P. Wilaipon / Journal of Agricultur al C hemistry and Environment 3 (2014) 1-4

Copyright © 2014 SciRes. OPEN A CCESS

drying [6]. Drying periods for the case of microwave, hot

air, and combined microwave-air drying methods were

studied. It was reported that the latter was accounted for

the shortest drying period.

The aim of this study was to evaluate the drying cha-

racteristics of combined microwave-air method for the

case of cassava. Furthermore, the mathematical model

parameters were also calculated by using regression

technique.

2. MATERIAL AND METHOD

2.1. Material

Rayong-9 cassavas with an initial moisture content of

61% on wet basis were obtained from a local factory in

Phitsanulok, Thailand. T heir initial moisture content val-

ue was examined, according to ASAE S358.2 DEC99

standard, by using a cabinet hot-air dryer (Memmert 600,

30˚C - 350˚C, 2400 W) and a digital balance (accuracy

0.001 g). Then, the material was cut into 10 mm thick

and 25 - 50 mm diameter with the cutting machine. All

cassavas used in the experiment were from the same

batch.

2.2. Drying Experiment and Data Analysis

The drying s ystem was comprised of two 86 × 43 mm

rectangular waveguides, two air-cooled magnetrons, and

a 44 × 51 × 93 cm cavity. Two 800 W-magnetrons used

in the experiments work at the frequency of 2.45 GHz.

They were installed in the waveguides mounted on the

same plane, the top of the cavity. Four heaters, 2 kW

each, were installed at the air inlet duct. A temperature

controller (Shimax MAC5D) and type K thermocouple

were utilized for temperature control purpose. In order to

record the sample weight loss, a 15 kg single-point loa d-

cell coupled to a load cell indicator (Primus CM 013)

was installed on the top of the cavity. Additionally, A

Testo 435, accuracy ±0.25˚C and ±2% RH, was used for

measuring the temperature and relative humidity of inlet

air.

In all experiments, approximately 2.5 kg of samples

were used. The samples were uniformly spread on a dry-

ing tray and placed in the drying cavity. The temperature

and velocity of hot air were set at 60 ˚C and 1 m/s respec-

tively. A temperature sensor was utilized for measuring

surface temperature of the sample. It was used as an in-

put for microwave power operation control. The experi-

ments were investigated at two levels of sample temper-

ature viz. 70˚C and 80˚C respectively.

The values of moisture ratio were calculated using the

following equation:

(1)

where :

MR is the moisture ratio;

Mt is the moisture content at 1 hour (%);

Me is the equilibrium moisture content (%);

Mi is the initial moisture content (%).

Several conventional drying models have been pro-

posed for determining the moisture ratio as a function of

drying time. In this research, the drying models of cas-

sava drying by using 2 planes magnetron microwave-air

drying system were investigated. Newton model [7],

Page’s model [8], logarithmic model [9], Henderson &

Pabis model [10], and diffusion model [3] were applied

to describe the characteristics of cassava drying.

(2)

(3)

(4)

(5)

() ()() ()

MRa expkt1aexpktb=− +−−

(6)

where :

k is the drying constant;

n is the power parameter;

a and b are parameters;

t is drying time (hour).

Coefficient of determination (R2), adjusted coefficient

of determination

, and standard error of esti-

mation (SEE) were utilized to evaluate the goodness of

fit of the tested drying models to the experimental data.

3. RESULT AND DISCUSSION

Effective mathematical model of drying characteristic

is crucial for cassava microwave-air drying kinetics in-

vestigation. The combination of microwave and hot-air

energy were able to reduce the sample moisture content

from 61% to 8% db in 300 - 340 minutes depending on

the levels of sample temperature set point. It was found

that as the set-point temperature increased, the drying

time was decreased. By using non-linear regression tech-

nique, the drying constants and coefficients of the five

models obtained are shown in Table 1.

In order to evaluate goodness of fit, coefficient of de-

termination (R2), adjusted coefficient of determination

, and standard error of estimation (SEE) were

also computed. The goodness of fit was determined by

the higher R2 and

values as well as the lower

SEE values. For all cases, it was found that R2 and

values were higher than 0.98, and SEE values

were lower than 0.029.

Furthermore, it was found that diffusion and Page’s

models gave the excellent fit results for all the experi-

mental data. For the case of diffusion model regression,

the values of R2,

and SEE for 70˚C - 80˚C set-