2490 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO.

4, AUGUST 2021

A Multi-Objective Control Strategy for Three Phase

Grid-Connected Inverter During Unbalanced
Voltage Sag
Liang Ji , Member, IEEE, Jiabin Shi, Qiteng Hong , Member, IEEE, Yang Fu , Member, IEEE, Xiao Chang ,
Zhe Cao, Yang Mi , Member, IEEE, Zhenkun Li , and Campbell Booth

Abstract—This paper presents a new multi-objective control considered as effective solutions for environmental issues and
strategy for inverter-interfaced distributed generation (IIDG) to energy crisis. Most of them are connected to the grid via power
ensure its safe and continuous operation under unbalanced voltage electronic converters, which have enormously different perf-
sags. The proposed control strategy can effectively improve the low
voltage ride through (LVRT) capability, reduce active power oscil- ormances compared with the conventional synchronous genera-
lations, and limit overcurrent simultaneously, which are marked as tors, especially during system disturbances. Thus, to maintain
the most important control objectives of IIDG during unbalanced the safe and continuous operation of IIDGs, the effective control
voltage sags. The advanced voltage support scheme, which utilizes methods of power electronic converters are required.
positive sequence component, is firstly proposed to maximize the The control systems of IIDGs are susceptible to voltage sags,
LVRT capability of IIDG during unbalanced voltage sags. Then, to
ensure the safety of IIDG, the active power oscillation suppression which may cause the disconnections of DGs and disrupt the safe
and current limitation algorithm are designed individually. Based and continuous operation [3]. Accordingly, low-voltage ride-
on the control algorithms of such objectives, the multi-objective through (LVRT) and voltage support abilities are required for
control method, including scenario classification and reference DGs [4], [5]. Besides, during unbalanced faults, DG can perform
current determination, is then presented to achieve such three undesirable output such as high overcurrent which may damage
objectives under various system conditions simultaneously. Finally,
case studies and evaluations based on MATLAB/Simulink are car- the inverter power electronic switches due to the thermal limit
ried out to illustrate the effectiveness of the proposed method. of the device. Moreover, the inverter output active power may
appear with large oscillations, which causes the dc-link voltage
Index Terms—Active power oscillation suppression, current
limitation, inverter interfaced distributed generation, low voltage
oscillations and threatens the safety of the dc-link capacitor.
ride through, multi-objective control, voltage sag, voltage support. Therefore, to ensure the safe and continuous operation of IIDG,
an effective control method of the inverter should satisfy the
three requirements, in terms of voltage support, active power
I. INTRODUCTION oscillation suppression and current limitation.
N RECENT years, increasing attention has been paid to Many researchers have put effort to solve such challenges
I the integration of renewable energy into power systems
to address the challenges of the global climate change [1],
of DG control during unbalanced voltage sag [6]–[23]. Many
of them are focusing on voltage support and current limitation
[2]. Renewable energy resources, such as solar and wind, are [6]–[18]. In [6]–[10], the voltage at the point of common cou-
pling (PCC) is well supported by injecting the constant reactive
Manuscript received February 14, 2020; revised May 23, 2020 and August 11, power. However, such methods have difficulty in selecting an
2020; accepted September 12, 2020. Date of publication September 18, 2020; appropriate power reference during the specific voltage sag
date of current version July 23, 2021. This work was supported in part by the
National Natural Science Foundation of China under Grant 61873159, in part
conditions, which may cause an overcurrent risk to the inverter.
by Shanghai Engineering Research Center of Green Energy Grid-Connected Therefore, Wang [11] proposed an LVRT strategy that provide
Technology under Grant 13DZ2251900, and in part by Electrical Engineering 1.5% of the reactive current per 1% voltage sag, while Lee [12]
Shanghai class II Plateau Discipline, and Shanghai Key Laboratory of Power
Station Automation Technology. Paper no. TPWRD-00240-2020. (Correspond-
injected 2% of the reactive current with the same voltage sag
ing author: Yang Fu.) according to the E. ON grid code [5]. Meanwhile, the phase cur-
Liang Ji, Jiabin Shi, Yang Fu, Zhe Cao, Yang Mi, and Zhenkun Li are rents are well controlled within the limits. However, the voltage
with the Department of Electrical Power Engineering, Shanghai University
of Electric Power, Shanghai 200090, China (e-mail: [email protected];
support capability was not the prior target of such methods. Thus,
[email protected]; [email protected]; [email protected]; miyangmi the voltage level cannot be effectively supported, especially
@163.com; [email protected]). during moderate or low voltage sag conditions. To address such
Qiteng Hong and Campbell Booth are with the Department of Electronic
and Electrical Engineering, University of Strathclyde, Glasgow G1 1XQ, U.K.
problems, the advanced voltage control techniques upon positive
(e-mail: [email protected]; [email protected]). sequence (PS) and negative sequence (NS) were developed in
Xiao Chang is with the State Grid Shanxi Electric Power Research Institute [13]–[15], which can maximize the voltage support and regulate
Taiyuan, Shanxi 030001, China (e-mail: [email protected]).
Color versions of one or more of the figures in this article are available online
the phase voltages within the desired range. The main disad-
at https://ieeexplore.ieee.org. vantage of such control method is the high requirement of the
Digital Object Identifier 10.1109/TPWRD.2020.3025158 inverter capacity. For the low capacity inverter, the controller
0885-8977 © 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
JI et al.: MULTI-OBJECTIVE CONTROL STRATEGY FOR THREE PHASE GRID-CONNECTED INVERTER 2491

may not work properly due to the relatively low voltage support
capability. By considering the effect of the DG capacity limit, a
method for low capacity DG was proposed in [16]. Although the
voltage support target is well ensured, the safety of the dc-link
capacitor of the inverter cannot be guaranteed since the active
power oscillation is ignored. Besides, a new solution based on the
Lagrange multiplier method was also investigated in [17], [18]
to maximize the positive sequence voltage support. Also, this
method is easy-to-implemented for different control objectives, Fig. 1. Simplified overall structure of a two-stage three-phase IIDG system.
e.g., minimize the negative sequence voltage and the voltage
unbalanced factor (VUF). However, the simultaneous realization
of multi-objective is not considered in these works, which is simultaneously realize and coordinate three critical control ob-
essential for the safe operation of the inverter duration the LVRT jectives, i.e., voltage support, current limitation and osci-llation
period. suppression, during all voltage sag scenarios. Meanwhile, the
Some researchers in recent years put effort to suppress active proposed method only applies positive sequence voltage con-
power oscillations [19]–[23]. These methods effectively damped trol, which can maximize the voltage support capability while
the active power oscillations by adequately regulating the ratio less-affected by the inverter capacity. Moreover, the proposed
of active and reactive current references of both positive and control strategy can make full use of the inverter capacity. In
negative sequences. In [19], [20], in addition to suppressing cases that the inverter has spare capacity after fully realizing
the active power oscillation, the non-MPPT operation mode such three objectives, the proposed method can use such spare
was also established for the boost converter. Therefore, both capacity to achieve extra control objectives, such as actively
oscillation and increase of the dc-link voltage can be avoided. improve system stability and voltage balance.
The main difference between [19] and [20] is the based reference This paper is organized as follows. Section I gives a brief
frames. In [21], a control strategy upon dq-frame was presented review of the state-of-the-art control strategies of IIDG. Sec-
to deal with the power oscillation. However, both active and tion II describes the operation of the inverter during unbalanced
reactive power references were user-defined, it is difficult to voltage sag. Section III discusses the active power control and
select a suitable value during a given voltage sag condition the proposed control objectives with their reference current
which cannot guarantee the safety of the inverter in all possible calculation procedure. Section IV presents a multi-objective
cases. In [22], [23], the control methods based on dq-frame are control strategy to achieves these objectives simultaneously. In
presented to improve the output performance of the inverter Section V, the effectiveness of the proposed control strategy is
by coordinating the current amplitude and power oscillation evaluated by various case studies. Finally, the conclusions are
constrains. However, due to the relatively low injected reactive briefly drawn in Section VI.
current, the voltage support capability in these methods may not
be fully exploited. In this way, the IIDG may face with high risk II. INVERTER OUTPUT UNDER UNBALANCED VOLTAGE SAG
of disconnection, which influences the safe and stable operation
This section describes the typical inverter operation during
of the power system.
unbalanced voltage sag conditions, which is vital for the cal-
By investigating the existing control methods and practical
culation of the reference currents and the realization of control
experiences, it can be concluded that the ideal control of IIDG
objectives in Section III. Fig. 1 describes the simplified structure
during the voltage sag should consider the above objectives
and control scheme of a typical IIDG system, which includes
simultaneously, in terms of voltage support, active power osci-
two stages: 1) the PV-side boost converter which extracts the
llation suppression and current limitation, as to ensure the safe
PV power and delivers to dc-link, and 2) the grid-side inverter
and continues operation of IIDG. In this paper, a new multi-
which delivers the extracted PV power to the ac grid.
objective control method based on the dq-frame for three-phase
Under unbalanced voltage sag, the voltage vector at PCC is
IIDG is proposed to achieve these control objectives. Firstly,
performed in the synchronous reference frame as
to fully exploit the inverter capability, the active power con-  + 
trol method is designed to inject the maximum allowed active − V cos(ωt + ϕ+ ) + V − cos(−ωt + ϕ− )
v =v +v =
+
pow-er. Secondly, by considering the phase voltage limits, a V + sin(ωt + ϕ+ ) + V − sin(−ωt + ϕ− )
positive sequence voltage control method is presented to maxi- (1)
mize the voltage support capability under various voltage sags. where v + and v − denote the positive and negative sequence
Thirdly, to ensure the safety of the inverter, the control algorithm voltages, V + and V − denote the amplitudes of v + and v − , and
to limit the active power oscillations and current amplitudes ϕ+ and ϕ− are their initial phase angle respectively. The output
are desig-ned individually. Finally, the multi-objective control current of the inverter in the SRF can be written as [13]
strategy involving scenario classification and reference current  + 
− Ip cos(ωt + ϕ+ ) − Iq+ sin(ωt + ϕ+ )
determi-nation is proposed, which realizes the above control i=i +i = +
+
Ip sin(ωt + ϕ+ ) + Iq+ cos(ωt + ϕ+ )
objectives simultaneously by coordinating the current references  − 
of each control objective. Compared with the existing methods, Ip cos(−ωt + ϕ− ) − Iq− sin(−ωt + ϕ− )
+ − (2)
the notable advantage of the proposed method is that it can Ip sin(−ωt + ϕ− ) + Iq− cos(−ωt + ϕ− )

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
2492 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO. 4, AUGUST 2021

where the subscripts “p” and “q” represent the active and reactive Thus, according to the known value of the generated PV power,
current components, and i+ and i− are the positive and negative the reference current for Ip+ can be obtained as
sequence currents.
Ppv
Based on the system structure, the relation of the sequence +
Ip_ini = (7)
voltages at PCC and grid side bus can be established as V+
+
where Ip_ini is the initial active current reference that delivers
di+
v + = vg+ + Lg + R g i+ (3) the initial PV power to the gird. The subscribe “ini” denotes that
dt this variable is an initial quantity, which may be modified due
di− to the phase current limitation. Ppv is the initial PV power at
v − = vg− + Lg + R g i− (4)
dt maximum power point (MPP).
where Rg and Lg represent the grid resistance and inductance
respectively, vg+ and vg− are the positive and negative sequence B. Objective 1: Positive Sequence Voltage Support
voltages at the grid side bus. When designing the proposed The conventional voltage support methods may not be able
control strategy, the grid impedances (Rg and Lg ) are required to fully support the voltage [6]–[12] or could be affected by the
to be known [13], which can be obtained by applying the online inverter capacity [13]–[15]. To address the above limitations, this
impedance estimator [24], [25]. In order to strengthen the control section develops a method to maximize the positive sequence
characteristic of the proposed method, the grid impedance is voltage support. Meanwhile, as to avoid overvoltage during the
assumed to be known in this paper. LVRT period, the following constraint must be fulfilled [7].
Considering the positive sequence voltage support objective
utilized in this paper, only equation (3) is used. Thus, by inserting Vmax = max{Va , Vb , Vc } ≤ Vupper (8)
(1) and (2) into (3), the relation among v + and vg+ can be where Vmax is the maximum voltage amplitude, and Vupper is
described with V + and Vg+ , which is written as [18] the phase voltage limit, which is 1.1 p.u. [14]. To achieve this,
 the phase voltage amplitudes are established at first, which are
2 2
Vg+ = (V + − ωLg Iq+ − Rg Ip+ ) + (ωLg Ip+ − Rg Iq+ ) ⎧ 
⎪
⎪ Va = [VL cos(γ)]2 + [VS sin(γ)]2
(5) ⎪
⎨ 
Vb = [VL cos(γ + π3 )]2 + [VS sin(γ + π3 )]2 (9)
III. CONTROL OBJECTIVES ⎪
⎪ 
⎪
⎩V = [V cos(γ − π )]2 + [V sin(γ − π )]2
In this section, the active power control is firstly described. c L 3 S 3
Then, three control objectives of the proposed methods are where
introduced individually (Please note that the coordination of the
VL = V + + V −
three objectives are introduced in Section IV), which are (10)
VS = V + − V −
- Objective 1: Positive sequence voltage support, which aims
to maximize the positive sequence voltage within the phase ϕ+ + ϕ −
γ= (11)
voltage limits. 2
- Objective 2: Active power oscillation suppression, which With equations (8)–(11), the maximum phase voltage ampli-
aims to limit the active power oscillations within the per- tude Vmax can be expressed as follow
missible range.
- Objective 3: Current limitation, which aims to avoid over-
2
Vmax = (V + )2 + (V − )2 + 2V + V − λmax (12)
current risk. where
Since the reference current calculation of objectives 2 and
2π 2π
3 depends on different system operation scenarios, the detailed λmax = max cos(2γ), cos 2γ + , cos 2γ −
calculation procedures are carried out in Section IV-C (reference 3 3
(13)
current determination). In this section, both objectives 2 and 3
Then, by replacing Vmax in (12) with Vupper , the maximum
are briefly discussed.
positive sequence voltage reference is obtained as

A. Active Power Control +
Vref = −V − λmax + (V − λmax )2 − [(V − )2 − (Vupper )2 ]
To fully exploit the DG capacity and mitigate the risk of (14)
sudden active power loss during the voltage sags, the proposed Finally, by solving (14) and (5), the initial positive sequence
control method is designed to inject the maximum allowed active reactive current that maximizes the voltage support is deduced
power of the inverter to the grid, which extracted from PV array. √
B − B2 − AC
To design the active power control, the average active power in Iq_ini =
+
(15)
A
the dq domain should be expressed as follow
where
P = V + Ip+ + V − Ip− (6) ⎧
⎨A = Rg2 + Xg2
For simplicity, the active power P in this paper is injected B = Xg Vref
+
(16)
⎩
via positive sequence active current Ip+ , while Ip− remains zero. C = (Vref − Rg Ip+ )2 + (Xg Ip+ )2 − (Vg+ )2
+

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
JI et al.: MULTI-OBJECTIVE CONTROL STRATEGY FOR THREE PHASE GRID-CONNECTED INVERTER 2493

+ where
In addition to maximizing the voltage support capability, Iq_ini
⎧
is also utilized to classify the scenarios in Section IV-B. Al- ⎪
⎪ IpL = Ip+ + Ip−
+
though the injection of Iq_ini can maximize the voltage support, ⎨
IpS = Ip+ − Ip−
it may fail to achieve objective 2 or 3 in some scenarios (e.g., (23)
⎪
⎪ I qL = Iq+ + Iq−
+
serious voltage sag). Thereby, Iq_ini should be modified to a ⎩
IqS = Iq+ − Iq−
lower value, which is discussed further in Section IV-C.
π π
δ= γ, γ +
,γ − (24)
C. Objective 2: Active Power Oscillation Suppression 3 3
The current amplitudes Ia , Ib , and Ic can be calculated by
During unbalanced voltage sag, the output active power may
inserting different δ in (24) to (22). As to inject the maximum
oscillate with double grid frequency, which leads to the dc-link
allowed current without overcurrent, equation (25) is defined
voltage oscillating with the same frequency. This oscillating
voltage can notably affect the long-term life of the dc-link max(Ia , Ib , Ic ) = Ilim (25)
capacitor and even endanger its safety. The suppression of the
dc-link voltage oscillations can be achieved by reducing the where Ilim denotes the phase current limit of the inverter. Ilim
active power oscillations [19]–[21]. The relation among the dc is an adjustable value, which depends on the capacity and
voltage and active power oscillations can be expressed as [22] thermal limit of the power transistors. The higher maximum
allowable current can bring stronger voltage support ability. In
P̃peak this paper, the value for Ilim is setting as 1.2 times of the rated
Ṽdc_peak = (17) current of the inverter [22]. With equations (22)–(25), the current
ωCdc Vdc
references can be calculated to avoid the overcurrent risk. The
where Ṽdc_peak and P̃peak are the amplitudes of the oscillating detailed calculation procedure for this objective is discussed in
dc-link voltage and active power respectively, Cdc is the dc Section IV-C (after the scenario classification).
capacitance, and Vdc is the rated dc-link voltage.
To ensure the safety of the capacitor, Ṽdc_peak must be con- IV. MULTI-OBJECTIVE CONTROL STRATEGY
trolled within the limit Ṽdc_lim . Based on the requirement of
a typical commercial film capacitor (e.g., MKP-B32674 [26]), The previous section has discussed the active power control
Ṽdc_lim should be within 20% of Vdc . The selection of this ratio and three control objectives individually. In this section, a new
mainly relates to the constraint of the dc-link capacitor and the control strategy to achieve multiple objectives is presented. In
requirement of IIDG users. In this paper, Ṽdc_lim is set as 10% this way, the inverter can maximize the voltage support capabil-
of Vdc , which is a compromise choice. Thus, the limit for P̃peak ity while ensuring its safety. Moreover, the inverter can actively
improve the system stability and voltage balance by full using
can be calculated upon (17), which is defined as P̃lim .
of the capacity for some operation scenarios.
To reduce the active power oscillations within P̃lim , the in-
stantaneous active power injected to the grid is formulated as a
function of the current and voltage vector [20] A. Introduction to the Process of the Control Strategy
Fig. 2 shows the process of the proposed multi-objective
p = v · i = (P̄ + P̃ ) (18) control strategy, which includes four main parts: 1) voltage
where the average and oscillating terms P̄ and P̃ are sag detection, 2) scenario classification, 3) reference current
determination, and 4) inner current controller. The details of
P̄ = v + · i+ + v − · i− (19) each step are as follow
P̃ = v + · i− + v − · i+ (20) Step 1 Voltage sag detection: The measured phase voltages
Vabc at PCC are transformed to positive and negative sequence
By applying equation (1) and (2), and replacing P̃ in (20) with voltages (V + and V − ), utilizing decoupled double synchronous
P̃lim , equation (20) can be rewritten as reference frame phase-locked loop (DDSRF-PLL) [27]. If V +
 falls below the preset border value (0.9 p.u. upon the Germany
2 2
P̃lim = (V + Ip− + V − Ip+ ) + (V − Iq+ − V + Iq− ) (21) LVRT standards [5]), the detection block generates a fault signal
Equation (21) is to calculate the current references that ensure to activate the proposed control strategy.
the active power oscillation suppression objective. The detailed Step 2 Scenario classification: Once the voltage sag has
calculate procedure is discussed in Section IV-C. been detected, this block is activated to classify the operation
+ +
scenario. Firstly, the initial reference currents Ip_ini and Iq_ini
are calculated upon equations (7) and (15). Then, the minimum
D. Objective 3: Current Limitation + +
reactive and active current references Iq_min and Ip_min are
As to avoid the overcurrent risk, the phase current amplitudes calculated by equations (26) and (29), which can evaluate the
Iabc should be described as (22) by developing equation (2) voltage sag severity and the initial PV power level (discussed in
 section IV-B). Based on the proposed scenario classification
Iabc = (IpL cos δ + IqL sin δ)2 + (−IqS cos δ + IpS sin δ)2 principle (judgement box in Fig. 2), three system operation
(22) scenarios can be classified. Each classified scenario matches

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
2494 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO. 4, AUGUST 2021

where
P̃lim
+
Iq_1 = (27)
V−
+
Iq_2 = Ilim (28)
+ +
where Iq_1 and Iq_2 denote the maximum allowed positive
sequence reactive current references that ensure objectives 2
and 3 respectively if only Iq+ is injected. They can be calculated
by setting Ip+ , Ip− and Iq− in (21) and (22) as zeros respectively.
It should be mention that the voltage support requirement in
this research has a higher priority than the active power delivery.
Consequently, the inverter can inject the active power into the
grid only in case that the voltage has been well-supported to
1.1pu.. In case that the voltage can be fully supported while
+ +
ensuring objectives 2 and 3 (Iq_ini < Iq_min ), the level of the
initial PV power should be evaluated. This can be done by
+
comparing Ip_ini with the maximum allowed active current
references, which are defined as follow
min = min(Ip_1 , Ip_2 )
+ + +
Ip_ (29)
where
 2

 P̃
=
lim 2
Fig. 2. Overview of the proposed multi-objective control strategy.
+
Ip_1 − (Iq_ini
+
) (30)
V−

2
+
Ip_2 = (Ilim )2 − (Iq_ini
+
) (31)
with a specific current injection mode, which can fully exploit + +
the inverter capacity and ensure its safety. where Ip_1 and Ip_2 denote the maximum allowed positive
Step 3 Reference current determination: Depending on the sequence active current references that ensure objectives 2 and
current injection mode of each scenario, four reference current 3 respectively when both Iq+ and Ip+ are injected. They can be
components are determined by the coordination of the three
+
calculated by inserting Iq_ini to (21) and (22), and setting Ip− and
− + +
objectives (discussed in section IV-C). Therefore, three control Iq as zeros respectively. By comparing Iq_ini with Iq_min and
objectives can be achieved at the same time, and the inverter + +
Ip_ini with Ip_min , three different operation scenarios can be
capacity can be fully exploited. obtained. The detailed discussions are given as follow:
Step 4 Inner current controller: This block is utilized to gen- + +
1) Scenario 1 (serious voltage sag): If Iq_ini > Iq_min , the
erate the voltage signals for PWM modulation by controlling the +
injection of the initial reference current Iq_ini will fail objectives
+ + − −
determined reference currents (Iq_ref , Ip_ref , Iq_ref , and Ip_ref ). 2 and 3, which endanger the safety of the inverter. Thus, this
Then, PWM drives the switches of the inverter and inherently scenario is defined as serious voltage sag scenario (left side of
realizing the control objectives. Some details of the inner current Fig. 2). As to fully support the voltage, only Iq+ is required to
controller are as shown in Fig. 3 of section IV-D, where four be injected in this scenario.
proportional and integral (PI) controllers are contained in the 2) Scenario 2 (moderate voltage sag and high initial PV
inner controller to regulate four current components. +
power): If Iq_ini +
< Iq_min +
, the injection of Iq_ini can maximize
the voltage support, and both objectives 2 and 3 are ensured.
Therefore, the voltage sag is defined as moderate sag. To fully
B. Scenario Classification
exploit the inverter capacity, the active power can be injected.
Once the voltage sag is detected, the initial current references + + +
However, due to that Ip_ini > Ip_min , the injection of Ip_ini will
+ +
Ip_ini (7) and Iq_ini (15) are computed. They are used to evaluate result in the overcurrent or high active power oscillations. So, the
the initial PV power level and the voltage sag severity respec- initial PV power is defined as high level (middle side of Fig. 2).
tively, thus to classify the operation scenarios. In this scenario, both Ip+ and Iq+ are required to be injected.
Considering the inverter safety constraints, the realization of 3) Scenario 3 (moderate voltage sag and low initial PV power):
the voltage support must ensure objectives 2 and 3 simultane- + + + +
If Iq_ini < Iq_min and Ip_ini < Ip_min , the combined injection of
ously. Therefore, the severity of the sag should be evaluated + +
Iq_ini and Ip_ini can fulfill all control requirements. Therefore,
+
by comparing Iq_ini with the calculated reference currents upon this scenario is defined as moderate voltage sag and low initial
objectives 2 and 3, which defined as follow PV power (right side of Fig. 2). To fully use the capacity and
decrease the negative sequence voltage, Iq− is injected combined
min = min(Iq_1 , Iq_2 )
+ + +
Iq_ (26) with Ip+ and Iq+ .

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
JI et al.: MULTI-OBJECTIVE CONTROL STRATEGY FOR THREE PHASE GRID-CONNECTED INVERTER 2495

Fig. 3. Control diagram of the two-stage IIDG system.

TABLE I 3) Scenario 3 (moderate voltage sag and low initial PV power):

PERFORMANCE OF IIDG WITH THE PROPOSED CONTROL STRATEGY
In this scenario, both Ip+ , Iq+ , and Iq− are injected. As the
+ +
injection of Iq_ini and Ip_ini can fulfill the voltage maximization
and initial PV power delivery, the current references for Ip+ and
+ +
Iq+ are chosen as Iq_ini and Ip_ini respectively. But the reference
current for Iq− must ensure objectives 2 and 3 simultaneously.
+ +
Firstly, by replacing Ip+ and Iq+ with Ip_ini and Iq_ini in
−
(21), and setting Ip as zero, the current reference that ensures
−
objective 2 is obtained, which defined as Iq_1
 2

−  P̃ V− +
2
C. Reference Current Determination − V +
Iq_1 = + Iq_ini −  lim
− Ip_ini (35)
V V + V +
Based on the operation scenario and corresponding current
injection mode, the final reference currents of each scenario can Then, by using the same way for equation (22), the current
be determined. Table I summarizes the performance of the IIDG reference that achieves objective 3 can be obtained as
with the proposed strategy under different scenarios. 
1) Scenario 1 (serious voltage sag): In this scenario, only Iq+ −
Iq_abc = A + (A)2 − B (36)
is injected. To fully support the positive sequence voltage and
ensure objectives 2 and 3, the final current reference for Iq+ where
should be determined as A = Iq_ini
+
cos(2δ) − Ip_ini
+
sin(2δ)
(37)
+
Iq_ref = Iq_
+
(32) B = (Ip_ini
+
)2 + (Iq_ini
+
)2 − (Ilim )2
min

− − − −
+
where Iq_min is obtained by (26), which is lower than Iq_ini +
. and Iq_abc = {Iq_a , Iq_b , Iq_c } represent the calculated neg-
Thus, the voltage support that increases Vmax to 1.1 p.u. is failed, ative sequence reactive current reference when the current in
as shown in Table I. Nevertheless, the inverter still maintains phases a, b and c reaches to Imax respectively, which can be
relatively high voltage support capability by injecting the max- calculated by inserting different δ in (24) to equation (36). To
+
imum allowed value of Iq_ref based on (1). ensure the current limitation objective in any condition, the
2) Scenario 2 (moderate voltage sag and high initial PV reference current should be defined as
power): In this scenario, both IP+ and Iq+ are injected. Since the −
Iq_2 −
= min(Iq_a −
, Iq_b −
, Iq_c ) (38)
voltage support can be maximized, the final reference current
for Iq+ can be chosen as Finally, as to achieve objective 2 and 3 at the same time, the
final current reference for Iq− should be chosen as the minimum
+
Iq_ref = Iq_ini
+
(33) value between Iq_1− −
and Iq_2 , which is
As to avoid the overcurrent and high active power ripple while − − −
Iq_ref = min(Iq_1 , Iq_2 ) (39)
delivering the active power to the grid, the final active current
reference for Ip+ must be chosen as the minimum value among As concluded in Table I, the voltage support maximization and
+
Ip_1 +
and Ip_2 described in (29) initial PV power delivery requirements are not always ensured
due to the phase current and active power oscillation limits. How-
+
Ip_ref = min(Ip_1
+ +
, Ip_2 ) (34) ever, the discussed three control objectives (e.g. voltage support,

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
2496 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO. 4, AUGUST 2021

active power oscillation suppression and current limitation) are TABLE II

SYSTEM AND CONTROL PARAMETERS
ensured simultaneously for all scenarios.

D. Configuration of the Control Scheme

The control scheme of the two-stage IIDG system is shown
in Fig. 3. During normal operation, the boost converter is op-
erating at MPPT mode. By adopting the dc voltage controller
which keeps the power balance and dc voltage stabilization,
the inverter can operate at constant power (PQ) control mode.
During the voltage sag, the proposed multi-objective control
strategy is activated to improve the voltage support capability
and ensure the safety of the inverter. If the inverter can handle
the maximum PV power (Ip_ref +
≥ Ip_ini
+
, e.g., in scenario 3),
the boost converter remains MPPT control mode. Otherwise,
+ +
if Ip_ref < Ip_ini (e.g., in scenarios 1 and 2), the inverter is
unable to deliver the maximum PV power to the grid due to the
safety constraints, the boost converter switches to non-MPPT
mode using the control method presented in [19], and reduces
the generated PV power to match the maximum allowed active
power (Pref ) of the inverter. In this paper, the design of the TABLE III
non-MPPT control is based on [19]. VOLTAGE SAG AND INITIAL PV POWER REFERENCE OF EACH CASE

V. CASE STUDY AND EVALUATION

To verify the effectiveness of the proposed multi-objective
control strategy, three case studies corresponding to the three
scenarios mentioned above are carried out. Firstly, the results of
the scenario classification of each case are presented. Then, the TABLE IV
RESULTS OF SCENARIO CLASSIFICATION
performance of the inverter under each scenario is discussed.
The case studies and simulations are upon MATLAB/Simulink.
Fig. 1 shows the tested two-stage system topology, including a
PV array, a boost converter, a 15 kVA, 400 V inverter with an LC
filter, which connected to the ac voltage source through the con-
ductor. The voltage sags are emulated by the ac voltage source
[12]–[17]. Fig. 3 shows the control scheme of the two-stage
system, including PV-side boost converter control and grid-side The evaluations of the three cases are discussed in follow-
inverter control. ing sections. The classical balanced positive sequence control
Table II lists the detailed system and control parameters. In all (BPSC) strategy [29] is presented as a compared method to fur-
tested cases, the phase current limit Ilim is setting as 1.2 p.u., and ther illustrate the advancement of the proposed control method.
the dc voltage oscillation limit Ṽdc_lim is setting as 10% of Vdc BPSC is an easy-to-implemented and widely-used method due
as stated in Section III-B. Therefore, the active power oscillation to the current balancing improvement during unbalanced volt-
limit P̃lim is calculated as 0.419 p.u. upon (17). age sags. The reactive current injection principle of the BPSC
method is upon the E. ON grid code [5], which requires 2% of
A. Case Study on Scenario Classification the injected reactive current for per 1% voltage sag.
To clearly show the operation mechanism of the proposed
control strategy, Table III lists the settings of the three cases and B. Case 1: Serious Voltage Sag (Only I +
q Is Injected)
Table IV displays the results of scenario classification.
− Fig. 4 shows the performance of the IIDG during serious
g = 0.45p.u., V g = 0.37p.u.): As Iq_ini >
+
Case 1 (V +
+
voltage sag. The proposed control strategy is activated at t = 0.1 s
Iq_min , this operation scenario is classified as serious voltage and lasts until the sag is cleared at t = 0.3 s. As shown in Fig. 4(b),
sag. the positive sequence voltage V + is supported from 0.45 p.u. to
−
g = 0.75p.u., V g = 0.25p.u.): As Iq_ini <
+
Case 2 (V + 0.58 p.u. However, due to the of the active power oscillation
+ + +
Iq_min but Ip_ini > Ip_min , this scenario is classified as moderate constraint, the voltage support is not maximized (Vmax <1.1
voltage sag and high initial PV power. p.u.), which is shown in Fig. 4 (a). However, by injecting the
−
g = 0.83p.u., V g = 0.17p.u.): As Iq_ini <
+
Case 3 (V + total reactive power to the grid, the voltage support capability is
+ + +
Iq_min and Ip_ini < Ip_min this scenario is classified as moderate still fully enhanced. Fig. 4 (c) and (d) indicate that both phase
voltage sag and low initial PV power. currents and active power oscillations are well controlled within

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
JI et al.: MULTI-OBJECTIVE CONTROL STRATEGY FOR THREE PHASE GRID-CONNECTED INVERTER 2497

Fig. 4. Simulation of the proposed strategy for Case 1: (a) phase voltages (b) Fig. 5. Simulation of the BPSC strategy for Case 1: (a) phase voltages, (b)
sequence voltages, (c) phase currents, (d) output power, (e) dc-link voltage. sequence voltages, (c) phase currents, (d) output power, (e) dc-link voltage.

Ilim and P̃lim respectively. As the maximum allowed active the maximum PV power due to the current limitation. Therefore,
power (Pref ) of the inverter in this scenario is zero, which is the non-MPPT of the boost converter is activated to reduce
lower than the initial PV power Ppv , the boost converter switches the generated PV power. As depicted in Fig. 6(d), the inverter
from MPPT to Non-MPPT mode. Therefore, by regulating the output active power P̄ decreased from 1 p.u. to 0.76 p.u., and
dc-link voltage, the average active power P̄ decreases to zero, the overcurrent risk can be inherently avoided. Meanwhile, the
and the average dc-link voltage V̄dc remains 1 p.u., which are active power oscillations are well controlled within the limits,
shown in Fig. 4(d) and (e). Moreover, due to the suppression of which is 0.301 p.u.. As a result, the dc-link voltage oscillation
the active power oscillation, the dc voltage oscillation is well is also limited to 7% of Vdc , which is lower than Ṽdc_lim .
controlled within Ṽdc_lim (10% of Vdc ). Fig. 7(a) and (b) demonstrate the voltage support capability
The results of the BPSC method are given in Fig. 5. As seen of the BPSC method is not fully exploited, where Vmax <1.1
in Fig. 5 (b), V + is supported by 0.1 p.u., which is lower than p.u. and V + is supported to 0.78 p.u. (lower than the proposed
the proposed strategy (0.13 p.u.). As compared Fig. 4(c) with method, 0.85 p.u.). As depicted in Fig. 7(c) and (d), both
Fig. 5(c), both methods can control the phase currents within the phase currents and power oscillations are within the constraints.
limit. However, the BPSC method does not as effective power Meanwhile, although BPSC strategy shows higher active power
oscillation suppression performance as the proposed method. injection ability (0.91 p.u.) than the proposed method (0.68 p.u.),
Fig. 5(d) and (e) show that both P̃peak and Ṽdc_peak beyond the the delivered reactive power is obviously low, which is 0.24 p.u.
allowed limit.
D. Case 3: Moderate Voltage Sag and Low Initial PV Power
−
p , I q , and I q are Injected)
(I + +
C. Case 2: Moderate Voltage Sag and High Initial PV Power
p and I q are Injected)
(Both I + +
Fig. 8 presents the performance of IIDG during moderate
Fig. 6 shows the output of IIDG during moderate voltage sag voltage sag and low initial PV power scenario. As shown in
and high initial PV power scenario. As shown in Fig. 6(a) and Fig. 8(a), the voltage support capability is maximized, with Vmax
(b), the voltage support capability is maximized, where V + is reaches to 1.1 p.u.. Meanwhile, V + is supported from 0.83 p.u.
supported from 0.75 p.u. to 0.85 p.u., and Vmax reaches to 1.1 to 0.95 p.u. as shown in Fig. 8(b). As observed in Fig. 8(c), the
p.u.. Fig. 6(c) shows that the phase currents are well limited inverter capacity is fully used ( Imax = Ilim ) by the combined
within Ilim . In this scenario, the inverter is still unable to handle injection of Iq− . And due to that, V − is decreased from 0.17 p.u.
Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
2498 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO. 4, AUGUST 2021

Fig. 8. Simulation of the proposed strategy for Case 3: (a) phase voltages, (b)
Fig. 6. Simulation of the proposed strategy for Case 2: (a) phase voltages, (b) sequence voltages, (c) phase currents, (d) output power (e) dc-link voltage.
sequence voltages, (c) phase currents, (d) output power, (e) dc-link voltage.

Fig. 7. Simulation of the BPSC strategy for Case 2: (a) phase voltages, (b) Fig. 9. Simulation of the BPSC strategy for Case 3: (a) phase voltages, (b)
sequence voltages, (c) phase currents, (d) output power, (e) dc-link voltage. sequence voltages, (c) phase currents, (d) output power, (e) dc-link voltage.

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
JI et al.: MULTI-OBJECTIVE CONTROL STRATEGY FOR THREE PHASE GRID-CONNECTED INVERTER 2499

to 0.15 p.u., which helps to reduce the voltage unbalanced factor [5] Grid Code High and Extra High Voltage. Bayreuth, Germany: E. ON Netz
(VUF). As revealed in Fig. 8(d), because the inverter can handle GmbH, Apr. 2006.
[6] F. Wang, J. L. Duarte, and M. A. M. Hendrix, “Pliant active and re-
the maximum PV power, the boost converter is operating at the active power control for grid-interactive converters under unbalanced
MPPT mode. Therefore, the initial PV power delivery in this voltage dips,” IEEE Trans. Power Electron., vol. 26, no. 5, pp. 1511–1521,
scenario is fulfilled with P̄ remaining at 0.4 p.u.. Meanwhile, May 2011.
[7] M. M. Shabestary and Y. A. I. Mohamed, “An analytical method to
the delivered reactive power Q̄ is relatively high, which is 0.93 obtain maximum allowable grid support by using grid-connected con-
p.u.. Moreover, as depicted in Fig. 8(d) and (e), the oscillations verters,” IEEE Trans. Sustain. Energy, vol. 7, no. 4, pp. 1558–1571,
of both active power and dc-link voltage are extremely low in Oct. 2016.
[8] Z. Shuai, C. Shen, X. Yin, X. Liu, and Z. J. Shen, “Fault analysis of
this scenario. inverter-interfaced distributed generators with different control schemes,”
As for the BPSC method, Fig. 9(a) and (b) indicate that the IEEE Trans. Power Del., vol. 33, no. 3, pp. 1223–1235, Jun. 2018.
voltage support ability in this scenario is extremely low, with V + [9] W. Guo, L. Mu, and X. Zhang, “Fault models of inverter-interfaced
distributed generators within a low-voltage microgrid,” IEEE Trans. Power
supported by 0.01 p.u. As demonstrated in Fig. 9(c), although Del., vol. 32, no. 1, pp. 453–461, Feb. 2017.
the phase currents remain balanced, the amplitude of the output [10] X. Guo, W. Liu, and Z. Lu, “Flexible power regulation and current-
current (Imax = 0.51 p.u.) is obviously lower than the limit Ilim , limited control of the grid-connected inverter under unbalanced grid
voltage fau-lts,” IEEE Trans. Ind. Electron., vol. 64, no. 9, pp. 7425–7432,
which means the inverter capacity is not fully used. Comparing Sep. 2017.
Fig. 9(d) with Fig. 8(d), it is obvious that the delivered reactive [11] Y. Wang, P. Yang, X. Yin, and Y. Ma, “Evaluation of low-voltage
power of BPSC method is lower than the proposed method. ride-through capability of a two-stage grid-connected three-level photo-
voltaic inverter[C],” in Proc. IEEE Int. Conf. Elect. Mach. Syst., 2014,
Moreover, due to the injection of Iq− the proposed strategy shows pp. 822–828.
lower active power oscillation (0.07 p.u.) than the BPSC method [12] C. Lee, C. Hsu, and P. Cheng, “A low-voltage ride-through technique for
(0.09 p.u.). grid-connected converters of distributed energy resources,” IEEE Trans.
Ind. Appl., vol. 47, no. 4, pp. 1821–1832, Aug. 2011.
[13] M. M. Shabestary and Y. A. I. Mohamed, “Asymmetrical ride-through
VI. CONCLUSION and grid support in converter-interfaced DG units under unbalanced
conditions,” IEEE Trans. Ind. Electron., vol. 66, no. 2, pp. 1130–1141,
System disturbances (e.g. voltage sags) can notably affect the Feb. 2019.
safe and continuous operation of IIDGs. Many existing LVRT [14] A. Camacho, M. Castilla, J. Miret, R. Guzmán, and A. Borrell, “Reactive
power control for distributed generation power plants to comply with
strategies show relatively weak voltage support capability and voltage limits during grid faults,” IEEE Trans. Power Electron., vol. 29,
may be influenced by the DG capacity. Moreover, there are no. 11, pp. 6224–6234, Nov. 2014.
very few research activities that have considered multi-objective [15] M. M. Shabestary and Y. A. I. Mohamed, “Advanced voltage support and
active power flow control in grid-connected converters under unbalanced
control of IIDG, which is necessary during the voltage sags. conditions,” IEEE Trans. Power Electron., vol. 33, no. 2, pp. 1855–1864,
Therefore, this paper has proposed a multi-objective control Feb. 2018.
strategy for IIDG to improve its dynamic performance, while [16] J. Miret, A. Camacho, M. Castilla, J. L. G. de Vicuña, and J. de la
Hoz, “Reactive current injection protocol for low-power rating distributed
being capable of well-adapting to various operation scenarios. generation sources under voltage sags,” IET Power Electron., vol. 8, no.
Three contributions have been made in this paper: 1) The pro- 6, pp. 879–886, Jun. 2015.
posed method can simultaneously realize and coordinate three [17] A. Camacho, M. Castilla, J. Miret, J. L. G. de Vicuña, and R. Guzmán,
“Positive and negative sequence control strategies to maximize the voltage
critical control objectives, i.e., voltage support, current limita- support in resistive-inductive grids during grid faults,” IEEE Trans. Power
tion, and oscillation suppression, during all voltage sag scenar- Electron., vol. 33, no. 6, pp. 5362–5373, Jun. 2018.
ios. 2) It only applies positive sequence voltage control, which [18] A. Camacho, M. Castilla, J. Miret, P. Martí, and M. Velasco, “Maxi-
mizing positive sequence voltage support in inductive-resistive grids for
can maximize the voltage support capability while less-affected distributed generation inverters during voltage sags,” in Proc. IEEE 42th
by the inverter capacity; 3) The proposed control strategy makes Annu. Conf. Ind. Electron. Soc., Oct. 2016, pp. 2343–2348.
full use of the inverter capacity. In cases that the inverter has [19] E. Afshari et al., “Control strategy for three-phase grid-connected PV
inverters enabling current limitation under unbalanced faults,” IEEE Trans.
spare capacity after fully realizing such three objectives, the Ind. Electron., vol. 64, no. 11, pp. 8908–8918, Nov. 2017.
proposed method can use such spare capacity to achieve extra [20] H. D. Tafti, A. I. Maswood, G. Konstantinou, J. Pou, and P. Acuna,
control objectives, such as actively improve system stability and “Active/reactive power control of photovoltaic grid-tied inverters with
peak current limitation and zero active power oscillation during unbal-
voltage balance. Various simulation results comparing with the anced voltage sags,” IET Power Electron., vol. 11, no. 6, pp. 1066–1073,
BPSC method are presented, which successfully validated the May 2018.
effectiveness of the new control method presented in this paper. [21] X. Du, Y. Wu, S. Gu, H. Tai, P. Sun, and Y. Ji, “Power oscillation
analysis and control of three-phase grid-connected voltage source con-
verters under unbalanced grid faults,” IET Power Electron., vol. 9, no. 11,
REFERENCES pp. 2162–2173, Sep. 2016.
[22] H. Chen, C. Lee, P. Cheng, R. Teodorescu, and F. Blaabjerg, “A low-
[1] J. Jia, G. Yang, and A. H. Nielsen, “A review on grid-connected converter voltage ride-through technique for grid-connected converters with reduced
control for short-circuit power provision under grid unbalanced faults,” power transistors stress,” IEEE Trans. Power Electron., vol. 31, no. 12,
IEEE Trans. Power Del., vol. 33, no. 2, pp. 649–661, Apr. 2018. pp. 8562–8571, Dec. 2016.
[2] N. R. Merritt, C. Chakraborty, and P. Bajpai, “New voltage control strate- [23] M. A. Garnica López, J. L. Garcia de Vicuña, J. Miret, M. Castilla, and
gies for VSC-based DG units in an unbalanced microgrid,” IEEE Trans. R. Guzmán, “Control strategy for grid-connected three-phase inverters
Sustain. Energy, vol. 8, no. 3, pp. 1127–1139, Jul. 2017. during voltage sags to meet grid codes and to maximize power delivery
[3] A. M. Stanisavljević and V. A. Katić, “Magnitude of voltage sags predict- capability,” IEEE Trans. Power Electron., vol. 33, no. 11, pp. 9360–9374,
ion based on the harmonic footprint for application in DG control system,” Nov. 2018.
IEEE Trans. Ind. Electron., vol. 66, no. 11, pp. 8902–8912, Nov. 2019. [24] N. Hoffmann and F. W. Fuchs, “Minimal invasive equivalent grid
[4] J. Fortmann et al., “Fault-ride-through requirements for wind power plants impedance estimation in inductive-resistive power-networks using ex-
in the entso-e network code on requirements for generators,” IET Renew. tended Kalman-filter,” IEEE Trans. Power Electron., vol. 29, no. 2,
Power Gener., vol. 9, no. 1, pp. 18–24, Jan. 2015. pp. 631–641, Feb. 2014.

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.
2500 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 36, NO. 4, AUGUST 2021

[25] A. Vidal et al., “A method for identification of the equivalent inductance Xiao Chang is currently a Senior Engineer with Elec-
and resistance in the plant model of current-controlled grid-tied con- tric Power Research Institute of State Grid Shanxi
verters,” IEEE Trans. Power Electron., vol. 30, no. 12, pp. 7245–7261, Electric Power Company, Taiyuan, China. His re-
Dec. 2015. search interests include new energy, power quality
[26] Metallized Polypropylene Film Capacitors. Mansfield, TX, USA: Mouser analysis, and energy storage application technology.
Electronics, Jun. 2018.
[27] P. Rodriguez, J. Pou, J. Bergas, J. I. Candela, R. P. Burgos, and D.
Boroyevich, “Decoupled double synchronous reference frame PLL for
power converters control,” IEEE Trans. Power Electron., vol. 22, no. 2,
pp. 584–592, Mar. 2007.
[28] V. L. Srinivas, B. Singh, and S. Mishra, “Fault ride-through strategy for
two-stage grid-connected photovoltaic system enabling load compensation
capabilities,” IEEE Trans. Ind. Electron., vol. 66, no. 11, pp. 8913–8924,
Nov. 2019.
[29] P. Rodriguez, A. V. Timbus, R. Teodorescu, M. Liserre, and F. Blaab-
jerg, “Flexible active power control of distributed power generation sys- Zhe Cao is working toward the M. Eng. degree with
tems during grid faults,” IEEE Trans. Ind. Electron., vol. 54, no. 5, the College of Electrical Engineering, Shanghai Uni-
pp. 2583–2592, Oct. 2007. versity of Electrical Power, Shanghai. His research
interest include microgrid control and protection.

Liang Ji (Member, IEEE) received the B.Eng. and

Ph.D. degrees from the University of Strathclyde,
U.K. He is presently an Associate Professor with the
College of Electrical Engineering, Shanghai Univer-
sity of Electric Power, Shanghai. His research inter-
ests include the modeling and simulation for power
system, power system protection and control.
Yang Mi (Member, IEEE) received the Ph.D. degree
from Northeastern University, Shenyang, China. She
is currently a Professor with Shanghai University
of Electric Power, Shanghai. Her research interests
include power system control and protection.
Jiabin Shi is working toward the M. Eng. degree with
the College of Electrical Engineering, Shanghai Uni-
versity of Electrical Power, Shanghai. His research
interests include power electronics and control of
power converters.

Zhenkun Li received the Ph.D. degree from Hohai

University, Nanjing, China. He is currently a Profes-
Qiteng Hong (Member, IEEE) received the Ph.D. sor with the Shanghai University of Electrical Power,
degree in electronic and electrical engineering from China.
the University of Strathclyde, in 2015. He is currently
a Chancellor’s Fellow with the University of Strath-
clyde, Glasgow, U.K. His main research interest is
on power system protection. He is a Regular Member
of the CIGRE Working Group B5.50, IEEE Working
Group P2004.

Yang Fu (Member, IEEE) received the Ph.D. degree Campbell Booth received the B.Eng. and Ph.D. de-
in electrical engineering from Shanghai University, grees in electrical and electronic engineering from the
Shanghai, China, in 2007. He is currently a Profes- University of Strathclyde, Glasgow, U.K., in 1991 and
sor with the Shanghai University of Electric Power, 1996, respectively. He is currently a Professor and
Shanghai, China. He is the corresponding author of Head of the Department for Electronic and Electrical
this paper. Engineering, University of Strathclyde. His research
interests include power system protection; plant con-
dition monitoring, and applications of intelligent sys-
tem techniques to power system monitoring, protec-
tion, and control.

Authorized licensed use limited to: Makerere University Library. Downloaded on October 25,2024 at 16:15:53 UTC from IEEE Xplore. Restrictions apply.