Drying Studies of Single Layer Thompson Seedless Grapes

International Solar Food Processing Conference 2009
Drying Studies of Single Layer Thompson Seedless Grapes.

R.L.Sawhney, D.R. Pangavhane* and P.N. Sarsavadia
School of Energy and Environmental Studies
Devi Ahilya University, Takshashila Campus, Khandwa Road
Indore -452001, (M.P.), INDIA

* Department of Mechanical Engineering

K.K. Wagh College of Engineering, Amrutdham, Panchvati,
NASHIK – 422003, (M.S.), INDIA

Key Words and Phrases: Chemical pretreatment; drying rate constants; empirical models;
nonlinear regression analysis; process variables; raisins.

Abstract

For determination of drying kinetics of Thompson seedless grapes a suitable experimental
unit for online measurement was designed and fabricated. The drying characteristics of oil
emulsion pretreated grapes were measured using ambient air under controlled air
temperatures (50 to 80ºC) and velocity (0.25 to 1.00 m/s) conditions. Out of the three models
considered (Page’s, Single term and Two term exponential) Page’s model was found to be
the most suitable for describing the drying behaviour of the grapes. The dependence of drying
constant K of the Page’s model on process variables (Temperature and Velocity of Air) was
analyzed using Arrhenius and Power Equations. It was found that the Arrhenius Equation
gives better values of K than the Power Equation. It is also found that the dependence of
another drying constant N of the Page’s equation on the process variables can not be
described in terms of Arrhenius or power Equations.

Introduction

The dried fruits have always been an important contributor source to the agricultural
economy. Raisins are one of the most important dried products obtained by drying of grapes.
The raisins are directly used as ingredients in the confectionery and in the form of raisin paste
applied in fillings, baked goods, sauces, microwavable coating and also for natural colouring
of other food products (Veronique and David, 1993). Thompson seedless and other varieties
like sultana, muskat and black coraith account for most of the world raisin production (Winkler
et. at., 1974). The raisins are generally produced either by traditional means (i.e. open sun or
shade drying) or in mechanical dryer.

For drying studies, Grapes is considered to be rather complex system with an outer waxy
cuticle and pulpous material inside. During drying of the grapes the waxy cuticle is main
obstacle which restricts and controls the moisture diffusion in the grapes. (Grnearevic and
Radler, 1971). Also the shrinkage of material during drying causes an increase in thickness of
the waxy cuticle which reduce the permeability of water through it. Chemical pretreatment (hot
or cold) is applied to the grapes to decrease the skin resistance for improving moisture
diffusion through waxy cuticle (Dudman and Grnearevic, 1963 ; Grncarevic et.al., 1968;
Ponting and Mebean, 1970). The hot dip pretreatment dries grapes more quickly while cold
dip pretreatment gives raisins an attractive golden brown colour without producing checks or
cracks on the grape berries (Grncarevic, 1963; Radler, 1964).





1

---

International Solar Food Processing Conference 2009
The type of chemical pretreatment and origin of the product significantly effect the drying
behavior of the grapes. For their delicate nature and in order to obtain the quality raisins
accurate prediction of the drying rates (as dependent on process variables, pretreatment and
origin) is very much required. For design optimization of any dryer, knowledge of the drying
constants and their dependence on drying air parameters is necessary input. A detailed study
on determination of drying parameters for grapes in all usable range of temperature and
velocity of the supplied ambient air has been made and the results are presented in this
paper. The drying constants K and N of the Page’s equation were determined from the
experiments conducted and dependence of these parameters on process variables
(temperature and velocity of supplied ambient air) is obtained in terms of the Arrhenius and
Power model.
Mathematical Models

In single layer drying of agricultural produce numerous models have been proposed to
calculate the rate of moisture loss with time. If we treat the moisture removal phenomenon
similar to the convective heat loss from hot bodies (Newton’s law of cooling) drying rate
should be proportional to the difference in moisture content between the material to be dried
and the equilibrium moisture content (EMC) at the drying air state (Hukill and Schmidt, 1960),
Mathematically it can be written as,

dM = - k (M-M
dt
e)



…………………………..(1)

Where M is the moisture content (kg/kg. on dry basis) at any time t; Me is material equilibrium
moisture content (kg/kg); k is drying constant (hrs-1) and t is time, hrs. The solution of the
above equation (1) yields one term exponential equation which is generally used to fit the
drying curves of various agricultural produce (Henderson and Pabis, 1961).

M ? M e =MR = A
M ? M
o exp (-ko t)

…………………………(2)
0
e

Where, MR is moisture ratio, Mo is moisture content kg/kg at time t = 0,

Sharaf-Eldeen et. al. , 1979; Noomhorn and Verma, 1986 found that the accuracy of
prediction of drying kinetic of material can be improved by adopting two term exponential
equation having the following from:

MR = A1 exp (-k1 t) + A2 exp (-k2 t)

…………………………(3)

For the composite nature of some agricultural produce above exponential models were found
to be inadequate for predicting loss of moisture from composite materials. For such materials
Page proposed a modified impirical relation (Page Equation) which gives better results (Misra
and Brooker, 1980; Dimante and Munro, 1991). The Page’s equation is given by

M ? M e
MR =
= exp (-K tN)
M ? M
o
e
…………………………………(4)

Where N is constant, t is time, hrs.

The value of equilibrium moisture content (Me). of raisins for different air temperature at
corresponding humidity conditions applicable to present study can be computed using well
known GAB equation (Singh and Singh, 1996). The dependence of Page’s drying constant K

2

---

International Solar Food Processing Conference 2009
and N on the experimental variables (temperature and velocity of the supplied ambient air) is
obtained through the empirical Arrhenius model and Power model, in the following form

? ? ?
? 2
Arrhenius type model : K or N = ?0 V?1 exp ??
?
?
?
…………………………(5)
Tab ?


Power model: K or N = ?o V?1 T?2 ………………………………………………(6)

Where V is the velocity of drying air (m/s); T is the temperature of air (ºC); Tab is absolute air
temperature (K).

Data Analysis

The nonlinear least square regression method was used in the present study to evaluate the
drying constants K and N, with the help of Scientific and Technical Graphics package
(Microcal Origin; 1991). The coefficient of determination (R2) and Chi2(X2) between the stated
model calculations and experimental observations were used to evaluate the goodness of fit.
The lower the value of chisquare, the better the model was taken to fit. The chi square is
defined as

N
?(MR ?MR )2
expi
cali
X2 = i 1
=
………………….…………………(7)
N ? n

Where MRexpi is the experimental moisture ratio of observations i; MRcali is the calculated
moisture ratio at that observation; N is the number of observations and n is the number of
constants, i.e. (N-n) is the degree of freedom.

Experimental Procedure

Thin layer experimental dryer

Successful collection of thin layer drying data depends on accurate measurement of moisture
removed from the sample through out the drying process. The dryer design was so developed
that the instantaneous weight of the sample at different times of drying process could be
measured without disturbing the sample position in the dryer. The developed experimental
dryer (Fig.1) consists of air supply and flow control section; Air heating and temperature
control section; Drying test chamber and Measurement units. The air required for drying was
taken from the ambient by a centrifugal blower. The required air flow rate was maintained
using manually operated flow control valve (V1). The air flow rate was measured using a flat
plate orifice and u-tube manometer provided in the line. The ambient air is heated to the
desired temperature in the heating chamber having four resistance electrical heaters of
capacity 1 KW. The air temperature in the drying chamber was regulated at the required state
using a dimmerstat. The entire flow section was well insulated to reduce heat losses.
Temperatures at two places in the supply line were measured using mercury bulb
thermometers (T1 and T2) and temperature of the air in the drying test chamber was
measured with the help of a Pt-100 digital thermometer.


3

---

International Solar Food Processing Conference 2009


Figure: 1 Experimental Laboratory Dryer Set-up

The drying test chamber is a vertical chamber made of 8 mm thick waterproof plywood of 360
mm x 360 mm cross section and 880 mm height. The drying test chamber was divided in two
parts, a plenum chamber in upper part and the sample test chamber in the lower part. Before
air reaches to the sample, air passes through two flow straightners between which a wire
mesh resistance is provided to ensure even pressure distribution. The straightners avoid the
twisting or turbulent air flow over the sample. The entire drying test chamber was innerlined
using Aluminium foil to provide thermally uniform condition in the drying chamber. A door with
glass window was provided on the front side of drying test section for loading and unloading
of sample tray. In the test chamber a wooden platform fitted at surface level of weighing pan
with a oil channel (width 35 mm and depth 22 mm having size 180 mm x 180 mm) filled with
light oil was fixed. Below the test platform the extended vertical chamber length helped in
straghtning the air flow and also isolating the sample tray from the ambient conditions. The
two hanger rods were suspended from a beam of a balance (capacity 200 gms. And 0.0001
resolution) placed above the test chamber. At the other end of rods a weighing pan is
attached on which the sample tray (100 mm x 100 mm) was placed. The frame of the sample
tray was made from Aluminum sheet and a wire mesh (5 mm x 5 mm) on which sample was
placed. The bottom edges of the weighing pan was dipped in the oil. By this arrangement all
the air was forced through the sample tray with oil providing an air seal.

Sample Preparation and Pretreatment

Fresh ripe hand harvested Indian Thompson seedless grapes from Nashik region were used
for the drying test. The procured grapes were stored at temperature 4ºC (± 1ºC). Before the
start of an experiment the grapes were kept open in the room for about two hours to bring
them at room temperature.

For ensuring the uniformity of the physical characteristics of the grapes dried. The berries for
each experiment were selected from the same bunch. The weight of the sample size
(consisting of 36±1 berries having an average berry diameter 18 mm ± 1 mm) was kept at 77
± 2.0 gms. After picking berries from the bunch, the characteristics such as size, sugar

4

---

International Solar Food Processing Conference 2009
content, moisture content and acidity of the grapes were determined. The initial moisture
content of each sample was obtained by vacuum oven method (AOAC, 1975). The selected
sample was cleaned with lap water to make sample free from dust and foreign materials. To
increase the water permeability through the waxy cuticle, the grapes were dipped for 3
minutes in a dipping solution consisting of a mixture of 2% proprietary dipping oil and 2.5%
K2CO3 in water having pH value of about 10-11 (Grncarevic and Radler 1971; Winkler, 1974),
After completion of the pretreatment the sample was immediately placed on the sample tray
for drying test.

Experimental Measurements

The drying time for bringing the moisture content of the raisin to 0.17 to 0.18 kg/kg. dry basis
(this is considered to be the safe moisture content for longer storage) was measured with the
help of experimental unit. In grapes drying, the drying time is dependent on many factors such
as grape variety; soluble sugars; chemical pretreatment and drying conditions. Except drying
conditions all factors are assumed to be constant in this study. Air temperature (T) and air
velocity (V) are taken to be the independent variables of the drying time (t). in open shade
drying the maximum temperature to which grapes are subjected in the tropical climate
reaches up to 45ºC. Hence 50ºC was selected as the lower limit of the temperature to which
air was heated in the experiment. From the earlier studies (Possingham, 1974; George Lof,
1962), it has been concluded that the open shade drying at lower temperature is not an
efficient method for the longer drying time. To dry the grapes temperature is not an efficient
method for the longer drying time. To dry the grapes efficiently in mechanical dryer the air is
generally heated to higher temperatures. For drying of sultana grapes the maximum
permissible air temperature is considered to be 77ºC (Van Arsdel and Copley, 1964; Pointing
and Mcbean, 1970). Hence the maximum temperature to which air was heated in the
experiment was taken to be 80º C. In addition to these two limits, experiments were also
carried out at two intermediate temperature levels of 60 and 70ºC.

It has been observed (An Arsdel and Copley, 1964) that for air velocities higher than 4 m/sec,
the effect of air velocity on drying rate is negligible. Taking in to account the earlier studies
(Eissen et. at., 1985; Tsamparlis, 1990) 1.00 m/s and 0.25 m/s were selected as upper and
lower limits of the air velocity for our experiments. 0.50 m/s and 0.75 m/s was also used as
intermediate air velocity levels. To obtain the steady state conditions in the drying test
chamber at the desired level of air temperature and velocity, the hot air was passed through
the test chamber for at least two hours before placing the grapes on the sample tray. The
pretreated sample was arranged in single layer uniformly on the sample tray which was then
placed on the pan in the drying test chamber, after it had acquired the steady state condition.
The loss in moisture of the sample was determined by taking the weight of the sample plus
tray when air was flowing trough the sample and also when the air flow through the test
chamber was stopped for a brief period of about 10 seconds. For this the hot air from the
heater box was released in to the atmosphere through the bypass arrangement using control
valves (V2 and V3) provided for the purpose.

The observations were taken for every half hour till the moisture content of the sample
reached about 17-18 % (dry basis). Air temperature at the location indicated in Fig. 1, just
above the sample tray was measured with the help of RTD sensor based digital temperature
indicator of 0.1ºC resolution. In total 16 experiments performed at four velocities and four air
temperatures ranges, the drying time for obtaining the raising was ranging in between 8 to 52
hrs.




Results and Discussion

5

---

International Solar Food Processing Conference 2009

The experimental drying conditions and elapsed drying time for each run is given in Table:1.
The effect of drying process variables (temperature and velocity) on drying time is shown in
Fig. 2 and Fig.3. It is clear from Fig.2 that the drying air temperature increases the
dehydration rate of the Thompson seedless grapes, owing to the increase in vapour pressure
of water and the permeability of the waxy cuticle. The effect of air velocity (keeping the
temperature of air constant) is shown in Fig.: 3 it is observed that at a given temperature
drying time decrease with increase in velocity of the air. The variation of the drying rate with
moisture content at different temperatures and at given velocity (0.5 m/s) is shown in Fig.4.
From Fig. 2-4, it is seen that the constant rate drying period generally observed in the initial
drying stages of some agricultural products is absent in the case of pretreated grapes for the
temperature range considered. From the initial stage itself the drying rate decreases
continuously with the moisture content/time. This agrees with earlier studies of Alvarez and
Legus, 1986; Riva and Peri , 1986; Tulasidas et. at, 1993. Initially the drying rate is higher
because of as initially water for evaporation comes from the regions near the surface.



Figure: 2 Effect of air temperature on drying time of grapes at air velocity 0.5 m/s with air
temperature 50ºC, ? 60ºC, 70º C, 80ºC, - page predicted.


6

---

International Solar Food Processing Conference 2009

Figure: 3 Effect of air velocity on drying time of grapes at air temperature 60ºC with air
velocities 0.25 m/s, ? 0.5 m/s, 0.75 m/s, 1.0 m/s- page predicted.

Table : 1

Experimental Drying Conditions and Drying Time to reach the Final Moisture Content (17-
18%) of Thompson Seedless Grapes.

Run
T
V
R.H.
Mo
t
No.
(ºC)
(m/s)
(%)
(% d.b.)
(hrs)
1 50 0.25
12.00
414.07 52
2 50 0.50
12.50
415.74 47
3 50 0.75
13.00
417.12 44
4 50 1.00
13.00
419.32 41
5 60 0.25
8.00
411.70 28
6 60 0.50
8.50
414.49 26
7 60 0.75
9.00
415.55 24
8 60 1.00
8.50
414.45 21
9 70 0.25
4.50
406.90 20
10 70 0.50
5.00 408.51 19
11 70 0.75
5.50 411.38 16
12 70 1.00
5.00 412.11 15
13 80 0.25
3.50 409.94 11
14 80 0.50
3.50 410.21 10
15 80 0.75
3.00 418.71 09
16 80 1.00
4.00 401.07 08


As drying progresses the drying rate decreases with decrease of moisture content, as the
water to the evaporated comes from parenchymal cells within the structure and must be
transported to the surface. The falling rate region is indicative of an increased resistance to
both heat and mass transfer through the inner cells and increased thickness of the crumpled
and shrunken skin.


7

---

International Solar Food Processing Conference 2009
For drying conditions of each run the experimental data is fitted in equations (2), (3) and (4),
the values of constant K0 and A0 of Eq. (2) : A1, A2 and K1. K2 of Eq. (3) and K and N of Eq.
(4), were obtained and are given in Tables: 2, 3 and 4 respectively. The goodness fit of each
model to the observed data is evaluated in terms of statistical parameters R2 and X2 and is
also given in these tables.

Table : 2

The results of Nonlinear Regression Analysis of the individual drying curve at different air
temperatures and velocities for Thompson Seedless grapes using single term exponential Eq.
(2).


Run
Temp.
Velocity
A0
K0 (x10-2
R2
Chi2 (x10-4)
No.
(ºC)
(m/s)
hrs-1)
1 50 0.25 1.03 5.52 0.997 2.27
2 50 0.50 1.03 6.07 0.997 2.18
3 50 0.75 1.03 6.52 0.996 3.14
4 50 1.00 1.03 7.07 0.995 3.95
5 60 0.25 1.03 11.49 0.996 3.20
6 60 0.50 1.04 12.48 0.995 4.05
7 60 0.75 1.04 13.28 0.994 4.84
8 60 1.00 1.04 14.14 0.994 5.24
9 70 0.25 1.04 15.47 0.995 4.52
10 70
0.50 1.04 17.36 0.995 4.64
11 70
0.75 1.04 19.56 0.995 4.62
12 70
1.00 1.04 21.48 0.994 5.19
13 80
0.25 1.02 33.81 0.998 2.17
14 80
0.50 1.02 36.50 0.996 2.71
15 80
0.75 1.02 39.84 0.997 3.53
16 80
1.00 1.02 45.72 0.996 4.42



It is evident from Tables: 2, 3, 4 and Fig. 5, that the Page’s model (Eq.4) fits best with the
experimental data for its better values of R2 and Chi2 compared to corresponding values for
Eq.(2) and (3). Hence for further analysis only Page’s model is considered.

The variation of Page’s constant N with air velocity and temperature is shown in Table 4 and
Fig. 6. It is seen that N increases with increase in air velocity but this increase is very
marginal (about 1.4%).

Table: 3

The Results of Nonlinear Regression Analysis of the Individual Drying Curve at Different Air
Temperature and Velocities for Thompson Seedless Grapes using Two Term Exponential Eq.
(3).

Run
Temp.
Velocity
A1
K1 (x10-2
A2
K2 (x10-
R2
Chi2
No.
(ºC)
(m/s)
hrs-1)
2 hrs-1)
(x10-4)
1 50 0.25
0.583
5.52
0.443
5.53
0.997
2.36
2 50 0.50
0.584
6.07
0.444
6.07
0.997
2.28
3 50 0.75
0.586
6.52
0.446
6.52
0.996
3.29
4 50 1.00
0.587
7.07
0.448
7.07
0.995
4.16

8

---

International Solar Food Processing Conference 2009
5 60 0.25
0.588
11.49
0.450
11.49
0.996
3.45
6 60 0.50
0.591
12.48
0.451
12.48
0.995
4.40
7 60 0.75
0.591
13.28
0.451
13.28
0.994
5.30
8 60 1.00
0.591
14.14
0.451
14.14
0.994
5.83
9 70 0.25
0.589
15.47
0.449
15.49
0.995
5.05
10 70 0.50 0.589
17.36
0.449
17.35
0.995 5.22
11 70 0.75 0.588
19.56
0.449
19.59
0.995 5.33
12 70 1.00 0.588
21.48
0.449
21.47
0.994 6.06
13 80 0.25 0.583
33.81
0.434
33.82
0.998 2.71
14 80 0.50 0.584
36.50
0.434
36.50
0.996 3.48
15 80 0.75 0.585
39.84
0.435
39.84
0.997 4.71
16 80 1.00 0.540
45.72
0.480
45.71
0.996 6.19


Also N increases with increase in air temperature between 50-70ºC while between 70-80ºC, it
decreases with increasing air temperature. But this dependence of N on temperature is also
very weak (only about 1-5%). The dependence of N on velocity and temperature of air was
also modeled in the form of Arrhenius and Power Equations (Eqs. 5 and 6), but the obtained
values of R2 (= 0.23) indicate very poor fit. Hence for very small variations of N with air
velocity and temperature it is tempting to ignore these variations and use the average value of
N ( =1.13) obtained for all velocities and temperature of the sixteen runs.


Figure 4: drying rate curves of grapes at air velocity of 0.5 m/s; with different air temperatures
50ºC, ? 60ºC, 70º C, 80ºC.



9

---

International Solar Food Processing Conference 2009


Figure: 5 Drying curves for goodness of fit, o experimental; - page equation, --------two terms
and …….single term.

Table : 4

The results of nonlinear regression analysis of the individual drying curve at different air
temperatures and velocities for Thompson seedless grape using Page’s Eq/ (4).

Run No.
Temp.
Velocity
N K
(x10-2 hrs-
R2
Chi2 (x10-5)
(ºC)
(m/s)
1)
1 50
0.25
1.10
4.02
0.999
7.59
2 50
0.50
1.10
4.37
0.999
5.24
3 50
0.75
1.12
4.46
0.998
8.36
4 50
1.00
1.13
4.67
0.998
0.10
5 60
0.25
1.13
8.08
0.999
4.71
6 60
0.50
1.15
8.52
0.999
6.08
7 60
0.75
1.16
8.90
0.999
7.67
8 60
1.00
1.17
9.49
0.999
6.46
9 70
0.25
1.15
10.87
0.999
6.40
10 70 0.50
1.15
12.46
0.998 0.10
11 70 0.75
1.15
14.26
0.999 6.49
12 70 1.00
1.16
15.69
0.999 8.16
13 80 0.25
1.08
30.13
0.998 0.12
14 80 0.50
1.09
32.01
0.998 0.12
15 80 0.75
1.12
34.16
0.999 9.31
16 80 1.00
1.14
38.87
0.999 5.75

10

---