Energy Conversion and Management 47 (2006) 2602–2611

www.elsevier.com/locate/enconman

Modeling and simulation of reduction zone of downdraft

biomass gasifier: Effect of char reactivity factor
B.V. Babu *, Pratik N. Sheth
Department of Chemical Engineering, Birla Institute of Technology and Science (BITS), Pilani, Vidya Vihar Campus,
Pilani 333 031, Rajasthan, India

Received 26 April 2005; accepted 30 October 2005

Available online 13 December 2005

Abstract

Biomass and waste are recognized to be major potential sources for energy production. Gasification enables the con-
version of these materials into combustible gas, mechanical and electrical power, synthetic fuels and chemicals. In view of
the considerable interest in the gasification process worldwide, it is necessary to model and predict the performance of the
gasifier in priori. Modeling of biomass gasification implies the representation of chemical and physical phenomena consti-
tuting pyrolysis, combustion, reduction and drying in a mathematical form.
In the present study, steady state composition and temperature profiles for the reduction zone of a downdraft biomass
gasifier are predicted. A model reported in the literature is modified by incorporating the variation of the char reactivity
factor (CRF) along the reduction zone of the downdraft biomass gasifier. Increasing the CRF exponentially along the
reduction bed length in the model gave better predictions of the temperature and composition profiles when compared with
the experimental data and earlier models reported in the literature.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Char reactivity factor; Biomass gasification; Downdraft gasifier; Modeling; Simulation; Reaction kinetics

1. Introduction

Biomass, mainly in the form of wood, is the oldest form of energy used by humans. Biomass generally
means a relatively dry solid of natural matter that has been specifically grown or has originated as waste
or residue from handling such materials [1]. The thermochemical conversion of biomass (pyrolysis, gasifica-
tion, combustion) is one of the most promising non-nuclear forms of future energy. It is a renewable source
of energy and has many advantages from an ecological point of view [2].
The pyrolysis process consists of the thermal degradation of the biomass feedstock in the absence of oxy-
gen/air, leading to the formation of solid, liquid and gaseous products [3]. Gasification enables the conversion
of this material into combustible gas, mechanical and electrical power and synthetic fuels and chemicals [4].

*
Corresponding author. Tel.: +91 1596 245073x205, 224; fax: +91 1596 244183.
E-mail address: bvbabu@bits-pilani.ac.in (B.V. Babu).

0196-8904/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2005.10.032
 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611 2603

Nomenclature

A cylindrical bed area (m2)

Ai frequency factor for reaction i (s1)
cx molar heat capacity (J/mol K)
CRF char reactivity factor
Ei activation energy of reaction i (J/mol K)
fp fraction of pyrolysis
Ki equilibrium constant of reaction i
Ln normalized length
Lin initial normalized length at which 85% of total composition change occurs
nx molar density of species x (mol/m3)
n summation of nx of all species
Px partial pressure of gaseous species x (Pa)
ri rate of reaction i (mol/m3 s)
Rx rate of formation of species x (mol/m3 s)
R gas constant (J/mol K)
T temperature (K)
v superficial gas velocity (m/s)
z axial distance (m)

Greek letter
q density (kg/m3)

Subscripts
x species N2, CO2, CO, CH4, H2O, H2
i reaction number

Gasification is a process of conversion of solid carbonaceous fuel into a combustible gas by partial combus-
tion. Understanding the interaction between the chemical and physical mechanisms during gasification is of
fundamental importance for the optimal design of biomass gasifiers. Downdraft gasification is a compara-
tively cheap method of gasification that can produce a product gas with very low tar content [5].
Giltrap et al. [5] developed a model of the reduction zone of a downdraft biomass gasifier to predict the
composition of the producer gas under steady state operation. The accuracy of the model is limited by the
availability of data on the initial conditions at the top of the reduction zone. Moreover, it is assumed that
the char reactivity factor (CRF), which represents the reactivity of the char and is a key variable in simulation,
is taken as constant throughout the reduction zone. Jayah et al. [6] developed a model incorporating MilliganÕs
[7] flaming pyrolysis sub-model along with a gasification zone sub-model. The gasification zone sub-model
assumes that a single char particle moves downwards along the vertical axis of the gasifier. The model is lim-
ited to considering the effect of packed char particles in the reduction zone.
In the present study, GiltrapÕs [5] model is modified by incorporating a variation of the char reactivity fac-
tor (CRF) along the reduction zone of the downdraft biomass gasifier. The CRF value is increased linearly as
well as exponentially along the reduction bed length in the model. The model is simulated with a finite differ-
ence method to predict the temperature and composition profiles in the reduction zone. The model predictions
are compared with the experimental data reported by Jayah et al. [6].

2. Model

Giltrap et al. [5] proposed a phenomenological model of downdraft gasification under steady state opera-
tion for the reduction zone of the biomass gasifier. The pyrolysis and cracking reactions were not considered in
2604 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611

this model as the number of possible pyrolysis products, along with all the possible reactions and intermediate
products, would make the model very complex. It assumes that all the oxygen from the air inlet is combusted
to CO2 and that the pyrolysis products are completely cracked. Solid carbon, in the form of char, was assumed
to be present throughout the reduction region. The model uses the reaction kinetic parameters developed by
Wang and Kinoshita [8], and the adopted reaction scheme is as follows:

Reaction-1: C + CO2 M 2CO (1,72,600 J/mol)

Reaction-2: C + H2O M CO + H2 (1,31,400 J/mol)
Reaction-3: C + 2H2 M CH4 (+75,000 J/mol)
Reaction-4: CH4 + H2O M CO + 3H2 (2,06,400 J/mol)

This model assumes a cylindrical gasifier bed of uniform cross-sectional area A with negligible radial var-
iation in the properties of both the bed and the gas. The molar and energy balances result in the following set
of equations. The species considered here are nitrogen, carbon dioxide, carbon monoxide, methane, steam and
hydrogen.
 
dnx 1 dv
¼ Rx  nx ð1Þ
dz v dz
!
dT 1 X dP dv X
¼ P  ri DH i  v P  Rx c x T ð2Þ
dz v x nx cx i
dz dz x
P P P  P  X !
dv 1 x nx c x x Rx i r i DH i dP v v x nx cx
¼P   þ  Rx cx ð3Þ
dz x nx cx þ nR n T dz T P x

The following empirical formula is used to find the pressure gradient of a fluid flowing through a bed of
solid char particles (supplied by Dr. Donna Giltrap on request, which was used but not reported in their article
[5]).
 
dP v2
¼ 1183 qgas þ 388:19v  79:896 ð4Þ
dz qair
The reaction rates were all considered to have an Arrhenius type temperature dependence and to be pro-
portional to the difference between the actual reactant/product ratio and the corresponding equilibrium ratio.
The reaction rates for reactions 1–4 can be written as follows:
   
E1 P 2CO
r1 ¼ n CRF A1 exp  P CO2  ð5Þ
RT K1
   
E2 P CO  P H2
r2 ¼ n CRF A2 exp  P H2 O  ð6Þ
RT K2
   
E3 P CH4
r3 ¼ n CRF A3 exp  P 2H2  ð7Þ
RT K3
  !
E4 P CO  P 3H2
r4 ¼ n CRF A4 exp  P CH4  P H2 O  ð8Þ
RT K4

Table 1
Frequency factor and activation energies values [8]
Reaction i Ai (1/s) Ei (kJ/mol)
1 3.616 · 101 77.39
2 1.517 · 104 121.62
3 4.189 · 103 19.21
4 7.301 · 102 36.15
 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611 2605

The values of the frequency factor and the activation energies are taken from Wang and Kinoshita [8] and
are shown in Table 1.

3. Method of solution and simulation of model equations

Eqs. (1)–(4) give a set of nine first order differential equations in the system variables nx (where x denotes
the six different gas species considered), P, v and T that is solved by an explicit finite difference method [9,10].

3.1. Initial conditions

The initial position in this model is the top of the reduction region and the end of the oxidation zone. Gases
coming from the oxidation zone are a mixture of pyrolyzed gas, incombustible CO2 and inert N2. The exact
proportion of each of these components depends upon the rate of air flowing into the gasifier relative to the
rates of combustion, pyrolysis and the cracking reactions. Instead of calculating this directly, Giltrap et al. [5]
introduced a variable Ôpyrolysis fractionÕ (fp), which can vary from 0 (no pyrolysis products) to 1 (pyrolysis
products only). The air flowing into the gasifier is a mixture of N2 and O2. It is assumed that all the O2 from
the air inlet has been consumed by combustion reactions with the char, while N2 remains inert. The pyrolysis
products are assumed to crack into an equivalent amount of CO, CH4 and H2O for a typical biomass
(CH3.03O1.17). The initial temperature, pressure and velocity for the model would be those of the gas coming
out of the oxidation zone. The temperature of 1400 K is taken in the model, which is reported by Jayah et al.
[6] for the gasification of rubber wood in the downdraft gasifier. The velocity has been approximated from the
airflow rate and the throat dimensions reported by Jayah et al. [6]. The operating pressure does not affect the
gasification efficiency significantly. The increase of pressure is limited on two counts: (1) due to the amount of
entrained dust (the gas, being denser, exerts a higher thrust on the dust particles) and (2) by the increased cost
of manufacturing [11]. Because of these limitations and to overcome the pressure drop offered by the bed, the
operating pressure should be above atmospheric pressure. The value has to be found by approximating the
pressure drop in the bed. The value of 1.005 atm is used in the simulations and a check has been made to
assure that the pressure at the outlet remains above atmospheric.
In the reduction zone, endothermic reactions occur, and the degree of temperature drop depends upon the
extents of the reactions. The extents of the reactions depend upon the reactivity of the char, which is repre-
sented by the char reactivity factor (CRF) value. The CRF, in turn, depends upon the thermal history, the
degree of burn-off and the number of active sites on the char and its particle size. As the char moves down-
wards in the reduction zone, the degree of burn-off increases. Because of the char-gas reactions and the shrink-
ing of the particles, the char size decreases and the porosity increases, so the gases would encounter more
active sites and the extents of the reactions increase. For higher CRF values, the reduction zone temperature
drops faster and reaction completion occurs rapidly. As a result of this, the gas compositions predicted by the
model change only in the initial length of the reduction zone. The above understanding leads to the fact that
the CRF value must be increased along the length of the reduction zone. The quantitative increment in the
CRF value depends upon the type of biomass and its physical characteristics. Giltrap et al. [5] simulated many
cases, each with a constant CRF value for the entire reduction zone. In their work, they varied the CRF value
from 1 to 10,000, and the results of their simulations are compared with the final composition of the producer
gas reported in the literature. However, the temperature and composition profiles are not compared. In the
present study, the CRF values from 1 to 10,000 are varied linearly and exponentially, and their effects on
the composition and temperature profiles are observed. When an exponential increment in the CRF value
is used in the simulation, it predicted the temperature and composition profiles better when compared with
the experimental data reported by Jayah et al. [6].

3.2. Results and discussion

The proposed model of the present study is tested with the data reported for a downdraft biomass gasifier
by Jayah et al. [6]. Table 2 shows the values of the parameters used in the model simulations.
2606 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611

Table 2
Parameters used in the model simulation
Parameters Values
Bed length 24.5 cm
v initial 1.175 m/s
T initial 1400 K
Moisture content 16% (dry basis)
CRF 1, 10, 100, 1000, Exp varying, Lin varying
fp 0.3

In Table 2, the top four quantities, i.e. bed length, velocity, temperature and moisture content are taken to
be the same as those reported by Jayah et al. [6]. The other parameters, CRF and fp, are new ones that are
incorporated in the present proposed model.
Fig. 1 shows comparisons of the simulated temperature profiles using the proposed modified model for dif-
ferent CRF values, the experimental data of Jayah et al. [6] and JayahÕs model predictions. The experimental
data shows that the temperature of the reduction zone decreases continuously along the length of the reduc-
tion zone, and the trend of the profile is concave downwards. JayahÕs model predictions significantly vary from
the experimental data and are of the convex type.
For CRF = 1, the temperature drop in the reduction zone is very small and does not agree with the exper-
imental data. As the CRF values are increased from 1 to 1000, the temperature drops increase, and the devi-
ations from the experimental data also increase. For CRF = 1000, the final temperature of the producer gas
reaches close to the experimental data, but when the profile is examined (Fig. 1), it indicates that the temper-
ature is reduced in the first 2.5 cm from the top of the reduction zone and reaches the final temperature at that
point. No significant change is observed in the remaining part of the reduction zone. The reason is that the
CRF value is increased and kept constant from the top of the reduction zone. As the CRF value is very high,
the reactivity of the char would also be very high, and the reactions would get completed in only the initial
part. On increasing the CRF value, the char would be more reactive, and the initial length in which the sig-
nificant temperature drop is observed also decreases. For the linearly increasing CRF value from 1 to 10,000,
the temperature profile is also of the convex type, and the deviations from the experimental data are also sig-
nificant. For the exponential increment in the CRF value, the temperature profile matches the experimental
data in a better way than any of the other profiles. The equation used for the exponential increment is
CRF = AeBx where A = 1 and B = 0.0037. This equation gives an increment from 1 to 8000 for a length of
2500 mm. Table 3 shows the standard deviations from the experimental data for each profile. It shows that
for the exponentially varying CRF values, the standard deviation (SD) is the least, i.e. 0.091151. The standard
deviation (SD) is calculated by the following formula:

2000
CRF = 1
CRF = 10
1800 CRF = 100
CRF = 1000
Jayah et al.'s expt data
1600 CRF Exp varying
Temperature (K)

CRF Lin varying

1400 Jayah et al.'s model data

1200

1000

800

600

0.00 0.05 0.10 0.15 0.20 0.25

Length (m)

Fig. 1. Temperature profile for different CRF values.

 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611 2607

Table 3
Standard deviation of temperature profile from the experimental temperature data for different CRF
Profile CRF = 1 CRF = 10 CRF = 100 CRF = 1000 Linearly varying CRF Exponentially varying CRF JayahÕs model
SD 0.346302 0.226614 0.145176 0.224383 0.224245 0.091151 0.147922

"P #12
N
 Y cal Þ=Y exp 2
i¼1 ½ðY exp
SD ¼ ð9Þ
N 1

Fig. 2 shows the comparisons of the product gas compositions coming out of the biomass gasifier for dif-
ferent CRF values with the experimental data reported by Jayah et al. [6]. It clearly indicates that the simu-
lated data differ widely from the experimental data when the CRF values considered are 1, 10 and 100. For the
cases of CRF = 1000, the linearly varying CRF and the exponentially varying CRF, the differences are very
much less, and hence, it is difficult to judge which is the best fit.
For the carbon monoxide component, the composition values for CRF = 1000 and the exponentially vary-
ing CRF fit well with the experimental data, whereas for hydrogen, the composition value for the linearly
varying CRF matches closely with the experimental data. So, to get the quantitative variation from the exper-
imental data, the standard deviation is determined. Table 4 shows the standard deviations of the end product
composition from the experimental data reported by Jayah et al. [6].
Table 4 shows that the standard deviation is the least for the composition values when the linearly varying
CRF is considered, and the next smaller deviations occur for the cases of CRF = 1000 and the exponentially
varying CRF, respectively. Based on the standard deviation for the temperature profile, the exponentially
varying CRF value should be considered, and based on the standard deviation for the end product composi-
tion; the linearly varying CRF should be chosen. To resolve this dilemma, the composition profiles for
CRF = 1000, linearly varying CRF and exponentially varying CRF are simulated and the results are pre-
sented in Figs. 3–5, respectively.
Figs. 3 and 4 show the composition profiles along the normalized length (ratio of actual length to total
length of reduction zone) for CRF = 1000 and the linearly varying CRF (from 1 to 10,000), respectively. Both
profiles show that the fractions of carbon dioxide, nitrogen and methane are continuously decreasing
and those of carbon monoxide and hydrogen are increasing. Nitrogen is an inert gas, so its amount is not

CRF = 1
0.6 CRF = 10
CRF = 100
Composition (mole fraction)

CRF = 1000
0.5
Jayah's expt data
Exponentially varying CRF
0.4 Linearly varying CRF

0.3

0.2

0.1

0.0
N2 CO2 CO CH4 H2

Fig. 2. Comparison of product gas compositions for different CRF values.

Table 4
Standard deviation of product composition from the experimental data for different CRF values
Profile CRF = 1 CRF = 10 CRF = 100 CRF = 1000 Linearly varying CRF Exponentially varying CRF
SD 2.808762 2.133514 1.533824 0.750679 0.118892 0.765496
2608 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611

0.8
N2
CO2
0.7 CO

Composition (mole fraction)

CH4
0.6 H2

0.5

0.4

0.3

0.2

0.1

0.0

0.0 0.2 0.4 0.6 0.8 1.0

Normalized Length (-)

Fig. 3. Composition profile for CRF = 1000.

0.8
N2
0.7 CO2
CO
Composition (mole fraction)

0.6 CH4
H2
0.5

0.4

0.3

0.2

0.1

0.0

0.0 0.2 0.4 0.6 0.8 1.0

Normalized Length (-)

Fig. 4. Composition profile for linearly varying CRF.

0.8
N2
0.7 CO2
Composition (mole fraction)

CO
0.6 CH4
H2
0.5

0.4

0.3

0.2

0.1

0.0

0.0 0.2 0.4 0.6 0.8 1.0

Normalized Length (-)

Fig. 5. Composition profile for exponentially varying CRF.

changing, but the amounts of the other gases are changing and, as a result, the dry basis fraction of nitrogen
is decreasing. The fraction of CO2 is decreasing as it is getting converted to CO according to reaction 1.
The fraction of CO is increasing almost in the same fashion as CO2 decreases. The H2 fraction increases
 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611 2609

Table 5
Initial normalized length Lin for different components
Sr. no. Component Lin (–)
CRF = 1000 Linearly varying CRF
1 N2 0.1020 0.1633
2 CO2 0.0612 0.1020
3 CO 0.0735 0.0816
4 CH4 0.3877 0.4898
5 H2 0.2653 0.5918

non-linearly as the gas travels through the reduction zone and its calorific value increases. The changes in com-
position for almost all the components are only within the initial length of the reduction zone.
The normalized length from the top of the reduction zone at which 85% of the total change in composition
occurs is defined as the initial normalized length Lin. For each component, Lin is found for CRF = 1000 and
for the linearly varying CRF, which are presented in Table 5. Table 5 shows that for nitrogen, carbon dioxide
and carbon monoxide, the value of Lin is less than 0.11 for the CRF value of 1000 and less than 0.17 for the
values of the linearly varying CRF. It indicates that only 11% of the bed length is used for completion of 85%
of the reaction, which is highly unrealistic in nature. The same phenomenon is also observed for the linearly
varying CRF where just 17% of the total length of the reduction zone is used.
It confirms again that a constant CRF value should not be used in the simulation as it gives high standard
deviation from the experimental data for the temperature profile (Table 2) and also gives highly unrealistic
composition profiles. These results also rule out the possibility of using linearly varying CRF values in the
modification of the model. For methane and hydrogen, the value of Lin is larger. This is due to the fact that
the rate of reaction 4, which is a totally gas phase, is significantly less, and even by increasing the CRF values,
no sudden change in composition is observed.
Fig. 5 shows the composition profiles for the exponentially varying CRF values. The changes in composi-
tion for all the components are non-linear in nature. The fractions of nitrogen, methane and carbon dioxide
are decreasing, and those of carbon monoxide and hydrogen are increasing.
Fig. 5 also indicates that the change in composition is spread over the entire length of the reduction zone,
which is different from Figs. 2 and 3. The results of the exponentially varying CRF are quite satisfactory and
realistic.
Another important observation is made during the simulation studies that the rate of reaction 1 becomes
negative at a temperature of 933 K. This represents that below this temperature, the reaction occurs in the
backward direction. The rate of reaction depends upon the equilibrium constant and the relative amounts
of the different reacting species. The concentrations of the reacting gaseous species depend upon the total

CRF Lin varying

CRF Exp varying
Rate of reaction (1) (mol/m3sec)

CRF = 1000
10

1100 1075 1050 1025 1000 975 950 925 900 875
Temperature (K)

Fig. 6. Rate of reaction 1 versus temperature (K).

2610 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611

pressure in the system and the equilibrium constant values. As the total pressure of the system remains con-
stant and the equilibrium constants are functionally dependent on temperature alone, the rate of reaction
effectively becomes a function of the bed temperature. As shown in Fig. 3, CO decreases and CO2 increases
at an Ln of 0.35. The same phenomenon is also observed in Figs. 4 and 5 at Ln values of 0.29 and 0.92, respec-
tively. That is because the temperature of the reduction zone falls below 933 K at that point, as can be viewed
in Fig. 1 also.
Fig. 6 shows the variation of the rate of reaction 1 with temperature for CRF = 1000, exponentially varying
CRF and linearly varying CRF. It clearly indicates that the rate of reaction becomes negative as the temper-
ature falls below 933 K. It also affects the calorific value of the gas as the CO decreases although the H2
increases. H2 is increasing continuously because the rates of reaction 2 and reaction 4 are positive for this
range of temperature. CH4 decreases continuously as the rate of reaction 3 is negative from the start of the
reduction zone because the CH4 concentration is high in the gas coming from the pyrolysis zone.

4. Conclusions

The steady state model discussed here predicts the composition and temperature profiles across the length
of the reduction zone. The char reactivity factor is the key parameter in modeling the downdraft gasifier, and it
directly represents the reactivity of the char in the reduction zone. The earlier model reported by Giltrap et al.
[5] is modified in the present study by incorporating the exponentially varying CRF. Simulations are per-
formed for varying CRF from 1 to 10,000 and linearly and exponentially as well. Simulations are also per-
formed for different values of CRF (1, 10, 100 and 1000), which were held constant throughout the
reduction zone. Based on the results obtained and the discussions in the earlier sections, the following conclu-
sions are drawn:

• The standard deviation for the temperature profile from the experimental data reported by Jayah et al. [6] is
the least when the exponential increment in CRF is considered.
• The composition profiles developed by considering constant CRF values show that the changes in compo-
sition occur only in the initial length of the reduction zone, which is unrealistic, e.g. for CRF = 1000, the
value of Lin is only 0.11.
• The value of the char reactivity factor (CRF), which represents the reactivity of the char, must be varied
along the reduction zone of the down draft biomass gasifier.
• The standard deviation for the temperature profile from the experimental data reported by Jayah et al. [6] is
high for the case when the linear variation of CRF is considered.
• The composition profiles developed by considering a linearly varying CRF value show that the changes in
composition occur only in the initial length of the reduction zone, which is speculative in nature, i.e. for a
linearly varying CRF, the value of Lin is only 0.17.
• For the case of the exponentially varying CRF, the standard deviation for the temperature profile is the
least, i.e. 0.091151. The standard deviation for the end product composition from JayahÕs experimental data
is 0.765496. In addition, the composition profile is changing over the entire length of the bed of the reduc-
tion zone.
• The simulated results obtained from the model developed in the present study are in very good agreement
with the experimental data of Jayah et al. [6], better in comparison with the mathematical model of Jayah
et al. [6], when the exponentially varying CRF value is considered.
• The rate of reaction 1 becomes negative, and the reaction occurs in the backward direction below the
temperature of 933 K when the pressure is slightly above atmospheric pressure.

References

[1] Babu BV, Chaurasia AS. Energy Convers Manage 2004;45:53.

[2] Babu BV, Chaurasia AS. Energy Convers Manage 2003;44:2251.
[3] Babu BV, Chaurasia AS. Energy Convers Manage 2003;44:2135.
 B.V. Babu, P.N. Sheth / Energy Conversion and Management 47 (2006) 2602–2611 2611

[4] Blasi CD. Chem Eng Sci 2000;25:2931.

[5] Giltrap DL, McKibbin R, Barnes GRG. Solar Energy 2003;74:85.
[6] Jayah TH, Aye L, Fuller RJ, Stewart DF. Biomass Bioenergy 2003;25:459.
[7] Milligan JB. PhD thesis, Aston University, Birmingham, UK, 1994.
[8] Wang Y, Kinoshita CM. Solar Energy 1993;51:19.
[9] Ghoshdastidar PS. Computer simulation of flow and heat transfer. New Delhi: Tata McGraw Hill Publishing Company Limited;
1998.
[10] Babu BV. Process plant simulation. India: Oxford University Press; 2004.
[11] Mathieu P, Dubuisson R. Energy Convers Manage 2002;43:1291.