INDUSTRIAL SECTION

Spouted Bed Drier’

W . S. PETERSON2

The first commercial spouted bed driers have been Les premieres unites commerciales de sechage
installed in Canada and one has been used to dry par technique “Spouted Bed” ont 6th installees au
three types of peas, lentils and flax. The data obtained Canada et l’une d’elles a 6th utilisee pour le sechage
from this drier have been compared and correlated de pois, de lentilles et de lin. Lea valeurs obtenues de
with previous drying studies in various laboratory ce sechoir ont 6th compardes avec des etudes ante-
spouted bed driers. An empirical relationship between rieures Bur sechoirs de laboratoire operant sur le
solids temperature and other variables has been meme principe. Une correlation empirique entre la
obtained: temperature des solides et lee autres variables a 6th
obtenue:

where T , and T., are respectively the particle and inlet

air temperature, d,, the particle diameter, D, the oh T , et T, sont respectivement lea tempkratures des
column diameter, F the feed rate and M I the feed particules et de l’air B l’entrhe; dr,, le diambtre dee
moisture content. particules; D. le diambtre de La colonne; F le taux
High drying capacity was achieved through the d’alimentation et M, l’humidit6 dans l’alimentation.
use of high air temperatures with the result that a Une forte capacit6 de sCchage a 6th obtenue en utili-
two foot diameter spouted bed heater (plug cooler) sant den temp6ratures d’air Blevbes. I1 a 6th possible
dried almost two tons per hour of peas through 8.8% de d c h e r de 8.8% d’humiditk pres de deux tonnes
moisture range. No damage was evident in materials par heure de poie dam un s6choir ayant 2 pieds de
dried. diamitre. Aucun dommage n’a 6th observe sur le
matdriel ainsi s6ch6.

I n 1951 after a wet fall in the Canadian West, large quantities

of wheat were harvested with moisture content well above
that considered safe for storale. A t the Lakehead terminal
recycled to the top. The air stream enters at a high velocity
and travels partially up the spout, but also spreads through the
elevators alone, about I10 milfion bushels were dried. There void space of the annular moving packed bed so that in all but
appeared to be a market for a small compact high capacity shallow beds most of the air moves up the annulus. Continuous
continuous drier for farm or small elevator. A spouted bed for feed and overflow of particles are possible.
contacting fluid and coarse particles was developed by Mathur Commercial Drier
and Gishler(1) following preliminary experiments on fluidizing
wheat. Results obtained by them with a one foot diameter T h e commercial drier had a 2 ft. dia. heater and a 3 ft. dia.
spouted bed wheat drier were published in 1955(2).Recently cooler side by side. T h e heater contained a 7 ft. deep spouted
~ ) used a ri orous mathematical treat-
Becker and S a l l a n ~ (have bed, while the cooler contained a moving packed bed. Light
B
ment for the spouted bed drying o wheat, where the drying
rate is controlled by moisture diffusion within the particles.
gauge galvanized steel was used in the construction. A 20 in.
pipe was used for the burner shell. A flow diagram is shown
Publications have also appeared on heat transfer(’! b), wood chip in Figure 2. W e t feed flowed through an adjustable slide from
a storage bin to an enclosed jog trou h and into the side of the
* drying(eo7),coal carbonization(8)and on the flow characteristics
and relationships in the spouted b e d ( g J o p 1 1 J 2 ) . W o r k has also B
heater. Diametrically oppositc the ecd inlet, the hot partially
been done on drying of barley(l8) and peas(l4’. As a result of dried product overflowed to the cooler where further drying
the latter work, the first commercial spouted bed pea driers occurred, with countercurrent air flow. T h e dried product
have been installed in Canada. One of these driers was used to flowed through a regulating orifice at the bottom of the cooler
obtain data on drying of peas, lentils and flax. to a bucket elevator and thence to storage. Fines entrained in
the moist air leaving the spouter were removed in a cyclone.
Dercriptlon of spouting T h e air for the spouter was supplied by a centrifugal blower
Flow pattern of partieles in the spouted bed is shown in and was heated by direct combustion of natural gas. T h e
Figure 1. T h e particles in the central core are lifted by the installation is shown in Figure 3.
high velocity inlet air jcr. As the particle velocity decreases
at the top, the particles fall onto the annular part of the bed. Operation
T h e particles in the annular section form a moving packed bed T h e unit was started by turning on the air blower and adding
with flow mainly downwards, but also laterally towards the feed until a bed was established. T h e burner was then started,
center, until the particles are once again lifted in the air jet and air flow readjusted, if necessary, and when the bed reached
................................................................................
. ,
operating temperature the feed was restarted. Bed temperature
'Manuscript received April 1; accepted M a y 31 1068.
was controlled by a thermostat which operated a modulating
*DfvWon of Applied Chemktw, Nntional Reaea;ch Coundl, Ottawa 2, Ont. valve on the burner. There was also a high temperature cut-off

226 The Canadian Jaurnal of Chemical Eniineering, October, 1962

 INDUSTRIAL SECTION

DRIED
UOWER PCAS

Figure 2 -Flownheet of commercial

pea drier.

Fipre 3 -Commercial pen drier

installation.

Rerub8
Attrition: During drying, a considerable amount of clcan-
ing took place due to attrition and entrainment, causing loose
hulls (on peas), dirt and mold to escape overhead to the cyclone
on the commercial drier. Ihsting rate could not be rncasurcd
on the commcrcial unit but it was estimated at 1 % or less, or
about the same as latxmtory driers. I t is noteworthy that
spouting appcars to be no morc violcnt than normal hulk
handling proccdiircs for these products.
Damage
The maximum particle tcmpcrattlrc which can be allowed
with these materials is dependent on their end use. On the
small scalc drying tests with pcas to bc used as food in soups
etc., no damage was cvidcnt cvcn in shallow twds below
Figure I-Spouted bed flow pattern. 190-200"~~.. hut at 220°F. the peas were scorclrcd acid shriveled.
However, in the same sized drier, the baking properties o f
wheat were affected by high inlct air tcinpcraturcs in con-
thermostat in the bed. Bcd levcl in the spouted bed was self junction with shallow beds (12 and 24 inchcs) resultiiig in
regulating at the height of the overflow opening. kernel temperatures of 124" and 129'1:. as measured in I k w a r
When the cooler was partially filled by overflow from the flasks(*). Becker and S a l l a n ~ ' presented
~) a correlation txtwcen
spouter, the cooling air was started and as the operating bed moisture content and critical grain tcmpcraturc above which
depth was reached, the bottom slide valve was opened and the the baking qualities of wheat arc thermally injured (ca. 1W"F.).
elevator started. There was considerable variation in cooler T h e relation is not conscrvativc for shallow I d s combined with
bed hcight with the manual setting of this valve. high inlet air temperature and so should be used with caution.
They also r e p r t that thc critical tcmpcrature for the prcscrvation
Labor used, except at start-up and shut down, on the dricr of viability of wheat as seed grain would bc cxpcctcd to 1~
itself was negligiblc. However, with thc box-pallet type of 20-30'F. lower than for preservation of brcadniaking qualities.
storagc used the labor required for this installation was 54 to 1 Germination tests following hark), drying, i n a 6 in. dia. colunin
man per shift.
with 18 in. deep bcd, show a inasiiiiuni safe avcragc kcrncl
temperature of I ZS°F.(*3).It is apparcnt that damage is dcpcndciit
Data on at least timc, temperature and moisture content.
Operating data were taken after conditions had stabilized. The seed pcas (Table 1) )\.ere dried safely with avcragc
Procedures were similar to those used in earlier studied2). kernel temperature of 114'' and 1 1 3'1. even though thc bcd
Pertinent temperatures were measured and air flows estimated temperature rose to 132°F. for a short timc at the end of a run.
by means of a calibrated pitot tube. Samples of feed, heater No damage was evident in the other inaterials dried in this dricr.
product and cooler product were taken at intervals for moisture
determinations. T h e dried product was weighed in 4 ft. cubical Heat Balunce
boxes and the fccd rate calculatcd from this. T h e rate of gas The heat balance for one run in the commercial dricr is
consumption in the burner was estimated from spot-timing of shown in Tahlc 2. Agrccnicnt tKtwcen input and output is
the calibrating dial on the utilities gas meter making possiblc excellent. In gcneral, howcver, heat halances at lower heat
a rough check on heat balances. The range of data collected inputs did not close as well as did this one due to inaccuracics
from the commercial drier is shown in Table 1 . involved in obtaining air flow, kernel temperature and avcragc
For correlating purposcs other spoutcd bed drying data have inlct air tcmpcraturc. 'Ihc latter mcasurcment was niadc
been included(2a13**4). difficult by rapid cycling of the burner under sonic conditions.

The Canadian Journal of Cheniical Engineering. October. 1962 227

 _-
I
Moisture u/o d.b. Heater Temperature O F . Air Flow - cfm.
Feed rate ____-_____
I I
~

Ib./hr. wet
Feed Product Inlet .4ir Particle f Ieater Cooler

A. Peas 3980 31.0

I
22.2
I 543
I
172
I
1210
I-
1310
Lentils 297.3 24.6 16.6 566 156 674 750
Flax 2226 34.7 13.8 , 836 153 730 183
Yellow Peas 3293 68.3 47.3 570 137 1390 585
2nd pass 3508 46.6 33.0 505 166 1310 585
Seed Peas 2729 26.7 22.6 . 257 114 1350 627
2nd pass 4462 22.7 20.9 255 113 1430 627

TABLE2 TABLE3
OVERALL - FIRSTPASS WET YELLOWPEAS
HEAT BALANCE EFFECTIVE PARTICLE DIAMETERS

-
B.t.u./hr. yo of Total i n 1 d,,
inches
I ahrn
inches
1 L"-.
inches
--
$
Heat in
Yellow Peas 0.297 0.295 0.292 0.96
Spouter - air 748,000 95.7 Seed Peas 0.294 0.295 0.303 ca.0.75
Cooler - air 3,200 4.3 A. Peas* 0.266 0.266 0.290 0.99
A. Peast 0.263 0.262 0.244 0.99
75 1,200 100.0 D. Peas' I 0.236 I 0.236 I 0.231
~~. 1 .oo
I I
~~ ~ ~

Heat out Barley 0.181 0.184 0.72

Lentils 0.174 0.179 0.238 0.81
Spouter - air
evaporated water
167,000
42 1,800
--
22.2
56.1
Wheat
Flax I
I
0.136
0.088 1
I
0.141
0.103 I
I
0.170
0.136
0.94 .
0.79
Total out of Spouter 588,800 78.3
*12 in. laboratory drier
Cooler - air 58,100 7.7 t24 in. commerrial
evaporated water 41,300 5.5
exit peas 65,400 8.7
__
Total out of cooler 164,800 21.9
Total out of spouter
and cooler 753,600 100.2
Spou ter Cooler
-~ __ _____
Material Column Air inlet Bed Column
Correlation diameter, diameter, depth, diameter, lief.
inches inches inches inches
Becker and Sallans have published a series of papers ___-~____--__________
concerned with wheat drying and have presented what appears

1 1
to be a satisfactory approach to the drying of wheat in a spouted Wheat 4 % 30 none -

I
used
bed"). Their method was attempted on the materials and columns Rarley
listed in Tables 3 and 4 which include not only the commercial
drier but smaller laboratory columns in which much preliminary
Wheat
Peas
Various peas*
tS 2 E
2
1$:8 9215
2 4 & 16 1 2 & 18
work had been done. Diffusion coefficient could be estimated Flax*
only for peas(") dried in 12 in. column and even here the data Lentils*
were not extensive enough for good accuracy. T h e expression

D' = 5 5 0 e x p (-;y).
- .. . . .., *commercial drier

was obtained and it appears that effective surface moisture T h e kernel temperature is shown plotted against
contents during drying in this column were close to 0.10 Ib./lb.
dry basis as found by Bccker for wheat. However the diffusion TAO63d,,O L1Dc0.28

equation presented by Becker and Sallans did not give drying P ''hfP.S@
rates close to those found for peas, even when used with Equa-
tion (1). Neither could the data be fitted to an empirical in Fi re 4. A straight line fits the data with an average deviation
equation of similar form. of 5& (max. 18%). Substitution of arithmetic for geometric
An expression similar to hlathur and Gishler's(2) was thus mean diameter makes little difference to the fit but changes
used to correlate the kernel temperature with other drying slightly the slope and intercept of the best fitting line.
variables. Exponent values were slightly changed and the cofumi Due to orientation of particles in the annulus (long axes
diameter and particle diameter were necessarily added as vari- normal to column axis), the particle characteristic dimension
ables in order to correlate the data available. Table 4 lists the I/', as recently proposed by' Pasternack and GauvincLS) for
materials and columns from which the data were obtained. \ single stationary oriented particles, was tried in the correlation.
The effective particle diameter has been given various defini- Normal and parallel orientation of major axes to air flow were
tions in the literature. Different effective particle diameters used. T h e parallel orientation fitted the data better than did
were tried and the geometric mean was used because it correlated the normal, but scatter was much greater than with the geometric
well and it was easy to obtain. mean diameter.

228 The Canadian Journal of Chemical Engineering, October, 1962

 INDUSTR'IAZ S E C T I O N
100 1

0 24111COMMERCIAL DRIER-PEAS
OF -FLAX
UL -LENTILS
0 IZIn - ARTHUR PEAS
0 12111 - DASHAWAY PEAS
A 6 1 n - BARLEY
A 12In - WHEAT
- 0

I
/
0 4 In PRELIMINARY-WHEAT-
NOT USED FOR CURVE
FITTING

O ? 10.0 O
/>
P

0 36in COMMERCIAL COOLER PEAS

1.0 O L 36111 w LENTILS
O F 36111 11 " FLAX
0 IZIn COOLER.PEAS
' 0 lain 11

0 8 in * -BARLEY
A 9in y -WHEAT
0 I5ln - #

Figure &Kernel temperature correlation for spouted bed

Ilea ter.
t
0 24m COMMERCIAL DRIER-PEAS Figure &Cooler vaporization correlation.
OF -FLAX
OL -LENTILS

0 121" - DASHAWAY PEAS seed pcas, whilc for coolers used in the laboratory the proportion
0 Sln BLRLEY
- WHEAT was as high as 50%. T h e inlet air temperature to the commercial
~

A 12ln
0 4 In . PRELIMLNARY-WHEAT-
NOT USED FOR CURVE
cooler was 32'F. and to the laboratory coolers was 70'to 80'1.'.
FITTINO
I t would bc expected that the aniount of water evaporated
10 0 (per hour) would depend on such things as inlet particle temper-
ature, feed rate, residence time, air to solids ratio, moisture
content and particle dianieter. However, from the data available.
only the kernel temperature and feed rate affected evaporation.
Figure 6 shows W , plotted against 7;F. It should bc noted that
for the laboratory p i drying experiments If', was found to be
directly proportional to 7'JtCo nrlio where I, was the wcight-
average residence time of the particles and JZ was the air to
solids weight ratio.
Only about 2S% of the sensible heat in thc particles was
A U - d - ~ L - I
L L L utilized in the commercial unit to cvapnratc moisture due mainly
'01 10' 10' Ion
to excessive air flow, low temperature of inlet air and perhaps
MFTA F to matdistribution of air. In the laboratory 25 to 50% of thc
Figure 5-Spouter vaporization correlation. heat was utilized to evaporate moisture from peas and S O to 80%
for wheat.
T h e amount of moisture evaporated in the various spouters Capacity and Efficiency
is plotted against feed moisture, inlet air temperature and feed T h e efficiency (B.t.u./lb. water evaporated) for the com-
rate in Figure 5 . Spouter operating conditions and capacity can mercial drier varied from 1800 for the yellow peas, to 3800 for
be determined with the aid of Figures 4 and 5 . Air flow required the seed pcas.
for spmting can be approximated from the following Icquation - T h e capacity (Ib. watcr evaporated/ft.a/hr.) in this drier
(LW:

U,,, = ( D , / D , ) (D,/D,)"S
[
2gL
(P,
~
p, PrJ]''* .. . . .. .(2)
varied from 1.5 for seed peas to 7 for yellow peas (including
freeboard above both beds). Typical values for farm batch
driers is about 0.5 and for rotary driers from 2 to 6. .Iconven-
tional cascading drier for drying pc?s would appear to have a
Becker(10) has also presented a method for finding minimum very much lower capacity than this.
spouting velocity which is perhaps applicable to larger columns
than is Equation (2). Cost
A sample calculation of a spouted bed wheat drier by Ueckcr T h e cost of drying at the rate of 1 ton peas per hour through
and Sallans(3) gave a feed rate of 1 5 0 0 Ib./hr. while using the 8.8% range was estimated at 0.2 cents per pound using the
present Figure 2 at a wheat temperature of 142'1;. gave 1460 method of Lapple, Clark and Dyhdal(16). Operating costs
lb./hr. account for more than 60% of this cost and utilities (gas and
power) aceounc for 15%. Sixty-five per cent of the original
Cooler cost of the unit was for blower and burner. A cascading type
In the commercial unit the proportion of drying in the cooler of drier was estimated to cost up to three times the cost of this
varied from 9% of the total, for wet yellow peas, to 30% for installation, while cost of drying should be about 0.2$/lb.

The Canadian Journal of Chemical Engineering, October, 1962 229

 TKe same blower and burner (with 3 ft. dia. spouter and
4 ft. dia. cooler) should dry 6750 Ib./hr. of wheat through 4%
moisture range with a cost of about 3%.ip/bushel.
U,
w,
wb
--minimum superficial fluid velocity for spoutin; ft./sec.
weight water evaporated, cooler, Ib./hr.
= weight water eva orated, heater, Ib./hr.
pj = fluid density, 1b.jt.a
Sise
The drying installation occupied a floor area 8 X 25 ft. and
p,
fi = sphericity -
= absolute density of solid, Ib./ft.*

surface area of sphere having same volume as particle

had a height of about 18 ft. exclusive of duct work and cyclone. surface area of particle
A cascading type of drier would have been about twice this
height. Rejerencer
(1) Mathur, K. B., and Gishler, P. E., A.1.Ch.E. Journal, 1,
Conclusion 157 (1955).
Results of drying several field crops have been correlated (2) hlathur, K. B., and Gishler, P. E., J. Appl. Chem. (Lon-
for spouted bed driers of different sizes. High drying rates don), 5, 624 (1955).
have been achievtd by use of high air temperatures. Drying (3) Becker, H. A., and Sallans, H. R., Chem. Eng. Sci., 13, 97
cost has bccn estimated at O.Z$/lb. of feed. (1961).
(4) Klassen, J., and Gishler, P. E., Can. J. Chem. Eng., 36, 12
Acknowledgement (1958).
Acknowledgment is made to T. W. Boyd and L. Pageau for work
(5) Ghosh, B., and Osberg, G. L., Can. J. Chem. Eng., 37,
done on the pea drying. Thanks is also due other members of the 205 (1959).
Chemical Engineering rection for assistance. (6) Cowan, C. B., Peterson, W. S., and Osberg, G. L., Pulp
and Paper Mag. (Dec.), 139 (1957).
Nomenclature (7) Cowan, C. B., Peterson, W. S., and Osberg, G. L., Eng. J.,
d,, = geometric mean particle diameter, inches 41, Issue No. I, 60 (1958).
D = diffusion coefficient, Equation (l), ft.*/hr. (8) Buchanan, R. H., and Manurung, F., Brit. Chem. Eng., 6,
D, = column diameter, It. 402 (1961).
Di = inlet diameter, Equation (2),ft.
D, = particle diameter, wheat, average of small diameters, (9) Thorley, B., Saunby, J. B., Mathur, K. B., and Osberg,
average screen opening for isometric, Equation (2), f t . G. L., Can. J. Chem. Eng., 37, 184 (1959).
F = feed rate, wet, Ib./hr. (10) Becker, H.A., Chem. Eng. Sci., 13, 245 (1961).
= 32.2 acceleration due to gravity, ft./sec*. (11) Madonna, L. A., Lama, R. F., and Brisson, W. L., Brit,
!, = bed depth, Equation (2), It. Chem. Eng., 6, 524 (1961).
particle surface area (12) Madonna, L. A,, and Lama, R. F., Ind. Eng. Chem., 52,
L“-. =u max. perimeter , ft. 169 (1960).
hl, = feed moisture, dry basis, % (13) Saunby, J. B., Unpublished-N.R.C. report Mch. 14, 1958.
R = air to solids weight ratio (14) Peterson, W.S., Unpublished-N.R.C. report No. C-95-
R‘ =c gas
- constant, R.t.u./mole O R . 59P.
(IS) Pasternack, I. S., and Gauvin, W. H., Can. J. Chem. Eng.,
weight average residence time in cooler (weig; bed), 38, 35 (1960).
min. (16) Lapple, W. C., Clarke, W. E., and Dybdal, E. C., Chem.
heater inlet air temperature, O F . Eng., 62, No. 11, 177 (1955). ,
particle temperature, O F .
particle temperature, Equation ( l ) , OR. * * *

CORRECTION. In the August, 1962 issue, page 178, Figure 1

illustrating the paper “A Method to Control Distillation of
Liquids Under Vacuum” by W. M. Conners was inadvertently
run upside-down.

230 The Canndian Journal oj Chervcical Engineering, October, 1962