Heat and Mass Transfer (2020) 56:671–690

https://doi.org/10.1007/s00231-019-02726-7

ORIGINAL

Energy use analysis of the variable refrigerant flow (VRF) system

versus the multi - split unit using TRNSYS
Y. H. Yau 1,2 & M. ‘Amir 1

Received: 11 February 2019 / Accepted: 9 August 2019 / Published online: 2 September 2019
# Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
The usage of air - conditioning in the residential and commercial buildings is becoming necessary because of the enormous
demand for thermal comfort and healthy indoor environment. In achieving human thermal comfort and healthy indoor environ-
ment in a large building, there are many types of air conditioning systems that can be used, including the multi - split type unit and
the variable refrigerant flow (VRF) system. The present research analyses the annual energy use of the existing ACMV system
(multi - split type ceiling cassette unit) installed in the examined building, which located in the tropical area and compares it with a
VRF system. The TRNSYS software is used to simulate both of these systems in accordance with the existing building’s
characteristics and its operation. The existing building’s characteristics and its operation are obtained by conducting the site
survey and field measurements in the building. In the current paper, the bin method is used to estimate the annual energy
consumption for these systems. Based on the simulation results, the existing ACMV system consumes more energy in compar-
ison to the VRF system by 13.62%. In addition, the annual operating cost of these systems has shown that the VRF system has a
lower annual operating cost in comparison to the existing ACMV system by 13.63%. Besides that, the VRF system can
accommodate an estimated payback period of around 6.6 years and an internal rate of return of 12.75%. The result has suggested
that the VRF system has a great potential in energy savings and could reduce the electricity consumption in a large building
significantly.

Nomenclature NPV Net present value

k Discount rate PBP Payback period
PI Power input (kW) SEER Seasonal energy efficiency ratio
Q̇ Cooling capacity (kW) TMY2 Typical meteorological year - 2
TRNSYS Transient systems simulation program
Abbreviations VAV Variable air volume
ACMV Air – conditioning and mechanical ventilation VRF Variable refrigerant flow
AHU Air handling unit VRF HP Variable refrigerant flow heat pump
COP Coefficient of performance
EEVs Electronic expansion valves
FORTRAN Formula translation program
FPFA Fan - coil plus fresh air
GSHP Ground source heat pump system 1 Introduction
HP Horse power
IIR Internal rate of return In a tropical climate, the usage of air - conditioning system is
important to support cooling in the building, especially in a
day by controlling both indoor temperature and humidity.
* Y. H. Yau However, a large quantity of energy needs to be consumed
[email protected] in order to provide cooling for the building. In the worldwide,
energy consumed in the building sector, including residential
1
Department of Mechanical Engineering, University of Malaya, and commercial buildings is more than 30% and has steadily
50603 Kuala Lumpur, Malaysia increased between 20 and 40% in the developed countries,
2
UM-Daikin Laboratory, Faculty of Engineering, University of which exceeds the industrial and transportation sectors due
Malaya, 50603 Kuala Lumpur, Malaysia to the spread of the air - conditioning and mechanical
672 Heat Mass Transfer (2020) 56:671–690

ventilation (ACMV) system, in which the ACMV system con- However, the main disadvantage of the VRF system
sumes about one-third to one-half of the total building energy is the cost [1, 13, 14]. The initial cost for this system
[1–4]. is quite expensive than the other ACMV systems [1,
In the U.S. and Korea, the ACMV system is responsible for 14]. Several cost comparisons have been made between
consuming about 47% and 48% of energy in the building, the VRF system and another ACMV system. For ex-
respectively [3]. However, for countries in the tropics, which ample, the total costs for VRF system are around 5%
are hot and humid, the ACMV system consumes more than to 20% more expensive than the chilled water system
50% of energy and it can even reach as high as 70% [5–7]. For with equivalent capacity, around 30% to 50% more
example, in Singapore, the ACMV system installed in the expensive than the single package ducted system with
single largest building consumes about 60% of energy [8]. the similar capacity and seasonal energy efficiency ra-
Moreover, United Nations Development Programme [9] indi- tio (SEER) of 13 to 14, and the VRF system costs
cates that in Malaysia, the ACMV system used in the building more than twice when compared to the packaged ter-
consumes for about 52% of energy and is expected to increase minal unit [13, 15].
throughout the years in between the year 2010 and 2040 by According to the comparison between the VRF system
56% as mentioned in the report on the present situation of the with the water - cooled chiller and the air - cooled chiller for
global energy consumption by the International Energy a 200 ton cooling system in a commercial building, the instal-
Outlook [10]. Thus, it is crucial to study the details about the lation cost for the VRF system is 8% more expensive than the
energy review on the variation of the ACMV designs in order water - cooled chiller, and about 16% more expensive than the
to assist the designers or the end users to choose the best one air - cooled chiller. However, the VRF system saves about 30–
among of them to be installed in the desired building, so that 40% of the energy, where it can accommodate an estimated
the energy saving goal can be achieved. payback period of about 1.5 years compared to the air - cooled
The previous literature reviews reveal that the variable re- chiller, and about 8 months compared to the water - cooled
frigerant flow (VRF) system consumes less energy when com- chiller [11, 13].
pared with other ACMV systems. For example, the compari- In addition, when comparing the VRF system with the
son of the VRF heat pump (VRF HP) system and the rooftop existing conventional ACMV system under the same indoor
VAV system in the medium office building at the 16 U.S. and outdoor conditions for the initial, operating, and mainte-
climate locations shows that the VRF HP system can save nance costs, it clearly shows that the VRF system is more
around 14% - 39% of energy, and after the conversion to the economical one since the installation and maintenance of this
source energy use, this system is expected to save about 2% - system is easier compared to the conventional ACMV system
32% of energy. Furthermore, from this comparison, it also due to the conventional ACMV system operates with many
clearly shows that the VRF HP system has a lower energy different equipment [11]. Besides that, by having many differ-
cost in the hot and mild climates while the VAV system has ent equipment to operate, the rate of malfunction of this sys-
a lower energy cost in several cold climate zones [4]. Another tem can be increased. In addition, the VRF system is provided
comparison study between the VRF system and the fan coil with its own automation equipment (the smart-integrated con-
system shows that the VRF system can save more energy. This trol), while the conventional system needs to add additional
is because the VRF system comes with the temperature preci- costs for adding the automation equipment and its software
sion ±0.5 °C due to the reason of having a modulating capac- into the system [11, 16].
ity control facility between 10% - 100%, while the fan coil Even though the initial cost for the VRF system is quite
system comes with the temperature precision up to 2 °C - 3 °C, expensive, the present system has a high potential for energy
which can increase the energy cost by 5% - 6% for every 1 °C savings, in which it can increase the possibility of the reduc-
[11]. tion of the payback period, and thus, the energy saving effect
Besides, Liu [12] analyzed the operational electricity con- of the VRF system is worth studying, especially in hot and
sumption between the VRF system with the centralized air - humid climates, in which published research reports are very
conditioning system in the two office buildings at the campus rare and hard to be found as indicated from the literature re-
based on the monitoring data obtained from a building energy view. Therefore, in the current research, the major aim is to
monitoring and management system. From this research, the estimate the energy performance of the VRF system by com-
rate of electricity consumption obtained was 0.48 kWh/m2 for paring the energy consumption of the system with the other
the VRF system and 0.42 kWh/m2 for the centralized air - ACMV system, which is the multi - split type unit system at
conditioning system, which can be concluded that both sys- the same indoor and outdoor conditions by using the bin meth-
tems have a great potential in energy savings. However, the od. The minor aim is to examine whether the bin method is
electricity consumption of the VRF system is better than the suitable to be applied in the simulation model by developing
centralized air - conditioning system due to the flexible adjust- and carrying out the numerical analysis of the energy simula-
ability of the VRF system [12]. tion for both systems in the TRNSYS coding.
Heat Mass Transfer (2020) 56:671–690 673

2 Differences between the VRF system units, where the separation tube has two branches while the
and the multi - split type unit system header has more than two branches [14]. The branch selector
box in the installation of the refrigerant pipes helps the VRF
A multi - split type unit system operates in the similar princi- system to operate with five different modes [13, 20]:
ples as a split type unit system. However, different from the
split type unit system, the multi - split type system has a & Cooling mode (only for cooling process).
capability to link a single outdoor unit to several indoor units, & Heating mode (only for heating process).
which is similar to the VRF system [1, 13–22]. & Main cooling mode (cooling is the main mode in the si-
However, the design of the VRF system contributes to the multaneously heating and cooling process).
simultaneous cooling and heating at divergent zones with the & Main heating mode (heating is the main mode in the si-
helps of heat recovery, and individualized zoning controls as multaneously heating and cooling processes).
shown in Fig. 1 [4, 14–17, 19, 20, 22–24]. The VRF system & Heat recovery mode (heat is balanced between indoor
has a capability for individualized zoning controls due to the units while the outdoor unit heat exchanger is closed).
usage of electronic expansion valves (EEVs). The EEVs help
the VRF system to operate separately in the different zones Moreover, the VRF system usually needs a separate venti-
and also some of them can be closed if the zones are unoccu- lation to provide a better indoor air quality and “free cooling”
pied or the desired indoor temperature is obtained while the because the system is dustless, while the multi - split type unit
others continue to operate [1, 13, 14, 18]. system is usually provided with the ventilation for “free
It is pertinent to mention that, there are two basic types of cooling” when the outdoor temperature is lower than the re-
VRF system, which are simultaneous cooling and heating at circulation air temperature [13, 16, 23, 25]. Besides, most of
divergent zones and the other type of this system is either used the VRF system manufacturers provide a centralized control
for heating or cooling by the helps of a four - way valve in the option, where the monitoring and controlling of the entire
outdoor unit [13, 14, 17, 18, 21, 22, 25]. It should be noted that system can be done either from a single location or by using
the heat recovery in the VRF system assists this system to the internet rather than using one master controller as in the
reject the heat from one zone in cooling mode, where this case of the multi - split type unit system [14, 16].
rejected heat can be used in other zones for heating mode
rather than being rejected to the outdoor [24].
Different from the multi - split type unit systems, the VRF 3 Examined building overview
system is provided with either two or three pipe configurations
to minimize the refrigerant path for increasing the efficiency 3.1 Building description
of the refrigerant flow in the system [1, 13, 20]. In the VRF
system, both separation tubes and headers are used to connect The target building is the R&D Centre of Daikin R&D Pte
the refrigerant pipes from the outdoor unit to several indoor Ltd., which is located at Sungai Buloh, Malaysia, where the

Fig. 1 Representative layout of

the VRF system used in the
building with several zones
674 Heat Mass Transfer (2020) 56:671–690

investigation activities on improving the existing ACMV Furthermore, this system operates from 8:30 AM until
products and procedures or the development of new ACMV 5:30 PM on weekdays and is assumed to be fully operated at
products and procedures are conducted. The building operates all time of the day throughout the entire year.
for a duration of 9 h, which is from 8:30 AM until 5:30 PM on
weekdays. 3.2.2 VRF system
The building has four floors and the assessments are only
conducted on the third floor. The third floor of this building The VRF system is used to compare with the existing ACMV
consists of a large hall, two toilets, an air handling unit (AHU) system installed in the examined building. A 36-HP
room, and 7 rooms as shown in Fig. 2. The only air - condi- (26.84 kW) VRF system is chosen to match up with the
tioned area is in the rooms, where each room has 73 m2 of net existing ACMV system’s horse power. Furthermore, the
area and 3 m of height. Figure 3 shows the plan view of one of VRF system is expected to operate at the same time just as
these rooms. It should be noted that each room accommodates the existing ACMV system, which is from 8:30 AM until
24 occupants who work mostly in front of the computer. 5:30 PM on the weekdays and is assumed to be fully operated
at all time of the day throughout the entire year. The installa-
3.2 ACMV system description tion of the VRF system is assumed to be exactly the same just
as the present ACMV system, which is the outdoor unit of this
3.2.1 Existing ACMV system installed in the target building system, and is placed on the rooftop and the outdoor air is
introduced on this floor through an AHU.
A multi - split type ceiling cassette unit serves the entire space
on the third floor of the target building with ventilation by
maintaining the indoor temperature, which is 24 °C in each 4 Methodology
room. The outdoor unit of this system is placed on the rooftop
of the building and is connected with several indoor units with The present paper analyses the annual energy use of the multi -
2.5 HP (1.86 kW) for each unit. Each room on the third floor split type unit system and the VRF system. The TRNSYS
of the building consists of two indoor units. The outdoor air is software is used to simulate both of these systems in accor-
introduced on this floor through an AHU. Figure 4 shows the dance with the data obtained, which will be explained later in
schematic diagram of the current system in each room. the next two sub-sections 4.1 and 4.2.

Fig. 2 Third floor of the

examined building’s layout

Rooms Staircase/Exit

Hall AHU room

Toilets
Heat Mass Transfer (2020) 56:671–690 675

Fig. 3 Top plan view of one of

12.7 m
the rooms on the third floor of the
examined building

4.08 m
1.52 m
13.94 m

Signs Descriptions
Circular ceiling diffusers
(200 mm diameter)
4 way diffuser grilles
(0.46 m x 0.46m)
Ceiling light panels
(1.22 m x 0.48 m)
Indoor unit (Ceiling Cassette Type)
(0.7 m x 0.7 m)
Window
(3.08 m x 1.75 m)
Inward double swing door
(2.1 m x 1.81 m)

TRNSYS software can simulate the energy for building’s period can be obtained [26–29]. Typically, in TRNSYS, a
heating or cooling systems on the basis of the hour - by – hour model is represented by components link together based on
analysis by using the hourly weather data for the whole year, their constant parameters and input or output variables, where
where the adequate simulation results for the long - term these components can be obtained in the TRNSYS library [30,

Fig. 4 The schematic diagram of

the existing ACMV installed in
one of the rooms on the third floor
of the examined building

Window

Door

Indoor unit Diffuser

Return air Cooling air

Supply air Return air

from AHU to AHU
676 Heat Mass Transfer (2020) 56:671–690

31]. However, if the components needed are not provided in performing a physical measurement, the Alnor
the TRNSYS library, these components can be created or velometer AVM440 - A is stabilized for about 5 min
bought from a third party, where a good knowledge of com- and then, in every thirty minutes, the indoor air relative
puter programming, which is FORTRAN software is neces- humidity, temperature, and velocity for occupied hours
sary in order to create the new components [26, 32]. The of this building are measured and recorded. It should be
ability to create new components is essential in this research noted that the procedures mentioned are based on the
because the VRF system cannot be simulated as a built - in study by Yahaya [5]. These data are taken for a day and
module by some of the simulation software, including assume to be the same for each day in the entire year.
TRNSYS, DOE - 2, BLAST, and EnergyPlus since these soft-
ware have not yet provided support for the VRF system due to
the complexity of the system’s operation [33]. 4.2 Data collection II
It is important to mention that the baseline simula-
tion developed in this research is based on the bin 4.2.1 Bin weather data
method. By using the bin method, the heating or
cooling energy can be obtained if the building design Since the simulation developed is based on the bin
loads are known, where these loads will be divided by method, the latest bin weather data for around 10-
the coefficient of performance (COP) of the ACMV years duration are needed. Thus, for the purpose of
system used in the examined building, and then these the research, the bin weather data in Petaling Jaya,
loads will be multiplied with the total hours of each Malaysia in years 2007 to 2016 adopted from Yau
bin temperature within a month or year, which clearly [38] as shown in Fig. 5 is used as the replacement of
shows that the bin method is suitable to be used for a the typical meteorological year - 2 (TMY2) weather
long-term periods, either monthly or annually and at data files in Petaling Jaya, Malaysia.
any different temperature intervals [27, 28, 34–36].
This clearly shows that this method can obtain ade-
quate results for the annual energy consumption for 4.2.2 Operating performance data of the ACMV system
the building and the cooling or heating system used
in the building, and also presents the analysis of the The operating performance data of the ACMV systems
energy change based on the variation of the bin tem- used in the current work, including the power input and
perature significantly more convenient, which can help the cooling capacity at different outdoor temperatures,
in sizing the proper cooling or heating system equip- are needed to ensure that the energy analysis of the
ment for the desired buildings. It is pertinent to men- ACMV systems used in the examined building can be
tion that the bin temperature is the variation of outdoor done by using the bin method. Through the findings of
temperature conditions (obtained by the bin weather the power input and the cooling capacity of the ACMV
data), where these temperatures are divided into dis- system, the COP of this system can be obtained as
crete groups (bins) in accordance with the weather mentioned below [34]:
conditions [26, 35].
COP ¼ Q˙ =PI ð1Þ
4.1 Data collection I
The function of the cooling capacity of the ACMV
In this simulation, the bin weather data, indoor temper- systems used in this research can be obtained by
ature, velocity, and relative humidity, coefficient of per- selecting two cooling capacities and outdoor tempera-
formance (COP) of the ACMV equipment, and the tures based on their operating performance curve as
building design loads, including solar, conduction and shown in Fig. 6. It should be noted that the operating
internal loads, and infiltration or ventilation load will performance curve for these systems is based on the
be acted as the inputs and parameters in the TRNSYS operating performance data of the systems as indicated
simulation [34, 36, 37]. The examined building enve- in Tables 7 and 8. It should be noted that these data
lope’s characteristics and building design details, which were obtained from the vendor who manufactures the
shown in Tables 1 and 2, respectively, can be obtained systems, where they only give permission to use these
by doing a site survey. data for academic purpose and nonprofit research only.
However, for obtaining the indoor air temperature, The similar procedures were used to find the linear
relative humidity, and velocity in the examined building, function of the power input for both systems. The op-
the physical measurements need to be measured by erating performance details of these systems are shown
using the Alnor velometer AVM440 - A. Before in Table 3.
Heat Mass Transfer (2020) 56:671–690 677

Table 1 Building envelope characteristics of the examined building

Envelope Element Description U – factor (W/m2.K) SC Unit Area (m2)

Roof • Outside air film 0.38 – 1 73

• 9.8” RC beam
• 1” Cement plaster (0.5″ on both sides)
• 2” Fiber glass
• 0.5″ Gypsum board
• Inside air film
Floor • Outside air film 2.89 – 1 73
• 9.8” RC beam
• Inside air film
Exterior Wall a • Outside air film 0.54 – 1 41.82
• 4.5” Brick wall
• 1” Cement plaster (0.5″ on both sides)
• 0.5″ Gypsum board
• 2” Fiber glass
• Inside air film
Exterior Wall b • Outside air film 0.54 – 1 38.10
• 4.5” Brick wall
• 1” Cement plaster (0.5″ on both sides)
• 0.5″ Gypsum board
• 2” Fiber glass
• Inside air film
Exterior Wall c • Outside air film 0.54 – 1 9.66
• 4.5” Brick wall
• 1” Cement plaster (0.5″ on both sides)
• 0.5″ Gypsum board
• 2” Fiber glass
• Inside air film
Exterior Wall d • Outside air film 0.54 – 1 13
• 4.5” Brick wall
• 1” Cement plaster (0.5″ on both sides)
• 0.5″ Gypsum board
• 2” Fiber glass
• Inside air film
Door • 35” Wood solid core flush door 1.48 – 1 3.80
Window • 1/4″ Double glazing, uncoated clear 0.71 0.61 4 1.786

4.3 TRNSYS simulation 4.4 The bias uncertainty and percentage uncertainty

The TRNSYS software associated with IISIBAT 3.0 software In the present research, the total annual energy con-
helps to assemble the appropriate TRNSYS baseline simulation, sumed for both systems using the bin method has also
where the current simulation is developed based on the bin meth- been calculated manually by multiplying the cooling
od as shown in Fig. 7. It should be noted that the red lines in Fig. load produced by the ACMV systems installed in the
7 represent the simulation of the examined building at the unoc- examined building and its design loads, including loads
cupied operating hours and the arrows in the simulation indicate for solar gain, conduction, internal energy and ventila-
the energy calculation flow. In addition, from Fig. 7, a simplified tion systems with the total hours of each bin tempera-
TRNSYS baseline simulation of the target building for occupied ture in a year, which can be obtained in the bin weather
hours is developed as shown in Fig. 8. It should be noted that the data as mentioned in section 4.2.1 [34, 37].
simplified TRNSYS baseline simulation of the target building for Nonetheless, by using the simulation approach, the re-
the unoccupied hours is the same as the occupied hours. sults can be obtained conveniently in comparison with
The components used in the current simulation are built using the manual calculation, especially when it involves en-
the FORTRAN code to ensure that the bin method can be applied ergy or load calculation on a yearly basis [26, 28].
in the TRNSYS software. It is pertinent to mention that the inputs The bias uncertainty analysis is required to be conducted to
and parameters for the developed components are based on the examine whether or not the errors present in the results ob-
data collection as mentioned in sections 4.1 and 4.2. The func- tained by manual calculation and simulated results are at a
tions of these components are summarized in Table 9. satisfactory level. The background details of the bias
678 Heat Mass Transfer (2020) 56:671–690

Latent Load (kW)

uncertainty can be found in the references [39, 40]. Note that
the results of the bias uncertainty analysis will be discussed
later in section 5.3.

1.08
4.5 TRNSYS simulation based on the bin method
Total Occupancy Load (kW)

Total Lighting Load (kW) as a benchmark for other studies

Sensible Load (kW)

The simulation in the present research is generic and

RH Outdoor (%)
can be used as a benchmark or a guideline for build-
ing energy analysis in other countries. In the present
2.714
1.680

research, this simulation is used for analyzing the en-

80
ergy consumption of the ACMV systems from the
previous studies [17, 41], and their simulation results
are compared with the actual results in order to prove
that the current simulation can be a benchmark for
Power / Unit (kW)
Latent Load (kW)

building energy analysis in other countries. Note that

RH Indoor (%)

the results of the comparison between simulation re-

sults with the actual results will be discussed later in
0.058
0.045

64.65

section 5.5.
All the data including the COP of the ACMV equip-
ment and the building loads from the previous studies
[17, 41] are taken and acted as the inputs and param-
*1Energy / Person (kW)

eters for the current simulation. Besides that, in order

Sensible Load (kW)

to perform the energy consumption estimation by using

Indoor Qair (m3/s)
*1Ballast Factor

the bin method, the bin weather data in references to

the hourly outdoor temperature need to be used
1.00375

[26–28, 34, 37]. Thus, the bin weather data of the

0.070

1.2

present research (refer to Fig. 5) need to be applied

in this simulation because the bin weather data for
the particular locations in the previous studies [17,
41] are not given.
Indoor Air Velocity (m/s)

It should be noted that these previous studies [17,

41] were chosen because the studies have complete data
Light work and sit

*1Heat Gain (kW)

as mentioned in sections 4.1 and 4.2. However, in the

Type of Activity

*2 Table of ideal - gas specific heats of various common gases [50]

previous study by Zhou [41], the comparison of the

simulation result with the actual result was done only
0.01375
*1CLF

for the VRF system due to the complexity installation

0.75
0.75

0.75
1

of the other two systems, which is FPFA and VAV

systems, where both systems have been installed with
additional features such as chiller, central fan, and water
Examined building design details

pump. The energy consumption for the ACMV system

*2Cpair (KJ/kg.K)
Fluorescent light

with additional devices can only be used in the present

*1 Data obtained from ASHRAE [49]

simulation if the additional components are created for

Quantity

these devices.
*1CLF

1.005
Type

In the same previous study just mentioned by Liu

25
25

1
1

and Hong [17], the comparison of the simulation result

with the actual result was done only for the VRF and
GSHP systems, which are used for either heating or
Occupancy Detail

Equipment Detail

Infiltration Detail
Lighting Detail

cooling because in the simulation, the systems that

Water Cooler

ρair (kg/m3)
Equipment

operate simultaneously for both heating and cooling

Quantity

Quantity

Desktop
Monitor
Table 2

Laptop

are quite difficult to simulate due to the complexity

1.225
24

24

in their thermodynamics processes and workflow.

Heat Mass Transfer (2020) 56:671–690 679

Fig. 5 The bin weather data

24 - 26 °C 26 - 28 °C 28 - 30 °C 30 - 32 °C 32 - 34 °C
details in Petaling Jaya in years 1200
2007 to 2016 (redrawn based on
reference [38])
1000

Frequency (from 2007 - 2016)

800

600

400

200

0
0000-0400 0400-0800 0800-1200 1200-1600 1600-2000 2000-2400
4 - Hours Shi

4.6 The payback period (PBP) and the internal rate The PBP is used to estimate the time needed to repay the
of return (IRR) of the VRF system initial cost of the VRF system as shown by the equation below
[42, 43]:
Based on the energy analysis from the previous section, the Initial Cost of New Unit
payback period (PBP) and internal rate of return (IRR) of the PBP ¼
Annual Operating Cost of Existing Unit−Annual Operating Cost of New Unit
VRF system can be determined, in which the result of the PBB ð2Þ
and the IIR of the VRF system will be discussed later in
section 5.5. where the existing unit is the multi - split type system
and the new unit is the VRF system. It is pertinent to
mention that the annual operating costs of both
Cooling Capacity Power Input
ACMV systems is based on the new current tariff,
10 4.00 in which it can be obtained from Tenaga Nasional
Berhad [44]. The calculation to estimate the electric-
8
3.00 ity usage over a 30 - day period of both systems
Cooling Capacity (kW)

Power Input (kWh)

6 serves as a guideline only, and the actual bill amount

2.00 may vary from that stated in the present research.
4
Besides that, since the initial costs of the VRF system
2
1.00 are for the whole building, while the assessments
were only conducted on the third floor of the exam-
0 0.00 ined building, and thus, it is assumed that the energy
19 25 30 35 40 46
Outdoor Temperature (°C)
consumption for every floor in the examined building
a) Performance curve for the existing ACMV system (multi - split type unit)
is the same as the third floor.

Cooling Capacity Power Input

116 35.00

114 30.00
Table 3 Operating performance details of the existing ACMVand VRF
Cooling Capacity (kW)

25.00 systems
Power Input (kWh)

112
20.00
110 Existing ACMV System (Multi - Split Type Unit)
15.00
108 Operating Performance A Linear Function (kW)
10.00
106
Cooling Capacity 14 [9.435–0.061(Tbin)]
5.00
Power Input 14 [1.21 + 0.038(Tbin)]
104 0.00
10 12 14 16 18 20 22 24 25 26 28 30 32 34 35 VRF System
Outdoor Temperature (°C)
Operating Performance A Linear Function (kW)
b) Performance curve for the VRF system Cooling Capacity 127.7–0.54 (Tbin)
Power Input 16.43 + 0.45(Tbin)
Fig. 6 Performance curve of the existing ACMV and VRF systems
680 Heat Mass Transfer (2020) 56:671–690

Type65a
Energy

COP
Type206
Qinternal

Qcond
Type201

Type202

Type203

Type204

Type205
Qsolar

Bin Weather
Data
*Red line = Unoccupied energy
*Venlaon load is assumed 0

Fig. 7 TRNSYS baseline simulation of the examined building Fig. 8 Simplified TRNSYS baseline simulation of the examined building
for occupied hours

The IRR of the VRF system can be obtained as shown in

the equation below [45]: 5 Results and discussion
ðCash FlowÞn
NPV ¼ −Initial Cost of New Unit þ ∑1n ð3Þ
ð1 þ k Þn 5.1 Simulated result
IRR ¼ k ðwhen NPV ¼ 0Þ ð4Þ
As mentioned earlier, the assessments were only con-
where n is the number of year(s). The number of years ducted on the third floor of the examined building,
has been defined according to the standard average where the only air - conditioned area is in the rooms,
lifespan of the ACMV system, which is approximately and thus the estimation of the energy consumption was
15 years. It pertinent to mention that the (Cash Flow)n done on the third floor only by using the room - by -
is the cash flow in the period (i.e. 15 years in the room method with an assumption that the size and the
present case). Noted that the cash flow is the difference room design details are the same for all rooms.
of the operating costs of the new and the existing In the present paper, the bin method is used for es-
ACMV systems, where the operating cost for both sys- timating the building’s total cooling energy consumption
tems are assumed to be the same for each year in the of the ACMV system. Thus, the simulated result for
15 years. Furthermore, k is the interest rate that has both systems is based on the energy consumption at
resulted the net present value (NPV) equals to zero, many outdoor temperatures (i.e. the bin temperatures)
where the NPV is the estimation of the future cash flow distributions for every 4 - h shifts in Petaling Jaya as
discounted back to the present moment. shown in Fig. 5. As shown in Figs. 9 and 10, the
Heat Mass Transfer (2020) 56:671–690 681

Fig. 9 Simulated result of the

existing ACMV system

simulation result for both systems is divided into occu- the bin temperature interval is 25 °C followed by
pied and unoccupied periods. It should be noted that 33 °C. The result suggests that sometime the tempera-
both occupied and unoccupied hours in the simulation ture in Malaysia can reach its highest peak at about
are 4380 h each. The occupied and unoccupied hours in 33 °C, which is extremely warm, and it rarely goes
the simulation are selected based on the relation be- below 25 °C. Thus, the finding strongly suggests that
tween the bin weather data used in the present work Malaysia has a high daytime temperature, which is be-
(refer to Fig. 5) with the examined building operation tween 25 °C to 33 °C.
hours. The red line in both figures represents the fre-
quency of the bin temperature interval based on the bin 5.2 The annual energy consumption comparison
weather data while the blue line represents the amount between the VRF system and the multi - split type unit
of energy consumption of the ACMV system in each system
bin temperature interval.
Then, the simulated results for both systems are sim- As shown in Fig. 11, the energy consumption of the VRF
plified as indicated in Fig. 11 by multiplying the fre- system is lower compared to the existing ACMV system
quencies for every midpoint of the bin temperature in- (i.e. the multi - split type unit) at almost every midpoint of
terval with the energy consumption at each of their own the bin temperature interval. The difference value is about
midpoint of the bin temperature interval and then, the 2.85–10.97 MWh. The difference in the total energy con-
result of the occupied and unoccupied periods are sumption for both systems is 37.70 MWh, and in term of
summed up for each system. It is observed from percentage, it is around 13.6%. Thus, it can be concluded that
Fig. 11 that the highest frequency of the midpoint of the VRF system is more energy efficient than the multi - split
the bin temperature interval is 27 °C. This shows that type unit system.
the average temperature in Malaysia is 27 °C. Furthermore, the highest total energy consumption for
Furthermore, the lowest frequency of the midpoint of both systems is at 27 °C, which is 78.13 MWh and
682 Heat Mass Transfer (2020) 56:671–690

Fig. 10 Simulated result for the

VRF system

89.10 MWh respectively, for the VRF system and the values at 21 °C and 23 °C in Fig. 11 are because the
multi - split type unit system, while the lowest total temperature in Malaysia has never reached below 23 °C
energy consumption for both systems is at 33 °C, which as shown in the bin weather data in Petaling Jaya,
is 17.07 MWh and 19.93 MWh respectively, for the Malaysia in years 2007 to 2016 (indicated in Fig. 5).
VRF system and the multi - split type unit system. It Thus, there is no energy calculation involved at the
should be noted that the result obtained is based on the midpoint of the bin temperature intervals of 21 °C and
temperature frequency in every 4 - h and the zero 23 °C.

Fig. 11 Comparison of the total

VRF EXISTING ACMV
energy consumption (MWh) of
the existing ACMV and the VRF 100.00
system based on the midpoint of 90.00
Total Energy Consumption (MWh)

bin temperature interval (°C) 80.00

70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
21 23 25 27 29 31 33
Midpoint of Bin Temperature Interval (ºC)
Heat Mass Transfer (2020) 56:671–690 683

5.3 The bias uncertainty and percentage uncertainty actual study was taken place in the subtropical country and
between the simulation results and manually experiences four distinct seasons. The bin weather data from
calculated results different locations are not suitable to be used because the
performance of the equipment that operates on the refrigera-
Table 4 shows the bias uncertainty and the percentage uncer- tion cycle, such as the ACMV system varies at different
tainty for the total annual energy consumed for the ACMV weather conditions [35]. The different weather conditions
systems used in the current research. As shown in Table 4, for have a different ΔT, where in most of the tropical countries
the simulation result, the energy consumed within a year by show a direct linear relationship between the ΔT and the
the VRF system is 257.92 MWh, while for the result of cal- cooling or heating loads while most of the countries, which
culation is 257.98 MWh. The percentage uncertainty for both experience four distinct seasons, have a varied non-linear re-
systems is 0.03%. As for the existing ACMV system (i.e. the lationship between the ΔT and the cooling or heating loads
multi - split unit), the simulation result for the energy con- [35]. Note that the ΔT is the difference in temperature of a
sumed within a year is 295.62 MWh, while for the result of heating or cooling medium as it enters and as it leaves a sys-
calculation is 295.27 MWh. The percentage uncertainty for tem that operates in the refrigeration cycle.
both systems is 0.06%.
The uncertainty analysis indicates that the error is insignif- 5.5 Analyses of the payback period (PBP)
icant and the findings are reliable, where the percentage un- and the internal rate of return (IRR) of the VRF system
certainty obtained for both systems is quite small, which is
below 1%. This clearly shows that the manual calculated re- From Table 6, the initial costs of the VRF system are based on
sults have agreed well with the findings of the simulation the current market price in Malaysia, where the equipment and
model. installation costs for the VRF system are RM 83000 (USD
19902) and RM 65000 (USD 15586), respectively [46].
5.4 TRNSYS simulation result on the previous studies However, the initial costs for multi - split type unit are not
provided since this system has been used for several years in
TRNSYS simulation from the present research is used for the examined building. Furthermore, the operating cost for the
modeling and simulating the energy consumption of the VRF system is cheaper than the existing ACMV system
ACMV systems from the previous studies [17, 41]. Then, installed in the examined building. The operating cost for
the simulation results are compared with the actual results as the VRF system is RM 151821.08 (USD 38000) while the
shown in Table 5. It should be noted that the simulation result operating cost for the existing ACMV system is RM
from Zhou’s [41] study (indicated in Fig. 12) as shown in the 174031.36 (USD 44000). The difference of the overall oper-
Appendix 3 also acts as the representative of the simulation ating cost of both systems is RM 22210.28 (USD 5600),
results from the other studies [17]. As mentioned in section which is about 13.63%.
5.1, the occupied and unoccupied hours in the simulation are The VRF system can accommodate an estimated PBP for
selected based on the relation between the bin weather data about 6.6 years. In other words, the initial cost for the VRF
used in this research (refer to Fig. 5) with the examined build- system is paid off before the average lifespan of this system,
ing operation hours. which is 15 years. Furthermore, the IRR of this system is
Based on the finding obtained, the simulation result from 12.75%, and according to the Central Bank of Malaysia, the
the Zhou’s [41] study at 1105.41 MWh is much lower com- actual national basic IRR is 3.1% per year. Since the obtained
pared to the actual result at 1533.33 MWh. The percentage IRR is at the minimum acceptable rate of return and the PBP is
difference for both results is 32.43% as indicated in Table 5. less than the average lifespan of the VRF system, it is clear to
Furthermore, based on the simulation validation for studies suggest that the VRF system is economically viable.
carried out by Liu and Hong [17], for the VRF system, the Based on the results obtained, it clearly shows that the VRF
simulation result is 25.93 MWh, while the actual result is system has greater energy savings potential, where this system
17.36 MWh, where the percentage difference in both results can save more energy and reduce significant electricity con-
is 39.60%. Meanwhile, for the GSHP system, the simulation sumption in the building located in the tropics when compar-
result is 14.17 MWh, while the actual result is 13.39 MWh, ing with the multi - split type system. This is mainly due to the
where the percentage difference in both results is 5.71%. advantages of this system, including the capability of simul-
Even though the percentage difference between the simu- taneous cooling and heating in different zones and the inde-
lation results and the actual findings are quite large for both pendent zoning controls, where this system is able to provide
studies, it clearly shows that the current simulation can be used different indoor temperatures at different zones. Besides that,
as one of the alternative ways for building energy analysis. this system can be shut down if one or more zones are unoc-
These differences only occurred because the bin weather data cupied or the desired indoor temperature is obtained while the
used in the simulation are from a tropical country, whereas the multi - split type unit system is operated with a fixed schedule,
 684

Table 4 Bias uncertainty and percentage uncertainty for the simulation results and calculated results for the existing ACMV and VRF systems

Midpoint of Bin Temperature Interval (°C) Existing ACMV System (Multi - Split Type Unit)
TRNSYS Calculation Average Total Energy (MWh) Bias Uncertainty Percentage Uncertainty (%)
Annual Total Energy (MWh) Annual Total Energy (MWh)
21 0 0 0 0 0
23 0 0 0 0 0
25 38.32 38.30 38.31 0.013 0.033
27 89.10 88.84 88.97 0.133 0.149
29 70.64 70.63 70.64 0.005 0.007
31 77.62 77.59 77.60 0.016 0.020
33 19.93 19.91 19.92 0.007 0.034
Total 295.62 295.27 295.44 0.173 0.058
Midpoint of Bin Temperature Interval (°C) VRF System
TRNSYS Calculation Average Total Energy (MWh) Bias Uncertainty Percentage Uncertainty (%)
Annual Total Energy (MWh) Annual Total Energy (MWh)
21 0 0 0 0 0
23 0 0 0 0 0
25 34.03 33.99 34.01 0.016 0.048
27 78.13 78.18 78.16 0.023 0.029
29 61.63 61.62 61.63 0.002 0.004
31 67.05 67.11 67.08 0.028 0.042
33 17.08 17.08 17.08 0.002 0.009
Total 257.92 257.98 257.95 0.072 0.028
Heat Mass Transfer (2020) 56:671–690
Heat Mass Transfer (2020) 56:671–690 685

Table 5 TRNSYS simulation results of previous studies

Paper by ACMV System Location Energy (MWh) Percentage Difference (%)

Simulation Result Actual Result

Liu & Hong [17] VRF Miami 25.93 17.36 39.60

GSHP 14.17 13.39 5.71
Zhou [41] VRF Shanghai 1105.41 1533.33 32.43

where it does not allow occupants to shut down the system if in the different zones with different temperatures at each zone
one or more of the zones are unoccupied or the desired indoor with the help of the EEVs in the system. Despite the limitation
temperature is obtained, resulting in a much longer operation of the comparison with only the split type unit system, the
time, and thus, consuming much more energy [1, 13]. findings have contributed significantly in the area of
the energy performance of the VRF system in terms of
the research methodology of the valuable site survey
6 Conclusions and the field measurements, as well as the simulations
are generic, and thus they are suitable and very useful
The energy performance of the VRF system has been success- to be used as an important ACMV design guide in the
fully studied by comparing the energy consumption of the buildings, not just in Malaysia, but also in other coun-
system with the multi - split type unit system in the present tries located in the tropics such as Singapore, Indonesia,
work at the same indoor and outdoor conditions by using the Thailand, Brunei and Philippines.
bin method. The results indicate that for conditioning the same
building with the same outdoor and indoor conditions in a Acknowledgements The authors would like to thank the University of
Malaya for providing RU Grants GPF004A-2018 and IIRG014A-2019 to
year, the VRF system saves about 13.62% energy consump- the authors for research work to be conducted at the University of Malaya.
tion and approximately 13.63% on the operating cost in com- Thanks are also extended to Daikin R&D SDN BHD at Sungai Buloh for
parison to the existing multi - split type unit. It clearly shows the permission given to conduct the research in the Daikin R&D Centre
that the VRF system has a greater energy saving potential, Building and sharing the operating performance table of the multi - split
type unit (model: FHC25EXV1 - R25CXV1) with the research team, and
where this system can save more energy and can reduce elec- also providing the partial financial assistance to the co-author, Mr. M.
tricity consumption significantly in the building located in the ‘Amir, via project PV018-2016, for conducting the research work at the
tropics. This is attributed to the usage of EEV control in the HVAC&R Lab in the Department of Mechanical Engineering, University
VRF system. Besides that, the VRF system can accommodate of Malaya. In addition, special thanks are also extended to Hisense -
Hitachi Sdn Bhd for providing the operating performance table of the
an estimated PBP for about 6.6 years and an IRR of 12.75%, VRF system (model: AVWT - 343TESZX).
which clearly shows that this system is economically viable.
The VRF system is most appropriate to be used in a build- Compliance with ethical standards
ing required a large cooling load and needs cooling in several
zones simultaneously with different temperatures at each Conflict of interest On behalf of all authors, the corresponding author
zone. Each indoor unit of this system can operate separately states that there is no conflict of interest.

Table 6 Payback period (PBP) and internal rate of return (IRR) of the VRF system

Initial Cost for VRF system

Equipment Cost (RM) [USD] 83,000 [19902]
Installation Cost (RM) [USD] 65,000 [15586]
Total Costs [USD] 148,000 [35488]
ACMV System Total Energy per Year (MWh) Operating Cost per Year (RM) [USD]
*(Current Tariff +6% GST)
Existing ACMV (Multi - Split Type Unit) 295.62 174,031.36 [44000]
VRF System 257.92 151,821.08 [38000]
Total Difference 22,210.28 [5600]
PBB (year) 6.6
IRR (%) 12.75

* Current Tariff +6% GST = Based on TNB [44]

686 Heat Mass Transfer (2020) 56:671–690

Appendix 1

Table 7 Operating performance of the VRF system [47]

MODEL: AVWT - 343TESZX Remarks: Q = Cooling Capacity; P = Power Input; To = Outdoor Temperature; TI = Indoor Temperature

Combination (%) To/TI 14 16 18 19 20 22 23 24

Q P Q P Q P Q P Q P Q P Q P Q P
100 10 65.9 10.20 71.1 10.64 95.4 15.26 100.4 25.05 104.5 25.42 109.8 25.87 112.3 25.97 115.2 26.17
12 66.3 10.20 70.9 10.64 95.3 15.26 100.3 25.11 104.4 25.49 109.7 25.87 112.1 25.97 115.0 26.17
14 65.8 10.33 71.2 10.70 95.6 15.33 100.7 25.18 104.3 25.55 109.4 25.93 111.8 26.04 114.7 26.24
16 65.9 10.45 71.3 10.83 95.4 15.51 100.5 25.30 103.8 25.68 109.6 26.07 111.9 26.18 114.7 26.37
18 66.0 10.58 71.5 11.01 95.8 15.64 100.3 25.49 104.2 25.87 109.6 26.34 111.9 26.41 114.6 26.58
20 66.0 10.83 71.6 11.26 95.3 15.95 100.5 25.74 104.2 26.13 109.5 26.54 111.7 26.62 114.3 26.79
22 66.1 11.14 71.7 11.58 95.4 16.20 100.7 26.00 104.3 26.39 109.8 26.87 111.9 26.96 113.9 27.14
24 65.9 11.45 71.7 11.89 95.5 16.58 100.7 26.37 104.2 26.78 109.4 27.21 111.9 27.30 114.3 27.48
25 65.7 11.64 71.4 12.11 95.3 16.77 100.5 26.60 103.8 26.97 109.4 27.44 111.9 27.54 114.2 27.73
26 65.7 11.91 71.4 12.41 95.3 17.15 100.5 26.97 103.8 27.35 109.4 27.82 111.9 27.92 114.2 28.10
28 65.6 12.57 71.4 13.01 95.3 17.90 100.5 27.73 103.8 28.10 109.4 28.58 111.9 28.67 114.2 28.86
30 65.6 13.17 71.4 13.62 95.3 18.65 100.5 28.48 103.8 28.86 109.4 29.33 111.9 29.43 114.2 29.61
32 65.5 14.08 71.4 14.60 95.3 19.90 100.5 29.69 103.4 30.03 108.3 30.50 110.2 30.56 112.0 30.67
34 65.5 15.06 71.4 15.58 95.3 21.14 100.5 30.90 103.0 31.20 103.0 31.67 108.6 31.69 109.9 31.73
35 65.6 15.55 71.4 16.07 95.3 21.77 100.5 31.50 102.8 31.78 102.8 31.78 107.8 32.25 108.8 32.25

Appendix 2

Table 8 Operating performance of the existing ACMV system [48]

Remarks: Q = Cooling Capacity; SC = Sensible Cooling Load; P = Power Input; To = Outdoor Temperature; TI = Indoor Temperature

TI/TO 19 25 30 35 40 46
Q SC P Q SC P Q SC P Q SC P Q SC P Q SC P
24 8.21 4.83 2.00 7.91 4.71 2.16 7.61 4.59 2.34 7.30 4.46 2.54 6.70 4.16 2.75 6.19 3.93 3.03
27 8.23 5.59 2.00 7.94 5.46 2.16 7.64 5.33 2.34 7.33 5.20 2.54 6.73 4.86 2.75 6.22 4.60 3.04
30 8.30 6.87 2.00 8.01 6.72 2.17 7.72 6.57 2.35 7.42 6.41 2.55 6.84 5.98 2.76 6.35 5.64 3.05
33 8.48 8.48 2.01 8.23 8.23 2.18 7.98 7.98 2.36 7.72 7.72 2.57 7.15 7.15 2.79 6.68 6.68 3.08
Heat Mass Transfer (2020) 56:671–690 687

Appendix 3

Table 9 Components used in the TRNSYS simulation

688 Heat Mass Transfer (2020) 56:671–690
Heat Mass Transfer (2020) 56:671–690 689

Fig. 12 Simulated result of the

VRF system based on Zhou’s
study [41]

References 8. Radhakrishnan N, Su Y, Su R, Poolla K (2016) Token Based

Scheduling for Energy Management in Building HVAC Systems.
Appl Energy 173:67–79
1. Hong T, Sun K, Zhang R, Hinokuma R, Kasahara S, Yura Y (2016)
Development and Validation of a New Variable Refrigerant Flow 9. United Nations Development Programme Achieving Industrial
System Model in EnergyPlus. Energy and Building 117:399–411 Energy Efficiency in Malaysia. 2006. UNDP: Malaysia
2. Zhu Y, Jin X, Fang X, Du Z (2014) Optimal Control of Combined 10. Hassan JS, Zin RM, Majid MZA, Balubaid S, Hainin MR (2014)
Air Conditioning System with Variable Refrigerant Flow and Building Energy Consumption in Malaysia: An Overview. Jurnal
Variable Air Volume for Energy Saving. International Journal of Teknologi 70(7):2180–3722
Refrigeration-Revue Internationale Du Froid 42:14–25 11. Özahi E, Abuşoğlu A, Kutlar Aİ, Dağcı Ŏ (2017) A Comparative
3. Moon HJ, Yang SH (2014) Evaluation of the Energy Performance Thermodynamic and Economic Analysis and Assessment of a
and Thermal Comfort of an Air Conditioner with Temperature and Conventional HVAC and a VRF System in a Social and Cultural
Humidity Controls in a Cooling Season. HVAC&R Research 20: Center Building. Energy and Buildings 140:196–209
61–71 12. Liu C, Zhao T, Zhang J, Chen T, Li X, Xu M, Yang X (2015)
4. Kim D, Cox SJ, Cho H, Im P (2017) Evaluation of Energy Savings Operational Electricity Consumption Analyze of VRF Air
Potential of Variable Refrigerant Flow (VRF) from Variable Air Conditioning System and Centralized Air Conditioning System
Volume (VAV) in the U.S. Climate Locations. Energy Rep 3:85–93 Based on Building Energy Monitoring and Management System.
5. Yahaya AZ, Boon TT, Yusmady MA (2015) Energy Analysis on Procedia Engineering 121:1856–1863
ACMV System for an Academic Building: Case Study in UTeM, in 13. Aynur TN (2010) Variable Refrigerant Flow Systems: A Review.
Mechanical Engineering Research Day 2015. Universiti Teknikal Energy and Buildings 42:1106–1112
Malaysia Melaka, Melaka 14. Bhatia A (2012) PDH Course M394: HVAC Variable Refrigerant
6. Saab R, Quabeh HA, Ali MIH (2018) Variable Refrigerant Flow Flow Systems. PDHonline; Available from: http://www.pdhonline.
Cooling Assessment in Humid Environment Using Different com/courses/m394/m394content.pdf
Refrigerants. J Environ Manag 224:243–251 15. Goetzler W (2007) Variable Refrigerant Flow Systems. ASHRAE J
7. Zhang Y, Wei Z, Long E, Zhang X, Guo S (2017) Outdoor Air 49(4):24–31
Thermal Plume Simulation of Layer - Based VRF Air 16. Thornton B, Wagner A (2012) Variable Refrigerant Flow Systems.
Conditioners in High - Rise Buildings. Energy Procedia 142: Prepared for General Services Administration by Pacific Northwest
3787–3792 National Laboratory: United State. pp. 1–68
690 Heat Mass Transfer (2020) 56:671–690

17. Liu X, Hong T (2010) Comparison of Energy Efficiency between 34. Elhelw M (2016) Analysis of Energy Management for Heating,
Variable Refrigerant Flow Systems and Ground Source Heat Pump Ventilating and Air-conditioning Systems. Alexandria
Systems. Energy and Building 42:584–589 Engineering Journal 55:811–818
18. Li Z, Wang B, Li X, Shi W (2016) Simulation on Effects of 35. University of Alabama. ME 416/516 Energy Calculations of Large
Subcooler on Cooling Performance of Multi-Split Variable Buildings. Lecture PowerPoint slides; Available from: www.me.ua.
Refrigerant Flow Systems with Different Lengths of Refrigerant edu/me416/LECTURE%20MATERIALS/energy%20cal%20for%
Pipeline. Energy and Building 126:301–309 20large%20buildings.pdf
19. Wang S (2014) Energy Modeling of Ground Source Heat Pump vs. 36. Papakostas K, Bentoulis A, Bakas V, Kyriakis N (2007) Estimation
Variable Refrigerant Flow Systems in Representative US Climate of Ambient Temperature Bin Data from Monthly Average
Zones. Energy and Building 72:222–228 Temperatures and Solar Clearness Index. Validation of the
20. Yildiz A, Ersöz MA (2015) Determination of the Economical Methodology in Two Greek Cities. Renew Energy 32:991–1005
Optimum Insulation Thickness for VRF (Variable Refrigerant 37. Wang Z, Ding Y, Geng G, Zhu N (2014) Analysis of Energy
Flow) Systems. Energy and Building 89:835–844 Efficiency Retrofit Schemes for Heating, Ventilating and Air-
21. Li YM, Wu JY (2010) Energy Simulation and Analysis of the Heat Conditioning Systems in Existing Office Buildings Based on the
Recovery Variable Refrigerant Flow System in Winter. Energy and Modified Bin Method. Energy Convers Manag 77:233–242
Building 42:1093–1099 38. Yau YH, Wong CM, Ong HC, Chin WM (2018) Significance of the
22. Kwon L, Hwang Y, Radermacher R, Kim B (2012) Field study of Bin Weather Data for Energy Analysis in the Tropics.
Performance Measurements of a VRF System with Sub - Cooler Unpublished Manuscript
in Educational Offices for the Cooling Season. Energy and 39. Chhetri KB (2012) Computation of Errors and Their Analysis on
Building 49:300–305 Physics Experiments. The Himalayan Physics 3:78–86
23. Afify R (2008) Designing VRF Systems. ASHRAE J 50(6):52–55 40. Department of Physics & Astronomy (2015) Averaging, Errors and
24. Karr M (2011) Ground - Source Variable Refrigerant Flow Heat Uncertainty. In Lab Manual: Undergraduate Labs. University of
Pumps: A Solution for Affordable Housing, Assisted Living, Pennsylvania: Pennsylvania
Hotels and Dorms. Washington State University Extension 41. Zhou YP, Wu JY, Wang RZ, Shiochi S (2007) Energy Simulation in
Energy Program, pp. 1–7 the Variable Refrigerant Flow Air - Conditioning System Under
25. Alshatti RA (2011) Analyses of Variable Refrigerant Flow and Cooling Condition. Energy and Building 39(2):212–220
Exergy in Air Conditioning Systems. In: Mechanical Engineering. 42. Zhang S, Jiang Y, Xu W, Liu Z, Sun D (2017) Key Prescriptive
University of South Florida: South Florida United State. pp. 1–73 Parameters Analysis of the New China Building Energy Code
26. Krarti M (2011) Energy Analysis Tools. In: Kreith F (ed) Energy Based on Saving to Investment Ratio Methodology. Indoor and
Audit of Building Systems: An Engineering Approach. CRC Press, Built Environment 26(1):78–91
New York, pp 4-1–4-14 43. Teke A, Timur O, Zor K (2015) Calculating Payback Periods for
27. Jin Z, Yezheng W, Gang Y (2006) A Stochastic Method to Generate Energy Efficiency Improvement Applications at a University
Bin Weather Data in Nanjing, China. Energy Convers Manag 47: Hospital. Çukurova University Journal of the Faculty of
1843–1850 Engineering and Architecture 30(1):41–56
28. Papakostas K, Tsilingiridis G, Kyriakis N (2008) Bin Weather Data 44. Tenaga Nasional Berhad (2018) Billing. Available from: https://
for 38 Greek Cities. Appl Energy 85:1015–1025 www.tnb.com.my/residential/billing
29. Eguía P, Granada E, Alonso JM, Arce E, Saavedra A (2016) 45. Leite GNP, Weschenfelder F, Araújo AM, Ochoa ÁAV, Neto NFP,
Weather Data Sets Generated Using Kriging Techniques to Kraj A (2019) An Economic Analysis of the Integration between
Calibrate Building Thermal Simulations with TRNSYS. Journal Air - Conditioning and Solar Photovoltaic Systems. Energy
of Building Engineering 7:78–91 Convers Manag 185:836–849
30. Megri AC (2014) TRNSYS as an Education Tool to Predict Indoor 46. Poh KS (2019) Personal Communication
Environment Temperature for Undergraduate Students. ASEE 47. Poh KS (2016) Personal Communication
Annual Conference & Exposition, Indianapolis 48. Wong CM (2016) Personal Communication
31. Shrivastava RL, Kumar V, Untawale SP (2017) Modeling and 49. American Society of Heating, Refrigerating, and Air Conditioning
Simulation of Solar Water Heater: A TRNSYS Perspective. Engineers (ASHRAE) (2014) ASHRAE Pocket Guide for Air
Renew Sust Energ Rev 67:126–143 Conditioning, Heating, Ventilation, Refrigeration, 8th edn.
32. Solar Energy Laboratory (2006) Getting Started. In in TRNSYS 16: ASHRAE, Atlanta
A Transient System Simulation Program. University of Wisconsin- 50. Çengel YA, Boles MA (2014) Thermodynamics: An Engineering
Madison, Madison, pp. 7–87 Approach, vol 7. McGraw-Hill Education, New York
33. Lin X, Lee H, Hwang Y, Radermacher R (2015) A Review of
Recent Development in Variable Refrigerant Flow Systems. Publisher’s note Springer Nature remains neutral with regard to
Science and Technology for the Built Environment 21:917–933 jurisdictional claims in published maps and institutional affiliations.