Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com
Procedia
Food Science
Procedia Food
Procedia FoodScience
Science 00 (2011)713
1 (2011) 000000
721
www.elsevier.com/locate/procedia

11th International Congress on Engineering and food (ICEF11)

Modelling of drying kinetic of potatoes taking into account

shrinkage
Marcelo S. Bacelosa*, Paulo I.F. Almeidab
a
Departamento de Engenharias e Computao, Centro Universitrio do Norte do Esprito Santo, Universidade Federal do Esprito
Santo, So Mateus-ES, Rodovia BR 101 Norte, Km. 60, Bairro Litorneo, 29932540, Brazil (marcelobacelos@ceunes.ufes.br
b
Departamento de Engenharia Qumica, Centro de Cincias Exatas e da Terra, Universidade Federal de So Carlos, Brazil

Abstract

A natural convection oven with controlled air temperature, vision and weighting system interfaced with computer
was used to obtain the drying kinetics data. Spherical potatoes samples with 10mm diameter are used in the
experiments. During drying, the operating conditions carries out inside the oven are an air temperature of 60C and
absolute humidity of 5 kgH2O/ 100 kg dry air. Shrinkage models are tested to describe gradually the particle size
reduction over time. Based on data analyses, a diffusive model, numerically solved by the method of finite control
volumes, to describe the drying kinetics of potatoes considering the effect of shrinkage on mass transfer is proposed.
The simulated data show a good agreement with the experimental ones as it can be proved by the significant
correlation coefficient (r = 0.9984) and low maximum experimental error (4.15%) achieved.

2011 Published
2011 by Elsevier
Published byB.V. SelectionLtd.
Elsevier and/orSelection
peer-review and/or
under responsibility
peer-review of 11thunder
International Congress
responsibility of ICEF11
on Engineering and Food (ICEF
Executive Committee Members 11) Executive Committee.

Keywords: model, drying, potatoes, kinetics, food, shrinkage;

1. Introduction

Potatoes are members of the Solanaceae family. Of the many tuber-forming Solarium species, the one
that is most widely cultivated is Solarium tuberosum. In 2007 total world production of potatoes was
more than 320 million tonnes, and about 66% were consumed by people as food. The other 34%
production is used as animal feed, and as potato starch in pharmaceuticals, textiles, adhesives industries.
In industry, the drying technologies have been widely used for processing food products. However,
during the drying food materials may undergo undesirable changes such as shrinkage that has a negative
effect on the required characteristics of dehydrated product for commercial usage. The analyses of the
drying kinetics data permits to understand the moisture transfer inside the foods considering the shrinkage

* Corresponding author. Tel.:+55-27-3312-1568 fax: +55-27-3312-1618.

E-mail address: marcelobacelos@ceunes.ufes.br.

2211601X 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of 11th International Congress on Engineering
and Food (ICEF 11) Executive Committee.
doi:10.1016/j.profoo.2011.09.108
7142 Marcelo S.Author
Bacelosname
and /Paulo I.F. Almeida
Procedia / Procedia
Food Science Food Science
00 (2011) 1 (2011) 713 721
000000

of samples and mechanisms involved. In the last fifty years, many studies were done to investigate the
drying of foods, especially potatoes [1-19]. Such analyses aid practitioners not only to decide how
significant is internal resistance as compared with external resistance to the moisture transfer but also
select a suitable air-food contactor assuring better quality attributes of final products. Thus, the aim of this
research is to present an experimental and theoretical study of drying kinetics considering the shrinkage
of spherical potatoes samples.

Nomenclature

X, X0, XE food moisture: punctual at time "t"; Initial, equilibrium, kg H2O/kg dry solid [ - ]
dimensionless free moisture [-]
rW, rP, rE position coordinates of method of finite volume control [L]
r position inside of cylinder [L]
*
r dimensionless posision [-]
R position on the surface of solid [ L ]
DEF Effective Diffusivity [ L2/T ]
s dry solid density, kg H2O/m3 dry solid [ M/L3 ]
a dry air density, kg H2O/m3 dry air [ M/L3 ]
KP partition coefficient, kg dry air /kg dry solid [-]
Y2, Y interface and bulk moisture content, kg H2O /kg dry solid [ - ]
t drying time, s [T]
aP, aE, aW, aK coefficients of numerical grid [ - ]
KM2 natural convective mass transfer coefficient [ L/ T]
2
A, A0 area punctual at time and initial [ L ]
1 first root: Bessel function of first order [ - ]
r* position between two any numerical grid points [ - ]

2. Modeling of drying kinetics taking into account the shrinkage

The model proposed for predicting the drying kinetics of potato taking into account the shrinkage of
samples is showed in Table 1, with the boundary conditions.
 Marcelo S. Bacelos and Paulo
Author name I.F. Almeida /Food
/ Procedia Procedia Food
Science 00Science
(2011) 1000000
(2011) 713 721 715 3

Table 1. Model of drying kinetics taking into account shrinkage of spherical potatoes

1.Physical considerations and control volume:

- spherical geometry;
- transient unidirectional mass transfer phenomena in radial
direction;
w P e
- material is homogeneous and isotropic;
- transient heat transfer: lamped capacitance model, which is r
based on the assumption of a spatially uniform temperature 2
distribution within a food (BiH <0,1);
r
- moisture distribution initially is uniform in the whole food
material; Figure 1- Control volume

-diffusion in only one direction (radial);

-evaporation occurs on the surface of the food; - The effective diffusivity varies with the moisture content;
- diffusion moisture flow inside the food material surface is same
as convection moisture flow on external surface.
3. Mass transfer
Mass balance for moisture in food: Dimensionless groups:
X 1 X (1)
r*
r K M a R X XE
DEF r 0r R ; t0 BioM
t r r r R DEF S K P X0 XE
Boundary and initial conditions:
t 0; 0 r R X X0 The mass balance for moisture in food can be rewritten:

X DEF *2 X (2)
r 0; t 0 X
0 r*
2
r * 0 r* 1 ; t 0
r t R r* r
r R; t 0 X Dimensionless boundary and initial conditions:
s DEF
r
1
a K m Y2 Y t 0; 0 r* 1

r* 0 t 0
0
r *
r* 1 t 0
Bio M
r *
4. Numerical solution: Method of finite control volume [20]
Full implicit discretization of the Equation (2) in the control The Equations ( 5 ), ( 6 ) are substituted in Equation ( 4 ) and this
volume according to Figure 1: equation may expressed as:

r * 3 * 3
E r W 3 r * 2
P r
* (7)

r *E t t * 2 (3)
r*W t r dt dr * By solving the integration (Equation 3) and substitute the
t Equations (5), (6) and (7) in (3), it results in following equation:
DEF * 2
tt t rr* E * r * * dr * dt aP (i) P (i) aW(i) W(i) aE(i) E(i) aK(i) (8)
R2 W
r r

Equation (8) is the general discretization equation present by

Relationships between dimensionless radius in the following
Patankar [20], using finite volume control to solve diffusive
* * *
positions: r E , r W and r P (see Figure 1) models.

r *E 3 r *W 3 r * *
E r W r *E 2 r *E r *W r *W 2 (4)
716 4 Marcelo S. Author
Bacelosname
and Paulo I.F. Almeida
/ Procedia Food /Science
Procedia
00Food Science
(2011) 1 (2011) 713 721
000000

Coefficients in the inside points of numerical grid

r * (5) 2 i N P 1 can be calculated as:

r *E r *P
2 aE(i) r* r* (9)
*2
r
r * E 2 r * P 2 r * P r * aW(i) r* r* ( 10 )
4
(6) r* r*2 R2 ( 11 )
r *
r *W r *
P aK (i) (k)
2 DEF t P
2

r *
r *W 2 r * P 2 r * P r *
4
5. Discretization of boundary conditions:
At the center: The boundary condition on the surface of numerical grid can be
a P 1. P 1 a W 1. E 1 a E 1. E 1 a K 1
discretizated by a second order approuch:
( 12 )
a P NP a P NP 2 a E NP r * BiM ( 14 )

It is assumed the first order approach in the center, resulting in the

following coefficients of numerical grid as:
aW ( NP ) a E ( NP ) aW ( NP ) ( 15 )
a P 1 a P 1 ( 13 )

a E 1 a E 1 aW 1 a K NP a K NP ( 16 )

a K 1 a K 1
aW 1 0 a E ( NP ) 0 ( 17 )

6. Effect of Shrinkage on mass transfer The Kilpatrick et al. [21] models fitted better the experimental
results, thus radius of samples can be expressed by:
1 ( 18 )
0.8 X 3
R R0
0.8 X 0
7. Effective diffusivity
Mclaughlin & Magee [9] relate the effective diffusivity with the For each new radius, determined by image analyses and expressed
shrinkage by using the definition of a drying constant (K): by equation (21), a new value of effective diffusivity was
( 19 ) obtained:
2 DEF
K
R2 2 ( 20 )
In the initial time R = 10 cm and DEF = 5.0.10 -10 2
m /s. Then the DEF c.(a b X 3)
valor of K can be determinate by Equation (19).
where, a, b and c are adjusted parameters to the fitted model.

3. Materials & Methods

Spherical potatoes samples with 10mm diameter were dried in the natural convenction oven at an air
temperature of 60C and absolute humidity of 5 kgH2O/ 100 kg dry air. For each test and before drying,
samples are blanched 5 minutes in 90C water, and then the surface water is removed with a filter paper.
The weight and volume changes of three samples of food are continuously measured by a vision and
weighting system interfaced with computer, as schematically shown in Figure 2.
 Marcelo S. Bacelos and name
Author Paulo/I.F. Almeida
Procedia / Procedia
Food Food
Science 00 Science 1 (2011) 713 721
(2011) 000000 7175

Probe Load cell

Oven
Camera

PT100
SP 60

Computer
Controller

Fig. 2. Experimental apparatus

Shrinkage models are tested to describe gradually the particle size reduction over time. Based on data
analyses, a diffusive model, numerically solved by the method of finite control volumes, was proposed to
describe the drying kinetics of potatoes taking into account the shrinkage. Details on model were showed
in Table 1.
The error between experimental and calculated values was obtained using the following equation:

exp calc (21)

E x100
exp

4. Results & Discussion

Figure 3 shows the shrinkage experimental data and adjusted models. Concerning the shrinkage
models tested, both Kilpaltrick et aL. (1955) and linear models fitted better the experimental data. This
can be proved by calculated statistics parameters such as correlation coefficient and maximum
experimental error depicted in Table2. The found results showed that samples presented a linear-
shrinkage behavior with moisture (A/A0 vs. X) until it reaches a limit moisture content of about 1 (dry
basis). From this point, the material stops to shrinkage. However, McLauglin e Maggee [6] found a limit
shrinkage at moisture content lower than 0.5 (dry basis), where samples became rigid and friable.

1.0 Experimental data 1.0 Experimental data

Kilpatrick et al. (1955) model's Linear model
0.9 0.9

0.8 0.8
A/A0

A/A0

0.7
0.7

0.6
0.6

0.5
0.5

0.4
1 2 3 4 5 0.4
1 2 3 4 5
Moisture content (dry basis)
(a) Moisture content (dry basis)
(b)

Fig. 3. Shrinkage models. (a) Kilpaltrick et al. (1955) models, (b) Lineal model
718 6 Marcelo S. Author
Bacelosname
and Paulo I.F. Almeida
/ Procedia Food /Science
Procedia
00Food Science
(2011) 1 (2011) 713 721
000000

Table 2. Shrinkage parameters fitted to experimental data

Models A B r E(%)
Kilpatrick 0.80000 0.80000 0.9986 9.92
Linear 0.16229 0.16322 0.9972 10.83

Figure 4 show the photographs of spherical samples during the drying in a natural convective oven. It
can be observed in Figure 4 that samples maintain the spherical shapes with the loss of moisture during
drying. Such results can be proved by the analysis of shape factor (not shown herein) along with the
drying time, during which this parameter remains close to the unit.

(a) (b) (c) (d)

(e) (f) (g) (h)

Fig. 4. Shrinkage of samples during drying time, a)0h (beginning); b)1h; c)2h; d)3h; e)4.5h; f)5.5h; g)7.35h; h)9.5h

Figure 5 presents the drying flux as a function of time considering and neglecting size reduction of
samples. Either neglect or consider the shrinkage, it can be noted a short constant drying flux followed by
long falling one. These results were in good agreement to those reported by Mc Minn and Magee [6].
0.5

without shrinkage
with shrinkage
0.4
Drying flux [kg/s.m )]
2

0.3

0.2

0.1

0.0
0 5000 10000 15000 20000 25000 30000 35000 40000
Time [s]

Fig. 5. Drying flux as a function of time

 Marcelo S. Bacelos and Paulo
Author name I.F. Almeida /Food
/ Procedia Procedia Food
Science 00Science
(2011) 1000000
(2011) 713 721 719 7

In addition, by using the method of finite control volumes to solve the differential equation of mass
transfer for moisture content inside the spherical potato (details in Table 1), the dimensionless free
moisture content profile along with radius axis was determined, as can be seen in Figure 6. Noted that, at
the beginning of drying (i.e., at 3600s) the moisture content inside the food remains almost spatially
uniform and the moisture content shape profile is sigmoid. From time of 10800s to 36000s the moisture
content shape profiles are parabolic, a fact that is consistent with the model equation.
Figure 7 depicts the experimental and simulated data of dimensionless free moisture content as a
function of drying time. In the beginning of drying, the diffusive model, not taking into account the
shrinkage, fits better the data, as can be seen in Figure 7. After, such model underestimates the internal
mass transfer phenomena as it does not take into account the changes in internal resistance promoted by
size reduction over time, leading to a significant effect on describing the drying kinetics data. However,
model taking into account shrinkage present a better agreement to the experimental data. This can be
proved by high significant correlation coefficient and low maximum error (e.g., r=0.9984, E=4.15).
Moreover, at the beginning of drying, the estimated Biot number (BiM) reaches the value of 20. After, this
value continuously changes until it reaches a value close to 30 for 10h drying. Noted that, as BiM equals
to 30 the equilibrium moisture content on the surface of samples is achieved (i.e., 0.02152, in dry basis).
Such results, for BiM are in agreement to those reported by Hernandez et al. [19].

1.0
Dimensionless free moisture content

3600 s
0.8 7200 s
10800 s
14400 s
[(X-XE)/(X0-XE)]

0.6 18000 s
21600 s
25200 s
0.4 28800 s
32400 s
36000 s
0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Dimensionless radius [r/R]

Fig. 6. Dimensionless free moisture content as a function of dimensionless radius axis at a given drying time

Experimental data
1.0
Model without Shrinkage
Model with Shrinkage
Adimensional de Umidade Livre

0.8
[(X-XE)/(X0-XE)]

0.6

0.4

0.2

0 7200 14400 21600 28800 36000

Tempo [s]

Fig. 7. Dimensionless free moisture content as a function of drying time

7208 Marcelo S.Author
Bacelosname
and /Paulo I.F. Almeida
Procedia / Procedia
Food Science Food Science
00 (2011) 1 (2011) 713 721
000000

5. Conclusion

The shrinkage of spherical potatoes can be described by both tested models: Linear and Kilpaltrick.
With respect to model of drying, at the beginning, the diffusive model neglecting the shrinkage of
samples fits well the experimental data. As the drying progresses, such model overestimates mass transfer
phenomenon as it does not take into account the modifications in the internal resistance on mass transfer.
However, the model considering the shrinkage presents a better agreement with the experimental data of
drying kinetics. This suggests that, although BiM, in order of magnitude, is not significantly changed (i.e.,
50% higher) from beginning to the end of drying (e.g., 20 to 30), the reduction of sample size has an
important effect on predicting drying kinetics. This suggests that model is highly sensitive to
dimensionless Biot number (BiM)

Acknowledgements

The authors would like to express their gratitude to Conselho Nacional de Desenvolvimento Cientfico
e Tecnolgico (CNPq) and Fundao de Amparo Pesquisa do Esprito Santo (FAPES) for the financial
support in carrying out this research.

References

[1] Karathanos V.T. & Belessiotisb V.G. 1999. Application of a Thin-Layer Equation to Drying Data of Fresh and Semi-
dried Fruits. Journal of Agricultural Engineering Research, 74(4), 355-361.
[2] McMinn, W. A. M., & Magee, T. R. A. (1997). Physical characteristics of dehydrated potatoes part I. Journal of Food
Engineering, 33(1/2), 3748.
[3] Suarez, C., & Viollaz, P. E. (1991). Shrinkage effect on drying behaviour of potato slabs. Journal of Food Engineering,
13(2), 103-114.
[4] Wang, N., & Brennan, J. G. (1993). The influence of moisture content and temperature on the specific heat of potato
measured by differential scanning calorimetry. Journal of Food Engineering, 19(3), 303-310.
[5] Wang, N., & Brennan, J. G. (1995b). A mathematical model of simultaneous heat and moisture transfer during drying of
potato. Journal of Food Engineering, 24(1), 47-60.
[6] MC MINN, W. A. M.; MAGEE, T. R. A.(1997). Physical characteristics of dehydrated potatoes-part I. Journal of Food
Engineering, v. 33, p. 37-48.
[7] MC LAUGLIN, C. P.; MAGEE, T. R. A. (1998a). The determination of sorption isotherm and the isosteric heats of
sorption for potato. Journal of Food Engineering, v. 35, p. 267-280, 1998a.
[8] MC LAUGLIN, C. P.; MAGEE, T. R. A. (1998b). The effect of shrinkage during drying of potato spheres and the effect
of drying temperature on vitamin c retention. Food and Bioproducts Processing, n. C3, p. 138-142, 1998b.
[9] MC MINN, C. P.; MAGEE, T. R. A. (1999). Principals, methods and applications of convective drying of foodstuffs.
Food and Bioproducts Processing, n. C, p. 175-193, 1999.
[10] Wang, N., & Brennan, J. G. (1995a). Changes in structure, density and porosity of potato during dehydration. Journal of
Food Engineering, 24(1), 61-76.
[11] MOHOR, K. H. (1994). Kinetics of moisture trans port in biological products under drying conditions. International
Chemical Engineering, v. 34, n. 2, p. 210-212.
[12] RAPUSAS, R. S.; DRISCOLL, R. H. (1995). The thin-layer drying characteristics of white anion slices. Drying
Technology, v. 13, n. 8/9, p. 1905-1931.
[13] ROSSELO, C. A. (1992). Simple mathematical model to predict the drying rates of potatoes. Journal Agriculture Food
Chemistry, v. 40, p.2374-2378.
 Marcelo S. Bacelos and name
Author Paulo/ I.F. Almeida
Procedia / Procedia
Food Food
Science 00 Science 1 (2011) 713 721
(2011) 000000 7219

[14] ROVEDO, C. O.; SUAREZ, C.; VIOLLAZ, P. (1998). Analysis of moisture profile, mass biot number and drying forces
during of potato slabs. Journal of Food Engineeiring, v. 36, p. 211-231.
[15] SARAVACOS et al. (1989). Effective thermal and moisture diffusivities of a granular starchs. Advances in Drying, v.11,
p. 314-319, 1989.
[16] SARAVACOS, G. D.; CHARM, S. E. (1969). A study of the mechanism of fruits and vegetable dehydration. Food
technology, v. 16, p. 78-81, 1962.
[17] YUSHENG, Z; POULSEN, K. P. (1988). Diffusion in potato drying. Journal of Food Engineering, v. 7, p. 249-262,
1988.
[18] ZOGZAS, N. P.; MAROULIS, Z. B. (1996). Effective moisture diffusivity estimation form drying data: a comparison
between various methods of analysis. Drying Technology, v. 14, n. 7/8, p. 1543-1573, 1996.
[19] HERNNDEZ, J. A.; PAVN, G.; GARCA, M. A. (1980). Analytical solution of mass transfer equation considering
shrinkage for modeling food-drying kinetics. Journal of Food Engineering, v. 45, p. 1-10, 2000.
[20] PATANKAR, S. V. (1980). Numerical heat transfer and fluid flow. New York: McGraw-Hill, 1980.
[21] KILPATRICK, P. W., LOWE, E, VAN ARSDEL, W. D. (1955). Tunnel dehydration for fruits and vegetables, in
advances in food research, Van Arsdel, W. D., (ed), Academic Press, New York, Vol.6, pp.152-156.

Presented at ICEF11 (May 22-26, 2011 Athens, Greece) as paper MCF1214.