BUILD SIMUL (2010) 3: 273 – 280

DOI 10.1007/s12273-013-0104-0

Modelling and performance prediction of an integrated central

cooling plant for HVAC energy efficiency improvement

Research Article
Vahid Vakiloroaya1 (), Jafar Madadnia1, Bijan Samali2

1. Center for Built Infrastructure Research, School of Electrical, Mechanical and Mechatronic Systems, University of Technology Sydney, PO
Box 123, Broadway, Sydney, NSW, 2007, Australia
2. Center for Built Infrastructure Research, School of Civil and Environmental Engineering, University of Technology Sydney, PO Box 123,
Broadway, Sydney, NSW, 2007, Australia

Abstract Keywords
The aim of this paper is to explore methods to reduce the energy consumption of a central cooling central cooling plant,
plant. To achieve this, we have developed a new design for heating, ventilating and air conditioning design improvement,
(HVAC) efficiency improvement. A storage tank together with an immersed heat exchanger is energy saving,
installed in the discharge line between the compressor and condenser. The heat exchanger uses the modelling,
make-up water of the cooling tower to reduce the refrigerant temperature entering the condenser. performance prediction
To investigate the potential of energy savings, we used a real-world commercial building with a
central cooling plant, located in a hot and dry climate, for our case study. The energy consumption
Article History
Received: 4 October 2012
and relevant data of the existing central cooling plant were acquired over the course of a typical
Revised: 2 December 2012
week in summer. The integrated system has been modeled and analyzed to achieve energy
Accepted: 18 December 2012
conservation. The performance of the proposed cooling system was simulated using a transient
simulation software package. Comparison of the proposed system with existing cooling plant is
© Tsinghua University Press and
included in this paper to demonstrate the advantages of our new configuration. Results show that
Springer-Verlag Berlin Heidelberg
up to 18% power saving can be obtained by using our design.
2013

1 Introduction are either expensive or very complicated to implement

and require constant monitoring. While optimizing the
Heating, ventilating and air conditioning (HVAC) systems mechanical design of the traditional HVAC system results in
consume significant amounts of electricity. The building extra upfront costs, these modifications can actually provide
sector represents 20%–40% of final energy consumption substantial savings in the long term by reducing ongoing

Building Systems and

in developed countries (IEA 2008). As reported by U.S.
Department of Energy (DoE), 31.2% of energy use by U.S.
residential buildings is by HVAC equipment (DoE 2001).
In addition, it has been accounted by DoE that energy use in
the built environment will grow by 34% in the next 20 years
at an average rate of 1.5% (DoE 2006). Therefore, finding
novel ways to reduce energy consumption in buildings
without compromising comfort and indoor air quality is an
ongoing research challenge. In recent years, different control
and optimization strategies have been used to improve the
energy consumption rates of air conditioning systems (Beghi
and Cecchinato 2011; Kusiak et al. 2011; Mossolly 2009; Lin
and Yeh 2007; Huang et al. 2009); however, these approaches
E-mail: vahid.vakiloroaya@engineer.com
maintenance costs associated with control and optimization
Components
strategies.
One proven way of achieving energy efficiency in HVAC
systems is to design systems that use novel configurations
of existing system components (Schoenfeld 2008). Recent
research has demonstrated that a combination of existing
air conditioning technologies can offer effective solutions
for energy conservation and thermal comfort. A 6.5 kW
prototype CO2 heat pump unit was constructed and equipped
with a counter-flow tripartite gas cooler in order to reject heat
by cooling of CO2 at supercritical pressure (Stene 2005).
The gas cooler was also used for preheating of the domestic
hot water. Their results showed that the integrated heat
2 Vakiloroaya et al. / Building Simulation

List of symbols

A cross-sectional area (m2) Subscripts

Cp heat capacity (kJ/(kg·℃)) a air
E power consumption (kW) ahu air-handling unit
f fraction amb ambient
h enthalpy (kJ/kg) cc cooling coil
l length (m) ch chiller
M mass (kg) comp compressor
m mass flow rate (kg/s) con condenser; condensing
N number of heat transfer unit ct cooling tower
N cell number of cooling tower cells operating db dry-bulb
p pressure (kPa) dsh de-superheating section
Q heat transfer capacity (kW) eva evaporator; evaporative
R capacity rate (kW/℃); capacity rate ratio exp expansion valve
(dimensionless) f fan
T temperature (℃) hx heat exchanger
t time (min) i inlet
U overall heat transfer coefficient (kW/(m2·℃)) m make-up water
V volume (m3) min minimum
VD displacement volume of the compressor (m3/s) o outlet
vsuc specific volume of the refrigerant at the opt optimum
compressor inlet (m3/kg) r refrigerant
vel velocity (m/s) sat saturated
Win electrical power consumption of the sc sub-cooled
compressor (kW) sh superheated
γ polytropic index of the refrigerant vapor st storage tank
ηcomp total efficiency of the compressor sup supply
ηv volumetric efficiency of the compressor tp two-phase
ε effectiveness w water
ρ density (kg/m3) wb wet-bulb

pump could work at higher coefficient of performance. Jeon
et al. (2010) analyzed the performance of a hybrid cooling
system that combines a screw water chiller with a ground
source heat pump. Their results showed that the hybrid
cooling system was able to stably provide the required cooling
capacity at high-load conditions. A hybrid air conditioning
plant obtained by combining an air dehumidification
system that uses a hygroscopic solution and hydrophobic
mamberanes with an inverse cycle vapor compression
device was analyzed by Bergero and Chiari (2011). Their
results showed that up to 60% power saving was possible
based on indoor latent load conditions. Deng et al. (2012)
evaluated the performance of the hybrid solar-assisted CO2
heat pump in which a solar thermal driven absorption chiller
was combined with the heat pump. In their system, a gas
cooler was located between the compressor of the heat pump
and evaporator of the chiller to decrease the temperature of
CO2. Their results demonstrated that the new proposed
system can meet the year-around cooling and heating
demand. However, no research has been done on whether
the performance of a vapor comparison central cooling plant
(CCP) is improved if the storage tank is installed after its
compressor to reduce the refrigerant temperature using the
make-up water of the cooling tower. This modified design
is the focus of our study.
This study provides a new design for the vapor com-
pression air conditioning systems to increase the refrigeration
system’s overall efficiency. Sub-cooling increases the system’s
efficiency by reducing compressor power and increasing
the system’s refrigeration effect. Simply stated, sub-cooling
lessens the amount of energy that is required to operate a
refrigeration system and, therefore, results in a cost saving
for the operator. There are various traditional methods to
achieve sub-cooling, all of which require additional power
Vakiloroaya et al. / Building Simulation
and thus result in additional system costs. In this study, the
make-up water of the cooling tower is used to reduce the
temperature of the superheated refrigerant gas before it
enters the condenser. In this system, the make-up water of
the cooling tower is first used to reduce the refrigerant
temperature entering the condenser: a horizontal cylindrical
storage tank, together with an immersed heat exchanger,
is used to reduce the temperature of refrigerant through
the energy transfer between the superheated refrigerant
vapor and make-up water of the cooling tower. The cooled
refrigerant then enters the condenser where an additional heat
reduction occurs. As a result, the refrigerant temperature
leaving the condenser decreases and the refrigerant subcool
temperature increases resulting in energy savings of 5%–18%.
An exact figure is dependent on the condensing and
evaporator temperatures of the individual’s refrigeration
system.
The objective of this project is to minimize the energy
consumption of building cooling systems while still satisfying
occupant comfort conditions. For this purpose, we used the
central cooling plant of a real-world commercial building
located in a hot and dry climate. Field tests were conducted
to quantify and determine the operational variables of the
central cooling plant. First, the mathematical model of the
system components and assumptions at which the study
was undertaken were developed. New modules for the
proposed system were developed by using collected
experimental data and were implemented in the transient
tool using a FORTRAN code. An iterative numerical
procedure was used to solve simultaneous heat transfer
equations for the various system components. The
experimental results were then compared with the
theoretical prediction from the mathematical model of
the system to ensure that it can be used for the intended
purpose. Because of the nonlinear, time-varying and
building-dependent dynamics of the system, the transient
software package, TRANSYS 16, was used to simulate and
predict the energy usage on an hourly and sub-hourly basis
for the building using the HVAC system. On the basis of
mathematical models and by using test data, simulation
modules for the central cooling plant were developed
and embedded for the building. The system integration
encompassed the water-cooled screw chiller, the counter flow
cooling tower, the air-handling unit and the storage tank
with immersed heat exchanger. The energy consumption of
the existing central cooling plant in the building was
measured in a typical week of summer to compare the
project data with the simulation results in order to analyze
the performance and feasibility of the proposed newly-
configured system. Results demonstrated that the benefits
of the new design lie in the system’s ability to operate at a
3
higher sub-cool temperature after the refrigerant passes
through the water-cooled condenser, significantly increasing
the overall system coefficient of performance.
2 System description
The schematic of the central cooling system with integrated
water storage tank is shown in Fig. 1 and its operation cycle
is shown in Fig. 2. As detailed above, the system comprises
a water-cooled screw chiller, an open circuit cooling tower,
an air-handling unit, chilled water and condenser water
pumps, water storage tank with its immersed heat exchanger,
valves and connection tubes. The cycle starts with a mixture
of liquid and vapor refrigerant entering the evaporator
(point 1). The heat from the water returning from the
air-handling unit is absorbed by the evaporator coil. During
this process, the state of the refrigerant changes from a
liquid to a gas and becomes superheated at the evaporator
exit. The superheated vapor enters the compressor (point 2),
where increasing pressure raises the refrigerant temperature.
A storage tank together with a serpentine tube heat exchanger
installed after the compressor uses the cooling tower’s
make-up water to cool down the refrigerant. The refrigerant
from the compressor goes through the copper coil inside
the tank and undertakes a heat exchange (point 3). In this
design, temperature of high pressure superheated refrigerant
gas can be decreased by exchanging heat with the make-up
water of the cooling tower. A storage tank is located in the
discharge line between the compressor and condenser for
this purpose. The incoming cold water is generally between
20℃ and 30℃ depending on the time of the year and the
geographical location. The make-up cold water then passes
through the storage tank and thus reduces the temperature
of refrigerant entering the condenser. This results in a more
substantial sub-cooling at the condenser outlet and therefore
reductions of refrigerant temperature entering the evaporator,
enhancing the system refrigeration effect and, in turn, its
coefficient of performance (COP). On the other hand, this
modification reduces the refrigerant temperature leaving
the condenser by pre-cooling the refrigerant before it
goes through the condenser for the final heat rejection. The
additional pre-cooling of the refrigerant then results in higher
sub-cooling temperature at the condenser outlet and thus
reduction of refrigerant temperature entering the evaporator.
The pre-cooling of the refrigerant means de-superheating the
refrigerant entering the condenser using make-up water to
reduce its temperature below the condensation point at the
corresponding vapor pressure after the condenser resulting in
higher sub-cooling temperature at the condenser outlet.
After leaving the storage tank, the refrigerant undergoes
a further temperature reduction in the condenser, causing
4 Vakiloroaya et al. / Building Simulation

Fig. 1 Schematic diagram of the newly-configured central cooling plant

it to be more de-superheated (point 4); thus, the refrigerant 3 Mathematical model

liquid is sub-cooled as it enters the expansion device (point 5).
Sub-cooling prevents flash gas formation before reaching the
expansion device and increases the evaporator refrigeration
effect. The high pressure sub-cooled refrigerant flows through
the expansion valve, which reduces both its pressure and
temperature.
Many models for central cooling plants have been developed
using various principles, see (Mortada et al. 2012; Tashtosh
et al. 2005; Lee and Lu 2010). To target the system’s energy
efficiency, our objective is to predict the system’s power
consumption while its thermodynamic transient performance
is analyzed using TRNSYS. A combined theoretical–
empirical modeling approach is developed in this section to
allow for performance prediction over a very wide range of
operating conditions.

3.1 Cooling tower

The performance model of the cooling tower can be

determined by combining the thermal and hydraulic
performance of the cooling tower. The optimum ratio of
cooling tower water flow rate to ambient dry airflow rate
can be determined as (Soylemez 2004)
5206.9
18.6 -
Tw
m w,ct Cp,a 3238.7hfg,w pamb e
( ) m a,ct opt
=
C p,w
+
é 18.6 -
Tw ù
5206.9 2
(1)
Fig. 2 A p–h diagram for the newly-configured central cooling Cp,wTw2 ê p - e ú
plant êë úû
Vakiloroaya et al. / Building Simulation 5

where hfg,w is the latent heat of the evaporation of the water
vapor and Tw is the mean temperature of the cooling tower
water. In a cooling tower, water enters the sump reservoir
from a water make-up source and operating tower cells.
The level of the sump is assumed to be constant in this
study because of its negligible effect on the cooling tower
performance. Therefore, the flow of the water make-up will
be equal to the total water loss from the cells. Furthermore,
the water volume of the sump is assumed to be fully-mixed in
all the cells so that the energy balance can be expressed as
dTsump é N cell ù
M w,ct = ê å( m w,o,ct (Tw,o,ct - Tsump ) )k ú
dt êë k =1 úû
é N cell
+ ê m w,i,ct - å (m w,o,ct )k ( Tm,o,st - Tsump ) ú (2)
( )
ê ú
ë k =1
3.3 Rotary compressor
Since the compressor can reach operating speed instant-
aneously, its transient characteristics can be neglected (Chen
and Deng 2006) and thus it can be modelled using algebraic
equations (Diaz 2007). Therefore, the compressor is modelled
with the use of an iterative procedure described by the
following equations for the polytropic work input to the
compressor, the refrigerant mass flow rate of the compressor
and the refrigerant enthalpy at the compressor exit. The
compressor mass flow rate is given by
VDηv
m r = (6)
vsuc
ù
The work input to the compressor is given by
û

where Tsump is the fully-mixed sump temperature and Tm,o,st γ -1

m r γ pcon ( γ )
is the temperature of water make-up to sump. For calculation Win = p v ( )(( )
ηcomp eva suc γ - 1 peva
-1 ) (7)
of the power consumption of the cooling tower variable air
volume (VAV) fan Ect,f, a quadratic equation is proposed as
Then the refrigerant enthalpy leaving the compressor can
Ect,f = a0 + a1Tw,o,ct + a2Tw,o,ct
2
+ a3Twb + a4Twb
2
+ a5Tw,o,ctTwb be calculated as
(3)
m r hr,i,comp + Win
where coefficients a0 to a5 are constant and can be determined hr,o,comp = (8)
m r
by curve-fitting the real data.
Finally, the refrigerant temperature leaving the com-
3.2 Evaporator pressor can be determined by using the compressor discharge
pressure and refrigerant enthalpy leaving the compressor
The evaporator in the plant is of the shell and tube type and and referring to the refrigerant thermodynamic properties
the heat exchange between the refrigerant and the water is in the superheated zone.
assumed to be counter-flow. The cooling duty required for the
evaporator process is calculated using the difference between 3.4 Integrated storage tank
inlet and outlet refrigerant enthalpy. If the evaporator inlet
and outlet refrigerant enthalpy respectively are hr,i,eva and The storage tank which contains an immersed serpentine tube
hr,o,eva, then the required cooling duty based on the principle heat exchanger is located between the compressor and the
of energy conservation is given by condenser. The make-up water supplied from the city water
enters the storage tank and interacts with the refrigerant in
Qeva = m r (hr,o,eva - hr,i,eva ) (4) the immersed coil, water stream which passes into and out of
the storage tank and with the environment. An energy balance
The enthalpy of refrigerant leaving the evaporator is
for the water storage tank yields
evaluated as
dTw,st
M w,st Cp,w = (m m Cp,w )(Tm,o,st - Tm,i,st ) + m r (hr,o,st
hr,o,eva = hr,o,sat + Cp,r (Tr,sh - Tr,o,sat ) (5) dt
- hr,o,comp ) + AstU st (Tw,st - Tdb ) (9)
where Tr,sh is the temperature of superheated vapor refri-
gerant leaving the evaporator which was measured during The energy balance principle for the immersed coil inside
the experimentation, hr,o,sat and Tr,o,sat are respectively the the storage tank gives
enthalpy and temperature of saturated refrigerant vapor at
AhxU hx
the evaporating pressure and can be calculated using the Tr,o,st = Tw,st + (Tr,i,st - Tw,st )exp -( m r Cp,r) (10)
refrigerant thermodynamic properties.
6 Vakiloroaya et al. / Building Simulation

3.5 Water-cooled condenser m r hfg,r

Ta,o,tp = Ta,i,tp + (18)
Ra
By the phase of the refrigerant the condenser can be divided
into three parts: de-superheating zone, two-phase (a liquid– where hfg,r is the enthalpy of the refrigerant at the two-phase
vapor) zone and sub-cooled liquid zone. We assume that zone. Finally in the sub-cooled liquid zone the temperature
the pressure drop along the condenser tube compared of the refrigerant decreases as it travels from the two-phase/
with total pressure head is negligible. Heat transfer in the liquid interface toward the condenser outlet. The sub-cooled
condenser is modeled in this section by using the NTU-ε fraction can be determined by
method. The Newton-Raphson method is used to solve the
equations in an iterative loop. f sc = 1 - f dsh - f tp (19)
The refrigerant entering the condenser after leaving the
water storage tank is either superheated or two-phase. In de- Therefore, the refrigerant temperature leaving the con-
superheating zone the refrigerant temperature decreases as it denser can now be calculated from
travels from the condenser inlet toward the vapor/two-phase
interface. The heat transfer in this zone can be evaluated form U sc f sc Acon
N sc = (20)
the following equations: Rmin

Rr,dsh (Tr,i,con - Tcon )

Rmin (Tr,i,con - Ta,i,con ) {
N 0.22
= 1 - exp dsh [ exp( -RN dsh
R
0.78
) -1 ] } εsc = 1 - exp { N sc0.22
R
[ exp( -RN sc0.78 ) - 1 ] } (21)

(11) εsc Rmin (Tcon - Ta,o,tp )

Tr,o,con = Tcon - (22)
Rr,sc
N dsh Rmin
f dsh = (12)
AconU dsh where Tr,o,con is the temperature of the refrigerant leaving
the condenser and influences significantly on the system
Qdsh = m r Cp,r (Tr,i,con - Tcon ) (13) overall coefficient of performance.

3.6 Expansion valve

Q
Ta,i,tp = Ta,i,con + dsh (14)
m a Cp,a For an isenthalpic process in the expansion valve, the
following equation is given:
where Qdsh is the heat transfer rate of the de-superheating
zone and Ta,i,tp is the air temperature entering the two-phase hr,i,exp = hr,o,exp (23)
zone and is equal to the air temperature leaving the de-
superheating zone. In the two-phase zone the refrigerant
where hr,i,exp and hr,o,exp are respectively the specific enthalpies
temperature is at its saturated value with no spatial variation.
of the refrigerant entering and leaving the expansion valve.
The following calculating procedure is applied in the two-
Therefore, the refrigerant temperature leaving the expansion
phase zone to determine the two-phase fraction as
valve can be determined after knowing the calculated
Ta,o,tp - Ta,i,tp refrigerant enthalpy and measured refrigerant pressure at
εtp = = 1 - exp(-N tp ) (15) the expansion valve exit.
Tcon - Ta,i,tp

N tp Ra 3.7 Air-handling unit

f tp = (16)
AconU tp
Using the energy and mass conservation laws, the dynamic
Therefore, the two-phase fraction can be obtained by change of the air and water temperature through the cooling
combining Eqs. (15) and (16) as coil is described as (Jin et al. 2006)

Ra Tcon - Ta,i,tp dTw,cc T - Tw,i,cc

f tp =
AconU tp
ln (
Tcon - Ta,o,tp ) (17) (
M w,cc Cp,w
dt
+ Cp,w (vel )w w,o,cc
l
= Qcc ) (24)

where Ta,o,tp is the air temperature leaving the two-phase zone dTa,cc Tsup - Ta,i,cc
and can be calculated from (
ρaVsup Cp,a
dt
+ Cp,a (vel )a
l
= -Qcc) (25)
Vakiloroaya et al. / Building Simulation
In the steady-state condition, Tw,cc and Ta,cc are constant
and thus by employing the method proposed in (Wang et
al. 2004), the cooling coil capacity can be obtained as
b0Cp,w m sup
Qcc = ( Tsup - Tw,o,cc ) (26)
m sup
1 + b1( )
m w,cc
where coefficients b0 and b1 can be determined from the
data set collected experimentally. The advantage of the heat
transfer model for the cooling coil expressed in Eq. (26) is
that in this model no geometric data of the cooling coil are
required. In order to estimate their values, a large number
of data points are required, often more than those given in
typical product catalogs and should be determined by real-
world test data.
Finally, a quadratic form for the air-handling unit VAV
fan is proposed to represent the total power consumption
in terms of the supply air temperature Tsup and chilled water
temperature Tw,o,ch:
Eahu,f = c0 + c1Tw,o,ch + c2Tw,o,ch
2
+ c3Tsup
+ c4Tsup
2
+ c5Tw,o,chTsup (27)
where coefficients c0 to c5 are constant and can be determined
by regression of the field-test data collected.
4 Case study
The simulation object is a commercial building together
with its central cooling plant. The gross floor area of the
building is 2500 square meters and the usable floor area is
1700 square meters. The building is three meters high. The
ground area to be occupied by the building is rectangular
in shape and stands in south-north direction. The building
is open from 8:00 a.m. to 10:00 p.m. The walls, windows,
floor and roof are modeled according to ASHRAE transfer
function approach. The occupant density in the building is
about one person per 5 m2 and the associated heat gain is
95 W and 90 W for sensible and latent loads, respectively.
The mechanical ventilation rate for acceptable indoor air
quality is two air changes per hour (ACH). The infiltration is
selected to be 1.5 ACH. Lighting power density is 20 W/m2,
a common value used in practice for a commercial building.
We assume the heat gain from the lights is 40% convective.
The equipment load inside the building is taken as 16 W/m2,
corresponding to a medium load type. The clothing factor
for people in the building is specified as 0.5 clo and relative
air velocity is fixed at 0.1 m/s. The metabolic rate is 1.2 met
and no external work is assumed. The weather data that drive
the project simulations are based on a typical meteorological
7
year. The data file distributed with TRNSYS 16 is generated
using “Meteonorm” including Synthetic hourly weather data.
The hourly meteorological data used in this study include
information about the month of the year, hour of the month,
direct normal solar radiation, global solar radiation, dry-bulb
temperature, humidity ratio, and wind speed.
The central cooling plant installed in the building consists
of one water-cooled chiller, one cooling tower, one air-
handling unit, two chilled water pumps, and two condenser
water pumps. The chiller has two screw compressors, each
with a nominal capacity of 175 kW, and uses refrigerant
R407C. For the shell and tube evaporator, the evaporator
temperature is set to be 4℃. The temperature of the supply
chilled water is taken at 7℃ at design conditions. The chiller
comprises two refrigeration circuits in parallel in which
each circuit includes one thermostatic expansion valve and
one screw compressor. The chiller can operate down to about
10% of its rated full load capacity via a modulating slide valve
in the compressor. The design airflow rate and the electric
power input of variable speed cooling tower fan at maximum
airflow rate are 10.4 kg/s and 1.5 kW, respectively. The
design airflow rate of the air-handling unit with variable air
volume fan is 13.5 kg/s and its rated power input is 12.8 kW.
The design water flow rate and electric power of each chilled
water pump is 11.4 kg/s and 1.7 kW, respectively. The design
water flow of the condenser water pump is 13.8 kg/s and its
electric power is 2.3 kW. All circulator pumps operate at a
constant speed. Photos of the screw chiller, air-handling
unit, and cooling tower are in the Electronic Supplementary
Materials (ESM1, ESM2 and ESM3) of this paper.
4.1 Experimental rig
In order to determine the variables of the cooling system, a
number of tests under different running conditions were
conducted. The test field procedure has been carried out
on-site. The cooling tower and air-handling unit are located
outside close to the building and the cooled air is carried
to the building using the ducting system. High precision
sensors/transducers were used for measuring all operating
variables. Manometers were used for obtaining supply
airflow rate readings. The temperature sensor for supply air
is of platinum resistance type with ±0.1℃ accuracy. Water
temperatures were measured by mercury thermometers
of ±0.2℃ precision. Five platinum resistance thermometers
of PT100 type with a calibrated accuracy of ±0.5℃ for a
temperature range from –50℃ to +260℃ are used to
measure the refrigerant temperature before and after of
each component. These sensors are in direct contact with
the refrigerant to achieve high accuracy.
Two pressure transmitters with a calibrated accuracy of
8 Vakiloroaya et al. / Building Simulation
1% are respectively installed in suction line and discharge
line. The measuring range for pressure transmitter located
in suction line is between zero to 2000 kPa and for pressure
transmitters located in discharge line is zero to 3500 kPa.
Powers of components were measured by a digital AC/DC
power clamp multimeter of ±3.5% tolerance.
All of temperature sensors, pressure transmitter, and
power clamp multimeter are integrated with a computer
based data-logger. Therefore the data of system power
consumption and other variables were logged for each
ten-minute period. Then the variables were stored and
arranged in database files so that the process performance
can be assessed using the proposed approach.
5 TRNSYS development
The mathematical model of the system components is
coded into the fully integrated visual interface known as
TRNSYS simulation studio using FORTRAN. Under real
conditions there are many factors that can influence the
HVAC system performance. In this study, all dimensions
of the system components and the dimension of connected
pipes are included in program code. The experimental
obtained data were entered through a dedicated visual
interface. The simulation information flow diagram is shown
in Fig. 3, where the icon “commercial building” involved the
building information file has been created by PREBID and
Fig. 3 Central cooling plant simulation information flow diagram in TRNSYS workspace
is compliant with the requirements of the ANSI/ASHRAE
Standard 140 (2007). The model included a subroutine to
evaluate the thermodynamic properties of the refrigerant
R407C. This simulation code was then used to calculate the
parameters of main interest such as the temperature of
refrigerant leaving the DX evaporator, compressor, water
storage tank, air-cooled condenser and expansion valve as
well as water temperature in the water storage tank. The idea
is to compare the behavior of the models with the observed
behavior of the real system.
The module of the screw chiller has been developed
to consider its different components separately. For this
purpose, the aforementioned mathematical model for each
part of the chiller, i.e., evaporator, compressor, condenser,
expansion valve, and water storage tank are coded and
implemented on the chiller module using FORTRAN. In
this module, a numerical algorithm is embedded to call the
determined parameters from the first component (evaporator)
and transfer them respectively to the next component in
each iteration. To begin the simulation, the program reads
the first set of operating data, such as the evaporative and
condensing pressure of the refrigerant and the temperature
of superheated refrigerant vapor leaving the evaporator.
The enthalpy and temperature of the refrigerant leaving the
evaporator are calculated by using the evaporator model. The
algorithm continues until all enthalpies and temperatures
of refrigerant are collected after it travels through each
Vakiloroaya et al. / Building Simulation 9

component of the system. In each iteration, the compressor
work, refrigerant sub-cool temperature and power con-
sumption of each component are calculated and stored in a
separate file. These data are then compared with the
conventional central cooling plant to measure comparative
system efficiency.
operated continuously from 8:00 a.m. to 10:00 p.m. Only 12%
of the total predicted data deviate from the actual data by
more than 10% and all of modeled data have a prediction
error of less than 15%, as shown in Fig. 4.

5.1 Model validation

According to the data provided by the field tests, the

corresponding coefficients of the models are obtained by
regression techniques which are done using the MINITAB
statistical software. The R-squared value of each model
indicates a good fit. The regression process yields the
coefficients according to variations of the overall energy
consumption. With standard SI constants used in
thermodynamics, these coefficients for the CCP described
in this study were obtained as Fig. 4 Comparison of the measured and simulated chiller power

a0 = 6.37 a1 = -0.47 a2 = 0.65 a3 = 0.53 6 Results and discussion

a4 = 0.37 a5 = -1.018
b0 = 1.93 b1 = 0.58 (28) Having described the specifications of the plant and coded
c0 = 6.03 c1 = -0.31 c2 = 0.006 c3 = -0.188 mathematical model, simulation results showing the electricity
c4 = 0.004 c5 = 0.001 saving potential are presented and discussed in this section.
TRNSYS is run to obtain a component-wise energy analysis
Then for the purpose of error analysis, the simulated throughout the summer. To predict the performance of
performance of the system is compared with obtained the system, the cooling load is calculated prior to the
experimental results. The simulation is run with a time simulation performed by using TRNSYS. The references
for indoor temperature and relative humidity during the
interval of 10 minutes, i.e., equal to the monitoring time step
cooling load calculation were set at 23℃ and 50%. The peak
in the real test process. This study uses the relative error
cooling load is estimated at 236 kW in the middle of July.
(RE), the root mean square error (RMSE) and the coefficient
In a vapor compression refrigeration cycle, the refrigerant
of variance (CV) to indicate how well a model fits the
temperature leaving the condenser has very high effect on
observations, defined as follows:
the system performance. In fact more degree of sub-cool
N
y pred,i - y data,i temperature is desirable as it increases the evaporator
åi =1 y data,i
refrigeration effect. The new configuration mentioned in this
RE = ´100% (29) study increases the sub-cool temperature of the refrigerant,
N
significantly increasing the system’s overall COP. This design
has a range of advantages over conventional systems. The
N
1 overall coefficient of performance for the system can be
RMSE =
N (å ( y
i =1
pred, i )
- y data,i )2 (30)
estimated as

RMSE
CV = ´100%
y data,m
where N is the number of collected data, ypred,i is the
predicted variable, ydata,i is the measured variable and ydata,m
is the average value of the measured variables.
The integrated simulation tool was validated by comparing
the predicted and measured power consumption of the
chiller for the first week of July, during which the chiller
refrigeration effect h2 - h1
COP = = (32)
(31) compressor work h3 - h2
From Eq. (32), it is clear of two possible modifications to
improve the system performance, namely, by increasing the
refrigeration effect or decreasing the compressor power. In
this study, as the refrigerant temperature in the condenser
outlet drops, the refrigerant temperature in the evaporator
inlet also drops and the enthalpy of the refrigerant entering
the evaporator is lowered. This reduction allows the
10 Vakiloroaya et al. / Building Simulation

refrigeration effect to increase significantly, thereby increasing

the evaporator capacity. Consequently, this approach can
increase the system COP. Furthermore, sub-cooling the
refrigerant leaving the condenser prevents flash gas forming
at the inlet of the expansion device.
The comparison study using simulation results indicated
the significant energy saving potential in the steady-state
condition as shown in Table 1. Here the enthalpy of the
refrigerant entering the condenser with and without the
new configuration is found to reduce from 455.02 kJ/kg
to 425.65 kJ/kg. However, the temperature of refrigerant
leaving the condenser is decreased from 32.72℃ to 29.65℃.
This temperature reduction yields to the lessening of the Fig. 6 Central cooling plant hourly power consumption
displacement volume of the compressor from 0.097 m3/s to
0.092 m3/s which drops the compressor work. At designed TRANSYS. It can be seen that the average power consumption
steady-state conditions, the compressor power consumption by using developed system is nearly 11.3% less than of the
for conventional cycle and developed system are respectively conventional system. It is noted that the water-cooled
83.2 kW and 78 kW which shows 6.25% energy savings. chiller and air-handling unit VAV fan power consumption
Furthermore, the enthalpy of the refrigerant entering the values for new system are less than those for conventional
evaporator by using developed system is nearly 11.94% less system while the cooling tower VAV fan power usage is
than that of the conventional system which means more higher than that for conventional system. The average energy
refrigeration effect for the evaporator in the developed savings potential of the proposed design for the chiller and
system. air-handling unit fan are respectively 9.2% and 4.1% while
Results show that both refrigerant temperatures entering the cooling tower fan power consumption increases by
and leaving the condenser will drop by using the new design 2.6%. The reason is that the reduction of the refrigerant
as shown in Fig. 5. The hourly power saving potential for the temperature entering the condenser yields reduction in
central cooling plant is shown in Fig. 6 as obtained from refrigerant temperature leaving the condenser which means
the refrigerant sub-cool temperature at the condenser outlet
Table 1 Comparison of the steady-state performance of the cycle is increased. Therefore, the refrigerant temperature entering
the evaporator is reduced which means increasing in the
p(kPa) T(℃) h(kJ/kg)
system refrigeration effect. A higher refrigeration effect causes
Conventional Developed Conventional Developed Conventional Developed
Point system system system system system system
more heat transfer in the evaporator which decreases the
supply chilled water temperature. Meanwhile, as the chilled
1 579 579 2 1.74 248.67 218.96
water temperature decreases, the heat transfer efficiency of
2 529 529 9 9 416.12 416.12
the air-handling unit cooling coil will be increased. Therefore,
3 1754 1754 69.19 69.19 455.02 455.02
less supply air will be required to provide the building cooling
4 — 1754 — 61.78 — 425.65
load which will slightly reduce the power consumption of
5 1754 1754 32.72 29.65 248.67 218.96
the supply fan. Furthermore, the lower refrigerant enthalpy
entering the evaporator tends to decrease the compressor
work and in turn avoids the unnecessary heat rejection in the
condenser. However, increasing make-up water temperature
of the cooling tower slightly increases the energy consumption
of cooling tower VAV fan.
The total energy usage of the whole system for each
summer month can be obtained by summation of the system
energy consumption in each working hour. Figure 7 shows
the simulation results for the average energy consumption
of the conventional and newly-developed air conditioning
Fig. 5 The temperature of refrigerant entering and leaving the systems and compares it in each summer month. It can be
condenser seen that the power consumption of the newly-configured
Vakiloroaya et al. / Building Simulation
air conditioning system is significantly less than the power
usage of conventional plant.
Fig. 7 Central cooling plant average monthly power consumption
6.1 Effect of storage tank area variation
Figure 8 shows the variation of the system energy con-
sumption as a function of the water storage tank volume
for summer. From this figure it can be seen that with the
increasing storage tank volume and therefore the amount
of the make-up water, the electricity usage decreases quite
rapidly initially and then more slowly subsequently. It is
interesting to note that the storage tank volume greater than
50 m3 does not have influence on the system energy usage.
Fig. 8 Variation of the system power consumption versus storage
tank area
7 Conclusions
In this paper a new configuration for the central cooling
plant has been developed, successfully demonstrating a
significant improvement in the system performance. In this
design, a storage tank is implemented in the discharge line
between the compressor and condenser. The storage tank
uses the make-up water of the cooling tower to reduce the
temperature of refrigerant entering the condenser. The
performance of the proposed system is then analyzed by
using a well-validated building transient simulation code.
The experimental results for the newly-configured system
demonstrated that the new design delivers higher system
11
efficiency due to the higher refrigeration effect in the direct
expansion evaporator which allows for a performance gain.
Thus, the new design is promising for an improvement of
the system performance while fulfilling the cooling demand
with guaranteed energy efficiency. The modified system can
save electricity up to 18%.
Electronic Supplementary Materials: supplementary
materials (photos of the screw chiller, air-handling unit,
and cooling tower) are available in the online version of this
article at http://dx.doi.org/10.1007/s12273-013-0104-0.
References
ANSI/ASHRAE Standard 140 (2007). Standard method of test for
the evaluation of building energy analysis computer program.
Atlanta, USA: American Society of Heating, Refrigeration and
Air-Conditioning.
Beghi A, Cecchinato L (2011). Modelling and adaptive control of
small capacity chillers for HVAC applications. Applied Thermal
Engineering, 31: 1125  1134.
Bergero S, Chiari A (2011). On the performance of a hybrid air-
conditioning system in different climatic conditions. Energy, 36:
5261  5273.
Chen W, Deng S (2006). Development of a dynamic model for a DX
VAV air conditioning system. Energy Conversion and Management,
47: 2900  2924.
Diaz G (2007). Controllability of cross-flow two-phase heat exchanger.
International Journal of Heat and Mass Transfer, 50: 4559  4567.
Deng S, Dai YJ, Wang RZ (2012). Performance study on hybrid solar-
assisted CO2 heat pump system based on the energy balance of
net zero energy apartment. Energy and Buildings, 54: 337  349.
DoE (2001). Energy Information Administration, Table US-1, Electricity
consumption by end use in U.S. households. U.S. Department of
Energy.
DoE (2006). Energy Information Administration, International Energy
Outlook. U.S. Department of Energy.
Huang G, Wang S, Xu X (2009). A robust model predictive control
strategy for improving the control performance of air-conditioning
system. Energy Conversion and Management, 50: 2650  2658.
IEA (2008). International Energy Agency, Online Energy Statistics,
Energy Balances. Available at http://www.iea.org/stats/.
Jeon J, Lee S, Hong D, Kim Y (2010). Performance evaluation and
modeling of a hybrid cooling system combining a screw water
chiller with a ground source heat pump in a building. Energy, 35:
2006  2012.
Jin GY, Cai WJ, Wang YW, Yao Y (2006). A simple dynamic model
of cooling coil unit. Energy Conversion and Management, 47:
2659  2672.
Kusiak A, Xu G, Tang F (2011). Optimization of an HVAC system
with strength multi-objective particle-swarm algorithm. Energy,
36: 5935  5943.
12
Lee TS, Lu WC (2010). An evaluation of empirically-based models for
predicting energy performance of vapor-compression chillers.
Applied Energy, 87: 3486  3493.
Lin JL, Yeh TJ (2007). Modeling, identification and control of air-
conditioning systems. International Journal of Refrigeration, 30:
209  20.
Mortada S, Zoughaib A, Clodic D, Arzano-Daurelle C (2012). Dynamic
modeling of an integrated air-to-air heat pump using Modelica.
International Journal of Refrigeration, 35: 1335  1348.
Mossolly M, Ghali K, Ghaddar N (2009). Optimal control strategy for
a multi-zone air conditioning system using a genetic algoruthm.
Energy, 34: 58  66.
Schoenfeld JM (2008). Integration of a thermoelectric subcooler into
Vakiloroaya et al. / Building Simulation
a carbon dioxide transcritical vapor compression cycle refrigeration
system. Master degree thesis, Maryland University, USA.
Soylemez MS (2004). On the optimum performance of forced draft
counter flow cooling towers. Energy Conversion and Management,
45: 2335  2341.
Stene J (2005). Residential CO2 heat pump system for combined space
heating and hot water heating. International Journal of Refrigeration,
28: 1259  1265.
Tashtoush B, Moihim M, Al-Rousan M (2005). Dynamic model of an
HVAC system for control analysis. Energy, 30: 1729  1745.
Wang YW, Cai WJ, Soh YC, Li SJ, Lu L, Xie L (2004). A simplified
modeling of cooling coils for control and optimization of HVAC
systems. Energy Conversion and Management, 45: 2915  2930.