Analysis of Continuous..

flow Grain Dryers

F.W. Bakker-Arkema*, M.D. Montross*, Liu Qiang* and D.E. Maiert

Abstract
The high-temperature continuous-flow dryer is the prevalent dryer type in the major grain-producing countries. The choice of a particular model is frequently based on the initial cost, rather than on technical factors
sucb as energy efficiency and grain quality. This has led at times to the employment of low-quality dryers,
and to the produetion of inferior-grade grain and the consumption of excessive fossil-fuel energy. This paper shows that the dryer-manufacturing industry currently markets reasonably-priced, energy-efficient dryers which are able to produce excellent quality grain.
Crossflow, mixed-flow, and concurrent-flow are at present the primary high-temperature dryer types.
Simulation modelling is routinely used today in the industry for analysis and design, resulting in their improved grain-quality and energy-efficieney characteristics.
The modem erossflow dryer is suitable for drying maize as feed; it is less expensive than mixed-flow and
concurrent-flow dryers. For the drying of rice (and food maize), mixed-flow and concurrent-flow dryers are
recommended because of their superior grain-quality characteristics. Of these, the concurrent-flow models
have, in general. the best energy efficiency.
Several unconventional high-temperature dryers are occasionally used commercially for the drying of grains.
Included in this group are the steam, the rotary, and the fluidised-bed dryers. Although each of these dryer types
has certain advantages, their high initial and operating eosts have thus far prevented market penetration.

country, and from crop to crop. For instance, a 1995

Chinese tender for the purchase of a series of high-temperature maize dryers ranks and weights the dryercharacteristics as follows (CNIIEC 1995): (I) initial
price-60%, (2) energy consumption--S%, (3) kernel
breakage-5%, (4) moisture gradient at dryer exit5%, (5) life expectancy-4%, (6) dryer type-3%, (7)
increase in stress-cracked kernels-3%, (8) level of
automatic control-3%, (9) kernel-temperature at dryer
exit-2%, (10) service record-- 10%. Thus, the price
factor (60%) far outweighs the technical factors (30%)
and the service factor (10%), and the least expensive
dryer is likely to be selected for the drying of maize in
northeastern China, regardless of the model's grainquality and energy-efficiency characteristics. In general, the weighting scale of the various economic/technical/service factors is decisive in the selection of the
high-temperature dryer at a grain facility.
Analyses of dryer designs are today routinely conducted by computer simulation; models of the major
dryer types have been available in the literature for a
decade (Brooker et al. 1992). Simulation modelling
has not led to revolutionary new dryer designs but has
resulted in evolutionary improvement of existing
dryer types. This is best illustrated by comparing the
energy-efficiency and grain-quality characteristics of

is one of the essential steps in the postharvest

technology of grains, especially of maize and rice
which almost always require mechanical drying.
Grains are dried on the farm and off the farm,
depending mainly on the country of production. For
instance, in the USA the bulk of the maize is farmdried; in France, almost all the maize is dried off the
farm at local grain depots. This paper assesses the
state-of-the-art of off-farm grain drying, and its specific needs for design improvement.
Off-farm dryers are considercd to be of the hightemperature continuous-flow type, and to have a
capacity of at least 12.5 tlhour. Usually, this excludes
all but the largest in-bin drying systems.
The main off-farm dryer types are: (I) the crossflow
dryer, (2) the mixed-jlow dryer, and (3) the concurrentjlow/counterjlow dryer (see Fig. I). Each dryer type is
able to dry maize rice and othcr grains. Each type has
specific advantages and disadvantages. The dryerselection criteria differ somewhat from country to
DRYING

" Department of Agricultural Engineering, Michigan State

University, East Lansing, MI, 48823, USA.
Department of Agricultural Engineering, Purdue University, West Lafayette, IN, 47907, USA.

123

1970-vintage crossflow dryers and 1990 models: significant improvement has occurred with respect to
both dryer-quality parameters.

The moisture range narrows and the standard deviation diminishes during storage (see Tables 3 and 4).

Dryer Types
CROSSFLOW

CONCURRENTFLOW

Crossflow dryers
Cross flow dryers are the most popular dryer type in
North America. They have a plenum surrounded by a
relatively thin grain column; hot air traverses the
grain perpendicular to the downward flow of the
grain. Cooling of grain takes place in the bottom section of the grain column. Cross flow dryers are often
called tower or column dryers.
Crossflow dryers do not dry grain uniformly. Significant moisture and temperature gradients exist
across the grain column at the moment the drying
process is discontinued. During the cooling cycle the
degree of non-uniformity decreases, but a definite
moisture differential among the kernels still exists
when the grain leaves the dryer, notwithstanding the
fact that the average moisture content may have
reached the desired level.
Recent design advances in crossflow dryers have
improved the grain-quality characteristics of this
model typc. Airflow reversal has been incorporated in
some crosstlow dryers in order to otTset the moisture
and temperature differentials in the grain column.
Grain inverters turn the overheatcd grain at the airinlet side to the air-exhaust side of the column, and
thus reduce overdrying/overheating. A new feature
added recently to the basic crossflow design-tempering-improvcs the quality of crossflow-dried grain.
Modern crossllow dryers with air recycle, grain
tempering, and grain inverting are able to dry wet
grain at high throughput and moderate energy efficiency, and can produce dried grain with moderate
moisture differentials among the kernels. For feed
grain, the cross flow dryer is a good choice.
In-depth analyses of crossllow-dryer designs can
be made by employing a differential-equation-based
model of the following type (Brooker et al. 1992):

-- --

-- --

moist
exhaust
air

warm

drying

air

t~

COUNTERFLOW

moist
exhaust
air

MIXED FLOW

moist
exhaust
air

moist
exhaust
air
warm
drying
air

moist
'----=~- exhaust

~
Figure 1.

air

warm
drying
air

---.... grain
_air

Schematics of the four major types of hightemperature grain dryers: crossflow. concurrent-flow, counterllow and mixed flow.

Grain Moisture Content

High-temperature grain dryers are commonly rated
for capacity on the basis of 10-point moisture
removal from maize, i.e. from 25 to 15% moisture
content (wet basis). No mention is made that the
quoted moistures are average moisture contents. It is
assumed that the variation in moisture content of the
maize kernels entering a dryer is small. However, this
is not the case, as is shown in Table I in which the
average moisture content, the moisture content range,
and the standard deviation of 50 kernels on an ear are
tabulated for the tip/middle/butt kernels of an early
and a late maize variety. The difference in the moisture content of the wettest and driest kernels is 27.5%
for the early variety, and 32.0% for the late variety.
The tip kernels are on average about 5% drier than the
butt kernels. Just after high-temperature drying, the
difference in moisture content between the kernels is
still large, as the data in Table 2 show.

aT
dx

0)

(1)

0)
(2)

(3)

aM =
124

. Ie - kerneId'
.
a smg
rymg equation

(4)

Table 1.

Average kernel moisture content (%), moisture content range, and standard deviation of 50 kernels on an ear of an
early and late variety maize in the midwestem USA.
Early variety

Late variety

Location
Average moisture
content

Moisture content
range

Standard
deviation

Average moisture
content.

Moisture content
range

Standard
deviation

Tip

25.0

8.5-36.5

6.20

20.8

15.5-32.0

3.40

Middle

28.3

10.5-38.0

6.44

24.3

15.5-35.0

5.04

Butt

30.4

12.5-40.5

9.03

26.1

10.0--37.5

6.88

Table 2.

Standard deviation and moisture content (%) range of maize dried in a commercial crossflow dryer.
Outlet

Inlet

Table 3.

Time
(hours)

A verage moisture
content

Standard
deviation

Moisture content
range

12.5-34.5

14.5

3.17

8.5-24.5

12.0-33.0

13.9

3.44

8.0--28.5

14.0-31.0

12.7

4.23

7.0--31.0

Standard
deviation

Moisture content
range

22.7

4.51

21.8

3.80

21.0

3.13

Average moisture
content

Table 4.

Short-term change in the moisture content (%)

range and standard deviation of maize after
crossflow drying (average m.C. ~ 14.5%).
Moisture content
range

Standard
deviation

9.0--30.5

3.75

10.0--21.5

2.63

20

10.0--24.5

2.38

27

10.0--18.5

2.08

45

11.0--17.5

1.74

50

9.5-16.5

1.63

68

10.0--17.5

1.50

92

11.0--17.0

1.38

114

10.5-16.0

1.21

122

11.5-16.5

1.24

450

11.5-16.5

1.03

Long-term change in the average moisture

content (%). moisture content range, and
standard deviation of maize stored in a 1500 t
bin under Michigan conditions.

Date

Standard
deviation

Average
moisture
content

Moisture
content

14.5

9.0--30.5

3.75

10 Jan. 94

16.4

11.0--18.5

1.30

17 Feb. 94

15.2

11.0--17.5

1.06

22 Mar. 94

13.6

10.0--15.0

1.15

25 Apr. 94

13.2

10.5-15.5

1.07

I Nov. 93-30 Nov. 93

range

Airflow reversal, tempering, and grain inverting

were first investigated by simulation (Pierce and
Thompson 1981), and have subsequently been
applied successfully to commercial crossflow dryer
designs (FFI Corporation, Indianapolis, Indiana,
personal communication, 1995).

Solution of the model for a specific crossflow

dryer establishes the approximate values of such
parameters as: (I) the grain retention time, (2) the
minimum and maximum kernel temperatures, (3)
the minimum and maximum kernel moistures, and
(4) the energy efficiency.
125

Concurrent-flow dryers

The equations for the concurrent-flow and crossflow drying models appear superficially to be similar
but close scrutiny reveals significant differences
(Brooker et al. 1992). The crossflow model consists
of a set of partial differential equations, while the
concurrent-flow dryer is represented by a set of ordinary differential equations:

The concurrent-flow dryer design is relatively new;

it has one or more concurrent-flow drying sections,
and one counterflow cooling section. In a concurrentflow section, the grain and drying air flow in the same
direction, in the counterflow cooler in the opposite
direction.
With the exception of a small, one-stage on-farm
concurrent-flow model, concurrent-flow dryers have
two or three concurrent-flow drying zones. A tempering zone is located between successive drying
sections. The ability to employ different air temperatures in the different stages is an inherent advantage
of this dryer type.
A concurrent-flow/counterflow dryer has alternate
concurrent-flow and counterflow drying sections.
Dryers of this type are of Chinese design, and usually
are multi-tower units. Because of the relatively poor
grain-quality characteristics of this dryer, it is not further discussed.
The most distinguishing feature of the concurrentflow dryer is the uniformity of the drying process.
Every kernel undergoes the same heatingldrying/tempering/cooling treatment, unlike in crossflow and
mixed-flow dryers. The temperature of the drying air
is much higher than in the other dryers because the
wet grain is subjected to the hot drying air not for
hours (crossflow dryers) or minutes (mixed-flow dryers), but only seconds. Therefore, the grain does not
approach the temperature of the drying air, as it does
in the other dryer types.
The uniform, relatively gentle drying and cooling
processes in concurrent-flow dryers, and the built-in
tempering treatment(s), result in dried grain of superior quality. The percentage of stress-cracked kernels
in concurrent-flow dryers is less than in mixed-flow
dried and in crossflow-dried grain.
A characteristic feature of the concurrent-flow
dryer is its use of ultra high drying-air temperatures, as high as 200-2S5C for maize. The high
evaporative and sensible heat loads at the inlet of
each drying section prevent the grain kernels from
reaching temperatures above 60--S0C (depending
on the type of grain).
Precleaning of the wet grain (recommended for all
high-temperature dryers) is essential for concurrentflow dryers because of the high operating temperatures.
Most dryer experts agree that concurrent-flow
grain dryers are theoretically, technically, and operationally superior to crossflow and mixed-flow dryers,
with respect to grain-quality characteristics and fuel
efficiency. However, the high-technology aspects,
general misunderstanding, and relatively-high initial
cost are definite disadvantages of this dryer type.

dT
dx

(5)

(6)

(7)

dM =

dt

. Ie - kerneI drymg
' equatIon
.
a slOg

(8)

Standard mathematical techniques (e.g. RungaKutta) can be used to solve the system of equations
(5)-(S).
The grain in a concurrent-flow dryer is cooled in a
counterflow cooler. The concurrent- and counterflow
drying/cooling models are similar except for some of
the signs in the equations (Brooker et al. 1992). However, the solution of the counterflow model is more
complex because it is a two-point boundary value
problem (instead of a one-point boundary problem as
is the case for the concurrent-flow model). Montross
(1995) recently developed a stable method of solving
the model of a counterflow grain cooler.
Computer simulation has contributed extensively
to the design of multi-stage concurrent-flow dryers, in
particular of optimum grain-bed depths and optimal
air-recycle patterns (Bakker-Arkema et al. 1992).

Mixed-flow dryers
Mixed-flow dryers are the predominant dryer type
in western Europe and Latin America. Grain is dried
in mixed-flow dryers by a mixture of crossflow, concurrent-flow, and counterflow processes. The grain
flows over a series of alternate inlet and exhaust air
ducts. This results in fairly uniform drying, and therefore in relatively uniform grain moisture content and
quality. The drying temperature in mixed flow dryers
is higher than in crossflow dryers because the grain is
not subjected to the high temperature for as long.
It has recently been shown that there is a significant
difference in the retention time and grain-temperature
history between the kernels as they pass through a
mixed-flow dryer (Liu 1993). This leads to a higher
126

Dryer Comparison

than expected spread in the moisture content and temexiting the dryer.
perature of the
The difference in design between different mixedflow dryer models centres around the duct size/spacing/pattern specifications. No comparative studies
have been published on these design modifications
with respect to fuel consumption, grain-quality characteristics, and capacity (i.e. throughput per unit of
dryer volume). Therefore, claims by manufacturers of
'best duct design' are impossible to verify.
Mixed-flow dryers are more expensive to manufacture and require more extensive air-pollution equipment than crossflow dryers.
During the mixed-flow drying process, the grain
kernels are subjected to a continuously-changing pattern of repeated crossflow concurrent-flow and co untertlow drying treatments. Therefore, a mixed-flow
dryer simulation model consists of a combination of
these three submodels (Liu \993).
The mixed-flow drying model has recently been
utilised for the design of tapered airducts in a mixedflow maize dryer (Cao 1993).

Drying temperatures
The drying-air temperatures employed in hightemperature grain dryers depend on the dryer type
and the grain variety. Table 5 contains values of the
temperatures measured recently in maize dryers at
elevator/grain-depot sites in the USA and China
(Montross et al. 1994; Liu et al. 1994). Clearly, the
concurrent-flow dryers operate at the highest temperature, the crossflow dryers at the lowest temperature.
The disparity in operating temperatures between
those measured in the USA and China is due to the
different heat sources used in the two countries: natural gas in the USA, coal in China.

Table S.

Miscellaneous dryer types

In addition to the three major grain dryer types, Le.
crossflow, mixed-flow, and concurrent-flow, several
other dryers are used occasionally for grains, often
only at the research level. Spouted-bed, fluidised bed,
and microwave dryers are examples; they have not
(yet) proven to be economical as high-temperature
dehydration devices for grains.
Rotary dryers are employed successfully in drying
rice, mainly parboiled rice. The uniform drying treatment of the kernels is an advantage. However, the initial and maintenance costs of these dryers are high,
and thus they do not compete with the more eonventional grain-dryer types.
The steam dryer is a recent Chinese invention, and
is popular in certain regions of northeastern China. A
steam dryer usually consists of 3 to 4 towers in series.
Each tower contains in its upper section a series of
steam pipes, and in its lower section a number of
inlet!outlet airducts. The grain is heated by conduction as it flows over the steam-heated pipes, and is
subsequently treated with ambient or slightly-heated
recycled drying-air. The grain retention time is long
(4--6 hours for lO-point moisture removal) due to the
relatively low grain temperatures, resulting in superior grain quality. The high initial cost of a steam
dryer is likely to prevent adoption of this dryer type
outside of China.
Rotary-dryer models have been developed in the
chemical engineering industry (Kelly 1987). However, they have not been applied to the design and
analysis of rotary grain-dryers.
The authors have not been able to find a steamdryer model for grain in the literature.

Typical drying-air temperatures employed in

different maize-dryer types in the USA and
China.

Dryer

USA (0C)

China (0C)

Concurrent-flow

205-290
I30--14O
85-120

150--160
120--J50
90--120

Mixedf1ow
Crossflow
Steam dryer

130-140

The grain-kernel temperature, not the drying-air

temperatures, determines the drying rate of the grain
in a dryer. Figure 2 shows the distribution of the kernel temperature in the crossflow, concurrent-flow,
mixed-flow, and countertlow dryers. The variation in
temperature (and in moisture content) between the
kernels in the crossflow dryer is large, in the mixedflow dryer relatively small, and in the concurrent!
countert1ow dryer non-existent. Also, in the erossflow dryer the grain kernels are subjected to the high
inlet-air temperature for a period of hours, in the
mixed-flow dryer for minutes, and in the coneurrentflow dryer for (only) seconds. Therefore, there are
different high-level limitations on the drying-air temperature in the three dryer types.

Maize quality
The authors have recently investigated the effect of
dryer type on maize quality (Montross et al. 1994; Liu
et al. 1994). In the USA, three dryer types were
tested, Le. crossflow, concurrent-flow and mixedflow models. In China, the same three types were
investigated, along with steam-drying and sun-drying
installations. Tables 6 and 7 show the results of both
studies.
127

4. Mixed-flow dryers fal] between the concurrenttlow and crosstlow dryer types with respect to the
stress-cracking of maize.
Comparing the data in Tables 6 and 7 shows that
the number of stress cracks recorded for the three
major dryer types in China is lower than in the USA.
This could be due to variety differences but is likely
to be caused by the higher drying temperatures used
in the USA than in China, Le. concurrent-tlow dryers
usually operate at 250-275C in the USA, but at 125150C in China.
A comparison between four dryer types is given
in Table 8 with respect to the drying-air temperature, the maximum temperature reached by the
grain, and the expected increase in stress-cracked
kernels, in drying maize by ten percentage points of
moisture.

at air exhaust
~

:::l

.~E

0(1)

Ec

>

t.

:::l

.it
0(1)

I \~
,
,

CO

.~o

El::

............

cO

'mu
Qj

(!)

at air inlet

air temp.

.=~
<1l:J
oc!S~

<:(1)

.~ ~

at a" "' ......

inlet..-'
mean

.~e:?

<1l:::l

(92

_---ai~; exhaust

o<!l
<:(1)

~
(92

grain temp.

distance

(a)

(b)

Dryer rating
High-temperature dryers are usually rated for
capacity only. In some cases the energy efficiency is
given. A buyer has difficulty interpreting the dryer
manufacturer's data due to the lack of a standard rating scheme.
The International Standards Organization (ISO)
has proposed the standard 'Agricultural Grain Dryers-Determination of Drying Performance.' Until
this standard has been approved, and accepted worldwide by grain-dryer manufacturers, it is impossible to
draw objective conclusions from a comparison of different dryer types and dryer models.

airlemp.

grain temp.

Dryer Control

Figure 2.

distance

distance

(c)

(d)

The moisture content of wet grain reaching a hightemperature continuous-tlow dryer over a 24-hour
period can vary greatly. This is due to the different
harvest-procedurc preferences, soil types, and variety
selections of individual farmers. At commercial elevators it is not unusual to encounter moisture content
differences of 10-15% in lots of maize received from
different growers. Yet all the grain must be dried to
approximately the same average moisture content.
The challenge presented to the dryer operator. or the
automatic controller, is to properly vary the speed of
the unload auger and thus the residence time of the
grain in the dryer.
Manual control of continuous-tlow dryers often
leads to significant overdrying or underdrying. Automatic control of continuous-tlow dryers is usually
designed to minimise these occurrences. Secondary
objectives are minimising energy consumption and
optimising dryer capacity, both necessarily subject to
grain quality constraints (Eltigani and BakkerArkema 1987).

Moisture and temperature changes during: (a)

crossflow drying; (b) concurrentflow drying;
(c) counterflow drying; and (d) mixed-flow
drying.

Several conclusions can be drawn from the data in

Tables 6 and 7:
I. Sun drying is able to produce maize with a minimum number of stress cracks if properly imple.
mented.
2. Steam-dried maize usually has only a small
number of stress cracks.
3. Of the three major high-temperature dryer types,
concurrent-flow dryers cause the smallest
increase in the number of stress cracks in maize
kernels while crossflow dryers generate the largest increase.
128

Table 6.

Average type of stress cracks, stress-cracked percentage and stress-crack index of maize dried in three dryer types
in the USA.
Stress cracks
None (%)

Single (%)

Multiple (%)

Checked(%}

Stress-cracked (%)

Se!"

Three-stage concurrent-flow

47.4

5.8

27.1

19.7

52.6

185.6

Mixed-flow

34.5

8.0

32.5

25.0

65.5

230.5

Crossllow

12.2

3.4

38.6

45.8

87.8

348.3

Dryer type

a SC! -1 * (% single) + 3* (% multiple) + 5* (% checked) (Gunasekanm et al. 1985).

Table 7.

Average type of stress cracks, stress-cracked percentage and stress-crack index of maize dried in five dryer types
in China.
None(%)

Single (%)

Multiple (%)

Checked (%)

Stress-cracked (%)

Concurrent-flow/crossflow

71.6

10.0

14.7

1.0

28.4

72.6

Mixed-flow

70.7

11.0

15.7

0.7

29.3

70.9

Crossflow

55.5

15.7

21.6

9.2

44.5

107.6

Steam

89.7

6.3

1.0

3.0

10.3

24.3

Sundrying

93.3

3.5

1.5

6.7

16.4

Dryer type

SC!"

a SCI _I * (% single) + 3* (% multiple) + 5* (% checked) (Gunasekaran et al. 1985).

Table 8.

The average effect of dryer type on the drying-air temperature, the maximum grain temperature, and the
percentage of stress-cracked kernels in maize.

Dryer type

Drying air temperature (OC)

Maximum grain temperature (.C)

Stress-cracked kernels (%)

Crossflow

80-110
100-130
200-285

80-110
70-100
60-80
40-50

70-85
40-55

Mixed-flow
Concurrent-flow
Steam

Ambient

Continuous-flow dryer control is of the classical

or the adaptive type. Either the exhaust-air temperature or the inlet and outlet moisture contents of the
grain are measured. and the speed of the metering
roles is adjusted according to a specific control law.
Temperature-based controllers are adequate for
small inlet moisture variations; moisture-based control is recommended for large swings in moisture
content.
A classical feedback controller reacts slowly when
the residence time of the grain in the dryer is long. An
optimal feedback control system requires a welldefined objective function, which is difficult to obtain
mathematically. A classical feed-forward controller
needs an accurate dynamic model of the drying process, which requires long computation time for on-line
calculations (Marchant 1985).

30-45

10-20

An adaptive controller has been shown to offer the

best technique for adequately controlling continuousflow grain drying (Moreira 1989). Adaptive feedback
and feed-forward control is able to minimise the fluctuation of the outlet moisture content even for large
variations in the inlet and ambient conditions.
Fuzzy or expert control is used when the objective
function is difficult to express mathematically. It
employs a set of heuristic rules based on expcrimental
knowledge and operator expertise (Zhang et al. 1990).
Two quality-control measures have been proposed
to evaluate different control performances. One is
based on the standard deviation of the grain moisture
content and the other on the percentage of off-specification product. By employing these control measures,
the success or the failure of a particular controller
type can be quantified.
129

Table 9.

Considerable computing power is required for continuous-flow dryer controllers. The required continuous or intermittent sensing of the grain moisture
needs sophisticated instrumentation. Therefore, the
cost of a control system for a continuous-flow grain
dryer is relatively high.
Notwithstanding the substantial costs, dryer control systems are economically justified on many
grain dryers.

Parameters used in the capital budgeting analysis

(in US$) of a concurrent-flow dryer to be fitted
with a preheater.

Input parameters
Discount rate

Dryer Economics

15%

Life of preheater

\0

Federal income tax

0,17,34%

Depreciation method

straight-line

Capital costs

The analysis of profitability of the purchase of a particular dryer is an essential part of the evaluation of
different dryer types. Analyses commonly used, such
as the payback period and rate of retum, do not adequately express the economics in the dryer-selection
process. The capital-budgeting (also called life-cycle
costing) analysis does allow the buyer to analyse the
cash-flows over time resulting from the purchase of a
specific dryer model. The capital-budgeting procedure
requires values of a series of parameters-including
the fixed and operating costs, the energy consumption,
the grain quality change, the fuel and maintenance
costs, the service life, the time value of money (i.e. the
interest to be paid on the loan), and so forth.
The capital-budgeting analysis provides a realistic
comparison of dryer types. For instance, the initially
costly dryer may in the long run be the better buy
because it may produce better quality grain at lower
operating costs and lower environmental pollution. A
payback-period analysis may not reveal these advantages; a capital-budgeting analysiS, spanning a 3-10
year planning horizon, will.
A capital-budgeting analysis of the purchase of a
dryer requires knowledge, or accurate estimates, of a
number of dryer-related parameters such as the dryer
capacity, energy efficiency, salvage value, etc. Many
of these parameter values are inadequately known,
and thus have to be researched by engineers and
economists.
As an example, the capital-budgeting costs of adding a grain-preheater to an existing concurrent-flow
dryer are analysed (Montross 1995). [The preheater
increases the capacity of high-temperature dryers by
10-15%.] The parameters used in the analysis are
listed in Table 9. The drying costs in the table refer to
the drying charges at a local grain elevator, and would
be incurred if no preheater had been installed.
Table 10 shows the tonnage to be dried and the
number of drying-operations to be required for the net
present value (NPV) to be zero, for three levels of elevator income-tax rate. Thus, at a tax rate of 17%, and
drying maize from 25 to 15%, at least 584 t of maize
have to be dried per year for the addition of the preheater to the dryer to be economical, or the dryer has
to be operated at least 255 hours.

Preheater cost (fans/burners)

$10,590

Dealer profit and

miscellaneous costs

$2,650

Installation labor ($23/hour)

$3,680

Crane time ($75Ihour)

$1.200

Salvage value

$2,000

Operating costs
Natural gas

$2.83/m3 ($0.401100 ft3)

Electricity (kWh)

$0.075

Drying costs
5 percentage points of
moisture removed
Capacity increase at 5 points
10 percentage points of
moisture removed
Capacity increase at 10 points

$3.93/1 ($O.IOlbu)
6.2 tlbr (245 bu/hour)
$9.82it (SO.25/bu)
2.3 t/br (90 bulhour)

Table 10. Maize to be dried (t/year) and annual operating

time (hours) required for the net present value
(NPV) of the prehealer to be zero at a variable
tax rale.
Tax rate and time required
Maize dried at 17% tax, t/year
Time required, hours
Maize dried at 34% tax, t/year
Time

hours

Drying
20 to 15%

Drying
25 to 15%

1,411

584

226

255

1,682

696

270

304

A capital-budgeting analysis of a high-temperature

dryer provides quick answers to the effect on the net
present value of: (I) the discount rate, (2) the fuelinflation rate, (3) the loan policy. (4) the local drying
cost, (5) the grain-quality premium and (6) the dryer
energy-efficiency. In short, a capital-budgeting analysis provides for better economic information on a
dryer, or on the add-on to a dryer, than can be obtained
from an initial-cost or payback-period comparison.
130

Liu, Q. 1993. Study on the drying mechanism, simulation

and test of mixed flow grain dryers. Ph.D. Thesis. Beijing
Agricultural Engineering University, Beijing, China.
Liu. Q., Bakker-Arkema. F.W. and Hines. R. E. 1994. Effect
of dryer type on maize quality in China. SI. Joseph. Michigan. American Society of Agricultural Enginecrs. Paper
Number 94-6589.
Marehant, 1.A. 1985. Control of high temperature continuous
flow grain driers. Agricultural Engineering, 40, 145-149.
Montross, M.D. 1995. Dryer performance enhancement
through grain pre-heating. M.S. thesis, Michigan State
University, East Lansing. Michigan.
Montross, M.D., Bakker-Arkema, F. W., Leppanen. C. L.
and Hines, R. E. 1994. Moisture content variatiou and
grain quality of corn dried in different high temperature
dryers. SI. Joseph, Michigan, American Society of Agricultural Enginecrs, Paper Number 94-6590.
Moreira, R.G. 1989. Adaptive control of continuous flow
grain dryers. Ph.D. thesis, Michigan State University,
East Lansing, Michigan.
Pierce, R. O. and Thompson, T. L. 1981. Energy use and
performance related to crossflow dryer design. Transactions of the American Society of Agricultural Engineers,
24,216-220.
Zhang, Q . Litchfield, J.B. and Bentsman, J. 1990. Fuzzy
predictive control system for corn quality during drying.
In: Food processing automation. Proceedings of the 1990
Conference, American Society of Agricultural Engineers.
SI. Joseph, Michigan, American Society of Agricultural
Engineers. Publication 02-90.

References
Bakker-Arkema, F.W., Maier, D.E., Raifsnider, R. and
Westelaken, C. M. 1992. Commercial dehydration of
soybeans. St. Joseph, Michigan, American Society of
Agricultural Engineers, Paper Number 92-6532.
Brooker, D.H., Bakker-Arkema, F.W. and Hall, C.W. 1992.
Drying and storage of grains and oil seeds. New York,
Van Nostrand Reinhold.
Cao, C.W. 1993. Experimental investigation of hot air drying mechanism of corn. In: Proceedings of the 1993 International Symposium of Grain Drying and Storage
Technology. Beijing, China, 15-18 September 1993.
CNlIEC (China National Instruments Import and Export
Corporation) 1995. Bid Number CICW956014: China
Grain Distribution and Marketing Project. Ministry of
Internal Trade, Beijing, China.
Eltigani, A.Y. and Bakker-Armena. F.W. 1987. Automatic
control of commercial cross flow grain dryers. Drying
Technology, 5,561-575.
Gunasekaran, S., Deshpande, S., Paulsen, M.R. and Shove,
G. C. 1985. Size characterization of stress cracks in corn
kernels. Transactions of the American Society of Agricultural Engineers, 28,1668-1672.
Kelly, J.1. 1987. Rotary drying. In: Mujumdar. A.S., ed.,
Handbook of industrial drying. New York, Dekker Incorporated.

131

In-store Drying and Grain Psychrometries

Robert Driscoll*

Abstract
Product drying, though conceptually simple, is the least understood postharvest operation. Successful drying produces a safe product, but drying problems can cause massive quality and actual losses. The interaction of a product with air can be represented on a psychrometric chart, a basic tool of the drying engineer,
and this allows estimates of drying performance to be made. Interaction of the product with the air also will
affect its chemical structure, microbial flora, and storage stability. A current trend in dryer models research
is towards quality models which can be incorporated in deterministic models of grain drying, and are conveniently expressed in terms of the temperature and moisture histories of the product. Such models allow
predictions of safe and optimal drying from simulated or historical weather data, as well as development of
appropriate control algorithms. The trend in Southeast Asia is towards mechanisation and increased collection of grain in bulk, allowing technologies on a more efficient scale to be introduced.

DRYING is one of the most common and least understood postharvest operations. Banga and Singh (1994)
stated that drying is too complex to fully analyse in a
rigorous mathematical sense, and that dryer design
still mainly depends on experimental data. Although
true, the application of basic science to situations
such as drying of agricultural crops has led to a revolution in our understanding of drying equipment and
strategies, but that revolution is only now starting to
impact on the massive postharvest grain drying problem of Southeast Asia.

fluidised-bed dryers, and conduction drum dryers

(I\oomhorm et al. 1994). Soponronnarit and Prachayawarakom (1993) have developed fluidised bed dryers for first stage drying of paddy, predicting a cost of
about US$80 per tonne water evaporated, and studying energy, quality and throughput rates of the tluidised-bed drying system. This has now been
commercialised in Thailand.
I worked on the dryer for the model grain depot at
Jilin in China (Newman 1992), and so was able to
observe developments there. China shows a slow trend
towards mechanisation, the main delay being that the
displaced labour force does not attract competitive
wages compared with agricultural employment, which
is currently 70% of employment. As the Chinese GNP
rises, this trend towards urban society and 10% rural
work force will accelerate (Schrock 1994). But all
countries in Southeast Asia reflect this trend.

Grain drying in Southeast Asia

The wet season harvests originated through the
development of high yielding, fertiliser-dependent
varieties of grain. Hybrid varieties, due to their
greater initial vigour, have helped to increase yields at
a time when land, labour, and water are decreasing
(Virmani and Dedolph 1994), and so have increased
pressure on the postharvest system. Drying the grain
became a major problem in the region, as sun drying
proved insufficient for wet season tropical climates_
Thus, mechanical dryers have been investigated as
an alternative. New types of dryers are now being
developed at a great rate; for example, spouted and

Interaction of grain with air

Drying inhibits product degradation by eontrol of
water activity only. Water activity determines the rates
of most biological and chemical reactions of relevance
to food systems. Thus, the design objective in drying is
not to control moisture but to control water activity. Yet
most dryer objectives are stated in terms of moisture.
The most common technique for dehydration is to
allow interaetion between low moisture air and the
product, moisture being transferred to the air and then
carried out of the grain mass in solution.

Department of Food Science and Technology, The University of New South Wales. Sydney, New South Wales
2052, Australia.

132

Other interactions between the grain and the air

occur, many related to quality. A secondary interaction is a thermal interaction between the air and the
grain, often referred to under constant aeration conditions as the initial transient. Some related effects are:
Grain can be preserved cheaply by cooling (Barth
1993), giving independence of weather conditions
and dryer delays, and at a low cost as cooling fronts
propagate rapidly through a grain mass. Few have
attempted to simulate more than the immediate
dryer, with the exception of Chung et al. (1991),
who simulated an on-farm grain storage system for
American conditions, finding that gas-modulated
burners gave a potential saving over on/off burner
control. Recirculation of air was also studied in this
simulation. My own work in simulating Malaysian
mills as a drying system demonstrated dramatically
the quality benefits that would result from cooling
grain in an aeration bin at receival, so that the effect
of dryer delays or bottlenecks within the complex
would be alleviated. A reduction of dry matter loss
from 0.8 to 0.5% was observed using typical harvest data, this being the difference between accepting and rejecting the grain for export. The cost
would have been that of a receival bin (bulk), aeration equipment and a small fan. No cooling system
was required.
Grain kept in bags absorbs moisture at a rate of
about 0.5% per month in tropical climates (Guritno
et al. 1991).
If grain is sealed inside a plastic bag, gases build up
due to respiration which are self-limiting, allowing
long term storage (several months) by this simple
technique (Kawashima and Siriacha \990).

where Pv is the vapour pressure exerted by moisture

in the air, and Ps is the vapour pressure of moisture in
saturated air at the same temperature. This definition
identifies RH with water activity aw , although relative
humidity is generally expressed as a percentage
whereas water activity never is.
Recent research has suggested that there are criteria
other than water activity and temperature which
determine reaction rates, and these alternatives will
be discussed briefly in a later section.
The result of the above identification is that we can
plot lines of constant moisture (isosteres) on a psychrometric chart, by holding moisture constant and
using the function/to calculate the relative humidity
as temperature varies. For most agricultural products,
temperature has less effect than moisture on the relative humidity, so that the isosteres tend to follow the
relative humidity lines. This accounts for the use of
moisture content for a particular product as the commercial determinant of the end point of drying.
Work has continued internationally on building up
a comprehensive library of product properties, with
individual researchers continuing to collect basic
thermophysical data.

Psychrometries
Definitions
Air is used as the transport medium for moisture in
the dryer. So as far as the dryer is concerned, air is a
two component mixture of a gas (primarily O2 and
N2 ) and a vapour (water). A limited volume of :iir
above a free water surface becomes saturated. the
amount of moisture present in the air depending on
the temperature only, and independent of the pressure
of the dry air. Relative humidity is a measure of the
proportion of vapour present to the maximum amount
the air can hold (at saturation). Absolute humidity is
the mass of moisture in the air relative to the mass of
the gas component of the air (called dry basis). Plots
of air state variables are called psychrometric charts
(Fig. I), of which the most important is a plot of temperature against absolute humidity.
The example chart shown also demonstrates relative humidity lines, enthalpy lines, and density lines.
All of these are important in grain drying. Air
enthalpy is a measure of the heat content of the air
and its contained vapour:

Relevance of isotherms
Drying is a two-phase dual component system. The
two phases are vapour and liquid, and the two components are water and water vapour. At equilibrium, the
number of degrees of freedom F is:
F

2 (phases) - 2 (components) + 2

=2
Since we have two degrees of freedom, if we define
two state variables of the system we have defined the
system state precisely. Choosing relative humidity
(RH) and temperature (1) as our state variables, we
can then write:
M=/(T,RH)

h = ma [CaT + H (cvT + /co)]

where/is a function, and M is the dry basis moisture

content. For a particular temperature, the relation
between M and RH is called an isotherm. and is product dependent. In turn, relative humidity is defined as
a partial pressure ratio:

where h is the air enthalpy (kJ), ma is the mass flow

rate of the air on a dry basis (gas component only), ca
and Cv are specific heats of the air and water vapour,
respectively, (kJ/kg.K) and /co is the latent heat of
water at OC.

RH =Pv l Ps
133

Enthalpy is important in drying because for an adiabatic, constant pressure system, enthalpy is conserved. Many dryers approximate this situation. Thus
on our psychrometric chart, drying can be represented using lines of constant enthalpy. In general,
the inlet air will start at a different enthalpy from the
initial product, and thus for the product to come to
equilibrium with the air, two balances are required, a
moisture balance and a thermal or enthalpy balance.
In general, thermal balances occur quickly (typically
more than 30 times faster than moisture balances).

tion-based solution method. In practice there is some

moisture change in the product in the enthalpy front
and some change in enthalpy in the moisture front for
both the air and the product. These changes are generally small enough to be neglected.
Many researchers have developed models of thinlayer drying. Some recent work is that of:
Parti (1990), who studied diffusion with a temperature and moisture-dependent diffusion coefficient;
Bala and Woods (1992), who combined internal
diffusion with surface evaporation for malt;
Weres and Jayas (1994), who in a recent evaluation
comparing a range of popular models found that
the two-term exponential model was superior for
describing the thin-layer drying rate of maize;
Martinez-Vera et al. (1995), who developed a thinlayer drying model of corn based on diffusion,
using finite element analysis.
Extension of this concept to deep bed dryers is discussed in the next section.

Drying on a psycbrometric chart

Thin-layer drying of a product can be represented
on a psychrometric chart using this information. First
the product (P) and inlet air (A) points are plotted
(Fig. 2). If the product is high in moisture, the product point will be close to the 100% relative humidity
(saturation) line. Construct a line from the air point
along its enthalpy line to intersect the 100% saturation line. Construct a second line at constant moisture
from the initial product state to intersect the air
enthalpy line, and call this point B. The line PB represents the enthalpy 'front' moving across the product,
whilc the line BA represents the product drying.
This graphical method is a good way of explaining
the drying process, but is not as accurate as an equa-

Other types of dryers

Practicall y all dryers are systems for interacting air
with the product. Although only in-store dryers were
considered, the principles can be extended. Some
examples showing specific aspects of grain/air interaction are given below.
Relatiye Humidlty,

Psycbrometric Cbart

81

ge

16

~.

68

5<J

f'Or humid air @ ]01325.6 pascals absolute.

.... .,.,
IM18

lOtb Novf.'"mber. 1995.

i).1)27
Wet~btdb

.....

Temperature. "C.

o,e16
O.82S

Psychrometrk data.
Saturation CUl'Ve; Wexler 197' (eq 168).
Wet bulb; Smithsonian Met. Tablt'S 1963 (FerreJ eq}.

' .13

Specific Vnlume-; Wllhelm 1976.

...u
"on

....

11019

-IU)l8

..Ot7
-0.016

tuns
D..014
O,OlJ

.s

0,012

3:

~
;:
:i'::I
is'

...
~
?

.... i?'"

-0.011

t.'"
.....
1..001

006

......

..004
-0.003
....2
.80t
f

.0

.s
Dry~Bulb

Figure 1.

Temperatun, "'c..

"

45

An example of a psychrometric chart (courtesy James Darby. Stored Grain Research Laboratory, CSIRO Division of Entomology, Canberra).
134

50

/1,"<
/~,;?-/

40

120/

/,,/

-;:-

Ol

-',

Cc

10~,/"

30

"0
0>

""
g

:aE
:::l

.c

20

/~"":--

<D

S
'0
fJ)

.0

40,,:'

,/""

'"

10

20/

,/,,,,~~

o
10

20

40

30

50

60

Temperature (0C)

Figure 2.

Graphical representation of grain drying on a psychrometric chart. See text for ex pJanation.

One aspect of air properties of great interest in

dryer design is solar energy, which can be used to
partially heat the air and so dry the grain (Oosthuizen
1992). Although technically possible and achieved
by, for example, Somchart Soponronnarit in ACIAR
Project 8308, it has not proven to be a commercially
successful approach.
Spouted-bed dryer dynamics are complex, especially when it comes to scale up, and hence simplified
arrangements are being tested such as two-dimensional spouted beds (Kalwar and Raghaven J993).
Both spouted-bed and fluidised-bed dryers use high
airflow rates to fluidise all (or a part) of the granular
product mass, allowing rapid heat transfer.

Thus, two fronts are generated for constant inlet

conditions to a deep bed of material. The first is an
enthalpy front, generated at the air inlet and propagating relatively rapidly through the grain mass. The second front is the moisture front, which is a drying front.
In predicting the rate of moisture removal, the difference between the absolute humidities of the air inlet
and zone B conditions determines the rate of drying.
The advantage of drying a deep bed rather than a
thin layer is that the drying air is exposed to a large
amount of product, and so has more time to come to
equilibrium with the product. The wettest product is
at the air exit. This gives high moisture efficiency
without the need to recirculate the air or stir the grain.
In contrast, thin-layer dryers give limited time for
contact between the air and the product, so that the air
leaves the product with unused moisture capacity.
Deep-bed drying has been modelled extensively
over the past three decades. In a recent review of theories (Cenkowski et al. 1993), only three types of
drying models were recognised: the equilibrium,
non-equilibrium, and logarithmic. In practice this
classification neglects by far the most successful
model, the so-called near-equilibrium model, which
is actually the assumption of thermal equilibrium
without also assuming moisture equilibrium. At low
temperatures, near equilibrium models were found to

In-store drying
A deep bed of product can be represented as a
number of thin layers superimposed, and so can also
be represented by points A, B, and P on a psychrometric chart. The first layer acts as a true thin layer
for constant inlet conditions. The second layer is
affected by the outlet conditions of the first layer,
resulting in a time delay before it is exposed to drying air. In this way, there will be a time delay
dependent on depth in the bed before layers deep
within the grain will be exposed to the same conditions as the first layer.

135

with its psychrometric environment which causes

changes in quality.
Recent studies of microflora in milled grain samples of rice from Southeast Asia have shown a predominance of field fungi, with few observations of
storage fungi (Pitt et al. 1994), indicating that the
milling process removes most of the microflora.
Great progress in dryer design has been achieved
with the discovery that fluidised-bed drying of paddy
does not necessarily result in poor quality grain, and
that fact, coupled with the ability to economically
recirculate the air inside a fluidised-bed dryer has led
to a rebirth of this technology in southeast Asia
(Ghaly et al. J984; Tumambing 1992; Magampon and
Elepano 1993). This success was due to the discovery
that high temperature drying does not damage grain
provided that the grain is initially at a high moisture
content (Tal\ada et al. 1994). Rather it is the steepness
of the moisture gradient within the grain which correlates to head rice yield (Siebenmorgen 1992), with the
greatest reduction occurring during the first 8 hours of
exposure to an absorption environment.
The results have come from studies on causes of
fissuring and breakage during milling of rice. Reduction in head rice yield is caused by moisture readsorption (Banaszek and Siebenmorgen 1990). Although
Japanese researchers in the 19308 had established that
fissure generation is due to adsorption rather than
drying, this information was not disseminated until
rediscovered by Kunze over a long series of careful
experiments (see Kunze 1991). This has resulted in
major rethinking of drying methodologies, with the
result that grains are now frequently dried under
much harsher conditions, yet with less loss of quality.
More generally, Siebenmorgen (1992) confirmed
that moisture reduction gradients caused fissuring, the
majority of fissuring occurring during the first 8
hours after exposure to an adsorption environment.
This suggested that maize could be dried using high
temperatures and high relative humidities (Estrada
and Litchfield 1993).
More detailed research on moisture distributions
led to the disturbing discovery that individual rice
grains vary widely in moisture content with their
growth stage, and that even position on the panicle
can have extreme effects. Did this matter? Unfortunately, the answer is yes. High moisture rice releases
its moisture to low moisture rice, causing that rice to
fissure, and hence reducing head rice yield (Siebenmorgen et al. 1992).
We have come to understand more about the chemistry of grain drying as well; for example, the effects
on nutrients (Barrier-Guillot el al. 1993), enzymes
(Chrastil 1990, 1993), and the location and causes of
fissuring in the grain structure (Juliano et at. 1993;
Juliano and Perez 1993; Peplinski et al. 1994). Theoretical calculations of fissure development have been

perform better in practice than non-equilibrium models, even for fast air situations such as fluidised-bed
drying (Nathakaranakule and Soponronnarit 1993).
Earlier reviews (e.g. Parry 1985) were perhaps a little more comprehensive.
The drying time ean be estimated by equating the
moisture removal rate with the amount of moisture in
the bed above equilibrium (Bowrey and Driscoll
1986).
Work to develop and improve models of deep-bed
drying continues, for example, on:
optimising integration step size-the work of Jindal and Seibenmorgen (1994) on the importance of
the mass ratio of grain to air in a control volume;
respiration-the heat and water generated by respiration were found to have a significant effect on
drying predictions for a deep bed (Soponronnarit
and Chinsakolthanakom 1990), with heat increasing the drying rate and the moisture produced
retarding it;
volume shrinkage-an example of a real bed effect
is volume shrinkage, for many years not included
in dryer models as being too difficult. Work on collecting empirical data on bed shrinkage is rare. An
example is
and Sokhansanj (1993) for wheat
and canola;
complexity-some aeration models are of great
complexity, covering three-dimensional flow using
volume averaging theory. Van Graver (1992)
reviews work by Thorpe, Wilson, and others.
The main problems with deep bed drying are:
creation of a moisture gradient between the top of
the bed and the inlet;
lost opportunity cost, as product is held up in the
dryer for long periods, so tying up capital.
compaction of lower layers at high moistures,
resulting in increased pressure drop, higher costs
and slower drying rates; and
loading and unloading times are comparatively long.
The main advantages are:
reduced quality deterioration;
Iow energy costs for drying under the right eonditions; and
reduced grain losses.

Psychrometries and quality

There are three parameters for assessing any process equipment: cost of operation (including capital
and equipment costs), throughput, and product quality. In this section quality is considered. Dryer models need to include quality models to allow the model
to be used for strategy development. For example, in
the in-store drying model developed by the University of New South Wales (Srzednicki and Driscoll
1995), four quality models are included, dry matter
loss, ergosterol increase, yellowing and seed viability. In most cases, it is the interaction of the product
136

 there is a wide choice for control objectives, for

example, uniform drying, safety, lowest cost in one
year, lowest cost over 20 years, maximum throughput, etc.;
there is always a trade-off with deep-bed drying
between reduced costs and increased dry matter loss.
In a detailed study of control and aeration strategies for near ambient systems for Canadian wheat,
Ryniecki (1991) analysed humidistat control, time
clocks, varying airflow rates, and burner strategies
using a stochastic (probability) model of weather and
drying. The drying model was based on Parti's formulation (Parti 1990) of the differential drying rate
across a layer, converted to stochastic form by
expressing temperature in the equation as a randomly
tluctuating variable linked to weather conditions
(Ryniecki et al. 1993; Ryniecki and Jayas 1992). This
allows a probabilistic analysis of 20 years of weather
data so that strategies which both guarantee and optimise aeration could be developed.
They found that varying the airflow rate and controlling the relative humidity by air selection worked
best with time-varying humidity limits, where the
band of acceptable air was increased if weather conditions continue unfavourable, to ensure enough
hours of aeration (Ryniecki et a1. 1993). This uses the
same principle as the Australian CSIRO's proportional controller, a mechanical device for ensuring
sufficient aeration hours.
Banga and Singh (1994) have also developed a
technique for optimising aeration conditions, based
on optimal control theory. What is especially interesting about their work is that instead of optimising from
the perspective of achievement of technical goals
(such as least cost, final moisture), they have also
optimised on quality, including maximum nutrient!
enzyme retention, minimum process time, and maximum energy efficiency as their control objectives.
This emphasis on final product quality is a growing
trend in grain storage research.
A further advance in modelling has bcen through
the association of expert systems with dryer models
(Kawamoto ct al. 1992). In this case a simulation of
harvest moistures based on weather data is coupled
with an expert system which predicts likely pest outbreaks and suggests solutions.
The area of aeration control is expanding rapidly as
new techniques are developed for reducing the major
cost of drying of grains. This is possible through psychrometrics.

attempted by several researchers, one of the most

successful being a finite element approach for maize
and soybeans by Irudayara et al. (1993), using van
Mises' failure criterion, and including thermal and
moisture expansion effects and a viscoelastic model
of the internal grain structure. Their model predicted
fissuring at high temperatures (> 70C), due to
stresses of the order of 20 atmospheres being generated near the grain surface.
Thus, we seem to have come full circle, with fissuring due to drying, not readsorption, being indicated by
the models. Studies at the University of New South
Wales, on moisture gradients in rice grains (Driscoll
1995) suggested that the gradients during drying are
not as steep as the gradients that are created by exposure of a dried grain to a humid environment. This
suggests that most fissuring could be prevented by
controlling the relative humidity of the air to which
the grain is exposed after drying to below 40%, until
moisture diffusion within the grain has allowed reequilibration to occur (at least 2 hours for paddy).
Other important quality models include the following:
A storage quality indicator developed for wheat
and rice was yellowing, which is discoloration during storage due to long-term chemical reactions
(Bason et al. 1990).
Several models exist for seed viability [e.g. Giner
et al. (1991), for wheat].
Pinto et al. (1991) and Gibson et al. (\994) have
modelled aflatoxin production in terms of water
activity and temperature.
Evranaz (1993) has modelled oxidation of fats in
peanuts, expressing it as a function of temperature
and moisture.
Many grain storage pests were found to be dependent on wet-bulb temperature [Desmarchelier (1988)
with species of beetles]. Other researchers found
that this was not always true [Beckett et aL (1994),
with insect pests].
Some efforts are now being made to correlate quality measures with physically measurable characteristics, as a guide to breeding.

Psychrometries and dryer control

Control of a grain store, whether for drying or aeration, depends critically on the interaction between the
product and the air. Criteria for an aeration control
system are that it should be accurate, quick to react,
stable, and tolerant of changes. Specific points to consider in aeration control are (Moreira and BakkerArkema 1992):
prevention of overdrying is a major concern;
feedback and adaptive feedforward can be successfully applied, and a recent work even combines
these two approaches;

Conclusions
Three points should be emphasised.
We have not yet reached the stage where a
mechanical dryer is the only solution. In most areas
137

in Southeast Asia, sun drying is the preferred option,

not just for the sake of tradition but in terms of pure
economics.
As Asian countries move towards increased GNP
and a resulting increased wage level, the trend will be
away from rural employment towards mechanisation,
and the current research on dryer design, control, and
optimisation will bear increasing fruit.
Reducing costs and optimising dryer performance
depends heavily on operators understanding their
product and its interaction with air. This is the realm
of psychrometry. High-temperature dryers are less
dependent on operator skills, at the cost of reduced
thermal efficiency at low moisture contents, whilst
near ambient systems such as in-store dryers are
heavily dependent on operator skills. For a large scale
drying complex, massive savings are possible through
a deeper understanding of the drying process, for
example, cooling the grain at receival, two-stage drying, and prevention of drying delays.

Cenkowski, S., Jayas, D.S., and Pabis S. 1993. Deep-bed

grain drying-a review of particular theories. Drying
Technology, 11, 1553-1581.
Chrastil, J. 1990. Influence of storage on enzymes in rice
grains. Agricultural and Food Chemistry, 38,1198-1202.
- 1993. Changes of oryzenin and starch during preharvest
maturation of rice grains. Agricultural and Food Chemistry, 41,2242-2244.
Chung, J.H., Verrna L.R., and Mailander M.P. 1991. Simulation of a rice drying system. Transactions of the American Society of Agricultural Engineers, 34, 2065-2072.
Desmarchelier, J.M. 1988. The relationship between wetbulb temperature and the intrinsic rate of increase of eight
species of stored product Coleoptera. Journal of Stored
Products Research, 24,107-113.
Driscoll, R.H. 1995. Energy efficiency in dryers and ovens.
Food Australia, 47, 310-314.
Estrada, J.A. and Litchfield, J.B. 1993. High humidity drying of corn: effect on drying rate and product quality.
Drying Technology, 11,65-84.
Evranuz, E.O. 1993. The effects of temperature and moisture content on lipid peroxidation during storage of
unblanched salted roasted peanuts: shelf life studies for
unblanched salted roasted peanuts. International Journal
of Food Science and Technology, 28, 193-199.
Ghaly, T., Sutherland, J.W., Ghaffar, R.A., and Hassan, A.
1984. Journal of Agricultural Engineering Research, 30,
337-345.
Gibson, A.M., Baranyi, J., Piu J.I., Eyles, MJ., and Roberts,
T.A. 1994. Predicting fungal growth: the effect of water
activity on Aspergillus flavus and related species. International Food Microbiology, 23, 419-431.
Giner, S.A., Lupano, C.E., and Anon, M.C 1991. A model
for estimating loss of wheat seed viability during hot-air
drying. Cereal Chemistry, 68, 77-80.
Graver, J. van S. 1992. Aeration computer programs. World
Grain, 10,28-29.
Guritno, P., Haque, E., and Chung, D.S. 1991. Moisture
adsorption of bagged grain under tropical conditions.
Transactions of the American Society of Agricultural
Engineers, 34, 2131-2134.
Irudayara, J., Haghighi, K., and Stroshine, R. 1993. Stress
analysis of viscoelastic materials during drying: 11 Application to grain kernels. Drying Technology, 11.929-959.
Jindal, V.K. and Siebenmorgen, TJ. 1994. Simulation of
low temperature rough rice drying and rewetting in shallow beds. Transactions of the American Society of Agricultural Engineers, 37, 863-871.
Juliano, B.O. and Perez, CM. 1993. Critical moisture content for fissures in rough rice. American Association of
Cereal Chemists, 70, 613-615.
Juliano, B.O., Perez, CM., and Cuevas-Perez, F. 1993.
Screening for stable high head rice yields in rough rice.
American Association of Cereal Chemists, 70, 650-655.
Kalwar, M.1. and Raghavan, G.S.V. 1993. Batch drying of
shelled corn in two-dimensional spouted beds with draft
plates. Drying Technology, 11,339-354.
Kawamoto, H., Sinha, R.N., and Muir, W.E. 1992. Computer simulation modelling for stored-grain pest management. Journal of Stored Products Research, 28, 139-145.
Kawashima, K. and Siriacha, P. 1990. Prevention of aflatoxin contamination in maize grains with high moisture

References
Bala, B.K. and Woods, J.L. 1992. Thin layer drying models
for malt. Food Engineering, 16,239-249.
Banaszek, M.M. and Siebenmorgen, T.J. 1990. Head rice
yield reduction rates caused by moisture adsorption.
Transactions of the American Society of Agricultural
Engineers, 33. 1263-1269.
Banga, J.R. and Singh. R.P. 1994. Optimization of air drying of foods. Journal of Food Engineering. 23. 189-211.
Barrier-Guillot, B.. Zuprizal. Jondreville C. Chagneau,
A.M., Larbier, M., and Leuillet M. 1993. Effect of heat
drying temperature on the nutritive value of corn in
chickens and pigs. Animal Feed Science and Technology,
41,149-159.
Barth, F. 1993. Cold Preservation of Grain. In: Naewbany,
J.O. ed., State of the art of the grain industry in ASEAN:
a focus on grain handling and processing. Proceedings of
the 15th ASEAN Technical Seminar on Grain Post Harvest Technology, Singapore, 1992.
Bason. M.L.. Gras, P.W., Banks, HJ. and Esteves, L.A.
1990. A quantitative study of the influence of temperature. water activity and storage atmosphere on the yellowing of paddy endosperrn. Cereal Science, 12, 193201.
BeckeU, SJ., Longstaff, B.C., and Evans, D.E. 1994. A
comparison of the demography of four major stored grain
coleopteran pest species and its implications for pest
management. In: Highley, E., Wright, E.l., Banks, HJ.,
and Champ, B.R., ed., 1994. Stored-product protection.
Procerdings of the 6th International Working Conference
on Stored-product Protection, Wallingford, U.K .. CAB
International,491-497.
Bowrey, R.G. and Driscoll, R.H. 1986. A procedure to estimate the supplemental heating needs for in-store paddy
drying. In: Champ, B.R. and Highley, E., ed., Preserving
grain quality by aeration and in-store drying. ACIAR
Proceedings, No. 15, 114-117.

138

content by using plastic bags. TARC Research Highlights. 13-15.

Kunze, O.R. 1991. Moisture adsorption in cereal graintechnology: a review with emphasis on rice. Transactions of the American Society of Agricultural Engineers,
7.717-723.
Lang. W. and Sokhansanj, S. 1993. Bulk volume shrinkage
during drying of wheat and canola. Food Process Engineering. 16. 305-314.
Magampon. M.T.e. and Elepano. A.R. 1993. A laboratory
study of fluidized bed drying of high moisture paddy.
Philippine Agricultural Mechanization Bulletin. I, 15-19.
Martinez-Vera. e., Vizcarra-Mendoza. 0., Galan-Domingo.
0., Ruiz-Martinez, R., and Maurer. S.G. 1995. Naturalair grain drying modelling and validation. Drying Technology, 13.333-350.
Moreira. RG. and Bakker-Arkema. F.W. 1992. Grain dryer
controls: a review. Cereal Chemistry, 69. 39Q....396.
Nathakaranakule. A. and Soponronnarit, S. 1993. Computer
application in grain drying simulation. In: Naewbany.
1.0. and Mamlay, A.A. Proceedings of the 14th ASEAN
Technical Seminar on Grain Post-Harvest Technology.
5-8 November 1991,1-22.
Newman, e. 1992. China, Australia develop model grain
depot. World Grain, 10,17-18.
Noomhorm, A., Premakumar, K., and Sabarez, H.T. 1994.
Design and development of a conduction drier for accelerated drying of peanuts. Food Engineering, 21, 411-419.
Oosthuizen, P.H. 1992. The selection of simple indirect
solar rice dryers. Proceedings of the 8th International
Drying Symposium. Montreal, 1992. 1465-1474.
Parry, J.L. 1985. Mathematical modelling and computer
simulation of heat and mass transfer in agricultural grain
drying: a review. Journal of Agricultural Engineering. 32,
1-29.
Parti, M. 1990. A theoretical model for thin-layer grain drying. Drying Technology. 8. 101-113.
Peplinski, AJ., Paulis. 1.W., Bietz, 1.A., and Prat!, RC.
1994. Drying of high-moisture corn: changes in properties and physical quality. Cereal Chemistry, 71,129-133.
Pinto, V.E.F" Vaamondc. G., and Montani, M.L. 1991.
Influence of water activity, temperature and incubation
time on the accumulation of aflatoxin B I in soybeans.
Food Microbiology, 8, \95-201.
Pit!, 1.1., Hocking, A.D., Bhudhasamai, K., Miscamble,
RP., Wheeler, K.A., and Tanboon-Ek, P. 1994. The normal mycoflora of commodities from Thailand. 2. Beans.

rice. small grains and other commodities. International

Journal of Food Microbiology, 23, 35-53.
Ryniecki. A. 1991. Control systems for near-ambient grain
drying. In: Cereal grain mycotoxins, fungi and quality
in drying and storage. Cereal Grain, 499-511.
Ryniecki, A. and Jayas. D.S. 1992. Stochastic modeling of
grain temperature in near-ambient drying. Drying Technology, 10, 123-137.
Ryniecki. A., Jayas, D.S . and Muir, W.E. 1993. A generalized control-strategy for near-ambient drying of wheat
under Canadian-prairie climate. Transactions of the
American Society of Agricultural Engineers, 36, 11751183.
Schrock, M.O. 1994. Grain harvesting in China as an example of the mechanization process. Agricultural Mechanization in Asia, Africa and Latin America, 25, 54-60.
Siebenmorgen. TJ. 1992. Relating moisture transfer rate in
rice to kernel quality. Proceedings of the 8th International
Drying Symposium, Montreal. 1992,58-73.
Soponronnarit. S. and Chinsakolthanakorn, S. 1990. Effeet
of heat and water from respiration on drying rate and
energy consumption. ASEAN Journal on Science and
Technology for Development, 7, 65--83.
Soponronnarit, S. and Prachayawarakom. S. 1993. Maintaining quality. maximizing throughput and minimizing
energy consumption in fluidized bed paddy drying. In:
Naewbany. J.O. cd . State of the art of the grain industry
in ASEAN: a focus on grain handling and processing.
Proceedings of the 15th ASEAN Technical Seminar on
Grain Post-Harvest Technology, Singapore, 8-11 September 1992.
Srzednicki, G.S. and Driscoll, RH. 1995. In-store drying and
quality maintenance in grain. In: Champ, B.R and Highley, E.. ed., Postharvest technology for agricultural products in Vietnam. ACIAR Proceedings, No. 60, 42-49.
Tallada, J.G., Elepano, A.R., and Quick, G.R 1994. Some
effects of high temperature drying on rice milling quality.
Paper presented at the PSAE Annual Convention. Manila.
Philippine Agricultural Mechanization Bulletin, J. 11-15.
Tumambing, J. 1992. Fluidised-bed drying. Unpublished
masters thesis, University of New South Wales, Australia.
Virnlani. S.S. and DedoJph, e. 1994. Reaping the benefits of
hybrid rice in the tropics. World Agriculture, 1994, 17-20.
Weres, J. and Jayas. D.S. 1994. Thin-layer drying of corn:
experimental validation of a new numerical structural
model. Canadian Agricultural Engineering. 36. 85-91.

139

Design of Aeration and Drying Systems

Design Parameters for Aeration and

In-store Drying Systems
C.J.E. Newman*

Abstract
This paper summarises some of the main engineering considerations and parameters for the design of aeration and in-store drying systems.
Both aeration and in-store dryers operate on the principle of forcing ambient air through a mass of stored
grain and. in many respects, the engineering design principles are much the same. Drying simply requires
a larger volume of air. and in some cases the addition of some low-temperature heat.
The paper covers the main aspects of designing systems for moving air through grain, most of whieh have
been well documented in the past. The paper covers methods for determining:
the fan operation time fraction;
specific airflow rates for aeration and drying;
selection of airflow direction;
selection of air-duct distribution pattern;
airflow resistance in ducts;
airflow resistance in grain;
selection of fans:
The paper presents some new ideas for aeration of large volume' squat' silos, but generally offers a guide
to designers based on well known practices and past experience.

drying and aeration both involve forcing

ambient (or near ambient) air through a grain mass.
It is well known that the forced movement of air
through grain sets up both temperature and moisture
fronts which move through the grain in the direetion of
the airflow-the temperature front moving very much
faster than the moisture front. Thus, grain aeration is
likely to result in some moisture change, and in-store
drying will cause changes in grain temperature.
However. while similar in principle, the two practices have very different aims:
In-store drying is carried out on relatively high
moisture grain. and its purpose is to remove moisture from the grain to make it 'safe' for storage
before fungal activity begins. Temperature control
is a secondary consideration, except in high humidity conditions when it may be necessary to increase
the air temperature with supplementary heat to
reduce its relative humidity to a level where it will
remove moisture.

Aeration is normally performed on relatively low

moisture grain with the purpose of eooling the
grain in order to preserve it by maintaining its
'condition'. Drying of grain is not a consideration
when aerating, since the airflow rates are normally too low to have much influence on moisture
content.
Cooling of grain by aeration significantly reduces
insect population growth, mould growth. and loss of
viability (or germination potential). In addition. aeration can prevent the occurrence of temperature differentials within the grain which can lead to moisture
migration; it can also inhibit the development of
odours and grain discoloration.
This paper summarises the basic design parameters for in-store aeration and drying systems. The
design of hot air continuous-flow dryers is beyond
its scope.
From an engineering designer's point of view, aeration and in-store drying systems differ principally in
the quantity of air that is required, and the possible
addition of supplementary heating (for drying).
Other engineering factors come into play when con-

IN-STORE

Chris Newman and Associates Pty Ltd, P.O. Box 7185.

Toowoomba Mail Centre. Queensland 4352, Australia.

143

Common practice (in Australia) is to select a fan

operation time fraetion of 50% for initial cooling of
newly harvested grain (to speed up the cooling process), and to reduce this to 15% for 'maintaining' the
grain and for further reducing its temperature, once
the initial cooling has been achieved.

sidering some of the more sophisticated options for

in-store drying, but the principal design parameters
remain the same, and involve selection of:
fan operation time fraction;
specific airflow rate;
airflow direction;
air-duct distribution pattern;
air-duct sizes; and
fan or fans.
The selection of the specific airflow rate is the most
significant of these design considerations. Compared
with this, the selection of duct configuration, duct
size, and fan type are relatively routine design matters. However, before the airflow rate is determined,
it is important to select the fan operation time fraction.

Specific Airflow Rates

General
The 'specific' airt10w rate is the quantity of air
passing through a grain mass divided by the volume
of grain it passes through (e.g. measured in litres of
air per second per tonne of grain, or m 3/minlm 3).
Adjacent to an air duct (Le. before the air passes
through any grain) the specific airflow rate is infinitely large. As the air passes through the grain mass,
the specific airflow rate becomes smaller and smaller,
until it reaches a minimum value at the surface of the
grain. It is this minimum value of specific flow rate
that governs the cooling or drying performance of a
system, since it defines the time that it takes for cooling and/or drying the grain mass.
Thus, the minimum specific airflow rate governs
the speed of the 'temperature front' in the case of aeration, and the 'moisture front' in the case of grain
drying. Estimating the speeds of both these fronts
depends on a number of factors, the principal factors
being-apart from the physical properties of the
grain-the temperature and moisture of the grain and
of the air being passed through it. Air temperature
and moisture are highly variable factors, and the
design engineer must often rely on the work of the
research scientist to assist in determining the airflow
rates applicable for aeration and in-store drying,
especially in subtropical and tropical regions in
which ambient conditions may often be only marginally useful for either purpose.

Fan Operation Time Fraction

The fan operation time fraction is the percentage of
time that fans operate when blowing air through a
grain mass.
In the case of an in-store drying system, the need
for controlling fan operation is relevant only when the
grain moisture has reached a level that is close to
equilibrium with the average daily relative humidity.
In normal drying operations, the fans are run 100% of
the time (even during periods of rain) in order to minimise the drying period.
In aeration systems, the fans are normally switched
on and off for the purpose of selecting only the
cool est air in order to reduce the temperature of the
grain. Various control systems are used-including
manual controls, fixed time settings, time proportional controls, and more sophisticated systems which
monitor both air and grain temperature and/or relative
humidity. Automated systems are far more efficient
than manual ones, both in temlS of minimising fanhours and grain damage.
The most commonly used controller in Australia is
the time~proportioning controller (developed by
CSIRO), which automatically turns the fans on when
ambient air temperatures go below a set point. and
turns them off when it goes above it. The controller
automatically adjusts the set-point to ensure that the
fans operate for a set proportion of time. By reducing
the proportion of time that an aeration system operates, the slower will be the rate of cooling of the grain
(because less air is being used), but because it selects
only the coldest air of the day, the controller ensures
that the average temperature of the selected air will
become lower as the aeration time fraction gets
smaller. Hence, a time-proportioning controller has
the effect of reducing the grain temperature as the
time fraction is reduced.

Aeration
Ideally, the airflow rate should be sufficient to
allow cooling of a newly harvested grain mass within
a period of about 4-5 weeks in temperate areas. The
cooling period needs to be much less than this in subtropical and tropical areas unless the grain moisture is
uniformly below an equilibrium relative humidity
level of around 70%. A value for the eooling period
can be estimated from equation (1) (from Hunter
1986):

e
q= GAOf
p

(1)

where q is the airflow rate in m 3/kg/sec;

. is the porosity of the grain bed (say 0.41 for
wheat);
144

Drying

A is the cooling front velocity ratio (i.e. cool-

ing front speed/interstitial air velocity) (say

0.75 x I 0~3 as a conservative figure);
e is the6 desired cooling time (in seconds-say
3 x 10 for five weeks);
G is the flow distribution factor (say 0.8 for
vertical storages and 0.5 for horizontal storages);
f is the fan operation time fraction (say 50%
for initial cooling); and
p is the bulk grain dcnsity.

In the case of in-store drying, the critical requirement is not to preserve the condition of the grain, but
to change its moisture content. The requirement is to
dry it quickly enough that deterioration by moulding
is prevented, and this may often require airflow rates
50 to 100 times those used for aeration.
The rate of moulding of grain is linked to its temperature and moisture content. Some species of
moulds are active even at very low temperatures, and
high moisture grain stored at or below freezing point
may experience moulding if left long enough in storage. At high temperatures and high moisture, moulding takes place very quickly as can be seen from
Table 3 (from Brooker et aI. 1992).
Thc rate of drying-and hence the rate of airflowmust be such that the drying front passes through the
grain mass before moulding starts.
The drying time can be roughly estimated from a
heat balance equation such as equation (2) (from
Brooker et al. 1992):

The above equation may give airflow rates that are

slightly lower than those commonly recommended.
Alrllow rates commonly recommended for eooling
aeration vary according to the climatic conditions and
storage type, and these are summarised in Table 1
(from Foster and Tulte 1982).

Table 1.

Recommended airflow rates for aeration.

Storage type

Temperate climate

Subtropical climate

Horizontal

0.8-1.6 Useclt

1.6 -3.2 Usec!t

Vertical

0.4-0.8 Useclt

0.8-1.6 Usecl!

Vhfg{DM)(Mo Mc)

- - -...- - - . - - - - - - - - - - - - - -

6OQc

a(Ta -Tg)

(2)

where t = time (hours);

n = specific volume of the moist air used for
drying (m 3 /kg);
Ca
specific heat of dry air (J/kgfDC);
(T" Tg) = dry bulb temperature drop through
the grai n mass;
hfg = heat of vaporisation at saturation (J/kg);
DM = dry matter weight (kg);
Mo = initial grain moisture level (%);
Me = equilibrium (final) grain moisture level
(%); and
Q airflow rate (m3/hour).

The need for higher rates in horizontal storages is

to allow for the fact that air distribution is less uniform. Higher airflow rates are recommended for
warmer subtropical climates to provide more rapid
cooling during the short periods when air temperatures are low enough to be used for aeration.
Ambient aeration effectively brings the wet-bulb
temperature of the grain into cquilibrium with the
wet-bulb temperature of the air. All the benefits
achieved through cooling of the grain through aeration (viz reduced insect activity, reduced moulding,
etc.) are directly related to reduction in the wet-bulb
temperature of the grain (or more specifically, the
wet-bulb temperature of the interstitial air which is in
equilibrium with the grain).
In order to evaluate the effectiveness of aeration on
a grain mass, dry-bulb temperature readings of the
grain should be taken regularly and converted to wetbulb readings using a psychrometric chart or conversion tables (such as Table 2 which applies to wheat
and maize). For optimum results, grain should (where
possible) be cooled to around 10-15C.
It should be clearly understood that the airflow
rates which are useful for aeration of grain are quite
insufficient to dry the grain other than to a very superficial extent. The benefits of aeration result only from
the cooling of the grain and in equalising temperatures within the grain mass.

Table 4 offers designers some general guidance

on minimum airflow rates for wheat and maize
(based on US. experience) in order to achieve drying rates fast enough to avoid moulding in temperate
climates.
Note that slightly increasing the temperature of
the air by adding heat will increase the rate of drying. However, the recommended airflow rates
should not be reduced to account for this, since the
increased drying air temperature also increases the
temperature of the moist grain above the drying
front, and thereby increases the rate of mould development. The airflow rates in Table 4 are therefore
recommended for both ambient air drying, and low
temperature air drying.
145

Table 2.

Tabulated values of equilibrium wet-bulb temperature for grain at varying moisture content.

Dry bulb grain temp

looe

5e

20 0 e

[5e

25e

30 0 e

40 0 e

35e

(a) Wheat
3.7

6.5

9.6

12.3

15.3

2.5

7.2

11.0

15.2

19.5

24.0

1.0

6.0

10.2

14.7

19.0

24.0

28.5

14%

2.7

7.4

12.0

16.5

21.0

26.0

31.0

16%

3.8

8.5

13.0

18.0

22.5

27.3

20%

4.9

9.8

14.6

19.3

24.0

29.0

8%
10%
[2%

(b) Maize
8%

2.2

6.0

10.0

13.9

18.0

22.4

26.8

10%

-1.0

3.5

8.0

12.2

16.4

21.0

25.8

30.7

12%

0.8

5.3

9.8

14.6

19.2

23.9

29.0

33.2

14%

1.9

6.7

11.6

16.2

21.1

26.0

31.1

16%

2.8

7.9

12.8

17.7

22.5

27.4

20%

4.0

9.1

14.2

18.8

23.8

28.8

Table 3.

Storage life of maize.

Maize temp (QC)

Storage life in days under aeration for varying maize moisture

-1.1

15.5%

18%

20%

24%

30%

2276

648

321

127

61

1.7

1517

432

214

85

40

4.4

1012

288

142

56

27

7.2

674

192

95

37

18

10.0

450

128

63

25

12

12.8

299

85

42

16

15.6

197

56

28

11

18.3

148

42

21

21.1

109

31

16

23.9

81

23

12

26.7

60

17

Table 4.

Minimum airflow rates for in-store drying in uSA.

Grain

Grain moisture (% wet basis)

Wheat

20

2.4

30

18

1.6

20

16

0.8

10

25

40

50

20

2.4

30

18

1.6

20

16

0.8

10

Shelled maize
and grain
sorghum

Minimum airflow rate

Source: USDA (1965).

146

Minimum airflow rate (L/sec/t)

The figures in Table 4 are consistent with those

recommended for in-store drying of rice in near-tropical conditions (such as in Southeast Asia) where it is
recommended that minimum airflow rates of 1.5 m3J
m3/min are adopted. In-store drying in such regions
should not be attempted unless the initial grain moisture is not more than 18% and the grain bed depth is
less than 4 m.

can be minimised by installing a small fan in the

storage roof.]
Increase in grain moisture levels around the aeration duct can occur if the average relative humidity
of the aeration air is above the equilibrium relative
humidity of the grain.

Downward airnow aeration

Advantages

Direction of Airflow
Consideration of direction of airflow direction is
more usually associated with aeration system design
than with drying systems. In-store drying systems
almost always involve upward movement of air
through the grain. Aeration systems, on the other
hand, can use either upward or downward airflow.
There are advantages and disadvantages in both
alternatives and there seems to be no universal consensus as to which option is better in any given circumstance. The advantages and disadvantages of
each are summarised below.

In cool conditions, including those applying to subtropical conditions, the risk of moisture condensation under the storage roof and in the upper levels
of grain is minimised.
In the event that high relative humidity air is used
for aeration, any moisture increase in the grain is
spread over the entire surface area, rather than concentrated around the air inlet duct. [Not relevant
where an elevated plenum floor is used.]
Greater cooling can be achieved because there is
no heating effect from the fan. This can be important in tall silos where high static pressures are
required.

Upward airnow aeration

Disadvantages

Advantages

There is no air temperature increase and reduction

in relative humidity caused by the fan. This can be
important in subtropical climates for the reasons
described above.
The slowest area to cool is at the bottom of the
grain bulk where monitoring of grain temperature
is difficult.
It is inappropriate where warm grain may be
loaded on top of grain that has already been cooled
by earlier aeration.
Grain dust can be drawn down through the grain
bulk and cause choking of the aeration ducts.
In silos with suspended steel floors: under freezing
conditions in temperate climates, air may be
warmed and moistened as it is drawn down through
the grain bulk. This can result in severe condensation when the air comes into contact with the silo
base and cause ice build up in the grain, preventing
emptying of the silo.
lt can be seen from the above that there are no clear
arguments in favour of one system or the other. However, in Australia, upward aeration is more common
than downward aeration, and most users in the USA
are firmly in favour of upward aeration for the reasons given above.

The compression of the air as it is blown into the

bottom of a storage, results in a small rise in air
temperature and thereby a reduction in the relative
humidity of the air. A temperature rise of 3-4C
can remove the danger of increasing grain moisture
content in high ambient relative humidity conditions. This can be particularly important in subtropical climates where the lowest temperatures are
recorded at night when relative humidity can be
close to 100%.
In situations where warm grain may be added to a
bin after aeration has begun, the use of upward aeration avoids the situation where air passes first
through the warm grain, and thence downwards to
warm the previously cooled grain underneath.
In hot conditions, air may be warmed under a storage roof. Upward aeration blows this warm air out
of the storage, rather than pulling it down through
the grain mass.
The last cooled layer of grain in the grain mass is
the top surface layer, where grain temperature and
moisture levels are easily checked.
Upward airflow results in more uniform air distribution, particularly in flat stores.

Disadvantages

Air Distribution Ducting

In cool conditions, the top surface of a grain mass

will cool naturally, as will the roof of the storage.
Upward airtlow through warmer grain may result
in condensation of moisture in the cooler upper
layer, or on the underside of the storage roof. [This

The method of distribution of aeration or drying air

into a storage is important from the point of view of
obtaining an acceptable uniformity of airflow in the
grain mass.

147

There are three main types of distribution systems:

elevated perforated storage floor;
above floor air ducts; and
in-floor air ducts.

fitted to the floor of the storage. In flat storages where

the grain height is low (less than 10 m), it is common
to install the ducts in a manner that makes them easily
removable to allow emptying of the storage with endloaders and other portable equipment-for instance,
round ducts may be secured with hoops or straps and
fixed to the floor with bolts screwed into recessed
sockets cast into the floor so as to avoid protrusions
which interfere with cleaning when the ducts are
removed (Fig. 2).
It is not possible to corrugate and roll plate thicker
than 1.6 mm, hence in deeper bins, where grain pressures are high, it is usually necessary to reinforce the
duets wilh steel framing (Fig. 3). In the case of sloping (e.g. conical) bin bottoms, it is necessary 10
securely anchor the ducts to the bin Iloor to prevent
them being dragged down the floor with the grain
(Fig. 4). Struetural analysis and design of these ducts
is an art rather than a science, since the load imposed
by,the grain on a duct is somewhat indeterminable
and, furthermore, the load capacity of a reinforced
corrugated steel duct is not easy to estimate accurately. Interestingly, in the writer's experience, few
duct failures have occurred in bins of up to 2500 t
capacity, whereas in larger capacity silos e5000 to
70000 t) built in the 1980s, even heavily reinforced
ducts have frequently failed as a result of unexpectedly high grain pressures.
An alternative to the use of corrugated rolled steel
ducts, is the use of ducts formed with stainless steel
'wedge-wire', as used for the manufacture of rotary
screens. Ducts formed from this material are
extremely strong, and hygienic in that they largely
prevent dust and brokens entering the duct through
the apertures. They are, however, very much more
expensive than corrugated perforated plate (Fig. 5).

Elevated perforated storage floor

The most efficient and uniform air distribution into
a grain mass is achieved with an elevated perforated
tloor in the base of the storage, which effectively provides a 100% contact area between the grain mass
and the incoming air (Fig. I). This type of floor is
effective only in flat floored storages, and is more
costly than the alternatives. It is, however, very commonly used in the central mid-west of the USA where
in-store drying is prevalent; some silo suppliers in the
region manufacture pre-formed Iloor panels which
lock together with light gauge support legs to form a
relatively low-cost elevated perforated storage floor,
suitable for cleaning out with a sweep auger.

Above-floor air ducts

More commonly, air is ducted into the grain mass
(or out of it, in the case of downward aeration) by
means of air-ducts that are either built into the storage
floor, or are placed on top of it. Above-floor ducts are
less costly, but make emptying of the storage more
difficult, particularly with end-loaders or sweep conveyors. Nevertheless, above-floor ducts are a practical engineering solution, particularly when retrofitting ducting to an existing storage Iloor, and they
can be particularly useful in self-emptying bin bottoms where their protrusion above the floor surface
does not inhibit the discharge of grain.
Above-Iloor ducts are usually made from corrugated perforated steel sheet which is rolled into circular or (more commonly) semi-circular shape and

elevated perforated metal flooring

from folded light gauge steel

folded light gauge plate floor

supports staggered as indicated

Figure 1.

Method of construction of elevated plenum floor of a type commonly used in the USA.

148

perforated duct

Figure 2.

angle bracket

Typical circular aeration duct often

used in horizontal
storages.

recessed fixing

half-round perforated duct

pipe supports and bracing
at (say) 600 centres

Figure 3.

angle base

half-

Reinforced

round ducting often

used in silo bases.

floor

.y

silo wall

steel shroud (reinforced)

air inlet duct

bolt flxings

Figure 4.

Reinforced

round

half-

ducling

as

used in conical silo

bases.

alt~rnative

'I

air inlet

J "

' \ silofloor

149

Duct Configuration
Horizontal storages

Figure S.

In 'horizontal' (rectangular) storages, ducts may be

aligned across the storage or along the length of the
storage. Both have certain disadvantages as follows:
Cross-floor ducts. In a storage where the grain is
peaked in the centre by use of a longitudinal conveyor, cross-floor ducts pass under varying depths
of grain. As a result, the airflow rate from the
outer ends of each duct will be higher than from
the middle unless the ducting is modified in some
way to minimise this, for instance by enlarging the
duct below the centre of the storage. This has the
combined effect of reducing the velocity head,
and thus increasing the static pressure. It also
increases the cross-sectional and surface areas of
the duct, thereby reducing losses in that region
(Fig. 8).
Longitudinal ducts. The floor of a large horizontal
storage is often only partially covered--e.g. when
the storage is partly filled. In such circumstances,
longitudinal ducting is exposed in the unfilled
areas, and large quantities of air will escape,
bypassing the grain mass.
Cross-floor ducting is usually the preferred option
for horizontal storages because it allows easier control of airflow into the grain.
Ducts should be spaced as follows:
For in-store drying. the duct spacing should be not
more than half the depth of the grain (Fig. 9).
(Grain depth must be uniform for in-store drying.)
For aeration, the duct spacing should be such that
the longest air-path from a duct to the grain surface
is no more than 50% more than the shortest path
from the same duct to the grain surface (Fig. 10).

Typical section through wedge-wire screen.

Alternative above-floor ducts can also be fonned

from 'solid' plate without perforations, in such a way
that allow air 10 escape around the bottom edges of
the duct (Fig. 6). These are not commonly used in the
writer's experience, but should work satisfactorily if
they are made strong enough to resist the grain pressures imposed on them (Fig. 6).

solid sheet cover

support
frame

air duct

Figure 6.

air escapes below

sides of duct

Non-perforated aeration duct, from which air

escapes from below side walls.

In-floor air ducts

In-floor ducts are generally more expensive than

above floor ducts, because the duct has to be formed
(usually in concrete), and covered by a perforated
steel plate placed flush with the floor surface to
allow unrestricted movement of equipment for cleaning the grain off the floor surface. The perforated
plate must thus be flat, and therefore substantial reinforcement must be provided, for example, in the
fonn of an open lattice grillage made from welded
steel flats (Fig. 7).
Alternatively, it should be possible to use the
folded perforated steel plate sections manufactured
for elevated silo floors to fonn covers for in-floor
ducts, although the writer has not seen this done.
Where portable equipment (such as end-loaders)
are used to remove grain from the storage, the wheel
loads imposed on the duct covers are often more
severe than the grain loads, and thus govern the
design of the cover supports.

steel grillage
.' -'

concrete floor

Figure 7.

air duct

Typical arrangement of in-floor ducting.

Flat-bottom silos
There are several options for air distribution systems in flat-bottom silos which are less expensive
than an elevated floor. The options include on-floor
ducts and in-!loor ducts.
Single radial duct. A single radial duct with an external fan can be used for aeration in small silos. However,
150

duct enlarged at shed centre increases

static pressure

Figure 8.

Arrangement of
cross-floor aeration ducts and

lateral
duct
enlargement in
a large shed.
place fans at
both sides of
wide storages

cross-floor aeration duct

storage building

Figure 9.

Duct spacing for

in-store drying.

min grain depth

(at side wall)

duct spacing < O.SH

H1 + 01 ~ H2 + 02 < 1.5L
H3 + 02 = H4 + D3 < 1.5L2

aeration ducts

151

Figure 10. Recommended

spacing of longitudinal ducts.

Quarter-round comlgated perforated steel plate was

used to form the ducts, and each section of duct was
hinged against the silo wall to permit easy access for
cleaning brokens and dust.
Circumferential ducts. In 'squat' silos (i.e. silos
with small wall-height to diameter ratio), the grain
depth varies significantly from the outer perimeter to
the centre of the silo. The length of airflow path (and
hence the static pressure needed to force the air
through the grain mass) thus varies significantly
from the silo wall to its centre. Parallel ducts placed
in the floor (even an elevated bin tloor), are likely to
result in a concentration of airtlow towards the outer
perimeter of the grain mass (where the airllow resistance is least) and substantially less airtlow in the
centre of the silo. The author has recently designed a
system of circumferential in-tloor ducts for grain
aeration in such silos, each fitted with its own fan (or
fans), and each duct and fan 'set' individually sized
to suit the mass of grain above it. Thus, the fan serving the centre of the bin has a relatively low airtlow
rate but high static pressure, while the fan (or fans)
serving the perimeter of the bin have a high airflow
rate (because of the larger volume of grain around
the perimeter) and a low static pressure (because of
the lower grain depth). The system is so far untried.
but is expected to be put into use on a project in
China in the near future (Fig. 14).

for large silos the duct size becomes excessive, and the
air distribution pattern is not adequate. Single radial
ducts are unlikely to be acceptable for in-store drying
in silos. Multiple radial ducts are seldom (if ever) used
because each duct requires a separate fan.
Parallel ducts. Parallel floor ducts are commonly
used for both aeration and drying, differing only in
their size and spacing. Duct spacing requirements
define the number and layout of the ducts. However.
the designer has to select the manifolding arrangement for the fan or fans, unless each duct is to be fitted with a separate fan. A 'tree' formation of
manifold and ducts is often adopted where multiple
ducts are fed from a single fan since this minimises
the effective length of the path for the ducted air
(Fig. 11).
V-ducts. V-formation ducts are commonly used to
distribute air from a single fan into the base of flat
bottom silos (Fig. 12). For in-store drying, this
arrangement is suitable for only relatively small
diameter bins, since in larger bins the space between
the ducts becomes excessive. V-formation ducts are
almost always recessed into the tloor of the silo.
Perimeter ducts. Perimeter ducts have been successfully used for grain aeration in some relatively
large (3000 t) tlat-bottom silos designed by the author
(Fig. 13). The aim was to provide good air distribution while leaving the silo floor free of protuberances.

manifolding duct

parallel ducls
in silo base

reclaim tunnel

Figure 11. Typical parallel duct arrangement in a flat floor silo base.

152

Figure 12. Typical V-duct arrangement in a flat-bottom silo.

Figure 13. Arrangement of perimeter

ducting in relatively large
flat-bottom silos.

153

aeration vent

middle core of grain

outer core of grain

undersilo
conveyer

intermediate pressurel
volume fans
grain discharge chute

low pressure high volume fan

high pressure low volume fan

ELEVATION

inner duct (one fan)

second duct (one fan)

....- third duct

(two fans)

- outer duct
(two fans)

PLAN VIEW

Figure 14.

Aeration

fan and duct arrangement for a 40 m

diameter,

Hopper bottom (self-emptying) silos

10 000 t silo.

silos since there is no special advantage (but large

extra costs) in building them in-floor.

Hopper bottom silos are not often used for in-store

drying of grain, since they are usually associated
with high throughput requirements, which are generally inconsistent with the requirements for in-store
drying. Aeration equipment is, however, often fitted
to hopper-bottom silos to allow the silos to be used
for 'holding' the grain for short periods without
deterioration before shipping. Aeration is also often
fitted to hopper bottom bins for cooling of grain after
drying.
Aeration ducting in hopper-bottom silos is usually
either perimeter-type or radial-type or a combination
of both. Above-floor ducts are used in hopper bottom

Vertical ducts
Vertical ducts attached to the silo walls have occasionally been used for aeration of bins in such manner
that air is blown (or sucked) horizontally across the
grain in the silo, out from one duct and into the other.
Apart from requiring long ducts (full wall height)
which are difficult to install, the biggest problem with
them is the same as with longitudinal ducts in horizontal storages-they must be covered with grain if
they are to work, and it is not possible to aerate a partially filled silo with them. There are few, if any, circumstances where they can be recommended.
154

Selection or air-duct sizes

Airflow Resistance in Ducts

Much work has been done recently in refining the

methods of detennining the optimum size of ducts for
distributing air into a grain mass, and software is now
available for carrying out these calculations (Wilson,
S.G. 1991. Duct: a PC Program for designing seed
store ducts. Personal communication). However, in
the absence of such tools, there are two well recognised 'rules of thumb' which can be used to estimate
the size of ducting that will minimise head losses.
The cross-sectional area of the duct should be large
enough such that the velocity of the air in the duct
is not more than 10 m/second.
The surface area of the duct should be large
enough such that the escape velocity of the air
from the duct into the grain is not more than 0.15
mlsecond (assuming a duct perforated area of
approximately 10%).
Commonly used perforations suitable for grain aeration are 2.5 mm diameter holes at 6.25 mm (triangulated) centres, giving an open area of around 15 %.

Air movement in ducts

Air passing through ducts will lose energy from
frictional effects associated with the duct shape, surface, turbulence, etc. This loss of energy results in a
gradual loss in 'total pressure' (or head) as the air
moves through the duct.
Total pressure is the sum of the static pressure and
the velocity pressure. Static pressure is the 'normal'
pressure that the air applies to the walls of the duct
that it passes through; velocity pressure is the component of pressure that results solely from the movement of the air, and is equal to V2/2g, where V is the
velocity of the air, and g is the acceleration due to
gravity. As air passes through a duct system, its
velocity will change where the duct size changes;
thus, where the duct size increases, the velocity pressure will decrease and the static pressure will
increase, and vice versa. Along the duct system, however, the total pressure (the sum of the velocity and
static pressures) will gradually diminish.
Airflow resistance in ducts depends on the size,
shape, and smoothness of the duct and the velocity of
air passing through it and the temperature of the air.
Friction losses in straight ducts are normally determined from charts. Elbows, bends, T-junctions, etc.
can add significantly to friction losses-for instance,
a single right-angle 'mitre' bend in a square duct is
equivalent to adding a straight duct with a length of
about 75 times the depth of the duct. (Note: giving the
bend a radius equivalent to the depth of the duct
reduces the friction loss to an equivalent length of
duct 1I times the depth.)
In most grain aeration and drying situations, duct
lengths are short and few bends are required. In most
circumstances, provided the air-velocity is kept low
enough (see below), friction losses in ducts are small
compared with the losses in the grain.

Airflow Resistance in Grain

There are various methods of detennining the resistance of airflow through a grain mass, from which the
fan characteristics and power requirements can be
determined. Holman (1966) presented a simple
graphical method in the form of curves for different
types of grain, relating bed depth to static pressure,
from which fan power could be detennined.
Shedd's formula (Equation 4) is another method
which can be used to calculate the airflow resistance
per unit metre of grain depth:
I'J" =

The pressure drop through a perforated duct covered with grain can be estimated from Equation (3)
(Brooker et al 1992):

l.oileOf J2
Qa

(3)

where: e ~ void space in grain (%)

and

(4)

where I:J.P' = pressure drop per unit depth of grain

Q a ~ airflow rate per unit area of floor
a and bare Shedd's constants which vary for
each type of grain. Constants for some common grains are given in Table 5.
Shedd's fonnula gives results that are satisfactory
for clean grain in small storage, but it does not reliably
account for dirty grain or for deep masses of grain. It is
suggested that the calculated pressure drops be
increased by at least 50% to allow for factors that
could restrict the airflow. Various studies have been
conducted under a range of grain conditions from
which Shedd curve multipliers can be detennined
based on variations in fines content, moisture content,
filling method, and airflow direction (Brooker et al.
1992). It appears from some of these results that:

Air movement through duct perforations

AP

aQ.;
fn(l + bOa)

Of - percentage opening in duct surface

(%).

where: Of> 10%, the pressure drop through the duct

can be ignored.
155

 a 3-5% fines content can increase airflow resistance by 50% above that for clean maize;
the pressure drop per metre of maize reduces by
about 50 PaJmetre per m 3/sec/m 2 of air flow (Qa)'
for each I % increase in moisture content; (the pressure drop for wheat changes by 200 PaJmetre per
Qa per 1% increase in moisture content);
the use of mechanical grain spreaders in the inlets
of silos ean reduce airflow resistance significantly
below Shedd's values; and
the resistance to airflow through grain in a horizontal direction is significantly lower than in the vertical direction.

Table 5.

normal and airflow rate higher than predicted-for

example, when the storage is only partially filled.
Fans with forward curved and straight blades will
draw more air (and current) under such conditions,
causing motor overloading.
Where a storage is designed so that it can be sealed
for fumigation, the designer should specify fans that
are readily sealable, and which have no exposed copper components which can be chemically attacked by
phosphine gas. In the case of centrifugal fans, this may
require minor modifications to the fan casing, for
instance to allow bolting of a 'blanking' plate over the
air inlet, and a gland or stuffing-box seal should be fitted where the impeUer shaft enters the fan casing.
Axial fans may need to be removed during fumigation
with phosphine unless the motor and wiring can be
sealed sufficiently to prevent entry of phosphine.

Shedd's constants for airflow resistance.

Grain

Wheat

2.70 x

8.77

Shelled maize

2.07 x 104

30.40

Paddy

2.57 x 104

13.20

Sorghum

2.12 x 104

8.06

Barley

2.14x 104

13.20
13.90

Oats

2.41 x 104

Rapeseed

3.99 x 104

4.20

Soybeans

1.02 x 104

16.00

Sunflower

2.49 x 104

23.70

Conclusion
The paper describes the basic principles for the
design of aeration and in-store drying systems. Space
precludes a complete coverage of the subject, and
designers should refer to some of the standard texts in
the reference list for more detailed information, particularly relating to particular circumstances. The
design parameters offered in this paper have been
used by the author in a number of installations, and
they have given good service in practice, which indicates that the parameters are at worst conservative
and at best reasonably sound.

Source: (ASAE Standards 1988)

Selection of Fans
References

Having determined the required airflow rate and the

pressure drop in the system, it is a relatively easy matter to select the appropriate fan to meet the performance criteria.
Selection is best made from performance curves
supplied by manufacturers. However, there are a few
basic guidelines that designers should be aware of:
Axial fans are most suited where static pressures
are predicted to be less than I kPa. Above 1 kPa,
centrifugal fans are likely to be more efficient.
Thus, axial fans are usually used for aerating shallow depths of grain, whereas centrifugal fans are
more commonly used on vertical silos.
Axial fans are not generally suited to grain drying
Axial fans are noisier than centrifugal fans because
they usually operate at around 3000 rpm (compared
with J 500 rpm for centrifugal fans). Sound attenuators may need to be fitted in noise-sensitive localities.
Centrifugal fans should normally be of the backward-curved-blade non-overloading type. Use of
this type of fan avoids overloading of the motor in
situations when the airflow resistance is lower than

ASAE (American Society of Agricultural Engineers) 1988.

Standards 1988. ASAE Data: ASAE D 272.1. St Joseph,
Ml., ASAE.
Brooker, D.B" Bakker-Arkema, F.W., and Hall, C.W. 1992.
Drying and storage of grains and oilseeds. New York,
Van Nostrand Reinhold.
Foster, G.B. and Tuite. J. 1982. Aeration and stored grain
management. In: Christiansen. c., ed .. Storage of cereal
grains and their products. St Paul, MN . American Association of Cereal Chemists.
Holman, L.E. 1966. Aeration of grain in commercial stOfages. Marketing Research Report No 178. United States
Department of Agriculture.
Hunter. A.1. 1986. Design of air distribution systems and
fan selection for grain aeration, In: Champ, B.R. and
Highley E . ed., Preserving grain quality by aeration and
in-store drying: proceedings of an international seminar,
Kuala Lumpur, Malaysia. October 1985. ACIAR Proceedings. No 15. 108-113.
USDA (United States Department of Agriculture) 1965.
Drying shelled corn and small grain. USDA Farmers Bulletin 2214.

156

Further Reading

tions. Journal of Agrieultural Engineering Research, 28.

437-450.

Bowrey. RG. and Driscoll, RH. 1985. A procedure to estimate the supplemental heating needs for in-store drying of
paddy. In: Champ. B.R. and Highley, E., cd., Preserving
grain quality by aeration and in-store drying: proceedings
of an intemational seminar. Kuala Lumpur, Malaysia,
October 1985. ACIAR Proceedings, No 15, 114-117.
Driscoll. RH. 1985. The application of psychrometries to
grain aeration. In: Champ, B.R. and Highley E., ed., Preserving grain quality by aeration and in-store drying: proceedings of an international seminar. Kuala Lumpur,
Malaysia. October 1985. ACIAR Proceedings. No 15,
67-80.
Hunter, A.J. 1983. Pressure difference across an aerated
seed bulk for some common duct and store cross-sec-

Hunter, AJ. 1986. Thermodynamic criteria for optimum

cooling performance of aeration systems for stored seed.
Journal of Agricultural Engineering Research, 33, 83-89.
Navarro, S. and Calderon, M. 1982. Aeration of grain in
subtropical climates. Rome, FAO Agricultural Services
Bulletin, No 52.
Sutherland, J.W . Banks, PJ. and Griffiths, HJ. 1971. Equilibrium heat and moisture transfer in air flow through
grain. Journal of Agricultural Engineering Research, 16,
368-386.
Wilson, S.G. 1994. A formulation that allows for variable
seed depth in calculating flow along seed-store ducts. Journal of Agricultural Engineering Research, 58,191-199.

157

Control Systems for Aeration and

Drying of Grain
G. Srzednicki*

Abstract
Automatic control systems are becoming increasingly popular in grain drying processes. They include
measurement, steering, monitoring and recording functions. The advantages include improvement in the
product quality, minimisation of energy use, optimisation of labour inputs, and increased effectiveness. A
good understanding and analysis of the drying process is essential for the design of a control system. The
control systems for in-bin aeration and drying systems aim to regulate the fan operations. They will also
control heater operations whenever heat is added to the drying air. In continuous-flow dryers, the parameters
controlled are the drying air temperature and the speed of augers. The most common control systems used
for grain drying are discussed. With increased interest in mechanical drying of grain under tropical conditions. the possibilities for introduction of automatic controllers are currently being investigated. This is occurring as a spin-off of the adaptive research in in-store drying that has been conducted over the last ten
years by several ACIAR projects in Southeast Asia. It can be expected that the increasing adoption of mechanical drying systems in the countries of the region will be followed by the introduction of control systems to optimise the use of the equipment.

instead of time-consuming building and testing of

prototypes.
The control systems originated in the eighteenth
century in the fonn of a speed controller for a steam
engine designed by James Watt. Modern control theory emerged between World Wars I and n, through
the work of Minorsky on ship steering controls that
could be derived from differential equations describing the system. Later, Nyquist developed procedures
to detennine the stability of closed-loop systems
based on the response of open-loop systems to the
response of steady-state sinusoidal inputs (Ogata
1990). A very rapid development of control systems
took place on the arrival on the market of affordable
digital computers. This led to the widespread use of
control systems in diverse areas including manufacturing, biological and biomedical applications, and
transport and entertainment.
A sound understanding of the process that is to be
controlled is essential for the design of a control system. The initial analysis examines the requirements
of production and the capacity of the equipment used,
the proposed control strategies, and the schedule of
operations. It also includes an initial analysis of the
potential benefits of the system.

AUTOMATIC control plays an increasingly important

role in the modem production processes and logistics.
Even a well-trained operator cannot optimise a process purposefully and unequivocally. This is because
production processes often involve non-linear relationships and dependencies. Rapid progress in the
theory and practical applications of control systems
has contributed to an improvement of productivity
and perfonnance optimisation. Also, reduction of
fatigue of operators performing monotonous and
repetitive tasks has resulted in a decrease in human
errors. Operations involving temperature, humidity,
pressure, speed, and viscosity will nonnally be controlled by microprocessor-based systems. Mathematical models are thc basis for computcr simulations that
enable engineers to define the statistical limits of values for parameters that are usually subject to random
effects. The great advantage of this approach is the
move from an empirical to a model approach, where
changes in values of parameters can be tested quickly

Department of Food Science and Technology. University

of New South Wales, Sydney, New South Wales 2052,
Australia.

158

Mechanical drying, as an industrial operation, has

seen applications of control systems for many years,
especially for large-scale operations. However, with
the decrease in cost of microprocessors, control systems have become increasingly accepted to capture
the same advantages that they have brought to other
areas of industrial production.

ever, some decisions have to be made by the operator

according to established procedures and translated
into action by adjusting setpoints or switches.
Automatic systems
Here, the function of the operator is largely taken
over by a microprocessor that is programmed to interact between the information supplied by the sensors
and the appropriate action to be taken by relays or
actuators. The system will automatically handle most
of the alarm conditions and also provide automatic
reporting. There is still need for an operator to reset
the system in case of a power failure or emergencies,
but their involvement will be minimal as compared
with the two previous systems.
Automatic control systems use algorithms based on
mathematical models representing physical and chemical characteristics of the process, and information
obtained through simulation or experiments. Mathematical simulations describe the process with sufficient accuracy to predict the result without the need for
use of empirical methods. The introduction of automatic controls results in optimisation of performance
of dynamic systems, in increased productivity, and in a
reduction in drudgery and the risk of human error.
Control theory uses specific terminology that is
defined in a number of textbooks (e.g. Shinskey
1988; Ogata 1990). The following terms are used
throughout this paper:
A controlled variable is a quantity or condition that
is measured and controlled.
A manipulated variable is the quantity or condition
that will be changed in order to achieve the desired
value of the controlled variable.
A closed-loop system or feedback control is a system that reduces error in order to produce required
output. The latter is generally referring to the product created by the process.
An open-loop system is one in which there is no
relation between control action and output.

Grain Dryer Controls

Drying of grain to reduce the water activity of the
product is the safest method of preserving grain from
further deterioration during storage. The main types
of dryers are in-bin dryers, using ambient air with or
without supplemental heat, and continuous-flow dryers using heated air. There are also combinations of
both systems, developed within the concept of twostage drying. In a two-stage drying system grain is
dried using a high-temperature fast dryer down to a
moisture content of about 18% wet basis (w.b.), and
then down to a safe storage moisture content (generally 14% w.b.) in an in-bin dryer using near ambient
air. The latter is called an in-store dryer since grain
will be stored after drying in the same bin or store
(Srzednicki and Driscoll 1994). During extended
storage, grain will usually also be aerated in order to
further reduce the risk of deterioration.
The two main types of grain dryers require specific
control systems related to their mode of operation.
However, the principal objectives of the control systems are the same:
uniform moisture distribution in the mass of grain
without over- or under-drying;
prevention of grain deterioration;
optimum grain quality;
maximum capacity; and
low energy consumption.
Controllers, relevant components of the dryer, and
the operator interact to achieve these objectives. The
level of the interaction is used to classify different
types of controls as follows (Pym and Adamczak
1986).

Control Systems for Continuous-flow

Dryers

Manual systems
Information about the process is collected by the
operator on dials or chart recorders and analysed
according to established procedures or experience.
The decision is made by the operator and translated
into action through manually operated switches or
levers.

Continuous-flow high-temperature dryers are used

for drying large quantities of grain, generally in several passes, from field moisture down to at least 18%
w.b. moisture content. This is a continuous operation
involving grain of different initial moisture contents
and therefore requiring appropriate adjustments of
the residence time. The main types of continuousflow dryers used in the grain industry are:
crossflow;
concurrent-flow;
counterflow; and
mixed-flow.

Supervised systems
These could also be called semi-automatic systems. The operator receives information about the
ongoing process on displays with adjustable setpoints
that are providing a partial feedback control. How159

A comprehensive review of control systems for

continuous-flow dryers can be found in papers by
Bakker-Arkema et al. (1990) and Moreira and
Bakker-Arkema (1992). The following cover the key
points.
The residence time of grain in the dryer is the factor that determines the final moisture content and
consequently the quality of the grain. Over-drying
and under-drying are undesirable and lead to reduced
profits of the plant. The operator or the control system has to adjust the speed of the unloading auger for
a given temperature and initial moisture content of
the grain. Control systems have been developed in
order to minimise over-drying and under-drying of
grain as compared with the manual intervention of
the operator. An evolutionary process led to development of five generations of controls.

Optimal feedback control

This type of control uses an algorithm based on a

process model approximating the drying model. The
technique used attempts to minimise the governing
equation by quadratic programming. In spite of the
great accuracy of this type of control, the difficulty in
defining the governing equation and a high computational requirement make this type of control unlikely
to be adopted on a commercial scale.
Feedforward control
The principle of this type of control is based on a
continuous or intermittent measurement of the main
load variable, i.e. the initial grain moisture in a continuous-flow dryer. The residence time is calculated
by the model based algorithm and the response time
is very short. Differential equations have been developed for the transient analysis of concurrent-flow,
crossflow, and mixed-flow dryers. Empirical models
have then been derived from the differential equations in order to simulate the operation of each of
these three types of dryer. Whitfield (1986) used a
steady-state simulation to predict unsteady states
caused by varying inlet moisture contents for controlling concurrent and counter-flow grain dryers.
Moreira and Bakker-Arkema (1990a) developed a
partial differential equation model that describes the
unsteady-state operation of a two-stage concurrentflow dryer for maize. This model was used to design
an automatic control system for such a dryer. Platt et
al. (1991, 1992) combined a feedforward and feedback control in order to incorporate the use of
optional features for cross-flow dryers, such as
inverters (aiming at reducing gradients in grain temperature and moisture across the dryer) and air recirculation. Bruce and McFarlane (1993) developed a
feedback-plus-feedforward algorithm that eliminates
the influence of variations in drying air temperature
in mixed-flow dryers.

Classical feedback control

In earlier days, the operator checked the inlet and
outlet moisture content on an hourly basis and adjust
the speed of the auger manually. The first automatic
controllers were temperature-triggered feedback controllers, measuring the temperature of the grain or of
the exhaust air at different locations in the drying column. However, because of the non-linearity of the
drying process, and lack of a consistent relationship
between the moisture content and the temperature of
the exhaust air, such systems were unreliable, especially if the moisture content changes exceeded 3%
w.b. For this reason, the temperature-activated systems have gradually been replaced by moisture-activated ones.
The classical feedback controls are based on proportional-integral (PI) or proportional-integral-derivative (PID) closed-loop controls using algorithms that
minimise the error between the output and the controlled variable. These controllers perform well if the relationship between the controlled and the manipulated
variable is linear and if the time response is short
(Whaley 1995). As already mentioned, nowadays, the
controlled variable in this case is the grain moisture
content. However, the classical feedback system is
unstable if the moisture range is wide. This is due
again to non-linearity of the drying process and the
grain flow. An attempt has been made by Whitfield
(l988a,b) to overcome this problem by modifying the
control algorithm so as to increase the bandwidth of
moisture range. This implies an initial tuning of
parameters by the operator as well as recalibration,
especially if there are differences in the bulk density of
the grain. Another problem, common to all feedback
controllers, is a slow response of the controls in case
of long residence time in the dryer. In recent times,
attempts have been made to increase the robustness of
the controls by using linearising transformations in the
control software (Courtois et al. 1995)

Adaptive control
The previously described feedback and feedforward
controllers are characterised by fixed parameters used
in the control law. Since a dryer operating under normal commercial conditions faces the problem of fluctuations in inlet grain moisture and temperature of the
drying air, this type of control system often results .in
inaccurate control, leading to increased variability of
the final moisture content in grain. The adaptive control system enables the controller to tune itself in spite
of such fluctuations. A number of researchers worked
on the process model for adaptive control. A moisture-activated feedforward controller for a crossflow
dryer has been developed by Forbes et al. (1984).
Nybrant (1988) developed an adaptive feedback controller for crossflow dryers, based on a linear model
160

using exhaust air temperature. Eltigani and BakkerArkema (1987) developed an adaptive model based
on feedforward control for crossflow dryers. It is a
two-term linear model with model parameters estimated by the sequential least square method. An adaptive feedforwardlfeedback control for crossflow
dryers has been developed by Moreira and BakkerArkema (1990b). Two linear models have been proposed by the authors, namely the generalised minimum variance controller and the pole placement
controller. The pole placement controller was found to
be faster, required fewer parameters to be estimated,
and was easier to implement. It controlled the final
moisture content of grain within O.I cC of the setpoint for a moisture content differential of 2.3% w.b.
in the incoming grain.

PM2

The term 'in-bin dryers' describes a drying system

that uses a fan and an air distribution system that will
force drying air through the bulk of grain. The grain
is contained in a bin that can be a metal or concrete
silo or a horizontal warehouse. There is a range of
types and sizes, varying in capacity from about 30 to
several thousand tonnes. There are also various types
of air distribution systems, such as perforated floors,
in floor or on-floor horizontal perforated ductings,
and vertical perforated ductings.
In-bin dryers can be one or other of two types. In
the high temperature type, hot air is blown through a
shallow depth of grain in order to achieve rapid drying. In the other type, air of near ambient temperature, or only slightly heated (usually by a maximum
of SOC above ambient) is passed through the grain
mass. The relativc humidity of the drying air corresponds to the equilibrium moisture content of grain
that can be stored safely without significant deterioration. Grain can be dried in deep layers in the same
bin, silo, or warehouse in which it will be stored. This
technique is also called in-store drying.
During extended storage, grain that has already
been dried, will need to be aerated in order to:
reduce the grain temperature and prevent it from
deterioration;
equalise the temperature throughout the bulk of
grain;
control insects by preventing formation of hot
spots; and
prevent formation of off-odours.
The control system for in-bin drying or aeration is
of the open loop type, which means the output is generally not measured or fed back to be compared with
the input. The control acts on the fan and heater in
high-temperature dryers and in-store dryers equipped
with heaters, or on the fan alone in in-store dryers
using only ambient air and in grain aeration systems.
The measured variable is the temperature or relative
humidity of ambient air, whereas the controlled variable is the temperature or the relative humidity of the
drying air. The grain temperature is an indication of

The complexity of the control objectives, especially when the quality aspects of grain are concerned, often make them difficult to achieve using
conventional controls based on a set of governing
equations. This is often due to insufficient understanding of such phenomena as breakage susceptibility. Therefore, research has been carried out in recent
years on fuzzy logic control and neural network strategies control that combine mathematical models of
the drying process and the experience of the operator.
The result is a set of rules, kept in a knowledge base,
that convert linguistic control strategies into automatic commands (Zhang and Litchfield 1993, 1994).
The rules are represented as membership functions
called fuzzy membership matrices. The difficulty in
implementing this type of control is that there is a
large number of rules associated with fuzzy membership matrices. In order to implement the fuzzy logic
control, the membership functions have to be finetuned using a wide range of process conditions,
which often proves very time consuming. Although
very promising at an experimental stage, fuzzy control usually requires a considerable amount of his tor ical process data in order to be included in industrial
applications.

Control performance assessment

Douglas et al. (1992) describe a method for quantification of control performance of moisture controllers for continuous-flow dryers. The authors
introduce two indicators of performance called performance measures PM I and PM2. They are defined
as follows:
=

SNToutIS out

(2)

Control Systems for In-bin Dryers

Fuzzy logic (expert) controllers

PMI

= OSallowabldOSactual

where OSallowable = percentage amount of acceptable off-specification product, and OSactual = percentage amount of actual off-specification product.
The performance measures reflect the product
moisture variation and the deviation of the mean
from the target. The higher the value of the two indicators, the better is the performance of the dryer.
Both performance measures can be used to compare
performance of various types of continuous-flow
dryers.

(I)

where SNT = outlet standard deviation of the moisture under 'no touch' control, and Sout = outlet standard deviation of the moisture with control.
161

the progress of drying. In grain aeration, grain temperature may become the measured variable, triggering the fan action.
There are various performance criteria associated
with the use of control systems in in-bin drying
(Bakker-Arkema et al. 1990; Ryniecki et al. I 993b).
Among the most commonly used are:
energy consumption;
ovcr-drying;
spoilage estimated by different methods (e.g. spoilage index, dry matter loss, etc.); and
drying time.
Over twenty strategies have been proposed to
meet these criteria (Moreira and Bakker-Arkema
1992). They involve a range of parameters such as
time, temperature, or relative humidity of the drying
air, target moisture content in the top layer of the
grain bulk, or estimated dry matter losses. Grain drying mathematical models estimating the average
moisture content of the grain bulk or grain quality
criteria have been used to write control algorithms.
Ancillary equipment aimed at improving the uniformity of drying (grain stirrers) and reduce the
energy consumption (recirculation) has been
included in the control systems with varying degrees
of success (Srzednicki and Driscoll 1994). Studies
aimed at optimising values of control parameters are
usually based on repeated computer simulations
involving a large amount of historical weather data.
Stochastic models of heat transfer in a thin-layer of
grain and of ambient air temperature variation have
been developed in order to further optimise the value
of control parameters (Ryniecki and Jayas 1992).
Most of the control strategies are based on experience, but some involve mathematical optimisation
techniques (e.g. Ryniecki 1991; Ryniecki and Nellist
199Ia,b; Ryniecki et al. I 993a,b). Ryniecki et at.
(I993a, b) compared two locations with different climalic conditions, one maritime (England), the other
dry continental (Canada). A large number of computer simulations using 20 years of weather data on a
hourly basis has been run using different control
strategies, based on relative humidity of the drying
air with fixed or variable power heater or without
heater, for near-ambient air temperature drying of
wheat. The above-mentioned performance criteria
have been used to assess the effectiveness of the
process. The authors found that a fan and heater
combination with variable power heater was the
optimum system for the humid maritime conditions
of England, whereas a fan-only system proved to be
the optimum for the dry continental conditions of
Canada. Furthermore, it was found that, for the fanonly system, variable airflow significantly reduced
over-drying, energy consumption, and subsequently
the drying cost as compared with fixed airflow.
However, a variable speed drive needed to vary the

airflow rate implies additional investment in the drying plant. As a result of this work, it appears that the
choice of the strategy depends very much on the climatic conditions of the site and requires a careful
study of the weather conditions before the selection
of a control strategy.
In high-temperature in-bin dryers the control is
focusing on the temperature of the drying air and is
the simplest of control systems for in-bin drying. As
for the near-ambient temperature drying, the following strategies have been commercially adopted:
continuous aeration with additional heat (max. ~T
= 6C);
relative humidity control with upper limit;
relative humidity control with lower and upper
limit; and
self-adjusting equilibrium moisture content.
Grain aeration being a technique to maintain the
quality of grain in storage consists in blowing cool air
through the grain bulk. In temperate countries this
process can involve cold ambient air during winter
months, cool air during the night, or artificially
cooled air if ambient air temperature is high. The
control strategies eonsist in aerating either at regular
intervals, using a timer, an air temperature based time
proportioning controller (Pym and Adamczak 1986),
or a grain temperature based system, activated by
increasing temperatures in the grain. The difficulty
with the time proportioning controller is that it disregards the relative humidity of ambient air and introduces the risk of grain rewelting.

Applications of Control Systems under

Humid Tropical Conditions
Southeast Asia is one of major grain-producing areas
in the humid tropics. Rice and maize are by far the
main cereal crops grown in the area. Two or more
crops per year are becoming a permanent feature in
most of the countries. This implies that some of the
annual crop will be harvested during the wet season.
The region is currently characterised by very rapid
economic growth resulting in higher purchasing
power and increased demand for quality grain. In
order to satisfy the domestic demand and the export
requirements, the grain industry is improving the productivity, resulting in increasing adoption of modem
technology in grain drying. Large grain complexes
and mills are nowadays using high-temperature dryers, and control systems are often supplied as an
option by the manufacturers.
The two-stage drying technique with in-store drying as a second stage is also being adopted by the
grain industry (Srzednieki and Driscoll 1995). The
technical feasibility of in-store drying under humid
tropical conditions has been established through col162

laborative work over the past 12 years between the

Australian team from the University of New South
Wales and research teams in the Philippines, Malaysia, Thailand and, more recently, Vietnam (Driscoll
1987; Rukunudin et al. 1988; Driscoll and Adamczak 1988; Srzednicki and Driscoll 1992). Already,
during the early stages of the research, some attention was given to control systems for in-store dryers
(Driscoll et al. 1989). As previously mentioned in
comparing the strategies used for humid maritime
and dry continental climate, in-store drying in the
humid tropics requires a fan and heater for the wet
season. Meanwhile, during the dry season, drying of
grain can be performed satisfactorily using a fan
only. The following four strategies have been
devised and tested:
a) time clock control (including continuous aeration);
b) drying air temperature control;
c) relative humidity control; and
d) modulated burner control.
Given the fact that mechanical drying is still a new
technique, slowly gaining ground against sun drying,
especially during the dry season, it is difficult to justify additional investment in automatic control systems. An experienced operator will often be capable
of switching on the fan at certain times or off when
relative humidity is outside a suitable range for drying. A thermostat will also provide a crude control of
the drying air. However, optimisation studies comparing drying cost for three of the above four strategies (a, c, and d) in different locations throughout the
region (Malaysia, Philippines and Thailand) in dry,
wet, and average years, for paddy and for maize, have
shown that relative humidity control was the least
expensive strategy followed by continuous aeration
(Srzednicki and Driscoll 1993). The trend to generate
supplemental heat by use of a rice hull furnace is
reducing considerably the cost of heat energy in Vietnam (Ph an Hieu Hien and Nguyen Le Hung 1995).
Rice hull furnaces for in-store dryers are also under
development in Thailand and the Philippines.
As far as aeration is concerned, a grain cooling system aiming at maintaining grain temperature at 1920C has been implemented at seventeen grain complexes of BERNAS, the recently corporatised rice
handling authority in Malaysia (Teoh and Hassan
1995). The control system is based on the grain temperature and cooling is triggered by the maximum
temperature grain is allowed to reach.

programmable logic controllers (PLC). PLCs handle

well a linear process with large time constants.
However, model based controls will often require a
large computational power and will require the use
of process computers. The prices of PLCs and
microcomputers have fallen considerably in recent
years, making them more affordable for potential
users. Sensors are an essential component of the
system. There is a wide range of qualities and
prices. Some of the sophisticated in-line moisture
meters for continuous-flow dryers may be worth an
equivalent of a medium capacity dryer. It is obvious
that their chance of adoption is limited to very large
capacity units, generally owned by large grain
authorities.
As far the control system for in-store drying is concerned, given that it is based on the temperature or
relative humidity sensors, plug-in data acquisition
cards offer a cost-effective alternative to PLCs. Plugin data acquisition boards plug directly into the
expansion bus of a Pc. Multi-function data acquisition boards include analog/digital converters, digital
input/output ports as well as counter/timer circuitry.
A multi-function card (Fig. I) can be interfaced with
a signal conditioning board for analog data input
(from sensors) and a relay board that triggers different actions required by the process (e.g. fan, heater,
dampers).

Conclusions
Control systems for grain dryers have shown rapid
development in the last fifteen years. For high-temperature dryers, the control parameters are exit air
temperature or the moisture content of the inlet and of
the exit grain on one hand, and the speed of the discharge auger on the other. The grain moisture based
controls are more accurate, but the cost of the system
is higher, primarily due to the cost of the sensor.
Among the commercially adopted systems are the
classical feedback control and the adaptive feedback!
feedforward control. The feedback control is generally slow if the residence time is long. The adaptive
feedbacklfeedforward controller provides fast
response and reduces the fluctuations of the moisture
content in the exit grain irrespective of fluctuations in
the moisture of the inlet grain.
As for the control system of in-bin drying, the controller acts on the fan and the heater, if the latter is
required. The climatic conditions of the location
determine which strategy is to be adopted. A single
strategy may only partially satisfy the performance
criteria (low energy consumption, prevention of overdrying, low spoilage, short drying time). Optimisation techniques are being used in order to reconcile
the performance criteria.

Hardware for Control Systems

The most commonly used process control methods
today incorporate digital controllers. Among them is
a considerable number of PID type feedback single
loop controllers. Larger control systems will use
163

Signal conditioning board

Computer plug-in 1/0 card

HIiH!..!!i!li'

Jo'igure 1.

Relay driver board

Data acquisition board with secondary boards (adapted from !DC 1995).
an international seminar held at Bangkok. Thailand, 10-12 Septembcr 1986. Canbcrra, ACIAR Proceedings No
19,58-68.
Driscoll, R.H. and Adamczak. T. 1988. Drying systems for
the humid tropics. In: Champ, B.R. and Highley. E .. ed.,
Bulk handling and storage of grain in the humid tropics.
Proceedings of an intemational workshop held at Kuala
Lumpur, Malaysia, 6-9 October 1987. Canberra. ACIAR
Proceedings No 22, 58-68.
Driscoll R.H., Adamczak, T., and Samsudin, A. 1989. Control options for in-store drying. In: de Mesa, S.M .. ed ..
Grain postharvest systems. Proceedings of the Tcnth
ASEAN Technical Seminar on Grain Postharvest Technology, Bangkok. Thailand, 19-21 August 1987. Bangkok,
1l1ailand, ASEAN Crops Postharvest Programme, 44-51 .
Eltigani, A.Y. and Bakker-Arkema, F.W. 1987. Automatic
control of commercial crossflow grain dryers. Drying
Technology, 5, 610--619.
Forbes. J.F., Jacobson, B.A., Rhodes, E., and Sullivan, G.R.
1984. Model based control strategies for commercial
grain drying systcms. Canadian Journal of Chemical
Engineering. 62, 773-779.
Moreira, R.G. and Bakker-Arkema. F.W. 1990a. Unsteadystate simulation of a multistage concurrent-now maize
dryer. Drying Tcchnology, 8, 1,61 - 75.
- 1990b. A feedforwardlfeedback adaptive controller for
commercial cross-flow grain driers. Journal of Agricul tural Engineering Research. 45, 107-116.
- 1992. Grain dryer controls: a review. Cereal Chemistry,
69(4),390--396.
Nybrant. T.G. 1988. Modelling and adaptive control of continuous grain driers. Journal of Agricultural Engincering
Research. 40, 165-173.
Ogata, K. 1990. Modem control engineering. Englewood
Cliffs. New Jersey, USA, l"rentice Hall, 963p.
Phan Hieu Hien and Nguyen Le Hung 1995. Drying
research and application in the Mekong Delta of Vietnam. Adoption of in-store drying technology in Southeast

Control systems are also being introduced for instore drying in the humid tropical climate of Southeast Asia. Relative humidity control and continuous
aeration prove to be the most successful techniques
for a range of climatic conditions in Southeast Asia.
However, the rate of adoption of controllers is slow,
except for large capacity dryers. This situation is
related to relatively high cost of control equipment as
compared with low margins for high quality grain.
Plug-in card based systems may become an attractive
option, promoting adoption of automatic controllers
and resulting in improvement of grain quality and
reduction of the cost of grain processing.

References
Bakker-Arkema. F.W . Kelley. R.L.. and Moreira. R.G.
1990. In-bin and continuous-flow dryer controls. Paper
90-6565. An ASAE meeting presentation. Chicago Illinois. USA. 18- 21 December 1990. SI. Joscph. Michigan.
USA. American Society of Agricultural Engineers.
Bruce. D.M. and McFarlanc. N.1.B. 1993. Control of mixedflow grain dryers: an improved fcedback-plus -fcedforward algorithm. Joumal of Agricultural Engineering
Research. 56. 225- 238.
Courtois. F .. Nouafo. J.L.. and Trystram. G. 1995. Control
strategies for com mixed-flow dryers. Drying Technology. 13.5-7.1153-1165.
Douglas, P.L., SuJlivan, G.R., and Whaley. M.G. 1992.
Control performance measures for grain drying proc esses. Transactions of the American Society of Agricultural Engineers. 35. 4, 1235- 1241.
Driscoll. R.H. 1987. Drying strategies for the humid tropics.
In: Champ, B.R .. Highley. E., and Remenyi, J.V., ed.,
Technological change in postharvest handling and transportation of grains in the humid tropics. Proceedings of

164

Asia. Paper presented at the 17th ASEAN Technical

Seminar on Grain Postharvest Technology, Lumut,
Malaysia, 25-27 July 1995.
Plat!, D., Rumsey, T.R., and Palazoglou, A. 1991. Dynamics
and control of cross-flow grain dryers. 1. Model development and testing. Drying Technology, 9(1), 27-60.
1992. Dynamics and control of cross-flow grain dryers.
2. A feedforward-feedback control strategy. Drying
Technology, 10(2),333-363.
Pym, G.R. and Adamczak, T. 1986. Control systems for the
aeration and drying of grain. In: Champ. B.R. and Highley, E., ed., Preserving grain quality by aeration and instore drying. Proceedings of an international seminar held
at Kuala Lumpur. Malaysia, 9-11 October 1985. Canberra, ACIAR Proceedings No IS, 118-128.
Rukunudin I.H., Samsudin, A .. Noon, S. and Kassim, B.
1988. Bulk in-bin storage of paddy under Malaysian
weather conditions. In: Champ, B.R. and Highley, E., ed.,
Bulk handling and storage of grain in the humid tropics.
Proceedings of an international workshop held at Kuala
Lumpur, Malaysia, 6-9 October 1987. Canberra, ACIAR
Proceedings No 22, 205-211.
Ryniecki, A. 1991. Control systems for near-ambient grain
drying. In: Chelkowski, J., ed . Cereal grain-mycotoxins, fungi and quality in drying and storage. Amsterdam
and New York, Elsevier Science Publishers B.V., 49951 J.
Ryniecki, A. and Jayas, D.S. 1992. Stochastic modelling of
grain temperature in near-ambient drying. Drying Technology, 10, I, 123-137.
Ryniecki A., Jayas, D.S., and Muir, W.E. 1993a. A generalized control-strategy for near-ambient drying of wheat
under Canadian-prairie climate. Transactions of the
American Society of Agricultural Engineers, 36(4),
1175-1183.
Ryniecki A., Muir, W.E .. and Jayas, D.S. 1993b. Optimisation of control systems for near-ambient drying of wheat
under maritime and continental climates. Postharvest
Biology and TechnOlogy, 2. 217-230.
Ryniecki. A. and Nellist, M.E. 1991a. Optimization of control systems for near-ambient grain drying: Part I, The
optimization procedure. 10urnal of Agricultural Engineering Research, 48.1-17.
- 1991 b. Optimization of control systems for near-ambient
grain drying: Part 2, The optimizing simulations. Journal
of Agricultural Engineering Research, 48, 19-35.

Shinskey, F.G. 1988. Process control systems. Application,

design. and tuning, 3rd ed. Singapore, MeGraw-Hill
International Editions, 526p.
Srzednicki. G.S. and Driseoll, R.B. 1992. In-store drying
trial in Malaysia. In: Naewbanij, 1.0., cd., Proceedings of
the 13th ASEAN Seminar on Grain Postharvest Technology, Bandar Seri Begawan, Brunei Darussalam. 4-7 September 1990. Bangkok. ASEAN Grain Postharvest
Programme. 197-210.
1993. Economic aspects of in-store drying in Southeast
Asia. In: Naewbanij. J.O., Manilay, A.A., and Frio. A.S.,
cd., Increasing handling processing and marketing efficiency in the grain postharvest system. Proceedings of the
16th ASEAN Seminar on Grain Postharvest Technology,
Phuket. Thailand. 24-26 August 1993. Bangkok, ASEAN
Grain Postharvest Programme. 252-264.
1994. jAdvances in research on in-store drying. In: Highley, E . Wnght, E.J . Banks, H.J., and Champ, B.R., cd.,
Stored product protection. Proceedings of the 6th International Working Confcrence on Stored-product Protection.
Canberra, Australia. 17-23 April 1994. Walling ford,
Oxon, U.K., CAB Intcrnational, 351-358.
- 1995. Adoption of in-store drying technology in Southeast Asia. Paper presented at the 17th ASEAN Technical
Seminar on Grain Postharvest Technology, Lumut,
Malaysia, 25-27 July 1995.
Teoh I.C. and Hassan, B.B. 1995. Paddy cooling in BERNAS grain complexes. Paper presented at the 17th
ASEAN Technical Seminar on Grain Postharvest Technology, Lumut, Malaysia, 25-27 July 1995.
Whaley, M.G. 1995. Using model-based controls in cereal
drying. Cereal Foods World. 40(1).19-22.
Whitfield, R.D. 1986. An unsteady-state simulation to study
the control of concurrent and counter-flow grain driers.
Journal of Agricultural Engincering Research, 33, 171178.
- I 988a. Control of a mixed-flow drier. Part I: Design of
the control algorithm. Journal of Agricultural Engineering Rescarch, 41, 4, 275-287.
- 1988b. Control of a mixed-flow drier. Part 2: Test of the
control algorithm. Journal of Agricultural Engineering
Research, 41, 4. 289-299.
Zhang, Q. and Litchfield, J.B. 1993. Fuzzy logic control for
a continuous crossflow grain dryer. Journal of Food Process Engineering. 16,59-77.
- 1994. Knowledgc representation in a grain drier fuzzy
logic controller. Journal of Agricultural Enginecring
Research, 57. 269-278.

165