EPIE&GY

AND
ELSEVIER
aUILDIPtG$
Energy and Buildings 22 (1995) 193-205

H y d r o n i c radiant c o o l i n g - preliminary a s s e s s m e n t *
Helmut E. Feustel, Corina Stetiu
Indoor Environment Program, Lawrence Berkeley Laboratory, Berkeley, CA 94720, USA

Abstract

A significant amount of the electrical energy used to cool non-residential buildings equipped with all-air systems is drawn
by the fans that transport the cool air through the thermal distribution system. Hydronic systems reduce the amount of air
transported through the building by separating the tasks of ventilation and thermal conditioning. Due to the physical properties
of water, hydronic systems can transport a given amount of thermal energy and use less than 5% of the otherwise necessary
fan energy. This improvement alone significantly reduces the energy consumption and peak-power requirement of the air
conditioning system. Radiant cooling has never penetrated the US markets significantly.The scope of this survey is to show
the advantages of radiant cooling in combination with hydronic thermal distribution systems, as compared to the commonly-
used all-air systems. The report descr~es the development, thermal comfort issues, and cooling performance of the hydronic
systems. The peak-power requirement is also compared for hydronic systems and conventional all-air systems.

Keywords: Hydronic radiant cooling; Radiant cooling

I. I n t r o d u c t i o n 30-

Cooling of non-residential buildings equipped with

all-air systems significantly contributes to the electrical
energy consumption and to the peak-power demand. 0 20-
Part of the energy used to cool buildings is consumed
by the fans that transport cool air through the ducts. o
This energy heats the conditioned air, and therefore ._1
adds to the internal thermal cooling peak load. Usibelli
et al. [1] found that, in the case of the typical office
building in Los Angeles, the external loads account for
only 42% of the thermal cooling peak (see Fig. 1). At
that time, 28% of the internal gains were produced by
lighting, 13% by air transport, 12% by people and 5% 0 ~.~.~%~o~ ' ., ~°°~o°" s
by equipment. The implementation of better windows, ~,~ "~'
together with higher plug loads due to increased use
Fig. 1. Peak cooling load components, typical office building, Los
of electronic office equipment, have probably caused Angeles.
these contributions to change to some extent since
then.
HVAC systems are designed to maintain indoor air DOE-2 simulations for different California climates
quality and provide thermal space conditioning. Tra- using the California Energy Commission (CEC) base
ditionally, HVAC systems are designed as all-air systems, case office building show that, at peak load, only 10-20%
which means that air is used to perform both tasks. of the supply air is outside air [2]..Only this small
fraction of the supply air is in fact necessary to ventilate
¢' This paper has been authored by a contractor of the US Gov-
the buildings in order to maintain a high level of indoor
ernment under contract No. DE-AC03-76SF00098. Accordingly, the
US Government retains a nonexclusive, royalty-free license to publish
air quality. For conventional HVAC systems the dif-
or reproduce the published form of this contribution, or allow others ference in volume between supply air and outside air
to do so, for US Government purposes. is made up by recirculated air. The recirculated air is

Elsevier Science S.A.

SSD1 0 3 7 8 - 7 7 8 8 ( 9 5 ) 0 0 9 2 2 - K
194 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205

necessary in these systems to keep the temperature height can be reduced, which offsets the initial cost of
difference between supply air and room air in the the additional system.
comfort range. The additional amount of supply air, The thermal storage capacity of the coolant in HRC
however, often causes draft 1 as well as indoor air systems helps to shift the peak cooling load to later
quality problems due to the distribution of pollutants hours. Because of the hydronic energy transport, this
throughout the building. cooling system has the potential to interact with thermal
All-air systems achieve the task of cooling a building energy storage systems (TES) and looped heat pump
by convection only. An alternative is to provide the systems.
cooling through a combination of radiation and con-
vection inside the building. This strategy uses cool
surfaces in a conditioned space to cool the air and the 2. Historical development of hydronic radiant
space enclosures. The systems based on this strategy cooling systems
are often called radiative cooling systems, although only
approximately 60% of the heat transfer is due to During the last decade, building inhabitants devel-
radiation. If the cooling of the surfaces is produced oped a critical attitude towards air conditioning systems.
using water as transport medium, the resulting systems Terms like complaint buildings and sick buildings were
are called hydronic radiant cooling systems (HRC sys- born. Several publications dealing with occupant sat-
tems). By providing cooling to the space surfaces rather isfaction in air-conditioned and naturally ventilated
than directly to the air, H R C systems allow the sep- buildings came to the conclusion that the number of
aration of the tasks of ventilation and thermal space unsatisfied occupants in air-conditioned buildings is
conditioning. While the primary air distribution is used significantly higher than in naturally ventilated buildings
to fulfill the ventilation requirements for a high level [4-6]. Esdorn et al. [7] state that "the existence of air
of indoor air quality, the secondary water distribution conditioning systems is actually only noticed when it
system provides thermal conditioning to the building. is not functioning properly".
HRC systems significantly reduce the amount of air Draft is one of the many problems associated with
transported through buildings, as the ventilation is the functioning of all-air systems. As the air from these
provided by outside air systems without the recirculating HVAC systems is normally turbulent in the occupied
air fraction. Due to the physical properties of water, zone, even low air velocities (less than 0.1 m s-l!)
HRC systems remove a given amount of thermal energy might cause an unwanted local cooling of the human
and use less than 5% of the otherwise necessary fan body [8]. In order to be able to remove cooling loads
energy. The separation of tasks not only improves from a building, the cooled air either exceeds in volume
comfort conditions, but increases indoor air quality and the outdoor air needed for ventilation (recirculating
improves the control and zoning of the system as well. air systems), or the temperature differences between
HRC systems combine temperature control of the room the supply air and the room air have to be so large
surfaces with the use of central air handling systems that the air supply might cause problems related to
[3]. the even distribution in the occupied zone (cold air
Due to the large surfaces available for heat exchange distribution systems). Because of comfort problems and
in HRC systems (usually almost a whole ceiling, and the excessive use of transport energy for all-air systems,
sometimes whole vertical walls), the temperature of new ventilation strategies were developed [9] such as
the coolant is only slightly lower than the room tem- displacement ventilation.
perature. This small temperature difference allows the The idea behind displacement ventilation is to over-
use of either heat pumps with very high coefficient of come the problems of mixing ventilation systems. Air
performance (COP) values, or of alternative cooling flows of low turbulent intensity supply clean air directly
sources (e.g. indirect evaporative cooling), to further to the breathing zone and displace contaminants [10].
reduce the electric power requirements. HRC systems If the air is displaced upwards, cooling is also achieved,
also reduce problems caused by duct leakage, as the but this strategy will result in an increase of the vertical
ventilation air flow is significantly reduced, and the air temperature gradient of the air. The natural driving
is only conditioned to meet room temperature con- forces of the vertical air transport are the heat sources
ditions, rather than cooled to meet the necessary supply in the space, as they create convective air currents
air temperature conditions. Furthermore, space needs (plumes 2). The resulting air flow pattern has greatly
for ventilation systems and their duct work are reduced improved ventilation efficiency (for definition of ven-
to about 20% of the original space requirements. Besides tilation efficiency, see Ref. [11]).
the reduction of space requirements for the shafts that
house the vertical air distribution system, floor-to-floor 2 Displacement ventilation should not be mistaken for 'plug flow'
or 'piston flow'; plume flow ventilation might be a better term for
1 Draft: air movement in an occupied enclosure causing discomfort. displacement ventilation.
 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205 195

Cooling D e q c ~
sensations of comfort between outdoors and indoors."
This example suggests that the surface temperatures
surrounding a person inside an enclosure have a great
influence on the thermal comfort of that person, and
therefore have to be carefully studied.
If people could not lose heat by radiation, and
convection were the only available heat loss mechanism,
high air velocities close to the human skin would be
required in order to produce a given heat loss. A
continuous increase in air velocities would eventually
lead to draft, and therefore, to uncomfortable condi-
tions. The possibilities of increasing the heat loss by
"L ake of F r ~ b Air" .~
respiration or excretion are very limited.
Heat loss by radiation is caused by the difference
between the body temperature and the mean radiant
Fig. 2. Air flow patterns in a room with a cooled ceiling. temperature, which depends on the temperatures of
the surrounding surfaces. The mean radiant temperature
Upward displacement ventilation shows a charac- is easy to define but quite complicated to calculate or
teristic temperature profile caused by the convective measure in practice. Due to the non-uniform distances
currents driven by the heat sources. As supply air enters and angles of persons in relation to the walls, floor
the room at floor level, the temperature gradient forms and ceiling of a space, each part of the space must be
a barrier that prevents low energy currents reaching considered separately in the radiation exchange. If a
high altitudes in the room. Due to comfort requirements, given surface is found to not be isothermal, it has to
the temperature gradient between feet and head cannot be divided into smaller isothermal surfaces. Each surface
exceed 3 °C, which limits the cooling capacity of dis- can be assumed to have high emissivity. The radiation
placement ventilation systems [10]. The fact that dis- emitted and reflected from any surface is distributed
placement ventilation systems use solely outside air as diffuse radiation, which is a good approximation for
further reduces their cooling capacity [12,13]. all normal non-metallic surfaces [15]. The enclosure
The most efficient way to use displacement ventilation surfaces often found in a normal room have a rectangular
is to associate it with a cooling source that does not shape and, therefore, the angle factor in the mean
require air transport inside the room. The logical choice radiant temperature calculations is defined between a
is the coupling of displacement ventilation systems with person and a vertical or horizontal plane. The body
hydronic radiant cooling, a strategy that allows the posture is also important. The mean radiant temperature
separation of the tasks of ventilating and cooling in in relation to a standing person is not necessarily the
the building. The theoretical air flow pattern in a room same as in relation to a seated one [15]. Likewise, the
with a cooled ceiling is shown in Fig. 2 [10]. location and orientation of the person inside the room
must also be known, .because the mean radiant tem-
perature often varies from point to point.
3. Thermal comfort The first experiments of thermal and comfort sen-
sations to radiation experienced by seated persons were
In order to maintain normal functions, the human conducted by Schlegel and McNall [16], and McNall
body needs to maintain the balance between heat gain and Biddison [17]. Fanger [15] defined mean radiant
and heat loss. Heat can be lost in different ways: temperature as follows: "The mean radiant temperature
radiation to surrounding surfaces, convection to the in relation to a person in a given body posture and
ambient air, conduction, evaporation, respiration and clothing placed at a given point in a room, is defined
excretion. The most important loss is due to radiation, as that uniform temperature of black surroundings which
followed in order by convection and conduction. Res- will give the same radiant heat loss from the person
piration and excretion have less influence on the heat as the actual case under study."
loss of a human body. The combined effects of radiation and convection
To explain the impact of radiation, Baker [14] gives inside an enclosure can be evaluated by using a pa-
the following example: "A person sitting out of doors rameter called the 'operative temperature'. The op-
under a clear sky on a summer evening may be chilly erative temperature is calculated by averaging the dry
although the air temperature is in the high 70s (°F). bulb temperature and the mean radiant temperature
Were he indoors at this same temperature, he probably inside the enclosure. The definition of the operative
would feel uncomfortably warm. The appreciable heat temperature shows that this parameter does not reflect
loss by radiation to the clear sky explains the different the presence of radiation asymmetry inside an enclosure.
196 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205

In the case when this effect is important, the use of scription of convective heat transfer from a cold flat
operative temperature in evaluating thermal comfort surface (downwards) as published in textbooks. Inves-
might lead to erroneous results. tigations dealing with cooled ceilings show overall heat
Air movement plays a special role among the comfort transfer coefficients of 9-12 W m -2 K. Given a heat
parameters. According to Esdorn et al. [7], air movement transfer coefficient for radiation of about 5.5 W m - 2
is the largest single cause for complaints (draft). Besides K (for a temperature difference of 10 °C), the resulting
the mean air velocity, the fluctuation of the air velocity convective heat transfer coefficient would be in the
has an important influence on the convective heat order of 3.5--6.5 W m - 2 K. These values for the con-
transfer of the human body. Mayer [18] relates comfort vective heat transfer coefficient are however reached
directly to the convective heat transfer coefficient, rather only if forced convection takes place (here forced means
than to the average air velocity. According to Mayer that phenomena other than the cooling at the ceiling
[19], at an air temperature of 22 °C draft is felt if the are responsible for driving the air flow).
convective heat transfer coefficient is above 12 W m -2 Radiant cooling elements extract heat from a room
K. This translates to average air velocities for laminar by cooling the air directly (convection) and indirectly,
flows of 1.35 m s -a, for transition flows of 0.15 m s -a, by cooling the surfaces of the room envelope. If there
and for turbulent flows of 0.10 m s -1. Lower air is only a small difference between the average surface
temperatures significantly reduce the acceptable air temperature of the room and the air temperature, the
velocities. two effects can be estimated jointly [22]. Under this
Extended explanations about the recommended levels assumption, the specific cooling power (per unit area)
of thermal comfort parameters can be found in the of a cooled ceiling can be expressed by the following
ASHRAE Standard 55-92, ISO 7730 and the 1993 (empirical) equation (see Appendix):
ASHRAE Handbook of Fundamentals, Ch. 8.
qtot=8.92(tair--tsu,~c~) 11 (1)
q,ot sum of convective and radiant heat transfer (W
4. Cooling power m - 2).
A survey of cooled ceilings [23] shows cooling output
The cooling power of HRC systems is limited due ranging from 40 to 125 W m - 2 . The data are however
to the fact that in operating these systems, the side- based on information from manufacturers and do not
effects associated with the presence of cold surfaces specify the boundary conditions for the measurements.
in the space have to be prevented or minimized. A This problem brings up the necessity of standards for
first effect to prevent is condensation. In theory, the both the measurement conditions and techniques, and
surface temperatures of the cooling elements must not several attempts have already been made to set standards
be lower than the dew point temperature of the air for testing radiant panels.
in the cooled zone. In practice, however, condensation A test facility and a method of testing was developed
prevention reduces the effective cooling temperature at the Department of Veterans Affairs [24]. The method
difference (between the cold surface and the air) by describes the testing procedure for thermal performance
a safety margin of approximately 2 °C. In the operation and pressure drop measurements in the test facility,
of the system the dew point can also be manipulated, as well as the accuracy of the instrumentation used.
by reducing the humidity content in the ventilation air. ASHRAE's technical committee TC 6.5 Radiant
A second and more serious concern is the comfort Space Heating and Cooling currently sponsors a com-
effect of the asymmetrical distribution of the radiant mittee on Methods of Testing/Rating Hydronic Radiant
temperature. Kollmar [20] shows that for offices, the Ceiling Panels (SPC 138P). The purpose of SPC 138P
lower limit for ceiling temperatures is approximately is to establish a method of testing for rating the thermal
15 °C. performance of hydronic radiant cooling panels used
To calculate the cooling power of a hydronic radiant for heating and/or cooling of indoor space [25].
cooling system, the heat transfer between the room In Germany, two competing testing procedures have
and the cold ceiling has to be evaluated. There are been recently published within five months. The Fa-
two components of the heat transfer: radiation and chinstitut Gebaeude-Klima (FGK) presented its testing
convection. While the radiation term is relatively easy procedure in December 1992 [26]. The FGK industrial
to calculate, the convective heat transfer is a function standard is based on a measurement in a box (2.4 m
of the air velocity at the ceiling level. This velocity is × 1.2 m × 1.5 m) with an internal operative temperature
dependent on the room geometry, the location and of 26 °C and water supply temperatures of 12, 14 and
power of the heat sources, and the location of the air 16 °C. The DIN standard was presented in April 1993
inlet and exhaust. [27]. It measures the performance of radiant panels in
Trogisch [21] compares heat transfer coefficients for the presence of natural convection. The test is based
cooled ceilings found in the literature with the de- on measurements performed in a closed test chamber
 H.E. Feuste~ C. Stetiu / Energy and Buildings 22 (1995) 193-205 197

(4 m × 4 m × 3 m) with a conditioned metal envelope. v

; +
The cooling load is simulated by 12 perforated tubes
containing three 60 W bulbs each. The measurements
are performed under steady-state conditions, with dif-
ferent coolant mass flows.
While testing procedures and future standards can
rate the performance of an H R C system with panels
under given boundary conditions, the efficiency of the
same system in a specific (but different) application is • ~r - ~.
difficult to determine. The difficulty arises from the
fact that the rated performance greatly depends on the
R :>eClton R~ Return
testing procedure. For example, a procedure for meas-
uring the efficiency of a cooled ceiling could use the
temperatures of the ceiling and of the exhaust air in
a test room as a measure for the convective heat transfer
T'J Per,ion o Pe~ TS
between the ceiling and the room air. If, in a hypothetical
situation, a shortcut between the supply and the exhaust
of the ventilation system in the test room provided
high air velocities on the ceiling surface, a large fraction
of the exhaust air would in fact be air that has been ~- - ~ ,I . . . . - - ~ - ~ -- .-I
cooled by the ceiling without having interacted with
j Pe~
the room loads at all. The difference between the
temperatures of the ceiling surface and the exhaust air
would in this case indicate lower convective heat transfer
than in reality, the measurements would appear as
Fig. 3. T e s t c h a m b e r u s e d to test the cooling p e r f o r m a n c e of h y d r o n i c
having been performed under low air flows, and the cooling.
ceiling would show high efficiency. The functioning of
the same ceiling in a normal situation (without the
forced convection) is very likely to give different results.
These considerations show that a building designer 287 z~ . . . . -:.~ - -~
Underside C e i l i n g 7
/
should carefully consider all the details in a testing
procedure before deciding to use a specific type of ExhaUst Air Opening r- / /
/

cooling ceiling panel. i

I"
i i i
2.0 I I
i i
5. Cooling performance
t ? I
I
Although several papers have been found that de- E I
o i i
scribe the cooling power of H R C systems, only a few i i
papers are found that investigate the performance of 1.0 I
I
these systems. Kuelpmann [28] reports on an experi- I
mental investigation in a temperature controlled test f
I
cell (see Fig. 3). In this experiment, the air was supplied
at floor level and exhausted approximately 0.2 m below
the ceiling level. Internal loads were simulated by i
Supply A i r Opening
electrically heated mannequins (dummies) standing next 0.0 .-.- f-- 71-- /

to a computer display, and by fluorescent lights. External 15 20 25

Average A i r T e m p e r a t u r e [*c]
loads were introduced by heating either one of the
long side walls, or the floor. For displacement ventilation
Symt)ol Cooling Room Temp.-Diff. P e r f o r m a n c e P a r t (-)
and no cooling with supply air, the room air temperatures Load Temperature Wails-Air Ceiling Ventilation
0N/m=} ('C) (K)
measured at different heights did not differ very much
0 35 21A 0.2 0.96 0.04
(see Fig. 4). The extraction of 100 W m -~ internal O 67 24.4 -0.2 1.00 0.00
& 101 1.00 0.00
load by hydronic radiant cooling caused temperature 26.7 -0,3

differences of approximately 2 °C between the supply Fig. 4. A i r t e m p e r a t u r e profile o v e r the r o o m h e i g h t as a function

and exhaust grilles. When the temperature difference of e x t e r n a l loads; t a b u l a t e d r o o m t e m p e r a t u r e is m e a s u r e d at 1.1
between the room air and the supply air was increased, m a b o v e the floor.
198 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205

the profile became more pronounced. Especially in the room air. Temperature measurements were also made
lower part of the room, these temperature differences in the supply and return pipes of the hydronie system
became close to, or exceeded, the comfort limits. and at three points on the ceiling surface. For outside
In all cases examined, the differences between the air temperatures of 30 °C, the air velocities measured
room air temperature and the surface temperatures of in the occupied zone were below 0.10 m s -1. This
the 'internal walls' were relatively small (_+0.4 °C). value shows that the risk of draft has been eliminated
Due to the radiation exchange with the cooled ceiling, in the tested zones. Below the ceiling, surface velocities
the floor surface temperature was usually below the between 0.10 and 0.15 m s -1 were detected. These low
wall surface temperatures. velocities assure that less than 40% of the heat transfer
Asymmetric or non-uniform thermal radiation may occurs by convection.
be caused in winter by cold windows, uninsulated walls
or heated ceilings. In summer, cooled ceiling panels
also produce asymmetric thermal radiation. Radiant 6. Numerical modelling
asymmetry due to a cooled ceiling causes less discomfort
than a warm ceiling. Based on Fanger's limit of 5% The evaluation of the theoretical performance of
uncomfortable as a rule for determining the acceptability HRC systems could most conveniently be made by
of a system, a radiant temperature asymmetry of 10 computer models. Energy analysis programs such as
°C is acceptable in the presence of a cool wall, and DOE-2 do not have the capacity of simulating hydronic
of 14 °C in the presence of a cooled ceiling (see Fig. cooling systems yet. There have been attempts to adapt
5) [291 . DOE-2 so that it roughly performs the task [31], but
Measurements of radiant temperature asymmetry at this approach involves laborious artifices, and is not
101 W m -2 cooling power in the reported investigation easily accessible to the average DOE-2 user. The con-
resulted in 8 °C difference at 1.i m above the floor sensus is that a separate module needs to be designed
level, in the middle of the room (see Fig. 4). This which models the specifics of hydronic radiant cooling.
corresponds to less than 2% of occupants dissatisfied Such a model is presently being developed at the
(see Fig. 5). Lawrence Berkeley Laboratory by the Energy Per-
Air flow velocities were measured at 1 m distance formance of Buildings Group. It is a SPARK-based
to the supply air grille, at 0.1 m height above ground. module (SPARK=Simulation Problem Analysis Re-
At an air exchange rate of 3.2 ACH and a supply air search Kernel) that performs most of the tasks of DOE-
temperature of 19 °C, exceptionally low values were 2, on the particular case of the HRC systems [32].
measured for the air velocity (0.12 m s l) and the
turbulence intensity (20%).
The performance of hydronic radiant cooling was 7. Energy savings
tested in two parliamentarian offices in Bonn, Germany
[30]. Dry-bulb temperatures and relative humidity were The use of HRC systems is an energy conserving
measured for the outside air, the supply air and the and peak-power reducing alternative to conventional
air conditioning, particularly suited to dry climates. A
100 significant amount of the electrical energy used to cool
WARM CEILING buildings by all-air systems is consumed by the fans
which are used to transport cool air through the ducts.
Part of this electricity used to move the air is heating
the conditioned air and, therefore, is part of the internal
thermal cooling peak load. The electrical cooling peak
load, if defined as the load from the fans and the
chillers, has a breakdown of approximately 37% for
~3 10 COOL CEILING WARM WALl
running the fans and 63% for using the chillers.
If the tasks of ventilation and thermal conditioning
of buildings are separated, the amount of air transported
through buildings can be significantly reduced. In this
case the cooling is provided by radiation using water
as the transport medium, and the ventilation by outside
air systems without the recireulating air fraction. Al-
though the supply air necessary for ventilation purposes
0 5 10 ~5 20 25 30 35 40
RADIAN] IFMPERATURE ASYMM~I-RY, % is still distributed through ducts, the electrical energy
Fig. 5. Measured percentage of people expressing discomfort due for fans and pumps can be reduced to approximately
to asymmetric radiation. 25% of the original value.
 H.E. Feustel, C. Stetiu I Energy and Buildings 22 (1995) 193-205 199

(100%) specified as supply air temperature. The temperature

adjustment depends on the pressure drop, fan efficiency
and volume flow. In our example, this air handling
(1.5%) temperature rise has been assumed to be 1.0 °C.
The electrical power for an all-air system amounts
in this example to
Zaelll-air system = 1270 W
*,d from A i r
~porl
9%)
In order to be able to compare the two systems, the
boundary conditions have to be equal. This includes
efficiencies of fans and motors, pressure losses for supply
Conventional Radiant Cooling and exhaust ducts, and chiller COPs.
HVAC System HVAC System
Whereas the all-air system removes the cooling load
Fig. 6. Comparison of electrical peak power load for an all-air system by means of supplying cold air, the HRC system removes
and an H R C system, based on the peak-load data shown in Fig. 1.
the load mainly by means of water circulation. The
Percentages for the H R C system are relative to the overall peak
power of the all-air system. tasks of the ventilation side of the system are thus to
supply the room with the necessary air exchange rate
for comfort reasons, and to control the dew-point in
The elimination of recirculation air also increases
the room, to avoid humidity buildup. In order to provide
the efficiency of air-handling luminaries, as the con-
a stable displacement ventilation, the supply air vo.lume
vective heat extracted from the light fixtures is not
flow should be about 3 °C below the room air tem-
recirculated as it partly would be in an all-air system,
perature. The required temperature of the supply air
but vented directly to the exterior. 50% of the required
is thus 23 °C, which reduces the hydronic cooling load
thermal cooling energy produced by lighting can be
by around 3 W m - L
removed in this way. Compared to a constant volume
In order to control humidity, the cooling of the
air system, an overall electrical cooling energy savings
outside air below the supply air temperature might be
potential of more than 40% seems reasonable (see Fig.
necessary. A reheater can be installed which warms
6). the air using waste heat from the compressor. The
warming of the air could be done even more efficiently
if the air were channelled through building components
8. Peak-power requirement before arriving to the room inlet. This would save the
power to reheat and provide some conditioning at the
In order to compare the electrical peak-power re- same time. The electrical power of the HRC system
quirement for conventional systems (all-air systems) amounts to
and advanced systems (HRC systems), the power re-
~]Q~i. . . . d . . . . . . = 909 W
quirement for a simple example has been calculated.
The example is based on an office with a floor area Table 2 shows the electrical power calculated for an
of 25 m, a two person occupancy, and a total heat gain all-air system and an HRC system. The values in the
of 2000 W. The specific cooling load amounts to 80 table show that the power required by the HRC system
W m -2, which is in the range of HRC systems. The is only about 72% of the total electrical power required
room temperature is set to 26 °C. Additional assumptions by the all-air system. The savings could be increased
and design considerations used for this example are by venting the light fixtures in the building.
shown in Table 1.
The all-air system supplies cooling to the room as
follows: a cooling coil dehumidifies the outside air 9. Economics
according to the required room conditions. ASHRAE
Standard 62-1989 requests a minimum air change rate Although manufacturers of HRC systems claim many
of 36 m s h -a per person, which means that for this installations of their systems, it is difficult to obtain
example the minimum air change rate is 72 m 3 h -1. information about the economics of the systems in-
In order to remove the internal load, a recirculating stalled. There are only a few papers that deal with the
air volume flow of 678 m 3 h - 1 is required. The assumed economics of HRC systems.
outside air condition of 32 °C leads to a mixing tem- Feil [33] compared different ventilation/cooling sys-
perature of 25.6 °C. tems for an office. In a first-cost comparison with a
After having mixed, the air enters a cooler. In order VAV system, the break-even point for HRC systems
to adjust for the temperature increase due to the fan is approximately at a specific cooling load of 55 W
work, the air has to be cooled further than the 18 °C m -2.
200 H.E. FeusteL C. Stetiu / Energy and Buildings 22 (1995) 193-205

Table 1
Assumptions used for the comparison of peak-power requirements for an all-air system and an HRC system

Both systems

Room conditions
Cooling load (W m -z) 80
Room air temperature (K) 26
Relative humidity (%) 50
Humidity ratio (g~,.tcr/kgun, air) 10.6
Number of people 2
Outside air conditions
Air temperature (°C) 32
Relative humidity (%) 40
Humidity ratio (gwater/kgdryair) 12.1
Enthalpy (kJ/kg) 63
All-air system HRC system

Design considerations
Outside air flow (m 3 h - I ) 72 72
Supply air flow (m 3 h -l) 750 72
Temperature differences
Room air - supply air (K) 8 3
Room air - ceiling (K) 0 8
Supply water - return water (K) 2
Effic&ncies
Fan: hydraulic/mechanical/electrical (%) 60/80/98 60/80/98
Water pump (%) 60
Pressure drop
Supply duct/return duct/water pipe (Pa) 500/250/- 500/250/40000
COP 3 3

Table 2 Table 3
Electrical power requirement to remove internal loads from a two- Yearly energy consumption (kWh m - t ) for an office building in
person office with a floor area of 25 m Europe

All-air system H R C system VAV system H R C system

Supply fan 222 W 21 W Heating 43 43

Air cooler 721 W DHW 4 4
Pre-cooler/dehumidifier 216 W 216 W Lights 34 34
Exhaust fan 111 W 11 W Miscellaneous 10 10
Water pump 20 W Ventilation 12 8
Water cooler 641 W Fans/pumps 31 24
Total 1270 W 909 W Cooling 7 8
100% 71.5% Sum 141 131

Hoenmann and Nuessle [34] estimated yearly energy Table 4

Space requirements for systems in office buildings
consumption for an office building in Europe (see Table
3). The building has 5000 m 2 of floor area distributed VAV system H R C system
over four floors. The specific cooling load is 50 W m-2.
The relatively low savings potential for the overall energy Shafts 25 m 2 18 m 2
consumption of the building (less than 8%), is due to Equipment rooms 165 m 2 107 m 2
Plenum height 0.4 m 0.2 m
the large energy consumption by heating and lighting.
Unfortunately, the authors do not provide consumption
data for cooling only. Furthermore, the outside-air-only
VAV system utilized an economizer mode, while the in the equipment rooms, followed by 28% for the
same savings potential has not been matched in the distribution shafts. For systems with false ceilings, the
HRC system by installing a water-economizer. reduction in height per floor is in the order of 0.15-0.20
The space requirements for the two systems are m. HRC systems which are integrated into the ceiling
shown in Table 4 [34]. The largest savings, 36%, appear produce even higher savings.
 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205 201

7%
I000

900

~00

700 j - I . |

|
•

•
•

•
• O

O
Q

Q
Q

O
• I
• I ~ Concrete
c:: 60O
- ~ ~ ~,c~ ..... | • • g

o 5o0.
l i - - , ~ - - - - Air Space

E 400-
[I Panels
I
300 -

200 -

100 --- [ - Direction of

0
heat flow
?(l 3O ao 50 60 7O .o 0~ ,00
Specific Cooling Load in W/m 2
Fig. 9. Heat transfer for panel system (cooling mode).
Fig. 7. Comparison between first cost for V A V systems and H R C
systems, as a function of the specific cooling load.

Fig. 8. Construction of a cooling panel system. Drawing by Fl~ikt

Lufttechnik GmbH.

For first cost calculations, Hoenmann and Nuessle

[34] show a break-even point for their aluminum panel
system at 50 W m -2 at an air exchange rate of 3 ACH
(see Fig. 7 [34]).

10. Systems
Fig. 10. Cooling ceiling in an office environment (courtesy of R E D E C
AG).
Most of the H R C systems belong to one of four
different system designs. The most often used system
is the panel system. This system is built from aluminum sulated panel backsides uncouples the thermal storage
panels with metal tubes connected to the rear of the of the slab in this arrangement. This arrangement
panel (see Fig. 8 [34]). improves the response time for startup conditions, but
The connection between the panel and the tube is loses the ability of smoothing cooling load peaks.
critical. Poor connections provide only limited heat The temperature profiles for the different ceiling
exchange between the tube and the panel, which results panel systems have been published by Graeff [35].
in increased temperature differences between the panel Cooling grids made of small plastic tubes placed
surface and the cooling fluid. Panels built in a 'sandwich close to each other can be imbedded in plaster, gypsum
system' include the water flow paths between two board or mounted on ceiling panels (e.g. acoustic ceiling
aluminum panels (like the evaporator in a refrigerator). elements) (see Fig. 11). This second system provides
This arrangement reduces the heat transfer problem an even surface temperature distribution. Due to the
and increases the directly cooled panel surface. flexibility of the plastic tubes this system might be the
In the case of panels suspended below a concrete best choice for retrofit applications. It was developed
slab, approximately 93% of the cooling power is available in Germany and has been on the market for several
to cool the room. The remaining 7% cools the floor years.
of the room above (see Fig. 9). When the tubes are imbedded in plaster, the heat
Fig. 10 shows a typical installation of suspended transfer from above is higher than in the case of cooling
ceiling panels. A closed panel arrangement with in- panels (Fig. 12). The heat transfer to the concrete
202 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205

17c7~

ion
te

1| 1-~*~NN°°°
-*"° ° ° °° °°o° I!lZ ~ (TubeSheatDirectionflowEmbedded)of
Fig.11.Constructionofa coolinggrid(courtesyofKaRo-Information
Service).
18% 9%
Fig. 14. Heat transferfor slabcooling.
Carpet
Floor A third system is based on the idea of a floor heating
system. The tubes are imbedded in the core of a concrete
Insulation ceiling. The thermal storage capacity of the ceiling
Concrete allows for peak load shifting, which provides the op-
portunity to use this system in association with alter-
i i~ Tubesembedded
~ I~ I inplaster native cooling sources. Due to the thermal storage
involved, the control of this system is limited. This
leads to the requirement of relatively high surface
temperatures to avoid uncomfortable conditions in the
Directionof case of reduced cooling loads. The cooling power of
heatflow the system is therefore limited [36]. This system is
particularly suited for alternative cooling sources, es-
Fig. 12. Heat transferfor ceilingwithcoolinggrid, pecially the heat exchange with cold night air. The
faster warming of rooms with a particular high thermal
load can be avoided by running the circulation pump
for short times during the day to achieve a balance
with rooms with a lower thermal load.
Due to the location of the cooling tubes in this system,
a higher portion of the cooling is applied to the floor
of the space above the slab. Approximately 83% of the
heat removed by the circulated water are from the
room below the slab, while 17% are from the room
above (Fig. 14).
A fourth system has been developed in Germany,
which is also commercially available in California. It
provides cooling to a raised floor. The floor provides
space for the tubes and the supply plenum. Air is
supplied below the windows, reducing the radiative
Fig.13.Coolinggridattachedto concreteceilingbeforebeingcovered effect of cold window surfaces in winter and hot window
withplaster(courtesyof KaRo-lnformationService). surfaces in summer [37].

couples the cooling grid to the structural thermal storage

of the slab. Plastic tubes mounted on suspended cooling II. Control issues
panels show thermal performance comparable to the
panel systems described above. Tubes imbedded in a As mentioned before, the cooling power of radiative
gypsum board can be directly attached to a wooden heat exchange is limited by the dew-point of the room
ceiling structure without a concrete slab. Insulation air. In order to avoid condensation the cooling surface
must be applied to reduce cooling of the floor above. is kept above the dew-point for all operation conditions.
Fig. 13 shows cooling grids attached to a concrete ceiling If the dew-point is reduced by dehumidifying the supply
before being covered with plaster. air, higher thermal loads can be removed by means of
 H.E. Feuste~ C. Stetiu / Energy and Buildings 22 (1995) 193-205 203

radiation. This means that the surface temperature of a space is not only dependent on the air temperature,
the cooled area can be reduced to increase the operative but also on the (dynamic) distribution of the surface
temperature difference. Precautions should however be temperatures in the space. As thermal building sim-
taken in order to keep the inside air within the specified ulation programs do not provide the data necessary
comfort limits. for evaluating the performance of radiant systems, the
Besides the option of reducing the dew-point in order development of dynamic models is needed to better
to avoid damage due to condensation, there is the understand comfort issues.
possibility of switching off the supply of cold water as Table 5 summarizes the features of H R C systems
soon as the relative humidity reaches 'dangerous' levels. and their effects.
An alternative to this is a type of control present in
a number of buildings, which consists of window contacts
that cut off the water supply when windows are opened. 13. Conclusions
This way, the ventilation system is from the start avoiding
operation outside design conditions. Hydronic radiant cooling systems have become an
Temperature controls for different H R C systems have alternative to conventional all-air systems in Europe.
very different response times. All of the systems working Although research results have been found which sup-
with thermal mass are relatively slow in response to port the data advertised by manufacturers in Europe,
load changes. If operation allows the room temperature there is not sufficient data available for proving that
to swing and cooling loads can be matched by these these systems could perform well in US climates. An
systems, they are however the most energy efficient analysis based on typical US climates is necessary for
systems available. Systems with water supply close to a full assessment of the possible applications of hydronic
the cooling surface and with little thermal mass (panel radiant cooling. There are reasons to believe that
systems) have a response time comparable to all-air hydronic radiant cooling could work even for climates
systems. other than the dry California or Arizona climates.
Although controlling the dew-point by dehumidifying Due to the limited cooling power available from these
the outside air requires lower cooler temperatures in systems, the building design has to be done carefully,
the air handling system than in the case when outside with focus on the reduction of the building cooling
air is used together with recirculation air for this task, peak power requirement. There is obvious potential
the process is usually more energy efficient because of for the systems to perform well in different California
the smaller amount of air which has to be cooled before climates, but a study outlining the building design
the dew-point is reached. measures associated with the implementation of the
systems should nevertheless be performed.
Several H R C systems have already been installed in
12. Summary California. There is some anecdotal evidence that some
of these systems do not perform to the satisfaction of
Although hydronic radiant cooling was implemented the occupants: they either seem to consume more energy
in the US, a significant market penetration has never than predicted, or to have problems providing thermal
been reached. Hydronic radiant cooling was more or comfort. These buildings should be studied to determine
less also abandoned in Europe, after some applications the status of the systems, and to unveil the causes of
in the late thirties and in the fifties. User complaints the performance problems.
about all-air systems have nevertheless changed the Current energy analysis programs such as DOE-2
designers' attitude towards these systems, and have led cannot model H R C systems. As a result, there is no
to new system designs with better control. Together way to predict the expected performance of these
with efficient ventilation systems and humidity control, systems, which inhibits their use. A computer model
H R C systems have several advantages when compared is currently being developed at the Lawrence Berkeley
to conventional HVAC systems. Laboratory that will allow users to calculate heat ex-
The reviewed literature shows that H R C systems traction rates and room surface temperature distri-
provide draft-free cooling, reduce space requirements, butions for radiant cooling systems.
increase indoor air quality, reduce the energy con- Performance test-methods should be developed to
sumption for thermal distribution and for space con- compare different products under standard conditions,
ditioning, and might even have lower first-cost, if specific and field tests should be performed showing the in-
cooling loads are above 55 W m -2. fluence of climate, building design and room layout on
Unfortunately, literature has not been found which hydronic radiant cooling performance.
describes the dynamic thermal behavior of the system As mentioned earlier, because they function at low
and the building. Dynamics might be an important issue temperature differences between the room air and the
in further studies, because the comfort temperature in coolant, hydronic radiant cooling strategies are obvious
 204 H.E. Feustel, C. Stetiu / Energy and Buildings 22 (1995) 193-205

Table 5
Summary of HRC systems

Features Effect

Separate ventilation and thermal conditioning reduce air movement

improve comfort
Transport cooling energy by means of water reduce transport energy
reduce peak-power requirement
Use large cooling surfaces cool at high temperature level
Eliminate recirculation air improve indoor air quality
Reduce convection improve comfort
Reduce size of thermal distribution system improve space usage
reduce building cost
Can utilize alternative cooling sources reduce energy consumption
reduce peak-power requirement
Limited cooling output need accurate sizing
Increased risk of condensation need good humidity control

candidates to be associated with cooling sources other L mean radiant temperature of an uncondi-
than compressors. Alternative cooling sources should tioned surface (K)
therefore be investigated, and their energy savings mean radiant temperature of a cooled sur-
potential, peak-power reduction and interaction with face (K)
the H R C systems should be determined. F, configuration factor (-)
F~ emissivity factor (-)
Acknowledgements
This shows that the overall heat extraction is a function
The research reported here was funded by the Cal- of the temperature differences between the cooling
ifornia Institute for Energy Efficiency (CIEE), a research panel and the air as well as the different surfaces. Both
unit of the University of California. Publication of convection and radiation can be expressed by means
research results does not imply CIEE endorsement of, of heat transfer coefficients. The combined heat transfer
or agreement with, these findings, nor that of any CIEE coefficient can be calculated using an empirical equation.
sponsor. This work was also supported by the Assistant Glueck (1990) developed the following equation based
Secretary for Conservation and Renewable Energy, on measurements of cooled ceilings:
Office of Buildings and Community Systems Building
Systems Division of the US Department of Energy oqot= 8.92(t.i~ - t~rf~ce)°'I
under Contract No. DE-AC03-76SF00098. where ato,=Sum of convective and radiant heat coef-
ficient (W/m 2 K).
Appendix: Cooling power This empirical equation is based on the assumption
that the mean surface temperature of the room differs
Radiant cooling elements extract heat from a room only slightly from the air temperature. With this as-
by cooling the air (convection) and by cooling the sumption, we can express the specific cooling power
surfaces of the room envelope. The two effects can be (per square meter) of a cooled ceiling by the following
described by: equation:
qtot = 8.92(ta~r--t~um~) m
for convection, and
q~o~=sum of convective and radiant heat transfer (W!
m2).

for radiation, where

References
qc heat transfer by convection (W/m 2)
Olc convective coefficient (W/m 2 K) [1] A. Usibelli, S. Greenberg, M. Meal, A. Mitchell, A. Johnson,
tnir room air temperature (°C) G. Sweitzer, F. Rubinstein and D. Arasteh, Commercial-sector
conservation technologies, Rep. LBL-18543, Lawrence Berkeley
tsurface surface temperature (°C)
Laboratory, Berkeley, CA, 1985.
qr heat transfer by radiation (W/m 2) [2t H.E. Feustel, Economizer rating, Final Rep., prepared for
Stephan-Bolzmann constant (W/m 2 K4) Southern California Edison Company, 1989.
 H.E. Feuste~ C. Stetiu / Energy and Buildings 22 (1995) 193-205 205

[3] Heatin~ Ventilating~ and Air-Conditioning Systems and Appli- [20] A. Kollmar, Die zulaessige Kuehldeckentemperatur aus waer-
cations, ASHRAE Handbook, American Society of Heating, mephysiologischer Sicht, Gesundheits-lngenieur, 88 (5) (1967)
Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, 137-140.
GA, 1987. [21] A. Trogisch, Kuehldecke und Lueftung, Clima Commerce lnt
[4] P. Kroeling, Gesundheits- und Befindensstoerungen in klimatis- 24 (4) (1991) 50.
ierten Gebaeuden, Zuckschwerdt Verlag, Munich, Germany, [22] B. Glueck, Leistung von Kuehldecken, Kuehldecke und Raumluft,
1985. Fachinstitut Gebaeude-Klima, Stuttgart, Germany, 1990.
[5] P.O. Fanger, Strategies to avoid indoor climate complaints, in [23] Anon., Marktuebersicht Kuehldecken, KI Klima-Kaelte-Heizung,
Proc. Third Int. Congr. BuiMing Energy Management, ICBEM No. 1/2, 1993.
'87, Vol. I, Presses Polytechnique Romandes, Lausanne, Switz- [24] Anon., Radiant Metal Ceiling Panels - A Method of Testing
erland, 1987. Performance, Department of Veterans Affairs (date unknown).
[6] M. Mandell and A.H. Smith, Consistent pattern of elevated [25] Anon., SPC 138P, M O T for Rating Hydronic Radiant Ceiling
symptoms in air-conditioned office buildings: a reanalysis of Panels, Handout at the SPC138P Meet., 24 June 1991, Indi-
epidemiologic studies, Am. Z Public Health, 80 (10) (1990). anapolis.
[7] H. Esdorn, H. Knabl and R. Kuelpmann, Air-conditioning, new [26] FGK Vorschrift KD 1, Vorschtiften des Fachinstituts Gebaeude-
horizons - new opportunities, in Proc. Indoor Air '87, Berlin, Klima e.V. fuer: Waermetechnische Messungen an Kuehldecke-
Germany, 198Z nelementen, Fachinstitut Gebaeude-Klima e.V., Bissingen-Bie-
[8] E. Mayer, Thermische Behaglichkeit und Zugfreiheit, phy- tigtheim, 1992.
siologische und physikalische Erkenntnisse, in Proc. XXII [27] DIN 4715 Entwurf, Raumkuehlflaechen; Leistungsmessung bei
Int. Kongr. Technische Gebaeudeausruestung~ Berlin, Germany, freier Stroemung; Pruefregeln, Beuth Verlag GmbH, Berlin, 1993.
1988. [28] R. Kuelpmann, Thermal comfort and air quality in rooms with
[9] G.M. Keller, Energieaufwand fuer den Lufttransport mindern, cooled ceilings -- results from scientific investigations, ASHRAE
Clima Commerce Int., 21 (2) (1988). Trans., 99 (2) (1993) preprint.
[10] E. Skaret, Displacement ventilation, in Proc. Roomvent '87, [29] P.O. Fanger, B.M. Ipsen, G. Langkilde, B.W. Olesen, N.K.
Stockholm, Sweden, 198Z Christensen and S. Tanabe, Comfort limits for asymmetric
[11] H. Sutcliff, A guide to air change efficiency, Tech. Note AIVC thermal radiation, Energy Build., 8 (1985) 225-236.
TN 28, Air Infiltration and Ventilation Centre, Coventry, UK, [30] H. Esdorn and M. Ittner, Betriebsverhalten yon Deckenk-
1990. uehlsystemen, HLH Heizung-Luef~ung-Hanstechnik, 41 (1990)
[12] H.M. Mathisen, Analysis and evaluation of displacement ven- 598--601.
tilation, Rep. No. 1989:31, Division of Heating and Ventilation, [31] G. Zweifel, Simulation of displacement ventilation and radiative
NTH; Ph.D. Thesis, 1989. cooling with DOE-2, ASHRAE Trans., 99 (2) (1993) preprint.
[13] C.W.J. COx, P.J. Ham, J.M. Koppers and L.L.M. Van Schijndel, [32] F.C. Winkelmann, Model to simulate the performance of hy-
Displacement ventilation systems in office rooms -- a field dronic/radiant cooled ceilings, in Proc. 1993Annual Conf. CIEE,
study, in Proc. Room Vent '90, Oslo, Norway, 1990. San Diego, CA, 1993.
[14] M. Baker, Improved comfort through radiant heating and [33] K.-H. Fell, Wirtschaftliche Betrachtungen zu Kuehldecken in
cooling, ASHRAE Z, 2 (2) (1960) 54-57. Bueroraeumen, in Kuehldecke und Raumlueftung, Fachinstitut
[15] P.O. Fanger, Thermal Comfort Analysis and Applications in Gebaeude-Klima e.V., Bietigheim-Bissingen, Germany, 1991.
Environmental Engineering, McGraw Hill, New York, 1972. [34] W. Hoenmann and F. Nuessle, Kuehldecken verbessern Ra-
umklima, in Kuehldecke und Raumlueftung, Fachinstitut Ge-
[16] J.C. Sehlegel and P.E. McNall, The effect of asymmetric radiation
baeude-Klima e.V., Bietigheim-Bissingen, Germany, 1991.
on the thermal and comfort sensations of sedentary subjects, [35] B. Graeff, Kuehldecke und Raumklima, in Kuehldecke und
ASHRAE Trans., 74 (1968) 144-154. Raumlueftung, Fachinstitut Gebaeude-Klima e.V., Bietigheim-
[17] P.E. McNall and R.E. Biddison, Thermal and comfort sensations Bissingen, Germany, 1991.
of sedentary persons exposed to asymmetric radiant fields, [36] R. Meierhans and M. Zimmermann, Slab cooling and earth
ASHRAE Trans., 76 (1970) 123-136. coupling, Proc. Innovative Cooling Systems, International Energy
[18] E. Mayer, Auch die Turbulenzen sind wichtig, Clima Commerce Agency, Energy Conservation in Buildings and Community
Int., 19 (10) (1985) 20. Systems, Solihull, UK, 1992.
[19] E. Mayer, Air velocity and thermal comfort, in Proc. Indoor [37] Anon., Advanced Hydronic Heating and Cooling Technology,
Air "87, Berlin, Germany, 198Z Leaflet, eht-Siegmund, Inc., Tustin, CA (date unknown).