Air Coolers for Meat Chilling

Technical Papers31st Annual Meeting

International Institute of Ammonia Refrigeration

March 2225, 2009

2009 Industrial Refrigeration Conference & ExhibitionThe Hyatt Regency

Dallas, Texas

ACKNOWLEDGEMENT

The success of the 31st Annual Meeting of the International Institute of Ammonia

Refrigeration is due to the quality of the technical papers in this volume and the labor of its

authors. IIAR expresses its deep appreciation to the authors, reviewers and editors for their

contributions to the ammonia refrigeration industry.

Board of Directors, International Institute of Ammonia Refrigeration

ABOUT THIS VOLUME

IIAR Technical Papers are subjected to rigorous technical peer review.

The views expressed in the papers in this volume are those of the authors, not the

International Institute of Ammonia Refrigeration. They are not official positions of the

Institute and are not officially endorsed

International Institute of Ammonia Refrigeration

1110 North Glebe Road

Suite 250

Arlington, VA 22201

+ 1-703-312-4200 (voice)

+ 1-703-312-0065 (fax)

www.iiar.org

2009 Industrial Refrigeration Conference & Exhibition

The Hyatt Regency

Dallas, Texas

IIAR 2009 1

Technical Paper #7

Extended Surface Air Coolers for Industrial Plants the Contractors Perspective

Stefan S. Jensen, B.Sc.Eng. MIEAustScantec Refrigeration Technologies Pty. Ltd.

Brisbane, Queensland, Australia

Abstract

In view of the very large range of evaporator coil geometries, coil material combinations, coil defrost methods and circuiting options available on the market, the industrial refrigeration contractor often faces difficulties deciding which evaporator design to use for a certain application.

Often selections are based on rules of thumb, e.g., an allowance of a certain number of square feet of coil surface area per pound of product. Considering the fact that coil material selections, choice of refrigerant, coil circuiting and coil geometry can influence heat transmission coefficients (u-values) by up to a factor of three or more, such an approach can lead to very poor results in practice.

Other issues are relative humidity and dehydration of the refrigerated space. There are thousands of practical applications where the quality and shelf life of the products stored or chilled are directly influenced by the equilibrium relative humidity inside the room in question. Yet often extended surface air coolers are selected with geometries to suit the manufacturer and not the application. Frequently, the result is a squat coil when the coil should have been shallow and with large face area.

The performance impact of fouling both on the inside, but also on the outside surfaces of cooling coils is often not highlighted sufficiently by manufacturers. Extended surface air cooling coils featuring comparatively high-heat transmission coefficients (u-values) in clean condition generally display a more rapid performance deterioration as a function of increasing fouling than an equivalent coil with lower heat transmission coefficient in clean condition. It is often not in the commercial interest of coil manufacturers to disclose this information, yet it can be of crucial importance to the refrigeration plant designer.

It is commonly known that cooling coil performance increases as a function of increasing face velocity. What is not normally available from manufacturers is optimization of face velocities for maximum plant energy efficiency. This is the contractors problem and the tools that the contractor needs to perform this task are often not readily available.

This paper will show a range of practical performance comparisons between various coil geometries, coil materials, circuiting options, refrigerant choices, flow patterns and refrigerant feed options. These performance comparisons will have one thing in common namely the software used to perform them. The comparisons therefore represent relative information valuable to the contractor because it is information not readily available from extended surface cooling coil manufacturers under normal circumstances.

2009 IIAR Ammonia Refrigeration Conference & Exhibition, Dallas Texas

Technical Paper #7 IIAR 2009 3


Introduction

In modern industrial refrigeration applications, the extended surface air cooler is

one of the most common components. In terms of importance for the success of an

installation, it ranks at the level of compressors, condensers and control systems.

Although many software packages exist for the selection and rating of air coolers,

these are all designed for certain proprietary geometries and manufacturing ranges.

They are therefore not readily suitable for relative performance comparisons between

one manufacturer and the other.

The expertise of industrial refrigeration contractors relates to the ability to

successfully combine key plant components to form a functional refrigeration system

to suit a certain technical application. The selection of extended surface air coolers

can be a complex matter particularly if the outcome specified goes beyond simple

coil performance issues and extends into factors such as product quality, weight loss,

defrost frequency, overall energy efficiency, noise pollution, hygiene, temperature

and air flow uniformity within a facility, freezing times, part load behavior and other

similar performance outcomes for the system as a whole.

Not all extended surface air cooler geometries and layouts are suited optimally to

all applications. To some this may be a statement of the obvious and to some extent

it is. The statement does not refer to basics like selecting the correct fin spacing for

a freezing application. The statement makes the point that within one particular

plant it may be necessary to employ a wide range of cooler geometries and layouts

to achieve the optimal combination of components suitable for the application. This

may mean mixing coil manufacturers within the same system.

The marketing emphasis of cooler manufacturers is generally that their particular

product offered covers the widest range of applications. This is true for those

manufacturers who offer a wide range of geometries and construction materials.

Those manufacturers who for one reason or another limit the offering to one or

4 IIAR 2009 Technical Paper #7

2009 IIAR Industrial Refrigeration Conference & Exhibition, Dallas, Texas

very few coil geometries cannot cover as wide a set of applications. This paper will

examine the contractors perspective when it comes to selecting optimal extended

surface air coolers for a range of applications.

The Extended Surface Air Cooler

The typical air cooler type which is the topic of this paper is shown in principle

(Figure 1). It consists of a tube bundle fitted with plate fins; air is forced through the

fin side, refrigerant flows through the tubes. The tube pattern can be either square

or triangular; the tube diameters, fin thickness, fin design, tube centre distances,

circuiting, face velocity, refrigerant, refrigerant feed and material selections are

variables. This paper will not cover the issues of turbulators in tubes or wavy fins

nor will it describe any other fluid on the fin side than humid air at atmospheric

pressure. Finally, the paper will generally be limited to the presentation of results

to maximize clarity.

The air cooler performance modeling methodology used for the various comparisons

in this paper is described in earlier papers by the author [Reference 1], [Reference 2].

Impact of Coil Geometry on Overall Heat Transmission (U-Value)

The term, coil geometry, refers in this context to the tube pattern and the centre

distances between the tubes. Cooling coil layout refers to the way a particular cooler

of a particular geometry is laid out, i.e., rows high, rows deep, circuiting, finned

length, fin spacing, etc.

In Figure 2 is shown a range of hot dipped galvanized steel coils of different

geometries. By using the same material throughout the range, the variation in heat

transmission coefficients from 34.6 to 18.9 W/m2K (6.09 to 3.33 Btu/h ft2 F) is the



result of variations in coil geometry only. All coils shown use 0.3 mm (0.0118") thick

fin material with a 0.112 mm (0.00441") thick layer of zinc.

Impact of Materials of Construction and Fin Thickness

Cooling coil performances are influenced by materials of construction and fin

thickness. This is demonstrated in Figure 3 where the coil geometry has been kept

constant, but materials and fin thicknesses have been varied to isolate the impact of

these variables. The variation in heat transmission coefficients from 13.2 to 37.5 W/

m2K (2.32 to 6.60 Btu/h ft2 F) is a direct result of the changes to thermal conductivity

of materials and fin efficiencies.

Circuiting

Circuiting is how the refrigerant is piped through the cooling coil. A coil can be

piped for physical parallel flow, physical counter flow, thermodynamic counter flow,

a mixture of physical parallel flow and counter flow and cross flow between air and

refrigerant. In addition, the circuits can be made long or short as dictated by the

tube diameter, refrigerant temperature, heat flux and refrigerant properties. If the

application requires relatively short circuits and the coil in question is large, then it is

often necessary to provide the coil with multiple headers or manifolds because there

is a manufacturing limitation with respect to the number of circuits which can enter/

leave a header. This makes the valve station more complicated and costly and is

generally not in the interest of the contractor.

The way the circuits run in relation to the force of gravity is dictated by a number

of factors. These include oil drainage, the type of refrigerant feed, orientation of the

coil, heat flux, type of defrost, thermostatic expansion valve sensor location and

general refrigerant side thermodynamics. Circuiting errors are most forgiving in liquid



overfeed applications and potentially most disastrous in dry expansion feed cooling

coils.

The greater the turbulence on the refrigerant side, the greater the surface heat

transfer and the greater the pressure drop. The greater the refrigerant side pressure

drop for a volatile refrigerant, the greater the refrigerant temperature drop as the fluid

passes through the coil. The greater the refrigerant temperature drop, the lower the

logarithmic mean temperature difference (LMTD) all other things being equal. Lower

LMTD translates into reduced coil capacity. This scenario presents an optimization

problem between achieving good heat transfer on the tube side of the heat exchanger

without jeopardizing LMTD. A relatively high refrigerant side temperature drop has

the capacity to reduce LMTD to such an extent that any gain achieved by improved

heat transfer is more than eroded by a reduction in temperature difference hence

reducing overall coil capacity.

It is generally not possible to circuit any cooling coil in an optimal manner for the

entire operating range that the coil will encounter during its operating life. All choices

of circuiting therefore represent a compromise. If the manufacturer is provided with

the range of design operating conditions applicable to the coil in question then it

is possible to determine the most reasonable circuiting for the operating envelope

nominated. In many cases, however, the manufacturer is only provided with one

operating point namely the maximum coil performance required. Particularly in

dry expansion applications this can in practice lead to poor results at part load (low

temperature difference) and/or at reduced face velocity [Reference 2].

By far the most common circuiting method for industrial applications is shown in

Figure 4. This shows a liquid overfeed coil with horizontal headers; liquid inlet at the

bottom, wet return at the top; horizontal air flow and four tube passes. In effect, this

represents a cross flow arrangement between air and refrigerant.



If a coil such as the one shown in Figure 4 is for a low temperature application and

is both long between the end plates and comparatively tall to meet the capacity

requirement, then the number of tube passes needs to reduce. This is to ensure that

the refrigerant side temperature (pressure) drop does not jeopardize LMTD to such

an extent that overall capacity reduces. It is generally less expensive for the coil

manufacturer to then supply a coil with multiple horizontal liquid and wet return

headers as opposed to having one inlet and one outlet header that accommodate all

the circuits. A typical multiple header coil is shown in Figure 5. The consequence to

the contractor of this header arrangement is that each evaporator segment needs to

be treated like an individual evaporator in terms of valve station design. Essentially,

the capital cost of the air cooler has therefore been minimized at the expense of

increased valve station costs. Often the result is an overall project cost increase

because the additional cost of the valve station more exceeds the saving in coil costs.

An alternative to multiple horizontal headers to achieve the correct circuit length

is vertical headers (Figure 6). In the great majority of cases, coil depth is less

than the coil height so the need for multiple headers can be eliminated by using

vertical headers. However, to ensure uniform distribution of refrigerant in a liquid

overfeed situation, each circuit inlet must have an orifice. If there are no orifices, the

refrigerant will take the path of least resistance and the upper circuits of the coil will

not perform. This may be illustrated with a simple example. If the refrigerant side

pressure drop for the coil illustrated in Figure 6 is 1K (1.8F) then this translates into a

pressure drop of 3839 Pa (0.56 psi) or 0.57 m (1.86 feet) of ammonia liquid column.

If the vertical liquid inlet header is 1.2 m (3.94 feet) tall, then gravity will force

the liquid refrigerant through the bottom half of the evaporator circuits leaving the

top circuits dry. The result will be more or less a halving of the anticipated cooling

capacity. The design driving force across the circuit orifices should be greater than

the height of the vertical inlet header. It follows therefore that the top orifices in the

header need to be larger than the bottom orifices to account for gravity. The mass

flows through circular orifices may be calculated as shown:



m = * A * SQR(2 * p * ) (1)

For the orifice contraction coefficient , a value of 0.7 may be used. For the top

orifices in the header a design pressure drop (p) of around 0.45 to 0.6 m (1.5 to

2 feet) of liquid column is practical and will generally not yield orifices which are

too small and hence susceptible to blockage. The design pressure drops measured

in liquid column for the bottom orifices are the same as for the top plus the height

of the header. Using larger design pressure drops will reduce the size of the orifices

to such an extent that they become susceptible to impurities. For the purposes of

streamlining manufacturing operations, some coil manufacturers only provide a

choice between either 3 or 4 mm (0.12" or 0.16") orifices. In most cases these

are far too large (unless the liquid overfeed ratio is very high) and do not provide

the appropriate graduation in size between top and bottom of the inlet header. More

appropriate orifice sizes are around 1.7 to 2.8 mm (0.067" to 0.11"). Even with

these orifice sizes, it is often found in practice that the liquid overfeed ratio needs to

be elevated to around 6 to 1 to ensure adequate refrigerant distribution to all circuits

during all operating conditions.

The orifices in a vertical inlet header not only restrict the flow of liquidthey also

restrict the flow of hot gas. This needs to be carefully considered when an evaporator

is to be defrosted with hot gas. It is often found in practice that coils with vertical

headers require longer time to defrost with hot gas than coils with horizontal headers

where the hot gas is entered in the top wet return header hence pushing the cold

liquid out of the bottom header relatively rapidly at commencement of defrost.

The circuit orifices in the coil shown in Figure 6 may be eliminated by turning the

coil 90 such that the air flows vertically up through the coil. In liquid overfeed

applications this then enables liquid entry through the bottom header and wet return

from the top thus creating physical parallel flow or thermodynamic counter flow.

Simply turning the coil 90 is naturally not always physically possible, but if it is,

then a circuiting problem can be converted to a circuiting advantage. This circuiting



method has been used successfully in a large number of automatic tunnel freezers in

Australia and New Zealand for about 50 years.

There are cases where circuit orifices may be required also in coils with horizontal

headers and horizontal air flow. In a situation where the coil is exposed to high ETD

for example a fresh air inlet cooling coil in a food processing plantthe circuits at

the air inlet to the coil may boil dry. This is due to the fact that high heat flux also

creates high refrigerant side pressure drop. At the air leaving side of the coil where

the heat flux is less, the refrigerant side pressure drop is relatively low. Without

circuit orifices the circuits at the air leaving side are then in equilibrium supplied

with excess refrigerant at the expense of the circuits at air entry, which are left

starved of refrigerant. The result is a significant reduction in overall coil performance

(Figure 7).

The impact of circuiting on coil performance is visualized in Figure 8. The number

of rows high, deep, fin spacing and finned length have all been kept constant.

The number of times that the refrigerant passes through the tube bundle has been

variedthis is referred to as the number of tube passes. The vertical axis shows

heat transmission coefficients and gross coil capacity; the horizontal axis shows

evaporating temperature and entering temperature difference (ETD) between air

and refrigerant. The ETD is calculated by subtracting the evaporating temperature

from the air on temperature of 18C (0.4F) shown. The graphs show the coil

performance as a function of the number of tube passes. It is clear that there is no

such thing as optimal circuiting for a wide range of operating envelopes. At 25C

(13F) evaporating temperature (7K or 12.6F ETD), 28 passes deliver maximum

coil performance. At 28C (18.4F) evaporating temperature, 14 passes yield

maximum performance. In a dry expansion application such as the one shown,

circuiting for a comparatively low refrigerant pressure drop at full load may deliver

optimal coil performance at that point. However, practical experience has also shown

that a coil circuited this way is likely to become difficult to control at low ETDs and



the result will be increasing liquid entrainment in the suction vapor leaving the coil.

This is also known as liquid flood-back.

It is possible to circuit both dry expansion coils and liquid overfeed coils such that

the reduction in LMTD by a relatively high refrigerant side temperature (pressure)

drop is minimized (Figure 9). Physical parallel flow translates into thermodynamic

counter flow. Using this circuiting method will, in dry expansion applications, give

rise to problems obtaining a suitable superheat signal for the expansion valve. These

problems may be overcome by circuiting the last tube pass through the air inlet side

(Figure 10). This circuiting method is termed reversed suction return. The difference

in coil performance between reversed suction return and physical counter flow for a

typical dry expansion coil is illustrated in Figure 11.

Fouling

Fouling is a layer of material with relatively poor thermal conductivity that has

settled on the heat transfer surface of a heat exchanger. This layer has the capacity to

create a thermal resistance and hence inhibit heat transfer. The result is a reduction

of the overall heat transmission coefficient. In industrial extended surface air coolers,

fouling can occur both on the refrigerant side and on the air side. On the refrigerant

side of air coolers in ammonia plants, the most common form of fouling is oil; on the

air side it is frost. There are other forms of fouling, but these are outside the scope of

this paper.

Not all air cooler geometries are affected the same way by a given fouling resistance.

The percentage coil performance reduction as a function of the internal and external

fouling resistances is shown in Figure 12 for two different coil geometries. It is

evident that the performance of geometry (a) deteriorates more rapidly than that of

geometry (b) as fouling increases. Due to the fact that fouling is almost unavoidable

in industrial ammonia refrigeration plants, this observation can be of crucial



importance to the contractor. A thin layer of oil of 0.05 mm (0.002") thickness on

the refrigerant side of an ammonia air cooler is not uncommon [Reference 3]. This

results in a fouling resistance of 0.0004 m2K/W (0.0023 ft2hrF/Btu). The impact of

this fouling resistance is a reduction in heat transmission coefficient of 13.5% in the

case of geometry (a). A frost layer of ~1 mm (0.04") throughout the coil represents

a fouling resistance of around 0.008 m2K/W (0.046 ft2hrF/Btu) [Reference 4]. The

impact of this is a capacity reduction of 27.8% and 21.4% for geometries (a) and (b)

respectively.

Refrigerant and Refrigerant Feed

A very common refrigerant feed method in industrial ammonia applications is liquid

overfeed. Figure 13 shows the performance of a typical ammonia liquid overfeed

coil as a function of liquid overfeed rate. Although the liquid overfeed rate has little

impact on the coil performance when viewed in isolation, excess liquid overfeed rates

can have a significant impact on the pressure drop in the wet return line downstream

of the evaporator. An increase in wet return line pressure drop increases the ammonia

pressure that the coil is exposed to at the suction connection and the result is a

reduction in coil capacity.

Common alternatives to liquid overfeed are gravity flooded and dry expansion feeds.

The principle of gravity flooded feed is shown in Figure 14. In equilibrium, the sum

of pressure gains and pressure drops in the refrigerant circuit must be zero. In Table 1

are shown the key data for a flooded evaporator with NH3 and CO2 refrigerants.

The air flow is vertically up (Figure 14).



Table 1. Key data for NH3 and CO2 flooded evaporators

Refrigerant NH3 CO2tE, C (F) 41.5 (42.7) 41.5 (42.7)tS, C (F) 42.4 (44.3) 41.8 (44.3)tE +PR, C (F) 39.7 (39.5) 41.46 (42.63)

Air on/off, C (F)35.0/37.8

(31.0/36.0)

35.0/38.2

(31.0/36.8)Relative humidity on/off, % 80/93 80/93Liquid overfeed rate 5.2 to 1 5.6 to 1Total cooling capacity, kW (TR) 124.8 (35.6) 143.2 (40.8)Overall heat transmission coefficient,

service, W/m2K (Btu/h ft2 F)31.6 (5.56) 32.1 (5.65)

d1, m (inches) 0.10226 (4.03) 0.0779 (3.07)d2, m (inches) 0.0779 (3.07) 0.0627 (2.47)d3, m (inches) 0.1282 (5.05) 0.0779 (3.07)

For both refrigerants the evaporator coils and surge drum heights HL are identical.

The interconnecting refrigerant pipe line diameters have been selected in accordance

with good design practice. The impact of the difference between pressure/

temperature gradients for the two refrigerants is evident in Table 1. The practical

consequence of the greater temperature change of NH3 is reduced LMTD across the

NH3 evaporator and hence lower performance.

Dry expansion refrigerant feed is very common in commercial and small to medium

size industrial applications. Some characteristics of dry expansion feed are less

favorable than those of liquid overfeed. This becomes evident at evaporator part-load

i.e., at reduced face velocity, reduced ETD and when frost accumulates on the coil.

A comparison between dry expansion and liquid overfeed evaporators is provided

(Figure 15). The leveling or declining capacity at high ETDs for dry expansion feed

coils is evidence that the refrigerant side pressure (temperature) drop becomes so



high that it starts to jeopardize LMTD and hence overall capacity. The liquid overfeed

coil capacity continues to climb when exposed to the same operating condition.

Air Pressure Drop

The air pressure drop through a finned air cooler is a function of air cooler geometry

and face velocity. Figure 16 provides some typical air pressure drop comparisons for

a number of geometries at identical face velocities. The geometries, coil layouts and

operating conditions are identical to those shown in Figure 2. The air pressure drop

increases with increasing face velocity. The rate of increase may be approximated as

the square of the face velocity increase.

Energy Optimization

The gross air cooler capacity increases with increasing face velocity. Gross and net

air cooler capacities are in this context defined as cooler capacities without and with

correction for fan power respectively. To force a certain quantity of air through a

cooling coil with a certain air pressure drop requires a certain amount of fan power.

In the majority of refrigeration, chilling and freezing applications, the fan power

is converted to heat, which enters the refrigerated space and in turn needs to be

removed again by the air cooler. In most cases it is therefore only the net cooler

capacity (gross capacity minus fan heat) that is available for refrigerating the space

in question.

Plotting net cooler capacity as a function of face velocity using the ETD as the

parameter reveals that the most energy efficient face velocity is not constant, but

varies depending on ETD, (Figure 17). For relatively high ETDs, it is more energy

efficient to operate with relatively high face velocities and vice versa. These data sets

may be generated for any combination of air cooler and fans. The data set is useful



for producing a look-up table used in the electronic control system for an evaporator

where the fans are controlled by means of a variable frequency drive (VFD). The line

shown through the optimum operating points is of significance for the establishment

of this look-up table. Measurement of refrigerant and air entering temperature defines

ETD and this in turn enables the control system to automatically set the most energy

efficient fan speed.

Humidity

The minimization of chilling weight loss is of vital importance during the cooling of

most perishable goods where the cooling process uses air as the secondary refrigerant

surrounding the goods (Figure 18). In addition, the maintenance of high relative air

humidity in spaces where perishable goods such as leafy vegetables, flowers and

fruits are stored is equally important to maximize shelf life. The importance of cooler

geometry and cooler layout is often overlooked in this context.

A typical average refrigeration load in a batch type beef carcass chiller with a holding

capacity of 24,000 kg (52,911 lbs) and a design chilling cycle time of 18 hours may

calculate as shown in Table 2 with an assumed chilling weight loss of 1%.

Table 2. Average heat load in a typical beef carcass chiller

Sensible Heat

[kW]/[Btu/h]

Latent Heat

[kW]/[Btu/h]Conduction 5.4 (18,442) 0Infiltration 0 0Product 35.6 (121,580) 9.3 (31,761)Fans 6.6 (22,540) 0Total 47.6 (162,563) 9.3 (31,761)



In Table 3 is shown a comparison of estimated chilling weight loss between two air

cooler types which have identical overall surface areas. Both coolers are shown in

end view (Figure 19) and both service the carcass chiller which is the subject of the

load calculation (Table 2). One cooler is squat; the other has a relatively large face

area and is shallow in the direction of air flow.

Table 3. Estimated theoretical chilling weight loss for two different air cooler

geometries and identical total heat exchanger surface areas.

Number of coils per chiller 2 3Estimated chilling weight loss, kg (lbs) 480 (1058) 270 (595)Chilling weight loss, % 2.0 1.1Value of loss, $/p.a. (1999 values) 240,000 135,000Approximate energy costs, $/p.a. (1999 values) 10,300 12,200

The higher energy cost for the low weight loss design is due to the greater circulated

air quantity. The difference in equipment costs is around $30,000 in 1999 values with

the low weight loss solution being the most expensive. Provided the operator gets

paid for the achieved weight gain, this differential investment may be returned in

a few months.

Discussion

Correct technical selection of an extended surface air cooler for an industrial

refrigeration application is a complex matter for the refrigeration contractor. It

requires careful evaluation of a large number of issues ranging from cooling capacity,

total/sensible heat ratios, relative humidity, water droplet carry-over, defrost

frequency, sound levels and air throw to oil drainage, defrost methods, operating

weight, material suitability, hygiene, corrosion resistance, delivery time and capital

costs. If the technical selection process being undertaken by the refrigeration



contractor is being complicated further by equipment manufacturers attempting

to influence purchasing decisions by way of technical marketing material displaying

varying degrees of relevance and indeed integrity, then it is not a surprise that the

whole process can become confusing, at times overwhelming and prone to an easy

way out approach.

This paper has shown variations in heat transmission coefficients of around a

factor of two as a direct result of changes to coil geometry only (i.e., variations in

tube diameters, tube centers and tube patterns). In addition, a variation in heat

transmission coefficients by a factor of almost three was shown as a direct result of

changes in materials of construction and fin thickness for constant coil geometry.

These two statements highlight the inherent danger of rules of thumb selecting air

coolers on the basis of simple surface area allocation.

An area of great significance in practice is refrigerant circuiting. The refrigerant

circuiting of an air cooler is dictated by heat transfer fundamentals. Yet there are

thousands of examples of air cooling coils performing poorly due to incorrect

circuiting. This is particularly the case in dry expansion applications where optimal

circuiting is only possible inside a relatively limited envelope of operating conditions.

The key question for the refrigeration contractor is how is it possible to evaluate

whether or not the potential supplier of a particular cooling coil is proposing the

correct circuiting? The first step is to request information. There are many examples

of the circuiting details not being disclosed by manufacturers at design stage. Once

the equipment is on site or installed it is usually too late.

Although the issue of fouling has been discussed, the issue of frost accumulation

has not. Frost accumulation on the external surfaces of the cooling coil will inhibit

air flow. A reduction in air flow will reduce cooling capacity and this has not been

accounted for in Figure 12, which focuses on the thermal resistance only. This

raises an important issue to consider for the refrigeration contractor. The issue is the



allowance of sufficient frost volume on the external surfaces of the cooling coil

particularly at coil inletto enable continued operation without inhibiting air flow.

It is commonly known that increasing the face velocity through a cooling coil

improves heat transfer and coil capacity. What may not be readily identifiable is that

a reduction in coil face velocity from 3.5 m/s (689 fpm) to 2.5 m/s (492 fpm) may

only reduce gross coil capacity by 18% yet reduce fan power by 60% and in many

cases leave net coil capacity (gross capacity minus fan power) almost unchanged.

This represents two potential traps for the refrigeration contractor. Firstly, there is

the likelihood of making a misdirected purchasing decision on the basis of gross coil

capacity. Secondly, excess fan heat may be detrimental to some perishable products

because the consequence will be a reduction in relative room humidity.

The message for refrigeration contractors generally and particularly with respect to

the selection of extended surface air coolers is to devote the topic the engineering

attention it deserves. A well known slogan from the oil industry is oil aint oils. This

slogan is equally applicable to the refrigeration industry in the slightly modified form

coils aint coils.



Nomenclature

LMTD [K] Logarithmic Mean Temperature Difference

ETD [K] Entering temperature difference

m [kg/s] mass flow

[-] Orifice contraction coefficient

A [m2] Orifice area

p [Pa] Pressure drop

[kg/m3] Density

p [-] Number of tube passes

tE [C] Evaporating temperature

tS [C] Saturated suction temperature

pR [K] Refrigerant pressure (temperature) drop

d [m] Diameter

H [m] Height

FeZn [-] Mild steel, hot dipped galvanized

SS [-] Stainless steel

Al [-] Aluminum

FIN [m] Fin thickness

kO [W/m2K] Heat transmission coefficient

u [Btu/h ft2F] Heat transmission coefficient

QE [kW] Total evaporator capacity



QE,SENS [kW] Gross sensible evaporator capacity

Q0 [kW] Net sensible evaporator capacity

QFAN [kW] Fan heat rejection

i [W/m2K] Inside film coefficient

WFACE [m/s] Face velocity

nFAN [-] Rotational fan speed



References

[1] Jensen, S.S. 1989. Extended Surface Steel Air Coolers for Industrial

Refrigeration, AIRAH Journal August 1989: 4151.

[2] Jensen S.S. 2006. Dry Expansion Feed in Dual Stage Ammonia Plants: Operating

Experiences in a Large Refrigerated Distribution Centre. Proc. 2006 IIAR

Ammonia Refrigeration Conference and Exhibition, Reno, Nevada.

[3] Koster, G.J. 1985. Grenco, sHertogenbosch, Netherlands. Energy Savings

in Ammonia Refrigeration Plant by Using Oil Scrubbers. Proc. Institute of

Refrigeration 198485. 4.1

[4] Mlhammar, . 1986. Frost Growth in Evaporators. Scandinavian Refrigeration

6/86: 314323.

[5] Jensen, S.S. 2004. Design and Selection of Industrial Finned Air Coolers for

Natural RefrigerantsA Comparison between NH3 and CO2, Proc. The Natural

Refrigerants Transition Board Design Seminar, Sydney, Australia.

[6] Jensen S.S. 2000. Carcass Chilling After Slaughter, Innovative Air Cooling and

Air Distribution Techniques. Proc. Australian Meat Council CEO Conference.



Figure 1. Typical extended surface stainless steel/aluminum air cooler



Figure 2. Impact of cooler geometry on heat transmission coefficient (u-value)

Figure 3. Impact of materials of construction and fin thickness on heat transmission coefficient



Figure 4. Common circuiting method with horizontal headers and bottom refrigerant inlet for industrial liquid overfeed applications

Figure 5. Multiple header arrangement



Figure 6. Circuit length reduction by employing vertical headers, coil shown in plan view

Figure 7. Equilibrium refrigerant flows through the circuits of a cross flow finned air cooler, high heat flux, ammonia refrigerant, liquid overfeed



Figure 8. Impact of refrigerant circuiting on coil performance

Figure 9. Physical parallel flow circuiting for thermodynamic counter flow



Figure 10. Coil circuiting with parallel flow and reversed suction return

Figure 11. Performance comparison for large low temperature finned air cooler with physical parallel flow and physical counter flow air/refrigerant flow patterns



Figure 12. Relative impact of fouling on coil performance; fi represents inside fouling, fo outside fouling



Figure 13. Coil performance as a function of liquid overfeed rate



Figure 14. Evaporator with gravity flooded refrigerant feed

Figure 15. Capacity of an NH3 evaporator as a function of entering temperature difference and face velocity. Dry expansion (DX) and liquid overfeed (LR), air on constant at 35C (31F)



Figure 16. Air pressure drops for different coil geometries; coil layouts and operating conditions as per appendix 1

Figure 17. Energy optimization at varying face velocities and entering temperature differences



Figure 18. Commonly used batch carcass chiller layout in side elevation.

Figure 19. End view of the two different evaporator geometries and layouts used for the weight loss estimate provided in Table 3

Notes:

