Applied Energy 112 (2013) 918–927

Contents lists available at SciVerse ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Optimisation of the energy efficiency of bread-baking ovens

using a combined experimental and computational approach
Zinedine Khatir a,⇑, Joe Paton a, Harvey Thompson a, Nik Kapur a, Vassili Toropov a,b
a
School of Mechanical Engineering, University of Leeds, Leeds LS2 9JT, United Kingdom
b
School of Civil Engineering, University of Leeds, Leeds LS2 9JT, United Kingdom

h i g h l i g h t s

" A scientific framework for optimising oven operating conditions is presented.

" Experiments measuring local convective heat transfer coefficient are undertaken.
" An energy efficiency model is developed with experimentally calibrated CFD analysis.
" Designing ovens with optimum heat transfer coefficients reduces energy use.
" Results demonstrate a strong case to design and manufacture energy optimised ovens.

a r t i c l e i n f o a b s t r a c t

Article history: Changing legislation and rising energy costs are bringing the need for efficient baking processes into
Received 24 September 2012 much sharper focus. High-speed air impingement bread-baking ovens are complex systems using air flow
Received in revised form 19 December 2012 to transfer heat to the product. In this paper, computational fluid dynamics (CFD) is combined with
Accepted 12 February 2013
experimental analysis to develop a rigorous scientific framework for the rapid generation of forced con-
Available online 13 March 2013
vection oven designs. A design parameterisation of a three-dimensional generic oven model is carried out
for a wide range of oven sizes and flow conditions to optimise desirable features such as temperature uni-
Keywords:
formity throughout the oven, energy efficiency and manufacturability. Coupled with the computational
Energy efficiency
Industrial bread-baking
model, a series of experiments measuring the local convective heat transfer coefficient (hc) are under-
High-fidelity computational fluid dynamics taken. The facility used for the heat transfer experiments is representative of a scaled-down production
Oven design oven where the air temperature and velocity as well as important physical constraints such as nozzle
Experimentation dimensions and nozzle-to-surface distance can be varied. An efficient energy model is developed using
a CFD analysis calibrated using experimentally determined inputs. Results from a range of oven designs
are presented together with ensuing energy usage and savings.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction been rising significantly over the past 20 years, but international
legislation is forcing manufacturers to reduce their carbon foot-
Bread is a mass-manufactured food product that is widely print in order to mitigate for climate change fears. These factors
consumed across the globe. Baking is a key step for production of are encouraging greater understanding of high-energy processes
bread on an industrial scale. The process of cooking dough, which and the introduction of ‘‘green’’ technologies into parts of industry
involves complex physical and chemical processes, is completed in where energy consumption has historically never been a major
the oven. This process ultimately determines many of the final concern. Annual energy use in the UK baking industry is currently
physical properties of bread, such as crust colour, crumb texture 2 TW h and produces of over 570,000 tonnes of CO2 per year. Half
and taste [1]. of the energy use in a bakery is in the oven [2,3], with other signif-
Commercial ovens have been designed in order to maximize icant contributors including mixers, provers and coolers. Of the
throughput and product quality. However, in recent years thermal energy supplied to the oven, waste heat accounts for
minimising energy use has been playing a larger part in the 46% of the total heat supplied [4,5]. Reducing this proportion of
thought process behind oven designs. Not only have energy costs waste heat is of paramount importance to bakeries and oven man-
ufacturers for both economic and environmental reasons.
Two types of oven are commonly used in the commercial bak-
⇑ Corresponding author. Tel.: +44 (0) 113 343 2197; fax: +44 (0) 113 234 9679. ing industry: indirect fired ovens and forced convection (direct
E-mail addresses: Z.Khatir@leeds.ac.uk, zkhatir@hotmail.com (Z. Khatir). fired) ovens. Forced convection ovens are the focus of the present

0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.apenergy.2013.02.034
 Z. Khatir et al. / Applied Energy 112 (2013) 918–927 919

Nomenclature

Abbreviation Q volumetric flow rate (m3/s)

CFD computational fluid dynamics Re Reynolds number (Re = unozd/m)
BC boundary condition s nozzle-to-nozzle spacing (m)
t time (s)
Symbols T temperature (K)
cp specific heat capacity (J/(kg K)) unoz nozzle velocity (m/s)
d nozzle jet diameter (m) Vin velocity inlet condition (m/s)
f relative orifice area x distance from centre of nozzle (m)
H nozzle-to-surface distance (m) a degree of gelatinisation
H/d dimensionless nozzle-to-surface distance e emissivity
hc heat transfer coefficient (W/(m2 K)) q density (kg/m3)
I turbulence intensity m kinematic viscosity (m2/s)
L characteristic length (m) r Stefan–Boltzmann constant (r = 5.67  108 W m2 K4)
k thermal conductivity (W/(m K)) s thermal diffusivity (m2/s) (s = k/(qcp))
Nu Nusselt number (Nu = hcd/k)
P power (kW) Subscript
p pressure (Pa) c convective
Pr Prandtl number (Pr = m/s) r radiative
q heat flux (W/m2) T total

study and are based on air jet impingement technologies that have Oven design, which is inextricably linked to the heat transfer coef-
been extensively used in the drying industry [6]. Within forced ficient, determines the baking time. Comparing the impact of both
convection baking ovens, heat is transferred to the dough via con- the bake time and heat transfer coefficient on energy use in terms
vection and radiation and the relative importance of each mode is of fans used to distribute the air within the oven (electricity) and
determined by the air temperature and velocity of the impinging heat used to cook the product (gas), enables an optimum value
air jets. Heat is then transferred through to the centre of the bread for heat transfer coefficient to be found. This can then be experi-
to cook the dough via conduction. mentally verified and can be used to drive oven design for energy
Experimental studies on jet impingement heat transfer are savings. Previous studies [4,9,15] have shown how quality assess-
numerous and detailed, most notably the seminal paper by Martin ments can be made based on temperature uniformity inside the
[7], who published heat transfer correlations for a number of dif- oven allowing the CFD model and the baking model (predicting
ferent types of nozzles and arrays of nozzles. Despite the detailed bake time) to be decoupled. Temperature uniformity can be further
studies in existence, it is impossible to know the exact value of confirmed by the CFD results laid out in Section 3.3. The prediction
the convective heat transfer coefficient (hc) for each scenario and of the bake time model accounts for convective heat transfer (i.e.
set of conditions. Thus, in this study the local heat transfer coeffi- convective heat transfer coefficient) and the temperature condi-
cient was measured experimentally for the specific oven operating tions around the surface of the bread, taken as uniform.
conditions. Computational fluid dynamics (CFD) is increasingly
being used to predict airflows for many applications in the food
2.1. Experimental apparatus
industry [8]. CFD techniques have been successfully used to predict
air flow and temperature distribution within baking ovens [9] and
The experimental rig depicted in Fig. 2 shows banks of nozzles
to optimise temperature uniformity at the bread surface for a bak-
that distribute hot air through nozzle plates. The total size of the
ing regime for energy efficiency [10].
pilot oven is 3 m wide, 4 m long and 2 m high. The rig is fully
Previous CFD studies of bread ovens have analysed how to: alter
adjustable, such that values such as nozzle-to-sample spacing
the temperature profile along the length of the oven to reduce mois-
(H), air velocity (unoz) and temperature (T) can be controlled accu-
ture loss [11,12], optimise temperature, heat transfer coefficient
rately. The nozzle plates can be changed to provide variations in
and bread radius (i.e. dough shape) for improving product quality
geometry for both slot and round orifice nozzles. Air can be sup-
[13], and used multi-objective optimisation and control vector
plied above and/or below the product through control of dampers
parameterisation approaches to optimise product quality for ratios
located in the supply ducting. A large number of experiments were
of different heat sources [14]. Previous numerical studies have esti-
conducted measuring heat transfer characteristics of jet impinge-
mated values for hc or quoted values from previously published
ment nozzles using over 50 different regimes relevant to the bak-
work in order to compute heat flux. This paper aims to provide
ing industry. From these experiments, a single set of conditions
the first system level bread baking optimisation methodology
was chosen that best reflects the heat transfer characteristics of
which accounts for all major energy flows and which encompasses
bread baking. The results of this experiment were used to inform
CFD-based analyses with experimentally-validated heat transfer
a CFD model.
coefficients. This model will then be used in conjunction with exper-
For this study, round orifices were used as these are most com-
imental validation to provide a framework for process optimisation.
monly used in commercial baking ovens. The dimensions are:
diameter, d = 0.012 m and nozzle-to-nozzle spacing, s = 0.2 m. A
2. Materials and methods commercially available sensor is used to measure heat transfer –
an RC01 heat flux sensor (Hukseflux), which measures both con-
Fig. 1 shows how the components of the analysis link together. vective heat flux (gold surface) and total heat flux (black surface),
A realistic value for the heat transfer coefficient feeds into a CFD allowing radiative heat flux to be inferred. Ambient air tempera-
model which predicts the temperature and flow field distributions. ture is measured with a K-type thermocouple, which has an oper-
920 Z. Khatir et al. / Applied Energy 112 (2013) 918–927

Fig. 1. Process diagram showing how energy savings link with oven design and computational analysis.

Recirculation
fan
Ambient
Heated
airflow
airflow

Burner
Nozzles

Heat flux sensor

Fig. 2. Sampling oven test rig.

ating range of 50–300 °C and has an accuracy of ±0.3% + 2 °C in the of flow openings and walls is solved numerically [15,17,19]. For
operating range used for this study. Nozzle exit velocity is mea- flow openings a combination of velocity inlet and pressure gradient
sured through use of a P200UL digital manometer and pitot tube boundary conditions are used, with temperature and convective
(Digitron Instrumentation Ltd.) and verified by a static pressure heat transfer boundary conditions defined along the walls (see Ta-
gauge readout mounted on the oven. ble 1). Turbulence intensity and length scale at the nozzle jets and
The sensor traversed the flow field of two complete banks of outlet boundaries are set to be 2% and 0.07D respectively [19].
nozzles. The heat flux sensor is mounted flush within a flat sheet The CFD mesh is generated using the blockMesh mesh genera-
of aluminium. The sensor is manoeuvred in and out of the oven tion utility of OpenFOAM [18]. Grid independency analysis is
by a reciprocating arm which can be set to a desired speed. undertaken for both grid cell numbers and grid distributions. A fi-
ner mesh is defined near the bottom wall such that y+ is below
unity for all cases to suitably resolve the near wall region with
2.2. Computational fluid dynamics methodology
large gradients. The grid generated for H/d = 6 and d = 12 mm is
outlined in Fig. 5. Grid independence is achieved with around
Following Khatir et al. [9,10,15], air flows in the oven are ana-
1.75 million cells.
lysed in three dimensions using the steady-state Navier–Stokes
equations solved using the SIMPLE algorithm [16]. Turbulence is
modelled using the realizable k–e transport model [9–11,17]. The 2.3. Energy analysis
continuity, momentum and turbulence transport equations are
solved computationally using OpenFOAM [18]. The bake time-heat transfer coefficient trend predicted by the
Following previous related studies, the flow field within the CFD model will relate to energy use by the oven. Whilst a faster
solution domain considered in Fig. 3 and shown in Fig. 4, composed bake allows throughput to increase, and thus thermal efficiency
 Z. Khatir et al. / Applied Energy 112 (2013) 918–927 921

0.01
12 m

0.6 m

0.02
22 m

s=0.2
2m

Fig. 3. Top view of the perforated plate showing the nozzle dimensions being considered here.

Fig. 6 shows the experimental results of hc as a function of Re,

where unoz was changed to vary Nu. The experimental results are
compared with the correlation developed by Martin for arrays of
round nozzles (ARNs) [7]. The experimental results match well
with heat transfer correlations (within 12%). Some discrepancy is
always expected due to the nature of differing experimental condi-
tions – Martin’s correlations were conducted using a contoured
nozzle exit, whereas in practical baking applications, and the
experiments outlined here, sharp edged orifices are used.
The thermal energy efficiency saving can be calculated by
assuming that approximately 19% of oven heat is lost to ambient,
using the methodology described in detail by Paton et al. [4,5].
Fig. 4. 3D view of the solution domain for the CFD analyses. Therefore for a faster bake time the specific energy loss (kJ of en-
ergy per kg of product produced) is reduced linearly with bake
time. By balancing these two factors, the optimum convective heat
Table 1 transfer coefficient for energy savings can be calculated.
Summary of boundary conditions.

Modelled Inlet Left/right Wall 3. Results and discussion

equation patch
Cells 624 3000 24,000 3.1. Experimental measurement of heat transfer coefficient
Energy T = 513 K T = 437 (Top) T = 437 K (bottom)
convective BC where hc to
The heat transfer coefficient and air jet velocities can be repre-
be specified
Momentum Vin = unoz Zero pressure No-slip wall function sented in non-dimensional form using the Nusselt number and
gradient Reynolds number respectively defined as Nu = (hcd)/k and
Turbulence I = 2% Zero gradient Zero wall function Re = (unozd)/m respectively. Using the set of conditions previously
intensity
listed, unoz = 19.87 m/s, H/d = 6, T = 437 K, the local heat transfer
coefficient distribution is shown by Fig. 7, where interaction be-
tween sets of nozzles is minimal due to the relatively large noz-
to increase, the higher air velocity required to increase convective
zle-to-nozzle spacing. The average value for hc was found to be
heat transfer results in a larger electricity load to power fans to dis-
22 W/(m2 K) (and is inferred from the Nusselt number in Eq. (2)).
tribute the air. The energy required to operate the fans can be esti-
The experimentally determined values of hc, unoz, H/d and T are
mated using:
used for boundary conditions in the CFD model.
P ¼ Dp  Q ð1Þ
3.2. Experimental analysis: convection vs. radiation
Eq. (1) gives the power required in kW to create the pressure re-
quired to give the flow rate, Q, of air necessary for a certain veloc-
Further experimentation on the heat profile, Fig. 8, has shown
ity. The required relationship between velocity and heat transfer
that the proportion of the total heat flux in the regime studied that
coefficient was correlated by Martin [7] in dimensionless form
is attributable to convection is approximately 93% and that due to
using Nusselt number and Reynolds number:
radiation is 7%. This matches well with the value previously found
pffiffiffi
Nu pffiffiffi 1  2:2 f 2 for the baking industry by Earle [20] and the following analysis
0:42
¼ KðH=d; f Þ  f pffiffiffi Re3 ð2Þ [21]. Following the Stefan–Boltzmann law and the radiant energy
Pr 1 þ 0:2ðH=d  6Þ f
exchange between the oven walls and the bread within a baking
where f represents the free area of the bank of nozzles. chamber, radiation heat transfer is calculated by:
922 Z. Khatir et al. / Applied Energy 112 (2013) 918–927

Fig. 5. Grid generated in the x–z plane x–z plane for H/d = 6 with d = 12 mm and 0 6 x 6 0.6 m using 1.753 million cells.

80

70 Martin (1977)

60 Experimental results

50
Nu/Pr 0.42

40

30

20

10

0
4000 6000 8000 10000 12000 14000 16000
Re

Fig. 6. Experimental data together with Martin’s correlation for arrays of round nozzles (ARNs) [7].

Fig. 7. Heat transfer coefficient distribution for a round nozzle jet in an array, where x/d = 0 represents the centre of a nozzle and x/d = ±8.33 is the midpoint between two sets
of nozzles.

qr ¼ reðT 1  T 2 Þ4 ð3Þ e ¼ 1=ðð1=e1 þ 1=e2 Þ  1Þ ð4Þ

where T1 and T2 are the oven and bread and temperatures respec- e1 and e2 denote the emissivity of the bread and the stainless
tively equal to 244 °C and 230 °C, r denotes the Stefan–Boltzmann steel oven walls, equal to 0.06 and 0.8.
constant which has value of 5.67  108 W m2 K4and e the emis- Using the experimentally measured convective heat transfer
sivity approximated as: coefficient, hc = 22 W/(m2 K), the convective heat transfer is equiv-
alent to:
 Z. Khatir et al. / Applied Energy 112 (2013) 918–927 923

Fig. 8. Measurement of the proportion of convective and radiative heat flux for a nozzle jet in an array, where x/d = 0 represents the centre of a nozzle and x/d = ±8.33 is the
midpoint between two sets of nozzles.

Fig. 9. CFD predictions demonstrating air flow field complexity through: (a) pathlines coloured by velocity magnitude and (b) front view of velocity contour plots at the plane
y = 0.011. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

qc ¼ hc ðT 1  T 2 Þ ð5Þ Fig. 8 shows that underneath the nozzle exit, total heat flux is
up to three times larger than away from the nozzles. This distribu-
Eventually, qT = qr + qc = 332.9 W/m2 with qr = 24.9 W/m2 and tion is similar to that shown by Mohanty and Tawfek for similar
qc = 308.0 W/m2, i.e. around 7.5% radiation. These results are in nozzle conditions [22]. As expected, the radiative heat flux remains
excellent agreement with the experimental data, and justify the relatively constant with local total heat flux fluctuations across the
decision to neglect radiation in the following CFD analysis. nozzle profile.
924 Z. Khatir et al. / Applied Energy 112 (2013) 918–927

Fig. 10. CFD predictions highlighting the high degree of temperature uniformity within the baking chamber through: (a) 3D view of temperature contour plots at plane z = 0,
at various x-planes (x = 0.1, x = 0.3 and x = 0.5) and (b) 2D view of temperature contour plots at the y-plane, y = 0.011.

3.3. Computational fluid dynamics results

Table 2
Temperature-dependent bread properties [24].
The complexity of the thermal air flow field within such high-
speed air impingement ovens is depicted by Fig. 9a where multiple Temperature (°C) Heat capacity, Thermal conductivity,
recirculation and vortical structures are clearly demonstrated by cp (J/kg K) k (W/m K)
pathlines coloured by velocity magnitude. This is further empha- 28 3080.0 0.85
sized through the front view of the velocity contour plots at the 60 2550.6 0.38
120 1774.3 0.17
plane y = 0.011 in Fig. 9b.
227 1514.1 0.16
The high degree of temperature uniformity as well as thermal
air flow structures within the baking chamber is obtained and con-
firmed by Fig. 10, at plane z = 0, at various x-planes (x = 0.1, x = 0.3
and x = 0.5) and the y-plane y = 0.011. These contour plots indicate
a small temperature variation of 0.04 K (i.e. between 436.96 K and for any points (x, y, z) in the domain D=[0, 0.25]  [0, 0.10]  [0, 0.12]
437 K). This is an important aspect of the baking process as this with the following initial  conditions, T(x, y, z, t = 0) = 39 °C and
will allow the bread to cook efficiently [23]. boundary conditions, k@T  ¼ hc ½Tð^
x; tÞ  T 1  where T1 = 513 K
@x ^
x
and ^x are points (x, y, z) on the boundary oD defined by the six planes
3.4. Bake time model bounding the rectangular computational domain D. Eq. (6) is solved
numerically using COMSOL [29].
The density of the bread is assumed to be constant with value Calculated cooking times for various values of the convective
q = 330 kg/m3 and temperature dependent thermal diffusivity heat transfer coefficient hc are summarized in Figs. 11 and 12. Note
a(T) = k(T)/(qcp(T)), where k(T) and cp(T) are listed in Table 2 [24], the asymptote in Fig. 12 at a cooking time of around 20 min, dem-
with a loaf of bread of 0.25 m length, 0.10 m width and 0.12 m onstrating that there is a minimum cooking time due to diffusion
height (i.e. mass of a loaf of bread of about 1 kg). It is also assumed limitations for any particular bread which depends on its proper-
that bread is cooked when its core temperature reaches 94 °C [13]. ties and dimensions (see Table 2). This shows that modifications
Complex models predicting the baking process have been to bread formulations which increase its thermal conductivity dur-
developed previously by other authors [25–27], however in this ing baking have the potential to improve energy efficiencies
work a simplified heat transfer model is used. This helps to illus- substantially.
trate the scope of potential energy savings within the bread-baking Fig. 13 shows the temperature distribution within the bread
industry. at an early stage and at the end of the baking process with a
The temperature inside the bread is modelled in three dimen- convective heat transfer coefficient hc equal to 20 W/(m2 K). As
sions by the heat equation [28]: expected the closer the bread surface temperature is to that of
surrounding hot air, the quicker the bread cooks with the tem-
1 @T
r2 Tðx; y; zÞ ð6Þ perature gradually increasing until the core-temperature reaches
aðTÞ @t 367 K (94 °C).
 Z. Khatir et al. / Applied Energy 112 (2013) 918–927 925

Fig. 11. Bread-core temperature vs. time for various heat transfer coefficient hc.

Fig. 12. Cooking time as a function of heat transfer coefficient, hc.

3.5. Energy efficiency analysis specific energy savings based on previous work using the value
hc = 10 W/(m2 K) [10,27] of 41.8 kJ/kg, which is composed of using
Fig. 14 shows specific energy use as a function of heat transfer an extra 2.9 kJ/kg of electricity in the fans to save 44.7 kJ/kg of heat
coefficient. It shows the optimum heat transfer coefficient for en- losses to ambient. The total energy saving is over 5% based on the
ergy efficiency of a commercial oven to be 35 W/(m2 K), shown fact that the energy needed to bake bread under typical operating
by the dotted line. Using these conditions, the energy model gives conditions is around 800 kJ/kg.
926 Z. Khatir et al. / Applied Energy 112 (2013) 918–927

Fig. 13. Temperature within the bread at an early stage and the end of the baking process when hc = 20 W/(m2 K) at: (a) t = 5 min and (b) t = 27 min.

Fig. 14. Specific energy use as a function of heat transfer coefficient, hc.

Table 3
Worldwide potential energy savings.

Region Annual production Percentage of production GW h saving Annual carbon

(,000’s tonnes) [30] classified as ‘industrial’ [30] saving (tonnes CO2e)
Asia–Pacific 8514.4 57% 55.86 10,054
Europe 50235.0 41% 241.48 43,466
UK (2768.5) (77.5%) (24.91) (4484)
France (3462.3) (27.9%) (11.22) (2019)
Denmark (unknown) (58.5%)
Sweden (58.0) (58.0%) (0.39) (70)
Finland (unknown) (90.2%)
Americas 28286.8 Unknown 37.49 6749
USA (5310.8) (61%) (37.49) (6749)
Canada (unknown) (56%)
Brazil (unknown) (26%)
Worldwide 94604.6 38.2% 419.61 75,530
 Z. Khatir et al. / Applied Energy 112 (2013) 918–927 927

The overall scaled-up energy benefit to the baking industry, [6] Ikin JB, Thompson HM. Drying air-induced disturbances in multi-layer coating
systems. Chem Eng Sci 2007;62:6631–40.
based on annual bread consumption in the UK of 2.8 billion kg
[7] Martin H. Heat and mass transfer between impinging gas jets and solid
per year [30], would yield potential energy savings for the UK of surfaces. Adv Heat Transfer 1977;13:1–60.
around 25 GW h/year. This equates to carbon emissions of [8] Norton T, Sun D-W. Computational fluid dynamics (CFD) – an effective and
4484 tonnes CO2e. Table 3 shows the opportunities for energy sav- efficient design and analysis tool for the food industry: a review. Trends Food
Sci Technol 2006;17:600–20.
ings globally, broken down into the main regions. As bread produc- [9] Khatir Z, Paton JB, Thompson HM, Kapur N, Toropov VV, Lawes M, et al.
tion differs widely across the globe, only the proportion of bread Computational fluid dynamics (CFD) investigation of air flow and temperature
which is produced on an industrial scale has been included in distribution in a small scale bread-baking oven. Appl Energy 2012;89:89–96.
[10] Khatir Z, Thompson HM, Kapur N, Toropov VV, Paton JB. The application of
the savings calculation. The total value for potential worldwide computational fluid dynamics and design optimisation in the British bread-
savings is at least 419.6 GW h or 36 ktoe (kilotons of oil equiva- baking industry. In: Proceedings of the 8th international conference on CFD in
lent). This shows a strong case both financially and environmen- the oil and gas, metallurgical and process industries, Trondheim Norway; 21–
23 June 2011.
tally to design and manufacture energy optimised baking ovens. [11] Therdthai N, Zhou WB, Adamczak T. Two-dimensional CFD modelling and
simulation of an industrial continuous bread baking oven. J Food Eng
2003;60:211–7.
4. Conclusions [12] Therdthai N, Zhou WB, Adamczak T. Three-dimensional CFD modelling and
simulation of the temperature profiles and airflow patterns during a
A novel methodology is presented which combines heat trans- continuous industrial baking process. J Food Eng 2004;65:599–608.
[13] Purlis E. Bread baking: technological considerations based on process
fer, thermal airflow and a bread baking model to predict theoretical modelling and simulation. J Food Eng 2011;103:92–102.
optimum operating conditions with respect to energy efficiency in [14] Hadiyanto, Boom RM, Van Straten G, Van Boxtel AJB, Esveld DC. Multi-
commercial baking ovens. Designing ovens with the optimum heat objective optimisation to improve the product range of baking systems. J Food
Process Eng 2009;32: 709–29.
transfer coefficient of 35 W/(m2 K) could potentially reduce energy [15] Khatir Z, Thompson HM, Kapur N, Toropov VV, Paton JB. Multi-objective
usage by 5%, on top of the 10% saving that can be gained through computational fluid dynamics (CFD) design optimisation in commercial bread-
optimisation for temperature uniformity [15]. Energy savings can baking. Appl Therm Eng 2012. <http://dx.doi.org/10.1016/j.applthermaleng.
2012.08.011>.
be further increased for smaller or denser products where thermal [16] Patakar SV, Spalding DB. Calculation procedure for heat, mass and
conduction to the core is better. This study shows thermal airflow momentum-transfer in 3-dimensional parabolic flows. Int J Heat Mass
and energy analysis for one set of commercially used nozzle de- Transfer 1972;15(10):1787–806.
[17] Boulet M, Marcos B, Dostie M, Moresoli C. CFD modeling of heat transfer and
signs. Further work to optimise other nozzle designs may result flow field in a bakery pilot oven. J Food Eng 2010;97(3):393–402.
in even larger energy savings for bread baking, and the drying [18] OpenFOAMÒ – the open source CFD toolbox. <http://www.openfoam.com>.
industry as a whole, and is currently on-going. [19] Banooni S, Mujumdar Arun S. Mathematical modeling of bread baking in an
impingement oven. In: Peng Xu, Zhonghua Wu, Arun S. Mujumda, editors.
Mathematical modeling of industrial transport processes, 2nd ed., Singapore:
Acknowledgements Transport Processes Research (TPR) Group National University of Singapore;
August 2012. p. 192–210. ISBN-978-981-08-6269-5.
[20] Earle RL. Unit operations in food processing. New York: Pergamon Press; 1983.
The financial support of the UK’s Engineering and Physical Sci- <http://www.nzifst.org.nz/unitoperations>.
ences Research Council through Grant EP/G058504/1 is gratefully [21] Toledo RT. Fundamentals of food process engineering. 3rd ed. In: Science,
acknowledged. The authors also thank Spooner Industries Ltd. for Springer; 2007.
[22] Mohanty AK, Tawfek AA. Heat transfer due to a round jet impinging normal to
providing access to their pilot plant baking oven and to Warbur- a flat surface. Int J Heat Mass Transfer 1993;36(6):1639–47.
tons Ltd. for their on-going support of this work. [23] Khatir Z, Paton JB, Thompson HM, Kapur N, Toropov VV, Lawes M, et al.
Thermal management in commercial bread baking – CFD investigation of
thermal air flow in a small scale bread-baking oven. In: The 1st PRO-TEM
References network conference – sustainable thermal energy management, SusTEM2010
Conference, Newcastle (UK); 2–3 November 2010. p. 146–55.
[1] Therdthai N, Zhou WB, Adamczak T. Optimisation of the temperature profile in [24] Wong SY, Zhou WB, Hua JS. CFD modeling of an industrial continuous bread-
bread baking. J Food Eng 2002;55:41–8. baking process involving U-movement. J Food Eng 2007;78(3):888–96.
[2] Thumann A, Mehta DP. Handbook of energy engineering. 6th ed. Lilburn GA, [25] Bollada PC. Expansion of elastic bodies with application in the bread industry.
Boca Raton FL: Fairmont Press; 2008. Mathmat Comput Model 2008;48(7–8):1055–67.
[3] Probert D, Newborough M. Designs, thermal performances and other factors [26] Jefferson DR, Lacey AA, Sadd PA. Crust density in bread baking: mathematical
concerning cooking equipment and associated facilities. Appl Energy modelling and numerical solutions. Appl Math Model 2007;31(2):209–25.
1985;21:81–222. [27] Zhang J, Datta AK. Mathematical modeling of bread baking process. J Food Eng
[4] Paton JB, Khatir Z, Thompson HM, Kapur N, Toropov VV. Thermal energy 2006;75:78–89.
management in the bread-baking industry using a system modelling approach. [28] Incroprera FP, Dewitt DP, Bergman TL, Lavine AS. Fundamentals of heat and
Appl Therm Eng 2012. <http://dx.doi.org/10.1016/j.applthermaleng.2012.03.036>. mass transfer. 6th ed. John Wiley & Sons; 2006.
[5] Paton JB, Khatir Z, Thompson HM, Kapur N, Toropov VV. Integration of system- [29] COMSOL MultiphysicsÒ engineering simulation software <http://www.comsol.
modelling and computational methods for energy reductions in the bread com>.
industry. In: Sustainable thermal energy management international [30] Bread & Rolls Industry Profile. Global, Bread & Rolls Industry Profile,
conference (SusTEM 2011), Newcastle upon Tyne (UK); 2011. p. 49–57. Datamonitor Plc: United Kingdom; 2011. p. 37.