Applied Thermal Engineering 41 (2012) 71e83

Contents lists available at SciVerse ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

The art of air blast freezing: Design and efciency considerations

Patrick Dempsey, Pradeep Bansal*,1
Department of Mechanical Engineering, The University of Auckland, Private Bag e 92019, Auckland, New Zealand

a r t i c l e i n f o
Article history: Received 16 March 2011 Accepted 6 December 2011 Available online 13 December 2011 Keywords: Blast freezer Batch Carton Refrigeration Low temperature Energy Efciency

a b s t r a c t
Air blast freezing is a common freezing technique used throughout the world to freeze various food commodities from carcasses to packaged goods. The New Zealand Cold Storage industry identied blast freezing as the most energy intensive operation in the frozen food storage industry, consuming 8.1 GWh of electricity in New Zealand in 2005. This paper presents an overview of various types of blast freezers, their common design aws, common energy saving measures and a best practice guide. A simulation model has also been presented to predict the performance and to design an optimal system under range of operating conditions. 2011 Elsevier Ltd. All rights reserved.

1. Introduction Air blast freezing is the process of taking a product at a temperature (usually chilled but sometimes at ambient temperature) and freezing it rapidly, between 12 and 48 h, to its desired storage temperature which varies from product to product (e.g. sh 20  C, beef 18  C). Typically, the evaporator temperature in a blast freezer refrigeration system ranges between 35  C and 52  C. Slow freezing produces large ice crystals, which grow through cell walls, permitting an accelerated penetration of oxygen, causing rancidity and browning of meat and enhancing the danger of higher drip on thawing. Therefore, rapid freezing is required to maintain food quality as it produces small ice crystals due to a higher number of nucleation points from which ice crystals form. Air Blast freezing is classied as a forced convection phenomenon where the use of fans increases the products surface heat transfer coefcient and produces a more uniform air temperature throughout the freezer. The air velocity, and hence heat transfer coefcient, can be altered with the use of variable speed drives (VSDs). The main detriment of forced convection in blast freezers is the use of large fans that add signicantly to the total heat load on the refrigeration system and running costs. Also, unwrapped foods

are prone to moisture loss during blast freezing as the absolute humidity of the bulk air is usually lower than that of the air at the surface of the food. Although air blast freezers have been used in industry since the 1950s, limited number of technical studies have been published on specic aspects of the topic in the open literature [1,3,12,24,26, 31,42,51,52], and there is hardly any study that summarises all aspects of blast freezers at one point in a single study. Therefore, this paper presents an overview of blast freezers of their working principle, historical background, different designs, efciency issues, a modelling perspective and a best practice guide. 2. Origin of air blast freezers The early freezing rooms typically consisted of bare pipe grids in the ceiling above rails on which sheep carcasses and beef quarters were hung. These freezing rooms relied on the natural convection of cold air, typically around 15  C, and resulted in freezing times up to three days. Following World War II the world faced a serious food shortage. A major New Zealand innovation was the air blast freezer which enabled rapid freezing for high export quantities. The air blast freezer used fans to blow air at low temperatures (down to 30  C) over carcasses reducing freezing times to between 10 and 24 h. This ability to freeze and transport food to distant markets made refrigeration a highly protable trade and in fact made New Zealand one of the richest countries in the world in the 1950s and 1960s. The New Zealand company Ellis Hardie Syminton Ltd patented the A189 air-blast freezer in about 1950 [1]. The concept was to use

* Corresponding author. Current address: Building Equipment Program, Oak Ridge National Laboratory, P O Box 2008, Oak Ridge, TN37831-6067, USA. Tel.: 1 865 576 7376. E-mail address: p.bansal@auckland.ac.nz (P. Bansal). 1 Tel.: 1865 576 7376. 1359-4311/$ e see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2011.12.013

72

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

nned surface evaporators to cool the air and large fans to direct the refrigerated air over the carcasses. During the 1950s blast freezing became the dominant method of freezing worldwide and the fundamental principles of carcass freezing changed little over the next 50 years. Prior to the 1950s the beef trade was dominated by quarter beef carcasses. The emergence of the hamburger in America opened a new market for frozen boneless beef suitable for direct processing. In response to the new boneless product demand, Ellis Hardie Syminton Ltd in conjunction with Bill Freeman constructed the worlds rst continuous carton freezer for 27.2 kg beef cartons near Palmerston North (New Zealand). The freezer blew refrigerated air as low as 40  C at a velocity around 3 m/s over the cartons and achieved freezing times of 24 h. Over the next 20e30 years, the air-blast freezer became universal in the New Zealand frozen food industry. Several variations were developed, including cross ow and vertical air ow systems and there was a move from batch to continuous production for larger through puts and reduced labour costs. By the 1980s, energy efciency became an important design parameter. Many potential energy saving initiatives were investigated in terms of both the refrigeration system and the system as a whole. Such energy saving initiatives included: improved air ow design by altering the product stacking arrangement and the use of bafes and turning vanes, varying the air velocities at different times throughout the freezing process and the effect of product packaging on freezing times. The easiest and most advantageous energy saving device today is the use of variable speed drives (VSDs) on evaporator fans. 3. Why blast freeze? The killing of bacteria is largest in the range 4  C to 10  C due to cold-shock, where their metabolism is disturbed, even stopped. When the freezing rate is slow, the bacteria have time to adapt to the new conditions, hence food needs to be frozen quickly. There are various methods available for food freezing, these include: air-blast freezers (batch and continuous), uidised bed freezers, impingement freezers, liquid immersion freezers, plate freezers, liquid nitrogen freezers and carbon dioxide freezers. The major advantage of the air blast freezer is its versatility. Since air is a low viscosity uid it has the ability to easily follow around irregular surface geometries, thus providing a more uniform freezing rate over the whole product. Other freezing methods such as plate freezing (contact freezing) offer faster cooling times [2] but can only be used with products of a suitable geometry, i.e. a at surface to match the plate bed. 4. Types of air freezers Air is the most widely used method of freezing food as it is economical, hygienic and relatively non-corrosive to equipment [4]. Various forms of air blast freezers are used in industry [5e16]. 1. Sharp freezer: or blast room freezer is a cold storage room that relies on natural convection and low air movement from evaporator fans to circulate the cooling air resulting in slow freezing times. This arrangement is sometimes used for bulk products like butter, beef-quarters and sh, but not for processed food products. 2. Tunnel freezers: the refrigerated air is circulated by large fans over the product conned in an insulated closed room. Meat carcasses are supported by hooks suspended from a conveyor or specially designed racks. The trays or spacers are arranged to provide an air space between each layer of trays. The air can either be cross ow or counter ow, depending on the type of tunnel freezer. Various forms of tunnel freezers exist including:

2a Batch Freezers: the product is stacked on pallets, or hung from hooks on slide rails in the case of carcasses, and loaded into the freezer using fork hoists. This is an on/off process where the freezer is loaded, run until the meat is frozen to its desired temperature, then pumped down and switched off for unloading. Batch blast freezers are suitable for small quantities of varied products [17]. Typically, the heat transfer coefcient is less than 50 W/m2 K (Fig. 1). 2b Mechanised freezers: the pallet racks are tted with casters or wheels. The racks or trolleys are usually moved on rails by a pushing mechanism, usually hydraulically powered. Such mechanised tunnel freezers are known as pushthrough tunnels or carrier freezers which have two tiers, one on top of the other. These freezers are designed primarily for packaged goods, as well as carcasses. Advantages of mechanised freezers over batch freezers include: improved air circulation over the product as it moves at a steady rate through the tunnel; labour costs are considerably decreased as pallets are not manually placed in the freezer; and there is added exibility of the facility by varying the freezing time with the speed of the ram [18]. Heat transfer coefcients in mechanised freezers are similar to batch freezers being less than 50 W/m2 K. 2c Belt freezers: the product is loaded on a continuous conveyor belt. Modern belt freezers usually employ vertical air ow to force air between the product items creating good contact with the product. Typically, the heat transfer coefcient of belt freezers varies between 25 and 80 W/m2 K. Multi belt freezers offer the advantage of smaller oor space compared to single belt freezers. There are several forms of belt freezers: i) Multi-tier belt freezers: consist of several conveyor systems positioned one above the other with fans and coils positioned above the top belt. The air ow in belt freezers can either be vertical or horizontal over the product. The most efcient ow is determined by the product characteristics, dimensions, packed or unpacked, as well as degree of processing and composition. ii) Spiral belt freezers: where the belt is coiled in numerous revolutions around one vertical central axis to optimise the use of oor space. The belt can stack 30 tiers or

Fig. 1. Schematic of a typical batch air blast freezer.

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

73

more, one above the other thus reducing oor space to a minimum. Spiral freezing is one of the most currently used methods in the freezing industry for large production needs due to its convenience, reduced oor space, exibility and efciency [19] (Fig. 2). 3. Fluidised bed freezers are used to freeze particulate foods of uniform size and shape such as peas, cut corn, diced carrots, and strawberries. The foods are placed on a mesh conveyor belt and moved through a freezing zone in which cold air is directed upward through the mesh belt and the food particulates begin to tumble and oat. This tumbling exposes all sides of the food to the cold air, thus the product is individually quick frozen (IQF). Typically, heat transfer coefcients range from 110 to 160 W/m2K (Fig. 3). 4. Impingement jet freezers are straight-belt freezers involving only one step where the top, or more generally, both faces of the product receive very high velocity air at low temperature via uniformly distributed nozzles. The jets break the stagnant boundary layer surrounding the product, leading to a considerable increase in the heat-transfer coefcient, up to 300 W/m2 K. The performance is comparable to cryogenic freezers in relation to freezing time and weight, but at a much lower cost (typically half the price). Table 1 summarises the characteristics and operating parameters of the freezers described above. The cooling air temperature for each freezers ranges between 30 and 45  C. With similar cooling air temperatures, it is the air velocity over the product that is the main factor affecting the heat transfer coefcient. 5. Products blast frozen Typical products frozen in air blast freezers include but are not limited to:  Meat e carcasses, cartons, large individually wrapped cuts, cured or processed, hamburger patties  Poultry e whole bird or pieces, processed or breaded products  Fish e whole or eviscerated, llets or small diced pieces, processed or breaded products, shellsh, prawns and shrimp  Fruits e small size (whole), large size (sliced), pure or pulp  Vegetables e small and medium size, leafy  Other e Cheese and butter, dough, bread and baked products, pre-cooked ready meals Selecting which freezing method to use is usually determined by quality specications, economics and availability. Each food product has its own unique characteristics which determine their appropriate freezing temperature and freezing rate. Seafood, such as prawns, requires faster freezing rates than red meat to maintain their texture and taste. As such, prawns are suited to Individual

Fig. 3. Schematic of a uidised bed freezer.

Quick Freezing (IQF) methods such as uidised bed freezers. Red meat however, does not require IQF freezing methods to maintain quality and can be adequately frozen in tunnel freezers. 6. Packaging It is common practice to freeze meat or sh products in their transport packaging. Packaging is important in air blast freezing as it prevents dehydration, freezer burn and adherence by freezing and oxidation. The detriment of packaging is a decrease in heat transfer and hence an increase in the freezing time due to the insulating properties of the packing material and excess enclosed air. Fig. 4 shows a typical temperature versus time graph of packaged and unpackaged meat products in an air-blast freezer. Fig. 4 clearly illustrates the increased freezing time as a result of packaging. Stage 1 refers to the sensible cooling from the product inlet temperature (usually chilled) to freezing, stage 2 the latent heat extracted during crystallisation and stage 3 the sensible cooling from the freezing temperature to the desired storage temperature. Table 2 shows the heat transfer resistance of packaging for frozen beef and sh. It may be noted from the Table that the heat transfer resistance due to packaging can account for up to 59% of the total resistance when sh is the product being frozen. For beef, the portion of heat transfer resistance from packaging is less at 38%, but still signicant. Food Packaging must perform three functions: i) Control the local environmental conditions to enhance storage life. This is usually met by the packaging layer closest to the food. Typical examples include sealed plastic lm and tin-plated cans. ii) Display the product in an attractive manner for the potential buyer. iii) Protect the product during handling and transit. Corrugated cardboard is commonly used which unfortunately is a very good insulator. To reduce freezing times, cartons should use single layer cardboard with a high heat transfer coefcient on the top and bottom as this is where the surface area is largest. 7. Air blast freezer operation and design Air blast freezers are designed to supply cool air over the food product with a uniform air velocity throughout the freezer [23e28]. Most operation problems are related to improper positioning of the pallet or cart in the freezer [29]. Therefore, it is imperative the pallets and products are stacked in such a way that the air is free to move over the entire product. The stacking method must enable the cold air to circulate between the trays or boxes unhindered. For carton freezing, a spacer up to 70 mm should be implemented to allow sufcient air velocity between cartons [30]. Boast [18] recommends an air space equal to approximately 50% of the product thickness. Air temperature must be at least 35  C, and in some cases 45  C [31]. This equates to a refrigerant evaporation temperature of 42  C and 52  C respectively.

Fig. 2. Schematic of a typical spiral belt freezer.

74 Table 1 Summary of forced convection freezing methods. Freezer type Product

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

Air velocity 1.5-6 m/s Typically z 4 m/s

H.T.Ca W/m2 C Capacity h < 50 [20] 1e80 tonnes

Advantages

Disadvantages i) Long freezing times ii) Relatively low H.T.C i) Requires additional space ii) Reduced freezing capacity due to frost on evap. coils i) More expensive than tunnel freezers ii) Hygiene issues i) Only suitable for small products of fairly uniform shape and size i) Only suitable for products of small thickness

Batch tunnel Useful for all foods but better for bulk items, particularly carcasses

i) Low capital cost ii) Versatile, can accommodate various product geometries Continuous Useful for all foods but better for 1.5e6 m/s h < 50 [20] 1000e20,000 i) Reduction in down time as tunnel bulk items. Mainly suited to Typically z 4 m/s kg/hr the freezer is not stopped for packaged product due to hygiene loading/unloading issues ii) Flexible with freezing times Spiral Suitable for most foods, packaged or 3e8 m/s h z 25e80 500e6000 kg/hr i) Compact unpackaged e.g. poultry, red meat, [21] ii) Capable of IQF sea, bakery product iii) Higher efciency than tunnel Fluidised bed IQF small products, .5e5 cm z30 m/s h z 110e160 100e20,000 i) Very fast freezing times, diameter, e.g. peas, French fries, [21] kg/hr comparable to cryogenic only shrimp, scallops, diced meat, meat cheaper balls ii) High efciency 10e100 m/s h z 250e350 Depends on i) Reduced moisture loss Impingement IQF. Meat patties, sh llets, Typically z 40 m/s [21] application, ii) Very fast freezing times, shrimp, French fries. Product can be up to similar to cryogenic thickness typically 0e25 mm1 1200 kg/h Heat transfer coefcient.

There are several optimum air velocities that are used for tunnel blast freezers in the open literature, depending upon the particular product being frozen; however, the generally accepted value is 4 m/ s. Although increasing the air velocity will increase surface heat transfer coefcient, it does not necessarily reduce cooling time due to the increased heat load from the fans because fan power, W f V3. This increase in fan power increases the running cost usually rendering the increased fan speed uneconomical when compared to slower speeds. Furthermore Kolbe et al [32] showed that increasing the air velocity above 5 m/s only barely increased the freezing rate. This is because partway through the freezing cycle when the surface layers are frozen, the rate of heat transfer is increasingly controlled by the internal conduction resistance, i.e. the Biot number becomes large. When sizing evaporators for tunnel blast freezers, a frost buildup factor must be considered with n spacing of no more than 4 ns per inch [34]. When air coolers are mounted above a false ceiling, logarithmically spaced air deectors can be installed to help deect the air through the 90 turns and help distribute a uniform airow over the products. Modern spiral freezer designs eliminate any type of structure and belt support and each tier is supported directly on the previous one (self-stacking belt). The temperature of the refrigerated air is

below 30  C, generally being closer to 40  C, with a circulation velocity ranging from 3 to 8 m/s. In simple designs the air ow direction relative to the belt can be horizontal, parallel or vertical (both upwards and downwards). Further design improvements implementing the use of bafes and ow dividers can provide air ow vertically upwards through the lower half of the stack and downwards through the upper half (controlled dual ow). This balances the heat transfer on the two sides of the food, and slightly decreases freezing time and weight loss. The current state of the art developments are focused on impingement freezers, dual air systems [35] and improving the air ow distribution throughout air blast freezers with the aid of computational uid dynamics (CFD). Various studies [36e43] with CFD application to air blast freezers have been performed in the open literature. CFD delivers detailed information e both in time and space e of the ow eld, the temperature and moisture distribution, the shear forces and the heat uxes. Furthermore, computer visualisation gives a direct insight in the process, which allows a fast interpretation of any possible problem. Finally, the model-based procedure allows the evaluation of many what if scenarios at little cost compared to the process of prototyping.

8. Product geometry Product geometry plays a signicant role in determining freezing time. Most meat plants use a standard carton depth of 160e165 mm. A reduction in carton depth can signicantly reduce freezing time. The most important areas of the carton are the top

Table 2 Heat transfer resistance of packaging [22]. Heat transfer resistance (m2K/W) Source of heat transfer resistance Convective boundary layer external to carton Carton wall Nominal 1 mm layer of trapped air between carton and product Product itself between surface and geometric centre Total Heat transfer resistance due to packaging system (%) Frozen sh .04 .06 .04 .03 .17 59 Frozen beef .04 .02 .04 .06 .16 38

Fig. 4. Freezing curves for packaged and unpackaged meat product in an air blast freezer.

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

75

and bottom as this is where the surface area and hence the heat transfer is largest. A New Zealand meat company decided, for logistic purposes, that the base dimension of their meat cartons should be the same. To maintain a constant weight for each carton, the cartons height was varied to cater for different product densities. Cleland [22] investigated the effects of the different carton heights on freezing time and found freezing time varied between a linear and a quadratic relationship with height if the air convection heat transfer coefcient was unchanged. The tall cartons obstructed the air ow channels in the freezer more than the short cartons, thus they had less air ow over them. Tan et al. [44] set out to determine the factors affecting the freezing process of tilapia llets of different geometries. A numerical model based on the continuity equation, momentum equations and energy equation was developed. Five different geometries of equal mass were tested, three sh cakes and two spherical: slab, elliptical, disc, spherical and cylindrical respectively. The freezers air velocity was set at 5 m/s and temperature at 35  C. The three sh cakes had very similar freezing times ranging from 1.167 (disc) to 1.233 h (slab) due to their similar thickness and surface area. The freezing time for the sphere was 3.7e3.9 times longer than that of the at shaped llets due to the difference in surface area and distance from surface to centre. The cylindrical shaped llets produced the longest freezing time of 5 h. 9. Single and two stage air blast freezers Schematic of a two-stage carton blast freezer operating in New Zealand is shown in Fig. 5. The system consists of a 63 kW Vilter VMC-440 two-stage compressor with desuperheating and sub-

cooling, a Miller air cooled condenser with ve pressure controlled fans, receiver, suction accumulator with sub-cooling, two 4.0 kW xed speed evaporator fans, two custom made four pass evaporators and seven Danfoss expansion valves. Six of the expansion valves, rated at 10.2 kW, are used on the evaporators (three per evaporator), the remaining valve is used for the intercooling. The system uses refrigerant R22. The pressure enthalpy (P-h) diagram for the two-stage blast freezer system shown in Fig. 5, is shown in Fig. 6, with actual data taken from the site, where the advantage of two-stage compression and sub-cooling is illustrated vividly on the P-h diagram in Fig. 6 (a) with reduced compressor work and an increased evaporator capacity. The volumetric efciency of a reciprocating compressor is inversely proportional to the compression ratio. The ability of the two-stage system to split the evaporating and condensing pressure difference over two stages means the compressors are operating with a lower pressure ratio, thus a higher volumetric efciency than the single-stage system. Furthermore, by desuperheating the LSC discharge refrigerant the HSC discharge temperature is lower than the single-stage compressor. This means the two-stage system is less prone to oil breakdown and fatigue on compressor components. Intercooling the liquid line enables the refrigerant to enter the evaporator at a lower quality. Since the heat transfer coefcient of a two-phase mixture is signicantly higher than the vapour, intercooling the evaporator leads to higher heat transfer coefcient than that of the single-stage system, thereby, increasing the evaporator efciency. Fig. 7 shows an older system, installed in approximately 1980. This is a single-stage system consisting of two stand alone refrigeration units, Sub-System One (SS1) and Sub-System Two (SS2).

Fig. 5. Schematic of a two-stage carton blast freezer operating in New Zealand.

76

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

Fig. 6. (a) Two-stage P-h diagram for system shown in Fig. 5 and (b) corresponding theoretical single-stage P-h diagram operating with the same suction and discharge pressures.

Each sub-system consists of two semi-hermetic reciprocating compressors, oil separator, water cooled shell-and-tube condenser, thermostatic expansion valves, evaporators and a suction accumulator. Both systems use refrigerant R22. The freezer has eight 4.0 kW evaporator fans and racks for suspending meat carcasses. Sub-System One has two Copelametic compressors, approximately 30 years old, rated at 26 kW each operating in parallel. These compressors have older reed valves. Sub-system Two has two DWM COPELAND compressors, approximately 10 years old, rated at 20.1 kW operating in parallel. These compressors have the more modern discus valves. Both SS1 and SS2 have three pass water cooled shell-and-tube condensers mounted underneath the compressor racks. The water cooling the condensers is fed from the mains which supplies

water for all the water cooled systems operating at the facility. The heated water at the outlet of the condensers is returned to the mains before entering the cooling tower. The water for all systems is driven by a single centrifugal pump. SS1 has three evaporators, one large and two small, each with different geometries. The large evaporator is fed by two TXVs, the two small evaporators are fed by one TXV each. SS2 has two large evaporators each fed by two TXVs. The COP for the two-stage and single-stage systems is respectively dened by Eqs. (1) and (2))Two-stage Coefcient of Performance:

COP2stg

_ LSC h2 h1 m _ LSC h2 h1 m _ HSC h4 h3 m

(1)

Fig. 7. Schematic of a single-stage carcass freezer operating in New Zealand.

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

77

COP2stg 1.73 Single-stage Coefcient of Performance:

calculating UA for a single tube (i.e. 1 of the 46 condenser tubes), then multiplying the UAsingle tube by the total number of tubes, i.e.:

COP1stg

_ ref h2 h1 m h h1 2 _ mref h1 h4 h1 h4

UASH;2Ph;SC Ntubes $UAsingle tube;SH;2Ph;SC (2)

(7)

COP1stg 1.30 Splitting the refrigeration cycle into two stages and sub-cooling resulted in 33% increase in COP from 1.30 to 1.73. 10. Simulation model of the two stage system A steady state simulation model of the two-stage system was developed to analyse the performance of the refrigeration System using the Engineering Equation Solver [33] package to optimise the system for maximum efciency. The main model was based on individual components including Compressors, condensers, evaporators etc. using equations and correlations from the open literature. 10.1. Compressor model The low and high stage compressors were modelled by the method proposed by Popovic and Shapiro (1995), where the geometry of both the low and high stage compressors was identical (bore, stroke, rpm etc.). The refrigerant mass ow rate is calculated by:

The model sequence starts by guessing the fraction of the tube length required to de-superheat the refrigerant, FSH (FSH (FSH < 1). To calculate the resistance over the superheated zone, each full resistance, that is the each of the ve resistances over the full tube length (Rfull), is divided by the fraction of the tube length (FSH) used for that particular section, The total resistance is the sum of the individual resistances. The conductance of the superheat zone, as calculated by the model, is given by:

1 UASH;model Ntubes $ Rtotal;SH

(8)

where Rtotal,SH is the sum of the superheat resistances. In order to verify the initial guess of the superheat fraction FSH. The actual rate _ SH ) is given by: of heat transfer over the superheat section (q

_ SH m _ 3 hin hsat ;x1 q

The NTU conductance is calculated by:

(9)

_ UASH;NTU NTUSH $C min

(10)

_ m

u$PD
60$vsuc

 1CC

Pdis Psuc

1=n  (3)

where piston displacement (PD) is the volume actually swept during one cycle. The clearance volume C is the refrigerant volume left in the cylinder after completion of the discharge process is calculated by:

Vdis Vsuc Vdis 2

(4)

The compressor work is calculated by:

0 13  n1 n n P dis _ m _4 1A5 W Psuc vsuc @ Psuc n1

(5)

The model conductance UASH,model is checked against the NTU conductance UASH,NTU. The initial guess of the fraction of the tube length required for desuperheating, FSH, is iterated until UASH,model and UASH,NTU converge. Upon convergence, the fraction of tube length required for desuperheating is known and the process is repeated for the two-phase section. The two-phase zone analysis starts by guessing the fraction of the tube length required for full refrigerant condensation (latent heat), F2Ph. The resistances are solved for the tube length L2Ph corresponding to F2Ph. The refrigerant saturation temperature and the HSC discharge pressure are used to determine the refrigerant properties. The average heat transfer coefcient during condensation in horizontal tubes, required to determine the refrigerant-side resistance in the two-phase zone, is calculated using the correlation suggested by Dobson and Chato [53]. The refrigerant side resistance is determined with the area corresponding to the two-phase length, L2Ph. The resistances are calculated in the same manner as described for the superheated region using L2Ph. The conductance of the two-phase zone, as calculated by the model, is given by:

The polytropic exponent n, in Eq. (5), is a function of the discharge specic volume (a model output), and is calculated from experimental data.

1 UA2Ph;model Ntubes $ Rtotal;2Ph

(11)

logPdis =Psuc logvdis =vsuc

_ 2Ph , is The actual rate of heat transfer over the two-phase zone, q calculated by:

(6)

_ 2Ph m _ 3 hx1 hx0 q

(12)

_ 1 and the The compressor model determines the LSC ow rate m _ 2 . The intercooler ow rate is calculated as the HSC ow rate m difference between the HSC and LSC ow rates. 10.2. Condenser model The condenser was modelled by dividing into zones corresponding to the refrigerant state (superheated, two-phase, subcooled). The condenser consisted of 46 individual tubes (Ntubes) and 6 tube passes (Npasses). The -NTU method is used to evaluate the heat exchanger performance. The conductance, UA, for each zone is determined by

where hx1 and hx0 refer to the vapour and liquid saturation enthalpies respectively. The air-side capacitance rate is the minimum capacitance rate in the condensing section as the capacitance rate of the condensing refrigerant is effectively innite. Therefore, the maximum possible heat transfer rate in the condensing zone is:

_ _ max;2Ph C q air ;sat TR;sat Tair ;in

The NTU conductance, UA2Ph,NTU, is calculated by:

(13)

_ UA2Ph;NTU Ntubes C minNTU 2Ph

(14)

78

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

The initial guess of the tube length required for the two-phase region F2Ph is iterated until UA2Ph,model and UA2Ph,NTU converge. Upon convergence, the fraction of tube length required for phasechange is known and the process is repeated for the sub-cooled zone. With FSH and F2Ph known, the fraction of tube length associated to sub-cooling is calculated by:

FSC 1 FSH F2Ph

(15)

The six resistances are calculated using the sub-cooling tube length given by:

a thermodynamic model is used to simulate the desuperheating and sub-cooling processes of the intercooler. Fig. 6a shows the log P-h diagram of the two-stage system with the state points used for modelling. State points 1, 2, 4, 5, 6 and the _ evap are calculated from the _ cond and m _ evap m mass ow rates m compressor, condenser and expansion valve models coupled with the evaporating and condensing pressures given as model inputs. The intermediate pressure is calculated as the geometric mean of the condensing and evaporating pressures given by:

Ltube;SC FSC $Ltube;total

The conductance of the sub-cooled zone is calculated by:

(16)

Pint

q Pevap $Pcond

(22)

State point 3 is determined from the fraction of superheat x, given by

UASC ;model Ntubes $

1 Rtotal;SC

(17)

h3 x h2 hsat ;Pint
where

(23)

The NTU over the sub-cooling section is calculated by:

NTUSC

UAmodel;SC _ C
min

(18)

0<x1

(24)

The rate of heat transfer over the sub-cooling section is determined by:

_ _ SC SC C q minTRef ;sat Tair;in

The fraction of superheat x is determined from the experimental data and can be adjusted for parametric analysis. The enthalpy at state point 7 is determined from an energy balance across the intercooler:

(19)

_ int h3 h6 m _ evap h2 h3 m _ evap h5 h7 Qint m

10.6. Complete system model

(25)

The enthalpy of the refrigerant leaving the condenser is determined by an energy balance across on the refrigerant side:

hout

_ q hsat;x0 SC _3 m

(20)

Finally, the total rate of heat transfer across the condenser is determined by:

_ SH q2Ph q _ SC _ cond;total q q

(21)

If FSC (from Eq. (15)) is less than one, the refrigerant leaves the condenser in a saturated state. Under these conditions the refrigerant enthalpy at the condenser outlet is calculated in the same manner as described for the sub-cooling zone. 10.3. Evaporator model The two-stage blast freezer has two identical wavy n-and-tube evaporators with a staggered tube layout. Each evaporator has three TXVs with a distributor further dividing the ow into 15 tubes, i.e. 15 tubes per TXV. This equates to 45 tubes per evaporator. Each tube has four passes. The evaporator is modelled in the same manner as the condenser. The evaporator is divided into zones corresponding to the refrigerant state. The refrigerant enters the evaporator in twophase state and leaves superheated. The air-side heat transfer coefcient is calculated for a six row heat exchanger, same as the condenser. The remaining resistances are calculated for four rows. 10.4. Expansion valves model The six evaporator expansion valves and the intercooler expansion valve are all of the mechanically controlled thermostatic type, TXV. The throttling of the liquid refrigerant is achieved by assuming the expansion is enthalpic. 10.5. Intercooler model The intercooler could not be modelled physically like the heat exchangers because the geometric data was unavailable. Therefore,

The system model consists of the individual component models linked together. Fig. 8 shows the ow chart of the complete system model. The condenser model uses two loops to determine the fraction of the tube length required for the superheat and twophase sections according to the condenser inlet air temperature, air mass ow rate and the refrigerant condensing pressure. The evaporator model has one loop used to determine the fraction of the tube required for the two-phase section according to the freezer air inlet temperature, air mass ow rate and the refrigerant evaporating pressure. The complete model has an additional loop on the superheat to ensure the energy balance between the evaporator, condenser and compressors is met. The model is used as a tool to analyse the thermodynamic performance of the system (i.e. system COP). 11. Model validation and results The two-stage compressor model consisted of two individual _ 3 ). The compressor _ 1 ) and HSC (m compressor models, the LSC (m model predicted the LSC mass ow rate to within 7% of the nine measured mass ow rates. Similarly, the compressor model predicted the intercooler mass ow rate to within 13% of the measured mass ow rates. The modelled condenser capacity agreed to measured capacity to within 5% of the 13 measured data points, as shown in Fig. 7.3 (Fig. 9). The evaporator model was validated using the measured refrigerant mass ow rate. The outputs of the evaporator model are the refrigerant superheat and the outlet temperature of the air. The modelled evaporator capacity agreed with the measured capacity to within 4%. The system model was validated with the recorded data from the unloaded freezer. The experimental data was collected over a 4 h period where the evaporator load dropped from 75 kW to 45 kW. The ambient air over this period varied from 17.4  C to 19.7  C. Fig. 7.6 shows the model predicted COP against the measured COP, where the model COP agreed to within 8% of the 11 measured data points (Fig. 10).

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

79

Fig. 8. System model ow diagram.

It was very encouraging to note that the model predicted the evaporator air outlet temperature very accurately to within an error of 2%. This established the accuracy and the validity of the model, which now could be used to predict performance of a blast freezer as well as a design tool for future blast freezers in order to achieve higher energy efciency. 12. Design improvements Odey [47] investigated performance enhancing measures of a batch air blast freezer. He found that generalised rules of thumb have been used for the design of air ow through blast freezers. Critical aspects of the design and implementation of the airow circuit are often excluded from the refrigeration contract, resulting in poorly implemented and underperforming facilities. Typically, the refrigeration contractors response to poor freezer performance is to increase the fan capacity and power. It was found that simply increasing the air ow by increasing fan speed did not necessarily increase the air speed through the cartons in the

freezer. The higher fan speed resulted in negative velocities at the fan inlet due to the formation of a large unstable vortex. As a result more heat was added to the freezer from the fans thus reducing the efciency. The following modications were installed on the air blast freezer:     Bafing on the top and sides of the freezing chamber Fan inlet cone and diffuser Air inlet and discharge vanes on corners Variable speed drive on fan

Prior to the modications the air ow entering the fan was highly unstable with signicant ow reversal. This turbulence reduced signicantly with the above modications. Most of the pressure drop in the unmodied freezer occurred at the 90 turning points, whereas the modied freezer had the largest pressure drop through the product pallets. As a result of the experiment, the existing 11 kW fan motors drawing 8.7 kW were replaced with

80

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

Fig. 9. Comparison of the condenser model capacity against measured condenser capacity. Fig. 11. Product heat load characteristics for hot-boned and cold-boned beef cartons [47].

3.5 kW motors drawing 4.0 kW. The fans were re-pitched from 30 to 22 to maintain drawn power within the motor capacity. After the modications satisfactory freezing was being achieved within the specied 48 h turnaround period. Kemp and Chadderton [48] performed a study on the performance of batch blast freezers used to freeze beef cartons and found that designs seem to be based on average product heat load which is insufcient to handle the initial peak heat load. This problem is especially prevalent with hot-bone meat. Insufcient cooling capacity generally occurs at the beginning of the freezing process when the product heat load is being released at a peak rate that far exceeds the average rate, see Fig. 5. This problem is compounded as very few blast freezer systems manage to maintain their design cooling capacity (Fig. 11). Bowater [34] states it is necessary to size evaporators at least 50% higher than the average refrigeration load for 24 h freezes to account for the high initial heat load, while this requirement is not so critical for 48 h freezes. Other factors effecting cooling time include overloading of the freezers that often results in higher cooling loads and therefore longer freezing times. Changes in product packaging have to be taken into account when sizing air blast freezer throughput. Mannapperuma et al. [49] found the surface heat transfer coefcient of whole, unpackaged chickens reduced by an order of magnitude when the chickens were wrapped in plastic and stored in vented boxes. Kemp and Chadderton

[48] surveyed a plant which changed the type of cardboard packaging used. As a result freezing time was increased by 8 h. The change of packaging was determined as the major cause of the plants freezing problems. The fan load in old carcass freezers can account for up to 60% of the total refrigeration load [50]. Wee et al [51] installed variable speed drives (VSDs) on the fans of a 4000 lamb carcass capacity blast freezer and reported a 44% energy saving. The VSD was controlled by a personal computer where the program continuously analysed data inputs such as air temperature and calculated the optimum air ow velocity. The VSD payback period was 2.1 years. Other benets included a more uniform product quality, improved power factor for the freezer fans and minimization of product weight-loss due to the lower air velocity. Kolbe et al [32] investigated the effects of bafing and various fan speed control on air blast freezer performance. The 8.7 tonne capacity freezer was used for 10 kg sardine cartons. The system had three 5.6 kW fans mounted in the false ceiling downstream of the evaporator. Typical freezing times were around 12.5 h. The following modications were applied to re-direct and channel the air ow:  Plywood on upper supply-side corners prevented air from sweeping around the upper end of the ceiling structure  The ceiling was lowered to 75 mm to reduce the gap between ceiling and product  Plastic sheeting on the supply-air side sealed the horizontal corner where the near-vertical and horizontal ceilings meet, and at the vertical corners between wall and cartons  Floor-to-ceiling plywood sheets, installed at the start of each freezing cycle, covered the ends of the racks and prevented end-around by-pass Prior to the modications, analysis of the ow velocities showed 35% of the air went through the carton racks, 15% owed over the top of the racks and 50% by-passed around the two sides. Air velocity ranged roughly from 1.5 to 4.0 m/s. After the modications velocities ranged from 3.0 to 4.0 m/s, hence the average velocity through the product increased. The results of the bafing are shown in Table 3. The bafing reduced maximum freezing time by 15%, fan energy usage by 6% and uniformity, the difference between maximum and minimum freezing times of individual cartons, improved signicantly.

Fig. 10. Comparison of model COP against measured COP.

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83

81

The second part of the experiment was to test effects of fan speed control on both freezing time and energy use. Tests for three fan schedules were performed:

Table 4 Results (averages) for fan loading trials [32]. Max. time Trial 1 Trail 2 Trial 3 11.4 11.4 11.9 Min. time 7.9 7.4 7.4 Difference 3.5 4 4.5 Fan energy (kWh) 151 167 197 Total energy (kWh) 819 873 847

Trial 1 all 3 fans come to full speed within 45 min. After 3.5 h, the fans are slowed to 75% speed for the remainder of the freezing period. Trial 2 fans came to full speed slightly faster than trial 1. Fan speed was halved after 7 h. Of these two trials, the rst saved more energy and would be the preferred option. Trial 3 the three fans started at 30 min intervals. After 4 h the centre fan was switched off, reducing fan power by 1/3. This method does not require the use of VFDs. The results for the three trials are shown on Table 4. Trial 1 performed the best reducing the maximum freezing time by an hour (8%) and reduced total energy use by 22%, based on the room performance prior to modications.

 Optimise defrost frequency and duration Other measures to improve blast freezing efciency include:  Improve air ow design to reduce fan power for the same effective air velocity over the product, e.g. use of air turning vanes, at inlet and outlet cones, bafes to prevent air ow short circuiting away from the product  Increase the time available to freeze the product so can operate the freezer at lower air velocities and higher air temperatures  Once freezing is completed, reduce fan speeds and increase temperature set points to storage temperature until unloading can occur  Load product so that the air ow distribution remains uniform throughout the freezer  Defrost coils a short period of time after loading a batch freezer so that the coils operate lightly frosted for most of the time Declining prot margins are forcing cold storage companies to employ energy savings initiatives, load management strategies and more efcient technologies. The most common energy saving measure is the use of off-peak electricity. Variable speed drives (VSDs) on compressors and blast freezer fans were identied as the most easily implemented energy saving new technology. Ambient air defrost systems are becoming more common rather than water or hot gas. Table 5 summarises the ndings of a report on best practise guide to industrial refrigeration produced by Sustainability Victoria [52]. The report emphasised taking the whole-system approach when designing new systems as this presents the greatest opportunity to incorporate energy efciency throughout the whole process, unhindered by the constraints that may be posed by existing equipment. The whole-system approach entails considering the system operation as a whole rather than just focusing on individual components as each component has owon effects that impact on other components, and therefore the efciency of the system as a whole. The report recommends the use of a control system that is responsive to the compressor head pressure. Electronic expansion valves should be used where possible and have their controls linked to the head pressure control system.

13. Energy usage and best practice guide for blast freezing Air blast freezing consumed 8.1 GWh of electricity in New Zealand in 2005 [45] and is the most energy intensive operation in the frozen storage industry. Apparent energy use for blast freezing was calculated as 133 kWh/tonne from regression analysis. This is 50% higher than the predicted value from theoretical best practice considerations. The New Zealand Cold Storage Industry identied blast freezing as an area where a 15% saving could be achieved for many sites, particularly related to reduction in fan power due to improved air ow design. Comparison with overseas survey results showed the NZ use was similar on average. If all facilities surveyed met the theoretical best practice energy consumption limit for blast freezing, this would represent an average energy saving of 33% per tonne of blast frozen product. This gure is supported by a survey on energy efciency of food refrigeration operations funded by the UK Government Department for Environment, Food and Rural Affairs (defra), who identied blast freezing as an area where a 20e30% energy saving could be achieved [46]. The New Zealand survey covered 13 sites that carried out blast freezing and recorded data over the duration of at least one year. These sites have a wide range of refrigeration systems from multi-stage pump circulation ammonia to single-stage direct expansion uorocarbon systems. The most common refrigerant was ammonia, used at 71% of the sites. The following measures were identied as potential energy saving solutions:     Reduce discharge pressure set points Raise suction pressure set points Variable speed drives (VSDs) for fans Improved door protections and management

Table 5 Potential energy savings for industrial refrigeration systems [52]. Method Electronic expansion valves VSD on compressor motors VSD condenser fans Reduced temperature lift Conversion from liquid injection to external oil coolers Refrigeration system replacement If over 10 years old Refrigerant selection Potential saving (energy, unless stated otherwise) 20% 20% 2e3% of total refrigeration cost 3e4% improvement for 1  C reduction Over 3% Up to 30e40% 3e10%

Table 3 Effect of bafing on blast freezer cell [32]. Max. time Unbafed Bafed Difference 12.5 10.6 1.9 Min. time 8.7 9 L0.3 Difference 3.8 1.6 2.2 Fan energy (kWh) 282 266 16 Total energy (kWh) 1054 924 130

82

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83 [27] A.B Pearson, R.A Lamb, 2001. Developments in refrigeration technology for spiral and blast freezers. Proceedings of the International Institute of Refrigeration e Rapid Cooling of Food, Section 2, pp. 153e161. [28] W.A Johnson, F.J Nicholson, A. Roger, G.D Stroud, 1994. Freezing and refrigerated storage in sheries. FAO Fisheries Technical Paper 340. ISBN 92-5103579-2. [29] G.C. Briley, Blast freezing, ASHRAE Journal 44 (7) (2002) 99. [30] J.E. Duiven, P. Binard, Refrigerated Storage: New Developments (2002) Bulletin of the IIR, No. 2002e2. [31] C.P. Mallett (Ed.), Frozen Food Technology, Blackie Academic & Professionals, 1994, ISBN 0 7514 0072 6. [32] E. Kolbe, Q. Ling, G Wheeler, 2004. Conserving energy in blast freezers using variable frequency drives. 26th National Industrial Energy Technology Conference, pp. 47e56. [33] F-Chart Software, Engineering Equation Solver, Wisconsin, USA retrieved from: (2011) www.owchart.com. [34] F.J. Bowater, 2001. Design of carton air blast freezing systems. Proceedings Of the International Institute of Refrigeration e Rapid Cooling of Food, Section 2, pp. 203e207. [35] D.W. Everington, 2001. Development of equipment for rapid freezing. Proceedings Of the International Institute of Refrigeration e Rapid Cooling of Food, Section 2, pp. 173e180. [36] B. Xia, Da-Wen Sun, Applications of computational uid dynamics in the food industry: a review, Journal of Computers and Electronics in Agriculture 34 (2002) 5e24. [37] J. Li, J. Xie, Numerical simulation of freezing time of shelled shrimps in an air blast freezer and experimental verication, Transactions of the Chinese Society of Agricultural Engineering 25 (4) (2009) 249e252. [38] Z. Hu, Da-Wen Sun, CFD simulation of heat transfer for predicting cooling rate and weight loss of cooked ham during air-blast chilling process, Journal of Food Engineering 46 (3) (2000) 189e197. [39] Z. Hu, Da-Wen Sun, Effect of uctuation in inlet airow temperature on CFD simulation of air-blast chilling process, Journal of Food Engineering 48 (2001) 311e316. [40] B.M. Nicolai, P. Verboven, N. Scheerlinck, M.L. Hoang, N. Haddish, 2001. Modelling of cooling and freezing operations. Proceedings Of the International Institute of Refrigeration e Rapid Cooling of Food, Section 3, pp. 211e215. [41] A. Foster, M. Swain, 2001. Using computational uid dynamics to predict airow in and through the entrances of cold storage rooms. Proceedings Of the International Institute of Refrigeration e Rapid Cooling of Food, Section 3, pp. 287e293. [42] Z. Hu, Da-Wen Sun, Predicting local surface heat transfer coefcients by different turbulent k- models to simulate heat and moisture transfer during air-blast chilling, International Journal of Refrigeration 24 (7) (2001) 702e717. [43] Z. Hu, D.W. Sun, J. Bryan, 1998. Modelling of an experimental air-blast freezer using CFD code. Proceeding of IIF/IIR commissions B2, C2, D1, D2/3, vol. 6, pp. 395e400. [44] F.-L. Tan, S.-C. Fok, Freezing of tilapia llets in an air blast freezer, International Journal of Food Science & Technology 44 (2009) 1619e1625. [45] S. Warner, F. Vaino, D. Cleland, Survey of Energy Use by the New Zealand Cold Storage Industry, College of Sciences Massey University, Palmerston North, New Zealand, 2005. [46] S.J. James, M.J. Swain, T. Brown, J.A. Evans, S.A. Tassou, Y.T. Ge, I. Eames, J. Misseneden, G. Maidment, D. Baglee, 2009. Improving the energy efciency of food refrigeration operations. The Institute of Refrigeration, Proc. Inst. R. 2008e09, 5.1. [47] M. Odey, 2006. Meat carton blast chilling/freezing cabinet performance improvements. Proceedings IIR-IRHACE 2006 Conference, pp. 118e124. [48] R.M. Kemp, T. Chadderton, 1992. Blasting ahead with cartoned beef. Refrigeration and Energy Section, MIRINZ Conference proceedings. [49] J.D. Mannapperuma, R.P. Singh, D.S. Reid, Effective surface heat transfer coefcients encountered in air blast freezing of whole chicken and chicken parts, individually and in packages, International Journal of Refrigeration 17 (4) (1994) 263e272. [50] A.K. Fleming, B.F. Edwards, D.W. Oliver, 1976. Evaluation of automatic carcass freezers. Proceedings 18th meat Ind. Res. conference, No 519, pp. 61e74. [51] H.K. Wee, M.P.F. Loeffen, Q.T. Pham, 1991. Fan speed control in a blast freezer using model reference control. Proceedings of 18th International Congress of Refrigeration, pp. 997e999. [52] Sustainability Victoria, Energy Efciency Best Practice Guide, Industrial Refrigeration, 2009 (2009).http://www.resourcesmart.vic.gov.au/documents/ BP_Refrigeration_Manual.pdf. [53] M.K. Dobson, J.C. Chato, Condensation in Smooth horizontal tubes, Journal of International Heat Transfer 120 (1) (1998) 193e212.

14. Conclusions This paper presented a snapshot of the use of blast freezers in the food industry, including historical background, modelling features, issues and recent advances that have enabled air blast freezers to play a signicant role in the meat industry due to their versatility and low capital cost. There seems to be a great potential to achieve energy savings in blast freezer industry by employing VSDs on fans that could result in with energy savings of up to 44%. Other, cheaper and simpler energy saving measures include air bafes, air turning vanes, fan inlet cone and outlet diffuser, and improved user operating procedures. When determining the heat load during the design process it is insufcient to use the mean product load. The peaky nature of meat, particularly hot boned meat, must be taken into account to ensure adequate freezing times are achieved. Evaporators should be sized 50% higher for 24 h freezers and hot boned meat. Also, the design capacity of the freezer will deteriorate with time. It is imperative the product is stacked with adequate spacing to enable air to pass through; failure to do so will result in localised regions of long freezing times. References
[1] A.C. Cleland, 2009. Refrigeration: underpinning the New Zealand economy for over 125 years. 3rd Australasian engineering heritage conference. [2] M.A. Hessami, 2004. A study of the efciency of plate freezing vs blast freezing of boxed boneless meat in an abattoir. Proceedings of the 2004 ASME Heat Transfer/Fluids Engineering Summer Conference, vol. 1, pp. 711e715. [3] Da-Wen Sun (Ed.), Handbook of Frozen Food Processing and Packaging, vol. 155, Taylor and Francis Group, 2006. [4] S.J. James, C. James, Meat Refrigeration, Woodhead Publishing in food and science and technology, 2002, ISBN 1 85573 442 7. [5] ASHRAE, Refrigeration ASHRAE Handbook (2006) 1-931862.87-7. [6] F. Toldr (Ed.), Handbook of Meat Processing, Blackwell Publishing, 2010 13:978-0-8138-2182-5/2010. [7] H. Ramaswamy, M. Marcotte, Food Processing: Principles and Applications, Taylor and Francis Group, 2006, 10:1-58716-008-0. [8] J. Rmy, Modern freezing facilities, International Journal of Refrigeration 10 (3) (1987) 165e174. _ Diner, Heat Transfer in Food Cooling Applications, Taylor and Francis [9] I. Publishers, 1997, ISBN 1-56032-580-1. _ Diner, Refrigeration Systems and Applications, John Wiley & Sons, Ltd, [10] I. 2003, ISBN 0-471-62351-2. [11] P.G. Smith, Introduction to Food Process Engineering, Kluwer Academic/ Plenum Publishers, 2003, ISBN 0-306-47397-6. [12] D.R. Heldman, R.W. Hartel, Principles of Food Processing. Chapman & Hall Food Science Book, Aspen Publishers inc, 1997, ISBN 0-8342-1269-2. [13] M. Kutz (Ed.), Handbook of Farm, Dairy and Food Machinery, William Andrew Publishing, 2007, ISBN 978-0-8155-1538-8. [14] G.D. Saravacos, A.E. Kostaropoulos, Handbook of Food Processing Equipment, Kluwer Academic/Plenum Publishers, 2002, ISBN 0-306-47276-7. [15] G.M. Hall (Ed.), Fish Processing Technology, second ed. Blackie Academic & Professional, 1997, ISBN 0 7514 0273 7. [16] G.S. Mittal (Ed.), Computerized Control System in the Food Industry, Marcel Dekker Inc, 1997, ISBN 0-8247-9757-4. [17] B.R. Becker, B.A. Fricke, Freezing Operations. Encyclopaedia of Food Sciences and Nutrition (2003) pp. 2712e2718. [18] M.F.G. Boast, Blast and Plate Freezing. Encyclopaedia of Food Sciences and Nutrition (2003) pp. 2718e2725. [19] C.O.B. Wallens, J. Venetucci, Emerging-Freezing Technologies. Elsevier Science Food Processing: Recent Developments (1995) pp. 227e240. [20] V.O. Salvadori, R.H. Mascheroni, Analysis of Impingement freezers performance, Journal of Food Engineering 54 (2) (2002) 133e140. [21] S. Gransson, G. Lndahl, 2001. Crustfreezing: a new method for rapidly chilling meat. Proceedings IIR, Com C2, Bristol, UK 2001. [22] A.C. Cleland, Package design for refrigerated food: the need for multidisciplinary project teams, Trends in Food Science & Technology 7 (8) (1996) 269e271. [23] Da-Wen Sun (Ed.), Advances in Food Refrigeration, Leatherhead Publishing, UK, 2001, ISBN 0 905748 83 2. [24] B.R. Becker, B.A. Fricke, Design Essentials for Refrigerated Storage Facilities, ASHRAE, 2005, ISBN 1-931862-74-5. [25] Z.B. Maroulis, G.D. Saravacos, Food Process Design, Marcel Dekker, Inc, 2003, ISBN 0-8247-4311-3. [26] K.J. Valentas, E. Rotstein, R.P. Singh (Eds.), Handbook of Food Engineering Practice, CRC Press, 1997, ISBN 0-8493-8694-2.

Nomenclature
_ LSC : low stage compressor mass ow rate, kg s1 m _ HSC : high stage compressor mass ow rate, kg s1 m hLSC,in: low stage compressor suction enthalpy, kJ/kg hLSC,out: low stage compressor discharge enthalpy, kJ/kg hHSC,in: high stage compressor suction enthalpy, kJ/kg

P. Dempsey, P. Bansal / Applied Thermal Engineering 41 (2012) 71e83 hHSC,out: high stage compressor discharge enthalpy, kJ/kg h1stgC in: single-stage compressor suction enthalpy, kJ/kg h1stgC out: single-stage compressor discharge enthalpy, kJ/kg hEXVin,1st: single-stage expansion valve inlet enthalpy, kJ/kg hEXVin,2nd: two-stage expansion valve inlet enthalpy, kJ/kg hsat: saturation enthalpy, kJ/kg ysuc: suction specic volume, m3/kg ydis: discharge specic volume, m3/kg Psuc: suction pressure, kPa Pdis: discharge pressure, kPa _ work (kW) W: COP2stg: two-stage coefcient of performance COP1stg: single-stage coefcient of performance UA: conductance, kW/ K R: thermal resistance,  K/W T: temperature,  K _ heat, kW q: _ capacitance, kJ/ K C: NTU: number of transfer units x: quality F: fraction, 0  F  1 N: number of

83