Journal of Food Engineering 77 (2006) 10371047 www.elsevier.

com/locate/jfoodeng

A CFD modeling system for airow and heat transfer in ventilated packaging for fresh foods: I. Initial analysis and development of mathematical models
Qian Zou a, Linus U. Opara
b

b,* ,

Robert McKibbin

a Food Systems and Technology, AgResearch Limited, Ruakura MIRINZ Centre, Private Bag 3123, Hamilton, New Zealand Department of Bioresource and Agricultural Engineering, College of Agricultural and Marine Sciences, Sultan Qaboos University, P.O. Box 34, Al-Khod, Oman Institute of Information and Mathematical Sciences, Massey University, Albany Campus, Private Bag 102 904, North Shore Mail Centre, Auckland, New Zealand

Received 1 December 2004; accepted 2 August 2005 Available online 30 September 2005

Abstract Ventilated packaging is commonly used in the horticultural industry to facilitate ecient cooling of fresh produce during forced-air cooling. The airow patterns and heat transfer processes inside such packaging systems are complex and research is needed to aid in the thermodynamic design and testing of such packaging systems to ensure adequate cooling and maintenance of produce quality. The aim of this research program was to develop a computational uid dynamics (CFD) modelling system of the airow patterns and heat transfer inside ventilated packages. In Part I of this article series, we rst describe the packaging systems modelled and then present the relevant mathematical models for both airow patterns and heat transfer inside such ventilated packaging systems based on the porous media approach. In the subsequent article (Part II), the computational solution and simulation software for the solution implementation are discussed, and then simulation results are compared with experimental data for model validation. 2005 Elsevier Ltd. All rights reserved.
Keywords: CFD modelling; Forced-air cooling; Ventilated packages; Heat transfer; Air ow

1. Introduction Forced-air cooling (pressure cooling) is the most common method for pre-cooling of horticultural produce to the optimum storage temperature. Ventilated packaging is required to achieve fast and uniform cooling. The cooling rates of produce mainly depend on heat transfer between cooling medium (air) and produce items in the packages. The heat transfer processes are closely related to airow transport within the packages. Materials and congurations of produce packaging system (trays, cartons, bins, palletisation patterns, stacking patterns, etc.) have major impacts on the heat transfer and airow pat*

Corresponding author. Tel.: +968 24415226. E-mail address: linus@squ.edu.om (L.U. Opara).

terns during force-air cooling. A packaging system needs to be carefully evaluated before implementation to ensure good cooling eciency. It is usually considered expensive, time-consuming and situation-specic to only use experimental methods for studying heat transfer and airow processes. Alternatively, mathematical modelling is overall a cost-eective strategy for predicting the airow patterns and temperature variation in the controlled environments such as ventilated packages. If information on packaging system, cooling conditions, produce properties is used as model input data, the results obtained can predict the eects of these factors on the airow patterns and cooling rate. Computational Fluid Dynamics (CFD) employs numerical methods to solve the fundamental uid transport equations that are derived from the laws of conservation of mass, momentum

0260-8774/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2005.08.042

1038

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

Nomenclature A empirical constant a empirical constant aap specic interfacial surface area, m1 ap-exposed exposed-to-air produce surface area per unit volume of air-produce region (m1) B empirical constant b empirical constant c constant Ca air specic heat at constant pressure, J kg1 K1 Cdis empirical constant Cp product specic heat, J kg1 K1 Cpack packaging material specic heat capacity, J kg1 K1 d constant deq equivalent mean diameter, m F Forchheimer coecient ht heat transfer coecient between air and solid surface, W m2 K1 K permeability, m2 Ka air thermal conductivity, W K1 m1 Kdis dispersion conductivity, W K1 m1 Kp product thermal conductivity, W K1 m1 Kpack packaging material thermal conductivity, W K1 m1 Nud Nusselt number, based on particle dimension p air pressure, N m2 hpia intrinsic phase average of air pressure, N m2 Ped Peclet number based on particle dimension Pr Prandtl number qPSur heat ux entering the outer surface of produce item per unit produce volume, W m3 qtray-Psur conduction heat ux from trays to the outer surface of produce item per unit produce volume, W m3 r spatial variable along sphere radius, m Red Rp Sp t Ta hTaia Tpack hTpip TPSur u huia v hvia Vp w hwia yn l / /1 /gap qa qB qp qpack Reynolds number product respiration heat, W kg1 mean surface area of produce item, m2 time, s air temperature, K intrinsic phase average of air temperature, K packaging material temperature, K intrinsic phase average of product temperature, K produce surface temperature, K air velocity component in the direction of x-axis, m s 1 intrinsic phase average of air velocity component in the direction of x-axis, m s1 air velocity component in the direction of y-axis, m s 1 intrinsic phase average of air velocity component in the direction of y-axis, m s1 mean volume of produce item, m3 air velocity component in the direction of z-axis, m s 1 intrinsic phase average of air velocity component in the direction of z-axis, m s1 distance from the wall, m air dynamic viscosity, N s m2 porosity average porosity estimated porosity of the vertical tunnel along package wall air density, kg m3 bulk density of container packed with products, kg m3 density of product, kg m3 packaging material density, kg m3

and energy. The increasing capacity and decreasing cost of modern computers have made the application of CFD modelling more ecient and popular. In general three types of models have been developed for modelling airow patterns and heat transfer in horticultural produce packages or refrigerated spaces during cooling processes. The rst type is zoned model (Amos, 1995; Tanner, 1998), in which the domains considered were divided into a number of zones. Airow was modelled by dening an airow pathway according to experimental data. Energy and water vapour mass balances were performed on each zone to determine air temperature, air humidity ratio, and the temperature of products and packaging materials. Zoned model requires much less computing eort, and it is easy to write computer codes for model solution. However, since the airow patterns were estimated from measured data for certain packages

or coolstores, this approach limits the model application under dierent package designs or coolstore arrangements. The second type is fully-distributed model, which applies numerical methods to solve 2D or 3D mass, momentum, and energy conversation equations (Wang & Touber, 1990; Zou, 1998). As the airow patterns are solved explicitly, no experimental data are required to run the model. If the model is used for the transport processes within a produce package, a complex body-tted grid system has to be generated to describe the complicated geometries inside the package, which could be a daunting task for most model users. The diculties in grid generation for detailing the geometries of dierent types of packaging systems largely reduced the accessibility of this type of model. The third type is porous-medium model (Tassou & Xiang, 1998; Xu & Burfoot, 1999), in which produce items inside the packages are treated as saturated porous media.

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

1039

Macroscopic volume-averaged transport equations are solved to nd the volume-averaged velocity and temperature. Since certain information with respect to microscopic structure is lost in the spatial averaging process, a set of empirical parameters is required for the closure of the macroscopic equations. These parameters are found in the expressions for porosity, permeability, Forchheimer constant, thermal and mass dispersion, and interfacial heat and mass transfer coecients. The volume-averaged approach eliminates the need to generate complicated meshes to describe the geometric details of the packaging systems. Therefore the porous medium models usually require less computing capacity than the microscopic models. However, existing studies (Tassou & Xiang, 1998; Xu & Burfoot, 1999) only dealt with some specic cooling conditions and bulk containers, and thus are not readily applicable to a wide range of packaging systems and horticultural crops. The aim of this research program was to develop a CFD modelling system for simulating airow and heat transfer processes, and therefore to predict airow patterns and temperature proles in ventilated packaging systems during the forced-air cooling of fresh produce. Such a modelling system can nd practical applications in evaluating forced-air cooling operations and assessing the cooling performance of alternative packaging designs for a range of horticultural commodities. In Part I of this paper series, the ventilated packages and forced-air cooling systems are analysed, then after discussing the general modelling strategies are discussed, the airow and heat transfer models are presented. In Part II, the computational solution adopted and software developed to integrate the overall modelling system are presented, and the model predictions for various packaging systems are compared with the experimental results for model testing. 2. Analysis of ventilated packaging and forced-air cooling systems 2.1. Ventilated packaging systems Based on the way products are packed in the containers, these ventilated packages can be divided into two main types, as shown in Fig. 1: Bulk packages, in which produce items are held in a bin or carton without any other packaging materials. Layered packages, in which produce items are placed on a stack of trays During forced-air cooling, bulk bins and cartons are grouped into pallets or stacks in front of fans or plenum. For secure palletisation, cross-stacked patterns may be used as shown in Fig. 1. To investigate the performance of a packaging system in terms of produce cooling eciency, both the characteristics of individual package (conguration, dimensions, vents, and packaging materials,

etc.) and the structure of the stack should be considered. Therefore this study took account of two domains of the packaging systems: individual package and stack of packages. 2.2. Forced-air cooling systems In most forced-air cooling systems (Mitchell, 1992; Watkins, 1989), the fans are closely positioned in front of one side of the stack, so the airow conditions inside the package stack are very similar, and can be approximately described as follows: On the stack side close to the fans, airow leaves/enters the vents with an approximately constant ow rate. On the stack sides other than the one close to the fans, airow pressure is approximately equal to the pressure of surrounding environment. If airow enters a vent, it has the temperature approximately equal to that of the air leaving the evaporator of the cooling system. Due to the similar airow conditions inside the package stacks in dierent forced-air cooling systems, this study focused on the transport processes taking place inside packaging systems, and therefore avoided dealing with minor details of various cooling systems. 3. Modelling strategies 3.1. Geometric features of ventilated packaging systems For both layered and bulk packaging systems, the domain inside an individual package or stack of packages can be divided into three types of regions as shown in Fig. 2: Produce-air regions (the void spaces and produce inside bulk packages and the void spaces and produce between trays in layered packages). Plain air regions (the spaces in the vents). Solid regions (package walls and trays). 3.2. General analysis of transport processes Since the air velocity is relatively large during forced-air cooling, the eect of buoyancy forces is considered negligible. By neglecting buoyancy forces, the heat transfer was assumed to have no eects on the airow mass and momentum transfer. Therefore, airow transport processes were treated as steady state, and the related airow transport equations were decoupled from the unsteady-state heat transfer equations. For most ventilated packaging systems, the vents on the bottom and top surfaces of packages are either blocked or do not exist when the packages are stacked, which indicates

1040

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

Fig. 1. Examples of layered and bulk packaging systems.

no air movement between dierent package layers in a stack. The heat transfer between the package layers is mainly achieved by heat conduction through the package surfaces, which is considered negligible compared with the heat transfer between airow and products during forced-air cooling. Therefore it was assumed there is no air movement and heat transfer between the package layers, and each package layer in a stack can be modelled independently. For airow in a vent, the dominant direction for air movement and heat transfer is perpendicular to the package wall with the vent, so one-dimensional airow and heat transfer in the vents were assumed. As produce items are packaged in boxes or bins, the individual produce item would likely receive minimal net radiative heat transfer (Tanner, 1998). Hence, it was assumed that eects of radiative heat transfer were negligible.

The range of air velocity (0.53.0 m/s) in forced-air cooling indicate that the possible changes in air temperature, pressure, and moisture content will not cause any signicant changes in most air properties. Thus it was assumed that air density, specic heat capacity, thermal conductivity and viscosity are constant. 3.2.1. Porous media treatment of produce-air region in bulk packages To avoid dealing with the geometric details inside the packages the porous media approach was adopted. Most fresh products are more or less sphere-shaped, and have relatively uniform sizes. The dimensions of bulk bins are generally at least one-order larger than the sizes of individual products. Therefore it was assumed that the produceair regions inside bulk packages are isotropic, rigid, saturated porous media with uniform spherical particles.

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

1041

Fig. 2. Regions in layered and bulk packages for fresh produce.

3.2.2. Pseudo-porous-media treatment of produce-air region in layered packages The geometry inside a layered package is more complex than that of bulk package. In general, the produce-air region inside the layered package is divided into several produce layers by the trays, as shown in Fig. 3. The distances between two neighbouring trays usually have the same order as the sizes of produce items, so the strict porous media approach cannot be used. A pseudo-porous-media treatment was employed to deal with the geometries of layered packages. 3.2.3. Airow and heat transfer within a produce layer The air movement along the vertical direction within a produce layer is usually negligible when compared with the air movement at horizontal directions, so it was assumed that the airow in each produce layer between two trays has only horizontal movements. Similarly the heat transfer within the airow in the produce-air region in the vertical direction was assumed to be negligible.

The thickness of a produce layer, i.e. the distance between two trays, is the same order as the sizes of produce items, so in theory the air-produce region in the produce layer cannot be treated as a porous medium. However, when examining the geometric characteristics of the layered packages carefully, it was found that the structure of the produce layer between two trays is very similar to that of a section cut from a packed bed since the trays are designed to tightly t the shapes of produce. Hence the produce-air regions between trays were treated as isotropic, rigid, saturated porous media with uniform spherical particles, if package length and width are at least one-order larger than sizes of individual products. As shown in Fig. 3, the package interior geometric structure with ve layers of product is transformed into six porous media layers. The additional bottom porous layer is used to represent the space between the package bottom wall and the bottom tray. The assumption of porous media is eective for air movement and heat transfer within the air in produce-air region, and the volume-averaged equation for air mass,

1042

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

Fig. 3. Illustration of the porous media approach for a layered package.

momentum, and energy conservation can be used. However, the produce items in the layered packages are generally not allowed to be in contact with each other so as to reduce the incurrence of produce mechanical damage, so it is unnecessary to use the porous media approach for produce energy conservation, and only the heat transfer within individual items should be considered. In Fig. 3, the trays were transferred into at plates. For the airow model, the at plates were only used for dening the positions of produce layers. For the heat transfer model, the actually tray areas (the area exposed to air, the area in contact with the produce item above, and the area in contact with the produce item below) were used, and the at plates were irrelevant. The above pseudo-porous-media treatment did not completely satisfy the conditions for classic porous media due to the existence of trays and the comparable dimensions of produce items and packages, which may cause extra modelling errors. However, this treatment can be considered adequate as long as it provides relatively accurate estimation on air distribution, which is the key for accurate prediction of he temperature proles of individual produce items. This approach enables us to avoid dealing with the situation-specic and complex geometries inside the packaging systems, and therefore facilitates the development of a general modelling system suitable for a wide range of packaging designs, produce types and stacking arrangement inside cool-stores.

3.2.4. Airow and heat transfer between produce layers Produce in layers are separated by trays, and the air movement between these produce layers is in the gaps between the tray edges and package walls, and mainly caused by pressure dierence. To simplify the treatment of air movement between produce layers, it was assumed that the air movement between produce layers is only along the direction of the package height. As shown in Fig. 3, the vertical airow from one product layer to the others was assumed to pass through narrow tunnels with the widths of the gaps between tray edges and package walls. Due to the complicated geometry in the near-package-wall regions in the tunnels, the resistance to airow caused by produce items close to the packaging was estimated by the Darcy and Forchheimer terms in the generalised macroscopic momentum equation for porous media (Section 4). Parts of the near-package-region in the vertical tunnel overlap with the product layers. The vertical airow between the product layers is related to the horizontal airow within the produce layers in the following two aspects: The momentum conservation equations for airow within the product layers share the pressure eld with the momentum conservation equation for airow between the product layers. The velocity of the vertical airow on the boundaries of the product layers should satisfy the mass conservation for each product layer.

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

1043

In terms of heat transfer, air in the gaps was treated in the same way as in the vents, so the gaps are plain air regions, and heat transfer in the gaps is one-dimensional and perpendicular to the trays. 3.2.5. Analysis of domain boundaries For the vent boundaries on the stack side near the auxiliary fans, assumptions were made as follows: Airow enters the vents with xed velocities in the direction perpendicular to the vent if the fans blow air into the packages. Airow leaves the vents with xed velocities in the direction perpendicular to the vent if the fans extract air out of the packages. For the vents on the stack sides other than the one near the auxiliary fans, it was assumed that air pressure is equal to the pressure of the surrounding environment. When considering the heat transfer on a vent boundary a xed temperature value was assumed if cold air enters the vent, and the air temperature was assumed to be invariant along the direction normal to the vent if air leaves the vent. Wall boundaries include inner surfaces of package walls, surfaces of trays, and surfaces around the perimeters of vents. No-slip boundary condition was used for airow transport equations. The heat transfer between airow and walls was modelled with Newtons cooling law. For the boundaries between the plain air regions and the produce-air porous regions, i.e. the interfaces between the vents and the regions inside package, it is adequate to maintain the mass, momentum, and energy conservation on the boundaries, as airow is perpendicular to the interfaces. The boundary conditions for single produce items need to take account of the heat convection between air and the exposed surfaces of produce items, the heat conduction between produce items in bulk packages, and the heat conduction between trays and produce items in layered packages. 4. Airow model formulation 4.1. Airow model in bulk packages The mathematical model for airow in bulk packaging systems consists of the equations presented as follows. 4.1.1. Continuity equation for air mass conservation in the plain air regions The air mass conservation in vents is described by the one-dimensional continuity equation (Bird, Stewart, & Lightfoot, 1960): du 0 dx dv 0 dy 1a 1b

where u is air velocity component in the direction of x-axis, m s1; v is air velocity component in the direction of y-axis, m s1. 4.1.2. Volume-averaged continuity equation for air mass conservation in the produce-air regions The air mass conservation in the produce-air regions is described by the volume-averaged continuity equation (Hsu & Cheng, 1990; Kaviany, 1995): o o o /huia /hvia /hwia 0 2 ox oy oz where / is porosity; huia is the intrinsic phase average of air velocity component in the direction of x-axis, m s1; hvia is the intrinsic phase average of air velocity component in the direction of y-axis, m s1; hwia is the intrinsic phase average of air velocity component in the direction of z-axis, m s1. 4.1.3. Equations for describing air momentum conservation in the plain air regions The air momentum conservation in vents is described by one-dimensional NavierStokes equations (Bird et al., 1960):   du d du dp l 3a qa u dx dx dx dx   dv d dv dp l 3b qa v dy dy dy dy where p is air pressure, N m2; l is the air dynamic viscosity, N s m2; qa is the air density, kg m3. 4.1.4. Volume-averaged momentum equations for air momentum conservation in the produce-air regions The conservation of air momentum in the produce-air regions is described by the generalised volume-average momentum equation (Hsu & Cheng, 1990; Kaviany, 1995; Vafai & Tien, 1981): o o o qa /huia huia qa /hvia huia qa /hwia huia ox oy oz       o oh ui a o oh ui a o oh ui a l/ l/ l/ ox oy oz ox oy oz 2 3  1 o/hpia l/ huia F / qa 2 2 2 2 p huia hvia hwia huia ox K K 4a o o o q /huia hvia q /hvia hvia qa /hwia hvia ox a oy a oz       o oh v i a o oh v i a o oh v i a l/ l/ l/ ox oy oz ox oy oz 2 3  1 o/hpia l/ hvia F / qa 2 2 2 2 p huia hvia hwia hvia oy K K 4b

1044

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

o o o qa /huia hwia qa /hvia hwia qa /hwia hwia ox oy oz       o ohwia o oh w i a o ohwia l/ l/ l/ ox oy oz ox oy oz 1 o/hpia l/2 hwia F /3 qa  2 2 2 p huia hvi2 h w i a a hwia oz K K 4c where K is the permeability, m2; F is Forchheimer coecient; hpia is the intrinsic phase average of air pressure, N m2. 4.1.5. Auxiliary algebraic equations for the porosity and permeability in the produce-air regions To account for channelling eects, the porosity near a package wall is modelled by an exponential decaying function of the distance to the wall (Amiri & Vafai, 1994, 1998):   by / /1 1 ae d eq 5 where / is porosity; /1 is the free stream porosity; y is the distance from the nearest boundary (m); a is the empirical constant (=1.7); b is empirical constant (=6). For packed beds of sphere-like particles, permeability K and Forchheimer coecient F can be expressed as (Ergun, 1952): K
3 d2 eq / 2

4.2.2. Volume-averaged continuity equation for air mass conservation in the produce-air regions between trays The two-dimensional volume-averaged continuity equation was employed (Hsu & Cheng, 1990; Kaviany, 1995): o o /huia /hvia 0 ox oy 9

4.2.3. Continuity equation for air mass conservation in the vertical tunnels (airow between produce layers) The air mass conservation in vertical tunnels is described by the one-dimensional continuity equation (Bird et al., 1960): dw 0 dz 10

where w is the air velocity component in the direction of z-axis, m s1. 4.2.4. Momentum equations for air momentum conservation in the plain air regions The air momentum conservation equation in the vents of the layered packages is the same as that of the bulk packaging system (Eq. (3)). 4.2.5. Volume-averaged momentum equations for air momentum conservation in the produce-air regions between trays The volume-averaged equations of motion for airow with only horizontal movements were written as follows (Hsu & Cheng, 1990; Kaviany, 1995; Vafai & Tien, 1981):   o o o ohuia qa /huia huia qa /hvia huia l/ ox oy ox ox     o oh ui a o ohuia l/ l/ oy oz oy oz 2 1 o/hpia l/ huia F /3 qa  2 2 p huia hvi2 a h ui a ox K K 11a o o o ohvia q /huia hvia qa /hvia hvia l/ ox a oy ox ox     o oh v i a o ohvia l/ l/ oy oz oy oz 2 1 o/hpia l/ hvia F /3 qa  2 2 2 p huia hvia hvia oy K K 11b 4.2.6. Momentum equations for air momentum conservation in the vertical tunnels (airow between produce layers) The momentum conservation equation for one-directional ow was employed, and the Darcy and Forchheimer terms were used to account for the resistance caused by  

A1 / B F q A/ 3

6 7

where A is empirical constant (=180); B is empirical constant (=1.8). The values of the empirical constants used in Eqs. (6) and (7) were taken for the work by Amiri and Vafai (1994, 1998). The equivalent mean diameter of produce items is calculated as: d eq 6 Vp Sp 8

where Vp is the mean volume of produce item, m3; Sp is the mean surface area of produce item, m2. 4.2. Airow model in layered packages The mathematical model for airow in the layered packaging systems consists of the equations presented as follows: 4.2.1. Continuity equation for air mass conservation in the plain air regions The air mass conservation equation in the plain air regions of a layered packaging system is the same as that of the bulk packaging system (Eq. (1)).

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

1045

produce items next to the package walls. (Vafai & Tien, 1981):       ow o ow o ow o ow l l l qa w oz ox ox oy oy oz oz
2 op l/gap w F /gap qa p jwjw oz K K

12

5.1.3. Volume-averaged air energy equation in the produceair regions The volume-averaged energy conservation equation for the uid phase in porous media was written for air energy conservation; the item on the RHS of the Eq. (14) represents heat exchange between produce and air (Hsu & Cheng, 1990; Kaviany, 1995). oqa /C a hT a ia oqa /huia C a hT a ia ot ox oqa /hvia C a hT a ia oqa /hwia C a hT a ia oy ox     o oh T a i a o ohT a ia /K a K dis /K a K dis ox  ox  oy oy o oh T a i a /K a K dis ht aaP T PSur hT a ia oz oz 15 where hTaia is the intrinsic phase average of air temperature, K; TPSur is the produce surface temperature, K; Kdis is the dispersion conductivity, W K1 m1; aap is the specic interfacial surface area, m1 (m2 soliduid/m3 porous medium); ht is the heat transfer coecient between air and solid surface, W m2 K1. 5.1.4. Volume-averaged product energy equation in the produce-air regions The volume-averaged energy conservation equation for the solid phase in porous media was written for produce energy conservation; the rst item on the RHS of Eq. (15) represents heat exchange between produce and air (Hsu & Cheng, 1990; Kaviany, 1995).     o qp C p 1 /hT p ip oh T p i p o 1 /K p ox ot ox     ohT p ip oh T p i p o o 1 /K p 1 /K p oy oz oy oz 16 aap ht T PSur hT a ia 1 /qp Rp where qp is the product density, kg m3; hTpip is the intrinsic phase average of product temperature, K; Kp is the product thermal conductivity, W K1 m1; Cp is the product specic heat, J kg1 K1; Rp is product respiration heat, W kg1. 5.1.5. Energy conservation equation in single produce items The energy conservation equation for a spherical solid was written for energy conservation of single produce items (Bird et al., 1960):   oqp C p T P 1 o 2 oT p K pr 2 17 qp R p r or ot or where r = spatial variable along sphere radius, m. 5.1.6. Auxiliary algebraic equations For a sphere-packed bed, specic surface area is (Dullien, 1979):

where /gap is the estimated porosity of the vertical tunnel along package wall. 4.2.7. Auxiliary algebraic equations for calculating porosity and permeability in the air-product regions As the number and position of produce items on each tray is pre-dened for a layered package, the porosity at any position on each tray can be calculated accordingly. The permeability K and Forcheimer coecient F are calculate with the same equations as for the bulk packages (Eqs. (6) and (7)). 5. Heat transfer model formulation 5.1. Heat transfer model in bulk packages The mathematical model for heat transfer in the bulk packaging systems consists of the equations presented as follows: 5.1.1. Air energy conservation equation in vents Air was treated as an incompressible uid, and the onedimensional air energy equations were written (Bird et al., 1960):   oqa C a T a oqa uC a T a o oT a Ka 0 13a ot ox ox ox   oqa C a T a oqa vC a T a o oT a Ka 0 13b ot oy oy oy where t is time, s; Ca is the air specic heat at constant pressure, J kg1 K1; Ta is the air temperature, K; Ka is the air thermal conductivity, W m1 K1. 5.1.2. Energy conservation equation in the solid regions The energy equation for package walls was written (Bird et al., 1960):   oqpack C pack T pack o oT pack K pack ox ot ox     o oT pack o oT pack K pack K pack oy oz oy oz 0 14 where qpack is packaging material density, kg m3; Cpack is packaging material specic heat capacity, J kg1 K1; Tpack is the packaging material temperature, K; Kpack is the packaging material thermal conductivity, W K1 m1.

1046

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

aap

61 / d eq

18

The heat transfer coecient between air and the surfaces of produce items was calculated with the following relations (Geankoplis, 1993): / Nud 2:876 0:3023 Pr0:66 :35 Red Red Pr Re0 d ht d eq Nud Ka  1 2 2 2 2 /q huia hvia hwia d eq Red l Cal Pr Ka 19 20

5.2.1. Air energy equation in vents or gaps between tray edges and package walls The energy conservation equation in vents of the layered packaging system is the same as that of the bulk packaging system (Equation 5.1). For the gaps between tray edges and package walls, energy conservation equation along z-axis was written:   oqa C a T a oqa wC a T a o oT a Ka 0 ot oz oz oz 27

21 22

where Nud is Nusselt number; Red is Reynolds number; Pr is Prandtl number. The dispersion conductivity is a tensor that is a function of Peclet number based on the supercial velocity in the main ow direction (Kaviany, 1995; Wakao & Kaguei, 1982). Since the main ow direction is dicult to dene, the tensor is simplied as a scalar, and assumed to be a linear function of the Peclet number based on the local supercial velocity: K dis C dis Ped K a 23  1 2 2 2 2 C a qa / huia hvia hwia d eq Ped 24 Ka where Ped is Peclet number based on particle dimension; Cdis is the empirical constant ($0.2). Product respiration heat was calculated as follows (Tanner, 1998): Rp cT p 255:35
d

5.2.2. Solid energy conservation equation in package walls and trays The energy conservation equation in solid regions of the layered packaging system is the same as that of the bulk packaging system (Eq. (13)). 5.2.3. Volume-averaged air energy equation in the produceair regions The volume-averaged energy conservation equation for the uid phase in porous media was written for the energy conservation of airow with horizontal movements (Hsu & Cheng, 1990; Kaviany, 1995): oqa /C a hT a ia oqa /huia C a hT a ia oqa /hvia C a hT a ia ot ox oy     o oh T a i a o ohT a ia /K a K dis /K a K dis ox oy ox oy   o oh T a i a /K a K dis ht ap -exposed T PSur hT a ia oz oz 28 where ap-exposed is exposed-to-air produce surface area per unit volume of air-produce region (m1). 5.2.4. Solid energy conservation equation in single items of produce The energy conservation of single items of produce in the layered packaging system is the same as that of the bulk packaging system (Eq. (16)). 5.2.5. Auxiliary algebraic equations All auxiliary equations in heat transfer model for the bulk packaging systems were the same as for the layered packaging systems apart from the outer surface boundary condition for the energy equation of single produce item, which includes the heat convection between air and exposed produce surface, and heat conduction between the produce item and trays. The heat conduction between the produce item and trays is related to the contact areas between produce item and trays: qPSur qtray-PSur ap-exposed ht T PSur hT a ia 29

25

where c is constant; d is constant. The values of constant c and d are associated with specic crops, and were summarised by Tanner (1998). The outer surface boundary conditions for the energy equation of single produce item takes account of heat convection between air and exposed produce surface, and heat conduction between the produce items:     ohT p ip oh T p i p o o qPSur 1 /K p 1 /K p ox oy ox oy   o h T i o p p 1 /K p aap ht T PSur hT a ia oz oz 26 where qPSur is the heat ux entering the outer surface of produce item per unit produce volume, W m3. 5.2. Heat transfer model in the layered packages The mathematical model for heat transfer in the layered packaging systems consists of the equations presented as follows:

where qtray-PSur = conduction heat ux from trays to the outer surface of produce item per unit produce volume, W m3.

Q. Zou et al. / Journal of Food Engineering 77 (2006) 10371047

1047

6. Conclusions Airow and heat transfer models in bulk and layered packaging systems have been developed based on a porous media approach. The areas inside the packaging systems were categorised as solid, plain air, and produce-air regions. The produce-air regions inside the bulk packages or between trays in the layered packages were treated as porous media, in which volume-averaged transport equations were employed. This approach avoids dealing with the situation-specic and complex geometries inside the packaging systems, and therefore facilitates the development of a general modelling system suitable for a wide range of packaging designs, produce types and stacking arrangement inside cool-stores. In Part 2, the solution methods for airow and heat transfer models will be discussed, and model simulation results for several package systems will be presented and compared with the experimental data. Acknowledgement The authors gratefully acknowledge the nancial support of Mr Collis Blake for a Research Award to Dr Linus U. Opara, and Massey University, New Zealand, for a postgraduate research award to Dr Qian Zou. References
Amiri, A., & Vafai, K. (1994). Analysis of dispersion eects and nonthermal equilibrium, non-Darcian, variable porosity incompressible ow through porous media. International Journal of Heat and Mass Transfer, 37, 939954.

Amiri, A., & Vafai, K. (1998). Transient analysis of incompressible ow through a packed bed. International Journal of Heat and Mass Transfer, 41, 42594279. Amos, N. D. (1995). Mathematical modelling of heat transfer and water vapour transport in apple coolstore. PhD Thesis, Massey University, Palmerston North, New Zealand. Bird, R. B., Stewart, W. Z., & Lightfoot, Z. N. (1960). Transport phenomena. New York: Wiley. Dullien, F. A. L. (1979). Porous media: Fluid transport and pore structure. New York: Academic Press. Ergun, S. (1952). Fluid ow through packed columns. Chemical Engineering Progresses, 48, 8994. Geankoplis, C. J. (1993). Transport processes and unit operations (3rd ed.). Boston: Allyn & Bacon. Hsu, C. T., & Cheng, P. (1990). Thermal dispersion in a porous medium. International Journal of Heat and Mass Transfer, 33, 15851597. Kaviany, M. (1995). Principle of heat transport in porous media (2nd ed.). New York: Springer. Mitchell, F. G. (1992). Cooling horticultural commodities. In A. A. Kader (Ed.), Postharvest Technology of Horticultural Crops. Davis, CA: University of California. Tanner, J. D. (1998). Mathematical modelling for design of horticultural packaging. PhD thesis, Massey University, Palmerston North, New Zealand. Tassou, S. A., & Xiang, W. (1998). Modelling the environment within a wet air-cooled vegetable store. Journal of Food Science, 38, 169187. Vafai, K., & Tien, C. L. (1981). Boundary and inertia eects on ow and heat transfer in porous media. International Journal of Heat and Mass Transfer, 24, 195203. Wakao, N., & Kaguei, S. (1982). Heat and mass transfer in packed beds. New York: Gordon and Breach Science Publisher. Wang, H., & Touber, S. (1990). Distributed dynamic modelling of a refrigerated room. International Journal of Refrigeration, 13, 214222. Watkins, J. B. (1989). Forced-air cooling (2nd ed.). Brisbane, Australia: Queensland Department of Primary Industries. Xu, Y., & Burfoot, D. (1999). Simulating the bulk storage of foodstus. Journal of Food Engineering, 39, 2329. Zou, Q. (1998). CFD modelling of airow and heat transfer in a ventilated carton. Master thesis, Massey University, Palmerston North, New Zealand.