This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
1
Smart Induction Motor Variable Frequency
Drives for Primary Frequency Regulation
Rasoul Azizipanah-Abarghooee, Senior Member, IEEEand Mostafa Malekpour, Student Member, IEEE

Abstract—This paper proposes an induction motor variable more efficient system operation [13]. In recent years, there has
frequency drive to investigate its potential in handling smart been a growing interest in the deployment of electric vehicles
grid’s frequency. This can thus mitigate the grid reliance on (EVs) to act as a mobile demand response in future smart
expensive power plants. To this end, a primary frequency grids. In [14], the potential of using EVs for PFR in the Great
controller is presented enabling the drive to reduce its power in Britain power system is studied through an EV load estimation
proportion to grid frequency drop. The dynamic limitation of the
tool. The economic benefit of EVs in providing PFR in the
drive due to load’s inertia is considered through a motor’s speed
rate limiter. Moreover, an appropriate inertia emulator is future low inertia grids is also quantified using an advanced
proposed for the smart drive by inspiring the inertial response of stochastic scheduling tool with explicitly modeling of inertia
a direct-on-line motor. The impacts of speed rate, maximum reduction due to high penetration of renewable energy sources
reserve power and motor driven load’s inertia on dynamic [15]. However, the PFR support from EVs has a noticeable
behavior of the smart motor load during the frequency support chronological distribution along a day.
are addressed. Besides, the effectiveness of the smart drive’s Recently, several researchers have focused on the potential
contribution in primary frequency regulation of the IEEE 39-bus of non-critical loads to provide demand response. In [16], a
test system is explored. In this regard, a critical droop coefficient smart load is formed by inserting a voltage compensator in
that guarantees the maximum reserve power delivery for the
series with the impedance load and grid. It provides governor
smart drives is analytically derived. Finally, the proposed
approach is extended to include the critical droop derivation for response by adjusting the magnitude of the voltage injected by
a set of smart drives with different sizes and priorities to rank the compensator according to the grid frequency. In [17], the
their participation in primary frequency regulation process. power of thermostatically controlled loads is coupled with the
Index Terms—Demand response, Inertia emulator, Primary grid frequency and its derivative by a linear controller and/or a
frequency controller, Induction motor variable frequency drive. preset shape controller using estimated disturbance size. A
thermodynamic model for representing 5000 domestic heat
I. INTRODUCTION
pumps is developed in [18] to change their consumption

P RIMARY frequency response (PFR) consists of the droop-

based governor response and inertial response provided by
synchronous generators [1]. In last decades, wind energy
power in response to grid frequency. The proposed static
smart load in [16] is extended to the variable frequency drives
(VFDs), i.e., dynamic smart loads as the collective rapid
conversion systems (WECSs) have been widely integrated into frequency responses in the Great Britain network [19]. This is
power grids [2]. As WECSs are connected by power conducted with subtle modifications to the constant volts-per-
converters, they neither obtain the inertial and governor hertz induction motor’s speed controller. These changes are
responses. Thus, the system’s inertia is significantly decreased comprised of a frequency deviation-based governor function
and spinning reserve must be increased to achieve a and a frequency derivate-based inertial response to amend an
satisfactory frequency control [3]. Several methods to equip induction motor’s power supply frequency. However, the
these power plants with PFR are proposed in [4], [5]. suggested motor speed ramp rate of 20 Hz/s may not be
However, the highest efficiency in energy harvesting cannot applicable particularly for the diode front-end drives with
be achieved using these methods due to required de-loaded high-inertia loads. Moreover, the droop gain of the suggested
operation [6]. Moreover, deployment of momentary increase smart drive is not tuned. Further, the inertia emulator does not
in the electrical power of WECSs to generate inertial response mimic the inertial behavior of a direct-on-line induction
results reduction in speed of wind turbines (WTs). This motor. Lately, a multi power-frequency droop characteristic is
extracts an extra PFR from grid to recover WT’s speed that proposed for a variable speed drive to minimize the impact of
may lead to the second frequency dip [7]-[9]. Battery energy the grid frequency variations on the speed of its load [20].
storage is another solution to meet the emergent challenges However, the efficiency of the drive on frequency response of
due to its fast ramp rate [10]. However, its extensive the power systems and their droop settings are not formulated.
installation is limited by high cost. Following the above-mentioned research studies, this paper
There is a growing motivation for loads to participate in focuses on the smart induction motor drive to enhance its
frequency regulation under demand response framework [11]. controllers and provide grid frequency support. The variable
Under-frequency load shedding (UFLS) is commonly used as speed drives are generally controlled by the constant volt-per-
a simplest demand response [12]. It is exposed that in the low hertz method for the below-rated speed [21]. The consumption
inertia grids, 1 MW load resources with under frequency power of an induction motor which is coupled with a constant-
relays providing full response within 0.5 second after dropping torque load has a nearly linear relationship with its rotor
frequency to a pre-set value is up to 2.35 times more effective speed. Based on this observation, the motor’s speed reduction
than 1 MW PFR from the generation side [3]. Furthermore, can be deployed to provide grid frequency support. Therefore,
the accurate online disturbance size estimation results in a a droop-based primary controller is developed to derive a
reserve speed resulting in a given power reduction for the
R. Azizipanah-Abarghooee is with the National Grid ESO, Warwick smart drive. Besides, the permissible speed down-rate limit is
CV34 6DA, U.K. (e-mail: [email protected]). determined by considering the motor driven load’s inertia to
M. Malekpour is with the Department of Electrical Engineering, avoid regeneration during motor’s speed reduction. Moreover,
University of Isfahan, Isfahan, Iran (e-mail: [email protected]).

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 2

the inertial response of a direct-on-line induction motor is vqs , ref  f ref (5)
properly mimicked for the smart drive. Finally, a methodology
where, 𝑓𝑟𝑒𝑓 is the reference frequency of the motor stator
is established to analytically set smart drive’s droop gain in
the large power systems. The main contributions of this paper voltage in per-unit. The steady-state characteristics of a 50 hp
can be summarized as follows: induction motor with this speed controller are shown in Fig. 1
[21]. The motor rotates a constant torque load. Thus, load’s
 A new primary frequency controller is developed for the
power is linearly reduced with its speed. This feature can be
smart drives and designed by considering the dynamic
deployed in the grid frequency regulation [19]. Let the motor
limitations imposed by motor driven load’s inertia and
initially operates at its nominal condition, denoting by Point
regeneration incapability of the diode front-end drives.
A. If the speed controller reduces the motor speed by 10%,
 An appropriate inertia emulator is suggested by inspiring
then, the Point B will be the new operating point. As a result,
from inertial response of the induction motors.
the input power of the drive is also reduced by about 10%. The
 A critical droop gain is derived for the smart drive to
dynamic behavior of the motor during speed reduction is not
achieve its highest frequency support capability.
shown in Fig. 1; however, it must be considered in practice.
Section II describes the induction motor variable speed
drives. The proposed smart drive is elaborated in Section III.
III. PROPOSED SMART VARIABLE FREQUENCY DRIVE
Section IV investigates dynamic behavior of the drive. The
efficiency of the proposed smart loads in grid frequency The proposed smart drive architecture is depicted in Fig. 2.
regulation is assessed in Section V. Finally, Section VI The consumption power of the drive, which is inherently
concludes the paper. decoupled from the grid supply, is mitigated based on the grid
frequency through adjusting the operating point of the motor.
II. INDUCTION MOTOR VARIABLE FREQUENCY DRIVES To this end, the inverter’s frequency is updated using the
A diode rectifier is commonly employed as a front-end in primary controller (PCO) and inertia emulator (IE) blocks.
the variable frequency drives [19]. However, an inverter A. Proposed Primary Controller for Smart IM VFD System
synthesizes the variable frequency voltage to control the
A reserve speed control scheme based on the governor
motor’s speed. These drives do not have energy reversibility
action of the power plants is suggested as PCO for the smart
capability. Thus, a dynamic braking chopper is required to
drive (See Figs. 2 and 3) [23]. The grid frequency measured
dissipate excessive energy in regeneration mode. The
by a phase-locked loop (PLL) and the motor’s speed are the
induction motors (IMs) are widely used in the variable speed
input signals for the PCO. The measured frequency deviation
drives as a means of converting electric power to mechanical
power [21]. The voltage and torque equations for a
Power (p.u.)

Motor Input A
symmetrical induction motor can be found in [21]. The
consumption electrical active power of the motor expressed in B
terms of d- and q-axes voltages (vds and vqs) and currents (ids
and iqs) of the motor’s stator is [21]: Motor Output Motor Driven Load
Pe  vds ids  vqs iqs (1)
Speed (p.u.)
The electromagnetic torque produced by the motor can be Fig. 1. Power versus rotor speed characteristics of a 50 hp induction motor.
written as follows:
Motor Pin
Te   ds iqs  qs ids (2) ~ Grid
Denote 𝜓𝑑𝑠 and 𝜓𝑞𝑠 the d- and q-axes stator’s flux linkages, M
~
respectively. The rotating speed of the motor’s shaft ωr is Load Pm Pe
governed by the equation of motion as follows: Dfie
Speed IE
d 1 PLL
r  Te  Tm  (3) vabc Controller
PCO
dt 2H t r Dr Df
with Tm as the mechanical torque. Parameter Ht is the total Fig. 2. Control structure of the proposed smart IM VFD system.
inertia constant in seconds. It may be expressed in terms of the
motor inertia constant Hm and load inertia constant Hl as [22]: Pres
max r1
1 J m  2eb  1 J l  2eb 
2 2
Df 1 1 Dr
Ht  H m  Hl     (4)
2 Pb  P 
1sTpco Rl Pres
2 Pb  P 
where, Jt is the aggregated inertia of the motor’s shaft Jm plus Df 1
1sTie Dfie
the motor driven load Jl in kg.m2. Denote 𝜔𝑒𝑏 the base
electrical angular frequency. The horsepower (hp) rating of the lim 2Dfdb0
machine is selected as the base power 𝑃𝑏 in volt-amperes.
D r Rate  fref
Denote P the number of poles. Limiter
The open-loop constant volt-per-hertz strategy is the most
popular method for the drive applications in industry due to its r,ref C1 C2 Dfie
cost-effectiveness [21]. In this control scheme, the reference Fig. 3. Block diagrams of the primary controller (top), inertia emulator
q-axis stator voltage is set as follows [21]: (middle) and the speed controller (bottom) shown in Fig. 2.

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 3

is firstly passed through a dead-band and then filtered by a low response. The temporary decrease in the motors’ input energy
pass filter (LPF). The resultant signal is divided by the droop reduces the rate of change of grid frequency instantly after
gain 𝑅𝑙 in order to obtain reserve power coefficient 𝑃𝑟𝑒𝑠 . This event. In addition to the inertial response, there exists a
coefficient is restricted to a maximum value and determines steady-state decrease in motor’s consumption energy labeled
the portion of the pre-event motor’s output power 𝑃𝑚1 , which as “G” in Fig. 4, which is pertinent to its load characteristic.
should be reduced in response to a sudden grid frequency However, the governor response of the DOL motors is very
drop. However, 𝑃𝑟𝑒𝑠 cannot be directly sent to the “Speed small compared to that of the proposed smart drives.
Controller” and it should be converted into a speed signal. According to the above explanations, the grid frequency
Accordingly, this variable conversion process is derived in the variation (e.g. 𝑓𝑛𝑒𝑤 − 𝑓𝑜𝑙𝑑 in Fig. 4) should be directly added
following manner. The motor speed and its output power can to the inverter’s frequency of a VFD drive in order to emulate
be updated following the frequency incident as follows: the inertial response of a DOL motor properly. As indicated
r 2  r1  Dr , Pm2  Pm1  DPm (6) earlier, this inertia emulation differs with that of the converter-
interfaced generations where time derivative of the grid
Dr  resr1 , DPm  Pres Pm1 (7)
frequency is deployed to model inertia emulator [24]. Whilst
Denote subscripts 1 and 2 the pre- and post-event steady the inertia emulation was traditionally implemented by
state quantities, respectively. The reserve speed factor 𝜔𝑟𝑒𝑠 , controllers in the converter-interfaced generations, it is
theoretically ranges from 0 to 1, determining the required conducted by both controllers and induction motor in the
reduction in the motor’s speed to provide a power reduction proposed smart drive. In the former case, the inertial power is
which is equal to 𝑃𝑟𝑒𝑠 𝑃𝑚1 . If the motor driven load is a firstly determined using the rate-of-change-of-frequency
constant-torque load, it can be yielded: (RoCoF), imitating the swing equation of a synchronous
 Pm1  Pres Pm1 r1  resr1   Pm1 r1 
1 1
(8) generator. Then, the derived power is added to the reference
power of the generation unit. On the other hand, the inertial
Thus, (7) can be appropriately expressed as follows: power of an induction motor in a VFD drive can be coupled
Dr  Presr1 (9) with grid frequency by transferring the frequency variations
Thus, the power and speed reserve coefficients 𝑃𝑟𝑒𝑠 and from the rectifier’s terminal to the inverter’s terminal. This
𝜔𝑟𝑒𝑠 are expectably identical for a constant-torque load. For strategy can be employed as a command variable in the
the sake of simplicity, the motor’s output power is deployed in inverter’s frequency while this is not the case for the
(6) to (9) for the reserve speed calculation instead of input converter-based generating units. As observed in Fig. 3, the
power of the drive. This is an acceptable way especially when proposed inertia emulator contains a dead-band and a low-pass
the induction motor is operating near to its rated speed. filter in order to remove the undesirable signals from the
measured frequency. There is no derivative block in this
B. Proposed Inertia Emulator for the Smart IM VFD System
inertia emulator in contrast with that of suggested in [19].
In order to design an inertia emulator for the induction
motor drives, what should be underscored is that whereas the C. Speed Controller of the Proposed Smart IM VFD System
drive is connected to grid through converters like variable A modified speed controller is suggested as shown in Fig. 3.
speed wind turbines, its inertia emulator design differs with In this design, ∆𝜔𝑟 determined by the PCO is firstly restricted
them. To describe this discrepancy, power trajectory of a by a rate limiter and then added up to the pre-event reference
direct-on-line (DOL) lossless induction motor, connected to an rotor speed 𝜔𝑟.𝑟𝑒𝑓 . Considering the inertia of the motor driven
infinite bus, following a step reduction in the bus frequency is load, the rate limiter has a key role in limiting the dynamics of
shown in Fig. 4. The power trajectory is plotted versus the motor in speed reduction process. In the next stage, the
frequency and time. Initially, the motor operates at point A. inverter electrical frequency 𝑓𝑟𝑒𝑓 is determined as follows:
After the infinite bus frequency decrement from 𝑓𝑜𝑙𝑑 to 𝑓𝑛𝑒𝑤 , f ref  C1 (r ,ref  Dr )  C2  Dfie (10)
the motor’s stator frequency is also identically changed. This
leads to the transient mitigation of the motor’s electrical power where, coefficients C1 and C2 can be calculated using the
following the operating point change from point A to B, while induction motor’s parameters. To this end, the relationship
its shaft speed has not been changed yet due to non-zero between operating IM’s rotor speed and its stator frequency
shaft’s inertia. Next, the motor’s speed decelerates to reach its for a given torque must be derived using torque-speed
new equilibrium point C since the motor’s electrical power characteristics. Then, the coefficients deployed in (10) can be
differs from its mechanical one. Meanwhile, some part of the determined using a least-square approximation.
shaft’s kinetic energy is consumed by the motor’s load. This
energy that is labeled as “I” in Fig. 4 is the motor’s inertial IV. VALIDATION OF THE PROPOSED SMART DRIVE

Power Power The performance of the proposed smart drive is validated

Load Power through the simulations in PowerFactory [25]. The smart
A A Electrical Power
drives are connected to a voltage source. Parameters of the
Steady-state reduction G drives are given in Table I. The motors’ parameters can be
C I
Transient reduction C found in [21]. The frequency dead-band ∆𝑓𝑑𝑏 is 15 mHz [26].
B B Mechanical Power A. Inertial Response of the Smart IM VFD Systems
fnew fold Frequency Time Let the voltage source frequency is abruptly dropped by
Fig. 4. Inertial and governor responses of a direct-on-line induction motor 1%. Here, the PCO of the smart drives is deactivated. The
with a constant torque motor driven load during a frequency drop event.

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 4

TABLE I
THE INDUCTION MOTORS AND SMART DRIVES PARAMETERS

r (p.u.)
Pb (hp) Jm (kg.m2) Jl (kg.m2) Jt (kg.m2) Hm (s) Hl (s) Ht (s)
50 1.66 30 31.66 0.8 14.3 15 500 hp
500 11.06 700 711.06 0.53 33.3 34
50 hp
𝑚𝑎𝑥
𝑃𝑟𝑒𝑠 (p.u.) 𝜔̇ 𝑙𝑖𝑚 (p.u./s) 𝑇𝑝𝑐𝑜 (s) 𝑇𝑖𝑒 (s) C1 C2

Pin (p.u.)
50 0.25 0.1 1 0.1 0.9862 0.0564
500 0.1 0.04 1 0.1 0.9937 0.0199

motor speed and its input power are portrayed in Fig. 5. It can
be clearly observed that the suggested inertial emulator for the Time (s)
smart drives can appropriately mimic the dynamic behavior of Fig. 7. The governor response of the smart drives with nominal settings.
the DOL motors. In addition, the amount of power nadir is
greater in case of larger motors due to its steeper power-slip power decrement is significantly higher than the maximum
characteristics around the synchronous speed. The input power reserve values. In other words, the motor’s speed reduction
versus slip and inertial energy versus rotor speed trajectories leads to the considerable dynamic behavior in terms of its
for the smart drive are given in Fig. 6. The steady-state input input power. However, the chopper limits the minimum
and output powers characteristics are also portrayed with rectifier power to zero in the diode front-end drives. Here, the
magenta dotted and red dash-dot traces, respectively. The load’s inertia is about 20 and 60 times larger than that of the
motor initially operates at point A with the slip of 1.5%. When motor, in 50 hp and 500 hp motors, respectively [27].
the electrical frequency is stepped down by 1%, the input C. Effect of Speed Rate Limits on Smart Drive Performance
power characteristic is shifted to the left side and the new
The dynamic behaviors of the smart drive with nominal and
operating point would be B. Thus, the motor’s input power
double-nominal speed rate limits are compared in Fig. 8.
decreases suddenly. However, the motor speed in points A and
Despite the initial faster speed reduction, it can be seen that
B are roughly identical due to the shaft’s inertia. Next, the
this process is not considerably accelerated while the speed
motor speed is reduced to reach new equilibrium operating
down-rate limit is doubled. However, it results in a substantial
point C. It is notable that the released kinetic energy of shaft
variation in motor’s electromagnetic torque, i.e. 75%. Whereas
𝐸𝑖𝑛𝑒𝑟𝑡𝑖𝑎𝑙 , is subtracted from the motor’s energy consumption.
the dissipated energy by chopper is amplified by about 20%,
B. Governor Response of the Smart IM VFD Systems the operating time of the chopper is decreased by around 20%.
The simulation results with the previously mentioned event The corresponding trajectories are portrayed in Fig. 9 where
are presented in Fig. 7. Considering the dynamic limitations of three steady-state characteristics associated with synchronous
the motors, it is assumed that the nominal rate of speed speeds such as 1, 0.875, and 0.75 p.u. are also shown with
reduction would be 0.1 and 0.04 p.u. per seconds for 50 hp dotted traces. At t  2.2 s, the motor’s torque and power
and 500 hp motors, respectively. The maximum reserve power approach the final post-event steady-state characteristic, that
is 0.25 p.u. and 0.1 p.u. for 50 hp and 500 hp motors, is, 𝜔𝑒 = 0.75 p.u., under a double-nominal speed rate denoted
correspondingly. The droop gain 𝑅𝑙 , is set to 0.05% in order to by A2. However, they reach to the characteristic of 𝜔𝑒 = 0.875
release maximum reserve power. As observed in Fig. 7, the p.u. under nominal speed rate limit (denoted by A1). Next, the
power consumption of the smart drives 𝑃𝑖𝑛 is permanently torque reaches the final steady-state characteristic with
scaled down by the maximum reserve power after about 8
seconds following the incident. However, the temporary
r (p.u.)

500 hp
r (p.u.)

50 hp
500 hp
50 hp
Pin (p.u.)

50 hp
Pe & Pin

A 500 hp
(p.u.)

C 500 hp
Time (s)
B Einertial  0.5(10.4)2  0.6 [p.u.s] Fig. 8. The governor response of the smart drives with nominal (solid) and
double-nominal (dash-dot) speed rate limits.
Time (s)
Fig. 5. Inertial response of the DOL IMs (dash-dot) and smart drives (solid). C C e 
0.875
Pe (p.u.)
Te (p.u.)

B2 B2
A1
e 
Pin (p.u.)

A
B1 e  B1 1
A2 0.75 A2 1
r (p.u.) r (p.u.)
Fig. 9. Torque and power versus rotor speed trajectories up to 2.2 s (A), 3.4 s
Slip (%) r (p.u.) (B) and 10 s (C) for the 50 hp drive system with nominal (blue solid) and
double-nominal (red dash-dot) speed rate limits.
Fig. 6. Pin trajectory (left) and inertial energy (right) of 500 hp smart drive.

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 5

nominal speed rate at t  3.4 s but under significantly lower

Pin (p.u.)
amplitude (denoted by B1). Afterwards, the operating point
goes up to the steady-state characteristics and establishes the 50 hp
system conditions determined by the PCO (point C). It can be
concluded that the speed down-rate limit has a meaningful
impact on the dynamic behavior of the smart drives and it Time (s)
should be taken into account in the designing stage. Fig. 12. Governor response of the smart drive with (solid) and without load’s
inertia (dash-dot) consideration.
D. Effect of Maximum Reserve on Smart Drive Performance
The dynamic behaviors of the smart drive with nominal and

Pe (p.u.)
Pe (p.u.)
double-nominal maximum reserve power are compared in Fig.
10. Doubling the reserve power results in twofold motor’s
rotor speed reductions since the motor rotates a constant
torque load. However, it does not have considerable effect on
the amplitude of the electrical torque in contrast to the effect r (p.u.) r (p.u.)
of the speed rate limit. Nevertheless, once the reserve power is Fig. 13. Power versus speed trajectories for the 50 hp (left) and 500 hp (right)
doubled the dissipated energy by the chopper and its operating drives with (solid) and without (dash-dot) load’s inertia consideration.
times are increased by about 120% and 80%, respectively. The plays a key role in the dynamic behavior of the drives. It can
corresponding trajectories along with three steady-state also be observed that the dynamic behavior of the motor
characteristics related to synchronous speeds of 1, 0.75, and during the speed reduction can be predicted adequately by the
0.5 p.u. are illustrated in Fig. 11. The motor reaches to the steady-state characteristics when the load’s inertia is ignored.
final post-event steady-state characteristic of 𝜔𝑒 = 0.75 p.u., However, this is not the case with the load’s inertia
at t  3.4 s (point A), while the reserve power is limited to the consideration. Looking more closely at the trajectories, the
nominal value. On the other hand, the time needed to approach underdamped behavior of the motor’s dynamics is changed to
this condition becomes twofold if the maximum reserve power overdamped one by modelling load’s inertia especially in
is doubled (point B2). Therefore, the chopper should be large motors. According to these observations, the parameters
correctly designed considering the maximum reserve power. of the primary frequency controller deployed in the smart
E. Effect of Load’s Inertia on Smart Drive Performance drives should be selected by considering the motor driven
load’s inertia, particularly for the diode front-end drives.
As previously pointed out, the significant motor’s dynamics
comes from the substantial inertia of the motor driven load. To
V. FREQUENCY REGULATION WITH THE PROPOSED SMART
understand this, the dynamic performance of the smart drive
DRIVES IN THE POWER SYSTEMS
with and without load’s inertia consideration is compared in
Fig. 12. The load’s inertia is 20 and 60 times larger than that The effect of the proposed smart variable speed drive on the
of the motor, in 50 hp and 500 hp motors, respectively. In the primary frequency regulation of the large power systems is
“W.O. load inertia” case, the shaft’s total inertia is calculated initially investigated. Then, a critical droop is analytically
neglecting the load’s inertia. The trajectories are depicted in derived for the smart drive as a function of disturbance size. It
Fig. 13. Inspection of the results reveals that the load’s inertia ensures that the reserve power of the smart drive is completely
released for a given disturbance size. Finally, the extended
expressions to calculate the critical droop for a set of smart
r (p.u.)

drives with discrepant sizes and priority rankings are provided.

500 hp
50 hp A. Description of the Studied IEEE 39-Bus Power System
The IEEE 39-bus system is considered as the studied power
network [25]. The system’s demand is about 6 GW, which is
Pin (p.u.)

supplied by 10 synchronous generators (SGs). All the SGs are

equipped with IEEE Type 1 exciters without power system
50 hp 50 hp stabilizer [22]. The nominal grid frequency is 50 Hz. The
Time (s) generator SG10 that is rotated by a hydro turbine conducts
Fig. 10. The governor response of the smart drives with nominal (solid) and primary frequency control. Its governor has permanent and
double-nominal (dash-dot) maximum reserve powers. temporary droops of 4% and 20%, respectively. In order to
keep RoCoF, frequency nadir, quasi-steady-state frequency
C2 C1 C1
B1 C2 and frequency recovery time within their standard boundaries
Pe (p.u.)
Te (p.u.)

B1 e 1 of 0.25 Hz/s, 49.2 Hz, 49.5 Hz and 60 seconds [26],

respectively, for a 5% infeed loss (i.e., 300 MW), the
A e  e  B2 A following modifications are made. The rating of the SGs is
B2 0.75 0.5 chosen as listed in Table II. Note that the values in Table II
r (p.u.) r (p.u.) denote initial settings for the SGs’ active power. The exact
Fig. 11. Torque and power versus rotor speed trajectories up to 3.4 s (A), 6.8 s active powers are determined through the load flow
(B) and 12 s (C) for the 50 hp drive system with nominal (blue solid) and calculation with the distributed slack generator. The reset time
double-nominal (red dash-dot) maximum reserve powers.
𝑇𝑟 of the hydro governor is decreased from 10 to 5 seconds.

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 6

TABLE II maximum reserve power of the drives (see Table III). Clearly,
PARAMETERS OF THE SG UNITS IN IEEE 39 BUS POWER SYSTEM [25]
this demand response mitigates the frequency support burden
Unit SG1 SG2 SG3 SG4 SG5 SG6 SG7 SG8 SG9 SG10
on the generation side. An important question that arises is
H (s) 6.5 5.28 4.75 3.12 5 4.35 3.23 3.46 4.06 3.25
what droop gain ensures the maximum reserve delivery for the
P (MW) 1000 710 650 632 254 650 560 540 830 250
S (MVA) 1250 700 812 790 317 812 700 675 1037 1000 smart loads. This is addressed in detail in the next subsections.

The SG5 is comprised of two 317 MVA generators. The C. Critical Droop Derivation for the Smart Drive
power of one of them is fixed to 254 MW and the other one is In this subsection, the critical droops of smart loads are
ranged from 60 to 300 MW to simulate discrepant infeed derived. To this end, let us first consider the impact of
losses of 1% to 5%. Five different smart drives, which totally decreasing droop gain of the smart loads. The COI frequency
consume 5% of the system demand, are integrated into the and the smart loads power reduction are illustrated in Fig. 15
grid at the terminals 3, 4, 8, 16 and 20. Their information are for the 𝑅𝑙 ranged from 5% to 1% following the 120 MW
given in Table III. Their speed rate limits are determined in generation loss. It can be seen that decreasing the droop 𝑅𝑙
such a way that the minimum consumption power of the results in more power decrement from the smart drives and
drives during the frequency support would be about 25% of pushes the COI frequency closer to its nominal value. As
their nominal values to avoid regeneration. The remaining observed, about 30% of maximum smart loads’ reserve power
95% of the system’s demand are modelled as static loads. is not used even with 1% droop. A more general assessment in
Their active and reactive powers are modeled as constant which the values of the post-event steady-state (at t  60 s)
current and constant impedance load types [22]. Furthermore, reserve power of the smart loads for different droop values 𝑅𝑙
their load damping constant is set to 2%/Hz [28]. and infeed loss sizes is presented in Fig. 16. As shown, the
B. Effect of Smart Drive on Primary Frequency Regulation critical droop 𝑅𝑙𝑐𝑟 increases with the size of infeed loss. Fig.
17 illustrates that this relationship is linear (the solid trace). As
The simulation results for a 120 MW infeed losses (2% of the post-event transients subside, it can be given:
l l Df  D 1  Pl  Df
the system demand) at terminal 34 are shown in Fig. 14. The
DPloss  Pg Rg1Df  PR 1
(11)
droop gain of the smart drives and the SGs are set to 5%. With
frequency support by the smart loads, the center-of-inertia where, ∆𝑃𝑙𝑜𝑠𝑠 , 𝑃𝑔 and 𝑃𝑙 represent the event size, generation
(COI) frequency is improved in terms of frequency nadir and side power and the smart loads power in per unit, respectively.
RoCoF. The smart loads’ power reduction reaches its Denote 𝑅𝑔 and 𝐷 the generation side’s droop and load
maximum 54 MW at roughly 3 seconds after the event. It thus damping constant, respectively. It is to be noted that 𝑃𝑔
reduces to its final value of 15 MW which is only 20% of total indicates generation side power with PFR capability (i.e. 1000
TABLE III MW (0.167 p.u.)). Solving (11) yields:
Df  DPloss  D 1  Pl   Pg Rg1  PR
l l 
PARAMETERS OF THE SMART LOADS IN IEEE 39 BUS POWER SYSTEM 1 1
𝑚𝑎𝑥 (12)
Drive Connected Power Pb Inertia (kg.m2) 𝑃𝑟𝑒𝑠 𝜔̇ 𝑙𝑖𝑚
No. bus No. (MW) (hp) Motor Load (p.u.) (p.u./s)
Smart drives COI Frequency

1 3 30 50 1.66 3.32 0.25 0.15 Smart drives droop decreases

2 4 45 50 1.66 6.64 0.25 0.09
(From 5% to 1%)
(Hz)

3 8 60 50 1.66 9.96 0.25 0.07

4 16 75 50 1.66 13.28 0.25 0.05
5 20 90 50 1.66 16.6 0.25 0.04
power reduction

Smart drives droop decreases

COI RoCoF SGs governor COI Frequency

(From 5% to 1%)
(MW)

With smart drives

(Hz)

Time (s)
(Hz/s) response (MW)

Fig. 15. Impact of decreasing smart drives droop on their participation in

primary frequency regulation after a 120 MW infeed loss.
Smart drives power
reduction
Pres (p.u.)

Without smart drives Rl (%)

Fig. 16. Post-event steady-state reserve power of the smart loads versus droop
𝑅𝑙 for infeed loss size ranged from 1% to 5% (of the system’s demand).
response (MW)

droop Rlcr (%)

SGs inertial

Analytical
Critical

Numerical

Time (s) Size of infeed loss (%)

Fig. 14. The simulation results for a 120 MW infeed loss. Fig. 17. Numerical (solid) and analytical (dash-dot) critical droops 𝑅𝑙𝑐𝑟 .

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 7

The power reduction of smart loads can be expressed as: their power can be decreased with different rates. Here, the
Df  Pl DPloss  Rl D 1  Pl   Rl Pg R  Pl 
1 1 1 drive with the fastest response is ranked with highest priority
DPl  PR
l l g
(13)
(see Table IV). Thus, the critical droops can be given as:
As defined earlier, the critical droop can be achieved while
DPlossj  i 1 Pli Presmaxi
j
∆𝑃𝑙 is equal or greater than its upper limit as follows: Rlcrj  j  1,..., 4 (24)
Pl DPloss  Rl D 1  Pl   Rl Pg R  Pl   PP
1 1  Pg Pli 
P  D 1  Pl    i  j 1
max 5
(14) max

 Rlcri 
g l res
resj
Hence, (18) can be rearranged as follows:  Rg

Rl   DPloss  PP
l res  Pres   D 1  P   P R 
max max 1 1 1 where, ∆𝑃𝑙𝑜𝑠𝑠 is defined in per-unit based on system demand:
(15)
l g g
DPloss  0.01 0.02 0.03 0.04 0.05 (25)
Subject to:
In other words, the critical droops are set in such a way that
DPloss  PP max
l res (16) the smart drive with the highest priority (i.e., with greatest
In fact, the calculated droop 𝑅𝑙 with equal sign in inequality ramp rate) is fully deployed for the lowest infeed loss. The
(15) indicates the critical droop 𝑅𝑙𝑐𝑟 . As previously observed, resultant critical droops obtained from (23) and (24) are listed
𝑅𝑐𝑟 has a linear relation with event size ∆𝑃𝑙𝑜𝑠𝑠 . It can be in Table IV. The droops calculated by (23) are provided for a
implied from (14)-(15) that if the droops are set to the values 5% infeed loss. By defining load’s priority, the load with
equal or lower than the critical droops, the reduction powers of highest priority has the lowest critical droop. In this regard,
the smart loads will be equal or greater than its maximum the results of the scheme defined by (24) are compared with
limit. The analytically derived critical droop is shown in Fig. that of (23) for various infeed loss sizes.
17. It is roughly 10% greater than the numerical critical droop. TABLE IV
The expressions derived in this case are valid for a single PRIORITY AND CRITICAL DROOPS OF THE SMART DRIVES
𝑚𝑎𝑥
smart drive with demand 𝑃𝑙 and reserve power 𝑃𝑟𝑒𝑠 , and/or a Drive number 1 2 3 4 5
set of identical drives that their total demand is 𝑃𝑙 . A more 𝑅𝑙𝑐𝑟 by Eq. (27) 2.5% 2.5% 2.5% 2.5% 2.5%
general approach is established in the next subsection. Priority 5 4 3 2 1
𝑅𝑙𝑐𝑟 by Eq. (28) 0.4% 0.83% 1.3% 1.9% 2.5%
D. Critical Droops for Smart Drives with Different Priority
COI Frequency

In practice, a set of smart drives may be constituted of

several drives with different sizes. This can be interpreted as: Rl (DPloss 2%)
(Hz)

Pl  Pl1  Pl 2  ...  Pl 5 (17)

where,
Pl1  Pl 2  ...  Pl 5 (18)
with 𝑃𝑙𝑖 as the consumption power of the i drive. The drives
th
Total power reduction of smart drives
(MW)

𝑚𝑎𝑥
can provide a total reserve power up to 𝑃𝑟𝑒𝑠 𝑃𝑙 . This can be
DPl

expressed in terms of individual drives’ reserve power as:

Presmax Pl  Presmax
1 Pl1  Pres 2 Pl 2  ...  Pres 5 Pl 5
max max
(19)
𝑚𝑎𝑥
where, 𝑃𝑟𝑒𝑠𝑖 denotes the maximum reserve power of the ith
(MW)

drive. Therefore, (12) can be rewritten as follows:

Pl1

 
1 Rl (Priority) Smart drive 03
Df  DPloss D 1  Pl   Pg Rg1   i 1 Pli Rli 1
5
(20)
where,
Rl (DPloss 4%)
Df  Rl1 Pres
(MW)

max
(21)
Pl2

The similar process of (13)-(15) is implemented in order to

derive the first smart drives critical droop as follows: Smart drive 04
DPloss  P P max
Rlcr1  l1 res1
(22)
 
(MW)

Pg P
 D 1  Pl     i  2 li
Pl3

5

max
P res1
 Rg Rlcri  Smart drive 08
This can be generalized for the first four smart drives as:
DPloss  i 1 Pli Presmaxi
j
(MW)

Rlcrj  j  1,..., 4 (23)

Pl4

 Pg Pli 
P  D 1  Pl    i  j 1
5

max
 Smart drive 16
Rlcri 
resj
 Rg
The fifth smart drive’s critical droop can also be determined
(MW)

from (23) by removing summation series at its denominator.

Pl5

The critical droops derived by (23) ensure that all the smart
drives deliver the maximum reserve power for the infeed loss Smart drive 20
∆𝑃𝑙𝑜𝑠𝑠 . Let us now assume that there is different load’s inertia Time (s)
in addition to diverse power ratings for smart drives. Hence, Fig. 18. Simulation results for a 60 MW infeed loss (∆𝑃𝑙𝑜𝑠𝑠 = 1%).

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 8

The simulation results related to 1% and 3% infeed loss VI. CONCLUSIONS

events are illustrated in Figs. 18 and 19. In the solid traces, the A smart induction motor variable speed drive is designed to
priority-based critical droop defined by (24) is used to set the support primary frequency regulation of the power systems. A
drives’ droops. This parameter is set using (23) in dotted and suitable inertia emulator is presented for the smart drive by
dash-dot traces with two different ∆𝑃𝑙𝑜𝑠𝑠 , which gives 0.5% inspiring from a direct-on-line induction motor. Besides, a
𝑚𝑎𝑥
and 1.8% droops, correspondingly. Recall that 𝑃𝑟𝑒𝑠 is 0.25 droop-based primary frequency controller is suggested for the
for all the drives, the first drive reaches to its limit for the 1% smart drive to decrease its power consumption following a
infeed loss when the load’s priority is considered. This ranked frequency drop event, considering the dynamic limitation of
contribution in frequency support can also be seen in Fig. 19 the drive due to its load’s inertia. The simulation results show
where three first drives’ power is reduced by 25% (solid that the proposed inertia emulator properly mimics the inertial
traces). Hence, the drives with low priorities sense lower response of a direct-on-line motor. Recall that the speed
power reduction in both post-event transient and steady-state down-rate limit has a meaningful impact on the dynamic
conditions (see the solid traces for 𝑃𝑙4 and 𝑃𝑙5 ). It is strikingly behavior of the smart drive where it becomes roughly twofold
observed that the priority traces have roughly identical by doubling this rate limit. On the other hand, the maximum
behavior with dotted (∆𝑃𝑙𝑜𝑠𝑠 = 2%) and dash-dot (∆𝑃𝑙𝑜𝑠𝑠 = reserve power of the drive has not a considerable impact on its
4%)) traces in the plots pertinent to 𝑃𝑙2 and 𝑃𝑙4 , respectively. dynamic behavior while it determines the required chopper
This observation is in agreement with (23) and (24). Besides, circuit ratings. The simulation results illustrate that the smart
by comparing the two plots of 𝑃𝑙1 , it is clearly revealed that drives can significantly enhance the frequency regulation of
the priority-based method causes the drive to release kinetic the large power systems even their total reserve power is just
energy faster once the infeed loss size increases. Finally, the 1.25% of the system demand. Finally, a new method is
total smart loads’ power reduction in the priority case tends to established to analytically derive the drives’ critical droops to
dash-dot traces. This is actually due to (24) which results in a guarantee the smart loads’ participation in primary frequency
1.65% equivalent droop for the drives. This equivalent droop control with their maximum capabilities. The critical droops
is obtained when ∆𝑃𝑙𝑜𝑠𝑠 inserted in (23) is 3.75%. are also developed for a set of drives with discrepant priorities
from the frequency support point of view. It is induced that the
COI Frequency

Rl (DPloss 2%) participation of the smart drives in frequency control can be

managed by the proposed priority-based critical droops.
(Hz)

VII. REFERENCES
[1]R. Boyer, “Primary Governing and Frequency Control in ERCOT,” in
IEEE Power Engineering Society General Meeting, Tampa, FL., 2007.
[2]R. Azizipanah-Abarghooee, M. Malekpour, T. Dragičević, F. Blaabjerg,
Total power reduction of smart drives and Terzija V. “A linear inertial response emulation for variable speed
(MW)
DPl

wind turbines. IEEE Trans. Power Syst., to be published.

[3]P. Du and J. Matevosyan, “Forecast system inertia condition and its impact
to integrate more renewables,” IEEE Trans. Smart Grid, vol. 9, no. 2, pp.
1531-1533, Mar. 2018.
[4]H. Xin, Y. Liu, Z. Wang, D. Gan and T. Yang, “A new frequency
regulation strategy for photovoltaic systems without energy storage,” IEEE
(MW)
Pl1

Trans. Sustain. Energy, vol. 4, no. 4, pp. 985-993, Oct. 2013.

Rl (Priority) [5]X. Yuan and Y. Li, “Control of variable pitch and variable speed direct-
Smart drive 03 drive wind turbines in weak grid systems with active power balance,” IET
Renew. Power Gener., vol. 8, no. 2, pp. 119-131, Mar. 2014.
[6]V. Knap, S. K. Chaudhary, D. I. Stroe, M. Swierczynski, B. I. Craciun and
(MW)

R. Teodorescu, “Sizing of an energy storage system for grid inertial

Pl2

response and primary frequency reserve,” IEEE Trans. Power Syst., vol.
Smart drive 04 31, no. 5, pp. 3447-3456, Sep. 2016.
[7]S. Engelken, A. Mendonca and M. Fischer, “Inertial response with
improved variable recovery behaviour provided by type 4 WTs,” IET
Renew. Power Gener., vol. 11, no. 3, pp. 195-201, Apr. 2017.
(MW)

[8]M. Kang, K. Kim, E. Muljadi, J. W. Park and Y. C. Kang, “Frequency

Pl3

control support of a doubly-fed induction generator based on the torque

Smart drive 08 limit,” IEEE Trans. Power Syst., vol. 31, no. 6, pp. 4575-4583, Nov. 2016.
[9]J. Van de Vyver, J. D. M. De Kooning, B. Meersman, L. Vandevelde and
T. L. Vandoorn, “Droop control as an alternative inertial response strategy
for the synthetic inertia on wind turbines,” IEEE Trans. Power Syst., vol.
(MW)
Pl4

31, no. 2, pp. 1129-1138, Mar. 2016.

[10] X. Li, D. Hui and X. Lai, “Battery energy storage station (BESS)-based
Smart drive 16 smoothing control of photovoltaic (PV) and wind power generation
fluctuations,” IEEE Trans. Sustain. Energy, vol. 4, no. 2, pp. 464-473, Apr.
2013.
[11] P. Palensky and D. Dietrich, “Demand side management: demand
(MW)
Pl5

response, intelligent energy systems, and smart loads,” IEEE Trans. Ind.
Rl (DPloss 4%) Informat., vol. 7, no. 3, pp. 381-388, Aug. 2011.
Smart drive 20 [12] C. Concordia, L. H. Fink and G. Poullikkas, “Load shedding on an
isolated system,” IEEE Trans. Power Syst., vol. 10, no. 3, pp. 1467-1472,
Time (s) Aug. 1995.
Fig. 19. Simulation results for a 180 MW infeed loss (∆𝑃𝑙𝑜𝑠𝑠 = 3%).

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2019.2952318, IEEE
Transactions on Energy Conversion
IEEE TRANSACTIONS ON ENERGY CONVERSION 9

[13] R. Azizipanah-Abarghooee, M. Malekpour, M. Paolone, and V.

Terzija. “A new approach to the online estimation of the loss of generation
size in power systems,” IEEE Trans. Power Syst., vol. 34, no. 3, pp. 2103-
2113, Nov. 2018.
[14] Y. Mu, J. Wu, J. Ekanayake, N. Jenkins and H. Jia, “Primary frequency
response from electric vehicles in the Great Britain power system,” IEEE
Trans. Smart Grid, vol. 4, no. 2, pp. 1142-1150, Jun. 2013.
[15] F. Teng, Y. Mu, H. Jia, J. Wu, P. Zeng, and G. Strbac, “Challenges on
primary frequency control and potential solution from EVs in the future
GB electricity system,” Appl. Energy, vol. 194, pp. 353-362, May 2017.
[16] Z. Akhtar, B. Chaudhuri and S. Y. Ron Hui, “Primary frequency
control contribution from smart loads using reactive compensation,” IEEE
Trans. Smart Grid, vol. 6, no. 5, pp. 2356-2365, Sep. 2015.
[17] V. Trovato, I. M. Sanz, B. Chaudhuri and G. Strbac, “Advanced control
of thermostatic loads for rapid frequency response in Great Britain,” IEEE
Trans. Power Syst., vol. 32, no. 3, pp. 2106-2117, May 2017.
[18] M. T. Muhssin, L. Cipcigan, N. Jenkins, S. Slater, M. Cheng and Z.
Obaid, “Dynamic frequency response from controlled domestic heat
pumps,” IEEE Trans. Power Syst., vol. 33, no. 5, pp. 4948-4957, Sep.
2018.
[19] D. Chakravorty, B. Chaudhuri and S. Y. R. Hui, “Rapid frequency
response from smart loads in Great Britain power system,” IEEE Trans.
Smart Grid, vol. 8, no. 5, pp. 2160-2169, Sept. 2017.
[20] J. Carmona-Sanchez, M. Barnes and J. Apsley, “Virtual energy storage:
Converting an AC drive to a smart load”, IEEE Trans. Energy Convers.,
vol. 33, no. 3, pp. 1342-1353, Sep. 2018.
[21] P. Krause, O. Wasynczuk, S. D. Sudhoff and S. Pekarek, Analysis of
electric machinery and drive systems, John Wiley & Sons, 2013.
[22] M. Malekpour, R. Azizipanah-Abarghooee, F. Teng, G. Strbac, and V.
Terzija, “Fast frequency response from smart induction motor variable
speed drives,” IEEE Trans. Power Syst., to be published.
[23] A. J. Wood, and B. F. Wollenberg., Power generation, operation, and
control, John Wiley & Sons, 2012.
[24] J. Ekanayake and N. Jenkins, “Comparison of the response of doubly
fed and fixed-speed induction generator wind turbines to changes in
network frequency,” IEEE Trans. Energy Convers., vol. 19, no. 4, pp. 800-
802, Dec. 2004.
[25] DIgSILENT PowerFactory. Gomaringen (Germany): DIgSILENT
GmbH, 2018.
[26] F. Teng and G. Strbac, “Assessment of the role and value of frequency
response support from wind plants,” IEEE Trans. Sustain. Energy, vol. 7,
no. 2, pp. 586-595, Apr. 2016.
[27] ABB, “Technical guide No. 7. Dimensioning of a drive system,
[Online]. Available: https://library.e.abb.com/public/a3ef20fdc69ccc9ac
12578800040ca95/ABB_Technical_guide_No_7_REVC.pdf. [Accessed 28
Feb. 2019].
[28] C. Concordia and S. Ihara, “Load representation in power system
stability studies,” IEEE Trans. Power App. Syst., vol. PAS-101, no. 4, pp.
969-977, Apr. 1982.
Rasoul Azizipanah-Abarghooee (S’15-M’18-SM’19) received his first PhD
degree in energy and power engineering with highest honor from Shiraz
University of Technology, Iran, in 2016. He received his second PhD degree
in electrical and electronic engineering with honor from University of
Manchester, U.K. in 2018.
He is currently a Senior Engineer at National Grid ESO providing business
support for ensuring the stable, secure, resilient and economic operation of the
modern energy system.
Dr. Azizipanah-Abarghooee serves as the Lead Guest Editor of IET
Generation, Transmission & Distribution and the Section Editor of
Encyclopedia of Electrical and Electronic Power Engineering, Elsevier. He is
the corresponding member of the CIGRE Working Group on the Impact of
Low Inertia Network on Protection and Control. He is also a Member of the
Iranian Association of Electrical and Electronic Engineers.
He is nominated in 2018 by Thomson Reuters to be the world’s top 1%
researchers in Engineering.

Mostafa Malekpour (S’18) received the B.Sc. degree from Isfahan

University of Technology, Isfahan, Iran, in 2010, and the M.Sc. degree from
Shiraz University of Technology, Shiraz, Iran, in 2013, all in power electrical
engineering. He is currently pursuing the Ph.D. degree in electrical
engineering at the University of Isfahan, Isfahan, Iran. His research interests
include power system dynamics and grid integration of renewable energy
sources. He serves as the Guest Editor of IET Generation, Transmission &
Distribution on “Challenges and New Solutions for Enhancing Ancillary
Services and Grid Resiliency in Low Inertia Power Systems”.

0885-8969 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.