Energy and Buildings 37 (2005) 389–397

www.elsevier.com/locate/enbuild

Modeling radiant heating and cooling systems: integration with

a whole-building simulation program
R.K. Strand*, K.T. Baumgartner
School of Architecture, University of Illinois at Urbana-Champaign, Urbana-Champaign, IL 61820, USA
Received 12 May 2004; received in revised form 13 July 2004; accepted 30 July 2004

Abstract

One of the primary goals of the radiant heating and cooling community for the past two decades has been the fair and accurate comparison
of radiant space conditioning systems with conventional forced air systems. The desire of the community was to give architects and engineers
the ability to compare the energy consumption of different system types while requiring the systems to produce equivalent thermal
environments. Unfortunately, radiant systems, though relatively simple in concept, are relatively complex from a modeling standpoint. A
rigorous model must include fundamental laws of thermodynamics and properly account for heat transfer due to radiation, convection, and
conduction. In addition, the model must do this in a generic way within a program that is capable of modeling buildings of any size and type
and their associated HVAC systems. Such challenges are not easily overcome and require careful attention to detail as well as a solid
foundation from which to build. This paper summarizes a model which addresses these concerns and provides a radiant system model within a
popular building energy simulation program.
# 2004 Elsevier B.V. All rights reserved.

Keywords: Simulation programs; HVAC systems; Heating and cooling systems

1. Introduction program adequately addresses fundamental heat transfer

processes such as conduction through the solid materials in
Determining the energy efficiency of a building before it the building envelope, convection between the envelope and
is even built requires sophisticated science. While simula- the surrounding air, and radiation between the envelope and
tion has the potential to reduce and has been shown to be the inside and outside environments. None of these
successful at reducing energy consumption, energy costs, processes are simple linear functions, and most are
and the impact of buildings on the environment, the process dependent not only on the material properties of the
through which a simulation program has to go to simulate a envelope but also the inside and outside environmental
building is inherently complex. Though computer technol- conditions. While some programs are capable of handling
ogy certainly helps architects and engineers have better most heat transfer effects for general building types and
access to improved predictions of energy efficiency, the fact materials, there are also challenges to modeling the way a
remains that simulation programs are, by their nature, building is heated or cooled. Most modern buildings have
complex and detail oriented. some form of HVAC system which has the function of
This is evident in the processes that take place within the maintaining thermally comfortable conditions inside the
building envelope as well as other parts of the building. The building regardless of the conditions on the outside of the
building envelope transmits heat to and from the inside and building. One of the more common methods of maintaining
outside environments that are experienced by a building. thermal conditions within a building is through the use of
Modeling a building properly requires that a simulation conventional forced air systems. These systems aim to
deliver conditioned air to a space to maintain a particular air
* Corresponding author. temperature since air temperature is one of the main factors
E-mail address: [email protected] (R.K. Strand). that determines thermal comfort. While in the case of most

0378-7788/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2004.07.009
390 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397

conventional forced air systems many of the system mathematical background of the algorithm and controls of
interactions have been well-studied and can be modeled, two distinct radiant system types: flow controlled systems
there are a wide variety of different system types and and temperature controlled (constant flow) systems.
applications. To be able to evaluate how well and efficiently
different system types will condition a building requires a
program that is capable of modeling various components as 2. Modeling radiant systems
well as various reasonable configurations of those compo-
nents. As was briefly mentioned in the preceding section, a
Radiant heating and cooling systems pose additional program that will accurately model radiant systems and
problems to simulation programs. The idea of using compare them to other system types based on equivalent
radiation to heat and cool environments is not new and is thermal (comfort) environments needs to be able to analyze
becoming more commonplace in Asia and Europe and to a a building from fundamental heat transfer and thermo-
lesser extent in North America [8]. Radiant systems aim to dynamic perspectives, the HVAC system, and the inter-
provide conditioning directly to space occupants by radiant connections between the two. In addition, to be successful, it
exchange between ‘‘conditioned’’ surfaces within the space needs to be general enough to handle any building type,
rather than indirectly via transferring conditioned air to the shape, or size and do so in a relatively efficient manner.
space which then provides comfort to the occupants via The program [3] chosen for implementing the radiant
convection. system model has incorporated all of the features needed to
While the concept of a radiant system seems relatively successful simulate buildings and conventional systems. In
simple—directly transfer heat to or from people—there are addition, it has features that make it an ideal application for the
significant challenges to properly modeling a radiant system integration of a radiant system model. From a heat transfer
and incorporating such a model into a whole-building energy perspective, it uses a fundamental heat balance approach [7]
simulation program. A radiant system, by definition, is both that applies the First Law of Thermodynamics to three key
an HVAC system as well as a surface within the space. This interfaces: the inside surfaces for walls, roofs, floors, etc.; the
means that it is part of the calculation to determine how outside surfaces for walls, roofs, floors, etc.; and the air
much heating or cooling is needed as well as the system that contained within the various zones in the building. The inside
actually responds to those space conditions. Since some and outside heat balances are done for individual surfaces
radiant systems have thermal mass associated with them, within a building and incorporate fundamental approaches for
controls and transient behavior of the system are critical handling convection, long wavelength radiation, and short
components of a realistic model. Moreover, since radiant wavelength radiation and their impact on each surface.
systems condition in a significantly different way in Moreover, heat transfer through each surface is modeled using
comparison to standard conventional forced air systems, a solution to the transient conduction equation known as
the system model must be able to evaluate thermal comfort conduction transfer functions (CTFs).
in the space and also account for the possibility of con- Another key feature of the program is the integrated
densation when the system is used for cooling. This high nature of the simulation [13]. In the program, the impact of
level of integration must then also account for a generic the HVAC system on various spaces is calculated based on
simulation environment that is established to ensure the the real capacity of the system. This means that when a
ability to simulate buildings of many different shapes, system is not sized large enough to meet all of the heating or
orientations, sizes, etc. cooling needs for the spaces it is conditioning, the program
It is also important that the resulting simulation not be so will simulate the deviation from the normal control band as
‘‘radiant-centric’’ that it cannot model more conventional it would occur in a real building. This type of simulation
systems or combined load-sharing systems that are a hybrid algorithm is particularly important when dealing with a
combination of radiant and forced air systems. One of the radiant system simulation since the radiant system is both a
most significant barriers to the specification of radiant surface within the building and an HVAC system. Many
systems in buildings is the inability to fairly compare the more details about the program [4] and the initial
radiant and conventional systems to each other or to implementation of the radiant system model [10,11] and
combinations of the two. This is really the most pressing thermal comfort models [6] in the program are available in
need of the radiant community today—to be able to simulate the literature. The main focus of this paper is two-fold: the
various system types and make energy and/or cost interconnection between the building and HVAC portions of
comparisons based on reliable simulation data. the simulation and how this connection is affected by the
The remainder of this paper focuses on the incorporation type of controls used with the radiant system.
of radiant system models into a whole-building energy There are two main categories of radiant system controls
analysis program sponsored by the US Department of [1]: variable flow/fixed temperature systems and fixed flow/
Energy. It will begin with a brief description of the variable temperature systems. The variable flow/fixed
capabilities of the program that are important for the proper temperature system has a central loop and pump that
modeling of radiant systems. It will then delve into the supplies water at a particular temperature to all of the
 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397 391

systems connected to the loop. Valves control how much

water flows through each portion of the system based on
some criteria relating back to conditions within the space.
The fixed flow/variable temperature system has a central
loop and pump but also has local pumps associated with each
portion of the system. The local pumps maintain a constant
flow rate through that portion of the system while valves
control the amount of recirculation in the local loops. A
central pump maintains flow so that there is enough fluid to
inject to each local loop to maintain the proper temperatures
at each local loop. While some systems have several local
loops or perhaps use slight variations on these main control
strategies, most systems can be categorized into one of these
two classes.
Fig. 1. Variable flow low temperature radiant system controls.
Regardless of which type of system (variable temperature
or variable flow) is being simulated, both systems are
essentially a heat exchanger between the primary/secondary to the variable flow radiant system. Graphical descriptions of
equipment loop and the building elements (floors, walls, the controls for the low temperature radiant system model in
ceilings, etc.). The primary/secondary side of the model is the program are shown in Fig. 1 for a variable flow radiant
simply a fluid loop that receives conditioned water from a system.
boiler, chiller, heat pump, etc. and then sends unconditioned In the constant flow-variable temperature systems, the
water back to the primary equipment. The building element controls are also considered piecewise linear functions, but
portion of the model connects to the heat balance of the in this case the user selects both the control temperatures and
surfaces via a modification of the conduction transfer function the water temperatures via schedules. This offers greater
(CTF) formulation that accepts heat sources and sinks [9]. This flexibility for defining how the radiant system operates
link is critical because the two sides of the radiant system though it may not model every situation. Fig. 2 shows how
are simulated in very different areas of the program code. The the ‘‘desired’’ inlet water temperature is controlled based on
next two sections discuss the details behind this integration user schedules. The user has the ability to specify the high
and the mathematical derivation that was necessary to make and low water and control temperature schedules for heating
the model generic enough to both fit in the program structure and cooling separately (a total of eight temperature
and model as many systems as possible and to make the model schedules). Note that this inlet temperature is a ‘‘desired’’
efficient by avoiding unnecessary iterations that can cause inlet temperature in that there is no guarantee that the system
excessive program runtimes. will provide water to the system at that temperature. The
model includes a local loop that attempts to meet this
2.1. Simulation control overview demand temperature through mixing and recirculation.

The controls for variable flow low temperature radiant

systems within the simulation program are fairly simple
though there is some flexibility through the use of schedules.
The program user is allowed to define a setpoint temperature
as well as a throttling range through which the system varies
the flow rate of water (or current) to the system from zero to
the user defined maximum flow rate. The flow rate is varied
linearly with the flow reaching 50% of the maximum when
the controlling temperature reaches the setpoint tempera-
ture. Setpoint temperatures can be varied on an hourly basis
throughout the year if desired. The controlling temperature
can be the mean air temperature, the mean radiant tem-
perature, the operative temperature of the zone, the outdoor
dry-bulb temperature, or the outdoor wet-bulb temperature,
and this choice is also left to the user’s discretion. Since
flow rate is varied, there is no explicit control on the inlet
water temperature or mixing to achieve some inlet water
temperature in the variable flow system. However, the user
does have the ability to specify on an hourly basis through a
schedule the temperature of the water that would be supplied Fig. 2. Variable temperature low temperature radiant system controls.
392 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397

tries to get as close as physically possible given the loop

conditions).
Note that Fig. 3 shows that multiple surfaces can be
attached to a particular local loop. This option is available
for both the variable flow and variable temperature hydronic
radiant systems in the program. This is intended to account
for potential situations where a local loop serving a
particular zone has multiple active surfaces within that
zone. For example, both the floor and ceiling of a space may
serve as part of the radiant system, several different ceiling
Fig. 3. Variable temperature low temperature radiant system component
details. surfaces may also be linked together, etc.

2.2. Modeling the connection between the heat balance

The constant flow (variable temperature) low temperature and HVAC system
radiant system model is actually a combination of mixing
valves, a pump (constant speed, but the maximum flow can As was mentioned earlier, the radiant system itself is a
be modified by a schedule), and the radiant system (surface, heat exchanger with a ‘‘stationary’’ fluid (i.e., the building
panel, or group of surfaces/panels). This is connected to the element) assumed to be at a single temperature on one side
main loop through the standard inlet connections as shown and a water loop on the other. The stationary fluid side
in Fig. 3. The system controls determine the desired inlet relates to the heat balance portion of the program and must
temperature and system flow rate while loop controls be integrated with the surface heat balances. The water loop
determine the flow rate and temperature of the loop. Note side is part of the primary plant simulation. A heat exchanger
that pump heat also factors into the model through a simple formulation merges the two sides as described below.
constant speed pump model and user input. As shown by [9], the standard conduction transfer
Control of the variable flow system is considerably easier function formulation of heat conduction through a surface
because it simply accepts whatever temperature water is can be adapted to include the effect of an internal heat
available and varies the flow rate based on the controls source/sink. This leads to in the following equation which is
defined by Fig. 1. There is also no recirculation to account used for the inside surface heat balance:
for nor is there a local pump that adds heat to the loop.
Control of the variable temperature systems is more X
M X
M
q00i;t ¼ Xm Ti;tmþ1  Ym To;tmþ1
challenging because of the desire to control temperature. m¼1 m¼1
There are four possible conditions (separate for heating
Xk X
M
and cooling) for the variable temperature radiant system. þ Fm q00i;tm þ Wm qsource;tmþ1 (1)
First, if the loop has adequate temperature and flow to meet m¼1 m¼1
system requests, then the model sets the radiant system inlet
temperature and controls to the desired values based on the A similar equation can be established for the outside surface
controls and simulates. This is the best condition and heat balance:
recirculation and bypass amounts are adjusted accordingly X
M X
M
based on radiant system outlet temperatures. Second, if the q00o;t ¼ Ym Ti;tmþ1  Zm To;tmþ1
loop temperature is adequate but the loop flow rate is less m¼1 m¼1
than the radiant system flow rate, we may or may not be able X
k X
M
to meet the desired inlet temperature since recirculation þ Fm q00o;tm þ Vm qsource;tmþ1 (2)
might lower the temperature below the desired temperature. m¼1 m¼1
In this second case, the model first simulates the radiant Additionally, we can characterize the temperature at the
system with the desired conditions and then resimulates it to source/sink location with a similar equation:
solve for the actual inlet temperature (see later in this
section) if it cannot achieve the desired inlet temperature. X
M X
M
Third, if the loop flow is greater than the radiant flow but the Ts;t ¼ xm Ti;tmþ1  ym To;tmþ1
temperature of the loop is not adequate, then there is no m¼1 m¼1

amount of mixing that will solve this problem. All of the X

k X
M

radiant flow comes from the loop and the loop temperature þ fm Ts;tm þ wm qsource;tmþ1 (3)
m¼1 m¼1
(after pump heat addition) becomes the radiant system inlet
regardless of the temperature controls. Finally, if both the In this series of equations, X, x, Y, y, Z, z, F, and f are
temperature and the flow of the loop are inadequate, then the conduction transfer function (CTF) coefficients that are
model simply solves for the actual radiant system inlet calculated once at the beginning of a simulation for each
temperature and does not try to meet the controls (merely unique surface construction. The V, W, and w terms are heat
 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397 393

source transfer function (QTF) coefficients that allow the results in:
evaluation of the response of a building element to an
q ¼ ðṁcp Þwater ðTwi  Two Þ (4)
embedded heat source or sink. Coefficients, temperatures,
and fluxes at time ‘‘t’’ are for the current moment in time where q is the energy transferred between the water
while terms at other times are history terms—values loop and the building element, ṁ the mass flow rate of
from past time steps which are known as a result of the the water, cp the specific heat of the water, Twi the
simulation. inlet water temperature, and Two the outlet water tempera-
Eqs. (1)–(3) serve as a starting point for the actual heat ture.
balance solution. These equations alone, however, are not The maximum amount of heat transfer that is possible
enough information to solve for what is happening in the according to the Second Law of Thermodynamics is:
radiant system at a given point in time. The heat source/
qmax ¼ ðṁcp Þwater ðTwi  Ts Þ (5)
sink term is not known directly for a radiant system since
this relates back to the inlet water temperature, the water where qmax is the maximum amount of energy transfer that is
flow rate, and the temperature of the radiant surface. The possible and Ts the temperature at the source location.
actual heat transferred between the building element and The effectiveness of the heat exchanger, e, is defined as
the water loop is related to the temperature of the building the ratio of the actual energy transfer to the maximum
element at the source location as well as the water system possible, or:
inlet and outlet temperatures. In the model, it is assumed q
that the inlet temperature to the slab (defined by a user e (6)
qmax
schedule and/or controls and the primary equipment
simulation) and the mass flow rate (determined by the For a heat exchanger where one fluid is stationary, the
control algorithm and/or user input) are known and that the effectiveness can be related to NTU, the number of transfer
remaining parameters must be calculated. However, the units, by the following equation [5]:
heat balance equations require the heat transferred to the e ¼ 1  eNTU (7)
building element from the water loop in order to calculate
the heat transferred from the element to the building where NTU is defined by:
environment. UA
Even though systems defined by this model can vary NTU  (8)
ðṁcp Þwater
somewhat, the same characteristic link between the system
variables exist. For modeling purposes, the overall water/ Since the water tubes were assumed to have no effect on the
slab system can be thought of as a heat exchanger. While heat transfer process, the only term present in the overall
in principle there are two alternative heat exchanger heat transfer coefficient, UA, is a convection term. Thus, the
methodologies, it is more convenient to use the effective- equation for UA is:
ness-NTU method in this case. Several assumptions will be UA ¼ hðpDLÞ (9)
incorporated into the heat exchanger analysis. It is assumed
that the building element that contains the water loop is where h is the convection coefficient, D the interior tube
stationary and that its temperature along the length of the diameter, and L the total length of the tube.
tubing is constant. The latter part of this assumption stems The convection coefficient can be obtained from internal
from assumptions made in both the one- and two- flow correlations that relate the Nusselt dimensionless
dimensional heat source transfer function derivations. In number to other flow properties. For laminar flow in a tube of
either case, the source was added at a single node that constant surface temperature, the Nusselt number is defined
was characterized by a single temperature. For consistency, by:
this assumption must be made again in the heat exchanger
analysis. Another assumption for the model is that the hD
NuD ¼ (10)
fluid in the tubing is water. Additionally, it is assumed k
that the thermal properties of the water do not vary
where k is the thermal conductivity of the water.
significantly over the length of the tubing. Finally, the
For turbulent internal flow, the Colburn equation can be
temperature at the inside surface of the water tubing is
used to define the Nusselt number:
assumed to be equal to the temperature at the source
location. In other words, it is assumed that the water tubing hD 4=5
NuD ¼ ¼ 0:023 ReD Pr 1=3 (11)
itself has no appreciable effect on the heat transfer process k
being modeled. where Pr is the Prandtl number of water and ReD the
Using these assumptions and the effectiveness-NTU heat Reynolds number which is defined by:
exchanger algorithm, several equations can be defined which
establish the relationship between the heat source and the 4ṁ
ReD ¼ (12)
water temperatures. First, a heat balance on the water loop pmD
394 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397

The parameter m is the absolute viscosity of water. For One can summarize the situation by hiding some of the
internal pipe flow, the flow is assumed to be turbulent for complexity for the purpose of deriving equations to solve the
ReD  2300. system of equations at hand. The three basic equations for
Knowledge of the flow conditions allows Eqs. (6)–(12) to the temperature at the inside and outside surface as well as at
be calculated. This fixes the value for e which can then be the location of the heat source/sink can be reduced to:
used to relate Eqs. (4) and (5) to each other. The controls and
Tinside ¼ Ca þ Cb Toutside þ Cc q00 (14)
the plant define the water mass flow rate and the inlet water
00
temperature, leaving two equations (Eqs. (4) and (5)) and Toutside ¼ Cd þ Ce Tinside þ Cf q (15)
three unknowns. The third equation that can be used in 00
Tsource ¼ Cg þ Ch q þ Ci Tinside þ Cj Toutside (16)
conjunction with Eqs. (4) and (5) is Eq. (3), which is the CTF/
QTF equation for the temperature at the source location. where Tinside is the current temperature at the inside surface,
Unfortunately, it has two additional unknowns—the inside Toutside the current temperature at the outside surface, Tsource
and outside surface temperatures. This requires the addition the current temperature within the radiant system at the
of modified forms of Eqs. (1) and (2) which adds two more location of the source/sink, q00 the heat source/sink from the
equations without the introduction of any further unknowns. radiant system water loop, Ca is all of the other terms in the
Eq. (1) is the standard conduction transfer function inside heat balance (solar, LW exchange, conduction history
formula for the inside surface of a building element with an terms, etc.), Cb the current cross CTF term, Cc the QTF
embedded source/sink of heat. In the actual simulation inside term for the current heat source/sink, Cd is all of the
program, the surface flux on the left hand side of the other terms in the outside heat balance (solar, LW exchange,
equation is replaced with a surface heat balance: conduction history terms, etc.), Ce the current cross CTF
term, Cf the QTF outside term for the current heat source/
X
M X
M
sink, Cg the summation of all temperature and source history
surface heat balance ¼ Xm Ti;tmþ1  Ym To;tmþ1
terms at the source/sink location, Ch the QTF term at the
m¼1 m¼1
source/sink location for the current heat source/sink, Ci the
Xk X
M
þ Fm q00i;tm þ Wm qsource;tmþ1 CTF inside term for the current inside surface temperature,
m¼1 m¼1 Cj the CTF outside term for the current outside surface
(13) temperature.
Eqs. (14) and (15) above can each be solved to remove the
The surface heat balance includes terms for incident solar other surface temperature. Then, substituting these new
energy, radiation heat transfer from internal heat sources equations for Tinside and Toutside as a function of C and q00 into
such as lights and electrical equipment, radiation between the equation for Tsource and simplifying, one obtains the
surfaces using Hottel’s Gray Interchange concept, and con- following equation:
vection to the surrounding air. A similar equation can be
Tsource ¼ Ck þ C1 q00 (17)
derived from Eq. (2) for the outside surface heat balance and
includes thermal forces such as solar radiation, convection to
the outside air, long wavelength radiant exchange with the Ci ðCa þ Cb Cd Þ þ Cj ðCd þ Ce Ca Þ
Ck ¼ Cg þ (18)
external environment, etc. It should be noted that the internal 1  Ce Cb
radiation heat balance is dependent on conditions at the
other surfaces within the space. Thus, an iterative loop is
Ci ðCc þ Cb Cf Þ þ Cj ðCf þ Ce Cc Þ
required to provide a more accurate estimate of the radiative Cl ¼ Ch þ (19)
exchange within the building interior. Thus, the terms of the 1  Ce Cb
heat balance on the left hand side of the equation have been Combining this with the heat exchanger analysis as shown
set with the only unknown quantity being Ti, the inside above, we eventually arrive at the following equation to
surface temperature at the current time step. On the right relate the flux to the slab to the water inlet temperature and
hand side of Eq. (13), most of the terms are already defined mass flow rate:
since they depend on known values from previous time steps Twater;in  Ck
(temperature, flux, and source histories). The only terms q00 ¼ (20)
ðCl =AÞ þ ð1=eðṁcp Þwater Þ
which are not defined are the inside surface temperature (Ti),
outside surface temperature (To), and internal heat source/ which includes all of the inside and outside heat balance
sink (qsource) of the current time step. Similar arguments can terms (‘‘hidden’’ in the Ck and Cl coefficients). Once the flux
be made for reformulation of the outside heat balance to the slab is known, the remaining terms of interest (outlet
equation and the base form of the heat source/sink location water temperature, inside and outside surface temperatures,
temperature equation (Eq. (3))—the only undefined terms etc.) can be calculated using the relatively simpler equations
are the inside surface temperature (Ti), outside surface shown above.
temperature (To), and internal heat source/sink (qsource) of The above development is valid for both the variable flow
the current time step. radiant system and for variable temperature radiant systems
 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397 395

where the inlet temperature is known (based on controls). based on flow fraction and results in Twater,out as shown in:
However, due to loop conditions and the presence of X
recirculation, it is not always possible to know the inlet Twater;out ¼ ½ðFlowFractionj ÞTwater;out;j (26)
j
temperature to the variable temperature radiant system
without actually simulating it. For cases where the main loop An energy balance on the mixing valve-pump group results
temperature and flow rates are adequate, one can assume an in the following equation that relates the radiant system inlet
inlet temperature and then determine the recirculation temperature (Twater,in) to the loop inlet water temperature
needed after the fact. Yet, when the main loop flow or (Tloop,in), the radiant system outlet temperature (Twater,out),
temperature is not adequate to meet the needs of the local and the pump heat addition
loop as determined by the system controls, one is faced with
a ‘‘chicken-and-egg’’ situation. One needs to know the inlet ṁloop ṁrecirc
Twater;in ¼ Tloop;in þ Twater;out
temperature in order to solve for the system output, but one ṁsystem ṁsystem
needs the system output in order to solve for the inlet PumpHeat
temperature. As a result, the model must either iterate using þ (27)
ṁsystem cp
a guess for inlet temperature as a starting point or perform
more mathematics to arrive at the inlet temperature as a Plugging in the definition of Twater,out based on the summa-
function of system characteristics. This implementation tion of Twater,out,j equations results in:
chose to avoid iteration and solved for the inlet temperature LoopTerm þ RecircTerm
as shown below. Twater;in ¼ (28)
TwiCoeff
In addition to resolving the iteration concerns, one must
also account for the fact that there could be several potential where
parts to a particular local loop as shown in Fig. 3. Each ṁloop PumpHeat
of these individual surfaces may have a different outlet LoopTerm ¼ Tloop;in þ (29)
ṁsystem ṁsystem cp
temperature. Thus, all of the surfaces that are part of an
individual loop must be solved at the same time to obtain ṁrecirc X
RecircTerm ¼ FlowFractionj Twater;out;j (30)
both the combined outlet temperature for the local loop ṁsystem j
and thus the inlet temperature based on the amount of  
recirculation needed. ṁrecirc X
TwiCoeff ¼ 1  ½FlowFractionj ð1  Cm;j Þ
The previous equation combines with the following ṁsystem j
equation which is valid for an individual surface in the
(31)
current local loop of a radiant system:
Once the actual water inlet temperature is calculated with
qj ¼ ṁj cp ðTwater  Twater;out;j Þ (21)
this equation, it is then possible to calculate individual outlet
where qj is the heat transfer to the jth surface in the radiant temperatures for each surface, the overall outlet tempera-
system, ṁj the mass flow rate only to this surface, and ture, and finally all of the necessary flow and loop quantities.
Twater,out,j the outlet temperature for the jth surface. Combin- This procedure avoids iteration but is somewhat complex to
ing the previous two equations results in: follow. However, this second mathematical process is only
needed for select cases of the variable temperature radiant
ej ðTwater;in  Ck;j Þ
Twater;in  Twater;out;j ¼ (22) system when the inlet temperature is not known explicitly.
1 þ ððej ṁj cp Cl;j Þ=AÞ With the proper establishment of input data, such cases can
where Ck,j and Cl,j are evaluated for each individual surface be avoided, though one cannot simply assume this for a
as shown in Eqs. (18) and (19). If for each surface in the comprehensive energy simulation program.
radiant system, we let: The proceeding development shows several important
ej points. First, it is possible to model the connection between
Cm;j ¼ (23) the heat balance and the HVAC portions of the simulation in
1 þ ððej ṁj cp Cl;j Þ=AÞ
a fundamental, generic way. Second, it is possible to handle
then the previous equations becomes the slightly less com- this connection without resorting to excessive iterations
plex: despite the intimate link between the two sides of the radiant
system. In the simulation program, iteration does take place,
Twater;in  Twater;out;j ¼ Cm;j ðTwater;in  Ck;j Þ (24)
but the radiant system model as described here does not
Rearranging to obtain the outlet temperature for the jth result in additional iteration loops that could lead to long
surface: execution times. Finally, the implementation works within
the current structure of the program—meaning that it does
Twater;out;j ¼ ð1  Cm;j ÞTwater;in þ Cm;j Ck;j (25)
not require special rules for this particular system and that it
The overall outlet temperature from the radiant system is just automatically has access to any improvement made to either
a simple mixing of all of the surface outlet temperatures the heat balance or HVAC portion of the simulation.
396 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397

2.3. Avoiding undesirable operating conditions incorporates all of the algorithm, control, modeling,
and other features described in the preceding sections.
Both the variable flow and variable temperature systems While this is a significant advance toward the ultimate
have checks within their algorithms to avoid operating the goal of modeling all features of radiant systems and having
radiant system under what might be considered undesirable the ability to compare them fairly and accurately to
conditions. One check is aimed at avoiding heat additions by conventional forced air systems, there are still features that
the system when cooling is required and heat removal by the needed to be added and are the focus of upcoming research
system when heating is required. While this likely will never work on the program. The two issues that will likely be
happen in most systems, such conditions might occur during addressed next are the incorporation of a more sophisti-
transitional seasons or changes in controls from heating to cated condensation limit for variable temperature radiant
cooling or cooling to heating. Thus, the system compares the systems and the modeling of ‘‘load-sharing’’ hybrid
water inlet temperature with the current temperature of the systems. Currently, the condensation check is a fairly
radiant system where the source/sink term will be applied. blunt strategy that shuts the system down when it perceives
If sending water to the radiant system will produce the that condensation may occur. This is based on a comparison
opposite of what is desired from a heat addition/removal between the dew-point temperature and the predicted
standpoint, the algorithm will interrupt flow to the radiant surface temperature if the radiant system is running. There
system. is a 1 8C tolerance that is presently used within the
Another check is related to the potential for condensation program. This was chosen so that the system will shut off
in a radiant cooling system. If moisture levels within the when the surface temperature would be close but still
space being cooled are high enough either due to indoor or higher than the dew-point temperature. This limit may be
outdoor conditions (or both), the surface temperature of the somewhat strict in that some systems that have the radiant
radiant system could drop below the dew-point temperature system connected to an evaporative condenser may
of the space. If this occurs, condensation will take place. actually trigger the condensation limit. Further study is
This can be a potential problem in radiant cooling systems, needed here to determine if this tolerance is, in fact,
especially in many of the climates found in North America. appropriate and whether this needs to be a parameter that
Currently, the program compares the dew-point temperature the user should be allowed to change. In addition, the
of the space with the predicted surface temperature of the variable temperature systems do not need to interrupt flow
radiant system based on what the normal controls dictate when condensation is predicted to occur. Since they have
should be happening within the radiant system. If the pre- the ability to increase the amount of recirculation, it would
dicted surface temperature drops below the dew-point be more appropriate to increase the water inlet temperature
temperature plus a small temperature tolerance, the first in an attempt to provide some amount of cooling
algorithm will stop flow to the radiant system. without causing condensation.
Concerns about condensation often result in the
2.4. Connecting to primary systems specification of a ‘‘load-sharing’’ or hybrid system. In such
a system, a radiant system is combined with a conventional
One of the advantages to having access to the HVAC forced air system to provide the conditioning needs of the
portion of the program through the primary plant water loop space. Generally, the radiant system provides most or all of
is that it allows the ability to try out different equipment and the sensible cooling while the air system handles the latent
various different ideas. One idea that has already been cooling. While it currently is possible to model a space being
addressed in the literature [12] is the possibility of hooking served by both a radiant and a conventional forced air system
the radiant system up to a water loop that is served only by an within the program, their interaction is serial rather than
evaporative condenser. Recent additions of ground and pond parallel. In the current structure of the program, one system
loops make it possible to investigate the energy efficiency of will run to capacity in an effort to meet the current space
using temperate water in a radiant system and determine conditioning needs of the zone. Then, additional systems
whether or not such a system can produce thermally will attempt to meet any remaining load by operating in
comfortable conditions inside the building. The beauty of series. True ‘‘load-sharing’’ systems are controlled differ-
such a simulation model is such alternative systems can be ently and thus this type of system is the subject of a current
tested for a variety of conditions and environments—even research project.
ones that would not be thought of as climates where such Despite the continuing work on the radiant model within
technology would be successful. the simulation program, the current model has many
capabilities and has been the subject of recent validations
studies [2]. Many current systems can already be modeled
3. Concluding remarks and compared to conventional forced air systems, and the
comparisons can be made on the more accurate basis of
The model described in this paper is already available to equivalent thermal comfort rather than simply equal air
the general public and has been available for some time. It temperatures.
 R.K. Strand, K.T. Baumgartner / Energy and Buildings 37 (2005) 389–397 397

Acknowledgment [6] J. Lee, R.K. Strand, An analysis of the effect of the building envelope
on thermal comfort using the EnergyPlus program, in: Proceedings of
the 2001 ACS A (Association of Collegiate Schools of Architecture)
The authors wish to thank the US Department of Energy’s Technology Conference, Austin, TX, July 2001, ACSA, Washington,
National Energy Technology Laboratory (NETL) for DC, 2001.
funding the work under Grant DE-FC26-01NT41250 which [7] C.O. Pedersen, R.J. Liesen, R.K. Strand, D.E. Fisher, L. Dong, P.G.
led to this paper. Ellis, Loads toolkitRP-987, ASHRAE, Atlanta, GA, 2001.
[8] K.W. Roth, D. Westphalen, J. Dieckmann, S.D. Hamilton, W. Goet-
zler, Energy consumption characteristics of commercial building
HVAC systems, vol. III: Energy savings potential, TIAX LLC Report
References for US Department of Energy Building Technologies Program,
Washington, DC, 2002.
[1] Invensys, An introduction to invensys radiant ceiling technologies, [9] R.K. Strand, Heat source transfer functions and their application to low
July 2002. temperature radiant heating systems, Ph.D. Dissertation, Department
[2] C. Chantrasrisalai, V. Ghatti, D.E. Fisher, D.G. Schaetzle, Experi- of Mechanical and Industrial Engineering, University of Illinois at
mental validation of the EnergyPlus low-temperature radiant simula- Urbana-Champaign, 1995.
tion, ASHRAE Transactions: Part 2 109 (2003), ASHRAE, Atlanta, [10] R.K. Strand, C.O. Pedersen, Implementation of a radiant heating and
Georgia. cooling model into an integrated building energy analysis program,
[3] D.B. Crawley, L.K. Lawrie, F.C. Winkelmann, W.F. Buhl, A.E. Erdem, ASHRAE Transactions: Part 1 103 (1997).
C.O. Pedersen, R.J. Liesen, D.E. Fisher, The next-generation in [11] R.K. Strand, C.O. Pedersen, Modeling radiant systems in an integrated
building energy simulation—a glimpse of the future, in: Proceedings heat balance based energy simulation program, ASHRAE Transac-
of the Building Simulation’97, vol. II, Prague, Czech Republic, tions: Part 2 108 (2002).
IBPSA, September 1997, pp. 395–402. [12] R.K. Strand, Investigation of a condenser-linked radiant cooling
[4] Crawley, B. Drury, L.K. Lawrie, F.C. Winkelmann, W.F. Buhl, C.O. system using a heat balance based energy simulation program, ASH-
Pedersen, R.K. Strand, R.J. Liesen, D.E. Fisher, M.J. Witte, R.H. RAE Transactions: Part 2 109 (2003).
Henninger, J. Glazer, D. Shirey, EnergyPlus: new, capable, and linked, [13] R.D. Taylor, C.O. Pedersen, D. Fisher, R. Liesen, L. Lawrie, Impact of
in: Proceedings of the 2002 International Green Building Conference, simultaneous simulation of building and mechanical systems in heat
November 2002, Austin, Texas, US Green Building Council, Washing- balance based energy analysis programs on system response and
ton, DC, 2002. control, in: Proceedings of the Building Simulation’91 Conference,
[5] F.P. Incropera, D.P. DeWitt, Introduction to Heat Transfer, Wiley, New Nice, France, International Building Performance Simulation Asso-
York, NY, 1985. ciation (IBPSA), 1991.