INTRODUCTION
In recent years, fish farming is gaining popularity in India due to providing
nutritional security to the food basket and earning foreign currency.
Fish is a very important foodstuff in developing countries, due to its
high protein content and nutritional value, supplying approximately 6%
of global protein requirement and 16% of the total animal protein (Ayyappan
and Diwan, 2003). Minor fish species consumed with bones and shell body
are good source of calcium, protein, vitamin B and vitamin B12 (Zakhia,
2000).
Fish drying is practice to preserve the fish for longer duration. The
spoilage reactions connecting on the death of fish proceed at very rapid
rate. Fresh fish contains up to 80% of water. It is highly perishable
material and having a short storage life (Bala and Mondol, 2001). Some
traditional techniques for improving preservation and storage are fish
salting/brining, open sun drying and smoking. Open sun drying is still
the most common method used for preserving food products in tropical and
subtropical countries. Drying is widely used industrial preservation method
in which water activity of food is decreased to minimize biochemical,
chemical and microbiological deterioration (Doymaz and Pala, 2002). With
the completion of sun drying, fish meat becomes condensed, saturated with
oil, becomes translucent and acquires an amber color, a typical flavor,
dense consistency and pleasant taste (Gerasimov and Antonova, 1979). Open
sun drying has some limitation, there are considerable losses and fish
quality is lowered because of over drying, insufficient drying and contamination
by foreign materials, insects and microorganisms as well as discoloring
by UV radiation (Tiwari and Sarkar, 2007). In comparison to open sun drying
the use of greenhouse dryer lead to reduction of the drying time up to
50% and to a significant improvement to the product quality in terms of
color, texture and taste (Esper and Muhlbauer, 1998). For onion flakes
the rate of moisture evaporation in case of greenhouse drying is more
than that of open sun drying during the off sunshine hours due to the
stored energy inside the greenhouse (Kumar and Tiwari, 2006).
The modeling of heat and mass transfer mechanism for solar drying of
fish is complex. The convective heat transfer coefficient is an important
parameter in drying rate simulation since the temperature difference between
the air and fish varies with this coefficient (Jain, 2006). Ratti and
Crapiste (1995) have evaluated the convective heat transfer coefficient
from the data on crop drying under forced mode of operation. The experimental
convective heat transfer coefficient for potatoes, apples and carrots
ranged from 25 to 90 W m^{2}. The convective heat transfer coefficient
of jaggery under solar drying has been evaluated by Tiwari et al.
(2004). Anwar and Tiwari (2001), as also Jain and Tiwari (2003) evaluated
the convective heat transfer coefficient for some crops (green chilies,
green peas, white gram, onions potatoes and cauliflowers) under solar
drying and developed a mathematical model for predicting the drying parameter.
The forced convection drying is better than natural convection drying
because it reduces the drying time of fish.
No such work has been reported so far on greenhouse fish drying under
forced convection mode. For the first time, experiment have been conducted
by Tiwari et al. (2006) to determine the coefficient of convective
heat transfer for greenhouse fish drying under natural convection mode.
However no one has studied the heat and mass transfer of greenhouse fish
drying under forced convection mode.
Therefore, the present studies were undertaken to determine the convective
heat and mass transfer coefficient at different durations of greenhouse
drying time of prawn under forced convection mode.
MATERIALS AND METHODS
Experimental Observation
For fish drying under forced mode, Indian minor carp prawn (Macrobrachium
lamarrei); invertebrate was considered for drying in greenhouse. Dead
and fresh fish was purchased from local market near IIT Delhi. The fresh
fish was washed with fresh water. Surface water was removed by blotting
with absorbent paper. A steel wire mesh tray was used during drying of
the fish. The dimension of tray was of 0.25x0.20 m. The fish were arranged
in a single layer in the drying tray. The tray with fish was kept on the
measuring balance. Experiments were conducted in July 2006 between 8:00
and 17:00h for forced drying and 10:00 to 17:00 h for natural drying under
the composite climate of New Delhi, India as described in Tiwari et
al. (2006). The solar radiation ranged during these hours between
150 and 900W m^{2}. Similarly Jain (2006) determined the convective
heat and mass transfer coefficients for solar (open) drying of fish.
Experimental Greenhouse
The photograph of experimental greenhouse dryer at Solar Energy Park,
IIT Delhi (Latitude28°35^{/} N, Longitude77° 12^{/}
E and an altitude of 216 m above mean sea level) for fish drying under
natural and forced mode is shown in Fig. 1a, b,
respectively. The plastic cover of greenhouse transmits the solar radiations
inside the greenhouse. The fraction of transferred solar radiation received
partly by the fish, floor, exposed tray area and remaining solar radiation
will heat the enclosed air inside the greenhouse. An even span roof type
greenhouse with effective floor area of 120x0.78 m^{2} was used
for experimental purposes. The orientation of the greenhouse dryer was
fixed with eastwest direction. The inclination of south and north roof
was 25.90°. The central height and sidewalls were raised to 0.60 and
0.40 m, respectively. For forced convection a fan of 120 mm sweep diameter
with air velocity 5 m sec^{1} was provided on the sidewall of
the greenhouse during the experiments. There were provisions of two vents
each of 0.2x0.1 m^{2} on the south and north roof for natural
ventilation purposes during over heating inside the greenhouse, if any.


Fig. 1a, b: 
Experimental setup for prawn drying inside greenhouse
under forced convection mode 
Instrumentation
For measurement of temperature, a noncontact thermometer (RaytekMT4)
having least count of 0≥1°C with accuracy of ±2% was used.
The range of thermometer was 18 to 260°C. A digital hygrothermometer
(model: Lutron HT3003)with least count of 0.1% was used to measure the
relative humidity inside greenhouse. A digital balance of 1kg weighing
capacity was used to weigh the sample during the drying with least count
of 0.1 g. The difference in weight gave the moisture evaporated during
drying. The solar radiation was measured with a solarimeter, in W cm^{2}
having a least count of 2 m W cm^{2} with + 2% accuracy. The
air velocity across the greenhouse section during the forced convection
mode was measured with an electronic digital anemometer (model : Lutron
AM4201). It had a least count of 0.1 m sec^{1} with ±2%
accuracy on the full scale range of 0.240.0 m sec^{1}. Ambient
air temperature (T_{a}) and just above the surface temperature
of fish inside greenhouse temperature (T_{g}) were measured by
calibrated alcohol  filled, glass  bulb thermometers (least count was
1°C).
Computation Methodology
Convective Heat Transfer Coefficient
The Nusselt number (Nu) is a function of Reynolds (Re) and Prandtl (Pr)
numbers for forced convection heat transfer.
Now, the forced convective heat transfer coefficient is determined by
using the following expression obtained from Eq. 1:
The rate of moisture evaporate is given (Malik et al., 1982) by:
After substituting expression for h_{c} from Eq.
2 in 3 we get:
From the above equation we can get the rate of mass evaporated:
After algebraic simplification:
After taking the logarithm of both sides of Eq. 6, we
get:
Equation 7 is of a straightline equation and is rewritten
as:
b_{1} 
= 
b_{o }are the independent and dependent variables.

X 
= 
ln (Re, Pr) 
b_{o} 
= 
ln C can be rewritten as C = e ^{bo}(8) 
Once the values of the C and n are known, the convective heat transfer
coefficient is computed by Eq. 2 using measured values
of the ambient air, inside greenhouse air, surface temperature of fish
and relative humidity in greenhouse condition during a given time period
(Table 2).
Computation Technique
The average surface temperature of fish
and inside greenhouse temperature above the fish surface
were calculated at hourly intervals for corresponding moisture evaporated.
The physical properties of humid air were evaluated for the mean temperatures
of and
using equations given in appendix. These physical properties were utilized
for calculating the values for the Reynolds (Re) and Prandtl (Pr) numbers.
The values of C and n in Eq. 2 were obtained by linear
regression analysis expressed in Eq. 8 at the increment
of every hour of observation and thus the mean values of h_{c}
were computed at the corresponding hour of drying. The convective heat
transfer coefficient in case of natural drying was computed following
the method given in Tiwari et al. (2006). The computer program
was prepared in the Excel.
Table 1: 
Observation on greenhouse fish drying under natural
convection (initial weight, 163 g; number of fish 250; monthJuly,
2006) 

Table 2: 
Observation on greenhouse fish drying under forced convection
(initial weight, 163 g; number of fish 250; monthJuly, 2006) 

RESULTS AND DISCUSSION
The input parameter required for the computation of convective heat transfer
coefficients for prawn (invertebrates) during greenhouse drying for both
the case is given in Table 1 and 2,
respectively. It took 12h (two sunny days) and 9 h (one day) to dry the
fish for natural and forced convection mode of drying, respectively. The
moisture evaporation rate was higher in the initial few hours of drying.
Changes in coefficients C and n are observed as the number of observations
and drying time increase. The values of C and n are 1.47, 1.00 and 0.26,
0.22 for 1st and 2nd days observations during natural drying and 1.28
and 0.26 during forced convection drying, respectively.
The convective heat transfer coefficient h_{c} for prawn drying
ranged from 9.21.23 W m^{2} °C in case of natural convection
and 211.5 W m^{2} °C in case of forced convection drying.
The convective heat transfer coefficient declined with a decrease in moisture
content of prawn as expected. Experimental convective heat transfer coefficients
during forced convection drying have been fitted by various mathematical
relations as a function of drying time in hours viz., Linear, Y = a+bxT,
LogLinear, Y = axT^{b}, Exponential, Y = axe^{bxT }and
Quadratic, Y = a+ bxT+cxT^{2}. A Quadratic curve exhibited best
relation between convective heat transfer coefficient and drying time
as it gave the maximum coefficient of determination (R^{2} = 96.53%)
.
The variation of convective heat transfer coefficient with respect to
drying for both the condition is shown in Fig. 2. From
the Fig. 2 it is observed that the heat transfer coefficient
is almost double in the case of forced convection drying then natural
convection drying and also it drying time for forced convection is almost
reduced to half.

Fig. 2: 
Variation of convective heat transfer
coefficient with drying time for fish under forced and natural convection
in greenhouse (July, 2006) 

Fig. 3: 
Moisture evaporated during the drying
of fish under natural and forced convection in greenhouse (July, 2006)

The amount of moisture evaporated during the drying of fish under natural
and forced convection mode is shown in Fig. 3. It is
observed that amount of moisture removed in the case of forced convection
drying is higher and faster then the natural convection drying.
CONCLUSION
The convective heat transfer coefficients of minor fish species like
prawn (invertebrates) have been determined under greenhouse drying for
both natural and forced convection condition at different drying times.
Convective heat transfer coefficient was a function of moisture removal,
physical properties of moist air, operating temperature and surface area.
The heat transfer coefficient can be increased by providing forced convection
drying. Moisture removed in the case of forced convection drying is higher
and faster then the natural convection drying.
Nomenclature
A_{t} 
= 
Area of fish (tray) (m^{2}) 
C 
= 
Constant 
C_{v} 
= 
Specific heat of humid air (J kg^{1} °C) 
C_{f} 
= 
Specific heat of fish (J kg^{1} °C) 
c 
= 
Coefficient 
Re 
= 
Reynolds No. (= ρ_{v}v d μ_{v} ) 
h_{c} 
= 
Convective heat transfer coefficient of fish (W m^{2} °C) 
K_{v} 
= 
Thermal conductivity of humid air (W m^{1} °C) 
X 
= 
Characteristic dimension (m) 
m_{ev} 
= 
Moisture evaporated (kg) 
Nu 
= 
Nusselt No. (= h_{c}L/K_{v}) 
n 
= 
Coefficient 
Pr 
= 
Prandtl number (= μ_{v} C_{v}/K_{v})

P(T) 
= 
Partial vapor pressure at temperature T (N m^{2}) 
Q_{e} 
= 
Rate of heat utilized to evaporate moisture (J m^{2} s) 
R^{2} 
= 
Coefficient of determination 
T_{f} 
= 
Surface temperature of fish (°C) 
T_{e} 
= 
Temperature of humid air above the fish surface (°C) 
T_{i} 
= 
Average of fish and humid air temperature (°C) 
t 
= 
Time (sec) 
γ 
= 
Relative humidity (%) 
λ 
= 
Latent heat of vaporization (J kg^{1} °C) 
μ_{v} 
= 
Dynamic viscosity of humid air (kg m sec^{1}) 
ρ_{v} 
= 
Density of humid air (kg m^{3}) 
Appendix
The expressions used for calculating values of various physical properties
of humid air, i.e., specific heat (C_{v}) in J kg^{1}
°C, thermal conductivity (K_{v}) in W m^{2} °C,
density (ρ) in kg m^{3}, dynamic viscosity (μ_{v}
) in kg m^{1} and the partial vapor pressure (P) in N m^{2}
are given below.
C_{v} = 999.2+0.1434 T_{i}+1.101x10^{4}
T_{i}^{2} 6.7581x10^{8} T_{i}^{3}
(Kyokai and Kogaku, 1978) 
K_{v}= 0.0244+0.6773x10^{4} (Kyokai
and Kogaku, 1978) 

μ_{v} = 1.718x10^{5}+4.620x10^{8}
(Kyokai and Kogaku, 1978) 

