See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/330081729

Comparative Performance Analysis of Induction & Synchronous Reluctance

Motors in Chiller Systems for Energy Efficient Buildings

Article  in  IEEE Transactions on Industrial Informatics · August 2019

DOI: 10.1109/TII.2018.2890270

CITATIONS READS

0 900

2 authors:

Felipe Oliveira Abhisek Ukil

University of Auckland University of Auckland
2 PUBLICATIONS   1 CITATION    189 PUBLICATIONS   1,686 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Voltage Sensitivity Analysis and Demand Dispatch Option of Electric Vehicle in Smart Grid View project

Microgrid View project

All content following this page was uploaded by Abhisek Ukil on 02 January 2019.

The user has requested enhancement of the downloaded file.


Comparative Performance Analysis of Induction &
Synchronous Reluctance Motors in Chiller Systems
for Energy Efficient Buildings
Felipe Oliveira, Member, IEEE, Abhisek Ukil, Senior Member, IEEE
Abstract—A significant portion (ca. 40%) of the world’s
total energy is utilized in buildings. Cooling of large buildings
worldwide is done by centrifugal chiller systems, where the
compressors, pumps, fans are driven by AC motors. These
motors are energy hungry parts of the chiller. All-variable
speed chillers, taking into account load variations, reduced the
energy demand significantly, compared to the constant speed
ones. Induction motors (IMs) are commonly used in modern
chillers. However, as the efficiency of the IM decreases with
speed, the overall efficiency drops, as the chiller operates most
of the time in part-load condition. In this paper, SimulationX
software is used to quantitatively investigate about the energy
efficiency improvement, by using synchronous reluctance motor
(SynRM) in all-variable speed centrifugal chillers. The system
was tested for three distinct building load profiles. The results
are judged in comparison with the state-of-the-art IE2 IMs. The
results provide quantitative evidence that synchronous reluctance
motor can significantly increase the energy efficiency in the chiller
systems and has a big potential to substitute induction motors in
these applications.
Index Terms—Building energy efficiency, centrifugal chiller,
compressor, fan, HVAC, induction motor, pump, smart grid,
synchronous reluctance motor.
I. I NTRODUCTION
WO-thirds of world’s electrical energy is utilized by
T electric motors [1]. A significant portion (ca. 40%) of
the world’s total energy is utilized in buildings [2]. Heat-
ing, Ventillation, and Air-Conditioning (HVAC) system is the
highest eletrical load (ca. 30–50%) in the buildings [3]–[5].
Components of the HVAC system, in particular the chillers,
namely, compressors, pumps, fans are motor driven. Therefore,
improving the energy efficiency of the fans, compressors and
pumps in the HVAC system is of particular importance, with
worldwide increased focus on building energy management
[6]–[9]. In this paper, we focus on the variable-speed cen-
trifugal chiller part of the HVAC system.
Modern chiller systems are typically driven by the induction
motor (IM)-based variable speed drive (VSD) [10]–[12]. Syn-
chronous reluctance motor (SynRM) is a relatively new high-
efficiency motor technology. It is being proposed for HVAC
and chiller applications [13]. In this paper, SynRM is exten-
sively investigated for chiller applications, in comparison with
F. Oliveira was Masters student in the Dept. of ECE. Univ. of
Auckland, 20 Symonds Street, Auckland, 1010, New Zealand, (email:
foli382@aucklanduni.ac.nz).
A. Ukil (corresponding author) is Senior Lecturer in the Dept. of ECE,
Univ. of Auckland, 20 Symonds Street, Auckland 1010, New Zealand, (email:
a.ukil@auckland.ac.nz).
1
the state-of-the-art high-efficiency IMs. Quantitative analysis
of energy savings in this kind of application is also presented.
SimulationX [14] software is used to develop the system model
of the all-variable speed centrifugal chiller system, along with
the building. Different building load profiles covering various
load utilization scenarios are selected for the energy consump-
tion simulations, at different ambient conditions. Comparative
results for the IM and the SynRM are presented and analyzed.
A. Contributions
The paper has the following main contributions. Firstly,
we give an overview of the centrifugal chiller systems, and
SynRM-based VSD, which is a promising new technological
development, compared to the traditional IM. Secondly, we
show relevant case-studies of different energy consumption
scenarios in the centrifugal chiller systems, demonstrating that
the SynRM-based VSD can achieve significantly higher energy
efficiency than the traditional IM.
B. Organization
The paper is organized as follows. Section II describes
the background information on the variable speed centrifugal
chiller, chiller efficiency, VSD, with other control methods,
and SynRM. Detailed modeling of the chiller systems along
with IM- and SynRM-based VSDs using the SimulationX [14]
software are described in section III. Results are presented
in section IV. Discussions on the result are presented in
section V. Finally, conclusions are provided in section VI.
II. BACKGROUND I NFORMATION
A. Variable Speed Centrifugal Chillers
Cooling of large buildings worldwide is done by centrifugal
chiller systems [11]. In brief, the heat from the buildings is
transferred to the chilled water, which is pumped through the
building, and then transferred to the refrigerant fluid in the
chiller. The refrigerant fluid goes through a compression cycle,
and transfers the heat to the cooling water. The cooling water
is finally pumped to a cooling tower, which blows atmospheric
air to the cooling water, to be refrigerated and to enter in the
process again, as shown in Fig. 1.
Large buildings have few chiller units installed, each chiller
unit containing several pumps and fans. The chiller system is
typically designed to provide the maximum cooling demand
during the hottest summer times [17], which, however, take
 2

Fig. 1: Schematic diagram of the centrifugal chiller system [?].

place only for a short duration. For majority of the time, the
chiller system does not run at full-load, rather in substantially
partial loading condition, especially during the winter. Effi-
cient control for dynamic operation of the chiller system can
significantly improve the energy efficiency, reducing energy
wastage due to over/under-cooling [10].
The flow control to achieve the required part-load conditions
can be implemented in different ways, e.g., with dampers,
throttles and bypasses, which can reduce the flow but do not
minimize the power consumption [15]. On/off cycle control,
using different devices to vary the chilled/cooling water flow
can to some extent reduce the electricity consumption [11].
The electrical power, the energy follows the affinity laws [16],
with a cubic relationship with the speed reduction (see Fig. 2).
Hence, the speed control is the most efficient way to control
the flow and reduce the energy consumption of pumps, fans
and compressors. A key point is that the motors should be
able to flexibly operate by reducing their speed in the part-
load condition, which is defined in the following section.
Fig. 2: Different controls for the variable-speed centrifugal
B. Efficiency of Variable Speed Centrifugal Chiller chiller system.
Coefficient of performance (COP), defined by the ratio of
the heat removed from the water by the evaporator (Qc ) over
the compressor mechanical work energy (W ), measures the
centrifugal chiller efficiency,
Qc
COP = . (1)
W
The ASHRAE COP [17] classification is shown in Fig. 3. An
excellent full-load COP value is considered to be over 5.0.
ARI standard 550/590 [18] defines the Integrated Part Load
Value (IPLV). IPLV is used to measure chiller performance in Fig. 3: ASHRAE COP classification.
the partial loading condition.
IP LV = 0.01A + 0.42B + 0.45C + 0.12D, (2) C. Variable Speed or Frequency Drive
where, the parameter A, B, C, D denote COP at 100%, 75%, The speed control of the electric motor is performed by the
50%, and 25% full-load condition respectively. From eq. (2), VSD, also known as, the variable frequency drive (VFD). 5%
it can be noticed that the calculation of the IPLV considers to 8% energy savings is reported for cooling fans using VSD
different weights for each part-load situation, the 75% and [19]. Applications of the VFD for controlling the primary-flow
50% being the most influencing ones. pumps resulted in overall energy savings ca. 2% and 5% [20].
 3

Following load-driven control methods and equal marginal this means that motor efficiency will decrease, resulting in
performance principle (EMPP), Hartman [17] demonstrated unwanted energy losses. Motor technology that can improve
28% energy savings for all-variable speed chillers. Reduction the efficiency under part-load condition, would be a huge
in annual electricity and water usage by ca. 19%, and 16% benefit to improve the energy efficiency at industrial scale.
could be achieved by adopting load-specific control of speed IEC 60034-30-1:2014 standard [21] defines the AC motor
in the chillers [16]. Energy savings of 30% could be achieved efficiency class. Fig. 4 shows the different classes, IE1–IE4,
in Taiwan’s industrial buildings by using VSDs [10]. which are known as ‘Standard,’ ‘High-efficiency,’ ‘Premium,’
‘Super Premium,’ and ‘Ultra Premium’. Standard IMs are
D. Comparison of Different Control Methods classified in IE2 or IE3, while, SynRM comes with higher
efficiency, IE4. With no permanent magnet, reduced torque
Following the discussion on the VSD, a quantitative com-
and current ripples, SynRM is aimed to replace in near future
parison of the different control methods (see Fig. 2) are
most of the AC motors [21].
described in this section. Referring to the Fig. 2, using curve-
fitting, we compute the input power requirements (Pinp in %)
for the different control methods, at the four loading conditions
in the IPLV definition (see eq. (2)). The values are shown in
Table I. For 25% loading condition, the values are calculated
by extrapolation of the fitted curves.
Considering the pump load requirement as the required
output, and using the values from Table I, we compute the
efficiency (η) of the different control methods at the four
loading conditions in the IPLV definition. The values are
shown in Table II. For example, at 25% load, VSD η =
Pinp (P ump)/Pinp (V SD) = 6.9/10 = 69%.
From Table II, it can be seen that most of the control
methods have relatively high efficiency at high load. How-
ever, except the VSD, all other control methods have poor
performance in terms of efficiency at low loading conditions, Fig. 4: IE efficiency classes for 4 pole motors at 50 Hz [21].
especially below 75%. At extremely low load like 25%, the
Specially designed rotor provides the maximum reluctance
drop in efficiency is significant. This demonstrates that VSD is
path in the SynRM between the direct (d) and the quadrature
the most optimal control method for variable speed centrifugal
(q) axes. The stator design is similar to standard IM. Based on
chiller loads.
the maximum reluctance, the mechanical torque is generated.
TABLE I: Different Control Methods: Input Power Compari-
With no rotor windings, SynRM has no rotor losses, compared
son
to other AC motors [22], [23]. This makes the SynRM
% of Pump VSD Mag. ON/ Throttle By- inherently more efficient than same rated IM [24], [25].
Full Coup. OFF pass
Speed Pinp Pinp Pinp Pinp Pinp Pinp
IM is a mature technology, developing over last 100 years,
(%) (%) (%) (%) (%) (%) while SynRM is commercially available from 2011 [26].
25 6.9 10 17 25 60 101 Being a new technology, feasibility of the SynRM in energy
50 13 18 34 50 79 101 efficiency applications, designs, control strategies, are not fully
75 43 47 63 75 95 101
100 100 105 103 100 104 101 proven, notwithstanding its huge potential from the theory.
Comparative views of the constructional details and losses of
the IM and SynRM are shown in Fig. 5 [26].
TABLE II: Different Control Methods: Comparison of Effi-
Increasing research works are being reported for SynRM.
ciency
Foo and Zhang [27] presented robust field-weakening algo-
% of VSD Mag. ON/ Throttle By- rithm for directly controlling the torque in SynRM. Bierhoff
Full Coup. OFF pass
Speed η η η η η
presented simple phase-locked loop (PLL) type algorithm for
(%) (%) (%) (%) (%) controlling the SynRM in [28]. Trancho et. al. [29] presented
25 69.00 40.59 27.60 11.50 6.83 sliding mode control and hybrid field weakening control for
50 72.22 38.24 26.00 16.46 12.87 the SynRM. Hadla and Cruz [30] used model predictive
75 91.49 68.25 57.33 45.26 42.57
100 95.24 97.09 100.0 96.15 99.01 control (MPC) strategy for SynRM.

III. M ODELING AND S IMULATION

E. Synchronous Reluctance Motors
High-efficiency induction motors are the workhorses in the
chiller systems. However, the short-comings of the IMs is
that the efficiency decreases significantly as the speed drops.
As chillers will run most of the time in part-load condition,
A. Modeling of Variable Speed Centrifugal Chiller
Fig. 6 shows a variable speed centrifugal chiller system,
simulated using the SimulationX [14] software. We have con-
sidered single chiller system. Detailed system specifications
are provided in Table III. Fig. 7 shows the actual viewing of
 4

The ambient temperature was kept at 25◦ C, which is typical

average value in Auckland, New Zealand. Besides the official
weather data [31],[32], this particular ambient temperature was
chosen following discussion with practicing authorities, e.g.,
building facility managers. We have also considered a higher
ambient temperature (35◦ C) for completeness of the results.
However, the change in the power required by the IM and the
SynRM would expectedly happen in similar fashion, our work
being focused on comparative analysis of the energy efficiency.
Fig. 5: Comparison of construction and losses of IM and
SynRM [26]. B. Modeling of Motors
Based on ABB Company data sheets [26], IE2 IMs and
the whole system, controls, and models implemented in the
IE4 SynRMs are modeled, considering the power and speed
SimulationX.
requirements for the required specifications. For both types of
The complete cooling system usually includes more than
motors, ABB efficiency figures (includes motor and VFD) are
one set of cooling tower, chiller and pumps. In fact, most of
used for uniform comparison.
the large systems contain two or more individual sets operating
Instantaneous torque and speed dictate the efficiency for
in parallel to achieve the maximum load. In a system with three
both the IE2 IM and SynRM. Fig. 8 shows the considered
chillers operating in parallel, when the load is lower than 66%
measurement points to assess the efficiency. Fig. 9 shows
of the maximum capacity, one set is switched off, while the
the efficiency of IM and SynRM and IM in pump and fan
other two feed the load. If the load goes lower than 33%, one
applications (200 kW). From Fig. 9, we can notice that the
more set is switched off, with just the last one feeding the
pump/fan power varies cubically with the speed, and SynRM
remaining load [17].
has better efficiency profile than the IM, especially in the low
However, the main objective of our work is to comparatively
load regime. Fig.10 shows the example of SynRM efficiency
analyze the performances of the IE2 IM against the SynRM for
map for any load characteristic, showing a rather uniform
centrifugal chiller application. Therefore, we have considered
profile, across different load torque and speed regions.
single chiller system, with different motor configurations,
focusing on the comparative energy efficiency analysis. With
multiple chiller systems, the energy analysis figures will be
C. Motor Control Methods
multiplicative.
TABLE III: System Details for Simulation A signal (0 to 1) is sent to the chiller sub-systems (compres-
sor, pump, fan). This signal is multiplied by the rated speed,
Chiller Parameters Full-load Values which allows to vary the speed proportionally to the full speed
Cooling Capacity (kW) 1183.69
Compressor Power (kW) 200 simultaneously for the three sub-systems. For ensuring the
Compressor Speed (RPM) 4000 adequeate flow of water into the heat exchangers, the pump
COP (full-load) 5.92 speed must attain 50% of full speed, which is a minimum
Refrigerant type R410a
Cooling Pump Parameters Full-load Values constraint [11]. Fig. 11 shows almost linear correlation be-
Cooling Pump Power (kW) 30 tween the control signal and the cooling capacity. However,at
Cooling Pump Speed (RPM) 1500 low-load region, there is some variations.
Cooling Water Volume Flow (l/s) 60
Cooling Water Mass Flow (kg/s) 62.17 For the VSD of the IM and SynRM, sensorless vector
Entering Cooling Water Temperature (◦ C) 35.18 control method is applied, namely direct torque control (DTC).
Leaving Cooling Water Temperature (◦ C) 40.96 DTC is well-established universal method to control the torque
Cooling Tower Parameters Full-load Values
Cooling Tower Heat Rejection (kW) 1407.23 and the speed of the IM [33], [34], as well as the SynRM
Fan Power (kW) 30 [27]. In DTC scheme, the motor voltage and current are
Fan Speed (RPM) 3000 measured. These values areused to estimate the torque and
Air Volume Flow (m3 /s) 78.96
Air Water Mass Flow (kg/s) 87.86
the magnetic flux. The estimated values are then compared
Chiller Water Parameters Full-load Values with the reference values. Based on the deviation of the flux
Chilled Water Volume Flow (l/s) 45 and torque, the transistors in the VSD are turned on or off,
Chilled Water Mass Flow (kg/s) 45
Entering Chilled Water Temperature (◦ C) 12.66
typically using the pulse width modulation (PWM) technique.
Leaving Chilled Water Temperature (◦ C) 6.38 A switching table is used for the flux and torque control in
DTC. Fig. 12 shows the algorithm.
The simulated chiller system has centrifugal compressor DTC is very fast-acting control method for torque and flux.
with VSD, one condenser unit with expansion valve, and an In commercial DTC-based VSD, the control cycle is in the
evaporator with shell and tube heat exchangers. Variable speed order of 10-30 µs. The switching losses are very low, as the
pump drives the cooling water loop through the condenser transistors are turned only when needed to keep the torque and
to the cross-flow type cooling tower. A constant speed pump flux within the hysteresis bands. The switching frequency of
drives the chilling water loop on the evaporator side. the transistors is not constant.
 5

Fig. 6: Simulated variable speed centrifugal chiller system.

Fig. 7: SimulationX model.

Fig. 8: Efficiency measurement points.

D. Modeling of Load Profiles
To run the comparisons, three different annual load profiles
were selected, namely commercial building in Hawaii (see Fig.
13) [35], office building 1 in Hongkong, China (see Fig. 14)
[36], and office building 2 in Hongkong, China (see Fig. 15)
[16]. The load profiles were selected to incorporate the part-
load conditions, as described in eq. (2).
These three different load profiles are very interesting for
this analysis because they have completely distinct charac-
teristics from each other. Load profile 2 (see Fig. 14) is
predominantly high loads, and load profile 3 (see Fig. 15) is
predominantly low loads, while load profile 1 (see Fig. 13) is
Fig. 9: 200 kW package efficiency curves in pump/fan duty
[13].
a kind of ‘bell shaped curve’ distribution, with more medium
loads and less extreme loads.
In order to achieve the component-wise energy consump-
tion, we first multiply the efficiency curve with the motor
power curve, followed by a time-integral of the result. The
energy consumptions are calculated at two different ambient
temperature conditions, namely at 25◦ C and 35◦ C.
 6

TABLE IV: Energy Performance Result for Load Profile-1 (at 25◦ C ambient temperature)
Load Profile-1 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 1653.19 1601.67 -3.12% 1592.66 -3.66%
Consumption (MWh)
Average 88.95% 91.81% +3.22% 92.33% +3.80%
Efficiency
Annual 182.71 131.19 -28.20% 122.18 -33.13%
Lossess (MWh)

TABLE V: Energy Performance Result for Load Profile-2 (at 25◦ C ambient temperature)
Load Profile-2 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 624.95 609.13 -2.53% 605.71 -3.08%
Consumption (MWh)
Average 90.39% 92.74% +2.60% 93.26% +3.18%
Efficiency
Annual 60.05 44.22 -26.36% 40.80 -32.06%
Lossess (MWh)

TABLE VI: Energy Performance Result for Load Profile-3 (at 25◦ C ambient temperature)
Load Profile-3 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 381.37 365.98 -4.04% 363.70 -4.63%
Consumption (MWh)
Average 86.44% 90.08% +4.21% 90.65% +4.87%
Efficiency
Annual 51.69 36.30 -29.77% 34.02 -34.18%
Lossess (MWh)

TABLE VII: Energy Performance Result for Load Profile-1 (at 35◦ C ambient temperature)
Load Profile-1 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 3185.7 3102.43 -2.61% 3080.2 -3.31%
Consumption (MWh)
Average 87.17% 90.89% +4.27% 91.41% +4.86%
Efficiency
Annual 352.1 254.11 -27.83% 236.29 -32.89%
Lossess (MWh)

TABLE VIII: Energy Performance Result for Load Profile-2 (at 35◦ C ambient temperature)
Load Profile-2 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 1204.28 1179.88 -2.03% 1171.44 -2.73%
Consumption (MWh)
Average 88.58% 91.81% +3.65% 92.33% +4.23%
Efficiency
Annual 115.72 85.65 -25.99% 78.91 -31.81%
Lossess (MWh)

TABLE IX: Energy Performance Result for Load Profile-3 (at 35◦ C ambient temperature)
Load Profile-3 IE2 IM SynRM Difference (%) SynRM Difference (%)
(All equipments) (Compressor only) (Col. 3 & 2) (All equipments) (Col. 5 & 2)
Annual 734.9 707.81 -3.69% 703.4 -4.29%
Consumption (MWh)
Average 84.71% 89.18% +5.28% 89.74% +5.94%
Efficiency
Annual 99.61 70.20 -29.53% 65.79 -33.95%
Lossess (MWh)
 7

Fig. 12: Direct torque control algorithm for IM and SynRM.

Fig. 10: 200 kW SynRM efficiency map for all applications.

Fig. 13: Load profile 1 for commercial building in Honululu,

Hawaii [17].

Fig. 11: Speed control curve.

3 yields the highest energy savings, 4.63%, and on the load
profile 1 giving 3.66% of savings, w.r.t. using the IE2 IM.
IV. R ESULTS Load profile 2 yields 2.53% (compressor only) and 3.08% (all
Due to the different load profiles and to the fact that they equipments) of energy savings, the lowest savings among the
were scaled to fit the maximum capacity of the designed three profiles. These differences were expected, and they hap-
simulation system, the result analysis consists of comparisons pen because the difference in efficiency between the SynRM
of the percentage of the electricity consumed using SynRM and the IE2 IM tends to increase, as the speed decreases.
with respect to (w.r.t) the IE2 IM. Firstly, simulations are done Therefore, systems that are operating highly concentrated on
for all equipments using IM and SynRM in a comparative low cooling loads tend to present higher energy savings, as is
manner. Besides, simulations are performed, using the SynRM the case of load profile 3. On the other hand, the load profile
only to run the compressor, since it is the component with the 2 presented the lowest energy savings, because it runs over
highest power in a centrifugal chiller system, while using the
IE2 IM on the other devices, to compare the energy savings
with the previous case.
Tables IV–VI, show the detailed figures for the annual
energy consumption (in MWh), average efficiency and annual
losses (in MWh) for the load profiles 1–3, at 25◦ C ambient
temperature, and Tables VII–IX, show the similar figures for
the load profiles 1–3, at 35◦ C ambient temperature. The annual
energy consumption for each of the load profiles at 25◦ C
and 35◦ C ambient temperatures are presented graphically
in Figs. 16 and 17 respectively, considering 100% of the
energy consumed by the IE2 IM (subplot i), and the relative
consumption when using SynRM applied to only compressor
(subplot ii), and all speed-controlled machines (subplot iii).
The results in the three different profiles are very significant. Fig. 14: Load profile 2 for office building in Hongkong, China
At 25◦ C ambient temperature, using SynRM in load profile [17].

Fig. 15: Load profile 3 for office building in Hongkong, China
[17].
80% of the total cooling load more than 50% of the year.
Similar trend can be observed at 35◦ C ambient temperature,
load profile 3 giving highest energy savings of 4.29%, while
load profile 1 and 2 result in 3.31% and 2.73% respectively.
Despite the differences in the load profiles, simulating
various operating conditions, energy savings between 2.53%
and 4.63% are quite significant. For the designed system of
these simulations, operating with the selected load profiles, the
calculated energy savings can be as high as 60,530 kWh/year
(e.g., for load profile-1).
Fig. 16: Comparison of annual energy consumption at 25◦ C
using (i) IE2 IM vs. SynRM in the (ii) compressor only, and
(iii) all machines.
With the use of SynRM only in the compressor, the total
energy savings reduced a little when compared to using
SynRM in all machines. The savings, in this case, are 3.12%
for load profile 1, 2.53% for load profile 2, 4.04% for load
profile 3, at 25◦ C ambient temperature. At 35◦ C ambient
temperature, the savings are 2.61% for load profile 1, 2.03%
for load profile 2, 3.69% for load profile 3. The ranking of
the savings remained the same among the three profiles, and
the difference in the savings can be explained likewise the
previous case.
Even if the SynRM is used only in the compressor, the
total energy savings are still relevant. The compressor is
the machine that consumes the highest amount of power in
8
Fig. 17: Comparison of annual energy consumption at 35◦ C
using (i) IE2 IM vs. SynRM in the (ii) compressor only, and
(iii) all machines.
a chiller system. Therefore, any improvements that increase
the efficiency of the compressor drive system is expected to
significantly influence the overall energy consumption.
For all the load profiles, using the SynRM, annual losses
are highly reduced as well. The annual losses are reduced
between 32%–34%, when SynRM is used in all equipments,
and between 26%–29%, when SynRM is used only in the
compressor. This is inherently due to the high-efficiency of
the SynRM compared to the IE2 IM, as discussed in section
II-E. Figs. 18 and 19 show the efficiencies of the different
motors on common X-Y axes for easy comparison, for IE2 IM
(all equipments), SymRM (in compressor) with IE2 IM (other
equipments), and SynRM (all equipments), at 25◦ C and 35◦ C
respectively.
Fig. 18: Comparison of motor efficiency at 25◦ C: (i) IE2
Induction Motor (ii) SynRM & IE2 IM , and (iii) Only SynRM.
V. D ISCUSSIONS
Discussion on the results are provided as follows.
1) Relative to the time in which this paper is written, the
comparison of costs (in US$) of the SynRM and the

Fig. 19: Comparison of motor efficiency at 35◦ C: (i) IE2
Induction Motor (ii) SynRM & IE2 IM, and (iii) Only SynRM.
TABLE X: Comparison of Costs: SynRM vs IM
Application Power SynRM IE2 IM
Type Rating Approx. Cost Approx. Cost
(US$) (US$)
Compressor 200 kW 11,500 8,850
Cooling Tower Fan 30 kW 1,980 1,840
Pump 30 kW 1,930 1,320
IE2 IM for applications in the different parts of the
chiller are presented in Table X. The figures shown in
Table X are average figures from Europe and Asian
markets, and do not signify any relative measure for any
manufacturers. From Table X, it can be observed that for
the compressor, which is the highest power consuming
part in a centrifugal chiller system, the SynRM cost is
approximately US$ 2650 more than the IM. On the other
hand, for parts with lower energy consumptions, e.g., the
cooling tower fan and pump, the SynRM is slightly more
expensive than the IM, in the range of US$ 150-600. IE2
IM is commonly used in pumps, which makes it rather
cheap. Currently, IM has a larger market share in the
chiller applications. However, the cost comparison shows
that there is not significant difference in the costs of the
two types of motors, especially when SynRM can result
in greater energy savings, as demonstrated in this paper.
With increased applications and market share, the costs
of SynRM would expectedly reduce in near future, as
often the manufacturing costs depend on the volume of
production.
2) The energy savings results are based on single chiller
system in this work. In practice, professional cooling
system usually includes more than one set of cooling
tower, chiller and pumps, as discussed in section III.
Therefore, based on the reported promising energy sav-
ings, the energy savings figures will be multiplicative
in nature with multiple chiller systems. This is of great
promise to improve the overall energy efficiency at larger
9
scale, e.g., big industrial plants, district cooling in big
shopping malls, buildings, datacentres, etc. Attaining en-
ergy efficiency at larger scale is often the interest for
the industry, city council, and government authorities for
significant cost-benefit analysis.
3) As discussed is section II-D, speed control using VSD
is most optimal for centrifugal chiller applications. And
accordingly, as discussed in section III-C, for the VSD
of the IM and SynRM, sensorless vector control method
is applied, implementing the DTC. Therefore, the control
performance of the VSD remains identical for IE2 IM
and SynRM, for a fair comparison.
4) Besides implementing the SynRM technology, advanced
dynamic control methods would be helpful. For example,
advanced multi-sensor approach [37], taking care of dy-
namic occupancy and ambient conditions, together with
multi-variable (e.g., temperature, air-flow rate) control
methods [8], [9], [38] could also be incorporated in future
works. In conjunction with the SynRM, these dynamic
optimizations would further increase the energy savings
in the chiller applications.
5) New Zealand has a strong focus on green energy and
a sustainable low carbon future. Over 180 countries
haven undertaken responsibility for carbon emission con-
trol, following The Paris agreement [39]. Buildings with
chiller and HVAC systems, has a significant share of
world’s total energy. Thus, it is a great sector to optimize
for larger energy efficiency. The quantitative results re-
ported in this work, should promote adopting more energy
efficient solutions like the SynRM in coming days.
VI. C ONCLUSION
Cooling of large buildings worldwide is done by centrifugal
chiller systems, consuming about 30–50% of energy. In this
paper, SimulationX [14] software is used to quantitatively
investigate about the energy efficiency improvement, by using
SynRM in all-variable speed centrifugal chillers. The system
was tested for three distinct building load profiles. The results
are judged in comparison with the state-of-the-art IE2 IMs.
The load profiles represent normalized distribution, from high
to low loads. SynRM reduces the motor-drive loss on average
by 33%, which significantly increases the energy efficiency.
This reduces the total annual energy consumption by 3.08%
to 4.63%, for the different load profiles. If SynRM is used
only in the compressor, which consumes the highest amount
of energy in the chiller, still the energy savings occur in
the range of 2.53% to 4.04%. SynRM usage in the variable
speed centrifugal chillers require changes at the drive level,
which is easy to achieve. The difference of costs of SynRM
and IM for different parts of the centrifugal chiller are not
too high. SynRM is slightly more expensive, which can be
well-justified for the resulting energy savings. Therefore, it
could be a rather easily implementable and cost-effective
energy efficiency solution for future smart and energy efficient
buildings.

R EFERENCES
[1] International Energy Agency (IEA), “Energy-efficiency Policy Opportu-
nities for Electric Motor-Driven Systems,” Paris, France, 2011.
[2] L. Chuan, A. Ukil, “Modeling and Validation of Electrical Load Profiling
in Residential Buildings in Singapore,” IEEE Transactions on Power
Systems, vol. 30, issue. 5, pp. 2800–2809, 2015.
[3] Y. Kim, “Optimal Price-Based Demand Response of HVAC Systems
in Multi-Zone Office Buildings Considering Thermal Preferences of
Individual Occupants,” IEEE Transactions on Industrial Informatics, vol.
PP, no. 99, 2018. (doi: 10.1109/TII.2018.2790429)
[4] A. Jindal, N. Kumar, J. J. Rodrigues, “A Heuristic-based Smart
HVAC Energy Management Scheme for University Buildings,” IEEE
Transactions on Industrial Informatics, vol. PP, no. 99, 2018. (doi:
10.1109/TII.2018.2802454)
[5] N. K. Dhar, N. K. Verma, L. Behera, “Adaptive Critic-Based Event-
Triggered Control for HVAC System,” IEEE Transactions on Industrial
Informatics, vol. 14, no. 1, pp. 178–188, 2018.
[6] Y. Zhang, Y. Jin, J. Yao, G. Zhu, B. Wen, “Power Peak Shaving with
Data Transmission Delays for Thermal Management in Smart Buildings,”
IEEE Transactions on Industrial Informatics, vol. PP, no. 99, 2018. (doi:
10.1109/TII.2017.2781256)
[7] J. Ploennigs, M. Maghella, A. Schumann, B. Chen, “Semantic Diagnosis
Approach for Buildings,” IEEE Transactions on Industrial Informatics,
vol. 13, no. 6, pp. 3399–3410, 2017.
[8] D. M. K. K. V. Rao, A. Ukil, “Computational fluid dynamics-based
thermal modeling for efficient building energy management,” In proc.
41st IEEE Annual Conference of Industrial Electronics Society-IECON,
pp. 481–486, Yokohama, Japan, 2015.
[9] D. M. K. K. V. Rao, A. Ukil, “Modeling of Room Temperature Dynamics
for Efficient Building Energy Management,” IEEE Transactions on Sys-
tems, Man, and Cybernetics: Systems, vol. PP, issue. 99, pp. 1–9, 2018.
(doi: 10.1109/TSMC.2017.2758766)
[10] F. J. Wang, T. B. Chang, M. T. Chen, “Energy Conservation for Chiller
Plants by Implementation of Variable Speed Driven Approach in an
Industrial Building,” Energy Procedia, vol. 61, pp. 2537–2540, 2014.
[11] F. W. Yu, K. T. Chan, “An alternative approach for the performance
rating of air-cooled chillers used in air-conditioned buildings,” Building
and Environment, vol. 41, issue 12, pp. 1723–1730, 2006.
[12] R. Saidur, S. Mekhilef, M. B. Ali, A. Safari, H. A. Mohammed,
“Applications of variable speed drive (VSD) in electrical motors energy
savings,” Ren. and Sust. Energy Rev., vol. 16, no. 1, pp. 543–550, 2012.
[13] F. Oliveira, A. Ukil, “Energy Efficiency in Variable Speed Centrifugal
Chiller Systems Driven by Synchronous Reluctance Motor,” In proc. IEEE
Innovative Smart Grid Tech-ISGT, Singapore, 2018.
[14] ESI ITI GmbH, “Reference manual,” SimulationX, ver. 3.8, Dresden,
Germany, 2016.
[15] R. Saidur, “A review on electrical motors energy use and energy
savings,” Ren. and Sust. Energy Rev., vol. 14, issue 3, pp. 877–898, 2010.
[16] F. W. Yu, K. T. Chan, “Environmental performance and economic
analysis of all-variable speed chiller systems with load-based speed
control,” Applied Thermal Engg, vol. 29, no. 8, pp. 1721–1729, 2009.
[17] T. Hartman, “All-variable speed centrifugal chiller plants,” ASHRAE
Journal, vol. 43, no. 9, pp. 43–53, 2001.
[18] Air-Conditioning, Heating, and Refrigeration Institute (ARI), Standard
for Performance Rating of Water-Chilling Packages Using the Vapor
Compression Cycle, no. 550/590-2003, 2003.
[19] H. Crowther, J. Furlong, “Optimizing chillers & towers,” ASHRAE
Journal, vol. 46, no. 7, p. 34, 2004.
[20] W. P. Bahnfleth, E. B. Peyer, “Energy Use and Economic Comparison
of Chilled-Water Pumping System Alternatives,” ASHRAE Transactions,
vol. 112, no. 2, pp. 198–208, 2006.
[21] Int. Eletrotechnical Commission (IEC), Standard for Efficiency classes
of line operated AC motors (IE code), no. 60034-30-1:2014, 2014.
[22] H. Mahmoud, G. Bacco, M. Degano, N. Bianchi, C. Gerada, “Syn-
chronous Reluctance Motor Iron losses: Considering Machine Non-
Linearity at MTPA, FW, and MTPV Operating Conditions,” IEEE
Transactions on Energy Conversion, vol. PP, no. 99, 2018. (doi:
10.1109/TEC.2018.2811543)
[23] J. Ikheimo, J. Kolehmainen, T. Knskangas, V. Kivel, R. R. Moghaddam,
“Synchronous High-Speed Reluctance Machine With Novel Rotor Con-
struction,” IEEE Transactions on Industrial Electronics, vol. 61, no. 6,
pp. 2969–2975, 2014.
[24] C. M. Donaghy-Spargo, B. C. Mecrow, J. D. Widmer, “On the In-
fluence of Increased Stator Leakage Inductance in Single-Tooth Wound
Synchronous Reluctance Motors,” IEEE Transactions on Industrial Elec-
tronics, vol. 65, no. 6, pp. 4475–4482, 2018.
View publication stats
10
[25] S. S. R. Bonthu, A. Arafat, S. Choi, “Comparisons of Rare-Earth and
Rare-Earth-Free External Rotor Permanent Magnet Assisted Synchronous
Reluctance Motors,” IEEE Transactions on Industrial Electronics, vol. 64,
no. 12, pp. 9729–9738, 2017.
[26] ABB, “Synchronous Reluctance Motor: Brochures,” 2011. Avail-
able: new.abb.com/motors-generators/iec-low-voltage-motors/frequency-
controlled-motors/synchronous-reluctance-motor-drive-packages
[27] G. H. B. Foo, X. Zhang, “Robust Constant Switching Frequency-Based
Field-Weakening Algorithm for Direct Torque Controlled Reluctance
Synchronous Motors,” IEEE Transactions on Industrial Informatics, vol.
12, no. 4, pp. 1462–1473, 2016.
[28] M. H. Bierhoff, “A General PLL-Type Algorithm for Speed Sensorless
Control of Electrical Drives,” IEEE Transactions on Industrial Electron-
ics, vol. 64, no. 12, pp. 9253–9260, 2017.
[29] E. Trancho, E. Ibarra, A. Arias, I. Kortabarria, J. Jurgens, L. Marengo,
A. Fricass, J. V. Gragger, “PM-Assisted Synchronous Reluctance Ma-
chine Flux Weakening Control for EV and HEV Applications,” IEEE
Trans. on Industrial Electronics, vol. 65, no. 4, pp. 2986–2995, 2018.
[30] H. Hadla, S. Cruz, “Predictive Stator Flux and Load Angle Control
of Synchronous Reluctance Motor Drives Operating in a Wide Speed
Range,” IEEE Transactions on Industrial Electronics, vol. 64, no. 9, pp.
6950–6959, 2017.
[31] MetService, “Auckland, New Zealand Weather Data,” 2018. Available:
www.metservice.com/towns-cities/auckland/auckland-central
[32] NIWA, “Auckland, New Zealand Weather Data,” 2018. Available:
www.niwa.co.nz
[33] I. M. Alsofyani, N. R. N. Idris, “Lookup-Table-Based DTC of Induction
Machines With Improved Flux Regulation and Extended Kalman Filter
State Estimator at Low-Speed Operation,” IEEE Transactions on Indus-
trial Informatics, vol. 12, no. 4, pp. 1412–1425, 2016.
[34] M. Manohar, S. Das, “Current Sensor Fault-Tolerant Control for Direct
Torque Control of Induction Motor Drive Using Flux-Linkage Observer,”
IEEE Tr. Industrial Informatics, vol. 13, no. 6, pp. 2824–2833, 2017.
[35] T. Hartman, “LOOP chiller plant dramatically lowers chilled water
costs,” In proc. Renewable & Advanced Energy Systems for the 21st
Century, pp. 137–144, Hawaii, 1999.
[36] F. W. Yu, K. T. Chan “An alternative approach for the performance
rating of air-cooled chillers used in air-conditioned buildings,” Building
and environment, vol. 41, no. 12, pp. 1723–1730, 2006.
[37] H. Zhang, A. Ukil, “Framework for Multipoint Sensing Simulation for
Energy Efficient HVAC Operation in Buildings,” In proc. 41st IEEE
Annual Conf. on Industrial Elect.-IECON, Yokohama, Japan, Nov. 2015.
[38] H. Zhang, A. Ukil, “Analysis of Ambient Temperature Effects and
Airflow Rate for Energy Efficient HVAC in Buildings,” In proc. IEEE
TENCON Conf., Singapore, Nov. 2016.
[39] Paris agreement, 2016. Available:https://unfccc.int
Felipe Oliveira (M’17) received his B.E. degree in
electrical engineering from Pontificia Universidade
Catolica de Minas Gerais, Brazil in 2016, and the
MS degree from the Dept. of Electrical & Com-
puter Engineering, University of Auckland, New
Zealand in 2018. He is currently product manager
at ABB, Brazil, responsible for electric motors and
generators. His research interests include electrical
machines, drives, energy efficiency, HVAC system,
smart grid, control methods.
Abhisek Ukil (S’05-M’06–SM’10) received the
Ph.D. degree from the Pretoria (Tshwane) Univer-
sity of Technology, Pretoria, South Africa in 2006.
Currently, he is Senior Lecturer (Assoc. Prof.) in
the Dept. of Electrical & Computer Engineering,
Univ. of Auckland, New Zealand. From 2006–2013,
he was Principal Scientist at the ABB Corporate
Research, Baden-Daettwil, Switzerland. From 2013–
2017, he was Assistant Professor in the School of
EEE, Nanyang Technological University, Singapore.
He is inventor of 10 patents, and author of more than 160 refereed papers,
a monograph, 2 chapters. His research interests include smart grid, energy
efficiency, HVDC, renewable energy, condition monitoring.