Related Commercial Resources

CHAPTER 41

LIQUID COOLERS
Types of Liquid Coolers................................................................................................................. 41.1
Heat Transfer ................................................................................................................................ 41.3
Pressure Drop ............................................................................................................................... 41.4
Vessel Design ................................................................................................................................ 41.4
Application Considerations .......................................................................................................... 41.5

A LIQUID cooler (hereafter called a cooler) is a heat exchanger

in which refrigerant is evaporated, thereby cooling a fluid
(usually water or brine) circulating through the cooler. This chapter
superheat, the remaining tubes must produce a higher superheat to
evaporate the liquid escaping into the suction line. This unbalance
causes poor overall heat transfer. Uniform distribution is usually
addresses the performance, design, and application of coolers. achieved by adding a distributor, which creates sufficient turbu-
lence to promote a homogeneous mixture so that each tube gets the
TYPES OF LIQUID COOLERS same mixture of liquid and vapor.
The number of refrigerant passes is another important item in
Various types of liquid coolers and their characteristics are listed
direct-expansion cooler performance. A single-pass cooler must
in Table 1 and described in the following sections.
evaporate all the refrigerant before it reaches the end of the first pass;
this requires long tubes. A multiple-pass cooler is significantly
Licensed for single user. © 2008 ASHRAE, Inc.

Direct-Expansion
shorter than a single-pass cooler, but must be properly designed to
Refrigerant evaporates inside the tubes of a direct-expansion
ensure proper refrigerant distribution after the first pass. Internally
cooler. These coolers are usually used with positive-displacement
and externally enhanced tubes can also be used to reduce cooler size.
compressors, such as reciprocating, rotary, or rotary screw compres-
A tube-in-tube cooler is similar to a shell-and-tube design,
sors, to cool various fluids, such as water, water/glycol mixtures,
except that it consists of one or more pairs of coaxial tubes. The
and brine. Common configurations include shell-and-tube, tube-in-
fluid usually flows inside the inner tube while the refrigerant flows
tube, and brazed-plate.
in the annular space between the tubes. In this way, the fluid side can
Figure 1 shows a typical shell-and-tube cooler. A series of baf- be mechanically cleaned if access to the header is provided.
fles channels the fluid throughout the shell side. The baffles create
cross flow through the tube bundle and increase the velocity of the
Fig. 1 Direct-Expansion Shell-and-Tube Cooler
fluid, thereby increasing its heat transfer coefficient. The velocity of
the fluid flowing perpendicular to the tubes should be at least 2 fps
to clean the tubes and less than the velocity limit of the tube and baf-
fle materials, to prevent erosion.
Refrigerant distribution is critical in direct-expansion coolers.
If some tubes are fed more refrigerant than others, refrigerant may
not fully evaporate in the overfed tubes, and liquid refrigerant may
escape into the suction line. Because most direct-expansion cool-
ers are controlled by an expansion valve that regulates suction

The preparation of this chapter is assigned to TC 8.5, Liquid-to-Refrigerant

Heat Exchangers. Fig. 1 Direct-Expansion Shell-and-Tube Cooler

Table 1 Types of Coolers

Type of Cooler Subtype Usual Refrigerant Feed Device Usual Capacity Range, tons Commonly Used Refrigerants
Direct-expansion Shell-and-tube Thermal expansion valve 2 to 1000 12, 22, 134a, 404A, 407C, 410A, 500, 502,
Electronic modulation valve 2 to 1000 507A, 717
Tube-in-tube Thermal expansion valve 5 to 25 12, 22, 134a, 717
Brazed-plate Thermal expansion valve 0.6 to 200 12, 22, 134a, 404A 407C, 410A, 500, 502, 507A,
508B, 717, 744
Semiwelded plate Thermal expansion valve 50 to 1990 12, 22, 134a, 500, 502, 507A, 717, 744
Flooded Shell-and-tube Low-pressure float 25 to 2000 11, 12, 22, 113, 114
High-pressure float 25 to 6000 123, 134a, 500, 502, 507A, 717
Fixed orifice(s) 25 to 6000
Weir 25 to 6000
Spray shell-and-tube Low-pressure float 50 to 10,000 11, 12, 13B1, 22
High-pressure float 50 to 10,000 113, 114, 123, 134a
Brazed-plate Low-pressure float 0.6 to 200 12, 22, 134a, 500, 502, 507A, 717, 744
Semiwelded plate Low-pressure float 50 to 1990 12, 22, 134a, 500, 502, 507A, 717, 744
Baudelot Flooded Low-pressure float 10 to 100 22, 717
Direct-expansion Thermal expansion valve 5 to 25 12, 22, 134a, 717
Shell-and-coil — Thermal expansion valve 2 to 10 12, 22, 134a, 717

41.1
Copyright © 2008, ASHRAE
 41.2 2008 ASHRAE Handbook—HVAC Systems and Equipment

Brazed- or semiwelded-plate coolers are constructed of plates sprayed to cover the tubes rather than flooding them. A mechanical
brazed or laser-welded together to make an assembly of separate pump circulates liquid from the bottom of the cooler to the spray
channels. Semiwelded designs have the refrigerant side welded and heads.
the fluid side gasketed and allow contact of the refrigerant with the Flooded shell-and-tube coolers are generally unsuitable for other
fluid-side gaskets. These designs can be disassembled for inspec- than horizontal orientation.
tion and mechanical cleaning of the fluid side. Brazed types do not In a flooded plate cooler (Figure 3), refrigerant vaporizes in ver-
have gaskets, cannot be disassembled, and are cleaned chemically. tical channels between corrugated plates with the liquid inlet at the
Internal leaks in brazed plates typically cannot be repaired. This bottom and the vapor outlet at the top (i.e., vertical upflow). The
type of evaporator is designed to work in a vertical orientation. Uni- warm fluid flow may be either counter or parallel to the refrigerant
form distribution in direct-expansion operation is typically achieved flow. Both thermosiphon (gravity feed) and pumped overfeed oper-
by using a special plate design or distributor insert; flooded and ation are used. Surge drums are required for pumped overfeed oper-
pumped overfeed operations do not require distribution devices. ation but usually not for thermosiphon operation because the
Plate coolers are very compact and require minimal space. corrugated plates demist flow under most conditions.
Most tubular direct-expansion coolers are designed for horizon-
tal mounting. If they are mounted vertically, performance may vary Baudelot
considerably from that predicted because two-phase flow heat trans- Baudelot coolers (Figure 4) are used to cool a fluid to near its
fer is a direction-sensitive phenomenon and dryout begins earlier in freezing point in industrial, food, and dairy applications. In this
vertical upflow. cooler, fluid circulates over the outside of vertical plates, which are
easy to clean. The inside surface of the plates is cooled by evaporat-
Flooded ing the refrigerant. The fluid to be cooled is distributed uniformly
In a flooded shell-and-tube cooler, refrigerant vaporizes on the along the top of the heat exchanger and then flows by gravity to a
outside of the tubes, which are submerged in liquid refrigerant in a collection pan below. The cooler may be enclosed by insulated walls
closed shell. Fluid flows through the tubes as shown in Figure 2. to avoid unnecessary loss of refrigeration.
Flooded coolers are usually used with rotary screw or centrifugal R-717 (ammonia) is commonly used with flooded Baudelot
Licensed for single user. © 2008 ASHRAE, Inc.

compressors to cool water, water/glycol mixtures, or brine. coolers using conventional gravity feed with a surge drum. A low-
A refrigerant liquid/vapor mixture usually feeds into the bottom pressure float valve maintains a suitable refrigerant liquid level in
of the shell through a distributor that distributes the mixture equally the surge drum. Baudelot coolers using other common refrigerants
under the tubes. The relatively warm fluid in the tubes heats the are generally direct-expansion, with thermostatic expansion valves.
refrigerant liquid surrounding the tubes, causing it to boil. As bub-
bles rise through the space between tubes, the liquid surrounding the Fig. 3 Flooded Plate Cooler
tubes becomes increasingly bubbly (or foamy, if much oil is present).
The refrigerant vapor must be separated from the mist generated
by the boiling refrigerant. The simplest separation method is pro-
vided by a dropout area between the top row of tubes and the suction
connections. If this dropout area is insufficient, a coalescing filter
may be required between the tubes and connections. Perry and
Green (2007) give additional information on mist elimination.
Another approach is to add another vessel, or “surge drum,” above
the suction connections. The diameter of this vessel is selected so
that the velocity of the liquid droplets slows to the point where they
fall back to the bottom of the surge drum. This liquid is then drained
back into the flooded cooler.
The size of tubes, number of tubes, and number of passes should
be determined to maintain fluid velocity typically between 3 and
10 fps for copper alloy tubing. Velocities beyond these limits may
be used if the fluid is free of suspended abrasives and fouling sub-
stances (Ayub and Jones 1987; Sturley 1975) or if the tubing is man-
ufactured from special alloys, such as titanium and stainless steel, Fig. 3 Flooded Plate Cooler
that have better resistance to erosion. In some cases, the minimum
velocity may be determined by a lower Reynolds number limit. Fig. 4 Baudelot Cooler
One variation of this cooler is the spray shell-and-tube cooler.
In large-diameter coolers where the refrigerant’s heat transfer coef-
ficient is adversely affected by the refrigerant pressure, liquid can be

Fig. 2 Flooded Shell-and-Tube Cooler

Fig. 2 Flooded Shell-and-Tube Cooler Fig. 4 Baudelot Cooler

 Liquid Coolers 41.3

Fig. 5 Shell-and-Coil Cooler t = thickness of heat transfer surface (tube wall thickness), ft
rfi = fouling factor of fluid side based on inside surface area,
ft2 ·h· °F/Btu
rfo = fouling factor of fluid side based on outside surface area,
ft2 ·h· °F/Btu
Note: If fluid is on inside, multiply rf i by Ao /Ai to find rfo.
If fluid is on outside, multiply rfo by Ai /Ao to find rf i .
These equations can be applied to incremental sections of the
heat exchanger to include local effects on the value of U, and then
the increments summed to obtain a more accurate design.

Heat Transfer Coefficients

The refrigerant-side coefficient usually increases with (1) an
increase in cooler load, (2) a decrease in suction superheat, (3) a
decrease in oil concentration, or (4) an increase in saturated suction
Fig. 5 Shell-and-Coil Cooler temperature. The amount of increase or decrease depends on the type
of cooler. Schlager et al. (1989) and Zürcher et al. (1998) discuss the
effects of oil in direct-expansion coolers. Flooded coolers have a rela-
tively small change in heat transfer coefficient as a result of a change
Shell-and-Coil in load, whereas a direct-expansion cooler shows a significant increase
A shell-and-coil cooler is a tank containing the fluid to be cooled with an increase in load. A Wilson plot of test data (Briggs and Young
with a simple coiled tube used to cool the fluid. This type of cooler 1969; McAdams 1954) can show actual values for the refrigerant-side
has the advantage of cold fluid storage to offset peak loads. In some coefficient of a given cooler design. Collier and Thome (1994), Thome
Licensed for single user. © 2008 ASHRAE, Inc.

models, the tank can be opened for cleaning. Most applications are (1990, 2003), and Webb (1994) provide additional information on pre-
at low capacities (e.g., for bakeries, for photographic laboratories, dicting refrigerant-side heat transfer coefficients.
and to cool drinking water). The fluid-side coefficient is determined by cooler geometry,
The coiled tube containing the refrigerant can be either inside the fluid flow rate, and fluid properties (viscosity, specific heat, thermal
tank (Figure 5) or attached to the outside of the tank in a way that conductivity, and density) (Palen and Taborek 1969; Wolverine
allows heat transfer. Tube 1984). For a given fluid, the fluid-side coefficient increases
with fluid flow rate because of increased turbulence and with fluid
HEAT TRANSFER temperature because of improvement of fluid properties as temper-
ature increases.
Heat transfer for liquid coolers can be expressed by the following The heat transfer coefficient in direct-expansion and flooded
steady-state heat transfer equation: coolers increases significantly with fluid flow. The effect of flow is
smaller for Baudelot and shell-and-coil coolers.
q = UA∆tm (1)
An enhanced heat transfer surface can increase the heat transfer
coefficient of coolers in the following ways:
where
q = total heat transfer rate, Btu/h • It increases heat transfer area, thereby increasing the overall heat
∆tm = mean temperature difference, °F transfer rate and reducing the thermal resistance of fouling, even
A = heat transfer surface area associated with U, ft2 if the refrigerant-side heat transfer coefficient is unchanged.
U = overall heat transfer coefficient, Btu/h·ft2 ·°F • Where flow of fluid or refrigerant is low, it increases turbulence at
The area A can be calculated if the geometry of the cooler is known. the surface and mixes fluid at the surface with fluid away from the
Chapter 3 of the 2005 ASHRAE Handbook—Fundamentals de- surface. For stratified internal flows of refrigerants, it may convert
scribes the calculation of the mean temperature difference. the flow to complete wetting of the tube perimeter.
This chapter discusses the components of U, but not in depth. U • In flooded coolers, an enhanced refrigerant-side surface may pro-
may be calculated by one of the following equations. vide more and better nucleation points to promote boiling of
refrigerant.
Based on inside surface area
Pais and Webb (1991) and Thome (1990) describe many enhanced
1 surfaces used in flooded coolers. The enhanced surface geometries
U = -----------------------------------------------------------------------------------------------------
- (2) provide substantially higher boiling coefficients than integral finned
1 ⁄ h i + [ A i ⁄ ( A o h o ) ] + ( t ⁄ k ) ( A i ⁄ A m ) + r fi
tubes. Nucleate pool boiling data are provided by Webb and Pais
Based on outside surface area (1991). The boiling process in the tube bundle of a flooded cooler
may be enhanced by forced-convection effects. This is basically an
additive effect, in which the local boiling coefficient is the sum of the
1
U = -------------------------------------------------------------------------------------------------------
- (3) nucleate boiling coefficient and the forced-convection effect. Webb
[ A o ⁄ ( A i h i ) ] + 1 ⁄ h o + ( t ⁄ k ) ( A o ⁄ A m ) + r fo et al. (1989) describe an empirical method to predict flooded cooler
performance, recommending row-by-row calculations.
where Based on this model, Webb and Apparao (1990) present the
hi = inside heat transfer coefficient based on inside surface area, results of calculations using a computer program. The results show
Btu/h·ft2 · °F
some performance differences of various internal and external sur-
ho = outside heat transfer coefficient based on outside surface area,
Btu/h·ft2 · °F face geometries. As an example, Figure 6 shows the contribution of
Ao = outside heat transfer surface area, ft2 nucleate pool boiling to the overall refrigerant heat transfer coeffi-
Ai = inside heat transfer surface area, ft2 cient for an integral finned tube and an enhanced tube as a function
Am = mean heat transfer area of metal wall, ft2 of the tube row. Forced convection predominates with the integral
k = thermal conductivity of heat transfer material, Btu/h·ft· °F finned tube.
 41.4 2008 ASHRAE Handbook—HVAC Systems and Equipment

Fig. 6 Nucleate Boiling Contribution to Total Refrigerant Haider et al. (1991) found little potential for fouling in such systems
Heat Transfer in a field survey. Experimental work with various tube geometries
by Haider et al. (1992) confirmed that negligible fouling occurs in
closed-loop evaporator tubes at 3 to 5 fps and 7 fps water velocities.
ARI Standard 480 discusses fouling calculations.
The refrigerant side of the cooler is not subject to fouling, and a
fouling factor need not be included for that side.

Wall Resistance
Typically, the t/k term in Equations (2) and (3) is negligible.
However, with low thermal conductivity material or thick-walled
tubing, it may become significant. Refer to Chapter 3 of the 2005
ASHRAE Handbook—Fundamentals and to Chapter 38 of this vol-
ume for further details.

PRESSURE DROP
Fluid Side
Pressure drop is usually minimal in Baudelot and shell-and-coil
coolers but must be considered in direct-expansion and flooded
coolers. Both direct-expansion and flooded coolers rely on turbulent
fluid flow to improve heat transfer. This turbulence is obtained at the
Fig. 6 Nucleate Boiling Contribution to expense of pressure drop.
Licensed for single user. © 2008 ASHRAE, Inc.

Total Refrigerant Heat Transfer For air-conditioning, pressure drop is commonly limited to 10 psi
to keep pump size and energy cost reasonable. For flooded coolers,
ASHRAE research projects RP-725 (Chyu 1995) and RP-668 see Chapter 38 for a discussion of pressure drop for flow in tubes.
(Moeykens et al. 1995) studied spray evaporation performance in Pressure drop for fluid flow in shell-and-tube direct-expansion cool-
ammonia and halocarbon refrigerant systems. Moeykens et al. ers depends greatly on tube and baffle geometry. The following
investigated shell-side heat transfer performance for commercially equation projects the change in pressure drop caused by a change in
available enhanced surface tubes in a spray evaporation environ- flow:
ment. The study determined that spray evaporation heat transfer can
yield shell-side heat transfer coefficients equal to or greater than New pressure drop =
New rate 1.8 (4)
those found with enhanced nucleate boiling surface tubes in the Original pressure drop ------------------------------
flooded boiling environment. Moeykens and Pate (1996) describe Original rate
an enhancement to shell-side heat transfer performance generated
with small concentrations of oil (<2.5%) in spray evaporation. They Refrigerant Side
attribute the improvement to foaming, which enhances heat transfer The refrigerant-side pressure drop must be considered for direct-
performance in the upper rows of large tube bundles operating in expansion, shell-and-coil, brazed-plate, and (sometimes) Baudelot
flooded boiling mode. coolers. When there is a pressure drop on the refrigerant side, the
Gupte and Webb (1995a, 1995b) investigated convective vapor- refrigerant inlet and outlet pressures and corresponding saturated
ization in triangular enhanced tube bundles. They proposed a mod- temperature are different. This difference changes the mean temper-
ified Chen superposition model that predicts the overall convective/ ature difference, which affects the total heat transfer rate. If pressure
vaporization coefficient as the sum of the single tube nucleate pool drop is high, expansion valve operation may be affected because of
boiling coefficient and a weighted contribution of a single-phase reduced pressure drop across the valve. This pressure drop varies,
convective coefficient. depending on the refrigerant used, operating temperature, and type
Casciaro and Thome (2001a, 2001b) provide a recent compre- of tubing. For flooded evaporators, Casciaro and Thome (2001b)
hensive review of thermal design methods for flooded evaporators. summarize prediction methods as part of ASHRAE research project
For direct-expansion evaporators, Kattan et al. (1998) proposed a RP-1089. For direct-expansion evaporators, Ould Didi et al. (2002)
flow-pattern-based method that includes effects of flow stratifica- describe seven prediction methods, compare them to data for five
tion at low flow rates and onset of dryout at high vapor qualities. refrigerants, and provide recommendations on the best choice.
Zürcher et al. (1998) in ASHRAE research project RP-800 added a
method to predict the adverse effect of oil on local flow boiling heat VESSEL DESIGN
transfer coefficients to the Kattan-Thome-Favrat model.
Mechanical Requirements
Fouling Factors Pressure vessels must be constructed and tested under the rules
Over time, most fluids foul the fluid-side heat transfer surface, of national, state, and local codes. The ASME Boiler and Pressure
reducing the cooler’s overall heat transfer coefficient. If fouling is Vessel Code, Section VIII, gives guidance on rules and exemp-
expected to be a problem, a mechanically cleanable cooler should be tions.
used, such as a flooded, Baudelot, or cleanable direct-expansion The more common applicable codes and standards are as fol-
tube-in-tube cooler. Direct-expansion shell-and-tube, shell-and- lows:
coil, and brazed-plate coolers can be cleaned chemically. Flooded
coolers and direct-expansion tube-in-tube coolers with enhanced 1. ARI Standard 480 covers industry criteria for standard equip-
fluid-side heat transfer surfaces tend to be self-cleaning because of ment, standard safety provisions, marking, and recommended
high fluid turbulence, so a smaller fouling factor can probably be rating requirements.
used for these coolers. Water quality in closed chilled-water loops 2. ASHRAE Standard 24 covers recommended testing methods for
has been studied as part of ASHRAE-sponsored research (RP-560). measuring liquid cooler capacity.
 Liquid Coolers 41.5

3. ASHRAE Standard 15 involves specific design criteria, use of should not be used in the presence of other metals in water cir-
materials, and testing. It refers to the ASME Boiler and Pressure cuits.
Vessel Code, Section VIII, for refrigerant-containing sides of Brines. Ferrous metal and a few nonferrous alloys are almost
pressure vessels, where applicable. Factory test pressures are universally used with sodium chloride and calcium chloride brines.
specified, and minimum design working pressures are given. Copper alloys can be used if adequate quantities of sodium dichro-
This code requires pressure-limiting and pressure-relief devices mate are added and caustic soda is used to neutralize the solution.
on refrigerant-containing systems, as applicable, and defines set- Even with ferrous metal, brines should be treated periodically to
ting and capacity requirements for these devices. hold pH near neutral.
4. ASME Boiler and Pressure Vessel Code, Section VIII, Unfired Ethylene glycol and propylene glycol are stable compounds that
Pressure Vessels, covers safety aspects of design and construc- are less corrosive than chloride brines.
tion. Most states require coolers to meet ASME requirements if
they fall within the scope of the ASME code. Some of the excep- Electrical Requirements
tions from meeting the ASME requirements are as follows: When the fluid being cooled is electrically conductive, the sys-
• Cooler shell inner diameter (ID) is 6 in. or less. tem must be grounded to prevent electrochemical corrosion.
• Pressure is 15 psig or less.
• The fluid portion of the cooler need not be built to the require- APPLICATION CONSIDERATIONS
ments of the ASME code if the fluid is water, the design pres-
sure does not exceed 300 psig, and the design temperature Refrigerant Flow Control
does not exceed 210°F. Direct-Expansion Coolers. The constant superheat thermal
expansion valve is the most common control used, located directly
Coolers meeting ASME code requirements have an ASME
upstream of the cooler. A thermal bulb strapped to the suction line
stamp, which is a U or UM inside a four-leaf clover. The U can
leaving the cooler senses refrigerant temperature. The valve can be
be used for all coolers, and the UM can be used for small coolers.
adjusted to produce a constant suction superheat during steady oper-
5. Underwriters Laboratories (UL) Standard 207 covers specific
Licensed for single user. © 2008 ASHRAE, Inc.

ation. If refrigerant pressure drop between the expansion valve and

design criteria, use of materials, testing, and initial approval by the thermal bulb is significant, an externally equalized expansion
Underwriters Laboratories. A cooler with the ASME U or UM valve should be used.
stamp does not require UL approval.
The thermal expansion valve adjustment is commonly set at a suc-
Design Pressure. On the refrigerant side, design pressure should tion superheat of 8 to 10°F, which is sufficient to ensure that liquid is
be chosen per ASHRAE Standard 15. Standby temperature and the not carried into the compressor. Direct-expansion cooler perfor-
temperature encountered during shipping of chillers with a refriger- mance is affected greatly by superheat setting. Reduced superheat
ant charge should also be considered. improves cooler performance; thus, suction superheat should be set
Required fluid-side (water-side) pressure varies, depending as low as possible while avoiding liquid carryover to the compressor.
largely on (1) static pressure, (2) pump pressure, (3) transients Flooded Coolers. As the name implies, flooded coolers must
caused by pump start-up, and (4) valve closing. have good liquid refrigerant coverage of the tubes to achieve good
performance. Liquid level control in a flooded cooler becomes the
Chemical Requirements principal issue in flow control. Some systems are designed critically
The following chemical requirements are given by NACE (1985) charged, so that when all the liquid refrigerant is delivered to the
and Perry and Green (2007). cooler, it is just enough for good tube coverage. In these systems, an
R-717 (Ammonia). Carbon steel and cast iron are the most orifice is often used as the throttling device between condenser and
widely used materials for ammonia systems. Stainless steel al- cooler.
loys are satisfactory but more costly. Copper and high-copper Level-sensing devices are another option. Typically, a high-side
alloys are avoided because they are attacked by ammonia when float valve or level-sensing system can meter flow to the cooler at a
moisture is present. Aluminum and aluminum alloys may be used controlled rate based on the condenser liquid level. For more direct
with caution. control of liquid level in the cooler, a low-side float valve or level-
Halocarbon Refrigerants. Almost all common metals and sensing system can be used to control flow of entering refrigerant
alloys are used satisfactorily with these refrigerants. Exceptions based on the level in the cooler itself. Because of the volatile nature
where water may be present include magnesium and aluminum of boiling inside the flooded cooler and the potential for false level
alloys containing more than 2% magnesium. Zinc is not recom- readings, consideration must be given to design and installation of a
mended for use with R-113; it is more chemically reactive than low-side liquid level sensing device.
other common construction metals and, therefore, it is usually
avoided when other halogenated hydrocarbons are used as refriger- Freeze Prevention
ants. Under some conditions with moisture present, halocarbon Freeze prevention must be considered for coolers operating near
refrigerants form acids that attack steel and even nonferrous metals. the fluid’s freezing point. In some coolers, freezing causes extensive
This problem does not commonly occur in properly cleaned and damage. Two methods can be used for freeze protection: (1) hold
dehydrated systems. ASHRAE Standard 15, paragraph 9.1.2, states saturated suction pressure above the fluid freezing point or (2) shut
that “aluminum, zinc, magnesium, or their alloys shall not be used the system off if fluid temperature approaches the freezing point.
in contact with methyl chloride,” or magnesium alloys with any A suction-pressure regulator can hold the saturated suction pres-
halogenated refrigerant. sure above the fluid’s freezing point. A low-pressure cutout can shut
Water. Relatively pure water can be satisfactory with both fer- the system off before the saturated suction pressure drops to below
rous and nonferrous metals. Brackish water, seawater, and some the freezing point of the fluid. The leaving fluid temperature can be
river water are quite corrosive to iron and steel and also to copper, monitored to cut the system off before a danger of freezing, usually
aluminum, and many alloys of these metals. A reputable water about 4°F above the fluid freezing temperature. It is recommended
consultant who knows the local water condition should be con- that both methods be used.
tacted. Chemical treatment by pH control, inhibitor applications, Baudelot, shell-and-coil, brazed-plate, and direct-expansion shell-
or both may be required. Where this is not feasible, more noble and-tube coolers are all somewhat resistant to damage caused by
construction material or special coatings must be used. Aluminum freezing and ideal for applications where freezing may be a problem.
 41.6 2008 ASHRAE Handbook—HVAC Systems and Equipment

If a cooler is installed in an unconditioned area, possible freezing Briggs, D.E. and E.H. Young. 1969. Modified Wilson plot techniques for
caused by low ambient temperature must be considered. If the obtaining heat transfer correlations for shell and tube heat exchangers.
cooler is used only when ambient temperature is above freezing, the Chemical Engineering Symposium Series 65(92):35-45.
fluid should be drained from the cooler for cold weather. Alterna- Casciaro, S. and J.R. Thome. 2001a. Thermal performance of flooded evap-
orators, Part 1: Review of boiling heat transfer studies. ASHRAE Trans-
tively, if the cooler is used year-round, the following methods can be
actions 107(1):903-918.
used to prevent freezing:
Casciaro, S. and J.R. Thome. 2001b. Thermal performance of flooded evap-
• Heat tape or other heating device to keep cooler above freezing orators, Part 2: Review of void fraction, two-phase pressure drop and
• For water, adding an appropriate amount of ethylene glycol flow pattern studies. ASHRAE Transactions 107(1):919-930.
• Continuous pump operation Chyu, M. 1995. Nozzle-sprayed flow rate distribution on a horizontal tube
bundle. ASHRAE Transactions 101(2):443-453.
Oil Return Collier, J.G. and J.R. Thome. 1994. Convective boiling and condensation,
3rd ed. Oxford University Press.
Most compressors discharge a small percentage of oil in the dis- Gupte, N.S. and R.L. Webb. 1995a. Shell-side boiling in flooded refrigerant
charge gas. This oil mixes with condensed refrigerant in the con- evaporators—Part I: Integral finned tubes. International Journal of
denser and flows to the cooler. Because the oil is nonvolatile, it does HVAC&R Research (now HVAC&R Research) 1(1):35-47.
not evaporate and may collect in the cooler. Gupte, N.S. and R.L. Webb. 1995b. Shell-side boiling in flooded refrigerant
In direct-expansion coolers, gas velocity in the tubes and suction evaporators—Part II: Enhanced tubes. International Journal of HVAC&R
gas header is usually sufficient to carry oil from the cooler into the Research (now HVAC&R Research) 1(1):48-60.
suction line. From there, with proper piping design, it can be carried Haider, S.I., R.L. Webb, and A.K. Meitz. 1991. A survey of water quality
back to the compressor. At light load and low temperature, oil may and its effect on fouling in flooded water chiller evaporators. ASHRAE
Transactions 97(1):55-67.
gather in the superheat section of the cooler, detracting from perfor-
Haider, S.I., R.L. Webb, and A.K. Meitz. 1992. An experimental study of
mance. For this reason, operating refrigerant circuits at light load for
tube-side fouling resistance in water-chiller-flooded evaporators.
long periods should be avoided, especially under low-temperature ASHRAE Transactions 98(2):86-103.
conditions. Some oil hold-up threshold measurements were pre- Kattan, N., J.R. Thome, and D. Favrat. 1998. Flow boiling in horizontal
Licensed for single user. © 2008 ASHRAE, Inc.

sented by Zürcher et al. (1998) in ASHRAE research project RP-800. tubes, Part 3: Development of a new heat transfer model based on flow
In flooded coolers, vapor velocity above the tube bundle is usu- patterns. Journal of Heat Transfer 120(1):156-165.
ally insufficient to return oil up the suction line, and oil tends to McAdams, W.H. 1954. Heat transmission, 3rd ed. McGraw-Hill, New York.
accumulate in the cooler. With time, depending on the compressor Moeykens, S.A., B.J. Newton, and M.B. Pate. 1995. Effects of surface en-
oil loss rate, oil concentration in the cooler may become large. hancement, film-feed supply rate, and bundle geometry on spray evapo-
When concentration exceeds about 1%, heat transfer performance ration heat transfer performance. ASHRAE Transactions 101(2):408-419.
may be adversely affected if enhanced tubing is used. Moeykens, S.A. and M.B. Pate. 1996. Effects of lubricant on spray evapo-
It is common in flooded coolers to take some oil-rich liquid and ration heat transfer performance of R-134a and R-22 in tube bundles.
ASHRAE Transactions 102(1):410-426.
return it to the compressor on a continuing basis, to establish a rate
NACE. 1985. Corrosion data survey—Metal section, 6th ed. D.L. Graver,
of return equal to the compressor oil loss rate. ed. National Association of Corrosion Engineers, Houston.
Ould Didi, M.B., N. Kattan, and J.R. Thome. 2002. Prediction of two-phase
Maintenance pressure gradients of refrigerants in horizontal tubes. International Jour-
Cooler maintenance centers around (1) safety and (2) cleaning nal of Refrigeration 25(7):935-947.
the fluid side. The cooler should be inspected periodically for any Pais, C. and R.L. Webb. 1991. Literature survey of pool boiling on enhanced
weakening of its pressure boundaries. Visual inspection for corro- surfaces. ASHRAE Transactions 97(1):79-89.
sion, erosion, and any deformities should be included, and any pres- Palen, J.W. and J. Taborek. 1969. Solution of shell side pressure drop and
sure relief device should also be inspected. The insurer of the cooler heat transfer by stream analysis method. Chemical Engineering Progress
may require regular inspection. If the fluid side is subjected to foul- Symposium Series 65(92).
ing, it may require periodic cleaning by either mechanical or chem- Perry, R.H. and D.W. Green. 2007. Perry’s chemical engineers handbook,
8th ed. McGraw-Hill, New York.
ical means. The manufacturer or a service organization experienced
Schlager, L.M., M.B. Pate, and A.E. Bergles. 1989. A comparison of 150
in cooler maintenance should have details for cleaning. and 300 SUS oil effects on refrigerant evaporation and condensation in a
smooth tube and a micro-fin tube. ASHRAE Transactions 95(1).
Insulation
Sturley, R.A. 1975. Increasing the design velocity of water and its effect on
A cooler operating at a saturated suction temperature lower than the copper tube heat exchangers. Paper 58, International Corrosion Forum,
ambient-air dew point should be insulated to prevent condensation. Toronto, Canada.
Direct-expansion coolers installed where the ambient temperature Thome, J.R. 1990. Enhanced boiling heat transfer. Hemisphere, New York.
may drop below the process fluid’s freezing point should also be in- Thome, J.R. 2003. Boiling. In Heat transfer handbook, pp. 635-717. A.
sulated and wrapped with heat tape, to prevent the fluid from freezing Bejan and A.D. Krause, eds. Wiley Interscience, New York.
and damaging the cooler during off periods. Chapter 23 of the 2005 UL. 2001. Refrigerant-containing components and accessories, nonelectrical.
ASHRAE Handbook—Fundamentals describes insulation in more de- Standard 207-01. Underwriters Laboratories, Northbrook, IL.
tail. Webb, R.L. 1994. Principles of enhanced heat transfer. John Wiley & Sons,
New York.
Webb, R.L. and T. Apparao. 1990. Performance of flooded refrigerant evap-
REFERENCES orators with enhanced tubes. Heat Transfer Engineering 11(2):29-43.
ARI. 2001. Remote type refrigerant-cooled liquid coolers. Standard 480-01. Webb, R.L. and C. Pais. 1991. Pool boiling data for five refrigerants on three
Air-Conditioning and Refrigeration Institute, Arlington, VA. tube geometries. ASHRAE Transactions 97(1):72-78.
ASHRAE. 2000. Methods of testing for rating liquid coolers. ANSI/ Webb, R.L., K.-D. Choi, and T. Apparao. 1989. A theoretical model for pre-
ASHRAE Standard 24-2000. diction of the heat load in flooded refrigerant evaporators. ASHRAE
ASHRAE. 2007. Safety standards for refrigeration systems. ANSI/ Transactions 95(1):326-338.
ASHRAE Standard 15-2007. Wolverine Tube, Inc. 2001. Wolverine engineering data book II. http://
ASME. 2007. Rules for construction of pressure vessels. ANSI/ASME www.wlv.com/products/databook/databook.pdf.
Boiler and Pressure Vessel Code, Section VIII. American Society of Zürcher, O., J.R. Thome, and D. Favrat. 1998. Intube flow boiling of R-407C
Mechanical Engineers, New York. and R-407C/oil mixtures, Part II: Plain tube results and predictions
Ayub, Z.H. and S.A. Jones. 1987. Tubeside erosion/corrosion in heat ex- (RP-800). International Journal of HVAC&R Research (now HVAC&R
changers. Heating/Piping/Air Conditioning (December):81. Research) 4(4):373-399.

Related Commercial Resources