ARTICLE

pubs.acs.org/IECR

Development of Efficient Designs of Cooking Systems. II.

Computational Fluid Dynamics and Optimization
Jyeshtharaj B. Joshi,*,†,|| Aniruddha B. Pandit,*,† Shirish B. Patel,*,†,§ Rekha S. Singhal,‡ Govind.K. Bhide,§
Kishore V. Mariwala,§ Bhagwat A. Devidayal,§ Sanjay P. Danao,†,^ Arijit A. Ganguli,† Ajitkumar S. Gudekar,†
Prakash V. Chavan,† and Yogesh H. Shinde†
†
Department of Chemical Engineering, Institute of Chemical Technology, Matunga, Mumbai
400019, India
‡
Department of Food Engineering and Technology, Institute of Chemical Technology, Matunga, Mumbai
400019, India
§
Land Research Institute, Second Floor, United India Bldg., P.M. Road, Mumbai
400001, India
)

Homi Bhabha National Institute, Anushaktinagar, Mumbai
400094, India

ABSTRACT: Sections 2
6 of Part I were devoted to the analysis of heat transfer characteristics of cookers. In all the experiments,
only water was employed as a working medium. Now, we extend such an analysis to the actual cooking process in order to arrive at an
improved cooking device. The major strategies for the optimization of energy utilization is to design appropriate insulation that has
been obtained by two cover vessels. In order to select an air gap, the flow and temperature patterns in the air gap have been
extensively analyzed using computational fluid dynamics (CFD). The flow pattern and heat transfer in cooking pots have also been
analyzed by CFD. This has enabled us to design suitable internals for minimizing the stratification of temperature. The
understanding of fluid mechanics has also given basis for selection of heat flux, gap between burner tip and cooker bottom, and
temperature of flue gases leaving the cooker. Chemical engineering principles have been used for modeling and optimization.
Kinetics have been obtained in batch cookers. The knowledge of kinetics, thermal mixing, axial mixing, and optimum selection of
insulation have been employed for the development of continuous cookers. The continuous mode of operation also helps in saving
of energy. Systematic data have been collected for the design and scale up of continuous cookers.

1. INTRODUCTION convection in cooling pots were comprehensively analyzed by

The importance of the development of cooking devices was computational fluid dynamics (CFD). In addition to the LPG
discussed in Part I. In Part II, we have optimized the cooking burning rate, three additional parameters were optimized: (i) the
system developed in Part I. The optimization exercise consisted gap between the burner tip and the bottom of base vessel, (ii) size
of selecting the appropriate rate of heat supply for burners using of the flame (projection and spread of flame on the bottom of the
LPG (liquefied petroleum gas) as the fuel. The rate essentially base vessel) with respect to the base area, and (iii) temperature of
depends upon the heat uptake rate by the cooking pots, which is flue gases leaving the base vessel. Further savings in energy have
supplied by condensing steam. Because the condensation heat been implemented by developing a continuous cooker. Thus, this
transfer coefficient is usually very high, it was important to paper considers the ways and means of minimizing the con-
understand the heat uptake rate by the contents of the pot. sumption of LPG in cooking. It also considers the optimum use
The pots contain a mixture of water and rice (or vegetables, of solid fuel stoves and solar energy. The analysis of cookers has
lentils, etc.), and the heat transfer inside the vessel occurs by been restricted to foods that can be cooked by boiling or steaming.
natural convection. It is known that the temperature needed for Apart from saving fuel (energy) and thereby reducing pollu-
the cooking of rice is about 74 C or greater and that for lentils is tion, the cooker produces food with better flavor and possibly
94 C or greater. Therefore, while investigating the natural con- higher nutrition values than other cooking methods. It also
vection, it is important to ensure the spatial temperature field in reduces the time a cook has to spend in the kitchen to monitoring
such a way that the temperature at all locations is at least above the cooking. An attempt has also been made to improve the
the temperature levels mentioned. This consideration sets a limit performance of gas burners (liquefied natural gas or LPG), solid
on the permissible dimension (diameter in particular) of the pots fuel stoves, and solar energy-based heating devices. Recommen-
and the aspect ratio. In order to increase the permissible dia- dations have been made for the selection and design of these
meter, it is important to reduce the extent of stratifications that energy providers. The areas of future research work have also
can be made possible by providing suitable internals. been highlighted and have been recommended.
The air gap provides insulation. In the range of geometries and
temperature differences under consideration, the values of the
Special Issue: Nigam Issue
Rayleigh number may exceed the transition value and cellular
convection can occur. Therefore, in order to optimize the air gap, Received: November 9, 2011
it is important to analyze the fluid mechanics for different cooker Accepted: December 19, 2011
sizes. The optimization of the air gap and the analysis of natural Published: December 19, 2011

r 2011 American Chemical Society 1897 dx.doi.org/10.1021/ie2025745 | Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

2.2. Optimization of Air Gap between the Two Metal

Covers. Minimizing thermal energy loss using an air gap as
insulation has been of interest to researchers during the last many
decades. Ganguli et al.3,4 have reviewed previous work on the
mathematical models from Batchelor6 and Elder7, experimental
work from Yin et al.8, Elsherbiny et al.9, and Wakitani10, and
numerical studies from Newell and Schmidt11, Korpela et al.12,
Lee and Korpela13, Le Quere14, Wakitani15, Wakitani16, Zhao
et al.17, and Lartigue et al.18, which have been performed on many
different situations to understand the complexities of flows inside
tall slender rectangular slots. The literature review suggests
that in recent years a considerable amount of research has been
carried out to understand the flow patterns generated in these
gaps, over a wide range of Rayleigh numbers (Ra) and aspect
ratios (AR), as these have been proven to be important parameters.
The conventional open pan cooking method has an efficiency
of about 15
25%, while pressure cookers have net thermal
efficiencies in the range of 35
40%, for instance, a pressure
cooker (outer vessel aluminum) with one or more steam whistles
during cooking. When one looks at conventional cooking
methods, it is easily observed that the lower efficiencies are due
to reasons like higher (or nonoptimum) fuel burning rates, use of
excess water (especially in the case of rice cooking), and
substantial and continuous heat losses to the surroundings.
Pressure cookers are available up to a maximum capacity of 30 L
but are rarely used beyond a size of 7
10 L because of various
safety issues. In this context, the cooker developed in Sections
4
6 has been shown to possess an overall thermal efficiency in
the range of 50
70%. It works at practically atmospheric pre-
ssure and is available in various capacities, from domestic (4 L) to
Figure 1. Schematic of cooker after phase iii development: (1) base, community cooking (120 L, about 350 people). For a typical
(2) bubble plate, (3) vessels, (4) inner cover, (5) outer cover, (6) gap heat transfer application across a solid wall, the resistance offered
between inner and outer cover, (7) burner, (8) distance between burner
by the wall and by the media on both sides determines the
tip and base.
maximum heat flux through the wall for achieving and maintain-
ing the required cooking temperature. In the case of cooking,
2. OPTIMIZATION USING COMPUTATIONAL FLUID heat transfer from the burning fuel, i.e., from the flame and flue
DYNAMICS gases to the cooking vessel base and effectively to the material to
be cooked is determined by the area of the cooking vessel base,
2.1. Introduction. A schematic diagram of a cooker is shown the temperature difference (ΔT), and the overall heat transfer
in Figure 1. It essentially consists of a base vessel, support stand, coefficient (U). For a given setup and application (cooking), the
set of cooking pots, metal enclosure for the pots, a second metal heat transfer area is constant. Because the flame temperature is
cover to create an air gap for providing insulation, and a gas above 1000 C, the value of ΔT remains practically constant
burner. The base vessel contains water. The pots cook rice, vege- during the entire period of preheating and partial cooking
tables, lentils, etc. The burner provides heat to the base vessel and (30
100 C) and subsequent cooking. In the case of cooking,
boils water. The steam generated, in turn, provides heat to the U is largely dependent on the inside heat transfer coefficient (hi).
cooking pots. The optimization exercise includes the selection of It is obvious that in the present case (as well as in the pressure
the appropriate rate of heat supply by the burner. The rate cooker or any other closed cooking device), stirring is not
essentially depends upon the heat uptake rate by the cooking possible (though intermittent stirring is followed in open pan
pots, which is supplied by condensing steam. Because the con- cooking). This puts an upper limit on the inside heat transfer
densation heat transfer coefficient is usually very high, it was coefficient (hi) and hence on the maximum heat uptake by the
important to understand the heat uptake rate by the pot contents material to be cooked. By selecting an optimum ratio of cooker
by natural convection. Further, the air gap (Figure 1) provides base to flame diameter (about 3, as determined in Section 3) and
insulation. the fuel burning rate, an attempt has been made to supply heat at
In order to quantify the natural convection in the air gap and in a rate that more or less corresponds to the heat uptake rate
the cooking pots, the technique of computational fluid dynamics (decided by the internal heat transfer coefficient). Now, it is
(CFD) was employed. The air gap involves single-phase flow, important that the heat losses to the surroundings from the side
whereas the cooking pots involve two-phase solid
liquid flows. walls are also minimized. In this context, the insulation provided
The solid phase generally forms a fixed bed where the particle size by the air gap between the inner and outer cover of the cooker
and the voidage depend upon the extent of cooking. All of the needs to be optimized.
guidelines needed for the formulation of governing equations, Let us consider an enclosed volume Ven surrounded by a
boundary conditions, discretization, and solution procedure are narrow air gap that acts as insulation. Details of the air gap can be
available in the published literature3
45 . found in Figure 1. Transient temperature variation occurs inside

1898 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

Table 1. Governing Equations Used for Natural Convection Problem

continuity ∂F
þ ∇ðFÆuk æÞ ¼ 0
∂t


momentum ∂ðFÆuk æÞ 2
þ ∇  ðFÆuk æÆuk æÞ ¼
∇Æpæ þ ∇  τk þ F τk ¼ μeff ∇Æuk æ þ ð∇Æuk æÞT
μeff ∇Æuk æI
∂t 3

energy ∂ðFÆTæÞ
þ ∇  ðFÆuk æÆTæÞ ¼ ∇  ðαeff ÆTæÞ
∂t

 
turbulent kinetic energy ∂ðFkÞ μ
þ Æuk æ∇  ððFÆkæÞÞ ¼ ∇  μ þ t ∇  k þ Gk þ Gb
Yk
∂t σk

Turbulent kinetic energy (TKE) Gk = νt|S|2 where |S| = (2Sij Sij )1/2 and |S| = 1/2((∂Æuzæ)/(∂r) + (∂Æuræ)/(∂z) + (∂Æuθæ)/(∂θ))
Generation of turbulence due to buoyancy Gb =
βg(νt)/(σt)(∂ÆTæ)/(∂z)
8
> 1, χk e 0
>
<" #
Dissipation of this TKE, Yk = Fβ∞*fβ*kω where β∞* = 0.09, fβ ¼ 1 þ 680χk 2 , χk > 0 and χk = (1)/(ω3)(∂k)/(∂z)(∂ω)/(∂z)
>
>
: 1 þ 400χk 2
energy dissipation rate (∂(Fω))/(∂t) + Æukæ3  ÆFωæ = 3[(μ + (μt)/(σω))3ω] + Gω
Yk Production of ω, Gω = (α∞)/(νt)Gk
equation where α∞,1and α∞,2 are constants.
α∞,1 = 1, α∞,2 = 0.52 and Yω = Fβrω2

the enclosed volume as a result of either heat supply or heat loss, (5) Length (in the third dimension) of the enclosure is suffi-
and a methodology needs to be developed using existing knowl- ciently large so that a two-dimensional motion can be
edge of the extent of insulation achieved using air gaps. Here lies assumed and cylindrical curvature may be neglected.
the main motivation for CFD simulation of the two different capa- (6) Variation in properties of air as a function of temperature
cities (120 and 700 L) that have been considered. We now des- are assumed to follow the following relations (Incropera
cribe the optimization procedure briefly. and Dewitt21)
2.2.1. Computational Fluid Dynamic (CFD) Simulation. The For air
following section describes the assumptions, governing equa-
tions, boundary conditions, and the method of solution for all of μ ¼ 5  10
8 T þ 2  10
6 ð3Þ
the CFD simulations carried out by Ganguli et al.3,4 Additional
details pertaining to CFD simulation can be obtained from Ekambara kt ¼ 8  10
5 T þ 0:0016 ð4Þ
and Joshi19 and Thakre and Joshi20
2.2.1.1. Assumptions. For the present simulations, the left
wall has been maintained at a constant temperature TC (equiva- Cp ¼ 0:07T þ 985:5 ð5Þ
lent to the atmospheric temperature of the outer cover) and the where, temperature T is in Kelvin (K)
right wall has been maintained at a constant temperature TH 2.2.1.2. Governing Equations and Method of Solution. The
(equivalent to steam temperature of the inner cover). Following transient governing equations of continuity, momentum, and
are the assumptions made in the present work: energy have been described in Table 1. The term τk in Table 1 is
(1) The fluid is assumed to be incompressible and New- transformed as τk = μ(3Æukæ + (3Æukæ)T). The ̅ boundary
tonian. Because of the small magnitude of variation in ̅
conditions have been described in Table 2. The laminar model
pressure, the density variation is also very small, and equations described in Table 1 have been solved using commer-
hence, the air can be approximated as incompressible. cial flow simulation software FLUENT (version 6.2).22 Transient
Hence, the flows can be accounted for by only buoy- simulations have been carried out with a time step of 0.0001 s.
ancy variation. For each time step, convergence criteria for the sum of normal-
(2) The flow is assumed to be two dimensional. ized residues have been set to 1  10
4 for continuity equations,
(3) All the working fluids are operated at temperature differ- 1  10
4 for momentum, and 1  10
7 for energy equations.
ences within the Boussinesq approximation. The well- Convergence has been ensured at every time step. The under-
known Boussinesq equation has been employed, and the relaxation parameters were set to 0.3 for pressure, 1 each for
density in the buoyancy term is assumed to vary with the density, energy, and body forces, and 0.7 for momentum equa-
temperature according to the following relation tions. In the cited study in ref 4, a segregated solver with implicit
F ¼ F0 ð1
βðT
T0 ÞÞ ð1Þ time discretization has been employed for obtaining the solution
of momentum equations. The momentum equations were dis-
cretized using the second-order upwind scheme (SOUS), and for
"
# the pressure equation, the PRESTO scheme was used. The
1 ∂F
where, β ¼ ð2Þ SOUS scheme in the commercial software FLUENT 6.2 pre-
F0 ∂T vents numerical diffusion while not dampening the disturbances
that cause instability. Thus, it ensures spatial accuracy.
2.2.1.3. Grid Sensitivity. The grid test has been performed for
(4) Heat transfer by radiation has been neglected, as relatively aspect ratio (AR = HE/L) of AR = 28 and AR = 174 to ensure grid
low temperatures and temperature differences have been independent results for the entire AR range. Two Ra numbers
considered. were chosen for AR = 174 (Ra = 907 and Ra = 4080) and for
1899 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Table 2. Boundary Conditions Used for the 2D Simulation of

Air Gap
Boundary Conditions

Initial velocity condition: u(x,y,0) = v(x,y,0) = 0 for 0 e x e L 0 e y e H

Initial Temperature condition: T(x,y,0) = 0 for 0 e x e L 0 e y e H
Initial x-direction boundary conditions: u(x,0,t) = v(x,0,t) = 0 for 0 e x e L
and t g 0
Final x-direction boundary conditions: u(x,H,t) = v(x,H,t) = 0 for 0 e x e L
and t g 0
Final y-direction boundary conditions: u(0,y,t) = v(0,y,t) = 0 for 0 e y e H
and t g 0
Final y-direction boundary conditions: u(L,y,t) = v(L,y,t) = 0 for 0 e y e H
and t g 0
TC = 300, TH = 370 for 0 e y e H and t = 0
TC e T(L,y,t) e TH for 0 e y e H and t g 0
(∂T(x,0,t))/(∂y) = (∂T(x,H,t))/(∂y) = 0 for 0 e x e L and t g 0
Figure 2. Variation in the insulation heat transfer coefficient (hins) with
gap width (5 mm e L e 25 mm; Δt = 70 K): (1) HE = 100 mm,
AR = 28 (Ra = 2.61  105 and Ra = 5.81  104). Three different (2) HE = 300 mm, (3) HE = 500 mm, (4) HE = 800 mm, (5) HE = 1000 mm,
grid distributions were chosen for each AR. For AR = 174, the grids (6) heat losses by only conduction.
chosen were (A) uniform grid 5  870, (B) non-uniform grid
10  870 with 75% of the nodes near the walls with the first grid
point at x+ = 0.1, and (C) non-uniform grid 10  870 with 50% heat transfer coefficient (hins) with gap width (5 mm e L e
of the nodes near the walls with the first grid point at x+ = 0.1. 25 mm; ΔT = 70 K). Line 6 is drawn to show the quantum of
Similarly for AR = 28, the same grid distribution was chosen; the heat loss by conduction only with a stagnant air gap. This line
numbers being (A) 20  560, (B) 40  560, and (C) 40  560. In (locus of values of all hins values representing pure conduction) is
all the cases, Case C was found to give the best results, both quanti- independent of the height of the air gap because no air currents
tatively and qualitatively. Hence, the grid distribution of Case C has are assumed in pure conduction. The remaining lines in Figure 2
been selected for all the simulations in the present work. represent the variation in hins with gap width for different heights.
2.2.2. Results and Discussion. 2.2.2.1. Preamble. This sec- It should be noted that the hins values remain constant after a
tion presents a discussion on the variation in the heat transfer certain gap width for any specified height over the entire range
coefficient (HTC) of the insulating air gap (hins) for a constant (500 mm e HE e 1000 mm) considered in this work, and the
ΔT (ΔT = 70 K) by varying the gap width (5 mm e L e 25 mm) observed constant value of hins is higher for lower heights. It is
and height (H = 560 mm and HE = 870 mm) of the selected expected that as gap width increases, insulation capability in-
geometry. The term hins is defined as the time average value of the creases (provided the air is stagnant), and hence, the hins should
HTCs from time t = 0 to the time when steady state (the con- decrease. However, it has been observed that after a certain width
vective circulation cell position and number after their formation there is no decrease in hins, and the hins becomes independent of
remains constant and do not affect the HTC, suggesting that the the gap width and then remains practically constant. This is
instability has ceased) or quasi-periodic state is reached (cells explained as follows. As the gap width increases [Figure 3; HE =
keep on moving and the number of cells keeps on changing, 560 mm, 870 mm; L = 15 mm (refer lines 2 and 3)], the local
affecting the HTC, mainly resulting in an increase in magnitude HTCs are enhanced due to flow fluctuations arising out of
of HTC). Simulations have been performed for the above range natural convection (formation of the multiple circulation cells
(ΔT and AR) and validated with the experimental results. in the central part of the air gap).4 Hence, when the height is low,
A procedure has then been suggested for estimation of the the hins is high. Further, as height increases, the magnitude of
optimum gap width over typical heating and retention periods HTCs at the top and bottom increases, but the hins remains
(up to 4 h). Using the same procedure, the temperature attained constant. This can also be explained in terms of flow patterns as
(resultant of supply and losses) after 30 min of heating period for represented in Figure 3, which shows flow patterns for different
the optimum gap has been predicted, as this is a typical time gap widths (5 mm e L e 20 mm; ΔT = 70 K; HE = 560 mm;
required for steam cooking of rice and lentils. 870 mm) Panel (A i, ii) of Figure 4 (L = 5 mm, 8 mm) and panel
2.2.2.2. Effect of Gap Width and Cooker Height on the Heat (B vii, viii, ix) of Figure 4 (L = 5 mm, 8 mm, 10 mm) show
Transfer Coefficient. Understanding the flow patterns of air unicellular patterns which represents that conduction is the main
inside the air gap is extremely important for the estimation of the mode of heat transfer, while panel (A iii, iv, v, vi) of Figure 4 and
net heat flux. Let us consider the maximum temperature at the panel (B vii, viii, ix, x, xi, xii) of Figure 4 represent multicellular
end of Ven to be 373 K (saturated condensing steam at atmo- patterns and a significant contribution of natural thermal con-
spheric pressure), while the minimum temperature at the other vection to the overall heat transfer. For a height HE = 500 mm,
end is 303 K (atmospheric room temperature). Thus, the maxi- multicellular cells start forming at L = 10 mm and above, while for
mum ΔT between inside and outside the ambient temperature is HE = 1000 mm, they start at L = 12.5 mm and above. However,
70 K. The flow patterns at this ΔT have been investigated for Figure 2 shows that minimum hins occurs at L = 12.5 mm (line 4,
different gap widths (5 mm e L e 25 mm) and height (HE = Figure 2, band of 10
12.5 mm). This is due to the contribution
560 mm, 870 mm). Figure 2 shows the variation in the insulation of heat transfer coefficients at the top and bottom of the air-gap
1900 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

for other gap widths, giving hins for L = 12.5 mm to be the

lowest.
2.3. Heat Transfer from Steam to Water. 2.3.1. Preamble.
A schematic diagram of pots in the cooker is shown in Figure 1.
The contents of the pots are heated by condensation of steam on
the outer surface of the cylinder. Steam is also condensed on the
bottom surface of the pot if suitable spacing is maintained
between pots. The transfer of heat from water results in natural
convection in the pool of water inside the pot. When the water
contains rice/lentils/vegetables, these immersed bodies provide
resistance to natural convection. It was thought desirable to
undertake CFD simulations of the heat transfer process so that
the optimum selection for the following are possible. Initially,
transfer of heat from condensing steam to water (single phase)
has been considered.4 The modifications in the flow pattern in
the presence of immersed particles can be estimated using the
procedure of Reddy and Joshi5 (Section 2.4). The following
stepwise procedure was employed:
(a) The estimation of the heat transfer rate for all the pots
gives the qualitative number for the rate of heat supply by
the burning of LPG. This exercise needs to be carried out
at various stages of cooking.
(b) The possibility of stratification is known to occur in the
process of heat transfer. Further, the food does not get
cooked in the low-temperature stratified region. This
feature puts a limit on the pot size and the aspect ratio.
(c) As mentioned previously, natural convection gets dampened
by the presence of rise/lentils/vegetables in the pool of
water. Therefore, it is desirable to understand the stratifica-
tion and heat transfer in the presence of these materials.
(d) The stratification can be lessened by providing suitable
internals in the pot. The modified geometries permit pre-
ssure and flow redistribution.
It was thought desirable to address all of the above issues using
CFD simulations. For this purpose, three-dimensional simula-
tions were employed against two-dimensional simulations of the
air gap in Figure 1. One-quarter of the pot was considered for the
modeling and is shown in Figure 4 together with the coordinate
system and grids.
2.3.2. Previous Work. Ganguli et al.4 and Gandhi et al.23 have
comprehensively reviewed the literature. The following is a brief
summary:
Aszodi et al.24 measured the temperature field in their system
with thermocouples and simulated the same using commercial
software CFX4, using the wall-boiling model included therein. It
was assumed that neither temperature nor gas fraction influences
the density of the liquid and that the momentum of the gas phase
was negligible. These assumptions are rather restrictive.20
Kang25 has experimentally investigated the thermal mixing in a
rectangular water tank being heated by steam. He has reported
the effect of tube orientation (i.e., horizontal or vertical) on natu-
ral convection. He has concluded that the time required for
achieving the saturation temperature of water is independent of
the tube orientation.
Krepper et al.26 have investigated natural convection in a large
pool using temperature measurements and CFD simulations
Figure 3. Variation of heat transfer coefficient with vertical distance for Δt = using commercial code CFX-4. They have used a laminar flow
70 K and L = 15 mm: (1) HE = 300 m, (2) HE = 560 mm, (3) HE = 870 mm. model for the system to investigate temperature oscillations due
to presence of a plume near the tube surface.
as shown in Figure 2. The fluid mixing at the top and bottom part Calcagni et al.27 have carried out experimental (thermocouples
of the air gap is lower in the case of a gap of 12.5 mm than that for temperature measurements and holographic interferometer
1901 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

for flow measurements) and numerical analysis of air in a square simulations were performed using the commercial code FLU-
enclosure approximating a two-dimensional vertical slot. The air ENT 6 assuming two-dimensional geometry. They have con-
was partially heated by a centrally located heating element from cluded that in a vertical slot, Rayleigh number (Ra) less than 104,
the bottom, with the vertical walls acting as cooling elements. heat transfer takes place mainly by conduction, while for Ra equal
The other surfaces were considered adiabatic. The numerical to 105, heat transfer takes place by convection.

Figure 4. Continued

1902 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

Figure 4. Flow patterns for various gap widths for Δt = 70 K. (A) HE = 560 mm, (i) L = 5 mm, (ii) L= 8 mm, (iii) L = 10 mm, (iv) L = 12.5 mm, (v) L = 15 mm,
(vi) L = 20 mm. (B) HE = 870 mm, (vii) L = 5 mm, (viii) L = 8 mm, (xi) L = 10 mm, (x) L = 12.5 mm, (xi) L = 15 mm, (xii) L = 20 mm.

Sharma et al.28 have investigated turbulent natural convection different boundary conditions (constant temperature, constant
in a rectangular enclosure heated from the bottom using numer- heat flux) on natural convection by performing steady state as
ical simulations of a two-dimensional geometry using commer- well as transient simulations. They have reported two counter-
cial codes (FLUENT, PHONIX, STAR-CD) as well as an in- rotating cells for the entire range of Ra numbers and for various
house code. They have considered air as the working fluid with values of heated width and have found that for isothermal
Ra ranging from 108 to 1012. They have investigated the effect of heating, the Nusselt number (Nu) increases with an increase in

1903 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

the heated width. The opposite was found to be true for isoflux
heating. They have concluded that the Nusselt number is a
weaker function of heated width for isoflux heating compared to
isothermal heating. They also concluded that the intensity of
buoyancy-induced flow in the case of isoflux heating is lower
compared to that from isothermal heating at identical values of
Ra. They have proposed correlations for the Nusselt number
(Nu) for both boundary conditions.
Ganguli et al.4 have performed CFD simulations and experi-
mental measurements for a centrally heated tube placed in a
cylindrical enclosure. They have measured the flow patterns by
PIV measurements and hot film anemometry (HFA). Tempera-
ture patterns were obtained using thermocouples. They have
performed CFD simulations for a two-dimensional axi-symmetric
geometry using commercial software FLUENT 6.3 with a user-
defined function (UDF) for incorporation of their phase change
(boiling) model. The value of Ra was varied in the range from
4.98  1012 to 3.764  1013. They have considered water as a
working fluid, and density variation has been modeled using the
Boussinesq approximation.
2.3.3. Numerical Procedure. 2.3.3.1. Governing Equations.
In order to model unsteady natural convection with possible
stratification, the governing equations (continuity, momentum,
and energy) with the appropriate Reynolds stress closure need be
solved. In the present work, a shear stress transport (SST) k -ω
turbulence model has been used. All the equations for single-
phase simulations (only natural convection) are listed in Table 1. Figure 5. Geometry used for simulation along with coordinate system
and grid size: (A) geometry used, (B) top view, (C) side view,
For solving the equations, open source CFD software Open- (D) geometry showing meshing details.
FOAM-1.6 was used. In the present case, the following assump-
tions have been made to model the heat transfer.
2.3.3.2. Model Assumptions 2.3.3.4. Boundary Conditions. The bottom wall of the tank
(1) Condensation occurring outside the wall is film conden- was considered with no slip boundary condition. Further, two
sation, and hence, the wall is at a constant temperature. cases of thermal conditions were considered for the bottom wall:
(2) Outside heat transfer coefficient varies along the length or (i) adiabatic and (ii) heat transfer from condensing steam. Free
height of the wall. slip and adiabatic boundary conditions were employed for the
(3) Fluid is Newtonian in rheological behavior and is top water surface. A constant temperature boundary condition
incompressible. has been given to the wall together with no slip. These details are
(4) The Boussinesq approximation is valid, i.e., density listed in Table 2.
differences are only important in producing buoyancy. 2.3.3.5. Method of Solution. All of the computational work
(5) There are constant fluid properties (such as viscosity), has been carried out using the open source software Open-
except in the formulation of the buoyancy term. FOAM- 1.6 with the buoyant BoussinesqPisoFOAM solver. The
2.3.3.3. Grid Independence. In the present work, a cylindrical second order implicit scheme has been used for time discretiza-
pot has been considered, and only a quarter of the geometry tion. In the case of the k -ω model, the QUICK discretization
(Figure 5) has been used for grid generation. Therefore, the total scheme was used for turbulence parameters. For the final sweep
computational time is reduced. Grid independence was investi- over each segment, upwinding has been performed using QUICK
gated by considering three different grid cases: (a) 450,000; with a second-order pressure scheme. The QUICK formulation
(b) 750,000, and (c) 850,000. A non-uniform hexahedral grid was has a third-order accuracy that helps to mitigate the unfavorable
used with a finer grid near the tube wall where gradients are more effect of artificial diffusion that can occur when using low-
important compared to zones away from the tube wall. Approxi- order upwinding schemes. All the solutions were considered
mately 70% of the nodes were located in the boundary layer, and to be fully converged when the sum of residuals was below
30% were located in the core region where (according to the 10
5. All the computations have been performed on an SGI
experimental results) the fluid is essentially quiescent and the cluster with quad processor nodes with a 2 GB RAM, 2.4 GHz
temperature varies linearly along the length of the pot wall. It was processor speed.
ensured that at least one should be in the viscous sublayer (y+ < 5) 2.3.4. Heating of Single Pot. For the initial transient period
and several nodes in the buffer and turbulent zones. For these (t = 30s), we see a strong temperature gradient near the vicinity
three cases, we have compared the results of mean quantities at of the wall and practically no gradient in the middle of the
different positions. All the chosen grids predict mean flow enclosure. Thus, the water near the wall gets heated and rises.
patterns effectively, but for further simulations a grid size of Water from the neighboring areas rushes to take the place of
750,000 cells has been used. Minor differences have been the risen water and motion of water starts in the pot. Figure 6
observed between 750,000 and 850,000. However, these were shows a schematic representation of the velocity variation with
found to be within the (5% average error. dimensionless lateral distance. We observe that there is a velocity
1904 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 6. Schematic representation of the axial velocity variation with

dimensionless lateral distance: (1) t = 30 s, (2) t = 100 s, (3) t = 200 s,
(4) t = 500 s.

Figure 8. Streamline plots at different times: (A) t = 50 s, (B) t = 150 s,

(C) t = 250 s, (D) t = 350 s.

maximum stratification. Mathematically, stratification at any time

can be expressed as
Æð∂T=∂yÞæ
S¼ ð6Þ
ð∂T=∂yÞmax
where

 "
#
ΔT 1 j
1 Tjþ1
Tj
Figure 7. Temperature contours at different times: (A) t = 50 s,
Δy
¼
j
1 j¼1 ∑Δy
ð7Þ
(B) t = 150 s, (C) t = 250 s, (D) t = 350 s.
 
ΔT Tmax
Tmin
peak near the wall, and the velocity decreases as we move away ¼ ð8Þ
from the wall. In the middle region, velocity becomes zero, and Δy max ðj
1ÞΔy
values becomes negative as we move toward the center of the pot. In simulations and experimental measurements, temperatures
Negative velocity implies the downward flow of water in the have been monitored at different axial locations, and the values of
central region. This shows that the material balance in the tank is (ΔT/Δy) and (ΔT/Δy)max have been calculated by eqs 7 and 8.
maintained at every point in time. Eventually, the boundary layer Runs have been carried out for a time of 1200 s. In both cases, the
flow adjacent to the wall begins. As time passes, hot water starts data have been analyzed using the same procedure. It was
accumulating at the top water surface because it is lighter than the observed that at any particular time instant the stratification
water in the middle region of the tank. number was found to increase with an increase in the Ra number.
Figures 7 and 8 show schematic representations of snapshots Thus, systems with a higher Ra number get stratified faster com-
of the temperature isotherms and flow pattern. It is observed that pared to those at a lower Ra number. Indirectly, sizes of all dead
the extent of stratification increases as time progresses. In zones increase with an increase in pot size.
particular, the size of the dead zone increases with time. 2.4. Heat Transfer from Steam to Water
Rice (Solid
2.3.5. Extent of Stratification. The stratification parameter has Particles) Mixture at Various Stages of Cooking. 2.4.1. Pre-
been proposed and defined by Han et al.29 It is defined as the amble and Previous Work. In Section 2.3, we analyzed the
ratio of the mean temperature gradient at any time to the maxi- problem of heat transfer from steam to water. In this case,
mum mean temperature gradient achieved at the end of the run. the steam condenses on the outside wall of the pot, and the
The stratification number varies from zero to one, with zero re- water in the pot is heated from room temperature to the de-
presenting no stratification (complete mixing) and one representing sired temperature.
1905 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 10. Temperature variation at two different axial locations:

(1) near bottom, (2) near top.

Figure 9. Streamline plots at different times in the presence of solid

particles: (A) t = 50 s, (B) t = 150 s, (C) t = 250 s, (D) t = 350 s.

When the material to be cooked is added to water in the

required proportions (for instance rice to water, 1:2, and lentils
to water, 1:3), the solid phase forms a fixed bed at the bottom.
The solid phase can be rice, lentils, potatoes, or vegetables. In
other words, at the beginning of cooking, the solid phase can have
different particle sizes, particle size distribution, and voidage.
During cooking, all the material swells, and the voidage (volume
fraction of free flowable water) changes according to the stage of
cooking. Thus, the height of the fixed bed also changes with
respect to cooking time. Water is present above the fixed bed and
also in the interstices of the fixed bed.
Thus, in the present case, heat transfer occurs from steam to a
fixed bed in the bottom region and steam to water in the top
region of the pot. The latter problem was addressed in Section Figure 11. Modified arrangement of pots/vessels; spacers placed in
2.3. Now, let us consider the problem of heat transfer from steam between pots/vessels.
to fixed bed.
2.4.2. CFD Simulation. Reddy and Joshi5,30 have performed
CFD simulations of the flow pattern in a fixed bed. The authors of the dead zone in the central bottom region increases. Because
have given details pertaining to the CFD simulations consisting this region is away from the location of condensing steam, the rate
of governing equations, boundary conditions, numerical procedure, of the rise of temperature was found to be relatively slow and did
grid size effect, etc. The results of these studies were used to under- not even reach the cooking temperature as shown in Figure 10 at
stand the convective flow pattern of the water in a cooking pot. this location. In fact, rice was found to be uncooked in this region.
2.4.3. Result and Discussion. As said previously, the cooking The intensity of the problem obviously increases with an increase
pot consists of a fixed bed at the bottom and free water at the top. in pot size and depends as well on the type of food and the water
The solid particles provide resistance to flow, and the flow patterns quantity.
as shown in Figure 8 are modified. These are shown schematically We need to ensure complete cooking in all parts of the pot in
in Figure 9. It may be pointed out that the height of the fixed bed as short a time as possible. For this objective, the dead zone in the
increases with respect to time in the cooking process. The value of central bottom region needs to be eliminated or at least some
the convective fluid velocity decreases in the presence of the fixed fluid convection needs to occur. For this purpose, two modifica-
bed. The size of the dead zone also increases. In particular, the size tions have been employed.
1906 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 13. Temperature variation at two different axial locations along

with internals: (A) smaller pot with perforations, (B) four perforated
quarter parts, (1) near bottom, (2) near top.

Food items were kept inside the perforated pot. Water

could remain present in the outside gap as well as in the
perforated pot. Temperature profiles measured at the
same locations (as shown in Figure 10) are shown in panel
(A) of Figure 13. Substantial influence can be seen on the
dead zone, where the temperature profile is observed to
reach the cooking temperature and also in a shorter time
(Figure 13A).
(2) In the second modification, four quadrant-shaped perfo-
rated pots (2 in Figure 12 A,B) were kept inside the
original pot with gaps on the side (2 in Figure 12 B),
middle, and bottom (Figure 12B). The temperature
Figure 12. Pot with different internals: (A) smaller pot with perfora- profiles in panel (B) of Figure 13 are also shown to
tions, (B) four perforated quarter parts, (1) vessel, (2) perforated vessel reduce the presence of dead zones, resulting in a faster
of quarter vessel. temperature rise.
Either of the two modifications permitted an increase of pot
size from 280 mm to 510 mm. Even larger sizes are yet to be ex-
In the first modification, spacers were placed between the pots amined and may allow an even higher initial charge of rice
(Figure 11). As a consequence, steam was allowed to enter and and water.
condense on the bottom outer side of the second and third pots.
This modification increases the total heat transfer area (which
otherwise is not easily modifiable) and also puts the enthalpy 3. COOKING OPTIMIZATION
source near the observed dead zones and uncooked material loca- The systematic development of a cooker was described in Part I.
tion. This modification reduces the size of dead zone. For further Sections 4, 5, and 6 have described the development of the
reduction in size of dead zones, two types of internals were cooker with optimization of energy efficiency. In all the earlier
attempted: cases, only water was used as the charge. It was thought desirable
(1) Another perforated pot (2 in Figure 12 A) was placed to evaluate the performance of the cooker developed earlier for
inside the original pot (1 in Figure 12 A) with a gap of cooking of rice and lentils and, if necessary, optimize the design
about 5 mm all round as well as on the bottom (Figure 12A). further.
1907 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Table 3A. Gas Consumption To Cook 100 g Rice in Cooker of Different Capacities

charge

S.N. cooker capacity (L) rice (kg) water (liter) LPG flow rate (mL/s) time (s) total G.C. (g) gas consumption per 100 g food item (g)

1 4.5 1.5 3.0 6.6 4220 64 4.3

2 6.0 2.0 4.0 7.3 4880 82 4.2
3 24.0 8.0 16.0 27.6 5200 330 4.1
4 72.0 12.0 24.0 83.3 2580 495 4.1
5 72.0 16.0 32.0 84.2 3340 646 4.0
6 72.0 20.0 40.0 84.8 4090 798 4.0
7 72.0 24.0 48.0 85.8 4820 950 3.95

Table 3B. Gas Consumption To Cook 100 g Lentils in Cooker of Different Capacities

charge

S.N. cooker capacity (L) lentils (kg) water (L) LPG flow rate (mL/s) time (s) total G.C. (g) gas consumption per 100 g food item (g)

1 4.5 1.1 3.3 6.6 4350 66 6.0

2 6.0 1.5 4.5 7.3 5120 86 5.7
3 24.0 6.0 18.0 27.6 5370 341 5.7
4 72.0 9.0 27.0 83.8 2620 505 5.6
5 72.0 12.0 36.0 84.8 3380 660 5.5
6 72.0 15.0 45.0 84.1 4270 826 5.5
7 72.0 18.0 54.0 86.9 4850 970 5.4

3.1. Cooking Performance of Cooker. A new set of cooking Using the traditional method of cooking for rice, lentils, and
experiments were performed with “Surti Kolam” variety of rice in potatoes, the mess earlier required about 14 LPG cylinders of
the proportion of 1:2 with water. Cookers of 4.5, 6, and 24 L were 14.2 kg net weight per month. Prior to the introduction of the
charged to the full capacity, and the cooker of 72 L capacity was new cooker, the food items were cooked by conventional
charged to 50%, 67%, 83%, and 100% capacity as described in methods (open pan cooking). For instance, lentils and potatoes
Section 6. Thermocouples were inserted in all the vessels and were cooked by boiling. Rice was cooked by the “drain method”
the water in the base. Gas was supplied at respective flow rates that is a common practice whenever rice has to be cooked on a
(similar to those employed for the heating profile) until the large scale. After the introduction of a new cooker, the LPG cylin-
charge temperature reaches 98 C. Retention was then allowed der requirement was found to be reduced from 14 to 6.
for 30 min, after which the cooker was opened. Rice grains in all 3.2.1. Experimental Setup and Procedure. Three types of
the vessels were tested for the extent of cooking. It was observed burners, namely, torch, sunbeam, and moonbeam, have been
that the grains were satisfactorily cooked to the required texture. examined (BIS standard31). The base of the 72 L model of the
Similar experiments were performed for “Arhar (Tuvar)” variety cooker was placed on the LPG burner. The required quantity of
of lentil (dal) in 1:3 proportion with the water. Lentils also were water was poured in the base. Three stands (for vessel stacks)
observed to be nicely cooked and could be mashed easily. Gas were placed in the base. The vessels with requisite quantities of
consumption figures for cooking of unit quantities of 100 g rice charge were placed in the stands. Thermocouple probes were
and 100 g lentils are listed in Tables 3A and 3B, respectively. inserted in water in the base and in the pots. Inverted plates were
From the knowledge that the rice grain cooking starts at 74 C kept on top of vessels (V4) in all three stacks. Wire stands were
onward (depending on the quality of rice) and lentil grain cook- provided to avoid the movement of vessels within each stack.
ing starts at 94 C (which also depends upon the quality of The LPG regulator was switched on, and the flow control valve
lentils), it can be observed that the gas consumption in the case of was adjusted in such a manner that the flame diameter was about
rice is less than that for lentils at their respective optimum ratios one-third of the base diameter. The cooker contents were heated
with water. It has also been observed that the gas consump- until they reached boiling point at which point the gas supply was
tion per unit quantity of food item reduces as cooker capacity switched off. Retention was then allowed, without opening the
increases. It may be pointed out that the pressure cookers are cooker. Gas flow rates during heating and temperatures of the
hardly available beyond 30 L capacity of the outer vessel. various components during heating and retention were measured
3.2. Standardization and Optimization of Parameters for at predetermined time intervals.
Cooker. In order to examine the performance of the present 3.2.2. Water in Base. The condition required during the
cooker, a 72 L capacity model was supplied in December 2002 cooking operation was that the water in the base should touch
to the Post Graduate (PG) Students Mess at the Institute of the bottom pots (V1) during the heating period, whereas the con-
Chemical Technology (ICT), Matunga, Mumbai, India, that tact between water and vessel V1 should be broken during
caters to 300 students at a time. Its regular requirement of food retention. In order to optimize the water quantity in the base, cook-
items at a time using the usual cooking method is 14 kg of rice ing experiments were carried out by keeping lentils in bot-
and 4.5 kg of lentils (tuvar). Potato requirement is about 6 kg . tom vessels (V1) and rice in the upper pots (V2, V3 and V4).
1908 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Table 4A. Optimum Water Quantity for Cooking of Rice in Metal Covered Vessela

water losses (g)

water quantity (g) gas flow rate (mL/s) time required (s) gas consumed (g) heating cooling result

200 4.3 1110 11 20 62 undercooked

7.3 800 13.5 20 65 undercooked
21.4 420 burnt without cooking
300 4.3 1210 12 22 75 cooked
7.3 890 15 22 80 undercooked
21.4 470 burnt without cooking
400 4.3 1320 13 21 70 cooked
7.3 980 16.5 23 84 cooked
21.4 510 burnt without cooking
500 4.3 1470 14.5 22 74 cooked
7.3 1070 18 24 88 cooked
21.4 610 30 26 233 cooked
a
Rice quantity = 100 g, Vessel capacity = 5.0 L, D = 180 mm, H = 220 mm.

Table 4B. Optimum Water Quantity for Cooking of Lentils in Metal Covered Vessela

water losses (g)

water Quantity (g) gas flow rate (mL/s) time required (s) gas consumed (g) heating cooling result

200 4.3 1420 14 45 118 Undercooked

7.3 950 16 48 120 undercooked
21.4 780 burnt without cooking
300 4.3 1620 16 51 119 cooked
7.3 1070 18 52 123 cooked
21.4 810 burnt without cooking
400 4.3 1720 17 52 124 cooked
7.3 1190 20 59 133 cooked
21.4 860 burnt without cooking
600 4.3 2120 21 60 142 cooked
7.3 1370 23 61 147 cooked
21.4 950 burnt without cooking
800 4.3 2530 25 61 166 cooked
7.3 1550 26 64 202 cooked
21.4 1040 51 75 382 cooked
a
Lentils quantity = 100 g, vessel capacity = 5.0 L, D = 180 mm, H = 220 mm.

During heating and retention, temperatures of charge in all the because the temperature in those vessels dropped considerably
pots were also noted. during the retention period resulting in undercooking. This indi-
During the experimentation, it was observed that for 10 L cates that the cooking pattern in the bottom pots is governed by
quantity of water in the base, rice in all the pots was nicely cooked. three parameters: (1) Quantity of water in the base should be
Lentils in two pots (bottom pots of two stacks) was undercooked. sufficient so that it is in continuously in contact with the bottom
Pots containing lentil that was undercooked showed final tem- vessels during heating. (2) There should be no contact between
peratures of 84 and 88 C. It indicates that these pots were not in the two during the retention time. (3) The base should be
contact with the water in the base for a sufficient time, resulting in perfectly horizontal if we want to use minimum water quantity
their inadequate heating. Hence, the water quantity in the base in the base (any inclination may result in different extent of
was increased to 12 L, keeping other conditions the same. It was contacts for different vessels).
observed that the rice in all the vessels was nicely cooked, but the From Section 4.6 of Part I, it can be observed that for the water
lentils in one vessel was undercooked. The temperature of the quantity of 12 L in the base, food items in all the pots were nicely
charge in this vessel was 91 C. The water quantity in the base was cooked. Comparing these results with the present one, it can be
further increased to 15 L, and cooking was carried out. It was concluded that the discrepancy in cooking of the charge in the
observed that rice in all the vessels was nicely cooked. However, bottom pots was because of the base that was not in a perfectly
lentils in all three vessels was undercooked, despite the tempera- horizontal position. Even in actual field practice, it may not be
tures of charges in these vessels having reached 98 C. This was possible, willingly or unwillingly, to keep it in a perfectly horizontal
1909 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

position because of the relatively large size of the base. Hence, Hence, the optimum ratio can be considered as 1:3. At a higher
a drain tap was provided in the base in order to drain excess water LPG flow rate of 21.4 mL/s, the ratio is 1:8, so that lentils are
at the commencement of the retention period. cooked after making provisions for water evaporation. (d) Stop-
Drain tap has a position in the base such that even if it is kept ping the gas supply after reaching the temperature of 98 C (i.e.,
open about 7 L of water is retained in the base. At this condition, early shut off condition) and then retaining the enthalpy of the
water in the base does not touch the bottom pots (V1). Fifteen system can further reduce gas consumption and the required
liters of water was sufficient to maintain contact with the bottom water quantity. Moreover, if the arrangement is such that the
pots, even if the base is in a slightly tilted position. At the start of food item does not come in direct contact with the vessel wall
retention, excess water can be drained by opening the tap. that receives heat directly from the flame, burning of the food
3.2.3. Gap between Pots. In this cooker, heat is transferred to item can be avoided.
the contents of the vessel via condensation of steam on the In view of these observations, some experiments were per-
outside surface of vessel. Thus, the arrangement of vessels V1, V2, formed for further optimization of these ratios, and it is recom-
and V3 plays a vital role in the cooking operation. The arrange- mended for the case of rice that the water ratio be 1:2.3 with a gas
ment of vessel decides the actual area available for steam to flow rate of 4.3 mL/sec. If a higher flow rate is used (for short-
condense and thus the rate at which heat energy is transferred to ening the time at the expense of efficiency), the recommended
the contents of vessels. Stacking of vessels one over another ratio is 1:3 at a flow rate of 7.3 mL/sec and 1:5 at a flow rate of
(Figure 1) blocks the bottom area of vessels V2 and V3. Contents 21.4 mL/sec. The same numbers of lentils are recommended to
of vessel V1 is heated at much faster rate compared to vessels V2 be 1:3, 1:4, and 1:8 for gas flow rates of 4.3, 7.3, and 21.4 mL/sec,
and V3, as V1 is always in physical contact with boiling water. respectively.
Then the content of the top vessel V3 is heated and finally the One important observation can now be emphasized. The
middle vessel V2. This difference in the rate of heating for vessels water ratio increases with an increase in heat flux. For instance,
V2 and V3 sometimes results in a difference in the quality of the value of ratios of water to rice are found to be 2.3, 3, and 5 for
cooked food. heat flux numbers of 15,200, 25,830, and 75,720 kcal/h m2,
This difference in heating rate is eliminated by simply adding respectively. Similarly for lentils, the ratio was found to be 3, 4,
spacers between vessels (Figure 11). This arrangement ensures and 8 for above-mentioned fluxes.
that the area available for steam to condense in vessels V2 and V3 We would like bring to attention one more important point.
remains same. A uniform rate of heating is achieved by this arra- The quantity of water required depends upon the quality of rice
ngement for vessels V2 and V3, and thus the quality of cooked (possibly related to the amylose content of rice). When an
food in those vessels is uniform as well. inferior quality of rice is used, the ratio of water nominally in-
3.2.4. Water Ratio on Smaller Scale. Different food items creases to 2.35 ( 0.05. This value has been used whenever an
require water in different proportions for cooking. To see the inferior quality of rice has been used.
effect of water proportion on cooking characteristics, experiments 3.2.5. Water Ratio on Larger Scale. After studying the effect of
were performed with different ratios, and these are described in water ratio for cooking on smaller scale, actual cooking experi-
the following sections. ments were conducted with rice, lentils, and potatoes with 72 L
In order to investigate the effect of the quantity of water, dif- cooker model, and the results are shown in experiments 1, 2, and
ferent quantities were added to 100 g of rice or lentils. Experi- 3 in Table 5. As described in Table 5 of Part I, the pot size is
ments were performed at gas flow rates of 4.3, 7.3, and 21.4 mL/s D = 255 and H = 140 mm. The variety of food items used was
using the 5 L vessel (D = 180 mm, HE = 220 mm) with lid and “Wada Kolam” rice and “Arhar (tuvar)” lentils. Potatoes available
insulating metal covers. The results based on the required texture from a local market were used. In experiment 1, the rice to water
and consistency of the cooked material are listed in Table 4A for quantity ratio of 1:2, lentil to water ratio of 1:3, and potato to
rice and in Table 4B for dal (lentils). water ratio of 1:2 were used, as these were the optimized ratios at
The following observations were made: (a) When it was a laboratory scale for these available varieties. Each pot was
observed that a food item gets stuck to the vessel base that is charged with 6 kg of material. Lentils were kept in the bottom
in direct contact with the flame during heating, further heating vessels (V1), rice was kept in vessels V2 and V4, whereas potatoes
was stopped as it would result in burning of the food item. This were kept in vessel V3. Twelve liters of water was added in the
was the case when the water quantity added to the food item was base. Heating was continued at the LPG flow rate of 90.4 mL/s
completely evaporated before completion of the cooking pro- for about 85 min. LPG consumed was 1060 g. Retention was
cess. This could be guessed easily for smaller flow rates, at which allowed for 30 min.
time the gas supply was stopped. It was considered as the primary Temperatures were noted during the heating and retention
stage of food burning. However, for the higher flow rate, periods. It was observed that lentils and potatoes in all the vessels
21.4 mL/s, early detection of primary burning was found difficult, were nicely cooked to the required texture. However, smaller
and the food item was almost burnt without cooking. (b) For a pockets of rice ranging between 100 and 200 g were observed
rice to water ratio of 1:3, rice was cooked at the LPG flow rate of undercooked in the central section of each of the pots. Rice in
4.3 mL/s, partially undercooked for the flow rate of 7.3 mL/s, other parts was cooked. Also, temperatures of these pots were
and burnt without cooking for the flow rate of 21.4 mL/s. about 94 C. This indicates that radial heat transfer was not
However, it remained undercooked for these flow rates for the effective up to the center of those pots. This was possibly because
1:2 ratio. Hence, the optimum ratio at lower flow rates is 1:3. At a the water quantity was inadequate to cook this variety of rice. No
higher flow rate of 21.4 mL/s, the optimum ratio is 1:5. (c) For a excess water was left available for heat transfer. This resulted in an
lentils to water ratio of 1:3, lentils were cooked at a LPG flow rate increased resistance to heat transfer to the center of the pot.
of 4.3 mL/s, partially undercooked for a flow rate of 7.3 mL/s, Hence, experiment 2 was performed using a rice to water ratio of
and burnt without cooking for a flow rate of 21.4 mL/s. However, 1:2.5, whereas other ratios were kept the same. Heating carried
it remained undercooked for lower flow rates for 1:2 ratio. out with a gas flow rate of 90.1 mL/s for 84 min consumed 1042 g
1910 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Table 5. Cooking Results of 72.0 L Cooker in PG Students Mess, ICT Mumbaia

vessel food + water (kg)

E.N. V1 V2 V3 V4 D R E.L. (kg) heat time (min) cool time (min) LPG FR (mL/s) G.C. (g) remark

1 D R P R 1.5 + 4.5 2.0 + 4.0 7.4 85 30 90.4 1060 uncooked rice pockets
2 D R P R 1.5 + 4.5 1.7 + 4.3 7.2 84 30 90.1 1042 cooked sticky rice
3 D R P R 1.5 + 4.5 1.8 + 4.2 6.9 83 30 89.8 1030 optimum water
4 D R R D 1.4 + 4.6 1.8 + 4.2 7.0 85 30 90.8 1068 ∼5 min more heating
5 D R R D 1.4 + 4.6 1.8 + 4.2 6.7 85 40 87.3 1024 ∼5 min more heating
6 D R R D 1.4 + 4.6 1.8 + 4.2 5.2 85 30 86.1 1010 slightly more heating
7 D R R R 1.4 + 4.6 1.8 + 4.2 5.8 85 30 88.1 1000 ∼5 min more heating
8 D R R R 1.4 + 4.6 1.8 + 4.2 5.1 82 35 86.6 980 adequate heating
9 D R R D 1.5 + 4.5 1.8 + 4.2 2.2 85 30 82.7 970 sufficient heating
10 D R R D 1.5 + 4.5 1.8 + 4.2 1.9 85 30 81.8 960 sufficient heating
11 D R R D 1.5 + 4.5 1.8 + 4.2 2.1 85 30 81.1 950 sufficient heating
12 R R R R
1.8 + 4.2 2.3 84 30 81.0 940 sufficient heating
13 D D D D 1.5 + 4.5
1.9 86 30 82.2 970 sufficient heating
14 D R P R 1.5 + 4.5 1.8 + 4.2 2.2 82 35 83.5 945 sufficient heating
a
V: vessel (from bottom to top). D: dal (lentils). R: rice. P: potatoes. E.L.: evaporative losses. G.C.: gas consumption. E.N.: experiment number. Variety
of food item: rice, Wada Kolam; dal (lentils), Tur (Arhar); potatoes, unchopped whole potatoes.

of LPG. It was observed that the lentils, rice, and potatoes were were allowed to presoak for some time. All these results are
nicely cooked. However, the cooked rice was sticky and not shown in Table 5 as experiment numbers 4
14. It is observed
completely free-flowing as required. This indicated that the water from this table that the optimum is to use 940 g of LPG to cook
quantity in the rice was in excess. Hence, experiment 3 was 22 kg of rice, 950 g of LPG to cook 9 kg of lentils and 11 kg of rice,
performed using a rice to water ratio of 1:2.3 keeping the other 945 g of LPG to cook 4.5 kg of lentils, 11 kg of rice, and 6 kg of
ratios the same. At a gas flow rate of 89.8 mL/s for 83 min, 1030 g potatoes, and 970 g of LPG to cook 18 kg of lentils. In general, gas
LPG was consumed. All the food items were nicely cooked. The consumption to cook a full charge is about 955 ( 15 g.
rice was not sticky and was completely free-flowing. Hence, these At these conditions, lentils could be easily mashed after
ratios were maintained for future cooking experiments when the cooking. Rice was very nicely cooked and was completely free-
same quality grains were used. Some observations pertaining to flowing. Potatoes were also nicely cooked and could be peeled
the ratio of water to rice (for small scale) have been given at the very easily. All the cooked food items had better taste, flavor, and
end of Section 3.2.4. Similar observations also hold for large scale appearance compared to those cooked in the conventional way.
cooking. This finding was based on a survey of the opinion of 300 student
The foregoing discussion brings out the optimum water members of the mess. Heating time is about 85 min and retention
quantity for 180 and 255 mm diameter pots. Some experiments time is 30 min. If the gas flow rate is further reduced, heating time
were performed on even larger scales, and pots of 280 and would increase, and that may act as a constraint in the overall
330 mm diameter were employed. Following are the observa- timetable of the hostel mess.
tions: (a) The 280 and 330 mm pots can cook 2 and 4 kg rice, The effect on thermal efficiency of a cooker with a (i) flow rate
respectively. If the rice quantity is less than 25%, then it remains of LPG and (ii) distance between the burner and cooker base
uncooked. (b) For all sizes (180, 255, 280, and 330 mm), when are discussed in more detail in Part I (Sections 4.5 and 5.5,
the initial rice quantity is more than 50% of maximum capacity respectively).
of the respective pots, the rice to water ratio was found to be 3.2.7. Extent of Cooker Usage. Taking into consideration that
2.35 ( 0.05. the cooker may not be used with a full charge every time,
In the foregoing discussion, the problem of the uncooked experiments were conducted for different charges, e.g., 3, 4,
portion was solved by using additional water. As discussed in and 5 L in each pot (i.e., 50%, 66%, and 83% of full capacity of the
Section 2.4, some portion of rice remains uncooked because of tem- cooker) and again temperature profiles were studied. It was
perature stratification. In view of this, some hardware changes observed that heating and retention cooling profiles were similar
have been proposed in place of using an additional quantity to those observed for a 72 L charge as shown in panels (A) and
of water. (B) Figure 14. At the LPG flow rate of 85 mL/s, the time to reach
3.2.6. LPG Consumption for Cooking. In order to optimize the the boiling temperature, the gas consumed, and the efficiencies
gas consumption to cook a full charge of the cooker, experiments for different capacities are shown in Table 6. It is observed that
were performed for different combinations such as (i) only rice, although the efficiency increases with an increase in charge, there
(ii) rice and lentils, (iii) only lentils, and (iv) rice, lentils, and is no significant difference between the efficiency values for differ-
potato. The total thermal load was 72 kg of cooking material and ent charges.
water and 15 L water in the base. Heating was carried at a 85 ( 3.2.8. Effect of Geometry. Experiments were also carried out
5 mL/s gas flow rate and respective gas consumptions were using various geometries of cooking pots/vessels such as cylindrical,
calculated. Temperatures of charges in the different pots were triangular, square, and rectangular. The dimensions are shown in
also noted. To reduce gas consumption further, lentils and rice Table 7. Initially experiments were carried out using cylindrical

1911 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

pots, which are readily available in the market. A cooker of 120 L 3.2.8.1. Cylindrical Vessel. It was observed that good quality
capacity accommodates 15 pots. Fourteen pots were charged rice (quality of the rice has been measured in terms of its market
with 8 L water, and one pot was charged with 2 kg rice with a price), about 50 cents per kg, is cooked nicely. However, low
1:2.5 as rice to water ratio. LPG flow rate was maintained at quality rice (Rs. 35 cents per kg) remains partially uncooked.
1.1 kg/h. Heating was continued for 15 min after a steam puff comes Therefore, attempts were made to cook this poor quality rice by
out from the base (approximately 60
70 min). The retention increasing the water to rice ratio. The increase in ratio, however,
period was maintained constant (40 min). results in mushy rice. It was thought that an increase in the heat
transfer area to volume of vessel (Table 7) may help to cook the
rice, and hence, various geometries were tried. Further, variation
of the rice to water ratio for a given quantity of rice was also
studied for low quality rice. All experiments were carried out
using low quality rice unless otherwise specified.
3.2.9. Other Aspects. In addition to the various parameters
affecting the cooker performance discussed, additional factors
were studied for making it user-friendly and more acceptable.
3.2.9.1. Pot Design for Cooking and Dispatch. In all devel-
oping countries including India, large numbers of charitable
institutions along with various government bodies offer free
food, mainly rice and lentils or khichdi (mixed food containing
rice, lentils or other cereals, vegetables, and spices), to school/
underprivileged children (about 110 million in India). The
process involves managing and handling of large food quantities.
Conventionally, large vessels are used to cook the rice by boiling
(open pan, drain method) as described before. It is then packed
in small containers (∼15 kg) and dispatched to the respective
destinations. The overall process including removal of cooked
rice from large cooking vessels and then distributing it again is a
labor intensive and particularly unhygienic step. This problem
could be addressed with better design of cooking pots that would
allow cooking and dispatch of the food in the same container.
Two methodologies were tested for this purpose.
Because the open pan method is widely followed at almost all
places with large burner flow rates, heat losses are significant.
This causes low thermal efficiency and high consumption of fuel
that in most cases of community cooking is LPG. For example, at
Ratnanidhi Charitable Trust, Girgaon, Mumbai, ∼200 kg of rice
and 125 kg lentils are cooked per day. Both rice and lentils are
cooked by the method described above, with three large burners
and three cooking vessels. Multiple cooking batches are required
daily for cooking of the said food quantities. The fuel consump-
Figure 14. Temperature profiles for water in base, vessels, and cover for tion is 75 LPG cylinders (capacity 14.2 kg) per month. Thus, the
sunbeam type burner: (A) heating, (B) cooling. cooking efficiency is in the range of 25%. The use of our cooker

Table 6. LPG Consumption and Efficiency for Different Charges in 72.0 L Cooker
S.N. charge in vessels (L) time of heating (min) LPG consumption (g) ηtherm (%)

1 36 45 524 71.4
2 48 55 643 71.5
3 60 67 788 71.8
4 72 82 960 72.1

Table 7. Particulars of Various Geometries of Cooking Vesselsa

dimensions volume surface area volume/surface area

3
3
geometry D/L/B (mm) height (mm)  10 3
(mm )  10 (mm ) 2
 10
1 (mm)

square 260/280 300 21870 3865 5.66

triangular 240 200 5985.95 1938.82 3.09
cylindrical 280 150 9231.6 1495.425 6.17
350 200 19232.5 2512 7.66
a
For square vessel, lower side = 260 mm, upper side = 280 mm. For triangular vessel, base = 240 mm, height = 200 mm.

1912 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

would save at least 45 LPG cylinders per month. Three cookers

of 120 L capacity each were supplied to this site, and a successful
cooking demonstration was given for one cooker. Though the
cooker worked nicely, it contained 15 of the 8 L pots. The 8 L
capacity was considered to be small. Because of the specific
requirement of cooking and transport in the same pot, pots with a
square base were designed similar to those of the 700 L cooker
(Figure 18 of Part I).
In addition to fuel saving, there is an advantage in food
distribution and transport. There is no need to remove the
cooked food from the cooking pots and reload it in other vessels.
Also, because of the specific pot design, they can be properly
arranged in the transport vehicle. A full charge cooking trial was
performed on a 160 L capacity cooker with these pots. Twelve
pots were arranged in four stacks of three pots each. Though the
full capacity of the pot is 14.4 L, it is necessary to allow for
swelling of the cooked food. Hence charging was done accord- Figure 15. Comparison of temperature profile for vessels for two
ingly. All of the four bottom pots were charged with 3 kg of lentils modes of heating. DSH: Direct steam heating. GH: Gas heating.
and 9 kg of water, 7 pots were charged with 3.5 kg of rice and
9.625 kg of water, and the remaining pot was charged with 3 kg of volume of about 0.6 and 1.7 L for small and large carriers,
rice and 8.1 kg of water. The rice used was of low quality with a respectively.
relatively low price. Twelve kilograms of water was charged in the Thermocouples were placed at the center of each container, in
base. It was found that both rice and lentils got cooked nicely, the base, and on the outer cover to measure temperatures. Water
though the cooked rice was sticky in nature. It can be said that quantities were monitored before and after the experiment for
this is mainly due to the quality of rice. It is always found that evaporative losses. Typical temperature profiles were found to
good quality rice is always free-flowing in nature and low quality a be similar to that for the cooker itself. On average, thermal
little bit sticky. This may be due to non-uniformity of initial efficiencies obtained for small and large tiffin carriers were 66%
moisture content in the low quality rice in addition to the mixing and 59%, respectively. These values are quite low compared to
of two varieties of rice. Thermal efficiency of the system was the possible efficiency of the standard cooker. This may be due to
found to be 73%. Replacement of present cooking practice by this the fact that the complete capacity of the cooker was not utilized.
cooker at Ratnanidhi Trust could result in LPG savings of 45 The height of small and large tiffin carriers is 140 mm and 217 mm,
cylinders/month worth $300/month (with rate of LPG at $6.67 respectively. When this is compared to the height of the cooker
per cylinder). It may be mentioned that the cost of one cooker is (575 mm), it is observed that the extra height provides additional
about U.S. $630. With 20% depreciation and 10% interest, the surfaces for heat loss. Three carriers occupy almost the entire
payout period works out to be less than eight months. base area, while a single large carrier uses less than half the base
Another option is cooking in small food containers like tiffin area. This combination when compared to the 24 L cooker
carriers. This can be useful at places where only rice and lentils suggests that heat losses remain the same in both cases and hence
(or a khichdi-like item) are to be supplied. In cooking of lentils, it result in the lowering of net thermal efficiency at lower inner
is normal practice to cook (boil) the lentils first and then perform container capacities. This limitation can be overcome with pro-
the tadka (frying process with vegetable oil followed by addition per sizing of cooker and tiffin carriers for such a cook-and-dispatch
of spices separately spluttered in oil). This process requires extra operation.
effort as well as extra fuel. Instead of this, the ingredients for tadka 3.2.9.2. Direct Steam Injection. Though the cooker is pre-
(oil, spices, etc.) may be mixed well and added to the cooking sently standardized with LPG as fuel, it can also be used with
material prior to cooking. This would avoid the reprocessing of other fuels like biomass or with an external steam supply. The
lentils after cooking, and hence, the only required operation option of external or direct steam supply was investigated keep-
would be to distribute the separately prepared tadka (or plain raw ing in mind that steam could be available from any other efficient
spices and oil depending on choice) prior to cooking along with steam generation method where the overall system efficiency
the soaked lentils in the containers. For applications involving would be higher. Also, another option for steam generation
khichdi-like preparations, it is necessary to distribute appropri- comes from the use of a renewable source such as solar energy.
ately mixed grains and other ingredients to all containers of The option of using solar generated steam would mean a com-
the tiffin carrier. Thus in both of these applications, there is no pletely CO2 free and highly efficient cooking practice.
secondary processing operation such as tadka, remixing, etc. An external steam supply option was investigated using a 160 L
Once the cooker is opened, the food is ready for direct serving. cooker with 15 vessels. Fifteen vessels were arranged in five
Two sizes of tiffin carriers (6 and 12 L capacities) in the 24 L stacks, each stach having three vessels, V1, V2, and V3. Steam was
cooker were used for testing. The 24 L cooker was used because it introduced inside the cooker through 2.5 cm diameter holes,
was found adequate to hold three small tiffin carriers. It should be drilled at the center of the inner and outer cover. Stacks of vessels
noted that a lot of space remained unutilized inside the cooker. were arranged such that steam was hitting the base and not any
A domestic LPG stove was used as a heating source with a large the vessels directly. Steam was generated at a rate of 25 kg/h
burner at maximum flow rate (∼3 g/min or 21.74 mL/s) for all using an oil-fired boiler. Figure 15 depicts the comparison of
experiments. Three carriers of small capacity were used at a time, temperature for vessels V1, V2, and V3 for two different mode of
while in the case of the larger one, a single tiffin carrier was used. heating: gas heating and introduction of direct steam-generated
Both small and large carriers have four containers each having a heat using a boiler.
1913 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Temperature profiles for the vessels clearly suggest that

10
15 min of cooking time can be saved by direct introduction
of steam. Vessel V1 is heated at a same rate for both modes of
heating in the initial few minutes.
Testing of the direct steam supply to the cooker was done
using the 24 L cooker and a tiffin carrier of 12 L. Steam used was
saturated steam at 100 C and was generated using a 6 L pressure
cooker after removing the cooking vessels and the whistle weight.
Steam coming out from the nozzle at the top of the pressure
cooker was fed to the cooker once a steady steam rate was
obtained. Because the heat is to be supplied by steam, a hole was
drilled through the two covers of the cooker for steam injection
through a pipe that released steam near the bottom of the cooker.
The diameter of the hole was selected so as to match the size of
the steam injection pipe. The steam rate obtained was 45 g/min.
It was observed that although the temperature of vessel V1
Figure 16. Schematic of chakson cooker.
rises faster than the other vessels, it is no longer as fast as that for
the regular cooker with heating by a flame below. The reason for
unlike the pressure cooker, and the 72 L capacity is being
a smaller difference between the temperatures of V1 and the
successfully operated at about 50 locations in India.
other vessels is that steam is available for heating to all the vessels
3.3.2. Chakson Cooker. A Chakson cooker is the patented
simultaneously. Whatever small difference is there is possibly due
product of Anna Aluminum Private Limited, Kerala. It utilizes a 5 L
to following reasons. Vessel V1 has the largest area available for
capacity vessel with lid and thermally insulated box. The
steam condensation (extra bottom surface), and steam was
schematic of the cooker is shown in Figure 16. The insulated
injected near the bottom pot. The temperature rise for vessel
box is double walled and consists of 12 mm thick thermocole
V3 is the slowest followed by vessel V4. The temperature of the
(polystyrene foam) as insulation between the walls. Food
base is actually the temperature of condensate accumulated.
material, along with an excess quantity of water, is put in the
Because the base is directly exposed to the surroundings, it also
pot, which is then heated to boiling point. The mixture is stirred
loses heat. The time required to reach the desired temperature
two or three times during heating as per requirement. When the
levels was 40 min. Actual heat input to the cooker was calculated
boiling point is reached, the pot with lid is put into the lower part
from the steam flow rate and time of experiment as
of the insulation box. Both parts of the box are tightly fitted
creating a closed system. Retention is allowed for a much longer
Heat Input ¼ rate of steam supply  latent heat  time period, and the box is opened when food is served.
3.3.2.1. Chakson Cooker versus Present Cooker. As both
The thermal efficiency of the cooker was calculated consider- cooking devices work on similar principles of boiling the contents
ing the heat given out by steam and that gained by the cooker and and retention, it was thought desirable to compare the retention/
was found to be 84.5%. The efficiency for pots (i.e., water kept in cooling processes. Figure 17 shows the cooling period compar-
the containers of the carrier) was 74.2%. For very large scale ison between vessels V2 and V3 of the 4.5 L capacity cooker and
cooking systems providing a few million meals per day, steam is the 5 L capacity Chakson vessel during the retention period. It is
generated using industrial boilers, and it is then used for cooking. observed that vessels of the present cooker cool slower than the
This experiment shows that if steam is available by external Chakson pot, thus showing better performance. In the Chakson
means, it can be used more efficiently in the cooker, and there is cooker, only one item can be cooked at a time compared to four
further scope for improvement in thermal efficiency in large scale items in the present cooker. Taste and nutrition is preserved in
cooking. both the models. Food can be kept warm for at least 20 h. Heating
3.3. Comparison of Present Cooker with Other Cooking efficiency of the Chakson pot varies from 50.5% at a 21.4 mL/s
Devices. Some of the cooking devices available in the market are flow rate to 56.7% at 5.6 mL/s flow rate, which is lower than that
the pressure cooker, Chakson cooker, Sarai cooker, and Mallick for the present cooker which in the range of 65
75%. The
cooker. These models are capable of cooking common food Chakson cooker can cook a maximum of 1 kg of rice at a time.
items such as rice, lentils, and vegetables. The performance of the This would require 90 g LPG to cook 1.5 kg of rice compared to
present cooker has been compared with pressure and Chakson 60 g for the present cooker at similar LPG flow rates.
cookers in terms of energy saving, capacity, scalability, ease of 3.3.2.2. Effect of Insulation Material. The present cooker
operation, and food quality. Other cookers use coal as a heating utilizes a 5 mm width air gap between the two covers for 24 L and
source and are not discussed here. 12 mm for 72, 120, and 160 L cookers as insulation to reduce
3.3.1. Pressure Cooker. A pressure cooker is a device that cooks convective heat losses. It was thought desirable to investigate the
food items using steam pressurized at 0.2 MPa and 120 C. It is effectiveness of the same insulation principle in a Chakson
mostly operated on commercial fuels such as LPG, kerosene, and cooker. Hence, a Chakson cooker was modified to accommodate
coal. It is available in the market in capacities ranging from 2 to 30 L a thermal insulation box having a 5 mm wide air gap between the
of the outer vessel. The actual capacity of the cooking vessel is two walls. A 5 L water charge was used in the vessel. The vessel
generally 70
75% the capacity of the outer vessel. Comparison was covered with a lid. Water was heated from room temperature
of the optimum water quantity to be added to the food items and of 27 C to 99 C at a LPG flow rate of 21.4 mL/s.
the energy saving aspects are discussed in Sections 3.2.5 and As soon as the water charge reached a temperature of 99 C,
3.2.6. The present cooker can be scaled up to large capacities, heating was discontinued. The vessel was put in the cooker box,
1914 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

to be used for cooking. However, nearly 75% of the world’s

population lives in developing countries, and the majority of it
continues to depend on traditional sources of energy for most of
its energy requirements. In the last few decades, these developing
countries have experienced a rapid depletion of natural forest
resources that has resulted in hardship for the people living in
rural areas, especially women and children who spend a con-
siderable part of their time and energy in search of fuel wood and
biofuels, and they often have to cover long distances. In parti-
cular, the domestic sector relies heavily on traditional sources of
energy, mainly for cooking, for which traditional stoves are often
used. These stoves are usually thermally as well as environmen-
tally inefficient and hence create drudgery as well as problems for
the users.
4.1.2. History. In earlier ages, cooking was presumably done
over an open fire with fuel arranged in a pyramid configuration.
This mode of cooking, primarily for roasting meat, had major
drawbacks: dispersion of the flames and heat during windy con-
Figure 17. Comparison of temperature profile during cooling. V2 and ditions, lack of proper control over the fire, exposure to heat and
V3 are the vessels in the cooker. smoke, as well as fire hazards. However, at the same time, heat
and smoke had also certain benefits such as food preservation
and the box was closed. Temperatures of water, air inside the box, and/or protection against large animals, insects/rodents, and
and the outer wall of the box were measured using thermo- providing warmth during the cold seasons. A major step toward
couples during a period of 6 h. the evolution of other cookstoves (Chulhas) was the develop-
In both cases, temperatures of the water charge after 6 h of ment of pots of various shapes and sizes. This necessitated the
retention showed a similar cooling trend with the temperature in modification of the open fire to create shielded fires in order to
the former case higher by 2 C over that in the latter case. Air balance the pot over the fire. The simplest form of the shielded
temperature where thermocole was used as insulation is higher fire was a three stone arrangement in which stones were arranged
compared to the air temperature using an air gap as insulation. at approximately 120 degrees to one another on level ground.
The outer wall temperatures of the air-insulated box were higher Besides allowing a cooking pot to rest firmly on it, this arrange-
by about 10 C compared to those of the thermocole-insulated box. ment also partly saved the fire from the vagaries of wind and
There was little difference between the water temperatures for slightly increased cooking efficiency. However, by and large, the
the two cases, indicating an equal rate of heat loss to the three stone fire still suffered similar drawbacks as the open fire.
surroundings. The maximum temperature of air and outer wall Subsequently, the shielded fire was changed to a U-shaped mud
for the air-insulated box are 64 and 46 C, respectively, whereas or mud/stone enclosure with an opening in the front for fuel
for the thermocole-insulated box these were 71 and 37 C, res- feeding and combustion air entry. Three small humps (made of
pectively. The outer wall temperatures in both cases are in line the same mud material) were put on the top rim of the enclosure
with the established industrial norms for an insulated outer surface. and acted as a pot rest, provided an induction point for secondary
According to optimization described in Section 2.2, the air gap air needed for better combustion of volatile matter, and provided
thickness that prohibits the formation of natural convective an exhaust gas exit. In order to conserve heat from the hot flue
currents in this temperature range is estimated to be in the range gases and to enhance cooking productivity, additional pot holes
of 5 mm. The difference between the respective inner and outer were later added. These pot hole enclosures were connected by a
wall temperatures of the two boxes pertains to the differences in tunnel. All the above-mentioned innovations in cookstove design
thermal conductivity and thickness of insulation. The thermal were made mainly by users in the light of their own experience.
conductivity of air is 0.028 W/m K (average) and of thermocole These innovations did increase the efficiency of the stoves to
(polystyrene foam) is 0.04 W/m K, and their respective thick- some extent, but health and other safety hazards remained.
nesses are 5 and 12 mm. Assuming similar heat transfer coeffi- 4.1.3. Improved Chulhas. In pursuit of the conservation of
cients, thermocole thickness equivalent to 5 mm air thickness is fuels, governments of the developing world launched national
about 7 mm. Compared to the air gap, the additional 5 mm thick- programs on improved chulhas (cookstoves). The following are
ness of thermocole improves performance in reducing heat the criteria for efficient cooking operation: (1) simple construc-
transfer from the inner wall to outer wall. This results in the tion and easy fabrication, (2) easy operation and maintenance,
difference between the respective inner and outer wall tempera- (3) easy ignition (4) complete fuel consumption, (5) higher
tures of the two different insulating materials. retention time of flame inside the hearth, (6) higher heat transfer
to the pot, (7) less pollution and soot formation.
4. RENEWABLE SOURCES OF ENERGY Under these programs, many designs were developed by
various agencies. Here, we discuss two that satisfy most of the
4.1. Solid Fuel. 4.1.1. Preamble. In the past, traditional above-mentioned criteria and can be considered as representa-
sources of energy such as fuel wood, charcoal, dung, etc. were tive of recent developments: Harsha stove and Oorja stove.
the only sources of energy used for all types of applications. It is We have selected these two cookstoves for further experi-
only during the last 250 years that fossil fuels such as coal, oil, and mentation for use with the cooker developed in Section 6 of Part I.
gas have emerged as major sources of energy in most developed A Harsha stove operates in continuous mode, which means fuel can be
countries, and it is during this time that electricity has also begun charged continuously until the cooking process is complete,
1915 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 18. Schematic of Harsha stove: (1) corrugated grate, (2) ash
scraper, (3) perforated combustion chamber, (4) air jacket, (5) fuel
feeding hole.

whereas the Oorja stove operates in batch mode, where a suffi-

cient quantity of fuel is loaded initially, so that a particular cook-
ing operation is completed. This was another reason for selecting
the Harsha and Oorja stoves, so that cooker can be tested with
both continuous and batch mode stoves.
4.1.3.1. Harsha Stove. The model was developed in 1990 at
the Technical Back-up Unit, Regional Research Laboratory
(CSIR), Bhubaneswar, Orissa. This stove is capable of burning
a variety of fuels containing 1
40% ash and high volatile materi-
als such as firewood, leaves, twigs, dung-cakes, agricultural waste,
raw coal, charcoal, etc. The stove is schematically shown in
Figure 18. It consists of a corrugated grate, ash scraper, perforated
combustion chamber, air jacket, and fuel feeding hole for flow of Figure 19. Oorja stove: (A) schematic, (1) combustion chamber,
air in primary and secondary streams. The combustion air is (2) solid fuel, (3) ash collector, (4) battery operated fan, (5) primary
preheated through the hot surfaces of the stove before it is drawn air, (6) secondary air, (7) annulus for flow of secondary air, (8) holes,
into the stove by natural draft during burning of the fuels, which (9) stand. (B) photograph.
occurs in two stages. Fixed carbon and volatile components of the
fuel burn in the corrugated grate and in the perforated combus- as preheated secondary air at the top of the bed, which comes
tion chamber of the stove, respectively, with the help of pre- through the annulus formed between the combustion chamber
heated primary and secondary air streams to obtain complete and the external wall of the stove. There are 18 holes of 5 mm dia-
combustion with low excess air. Fuel can be charged continu- meter for the emergence of secondary air into the fuel chamber at
ously through the fuel-feeding hole. The ash scraper helps to the top. To keep a pot on the stove, a stand is provided at the top
remove ash intermittently during burning of high ash fuels. of the stove. The Oorja stove is shown in panel (B) of Figure 19.
Otherwise, ash discharges automatically from the stove. A high 4.1.4. Experimental Setup and Procedure. The experimental
temperature clean flame (800 C) with reduced smoke is ob- setup is essentially similar to that in Figure 1 of Part I except the
tained during the combustion of solid fuel. gas burner was replaced by a stove burning solid fuel. To simulate
4.1.3.2. Oorja Stove. Panel (A) of Figure 19 shows the cooking performance, experiments were performed with water as
schematic of the Oorja stove. It consists of a combustion chamber the charge. Casuarina wood was tested as fuel for the Harsha
of an inner diameter of 108 mm and a height of 175 mm. stove, and biomass pellets were used as fuel for the Oorja stove.
A sufficient amount of solid fuel (biomass pellets) is placed inside Wood was cut into 50 mm  50 mm  500 mm pieces. Biomass
the combustion chamber initially for a particular cooking process. pellets used had a cylindrical shape, with 11 mm diameter and
The ash collector is inserted in the combustion chamber for the length varying up to 60 mm.
collection of ash. This stove works on the principle of forced draft, The 24 L cooker was used with the Harsha stove, and the 5 L
which is created by using a battery operated fan at the bottom of cooker was tested with the Oorja stove. The standard Harsha
the combustion chamber. The fan can be operated at different stove is too big for the 5 L cooker, and the standard Oorja stove is
speeds by using a voltage regulator. Although the forced draft is too small for the 24 L cooker, which is why different sizes of
upward, the flame front travels from the top down. cookers were used for the two different stoves. Four pots of 7 L
Air passes through the combustion chamber in two ways: first, were used with the 24 L cooker, and three pots of 1.5 L each were
as primary air through the bottom of the fuel bed, and second, used with the 5 L cooker.
1916 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Small wood pieces were kept over cotton waste soaked in Sunlight enters the oven through the glazing and heats the inside
kerosene in the combustion chamber of these stoves for ignition. black-colored high absorptivity and low emissivity surfaces.
After ignition in the Harsha stove, normal sized feed was placed Cooking pots are also painted on the outside with high absorp-
over the fire for burning. In the Oorja stove, the fuel was already tivity and low emissivity materials so as to absorb the solar
in place under the cotton waste. Once the emanating smoke energy. Glass’s opacity to the blackbody radiation emitted by the
stopped (after about 30
40 s in the case of both wood and contents and interior of the oven results in increasing of tem-
biomass pellets) and there was a steady flame, the cooker was perature inside the oven. This heat is utilized for cooking of the
placed on the stove. Thermocouple probes were inserted in the food items. The advantage of these cookers is simplicity of opera-
water in the base and in the pots to measure temperatures of the tion and economy, as they are affordable to common people.
charge inside. However, these cookers cannot be used to generate higher
The flame diameter in case of both wood and biomass pellets temperatures. The efficiency of the system is also low, as the
was about two-thirds of the base diameter of the cooker. Fuel was system cannot track the movement of the sun. In order to enha-
charged from time to time according to requirement for the nce the performance of the cooker, one or more of reflectors
Harsha stove, such that continuity in the flame was maintained. (boosters) need to be added. Danao and Joshi33 have described
Initial water temperatures in all cases were 30 ( 1 C. The cooker the pertinent aspects of box type cookers.
contents were heated until they reached near boiling point. In 4.2.2. Concentrating Type Solar Cookers. This type of solar
this case, temperatures of 98 ( 1 C for vessel V1, 97 ( 1 C for cooker consists of a concentrating reflector that focuses the
vessel V4, and 95 ( 1 C for vessels V2 and V3 were allowed to be incident solar rays at a particular point or locus of points. The
achieved, just like the end-of-heating condition when testing the receiver or heating target is kept at the focus and receives the rays.
cooker with LPG as a fuel. At this point, unburnt fuel was taken Usually, the receiver is coated with high absorptivity material.
out and immediately extinguished with a little quantity of water. The heat received is transferred to the material inside the receiver
The cooker was immediately opened to avoid condensation of that is heated. Higher temperatures can be achieved in these
the steam formed. Initial and final quantities of fuel as well as cookers, and the cooking time can be reduced. The temperatures
water in all the pots and the base were measured to a precision of are proportional to the ratio of the aperture of the concentrating
0.005 kg on an electronic balance to estimate the fuel consump- reflector to the focusing area. However, these type of cookers
tion and water loss due to evaporation during heating. Tempera- need tracking to follow the sun during the operating time interval
tures of the various components during heating were measured at so as to intercept the maximum amount of solar energy. Hence,
fixed time intervals to evaluate the heating performance of the their operation becomes relatively difficult and needs skill. Para-
water charge with a precision of 0.1 C. The overall thermal boloidal concentrators with a vertical cylindrical receiver have
efficiency was estimated using eq 1. been successfully used are very popular for the direct application
4.1.5. Results and Discussion. Calorific value for Casuarina of cooking. Danao and Joshi34 have discussed the design and
wood is 4425 kCal/kg and that for biomass pellets is 3900 kCal/kg. optimization features of this type of solar cookers.
4.1.5.1. Harsha Stove. The first experiment was carried out Wolfgang Scheffler is the inventor and promoter of Scheffler
using 50 mm  50 mm  500 mm sized fire wood pieces for Community Kitchens (large parabolic dishes for community
which the burn rate was found to 1.9 kg/h, which corresponds to kitchens, bakeries, etc.35). Over 2000 large cookers have been
63,000 kCal/h m2. With this flux, the efficiency of the cooker built, commissioned, and distributed worldwide by 2008. The
obtained is 31%. The optimized flux for the cooker is 21,000 small scale Scheffler dish-based cooking system consists of a dish
kCal/h,m2 when it is operated with LPG as the fuel. The (which is a segment of a paraboloid), a secondary reflector, and
obtained flux (63,000 kCal/h.m2) is much higher, so another the cooking vessel. Incident rays from the sun are reflected from
experiment was carried out with a reduced burn rate of 1.1 kg/h the Scheffler dish onto the secondary reflector that directs these
(38,000 kCal/h m2), for which the efficiency of the cooker was rays onto the bottom of the cooking pot. Community-scale
also 31%. cooking systems consist of a number of Scheffler dishes with their
4.1.5.2. Oorja Stove. The 5 L cooker was tested with the Oorja focus on the receiver where steam is generated. The steam is
stove, using biomass pellets as a fuel. The burn rate for the carried through a common header to the kitchen for use in the
experiment was 0.48 kg/h, which corresponds to a flux of 63,000 cooking process. During the past 80 years, several designs of
kCal/h m2. With this flux, the efficiency of the cooker was found cookers heve been proposed. The focus of development has been
to be 40%. to improve the energy efficiency and also to increase the cost
4.2. Solar Energy. During the past 80 years, several designs of effectiveness of the solar cooking systems.
cookers have been proposed. The focus of development has been In Section 3, LPG gas was employed as a fuel. In this case, a
to improve the energy efficiency and also to increase the cost detailed optimization exercise has indicated that the flame size
effectiveness of the solar cooking systems. Solar cooking has been should be one-third of the diameter of the cooker base. It was also
found to produce more nutritious food than is obtained by found that the optimum flux is about 25,000 kCal/h m2. It was
conventional methods and in the pressure cooker.32
34 There thought desirable to investigate the possibility of solar energy. An
are two types of solar cooking devices: oven type solar cookers attempt has been made to achieve the optimum value of heat flux
(solar box cooker) and concentrating type solar cookers. using solar energy. It may be noted that solar energy can be
4.2.1. Oven Type Solar Cookers. This type of cooker is focused on a concentrated hot spot geometrically similar to the
essentially an insulated box with sides blackened from the inside projection of the flame produce by stove (burning either LPG or
and a transparent glass cover on top. These are the closed units solid fuel).
working on the principle of the greenhouse effect. Glass is trans- The schematic arrangement of solar energy is shown in
parent to energy in the solar spectrum, which is mostly in the Figure 20. It consists of a primary reflector, secondary reflector,
visible and short wavelength infrared regions, and glass is opaque and cooker of 24 L size. During operation, there was a need for
to blackbody radiation, which is long wavelength infrared. tracking that was carried out every 10 min. For the 24 L cooker
1917 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 20. Schematic arrangement of a cooker using solar energy: (1) primary reflector, (2) secondary reflector, (3) cooker, (4) rays coming from sun,
(5) rays reflected from primary reflector, (6) rays reflected from secondary reflector, (7) hot spot at cooker bottom.

(base diameter = 400 mm), the optimum heat transfer rate is adsorption process in which the gelatinized starch is adsorbed by
calculated to be about 2,500 kCal/h or 2.9 kW. The average the rice grains.
efficiency of the reflectors was found to be 90%, and the efficiency The process developed by Miyagawa et al.38 claims to be
for absorbers was found to be 50
60%. Thus, the overall effi- adaptable for various kinds of raw rice and a variety of textures for
ciency is calculated to be in the range of 45
55%. Thus, for a the finally cooked rice. It essentially uses a conveyor belt
power requirement of 2.9 kW and a DNI value of 600 W/m2 ,the arrangement for moving the rice. The authors have used water-
area of the primary reflector would be about 10 m2. Further work soaked rice for experimentation. The system has three sections.
is in progress in regard to increasing the energy efficiency. In the first section, the presoaked rice was preheated to cooking
temperature. In section 2, rice was allowed to cook, and steam
supply was not needed. However, in section 3, some steam was
5. CONTINUOUS COOKING provided for the completion of cooking.
Most of the above processes require sophisticated equipment
5.1. Preamble. Section 3 shows that the energy required for with a steaming chamber and moving belt type mechanisms for
cooking is mainly needed to increase the temperature from rice handling. Further, in the previously proposed hardware, a
ambient to about 98 C. The actual energy required for the cook- scheme for the recovery of sensible enthalpy was not included.
ing reaction is a small fraction of the sensible enthalpy. Because Therefore, in the present work, an attempt has been made to
cooked food is available at 98 C and is consumed at tempera- eliminate both of these limitations.
tures below 50 C, a major portion of the sensible enthalpy can be 5.2. Equipment. The continuous cooker consists of two
recovered. A brief summary of the previous work on continuous sections: (a) cooking and (b) heat recovery. It is schematically
cooking will now be described. shown in Figure 21. Both sections consist of a U-trough with a
Various methods for cooking of grains (especially rice) on a single screw (primary screw) for moving the material in the
continuous basis have been patented. Sugimura36 developed a forward direction. The trough has a jacket for providing the heat
process for continuous cooking of rice by steam. This process through condensing steam. Heat can also be supplied by open
involves first heating of rice in a steamer thereby promoting its steam injected through nozzles. The speed of rotation of the
gelatinization, followed by heating in hot water, and again steam screw can be adjusted through a variable frequency drive (VFD)
cooking for a short time in a continuous manner. The process attached to a motor. Rice or lentils or a mixture of both are fed
claims to obtain full grain well-cooked free-flowing rice. through a hopper. The cooked material overflows from the trough
An automatic and continuous rice cooking system developed through an opening at the far end with an adjustable screw ar-
by Watanabe37 consists of a spiral rotary steam vessel with an elonga- rangement (secondary screw weir for maintaining the water level
ted steam jacket having a semicircular bottom, screw conveyer, inside the trough and for positive displacement of cooked material).
and steam tunnel passage provided with a mesh conveyer. The The cooked material falls into a heat recovery unit having prac-
overall process involves three steps: (a) steaming process in tically the same arrangement as the cooking section except that
which starch is released from the rice, (b) a gelatinization process the jacket is replaced by a limped coil. In addition, the steam nozzles
in which the starch from the rice is gelatinized, and (c) an are absent in the heat recovery section. Cooking experiments were
1918 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

Figure 21. Schematic of continuous cooking equipment: (1) U-trough, (2) single screw for moving material in forward direction, (3) jacket for
condensing steam, (4) nozzles for direct steam injection, (5) variable frequency drive, (6) motor, (7) hopper for addition of rice/lentils, (8) limped coil,
(9) rice/lentils outlet.

performed with a screw diameter of 108 mm and length of 1850 mm water was found to be 1:3. It was concluded from these trials that
with a central shaft of 33 mm diameter. Total length of the a residence time of 20 min was sufficient for cooking of rice.
trough is 2000 mm and width is 120 mm. The pitch of the screw The construction and the performance evaluation of the heat
is 40 mm for the initial 400 mm length and 80 mm for the rest of recovery unit is in progress. In one trial the flow rate of ambient
the screw length. As shown from the schematic (Figure 21), the water was 60 kg/h, ambient water enters at (30 ( 1 C), and the
trough extends above the screw for accommodating extra water if hot water leaves at about (83 C). The cooked rice leaves at
necessary. A rectangular plate with two handles on the top sur- about (50 C) and is ready for continuous packaging. The overall
face and hinges on one side are provided to cover the trough from heat balance indicates that 70% of the sensible enthalpy can be
above. The cover arrangement is handy during cleaning operation. recovered in the form of hot water, which in turn can be added to
Because the cooked rice was to be removed from the equip- the feed rice. In reality, the sensible enthalpy needed for cooking
ment in suspended form, initial trials were carried out to under- has been provided to increase the temperature of inlet water and
stand the effect of speed of rotation of the screw. A higher speed rice (∼30 C) to the outlet cooked rice temperature of 50 C.
would result in the breakage of cooked rice, while a very low Thus, the energy requirement compared with the efficient batch
speed can result in improper suspension. In the first trial, 30 kg of operation (Section 3) is about 45%. It may be further noted that
water was charged and brought to the boiling point, and then rice the batch operation reached 70% thermal efficiency compared
was charged from the feed end with screws rotating at 0.5 r/s. It with the range of 10
20% obtained in the conventional open
was observed that the rice grains moved gradually along the pan cooking. It is interesting to note that for the estimation of
length of the trough. A sample of rice water slurry was removed efficiency (eq 1), the denominator is the amount of enthalpy to
from the other end. From the samples collected, it was found that heat the material (rice and water) from 30 to 100 C. Because we
cooked rice started coming out after an average residence time of have recovered 70% of this heat, the energy needed in the
about 25 min. continuous process is about 12
15 times lower than the con-
Though the rice was cooked completely, there was large ventional open pan cooking. This is a significant result. Con-
breakage of grains indicating the need to lower the speed of sidering the enormous potential of a continuous cooker, further
rotation. In another trial, the rotation speed was decreased to work is in progress for the scale-up and development of a reliable
0.4 r/s with rice addition of 10 kg at the start, followed by 2 kg/h procedure for large scale cooking.
on a continuous basis. Samples of rice coming out from the
discharge end were continuously checked for cooking, and the
rice was found to be cooked satisfactorily During the final trial, 6. CONCLUSIONS
the speed of rotation was further reduced to 0.25 r/s. Ten (1) Efficient designs for cooking systems have been devel-
kilograms of rice was charged initially in heated water, and oped using LPG with sizes of 4, 6, 24, 30, 72, 120, 160, and
further addition was continued at the rate of 13 kg/h. Consider- 700 L, catering to the food needs for people in the range of
ing that rice requires water in the proportion of approximately 6
2000.
1:2.5 for cooking and some water is required for suspension, (2) The cooking pots gain energy from condensing steam on
water was charged at the rate of 71 kg/h. The operation was the outside surface. The cooking charge (water + rice or
carried out for 2 h. Rice coming out after the initial 20 min time lentils and/or vegetables) receive heat by the mode of
was found to be cooked nicely. The ratio of cooked rice to free natural convection. The flow and temperature pattern has
1919 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

been analyzed by computational fluid dynamics (CFD). development of suitable internals so the extent of strati-
The CFD prediction and experimental measurements indi- fication be reduced to permit cooking at all the locations.
cate that the optimum heat flux (on the basis of LPG This exercise will also allow for the maximization of the
burning) is in the range of 16,000
25,000 kCal/h m2, size of the cooking pot, and hence, reduce the number of
where m2 is the bottom surface area of the cooking system. pots required, which will lessen the drudgery for cooking
(3) The heat losses to the surroundings have been reduced by personnel.
providing air gap insulation between the covers. The flow (b) On the basis of mathematical modeling and CFD analysis,
pattern in the air gap has been comprehensively analyzed we need to find the flow and turbulent structures.40
53 It
by computational fluid dynamics. The optimization has is known that such information is useful for understand-
shown that the air gap should be in the range of 10
12 mm. ing the transport phenomenon in general and heat trans-
(4) The optimum gap between the burner top and the cooker fer in particular. The later is useful in the present work.
bottom was found to be 3 mm for all cooking systems (c) The flow pattern around the particle should also be
covered in this work. For instance, for a 6 L cooker, an investigated using computational fluid dynamics with
increase in the gap to 7, 12, and 15 mm reduces the ther- direct numerical simulation (DNS).54
mal efficiency by 14, 22, and 23%, respectively. Further, it (d) It is shown in Section 2.3 that the heat flux absorbed by
was also found that the effect of the gap was found to be the cooking pot diminishes as the temperature appro-
relatively less severe when the cooker size is increased. aches boiling point. For implementing this observation, a
Thus, an increase from a 3 to 6 mm gap results in a reduc- control strategy needs to be devised that shall be simple in
tion in thermal efficiency of 5% compared to about 10% construction and operation.
for a 6 L cooker. (e) In regard to use of solid fuels, substantial additional work
(5) The cooking reaction is practically thermally neutral. is needed to achieve desired burning rate over a sustained
Therefore, when the contents are heated to boiling tem- period. Rational procedures need to be developed for
peratures, the energy supply can be stopped by switching optimum design and scale-up.
off the fuel supply. After this point, the temperature was (f) Section 3 brought out the huge potential in continuous cook-
found to decrease by less than 5 C in 30 min. This is ing. Additional work is needed to attain free flow of solids until
because of the quality of insulation provided by the air gap outlet. We need to measure residence time distribution of
and the loss of contact between the water in the container solid phase for attaining near plug flow. An attempt needs
vessel and the bottom pot. Thus, after the fuel shut off, the made for optimization of energy utilization, which means
temperature of cooking material remains above the cook- partial cooking of cooked food against preheating of water.
ing temperature. (g) It will prove useful to undertake a systematic investigation
(6) The combined effect of all of the optimization exercises of the conventional cooking practices (as many as possible
has given a value of thermal efficiency in the range of from different countries) for measurement of thermal and
65
75% and useful efficiency in the range of 50
60%. useful efficiencies.
These efficiencies have been explained in eq 1 in Part I. It (h) In order to establish the amount of heat actually used for
is noted that conventional cooking methods give thermal the cooking reaction, systematic research work is needed
efficiencies in the range of 10
20%. for estimating the heat of the cooking reactions for the
(7) More than 50 cookers (72
160 L) are in actual practice variety of materials being cooked in practice.
at various locations. LPG consumption has found to be (i) The implementation of improved cooking systems devel-
reduced by about 60%. oped in this work is expected to have substantial impact on
(8) At present, the cost of a cooker is about U.S. $630, which the energy footprint of developing countries. Success will
caters the food for about 300 people. With 20% depecia- lie in the dissemination of knowledge and large scale
tion and 10% interest on capital cost, the pay out period implementations. For this purpose, concerted efforts are
works out to be eight months at a LPG price of 44 cents needed with strong collaboration between scientific in-
per kg. stitutions, NGOs, and governments.

’ AUTHOR INFORMATION
7. SUGGESTIONS FOR FUTURE WORK
(a) Simulations using computational fluid dynamics is dis- Corresponding Author
cussed in Section 2. The work needs to be extended for *Phone: +91 22 33611111 (J.B.J). Fax: +91 22 33611020 (J.B.J).
detailed analysis of temperature and flow patterns in E-mails: [email protected] (J.B.J), [email protected].
cooking pots. Investigations are needed to understand in (A.B.P.), [email protected] (S.B.P.).
the sensitivity of pot diameter, extent of filling, ratios of Present Addresses
rice/lentils/vegetables to water, and the extent of cooking. ^
AISSMS College of Engineering, Pune
411001, India
It may be pointed out that the porosity, average particle
size, and level of cooked material change with respect to
time. The CFD simulations need to take into account the ’ NOTATIONS
flow through porous media with variable thickness. Further, A = area of heat transfer (m2)
the extraction of starch in water may import viscous or non- AR = aspect ratio (H/L) (
)
Newtonian behavior to the liquid phase, and the descrip- A0 = pre-exponential constant in eq 14 (s
1)
tion of interface forces need to be described accordingly39. C = quantitative value of the product of degradation under
Detailed analysis needs to be carried out regarding tem- consideration (mol L
1)
perature stratification. This information should enable [C]0 = value of the product under consideration at time 0 (mol L
1)
1920 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922
Industrial & Engineering Chemistry Research ARTICLE

[C]t = value of the product under consideration after reaction α∞,1 = constant in Table 1
time t (mol L
1) α∞,2 = constant in Table 1
ΔC = amount degraded during ti (mol L
1) β = coefficient of thermal expansion (K
1)
CP = specific heat capacity (kJ kg
1 K
1) βr = constant in Table 1
CV = calorific value (kJ kg
1) β/∞ = constant in Table 1
D = diameter of pot/cooking vessel (m) F = density of fluid (kg m3)
Ea = activation energy of the reaction (J mol
1) F0 = density of fluid at reference temperature (kg m3)
fβ = constant as in Table 1 Δ = difference in quantity, e.g., temperature
Gb = generation of turbulence due to buoyancy (m4 s
1) ε = turbulent kinetic energy dissipation rate per unit mass
Gk = generation of turbulence kinetic energy due to mean (m2 s
3)
velocity gradients (kg m
1s3) λ = latent heat of vaporaization (kJ kg
1)
Gω = production of ω μ = viscosity (m2 s
1)
g = gravitational constant (m s
2) μt = turbulent viscosity (m2 s
1)
H = height of the pot/Cooking vessel (m) μeff = effective viscosity of fluid (m2 s
1)
HE = height of the enclosure (m) νt = turbulent kinematic viscosity (m2 s
1)
hi = inside heat transfer coefficient (W m
2 K
1) ω = specific dissipation rate (s
1)
hins = time-averaged inside heat transfer coefficient (W m
2 K
1) σk = turbulent Prandtl number for energy dissipation rate (
)
ho = outside heat transfer coefficient (W m
2 K
1) σt = turbulent Prandtl number for kinetic energy (
)
I = turbulence intensity (%) χk = constant in Table 1
k = turbulent kinetic energy per unit mass (m2 s
2) η = efficiency
kt = thermal conductivity (W m
1 K
1)
k1 = reaction rate constant (s
1) Subscripts
ki = rate constant at time ti (s
1) B = quantity in base
L = gap width of enclosure (m) C = contents
M = mass (kg) G = LPG gas
m = quantity of steam evaporated (kg) i = inner diameter
n = order of the reaction lm = logarithmic mean
Nu = Nusselt number (
) o = outer diameter
p = pressure (Pa) therm = thermal
R = ideal gas constant (Jmol
1 K
1) use = useful
Ra = Rayleigh number (gβΔTW3/να) (
) V = vessel
S = stratification number (
) W = water

1
|S| = mean strain rate (s ) Abbreviations

1
Sij = strain rate (s ) CFD = computational fluid dynamics
T = temperature (K)
DNS = direct numerical simulation
TC = temperature of the outer cover (K)
SS = stainless steel
TH = temperature of inner cover (K)
T0 = reference temperature (K)
Tmax = maximum temperature of fluid (K) ’ REFERENCES
Tmin = minimum temperature of fluid (K) (1) Legros, G.; Havet, I.; Bruce, N.; Bonjour, S. The Energy Access
t = time (s) Situation in Developing Countries. WHO and UNDP, 2009. http://
ÆΔTæ = average temperature driving force (K) www.who.int/indoorair/publications/energyaccesssituation/en/index.
U = overall heat transfer coefficient (W m
2 K
1) html (accessed January 3, 2012).
ur = mean radial velocity (m s
1) (2) Wikipedia. http://wiki.answers.com/Q/What_percentage_of_
uz = mean axial velocity (m s
1) the_world's_population_lives_in_developing_countries (accessed January
Æuæ = time averaged mean velocity (m s
1) 3, 2012).
uθ = mean annular velocity (m s
1) (3) Ganguli, A. A.; Pandit, A. B.; Joshi, J. B. CFD simulation of heat
u = horizontal velocity (m.s
1) transfer in a two-dimensional vertical enclosure. Chem. Eng. Res. Des.
2009, 87, 711–727.
V1 = standard pot/vessel (4) Ganguli, A. A.; Sathe, M. J.; Pandit, A., B.; Joshi, J. B.; Vijayan,
V2 = tall pot/vessel P. K. Hydrodynamics and heat transfer characteristics of passive decay
V3 = standard pot/vessel heat removal systems: CFD simulations and experimental measure-
V4 = tall pot/vessel ments. Chem. Eng. Sci. 2010, 65, 3457–3473.
Ven = enclosed volume (m3) (5) Reddy, R. K.; Joshi, J. B. CFD modeling of solid-liquid fluidized
X = amount of rice (kg) beds of mono and binary particle mixtures. Chem. Eng. Sci. 2009, 64,
x = abscissa (m) 3641–3658.
Y = ratio of water to rice (kg/kg) (6) Batchelor, G. K. Heat transfer by free convection across a closed
Yk = dissipation of turbulent kinetic energy (s
1) cavity between vertical boundaries at different temperatures. Q. Appl.
Math. 1954, 12, 209–233.
y = ordinate (m) (7) Elder, J. W. Laminar free convection in a vertical slot. J. Fluid
Mech. 1965, 23, 77–98.
Greek Symbols (8) Yin, S. H.; Wung, T. Y.; Chen, K. Natural convection in an air
α= thermal diffusivity (m2 s
1) layer enclosed within rectangular cavities. Int. J. Heat Mass Transfer 1978,
αeff = effective thermal diffusivity (m2 s
1) 21, 307–315.

1921 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922

Industrial & Engineering Chemistry Research ARTICLE

(9) Elsherbiny, S. M.; Raithby, G. D.; Hollands, K. G. T. Heat (35) Scheffler, W. Introduction to the Revolutionary Design of
transfer by natural convection across vertical and inclined air layers. Scheffler Reflectors. Solar Cooker International Conference, 2006.
Trans. ASME, J. Heat Transfer 1982, 104, 96–102. http://www.docstoc.com/docs/100239025/Introduction-to-the-Revo-
(10) Wakitani, S. Experiments on convective instability of large lutionary-Design-of-Scheffler-Reflectors (accessed January 4, 2012).
Prandtl number fluids in a vertical slot. J. Fluid Mech. 1994, 116, 120–126. (36) Sugimura, T. Process for Continuous Rice Cooking by Steam-
(11) Newell, M. E.; Schmidt, F. W. Heat transfer by laminar natural ing and Apparatus Thereof. U.S. Patent 4571341, 1986.
convection within rectangular enclosures. Trans. ASME, Ser. C 1970, 92, (37) Watanabe, K. Automatic and Continuous Rice Cooking Sys-
159–167. tem. U.S. Patent 4934259, 1990.
(12) Korpela, S. A.; Lee, Y.; Drummond, J. E. Heat transfer through a (38) Miyagawa, T.; Ishii, Y.; Tanaka, T. Method and apparatus for
double pane window. Trans ASME, J. Heat Transfer 1982, 104, 539–544. continuously steaming and boiling rice. U.S. Patent 6056986, 2000.
(13) Lee, Y.; Korpela, S. Multicellular natural convection in a vertical (39) Lali, A. M.; Khare, a.S.; Joshi, J. B.; Nigam, K. D. P. Behavior of
slot. J. Fluid Mech. 1983, 126, 91–124. solid particles in viscous non Newtonian solutions: settling velocty, Wall
(14) Le Quere, P. A note on multiple and unsteady solutions in two- effects and bed expansion in solid liquid fluidized beds. Powder Technol.
dimensional convection in a tall cavity. Trans. ASME, J. Heat Transfer 1989, 57, 39–50.
1990, 112, 965–973. (40) Joshi, J. B.; Ranade, V. V.; Gharat, S. D.; Lele, S. S. Sparged loop
(15) Wakitani, S. Formation of cells in natural convection in a reactors. Can. J. Chem. Eng. 1990, 68, 705–741.
vertical slot at large Prandtl number. J. Fluid Mech. 1996, 314, 299–314. (41) Joshi, J. B. Computational flow modeling and design of bubble
(16) Wakitani, S. Development of multicellular solutions in natural column reactors. Chem. Eng. Sci. 2001, 5, 5893–5933.
convection in an air-filled vertical cavity. Trans. ASME, J. Heat Transfer (42) Tabib, M. V.; Roy, S. V.; Joshi, J. B. CFD simulation of bubble
1997, 119, 97–101. columns: an analysis of interface force and turbulence models. Chem.
(17) Zhao, Y.; Curcija, D.; Gross, W. P. Prediction of multicellular Eng. J. 2008, 139, 589–614.
flow regime of natural convection in fenestration glazing cavities. (43) Bhole, M. R.; Joshi, J. B.; Ramakrishna, D. CFD simulation of
ASHRAE Trans. 1997, 103, 1–12. bubble columns incorporating population balance modeling. Chem. Eng.
(18) Lartigue, B.; Lorente, S.; Bourret, B. Multicellular natural Sci. 2008, 63, 2267–2282.
convection in high aspect ratio cavity: Experimental and numerical (44) Dijkhuizen, W.; Roghair, I.; Van Sint Annaland, M.; Kuipers,
results. Int. J. Heat Mass Transfer 2000, 43, 3157–3170. J. A. M. DNS of gas bubbles behaviour using an improved 3D front tracking
(19) Ekambara, K.; Joshi, J. B. CFD Simulation of flow in bubble model: Model development. Chem. Eng. Sci. 2010, 65, 1427–1437.
columns: Laminar and transition regimes. Ind. Eng. Chem. Res. 2005, 44, (45) Dijkhuizen, W.; Van Sint Annaland, M.; Kuipers, J. A. M.
1413–1423. Numerical and experimental investigation of the lift force on single
(20) Thakre, S. S.; Joshi, J. B. CFD simulation of bubble column bubbles. Chem. Eng. Sci. 2010, 65, 1274–1287.
reactor: Importance of drag force formulation. Chem. Eng. Sci. 2001, 56, (46) Joshi, J. B.; Sharma, M. M. Mass transfer characteristic of
5893–5933. horizontal sparged contactors. Trans. Inst. Chem. Eng. 1976, 54, 42–53.
(21) Incropera, F. P.; Dewitt, D. P. Fundamentals of Heat and Mass (47) Joshi, J. B.; Sharma, M. M. Liquid phase backmixing in sparged
Transfer, 5th ed.; Wiley & Sons: NewYork, 2005. contactors. Can. J. Chem. Eng. 1978, 56, 116–119.
(22) User’s Manual to FLUENT 6.2; Fluent, Inc.: Centrera Resource (48) Joshi, J. B.; Vitankar, V. S.; Kulkarni, A. A.; Dhotre, M. T.;
Park, Lebanon, NH, 2005. Ekambara, K. Coherent flow structures in bubble column reactors.
(23) Gandhi, M.; Sathe, M. J.; Joshi, J. B.; Vijayan, P. K. Two phase Chem. Eng. Sci. 2002, 57, 3157–3183.
natural convection: CFD simulations and PIV measurement. Chem. Eng. (49) Mathpati, C. S.; Joshi, J. B. Insight into theories of heat and mass
Sci. 2011, 66, 3152–3171. transfer at solid
fluid interfaces using direct numerical simulation and
(24) Aszodi, A.; Krepper, E.; Presser, H. M. Investigation of Heating large eddy simulation. Ind. Eng. Chem. Res. 2007, 46, 8525–8557.
up and Evaporation Processes of Fluids in Storage Tanks by Experiments and (50) Tabib, M. V.; Joshi, J. B. Analysis of dominant flow structures
by Numerical Simulation. Second International Symposium on Two and their flow dynamics in industrially relevant equipment using proper
Phase Flow Modeling and Experimentation, Pisa Proceedings, 2000, pp orthogonal decomposition. Chem. Eng. Sci. 2008, 63, 3695–3715.
1667
1674. (51) Deshpande, S. S.; Joshi, J. B.; Kulkarni, V. R.; Kulkarni, B. D.
(25) Kang, M. Thermal mixing in a water tank during heating Identification and characterization of flow structures in chemical process
process. Int. J. Heat Mass Transfer 2002, 45, 4361–4366. equipment using multi-resolution techniques. Chem. Eng. Sci. 2008, 63,
(26) Krepper, E.; Hicken, E. F.; Jaegers, H. Investigation of natural 5330–5346.
convection in large pools. Int. J. Heat Fluid Flow 2002, 23, 359–365. (52) Joshi, J. B.; Tabib, M. V.; Deshpande, S. S.; Mathpati, C. S.
(27) Calcagni, B.; Marshili, F.; Paroncini, M. Natural convective heat Dynamics of flow structures and transport phenomena. I. Experimental
transfer in square enclosures heated from below. Appl. Therm. Eng. 2005, and numerical techniques for identification and energy content of flow
25, 2552–2531. structures. Ind. Eng. Chem. Res. 2009, 48, 8244–8284.
(28) Sharma, A., K.; Velusamy, K.; Balaji, C. Turbulent natural (53) Mathpati, C. S.; Tabib, M. V.; Deshpande, S. S.; Joshi, J. B.
convection in an enclosure with localized heating from below. Int. J. Dynamics of flow structures and transport phenomena. II. Relationship
Heat Mass Transfer 2007, 46, 1232–1241. with design objectives and design optimization. Ind. Eng. Chem. Res.
(29) Han, Y. M.; Wang, R. Z.; Dai, Y. J. Thermal stratification 2009, 48, 8285–8311.
within the water tank. Renewable Sustainable Energy Rev. 2009, 13, (54) Reddy, R. K.; Jin, S.; Nandakumar, K.; Minev, P. D.; Joshi, J. B.
1014–1026. Direct numerical simulation of free falling spheres in creeping flow. Int. J.
(30) Reddy, R. K.; Joshi., J. B. CFD modeling of pressure drop and Comp. Fluid Dyn. 2010, 24, 109–120.
drag coefficient in fixed beds: Wall effects. Particuology 2010, 8, 37–43.
(31) BIS Sandard: IS 5116, 1996.
(32) George, R. Solar Cooking for Healthy Living- Solar Cooker: A
Device for Nutritious Food; Faculty of Home Science, The M.S. University
of Baroda, Vadodara, India, 2001.
(33) Danao, S. P.; Joshi, J. B. Solar Cooking: Optimisation of
Paraboloidal Concentrator. Proc. Indian Natl. Sci Acad. 2011, 77 (4),
373–388.
(34) Danao, S. P.; Joshi, J. B. Optimisation of Solar Box Cooker. Proc.
Indian Natl. Sci Acad. 2011, 77 (1), 1–18.

1922 dx.doi.org/10.1021/ie2025745 |Ind. Eng. Chem. Res. 2012, 51, 1897–1922