Project Completion Report: Investigation of rice kernel cracking and its control in the field and during post-harvest processes in the Mekong Delta of Vietnam - APPENDIX 1 " pdf

63


Ministry of Agriculture & Rural Development

Project Completion Report

MS14: PROJECT COMPLETION REPORT

026/05VIE
Investigation of rice kernel cracking and its control in the field and
during post-harvest processes in the Mekong Delta of Vietnam

APPENDIX 1

Influence of harvesting time around grain maturity on rice cracking and head
rice yield in the Mekong River Delta of Vietnam





April 2010




64

APPENDIX 1. Influence of harvesting time around grain maturity on rice
cracking and head rice yield in the Mekong River Delta of Vietnam

ABSTRACT
Timely harvesting plays an important role in controlling rice cracking. Reduced whole rice grain
yield due to cracking causes the value loss and reduces the farmers’ income. Field experiments
were carried out to study the effect of harvesting time around crop maturity on rice cracking and
head rice yield for seven common rice varieties (OM1490, OM2718, OM2517, OM4498, AG24,
IR50404 and Jasmine) in three different locations during two cropping years (2006-2008) in the
Mekong River Delta, Vietnam. The results showed that the rice cracking was strongly influenced
by both the variety and time of harvesting around maturity. There was a general trend of increase
in percentage of cracked rice with late harvesting in relation to estimated grain maturity date.
The head rice yield also followed the same trend in response to delayed harvesting. A delay of 4-
6 days reduced the head rice yield by 11.3 % an average and up to 50 %. Similar trends were
observed in both wet and dry seasons. The large varietal difference in percentage of cracked
grain (0.9 to 60.5%) on 6 days after maturity date indicated that the level of rice cracking caused
by late harvesting time can be minimized by the selection of suitable varieties.

INTRODUCTION
Head rice yield, which is defined as the weight percentage of rough rice that remains as head rice
(the kernels that are at least ¾ of the original kernel length) after milling, is considered as the
main quality indicator because the broken rice has often half the commercial value of whole
grain rice. It has been shown that timeliness of harvesting can influence milling yield
significantly. Harvesting rice at crop maturity can give a maximum head rice yield (Kester et al.
1963, Bal and Oiha 1975). Any delay in harvesting time causes reduction of head rice yield (Bal
and Oiha 1975, Ntanos et al. 1996, Berrio et al. 1989) and extended delay in harvesting can lead
to significant losses in head rice yield. Berrio et al. (1989) showed that among 16 investigated
rice varieties studied the whole-milled grain was reduced by 18% when harvesting was delayed
by 2 weeks. However, it was also found that there was no impact of harvesting time on sensory
perception of rice (Champagne et al. 2005, Chae and Jun 2002).

The incidence of rice fissuring in the field has a potentially significant impact on head rice yield.

Cracking can develop in the field as a result of changes in grain moisture after the rice matures
due to hot sunny days followed by humid nights. Harvesting time affects proportion of cracked
rice and hence head rice yield. Large quantities of immature rice kernels can be detected in early
harvested rice. Immature kernels are usually thinner and defective, and are easily cracked during
subsequent milling (Swamy and Bhattacharya 1980). In contrast, late harvested grain is often
associated with a grain product that is too dry and more prone to fissuring. Investigations by
65

Chau and Kunze (1982) showed that cracking can develop in low-moisture content kernels (13%
or 14% wet basis) before harvesting as a consequence of the swings in relative humidity in the
atmosphere. Furthermore, improper post-harvest practices, such as a delay in threshing when rice
stacks are left in the field, can also provide the potential of moisture adsorption due to an uneven
moisture content and uneven maturity within the bulk rice (Kunze and Prasad 1978).

Reduced whole rice grain yield due to cracking is one of the major issues that directly reduce
income and availability of staple food to the farmers in the Mekong River Delta of Vietnam.
Mekong River Delta is the largest rice production region in Vietnam. The cracking or partial
fissuring of rice kernels may occur right in the paddy field due to incorrect harvesting time and
improper harvesting practices, and occur also due to adverse post-harvest drying conditions and
inappropriate milling operations. The weather pattern (temperature and humidity) in Mekong
River Delta is unique. The rice is grown and harvested in both wet and dry seasons. Weather
conditions at around harvesting period are different between the two seasons and this can impact
the rice fissuring and cracking during milling. However, there is no experimental data available
on the impact of harvesting time on rice cracking and head rice recovery on the rice varieties
grown at different seasons in the Mekong River Delta. This research work is an attempt to
systematically collect the rice cracking and head rice yield data based on field experimentations
in four consecutive harvesting seasons between 2006 and 2008. The main factor considered in
this study during the collection of data was harvesting time- before and after grain maturity. The
objective of this experiment was to evaluate the effects of harvesting time of several rice
varieties on the level of rice cracking and head rice yield in different seasons. This study will

assist to determine the optimal harvesting time for various rice varieties grown in the Mekong
River Delta.

MATERIALS AND METHODS
Rice samples
Experiments were carried out at three locations, namely Seed Centre (An Giang Province), Tan
Phat A Cooperative (Kien Giang Province) and Tan Thoi 1 Cooperative (Can Tho City) in four
consecutive harvesting seasons during two years (2006-2008). Seven rice varieties commonly
cultivated in these cooperatives and seed centre were selected for field experiments as shown in
Table 1. The grain maturity date of these rice varieties provided by the local extension centers
were in the range of 86-98 days (Table 1). The maturity date is defined here as the harvesting
date expressed in days after sowing (DAS) planned by the farmer as recommended by the
extension centre based on the predicted physiological maturity of the grain.




66

Table 1. Rice varieties and their maturity dates used for this study.
Variety Crop season Recommended MD
†
Experimental MD
††

Wet
92
OM1490
Dry
87-92

92
Wet
92
OM2718
Dry
90-95
92
Wet
90
OM2517
Dry
85-90
86
Wet
90
OM4498
Dry
90-95
91
Jasmine
Wet
95-105
98
AG 24
Wet
90-95
90
IR50404
Wet
90-95

92

†
Recommended maturity date (days after sowing) given by local extension centre.
††
Maturity date (days after sowing) chosen for this study.

Experimental design
Each experiment consisted of seven treatments corresponding to the harvesting times prior and
after expected maturity date for each of seven varieties. These varieties were grown in different
paddies at the three locations. There were seven harvesting times, two days apart commencing
six days prior to maturity date (MD) and ending at six days after MD. Experiments were
designed in RCBD (Random Complete Block Design) method with five blocks (Table 2).
Table 2. Treatments (harvesting days) in relation to the maturity date (MD). 0, +2, +4, +6 and -2, -4, -6 are
harvesting days after and before estimated maturity, respectively. A, B, C, D, E are the replication blocks.
Block
Treatment
(Days vs. MD)
A B C D E
1
(-6) -6A -6B -6C -6D -6E
2
(-4) -4A -4B -4C -4D -4E
3
(-2) -2A -2B -2C -2D -2E
4
(0) 0A 0B 0C 0D 0E
5
(+2) +2A +2B +2C +2D +2E
6

(+4) +4A +4B +4C +4D +4E
7
(+6) +6A +6B +6C +6D +6E
67


Experimental procedure
Some rice fields of selected varieties were chosen and used for the experiment. Wet season
experiments were sown in March-April and harvested in June-July (in 2006 some varieties were
grown in late wet season and harvested in September), while sowing and harvesting for dry
season experiments were November-December and February-March, respectively. Figure 1
depicts plot layout for harvesting time experiment of each rice variety. Grains were harvested
from 35 sub-plots of 1 m x 2 m (total harvesting area is 70 m
2
) at 7 harvesting dates according to
the treatments from 6 days before to 6 days after maturity date (MD) with five replications for
each treatment (Figure 1). Cutting and threshing operations were done manually using a sickle
for cutting the rice stalk. Rice was harvested in the morning to avoid intense sun light, aiming to
reduce natural cracking due to sudden change of moisture distribution inside the kernel when it
goes between wet night and dry day. After cutting, rice was transferred into a shaded area for
manual threshing and cleaning in which most bulk straw, chaff, immature grain, and very light
and fine impurities were separated from the grains. The straw and chaff were manually separated
and the grain was dropped though a cross-wind to remove the lighter impurities.

Samples were then transferred to the dryer after undertaking moisture determination. Samples
were gently dried at 35
o
C using a laboratory tray drier developed by Chemical Engineering
Department, Nong Lam University Ho Chi Minh City, Vietnam until the moisture content
reached 14 % wet basis. Then samples were once again cleaned to remove residual immature

grains, measured for moisture content using grain moisture tester (Kett Co. Ltd., Japan), packed
in nylon bags, and transferred into lab for rice cracking and head rice yield analyses.


1A 3B 4C 5D 7E
2A 1B 3C 6D 4E
3A 5B 1C 7D 6E
4A 2B 6C 3D 5E
5A 6B 7C 1D 2E
6A 7B 2C 4D 3E

7A

4B

3C

2D

1E



Figure 1. Plot layout of harvesting experiments for each rice variety.
Each plot is 2 m long and 1 m wide and the border around harvesting area is 1.5 m.
2
m

1 m
1.5 m

1 m
68


Measurements
Cracking of paddy rice
This is the direct indicator for effects of harvesting time on cracking. Three 150 g paddy samples
were taken from each plot to ensure the repetition of each plot. Grains were dehulled by hand to
avoid cracking developing during this procedure. Fifty dehulled grains were randomly checked
to count cracking grains under microscope, and cracking fraction calculated.

Head rice yield
Exact 180g of paddy was dehusked and then 100g of brown rice was whitened for 60 seconds
using a laboratory milling system. Whole kernels were separated by grader from broken kernels,
to determine head rice yield which is defined as the ratio of the mass of unbroken kernel to the
total mass of paddy rice. The head rice is composed of grains which maintain at least 75% of
their length after milling.

Statistical analysis
Data were analysed by statistical software Statgraphics® 3.0 (StatPoint, Inc.) using ANOVA
(Analysis of Variance) procedure.

RESULTS AND DISCUSSION
Level of rice cracking
Percentages of cracked grain before husking obtained from seven varieties in four consecutive
crop seasons, i.e., wet season 2006, dry and wet season 2007, and dry season 2008, are shown in
Table 3. For each rice variety, level of rice cracking before husking were significantly different
among harvesting dates (P<0.05). Early harvesting (before maturity date) showed lesser
proportion of cracked grains. There was a general trend of increase in the percentage of cracked
rice corresponding to delayed harvesting day in relation to the time of grain maturity (day 0). For

example, the highest level of rice cracking for all rice varieties was recorded at sixth day after
maturity date, which was the last harvesting date in this experiment (Table 3). This indicates how
important it is to harvest the paddy at correct time in its maturity period. The result of this
research work on Vietnamese rice varieties is in agreement with previous studies showing
negative impact of delayed harvesting to rice quality with respect to level of rice cracking
(Ntanos et al. 1996, Berrio et al. 1989). Any over-drying in the field (or in the panicle) can result
in increased number of cracked grains.


69


Table 3. Degree of rice cracking of seven rice varieties before and after maturity dates during two crop years.
Cracked grain (%) before and after maturity day
Variety Crop
season
-6 -4 -2 0 +2 +4 +6
Wet ‘06
0.80
a
3.20
a
9.60
bc
4.80
ab
10.80
bc
15.20
c

23.60
d
OM1490
Dry ‘07
1.87
a
0.53
a
2.27
a
2.80
a
5.60
a
14.40
b
22.40
c
Wet ‘07
2.00
a
2.13
a
2.27
a
1.07
a
1.33
a
2.13

a
2.40
a
Wet ‘06
0.40
a
0.40
a
1.20
a
2.80
a
10.80
b
4.00
a
5.20
ab
0M2718
Dry ‘07
2.40
a
0.67
a
6.27
b
2.00
a
3.20
a

7.20
b
8.53
b
Dry ‘07
1.47
a
2.00
a
3.60
a
5.73
a
16.00
b
33.60
c
60.53
d
Wet ‘07
3.47
a
10.27
b
15.73
bc
18.67
c
12.13
b

12.67
b
20.27
c
OM2517
Dry ‘08
0.67
a
1.73
a
3.33
a
8.13
b
9.33
b
14.13
c
25.73
d
Dry ‘07
3.73
a
1.07
a
1.47
a
1.47
a
1.07

a
2.93
a
9.33
b
OM4498
Wet ‘07
2.53
a
3.73
ab
3.87
ab
4.67
ab
8.93
b
10.40
c
8.13
ab
Wet ‘06
†

1.33
a
0.13
a
1.60
a

0.53
a
1.33
a
5.47
b
5.47
b
AG24
Dry ‘08
6.50
a
18.17
bc
16.44
bc
17.67
ab
21.47
bc
32.40
c
53.07
d
Wet ‘07
1.47
b
1.60
b
1.07

b
0.67
a
0.93
ab
0.4
a
1.33
b
IR50404
Dry ‘08
0.80
a
1.47
a
2.80
a
1.07
a
1.73
a
1.60
a
12.27
b
Jasmine Wet ‘06
†

4.00
a

3.90
a
5.18
ab
5.14
ab
6.00
ab
8.66
c
7.60
bc

All data represent mean values of five replications. The same superscripts in the same row indicates that the values
are not significantly different (P>0.05).
†
harvested in ‘late wet season’ which was in September 2006.

Increased rice cracking due to delayed harvesting also depended on the variety. There was a
large amount of cracked grains after maturity date for OM2517 (16.00-60.53%) and AG24
(21.47-53.07%) in dry season 2007 and dry season 2008, respectively. In contrast, percentages of
cracked grain of IR50404, OM2718, and OM4498 varieties had lower values in both wet and dry
seasons (in the range of 0.4-12.27%, 3.20-10.80% and 1.07-10.40%, respectively) after maturity
date. This implied that there is a varietal difference on rice cracking and hence the selection of
variety is important in decreasing cracked grain percentage.
The cracking behavior of the rice in the field is expected to depend on the season due to the
different patterns of temperature fluctuation during day and night, solar radiation intensity,
sunshine hours and frequency of rain. During the rainy season, the rice grain can develop cracks
during the late maturity stage due to rewetting. At the same time, during dry season it is likely
that the grains over-dry if not harvested by its maturity. However, data on Table 3 obtained from

four consecutive crop seasons (wet 2006, dry and wet 2007, and dry 2008) showed that crop
seasons did not have much impact on level of rice cracking as similar trend was observed in both
wet and dry seasons.

70

Head rice yield
The head rice yield as a function of harvesting time for seven rice varieties is presented in Table
4. The head rice yield was generally less at late harvesting time. A delay of 4-6 days reduced the
head rice recovery by up to 50% of the head rice yield at the expected maturity. The head rice
yield followed the opposite trend to rice grain cracking, indicating that the presence of cracks in
the original paddy reduced the head rice recovery.

The overall results as influenced by harvesting time are presented in Table 5. It should be noted
that the head rice yield is affected by a laboratory milling system, as it is a function of milling
efficiency. Therefore, the head rice yield data is presented in relative term in Table 5, with the
recovery on the harvesting at maturity (0 day) being assigned a value of 100%. In addition, due
to the limited number of experiments undertaken, the values are presented as a range for each
variety.

Table 4. Change in head rice yield of seven rice varieties at different harvesting time (days after expected
maturity date).
Head rice yield (%) before and after maturity day
Variety Crop
season
-6 -4 -2 0 +2 +4 +6
Wet ‘06
51.06
cd
52.30

d
50.73
cd
48.08
c
42.23
b
36.51
a
34.53
a
Dry ‘07
63.13
bc
66.21
c
66.93
c
67.90
c
64.57
bc
60.25
ab
56.35
a
OM1490
Wet ‘07
50.03
a

45.10
a
52.15
a
45.56
a
49.81
a
49.26
a
49.01
a
Wet ‘06
45.41
c
51.47
d
43.54
bc
43.91
bc
38.76
ab
36.83
a
40.72
abc
0M2718
Dry ‘07
67.93

b
67.01
b
66.40
b
67.48
b
66.22
b
63.81
a
62.41
a
Dry ‘07
64.58
d
41.09
b
45.19
b
56.68
c
53.18
c
43.74
b
28.63
a
Wet ‘07
48.01

c
44.16
bc
37.88
a
42.19
ab
44.47
bc
49.24
c
44.34
bc
OM2517
Dry ‘08
65.68
c
65.36
c
64.67
c
59.84
c
60.55
b
55.29
a
52.90
a
Dry ‘07

43.80
a
54.35
bc
54.02
bc
58.33
d
56.95
cd
53.78
bc
52.55
b
OM4498
Wet ’07
36.64
a
37.77
a
35.83
a
39.35
ab
37.87
ab
42.42
b
35.35
a

Wet ‘06
†

40.35
b
42.35
bc
40.76
b
43.50
bcd
46.99
d
35.90
a
35.35
a
AG24
Dry ‘08
61.66
c
55.42
bc
52.38
b
42.62
a
43.55
a
36.48

a
37.94
a
Wet ‘07
58.08
c
56.94
b
57.79
c
53.27
a
56.54
bc
55.67
abc
54.55
ab
IR50404 Dry ‘08
64.28
de
61.75
cd
64.57
e
60.28
c
57.40
b
56.99

b
51.68
a
Jasmine Wet ‘06
†

41.59
a
54.65
c
51.82
bc
55.36
c
54.59
bc
48.15
b
49.46
bc
All data represent mean values of five replications. The same superscripts in the same row indicates that the values
are not significantly different (P>0.05).
†
harvested in ‘late wet season’ which was in September 2006.

In general, the optimum harvesting time presented in Table 5 is similar to the maturity time
shown in Table 1 for all varieties used in this investigation. Suggested optimum harvesting times
in wet season for OM1490 (94 days) and OM2517 (94 days) are 2-4 days longer than
recommended maturity day by local extension centre. It can be concluded that (1) even if the rice
varieties were harvested about the right time, varieties differ considerably in the cracking and

71

hence intervention opportunity of growing low cracking varieties such as OM2718 for farmers
and developing such varieties for rice breeders, (2) harvesting at optimum harvest time had
rather small cracking problem but delay of 6 days can cause major problem, and hence
intervention opportunity here to ensure harvesting at the right time, and (3) varieties differ in
their response to time of harvesting hence time of harvesting is more critical for some varieties
than others, for example OM2517 was the most sensitive variety, and hence there is an
opportunity for intervention to ensure quick harvesting of particular varieties.

Table 5. Seasonal trend of effect of harvesting time before and after maturity (4-6 days prior and 4-6 days
later than the expected day of maturity) on the proportion of cracked grains (prior to milling) and head rice
recovery. Head rice yield is expressed as relative to the yield on maturity day.
Proportion of cracked grain % Relative head rice yield %
Crop
season
Rice
variety
Before maturity After maturity Before maturity After maturity
Optimum
harvesting
date
OM1490 0.8-9.6 1.1-23.6 101-109 72-88 94
Wet
OM2718 0.4-1.2 4.0-10.8 103-117 84-93 92

OM2517 3.5-15.7 12.1-20.3 90-114 105-117 94

OM4498 2.5-3.9 8.1-10.4 91-93 96-108 94

AG24 0.3-1.5 1.1-4.1 93-97 83-108 94
IR50404 1.1-1.5 0.4-1.3 103-105 99-106 90
Jasmine 4.0-4.5 6.0-7.7 75-99 87-99 98
OM1490 0.5-2.3 5.6-22.4 93-99 83-95 92
Dry
OM2718 0.7-6.3 3.2-8.5 98-101 92-98 92
OM2517 0.7-3.6 9.3-60.5 77-106 51-97 86

OM4498 1.1-3.7 1.1-9.3 75-93 90-98 91

AG24 6.5-16.4 21.5-53.1 133-145 86-102 88

IR50404 0.8-2.8 1.7-12.3 105-107 86-95 88


CONCLUSIONS
A few days early harvesting (before maturity) is better than late harvesting by 4 to 6 days
because late harvesting will make the grain more sensitive to cracking. Therefore, any delay or
longer harvesting time can cause more losses, as is often the case of harvesting by hand. The
degree of harvesting time effect is also dependent on the variety.

REFERENCES
Bal, S., & Oiha, T. P., 1975. Determination of biological maturity and effect of harvesting and
drying conditions on milling quality of paddy. Journal Agricultural Engineering
Resource, 20, 353-361.
Berrio, L. E., & Cuevas-Perez, F. E., 1989. Cultivar differences in milling yields under delayed
harvesting of rice. Crop Science, 24, 1510-1512.
72

Calderwood, D. L., Bollich, C. N., & Scott, J. E., 1980. Field drying of rough rice: Effect on

grain yield, milling quality energy saved. Agronomy Journal, 72, 644-653.
Chae, J. C., & Jun, D. K., 2002. Effect of harvesting date on yield and quality of rice. Korean J.
Crop Sci., 47(3), 254-258.
Champagne, E. T., Bett-Garbet, K. L., Thompson, J., Mutters, R., Grimm, C. C., & McClung, A.
M., 2005. Effects of Drain and Harvest Dates on Rice Sensory and Physicochemical
Properties. Cereal Chemistry, 82(4), 369-274.
Chau, N. N., & Kunze, O. R., 1982. Moisture content variation among harvested rice grains.
Transactions of the ASAE, 25(4), 1037-1040.
Kester, E. B., Lukens, H. C., Ferrel, R. E. M., A., & FIinfrock, D. C., 1963. Influences of
maturity on properties of western rice. Cereal Chemistry, 40, 323-326.
Kunze, O. R., & Prasad, S., 1978. Grain fissuring potentials in harvesting and drying of rice.
Transactions of the ASAE, 21(2), 361-366.
Ntanos, D., Philippou, N., & Hadjisavva-Zinoviadi, S., 1996. Effect of rice harvest on milling
yield and grain breakage. CIHEAM-Options Mediterraneennes, 15(1), 23-28.
Swamy, Y. M. I., & Bhattacharya, K. R., 1980. Breakage of rice during milling- Effect of kernel
defects and grain dimension. Journal of Food Process Engineering, 3, 29-42.











73

Ministry of Agriculture & Rural Development

Project Completion Report

MS14: PROJECT COMPLETION REPORT

026/05VIE

Investigation of rice kernel cracking and its control in the field
and during post-harvest processes in the Mekong Delta of
Vietnam

APPENDIX 2A

Study on the flat-bed dryer in the Mekong River Delta of
Viet nam





April 2010


74

Appendix 2A. Study on the flat-bed dryer in the Mekong River
Delta of Viet nam
ABSTRACT
The study, including experiments and survey on the flat-bed dryer, focused on the cracking of

paddy grains, and on comparing the air reversal mode. Results showed that, in both the 8-ton
production-scale dryer and the 20-kg laboratory dryer, the effect of air reversal was very
apparent in reducing the final moisture differential; however, its effect on the drying time or
the drying rate was not statistically significant. Mechanical drying, whether with or without
air reversal, was superior to sun drying in terms of reducing rice crack. However, compared
to shade control drying, drying (with or without air reversal) did decrease the head rice
recovery and increase the crack; the causing factor was not apparent, most suspected reason
was the drying rate. The decrease in head rice recovery was inconsistent, slightly lower or
higher in each specific pair of experiments with and without air reversal; this was not
expected in line with data on the final moisture differential. Testing of a 4-ton dryer at Long-
An equipped with the solar collector as supplementary heat source resulted with good grain
quality and confirmed the good economic potential. Major findings from the survey on the
current status on the use of flat-bed dryers in 7 Provinces were: The trend for increased
drying capacity, the role of local manufacturers and local extension workers, government
support with interest reduction for dryer loans, the drying during the dry-season harvest, and
especially the unbalance between drying costs and drying benefits.

INTRODUCTION
Flat-bed dryers have been with the rice agriculture of the Mekong Delta of Viet Nam for a
long time. From the first flat-bed dryers in the 1980’s to about 6500 units in 2007 is quite a
good progress. But not all is optimistic. Acceptance varies among provinces, even among
districts or communes in the same province. Finding the interrelated factors affecting the
dryer acceptance is quite complex. Within the context of the CARD Project 026/VIE-05 with
focus on the cracking of paddy grains in the area, the study on the flat-bed dryer from 2006 to
2008 included the following objectives:

• Conduct experiments under laboratory controlled drying conditions and under actual
production conditions to evaluate the effect of air reversal on the rice crack and other
drying outputs.
• Conduct experiments on the 4-ton flat-bed dryer, using solar energy as supplementary

heat source.
• Conduct a Participatory Rapid Rural Appraisal (PRRA) survey to update on the use of
flat-bed dryer in the Mekong Delta.
75


REVIEW OF RELEVANT INFORMATION
The following information is based on data by various Provinces presented during different
seminars, on an integrated assessment study by the Ministry of Agriculture-Rural
Development in collaboration with DANIDA in 2004, and on the first author’s working
experience with flat-bed dryers in the past 25 years.

Development of the flat-bed dryer
The Mekong Delta in Southern Viet Nam, with 2.7 million hectares of rice land, is producing
about 50 % of Viet Nam total rice output. With 16 million people or about less than 20 % of
the total population, this region has accounted for more than 90 % of Vietnamese rice export
in the past decade. Average farm size is about 1 ha per household, although in some newly-
reclaimed districts, 3 - 10 ha per household is not uncommon.
Rice drying became an issue in Mekong Delta in early 1980’s when a second crop was
promoted, of which the harvest fell into the rainy season. Different dryer models were tried
by various agencies; only one model was accepted by the production sector, namely the flat-
bed dryer (FBD). The first FBD was installed in Soc-Trang Province in 1982 by the
University of Agriculture and Forestry (now renamed Nong-Lam University NLU). Farmers
in Soc-Trang copied/ modified/ improved this FBD using cheap local materials. In 1990,
there were about 300 FBD units in the Mekong Delta, half of which were in Soc-Trang.
Other Provinces began to adopt these dryers. In 1997, a survey conducted by a Danida-
assisted Project reported a total of 1500 FBD in all Mekong Delta, with 3 leading Provinces
(Kien Giang, Soc-Trang, Can-Tho) accounted for 850 units; all remaining 10 Provinces
shared the balance of 650 units.
The above Danida-assisted Project in Can-Tho and Soc-Trang doubled the FBD in each

Province from about 250 to 500 units in the two-year span of 1998-1999, through a credit
scheme and extension activities. The Project terminated in 2001, and replaced by a Program
managed by the Ministry of Agriculture, but still assisted by Danida, then with only extension
activities. The Program terminated in mid-2007. The number of FBD dryer rose rapidly,
about 3000 units in 2002 and 6200 units in 2006. The dryers in the Mekong Delta account
for more than 95 % of all dryers in Viet Nam.
The technical development of the FBD in the past 25 years followed an interesting pattern.
First, a design was released by a research institution, NLU in this case. Next, farmers/
mechanics copied/ modified/ improved the design. Next, NLU monitored those modifications
and came up with a major design change and improvement. The cycle repeats.
The landmarks for these major design releases by NLU have been:
1982: Conventional FBD with central air inlet to the plenum chamber, using flat-
grate rice husk furnace with precipitation chamber (Fig.1).
1994: Conventional FBD with side-duct plenum (Fig.2), rice husk furnace with
vortex and central-pipe precipitation chamber (Fig.3).
2001: Reversible FBD (Fig.5 & 6).
(expected):
2006: Automatic rice husk furnace (model NLU-IRRI-Hohenheim, Fig.4)
76

2007: Solar collector for FBD
Major modifications /improvements by farmer-mechanics have been:
1987: Rice husk furnace with inclined grate.
2004: Drying bin for reversible dryer, with distributed central inlet.
2006: Raking mechanism under the rice husk hopper for more uniform husk feeding.

Figure 1. Conventional FBD with central air inlet to
the plenum chamber.

Figure 2. Conventional FBD with side-duct

plenum.

Figure 3. Rice husk furnace with vortex and
central-pipe precipitation chamber.

Figure 4. The automatic rice husk furnace for
SRA-4 reversible flat-bed dryer.

Drying Air
UP
Grain
CONVENTIONAL SHG
FLAT-BED DRYER
Floor: 50 sq.m / 8 ton
0.3m

Drying Air
UP
Drying Air
DOWN
Grain Grain
REVERSIBLE SRA DRYER
0.6m
Floor: 25 sq.m / 8 ton

Figure 5. Principle of reversible-air dryer.

Figure 6. The SRA-10 reversible air dryer (10 tons per batch).
77

NLU have taken the leading role in releasing efficient dryer fans, both for conventional and
reversible dryers, with transfer of design and fabrication technology to 15 manufacturers in
the Mekong Delta, among them 7 have built fan test ducts according to JIS Standards.

Quality of paddy dried by the flat-bed dryer
The quality of dried paddy is judged by several criteria:
 The paddy is not contaminated with black ashes from the furnace.
 The paddy final moisture content is uniform at the desired level for storage.
 For seed grain, the germination is high.
 For commercial grain, the dried grain crack is minimized.

The first criterion (no ash mixed with grain) has been met after some years, due to gaining of
experience in furnace building, and competition among furnace builders, to deal with
farmers’ first-visual reactions.

The second criterion is difficult to meet due to the inherent principle of flat-bed drying. A
final moisture differential of 1.5 % (between the top and the bottom layer) is considered
good, while in continuous-flow mixing-type dryer, 1.0 % is normal. For flat-bed dryers,
farmers rely on manual mixing. Technically, a good high airflow rate at moderate
temperature (below 44
o
C) helps reducing the differential. The air-reversal principle
introduced since 2002 also reduces the non-uniformity. All these technical features should be
re-evaluated / confirmed in this current CARD Project.

The preserved seed germination is well established by Seed Companies in using a safe drying
temperature below 42
o
C, and most importantly to dry the grain within 12 hours after harvest.
For commercial grain, the dried grain crack is a big issue. A report (Phan Hieu Hien, 1998)

based on surveys of a few rice mills in Can-Tho and Long-An showed a reduction of 5 to 7 %
of farmers’ profit due to more broken rice due to improper drying. This high loss was due to
the habit of field drying in the dry-season harvest, and estimated to be about 20 million US$
per harvest in the Mekong Delta. However, data and estimates were based on a few
interviews, and not on systematic testing. Thus, in this CARD Project, the need is to confirm
or reject based on solid test data.

MATERIALS AND METHODS
Testing
Testing of dryers followed standard procedures described in RNAM (1991) and ASABE
(2006). Measurement equipment included different thermometers, moisture meter and drying
oven, power meter etc.
For the 8-ton dryers, the drying temperature was at 2 levels: a) Constant at 43
o
C; and b) At
50
o
C for the first hour, and afterwards constant at 43
o
C. In reality, due to the furnace
configuration, the temperature rarely exceeded 50
o
C, and was about 48
o
C at most. In all
tests, the focus was to compare two drying modes: WITH air reversal, and WITHOUT air
78

reversal. Some experiments also compared with sun drying on the cement drying yard with a
7-cm paddy layer, as popularly practiced by local farmers.

The crack analysis and head rice analysis was first done at the Vinacontrol, an accredited
agency in charge of certifying rice quality for export, and later at the Rice Quality Laboratory
of the NLU Chemical Technology Department, following procedures adopted by
International Rice Research Institute and the University of Queensland. Each treatment was
analyzed by 3 samples, each consists of 50 grains taken at random; each paddy grain was
hand-husked and examined under the magnifying glass for fissure. The increase in crack or
decrease in head rice of each treatment based were on the control shade drying (or further
shade drying to 14 %MC).

The biggest problem for testing has been the input paddy. We encountered severe difficulties
in securing batches of the same quantity or initial moisture content. This was apparent with
the 8-ton batches. But even scale-down to 1-ton batch, the 3-factor experiments could not be
run, due to different initial MC. Finally, from the “lumpsum” conclusions with 8-ton dryers,
we had to concentrate on and be contented with 20-kg batches in paired experiments (block)
of Air reversal and No air reversal.

For experiments on the use of solar heat for paddy drying, a 4-ton popular flat-bed models
fabricated by a local mechanical shop was selected, and added with a solar collector designed
at the NLU Center for Agricultural Energy and Machinery.

Survey
The objectives of the surveys were: (i) to update the role of flat-bed dryers in reducing post-
harvest losses and in preserving rice quality; (ii) to identify operating factors of the flat-bed
dryer which contribute to the reduction of rice crack; and (iii) to identify problems with the
flat-bed dryer that the CARD Project could possibly help.

The survey used the Participatory Rapid Rural Appraisal (PRRA) method, through
interviewing different people class, from farmers to rice millers to governmental officials etc.
But it also relied heavily on both available data gathered in the past 10 years by various

agencies, and on personal experience of the people involved with the dryer at NLU over the
past 20 years.

Four Provinces were selected in 2006, namely Can-Tho City, Kien-Giang, Long-An, and
Tien-Giang. The first three Provinces have sites which had been selected by the CARD
Project for all related experiments, demonstrations, and extension activities. The fourth
Province is adjacent to Long-An, and also planned as site for rice milling survey, so facts and
data on the dryer would be relevant. In 2007, we visited more Provinces such as Hau-Giang,
An-Giang, Kien Giang, Soc-Trang…, with resulted with additional findings.
79

RESULTS AND DISCUSSSION
TESTING
Experimental results on the 8-ton dryer, the laboratory dryer, the solar-assisted dryer, as well
as the survey results are presented in the following sections.
The 8-ton dryer
Two 8-ton dryers were selected for experiments. One was a NLU-designed air-reversible
dryer installed at Tan-Phat-A Cooperative, Tan-Hiep District, Kien Giang Province in July
2006 (Figs. 7&8). The other was an air-reversible dryer made by a local manufacturer
installed at Tan-Thoi Cooperative in Can-Tho Province, with the design patterned on the
SRA-8 of NLU; the difference was the under-plenum duct inside the drying bin, “ong gio
chim” in Vietnamese (Fig.9), in order to distribute the airflow evenly.

Figure 7. The 8-ton dryer at Tan-Phat-A
Cooperative, Kien Giang

Figure 8. The 8-ton dryer with the air for
downward direction .

Figure 9: The 8-ton flat-bed dryer at Tan-Thoi Cooperative, Can-Tho Province

Experiments from Kien-Giang were under more control thus more results are reported here,
while results at Can-Tho are supplementary. Refer to Phan Hieu Hien (2006, 2007, 2008) for
testing details.
In Kien-Giang experiments were conducted in two wet seasons (July 2006, and July- August
2007), and two dry-seasons (March 2007, and March 2008). Major findings are as follow:

• The drying temperature is stable and can be kept within ± 3
o
C, usually from the nominal
value of 43
o
C.
80

• The effect of air reversal was very apparent in reducing the final moisture differential.
When operated correctly, this differential was less than 2.2 % with air reversal, but over
4.6% without air reversal. More MC differential means more rice cracking during
milling. This explains why dryers installed since 2003 have been more and more of the
reversible principle.
• However the effect of air reversal on the drying time or the drying rate was not clear
because of several other factors involved (Fig.10).

Data on the crack of rice upon milling in March 2007, and July 2007 with three pairs of
drying batches (With Air reversal, and Without air reversal) showed that:
• Mechanical drying, whether with or without air reversal, was superior to sun drying in
terms of less crack percentage or more head rice recovery. About 3- 4 % less cracking,
and about 4 % more head rice recovery were main data obtained from .March 2007
experiments.
• The grain cracking in Air-reversal batches were lower than No-air-reversal batches
(Fig.11). This is a basic result.
• However, the decrease in head rice recovery was inconsistent, slightly lower or higher in
each specific pair (Fig.12). This is confirmed by the statistical comparison on the head
rice recovery with t-test between batches of Air reversal and No air reversal, which did
not show significant difference at 5% level. This was not expected in line with the above
data on Final MC differential. The reason was probably due to the sample milling; the
whitening time was only 1 minute, thus some slightly cracked kernels might not be
broken during milling.
• In both cases (Air reversal and No air reversal) drying did decrease the head rice recovery
and increase the crack. The causing factor was not apparent, due to so many factors
involved in a large mass of 8 tons of grain: paddy non-uniformity, drying rate…. Most
suspected reason was the drying rate (Fig.13), data pointed to an optimum drying rate in
the 1.0– 1.2 %/hr range, but this has to be confirmed by further elaborate experiments, or
from laboratory scale experiments.
Dr ying r ate
0.0
0.5
1.0
1.5
2.0

2.5
3.0
20 22 24 26 28 30
Initial MC, %wb
Air Reversal
No air reversal

Fig.10: Effect of air reversal on the drying rate.
81

Crack % INCREASE (Kien Giang 2007 wet-season)
0
5
10
15
20
25
30
35
40
B2 & B5 B1 & B6 B9 & B6 Ave(3batches)
Batches
Crack %
Air reversal No air reversal

Figure 11. Crack% INCREASE, Kien-Giang, wet-
season 2007.
Head rice, Kien Giang 2007 Wet-season
(AR = Air Reversal; NAR = No air reversal. B2 = Batch No2)
0

10
20
30
40
50
60
70
AR B2
AR B9
NAR B5
NAR B6
StDev(AR
)
Head Rice Before drying, % Head Rice After drying, %
Figure 12. Head rice Before and After drying.
Effect of Drying rate (AR & NAR)
0
4
8
12
16
20
24
28
32
36
0.8 1.0 1.2 1.4 1.6 1.8 2.0
Drying rate, % /hr
Crack Increase,
Head rice Decrease, %

Grain Crack Increase, % Head Rice Decrease , %

Figure 13. The effect of the drying rate on the crack increase or the head rice recovery.


The laboratory dryer
The effects of two factors were studied: Factor A was the final MC with two levels (14%
coded X14, and 17% coded X17). Factor B was the air reversal mode with two levels (Air
Reversal AR, and No Air reversal NoAr). Each four treatments (or factor combinations)
were in one block of experiment that is, conducted at the same time. This was possible
thanks to 2 identical laboratory dryers running in parallel. Each batch contains 20 kg of
paddy. Four replications (or 4 blocks) were made.
The total thickness of the paddy layer in the AR batches was 0.51 m while that of NoAr
batches was 0.31 m. Paddy was sampled at three layers –Bottom, Middle, and Top layer— in
3 specific trays, with other buffer trays in-between.
In each block of experiment, the dependent variables were: the drying rate (shown by the
drying curve), the uniformity of the final MC (shown by the MC of the bottom, middle, and
top layers), the head rice recovery, and the grain crack. Data in one typical block are graph in
Fig.14, 15, 16&17. Results are statistically analyzed as a RCBD (randomized complete block
design) with data compiled in Table 1.

From the results and the statistical analysis, the following remarks can be derived:
a. The final MC differential:
82

The effect of both the reversal mode and the final MC was statistically significant at 5%
alpha level. Air reversal yielded less final MC differential than No air reversal (Table 1,
Fig.14). Also, drying stop at 14% MC gave less final MC differential than at 17% MC.
However, since the interaction effect is significant, comparisons should be made from each
treatment, that is between factor combinations. For example in Table 1, treatment NoArX14

and AR_X17 had similar MC differential.
20-8-2008I: 43
o
C
10
11
12
13
14
15
16
17
18
19
20
21
22
23
AR X14 NoArX14 AR X17 NoArX17
AR = Air Reversal; NoAr = No Air reversal.
X14 = Average Final MC 14%
X17 = Average Final MC 17%
Final MC, %w
b
Upper
Middle
Lower

Figure 14. Moisture non-uniformity.

20-8-2008I: 43
o
C. Final MC 17%
10
12
14
16
18
20
22
24
26
28
30
024681012
Drying time, hr
MC , % wb
NoArX17-Bottom
NoArX17-Middle
NoArX17-Top
AR X17-Bottom
AR X17-Middle
AR X17-Top Layer

Figure 15. Drying curves down to 17% MC of the Top, Middle, and Bottom layers.
AR = Air Reversal; NoAr = No Air reversal.
X14 = Average Final MC 14%. X17 = Average Final MC 17%
83

20-8-2008I: 43

o
C. Final MC 14%
10
12
14
16
18
20
22
24
26
28
30
024681012
Drying time, hr
MC , % wb
NoArX14-Bottom
NoArX14-Middle
NoArX14-Top
AR X14-Bottom
AR X14-Middle
AR X14-Top Layer

Figure 16. Drying curves down to 14% MC of the Top, Middle, and Bottom layers.

b. Drying rate:
The effect of both the reversal mode and the final MC was not statistically significant at 5%
alpha level. However, at 10% alpha level, drying down to 14% MC was significantly at
slower rate than down to 17% MC (Table 1, Fig.15 &16).

-6.38
-8.56
-12.92
22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
NoArX14 AR X14 NoArX17
A
Decrease in Head rice
r
y, compared to shade drying,
% (decrease = - )
AR = Air Reversal; NoAr = No Air reversal.
X14 = Average Final MC 14%
X17 = Average Final MC 17%

Figure 17. % Decrease in head rice recovery.

Table 1. Data from experiments with the laboratory dryers. AR = Air Reversal; NoAr = No Air reversal;
X14 = Final MC 14%; X17 = Final MC 17%.
84

DRYING RATE,
%/hr
FINAL MC
DIFFERENTIAL, %
% DECREASE IN
HEAD RICE
RECOVERY
GRAIN CRACK, %
A =
Final
MC

B =
Reversal mode
A =
Final
MC
B =
Reversal mode
A =
Final
MC
B =
Reversal mode
A =
Final
MC
B =Reversal
mode

NoAr AR

NoAr AR

NoAr AR

NoAr AR Ctrl
X14
1.00 0.88
X14
4.20 2.00
X14
-2.99 -5.96
X14
8.00 11.33 11.33
1.37 1.54 6.30 4.00 -10.02 -9.26
0.00 7.33 4.00
1.46 1.35 7.00 4.90 -7.50 -9.32
0.00 2.00 0.67
1.59 1.53 7.89 3.90 -5.02 -9.70
0.67 8.00 1.33
Av: 1.35 1.32 Av: 6.35 3.70 Av: -6.38 -8.56 Av:
2.17 7.17 1.83
X17
1.11 1.05
X17
5.30 3.70
X17

-14.18 -5.70
X17
18.67 24.00 3.00
1.47 1.49 11.70 9.60 -10.27 -10.18
19.33 8.67 0.67
1.60 1.36 8.90 4.80 -21.83 -25.70
4.00 9.00 0.00
1.59 1.71 10.30 6.60 -5.77 -8.41
2.00 1.33 0.00
Av: 1.55 1.52 Av: 10.30 7.00 Av: -13.01 -12.50 Av:
11.00 10.75 0.92
Statistical Analysis Results (at 5% significance level):
Interaction AB:
No
Interaction AB: Yes Interaction AB: Yes Interaction AB: Yes
A: non-significant LSD = 2.14% LSD = 7.70% LSD = 7.73%
B: non-significant (between treatments)


c. Decrease in head rice recovery and Rice crack
Drying, whether with air reversal or not, did decrease the head rice recovery and increase the
rice crack compared to control shade drying. The reason was suspected as too high drying
rate (over 1.3 %/hr). But the regression analysis and graphing (Fig.18) showed no definite
trend. More crack and lower head rice recovery at 17% as expected because paddy was
milled with the laboratory mill at that MC which was not the optimum.

85

Effect of Drying rate on rice crack
0

5
10
15
20
25
0.8 1.0 1.2 1.4 1.6
Drying rate , %wb /hr
Rice crack, %
NoAr
AR

Figure 18. Effect of the Drying rate on Rice crack.

In theory, more head rice recovery corresponds to less rice cracks; and more head rice and
less crack correspond to more MC differential. The less final differential with Air Reversal
compared to No Air reversal was expected to go with less decrease in head rice and less
crack, but data showed the contrary (Table 1).
In summary even with the laboratory experiments, the experimental data were still hard to
analyze for balanced and good-looking results, with the possible reason traced to the
variability between the individual grains.

The 4-ton solar-assisted dryer
A 4-ton popular flat-bed models (named SDG-4 dryer) fabricated by a local mechanical shop
was selected. This is a collapsible unit, which consists of the following components:
i.) A two-stage axial fan, a design transferred by NLU, powered by a 15 HP Chinese diesel
engine.
ii.) A coal furnace, with coal consumption adjustable within 5 to 12 kg/hr.
iii.) A drying bin, with the grain floor size 4.50 m *3.27 m made from bamboo slat and nylon
net. The bin is supported on 7 metal legs, thus can be easily installed on rough land. The
airflow can be upwards (Fig.19), or downward (Fig.20) with a covering tarpaulin.

iv.) A solar collector (designed at NLU) consisting of 2 cylindrical plastic collector (Fig.19
&20). Each cylinder is φ1.0 m * 27 m long. Inside the transparent plastic layer is the
black PE layer for absorbing heat. The two cylinders converged into a transition box,
which also received heat from the coal furnace. The collector used cheap materials such
as bamboo slats and plastic wires, and was installed on the open ground instead on the
rooftop, thus the investment cost was significantly reduced compared to the steel-frame
collector of the macaroni dryer (Phan Hieu Hien et.al, 2007).
The solar collector and the coal furnace can be used separately or in combination. Tests
were done at Long-An Province in March 2007, the driest month of the year.
86


Figure 19. The SRA-4B dryer with the upward
airflow.

Figure 20. The SRA-4B dryer with the downward
airflow, using solar heat.

Five drying batches were tested in March 2007: Batch 1 with heat from coal only; Batches 2
& 3 with heat from solar energy only, Batches 4 & 5 with heat combined from both coal and
solar energy.

Results are summarized as follow:
• The capacity was 3.8 – 4.1 ton per batch of 7- 12 hours, with moisture reduction (average
± standard deviation) from 23.8 ±1.7 % MC down to 14.2 ± 0.8 % .
• The drying temperature could be adjusted within 38- 44
o
C using coal. With solar heat,
the drying temperature could reach 38
o

C with good sunshine (over 800 W/m
2
radiation),
or only 36
o
C in cloudy weather (about 500 W/m
2
radiation, which is also typical in the
wet-season). With less sunshine, the 12-hr drying time as in Batch 3 was expected.
• The combination of solar and coal heat is handy in ensuring to finish one batch within the
day. The harvest season in one village or commune usually lasts less than 25 days, thus
can not allow the “luxury” of 2-day drying batch.
• The head rice recovery in all batches were comparable to “shade” drying; or even slightly
better with 2 batches with solar energy, possibly to slightly lower drying temperature.

The contribution of solar energy is analyzed using data of Batch 4 and Batch 5, both with
combined heat from coal and solar energy; the following can be drawn:
 Solar energy could contribute to a cost saving of 43– 78 % from reducing the coal
consumption.
 The saving translated into US$ 3– 5 per batch, or US$ 0.7– 1.3 per ton)
# # #

 For estimation, assume that in one year, the dryer is used for 100 batches or 400 tons, of
which ½ totally use solar energy, and ½ use supplementary solar energy with 50 %
saving, or US$1.6 and US$0.8 per ton respectively. Thus the total saving would be
US$480 per year.
 Compared to the additional investment for the solar collector of about US$ 560, with the
replacement of the plastic sheet costing about US$120 after every 7 months, the payback
period is about 2 years.

# # #
Converted from 16 300 VN dong
≈
US$1 (In 2007).
87

 For stand-alone dryer owner, he/she might not be able to dry for 100 batches per year. In
contrast, dryer owner-cum-rice miller can surpass that quantity easily. Thus the solar
collector would aim practically more for the rice milling compound.

In the Mekong Delta of Viet Nam, farmers currently use the flat-bed dryer mostly for paddy
harvested in the wet-season. For the dry-season harvest, people mainly rely on the pavement
natural sun drying to save the cost of fuel for drying. Thus paddy crack in the dry-season
harvest is even more severe, as repeatedly warned by research and extension agencies without
many results, obviously due to the very low drying cost under sunshine.
Thus, from test data, solar energy has been used to dry paddy at a production scale; early
attempts in the 1980’s dealt with 50- 300 kg/ batch lasting 2 days The test has proved the
quality of the dried paddy. Economically, it could refute the popular saying that “Solar
energy is free but not cheap” with the fact that the owner can recover the additional
investment for the solar collector in about 2 years. Environmentally, solar energy is clean.
The problem remains to introduce the solar heat to the integrated rice mill with dryers. The
test has proved the function of the solar collector in saving fuel cost, especially in the dry-
season harvest.

SURVEY
Background data
The 4 Provinces under study have similar data in terms of climate and other agricultural
featuresm typical of the Mekong Delta. All have the average monthly temperature of
27- 28

o
C, with the average maximum of 29
o
C in April and minimum of 25
o
C in January.
But the temperature difference between daytime and night time is more pronounced, say
between 25 and 36
o
C in hot months, or 23 and 33
o
C in cooler months.

The rainy season in the region occurs from May to October, the remaining months are dry
season and there is not four seasons such as the Spring, Summer, Winter like in Northern
Provinces. The annual rainfall is 1 400 mm in Long-An, and higher in Can-Tho and Kien-
Giang, 1600 and 1800 mm respectively.

The average annual relative humidity is 80- 82 %. This just says that is typical tropical
humid climate, and not specific enough about its significance in post-harvest. Earlier
compilarion (Phan Hieu Hien, 1998) showed that in a typical day of March (dry season) and
of August (rainy season) in of the Mekong Delta, the relative humidity during the night time
(21h00 PM to 7h00 AM) is very high, over 90%. This is totally different with Australia,
where the Rh is below 70 % even in night time. The implication is the moisture re-absorption
of the grain during storage.

Specific data pertaining to each Province are shown in Table 2.