% Thus N was observed to be very much higher for the propeller than for s either the turbine or the simple paddle. Furthermore, the mean value of N for the turbine was significantly higher than for the paddle. s Typical values of Pmin f°r all three impellers tested are shown in Figure 5, which incorporates the full range of baffle configurations examined at an arbitrarily chosen particle concentration level of 2.25%. The results clearly show that the power requirement for the propeller is substantially less than for either the paddle or the turbine and that this trend is more evident as the number of baffles is increased. It is also evident that drawdown is generally achieved with lower pov/er requirements as the number of baffles is reduced. This trend is more noticeable for the paddle and the turbine than for the propeller. Effect of Baffle Configuration As evidenced by the data shown in Figure 2 which includes results for different numbers of baffles and for the baffles in the raised and lowered positions (configurations 1/0, 1/1, 2/0, 2/2, 3/0, 3/3 and 4/0, 4/4), the baffle configuration has little or no effect on the minimum suspension speed. Similar results were found for the propeller and the simple paddle. In the absence of any baffles, the propeller gave a significantly lower value of N than for any of the baffled arrangements, as shown in Table 2. s The latter also shows that there was a similar, although rather less sig­ nificant effect for the paddle. In the case of the turbine, complet esus­ pension could not be obtained in the absence of baffles except at a parti­ cle concentration of 3.75% with the lower impeller clearance. 5 Symposium Series No. 89 For each impeller the power requirement tended to increase as the number of baffles was increased and was generally greater when one or more of the baffles were in the raised position. As noted previously, however, con­ siderable scatter was observed in the basic data, for Pmirw see Figure 5. Effect of Impeller Clearance Measurements of Ns and Pmin were obtained for the turbine and propeller, but not the paddle, with the impeller raised to give a clearance of 2/3 Qf the liquid depth above the base of the vessel. A wide range of baffle configurations was examined using a particle concentration of 3.75'*. w/w. The effect of impeller clearance at this concenti iticr: I s shv ile 3 for a selection of baffle configuration-.. TABLE 3 Influence of Impeller Clearance (for X = .>.7s3 N (c = ^3 H) P . (c = L/3 H) Baffle min Impeller Configuration N (c = 2/3 H) P~~ (c = 2/3 H) s mm Turbine 1/0 1.18 1.52 1/1 1.21 1.25 2/0 1.00 0.94 2/2 1.12 1.39 3/0 1.17 1.47 3/3 1.21 1.74 4/0 1.44 2.70 4/4 1.32 2.31 Propeller 1/0 1.21 3.90 1/1 1.35 3.40 3/3 1.03 1.15 4/0 1.10 1.70 4/4 1.05 1.19 These results indicate that whilst the effect on Ns of changing the clear­ ance is relatively small there is a substantial effect on the power requi­ rement. Thus, reducing the clearance from 2/3 to 1/3 of the liquid depth increased the values of Ns and Pmin by average factors of 1.2 and 1.7 for the turbine and 1.15 and 2.3 for the propeller. From Table 3 it can be seen that increasing the clearance above the base gives the most significant power saving for the turbine when four baffles are used. For the propeller the greatest power reduction is observed for single baffle systems. Joosten et al (5) also observed a small effect of clearance upon Ns. Further, they also found a general decrease in power input with increa­ sing clearance (except for their 6 bladed inclined paddle). 6 Symposium Series No. 89 CONCLUSIONS a) Measurements have been made in a 1.22 m vessel of the impeller speed and power requirement for the drawdown of 1.4 - 3.35 mm polyethylene chips into water. For such large particles the inter-particle forces are negligible and there is no problem in wetting clusters, of parti­ cles after drawdown. b) The effect of particle concentration upon both the minimum drawdown speed and the associated power input is insignificant over the range of concentrations used (0.73 - 3.75% w/w\. c) The lowest power requirements were found using propellers and in gene­ ral reducing the number of baffles also reduced the power input. In the case of zero baffles, the power requirements were so low that they could not be measured with great accuracy. This confirms the often quoted teclinique of "pouring solids down the vortex into a vessel". However, it should be noted that since the power input in such vortex- ing situations is very low, the turbulent stresses in the fluid may not be sufficient to break up clumps of particles when fine powders are added. d) The position of the impeller above the base of the tank can signifi­ cantly influence the power required for drawdown. SYMBOLS c clearance of impeller above base d particle size P D impeller diameter g gravitational acceleration H liquid depth N. minimum suspension speed N minimum drawdown speed p . power at minimum drawdown speed Po power number s constant in Equation (1) T vessel diameter X solids concentration p liquid density Δρ density difference between solid and liquid v kinematic viscosity 7