Alexandria Engineering Journal (2023) 74, 535–545

H O S T E D BY
Alexandria University

Alexandria Engineering Journal

www.elsevier.com/locate/aej
www.sciencedirect.com

ORIGINAL ARTICLE

Robust hybrid nonlinear control approach for

stability enhancement of a constant power load
DC-DC boost converter
Zubaer Alam a, Subarto Kumar Ghosh b, Abdulhameed Fouad Alkhateeb c,
Tushar Kanti Roy a, Md. Shofiqul Islam c, Sajeeb Saha d,*, Mohammad Asif Hussain c

a
Department of Electronics & Telecommunication Engineering, Rajshahi University of Engineering & Technology, Rajshahi,
Bangladesh
b
Department of Electrical & Electronic Engineering, Rajshahi University of Engineering & Technology, Rajshahi, Bangladesh
c
Department of Electrical and Computer Engineering, Faculty of Engineering, King Abdulaziz University, Jeddah 21589,
Saudi Arabia
d
School of Science, Technology and Engineering, University of the Sunshine Coast, Queensland, Australia

Received 23 October 2022; revised 20 March 2023; accepted 12 May 2023

KEYWORDS Abstract This paper presents a robust hybrid super-twisting reaching law-based sliding mode con-
DC distribution network; troller in conjunction with a complementing recursive backstepping controller for DC-DC boost
Robust hybrid controller; converters with constant power loads (CPLs). The key control goal is to keep the DC-bus voltage
Lyapunov stability theory; steady and constant while continuously powering the CPL. The nonlinear dynamical model of the
Constant power load; DC-DC boost converter is translated into Brunovsky’s canonical forms, utilizing the precise feed-
Negative resistance back linearization technique to attain this control goal. It is worth mentioning that the dynamical
characteristics model has been modified with bounded external uncertainties to demonstrate the restraint perfor-
mance for huge loads and input voltage variations. The proposed hybrid controller is then con-
structed using this model. Furthermore, to ensure the large signal stability of the whole system
under the designed control law, the Lyapunov stability theory is used. Numerical simulation anal-
ysis are carried out in MATLAB/Simulink to validate the performance of the designed robust con-
troller for a DC–DC boost converter feeding a CPL when the operating point change. The results
show that the designed controller outperforms both existing nonlinear controllers and a traditional
PI controller in terms of ensuring the system’s stability and fast response in a variety of operating
modes. Finally, in order to confirm the theoretical design and simulation results of the designed
robust hybrid controller, experimental results are also provided in this paper.
Ó 2023 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Alexandria
University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).

* Corresponding author.
E-mail address: [email protected] (S. Saha).
Peer review under responsibility of Faculty of Engineering, Alexandria University.
https://doi.org/10.1016/j.aej.2023.05.041
1110-0168 Ó 2023 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Alexandria University.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
536 Z. Alam et al.

1. Introduction to the CPL is not taken into account in [17], this controller can-
not assure transient stability under significant shocks. To over-
Because of their voltage step-down or step-up capabilities, power come this issue, MPC based fuzzy and hybrid MPC controllers
electronic converters are frequently employed in renewable are presented in [18–20], respectively. However, these MPC tech-
energy source (RES)-based systems (for example, AC microgrids, niques can only deliver the necessary results if a precise dynam-
DC microgrids, and so on) [1,2] and vehicle power grids (for ical model is available, which is virtually unattainable. As a
example, land, space vehicles, and so on) [3,4]. RESs-based dis- result, the tracking error arises when the operational point varies.
tribution power networks, on the other hand, have recently Furthermore, due to the high computing costs of these MPC
received a lot of interest owing to the multiple benefits described techniques, their applicability is limited.
in [5]. Even though power electronics-based DC distribution net- Several nonlinear controllers have been documented in the
works (DDNs) are becoming increasingly popular, the system existing literature to mitigate the aforementioned limitations
still confronts two major obstacles to remain stable and depend- while improving the stabilization of DC-bus voltage under
able. The first issue is that RESs have a low and variable output the highly nonlinear properties of CPLs and the non-
voltage, but the DC-bus requires a constant and high supply minimum phase issue of boost converters. Ref. [21] presents
voltage. To alleviate this issue, a DC-DC boost converter is a passivity-based controller (PBC) to reduce the threat of
widely employed as a bridge between RESs and the DC-bus instability caused by CPLs while enhancing the stable opera-
[6]. The system’s reliability and stability are put to test when tion of the DC-bus voltage. But, it does need precise system
the interface consists of a vast number of DC-DC converters. parameters, which are not resilient in the face of system uncer-
The continual power demand also causes instability, which is a tainty. To overcome the parameter variation problem, an
worry. The system may become unstable or cause a blackout adaptive PBC scheme is presented in [22,23]. On the other
due to the negative-impedance properties of CPLs [6–8]. There- hand, the non-minimum phase problem of boost converters
fore, a robust controller that not only addresses the problem has not been solved, although it can compromise system stabil-
of instability but also offers fast dynamic stability while main- ity in the face of massive shocks. A time-scale model is also
taining system stability is required [9]. necessary for the development of these controllers, which are
Various control solutions have been published in the current challenging to cope with in practical systems.
literature to solve the instability concerns generated by CPLs. A As mentioned in [24], noise/uncertainty decoupling is an
passive damping-based controller is presented in [10,11] to coun- essential aspect of the nonlinear feedback linearization controller
teract the destabilizing impact of CPLs. Even though this (FBLC). Based on this merit, [25] proposes a nonlinear FBLC to
method efficiently mitigates the instability issue of a CPL- stabilize the voltage of the DC-bus by addressing the non-
based DC-DC boost converter, its high cost and huge size limit minimum phase issue associated with boost converters and the
its practical usage. To get around this constraint, active instability issues in the presence of CPLs. However, while the
damping-based methods [12,13] like virtual resistance, virtual destabilizing effect produced by CPLs has been overcome, it is
impedance, and virtual capacitor are commonly employed to unclear how to fix the boost converter’s non-minimum phase
reduce nonlinearity and instability. Through the adjustment of problem. A comparable controller is presented in [26] to alleviate
an extra control loop for decreasing the destructive effect of the non-minimum phase concern. The inductor current is used as
CPLs, the unstable poles have been driven to shift towards the an output function in this study to solve the non-minimum phase
stable area in the active damping technique. Moreover, it is nec- problem. However, the control rule is solely developed for a
essary to change the original converter control loop, which has a resistive load, even though the gradual negative-impedance prop-
considerable influence on the system’s dynamic responsiveness. erties of CPLs cause the system to operate in an unstable man-
To address the aforementioned flaws, [14] proposes a dual-loop ner. Furthermore, the above-mentioned controllers’ satisfactory
controller based on the droop hypothesis. While enough damp- performance is strongly dependent on accurate system parameter
ing force can be delivered into networks to stabilize the system, information, despite the fact that knowing the correct informa-
a little fluctuation in the gain of the controller can lead to erro- tion is nearly unattainable.
neous power-sharing in networks. An adaptive-based voltage– As mentioned in [27], the sliding mode control (SMC)
current (V-I) droop controller is presented in [15] to address this scheme can be a viable solution for handling the parameter
problem. However, as these methods are built based on a model sensitivity problem of FBLCs in both linear and nonlinear sys-
of tiny signals under modest disturbances around a real operat- tem models because it may provide a promising solution to
ing area, the system’s stability can only be maintained near nom- reject disturbances and adjust for uncertainty. To ensure
inal operating points. Hence, when significant disruptions occur, steady-state and transient stability in DC microgrids, [28] pro-
these controllers may lose their effectiveness, consequently poses a robust pulse width modulation (PWM)-based SMC for
endangering the global stability of a DC-system. DC-DC boost converters feeding CPLs. A different SMC
Nonlinear control techniques have been extensively strategy with the same control goal is proposed in [29,30]. Even
researched to address the concerns outlined in the case of linear though SMC can handle parametric and external uncertainty
control systems [16]. Because of the instability and nonlinearity [31], the necessity for a variable switching frequency and high
impacts induced by CPLs, nonlinear controllers must be gain for the SMC scheme in real-world applications may gen-
designed to not only resolve the shortcomings of linear con- erate unwanted chattering in the control signal. Furthermore,
trollers but also to improve the system’s transient and dynamic the non-minimum phase issue of the boost converter has not
performance. To overcome the aforementioned restrictions, been described by the existing SMC control scheme [28–31].
model predictive control (MPC) might be utilized. In [17], an To overcome the shortcomings of nonlinear SMCs, a non-
MPC is presented for improving DC-bus voltage stability in linear backstepping controller (BSC) to control the DC-bus
DC microgrids using CPLs. However, as the uncertainty owing voltage in CPLs-based DC-DC boost converters is presented
Robust hybrid nonlinear control approach for stability enhancement 537

in [32]. During implementation, however, correct system
Using an exact feedback linearization methodology, the
parameters are required since neither model uncertainty nor nonlinear dynamical model is converted to Brunovsky’s
external interruptions are taken into consideration. To circum- canonical form. To avoid associated non-minimum phase
vent this restriction, an adaptive BSC is proposed in [33] for issues with the DC-DC boost converter, instead of the
preserving the stability of the boost converter in conjunction capacitor voltage or inductor current, an energy function
with CPLS under large disturbances. However, the boost con- is employed as the output function.
verter’s non-minimum issue is not addressed in [33], although
If the disturbance is not effectively mitigated, the DC-bus
it can hamper the system’s transient and dynamic stability. voltage, and hence the entire system, may become unstable.
As a result, the proposed controllers are unable to maintain As a result, a stable controller is required to reduce the
system stability when large-signal disturbances occur. To solve instability induced by CPLs while improving transient per-
the aforementioned constraints, [34] designed a hybrid con- formance. To meet this objective, Brunovsky’s canonical
troller by combining the FBLC and adaptive BSC systems. form mathematical model is altered by taking into account
Although the boost converter’s non-minimum phase issue disturbance, which is the incorporation of modeling uncer-
and negative-resistance implications due to CPLs are tainty and incremental negative-impedance induced by
addressed, the internal stability of the controllers for other CPLs. The suggested hybrid controller is designed based
state variables is not considered in these articles. In [35], a sim- on the overall asymptotic stability, which is shown using
ilar strategy is adopted to design another hybrid controller that the Lyapunov stability theory and combines the benefits
combines SMC with adaptive BSC to address the aforemen- of a BSC and an STSMC.
tioned concerns. The chattering effect is decreased by lowering
A super-twisting reaching law is being used to reduce the
the disturbance rejection skills. As a result, for a CPL-based high frequency chattering while receiving fast convergence
DC-DC boost converter, eliminating chattering caused by as compared to conventional reaching laws.
variable switching frequency control action while retaining
The reachability and stability of the proposed BSTSMC are
high precision performance remains an unsolved challenge. theoretically and empirically examined to put it into practi-
To rectify the aforesaid limitations of earlier studies, this cal use.
paper proposes a hybrid nonlinear controller that combines
Numerical simulation analysis and a real-time experimental
the benefits of recursive backstepping with the super-twisting result with severe disturbances are used to validate the effi-
sliding mode control (BSTSMC) algorithm. It is worth men- cacy of the proposed controller.
tioning that BSTSMC is resistant to all parametric fluctuations
as well as external interruptions since the sliding surface and its
derivative converge to zero in a finite amount of time. The 2. Dynamical modeling and problem statement
main purpose is to improve the DC-bus voltage stability when
the input voltage fluctuates and there is a load disturbance. The overall layout of a distribution system is depicted in Fig. 1,
This work’s uniqueness can be summed up in the following where the DC source is often RE-based distributed generating
ways: units (DGs), such as PV panel, rectifier injunction with wind

Fig. 1 General structure of a DC distribution network.

538 Z. Alam et al.

power production systems, and so on. The DC source in a the average dynamical model of a converter with a CPL, as
DGS is commonly interfaced with the main DC-bus using illustrated in Fig. 2, may be stated as follows [31]:
power converters to fulfil the terminal voltage requirement of Vg vdc
the DC load on the DC-bus. Constant voltage loads, or CPLs,
diL
dt
¼ Lb
þ vLdcb l
ð3Þ
are one type of DC load. As a result, the main DC-bus may dvdc
¼ C1dc ðiL  RvdcL  PvCPL Þ  CiLdc l
dt dc
deliver both constant power to the CPL and a constant voltage
to the resistive load to ensure system functions under varying where l is the duty ratio, iL is the inductor current, vdc is the
loads. However, variations in the DC-bus voltage might be main DC-bus voltage, Cdc is the equivalent capacitance of
caused by changes in the DG power supply or loads, as well the DC-bus, and RL is the total resistive load. The system is
as the negative impedance of CPLs. All of these in detail are highly nonlinear due to the nonlinear term PvCPL , as can be
dc
discussed in the following. The following expression is being shown from (3). Furthermore, as CPLs have a negative incre-
used to describe a CPL’s electrical characteristics mental impedance as can be seen from (2), the system with
these loads is prone to instability. However, because of these
PCPL
iCPL ¼ ð1Þ nonlinearities caused by the CPL’s negative impedance charac-
vdc
teristics and the converters’ non-minimum phase issue, con-
ventional controllers (such as SMC and PI) are unable to
where iCPL ; PCPL , and vdc denote the current, power of a CPL
provide adequate damping to assure correct operation.
and DC-bus voltage, respectively. After doing the small-
Designing a robust hybrid nonlinear control strategy with ade-
signal disturbance analysis around the operating point in (1),
quate damping to counteract the destabilizing impact of CPLs
the associated resistance of a CPL may be written as:
is the aim of this paper. The next section goes over each of
@vdc PCPL these points.
RCPL ¼ ¼ 2 ð2Þ
@iCPL ICPL
3. Controller design
where ICPL is the current drawn by the CPL in steady-state
condition. The negative resistance feature of a CPL has a neg- This section is concerned with the design of a nonlinear com-
ative influence on global stability, i.e., it reduces the damping posite controller for a boost converter with a specified CPL.
of the system or even causes system instability, especially when The fundamental control goal is to stabilize the whole system
it is linked to a source converter, as shown in (1). Standard lin- while attaining zero tracking error of the main DC-bus volt-
ear controllers cannot ensure system stability under severe dis- age. However, as can be seen from (3), the mathematical model
turbances since nonlinearity is introduced into the system by is insufficient for designing the suggested composite controller.
CPLs. Hence, to counteract the nonlinear impact of CPLs, a Therefore, it is important to adapt this model to one that is
nonlinear controller is required. However, to obtain such a appropriate for the proposed controller, as detailed in the fol-
goal from a nonlinear controller, the controller must be lowing subsections.
designed using a suitable dynamical model. Hence, to fulfil
the controller design requirement, a dynamical nonlinear 3.1. Coordination transformation
model for DC-DC boost converters caused by CPL loads is
constructed in this section. The simplified configuration of a
The coordinate transformation of the original nonlinear sys-
DC-DC boost converter feeding CPLs is described in Fig. 2.
tem as described in (3) using a exact feedback linearization
The DC voltage Vg will be fed to the main DC-bus through
approach can be represented as follows [24]:
a DC-DC boost converter, as shown in Fig. 2, and the CPL
and resistive load will be linked in parallel to the main DC- z_1 ¼ z2
bus. However, in this research, the DC load is treated as a pure ð4Þ
z_2 ¼ m ¼ aðxÞ þ bðxÞl
CPL, which has the greatest influence on stability when com-
pared to a resistive load. The boost converter is assumed to where z1 ; z2 are the transformed linearized system’s state vari-
be in continuous conduction mode for simplicity. As a result, ables, t is the converted linear system’s control variable,

Fig. 2 Simplified structure of a DC distribution network.

Robust hybrid nonlinear control approach for stability enhancement 539

Vg ðVg x2 Þ
aðxÞ ¼ L2f hðxÞ ¼ Lb
 R2x 2
L Cdc
ðx1  Rx2L  PxCPL
2
Þ, and bðxÞ ¼ Step 2: The second error dynamic, as seen in (8) could be
Vg x 2 expressed as follows:
Lg Lf hðxÞ ¼ Lb
þ RL Cdc . By including the influence of
2x1 x2
uncer-
tainty terms as external disturbances, (4) can be modified as e_2 ¼ z_2  w_ ð15Þ
follows:
Eq. (15), using (5) and (14), can be written as:
z_1 ¼ z2 þ ED1
ð5Þ e_2 ¼ aðxÞ þ bðxÞl þ ED2  s þ r ð16Þ
z_2 ¼ m ¼ aðxÞ þ bðxÞl þ ED2
Now, to build the suggested hybrid controller, a sliding surface
where EDi with i = 1,2 is expressed as external disturbances. in terms of tracking error e2 can be chosen as follows:
In the next section, the prospective composite control scheme
will be developed using the model indicated by (5). S ¼ e2 ð17Þ
The time derivative of S may be stated as using the value of e_2
3.2. Design methodology of the proposed hybrid controller
S_ ¼ aðxÞ þ bðxÞl þ ED2  s þ r ð18Þ
In this subsection, the control input will be derived utilizing the The reaching rule is well recognized for its importance in
suggested hybrid control technique. The main goal is to keep decreasing chattering while improving convergence time. To
track of the desired values of z1 and z2 , which guarantees that attain this control purpose, this work presents a super-
the DC-bus voltage is accurately tracked under the impact of twisting-based reaching law, which may be stated as follows:
EDi . In the following, the design procedure of the proposed Z
hybrid controller is presented in detail. S_ ¼ g1 jSjb sgnðSÞ  g2 sgnðSÞdt ð19Þ
Step 1: The tracking error between z1 and its reference value
z1ðrefÞ is calculated as follows: where g1 and g2 are positive constants and 0 < b < 1. From
(19), it can be seen that the super-twisting-based reaching
e1 ¼ z1  z1ðrefÞ ð6Þ R
law is continuous since both g1 jSjb sgnðSÞ and g2 sgnðSÞdt
where z1ðrefÞ ¼ 12 Lb i2LðrefÞ þ 12 Cdc v2dcðrefÞ with iLðrefÞ ¼ PVCPL
g
is the terms are continuous. Moreover, it can be seen that the high
inductor current’s reference value for the pure CPL and frequency switching term sgnðSÞ conceals under the integral,
vdcðrefÞ is the DC-bus voltage’s reference value. Now, e1 ’s and consequently, it reduces the chattering. The following
dynamic using (5) can be expressed as: equation can be formed by equating (18) and (19):
Z
e_1 ¼ z2 þ ED1  z_1ðrefÞ ð7Þ aðxÞ þ bðxÞl þ ED2  s þ r ¼ g1 jSjb sgnðSÞ  g2 sgnðSÞdt
In (7), z2 will serve as an intermediate variable according to the ð20Þ
backstepping control theory. The ultimate error may be stated
as if is used as the stabilizing function for z2 At this point, the real control law from (20) can be adopted as
follows:
e2 ¼ z2  w ð8Þ Z
1 1
By putting (7) and (8) together, one get l¼ ½aðxÞ  s þ e1 e2 þ g1 jSjb sgnðSÞ þ g2 sgnðSÞdt
bðxÞ S
e_1 ¼ e2 þ w þ ED1  z_1ðrefÞ ð9Þ ð21Þ
At this point, to examine the stability of e_1 , a Lyapunov func- Inserting (21) into (18), it yields
tion (LF) can be expressed as follows: Z
1
1 S_ ¼ g1 jSjb sgnðSÞ  g2 sgnðSÞdt  e1 e2 þ ED2 þ r ð22Þ
W1 ¼ e21 ð10Þ S
2
The LF can now be selected as follows to ensure overall system
Its dynamic use of (9) may be expressed as: stability:
W_ 1 ¼ e1 ½e2 þ w  z_1ðrefÞ  þ e1 ED1 ð11Þ 1
W2 ¼ W1 þ S2 ð23Þ
The stabilizing function w, from (11), can be selected as: 2
The time derivative of (23) may be expressed as using (13) and
w ¼ k1 e1 þ z_1ðrefÞ ð12Þ
(22):
where k1 is a positive gain. Putting the value of w from (12) Z
into (11) yields W_ 2 ¼ k1 e2  g1 jSjb sgnðSÞ  g2 sgnðSÞdt þ ED2 þ D1
1

W_ 1 ¼ k1 e21 þ e1 ðe2 þ ED1 Þ ð13Þ ð24Þ

The derivative of W1 will be negative semi-definite if e1 ¼ 0, as where D1 ¼ r þ e1 ED1 . To make W_ 2  0, the following condi-
shown in (13). However, as the actual control law has not yet tions should be hold
been specified, the second stage is necessary to assure the sys-
tem’s stability. However, the dynamic of w is required to finish jD1 j < g1 andjED2 j < g2 ð25Þ
the following phase, which may be stated as follows: At this stage, it is essential to analyze the reachability
of the proposed controller. To do that the dynamic of the
w_ ¼ s  r ð14Þ
compensating reaching law under the conditions of (25), can
where s ¼ k1 ½z2  z_1ðrefÞ  þ z€1ðrefÞ and r ¼ k1 ED1 be expressed as follows:
540 Z. Alam et al.

dS The designed controller’s effectiveness is validated in this

¼ g1 jSjb sgnðSÞ ð26Þ
dt section under a wide range of operating situations, taking into
Eq. (26), can be written as follows: consideration variations in supply voltage in terms of input
power and CPL reference power. External disturbances are
dS also incorporated into all the proposed controller’s feedback
g1 dt ¼  ð27Þ
jSjb sgnðSÞ variables. During the simulation investigation, white Gaussian
Integrating both sides of (27), will yield noise is employed as an external disturbance. The designed
controller’s performance is now illustrated using the three case
1 studies listed below:
t¼ ½Sð0Þ1b  jSj1b  ð28Þ
g1 ð1  bÞ Case I: Variations in supply voltage are considered to test
where Sð0Þ is the initial value of the sliding surface S when the the controller’s performance,
reaching phase is started. If tr is considered the reaching time Case II: Variations in DC-bus reference voltage are consid-
then the sliding surface value at tr will be zero. Thus, the reach- ered to test the controller’s performance, and.
ing time tr from (28), can be expressed as follows: Case III: Variations in the power of lumped CPL are con-
sidered to investigate the controller’s performance.
Sð0Þ1b The effectiveness of the suggested nonlinear controller is
tr ¼ ð29Þ
g1 ð1  bÞ also compared to that of existing SMC (SMC) and adaptive
backstepping SMC (BSMC) as explored in [28,35], respectively
Thus, it is confirmed that the tracking error will converge to
and a traditional PI control scheme to illustrate its superiority.
zero in a finite time by using the proposed controller. It is clear
Case I: The evaluation of the controller’s effectiveness
from Eq. (29) that the reaching time is shorter than the reach-
under variation of the supply voltage.
ing time predicted by the constant reaching law. If the value of
Because of its intermittent features, the output power of
g1 is larger, the reaching time will be faster and the chattering
RESs is constantly changing in practice. The unpredictability
degree will be increased, and vice versa. Therefore, the pro-
of the input power is used in this scenario to highlight the
posed control scheme can decrease chattering while improving
impact on the distribution network. At t = 1 s and t
convergence time. In the next part, the usefulness of the devel-
= 2.5 s, the DC-voltage input voltage climbs from 60 V to
oped controller is proven.
90 V and then lowers from 90 V to 40 V for this purpose.
The system’s dynamic response and stability are depicted in
4. Controller performance evaluation Fig. 3. When input voltage varies, the proposed and existing
controllers can correctly follow the voltage of the main DC-
A comparable simulation model, as illustrated in Fig. 2, was bus and input current of the inductor, as illustrated in Fig. 3
developed using the MATLAB/Simulink platform to validate (b) and Fig. 3 (c). In comparison to the SMC, BSMC and stan-
the efficiency of the designed composite control technique. dard PI control techniques, the developed controller has a
The distribution network’s main-DC bus is assumed to have quicker dynamic reaction speed, less tracking error, and
a nominal voltage of 220 V, and DC loads have a rated power reduced overshoot/undershoot. Furthermore, the CPL stays
of 1.5 kW. To capture the greatest destabilizing influence on unaltered in the equilibrium condition of 1.5 kW, as seen in
the DC distribution network’s DC-bus voltage, the pure Fig. 3 (d).
lumped CPL is studied without taking into account the con- Case II: Controller effectiveness evaluation at changes in
stant voltage and current load. During the deployment of the the reference DC-bus voltage.
designed hybrid controller, the sampling frequency is set to In this study, with the CPL set to 1.5 kW, the reference
100 kHz whereas the boost converter’s frequency is considered voltage of the DC-bus is decreased from 220 V to 120 V at
as 10 kHz. It is common knowledge that knowing the system t = 1.4 s and then it is increased from 120 V to 180 V and
parameters is critical while implementing controllers. There- 180 V to 220 V at t = 2.8 s and t = 3.8 s, respectively. In
fore, Table 1 shows the nominal parameters of the proposed Fig. 4, one can see the relevant system reactions. The PI con-
distribution network that are employed in the numerical inves- trol scheme fails to track the reference voltage of the common
tigation. The control parameters of the existing nonlinear bus at both conditions whereas SMC and BSMC cannot main-
SMC, adaptive backstepping SMC (BSMC), and PI controller tain the faster tracking of the DC-bus voltage, however, the
are selected based on the guidance documented in [28,35,36], designed controller can swiftly track the DC-bus voltage
respectively. because the overshoot and settling time are near to zero, as
shown in Fig. 4 (a). Similarly, as mentioned in Fig. 4 (b), the
developed BSTSMC controller outperforms the BSMC and
PI in terms of dynamic performance whereas the existing
Table 1 Simulated parameters of the proposed DC distribu- SMC provides less overshoot. Overall, the proposed controller
tion networks. provides better dynamic performance as compared to the exist-
Parameters of DC system Value ing control approach. Because of the electrical properties of
the CPL (shown in Fig. 4 (c)), the load power remains constant
Supply voltage, Vg 60 V
throughout the simulation investigation.
Main DC-bus voltage, vdc 220 V
Constant power load, PCPL 1:5 kW Case III: Controller effectiveness evaluation under
Switching frequency of the converter, (fsw ) 10 kHz changing in the power of the lumped CPL.
Inductance of the boost converter, Lb 5 mH The change in CPL reference power is taken into account in
Main DC-bus capacitance, Cdc 6000 lF this scenario to demonstrate the effectiveness of the existing
and proposed controllers. The CPL’s rated power is 1.5 kW
Robust hybrid nonlinear control approach for stability enhancement 541

Fig. 3 Demonstration of the dynamic stability of the (b) DC-bus voltage, and (c) input current of the inductor under wide variations in
(a) input voltage.

150
1000 BSTSMC
BSTSMC
PI
PI 60
800 220 40 SMC
SMC 40
200 100 BSMC
180 220 BSMC 20 20
600
IL (A)

160
Vbus (V)

Vref
140 200 0 0
120
1.4 1.5 1.6 1.7 3.8 3.85 3.9
400 1.4 1.5 1.6 1.7 180 50
3.8 4 4.2

200
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0 1 2 3 4 5 Time(s)
Time (s)
(a) (b)

Fig. 4 Demonstration of the dynamic stability of the (a) DC-bus voltage (b) input current of the inductor (c) power in CPL under wide
operating point variations in reference DC-bus voltage.

at the start of the simulation, but it rises to 4 kW at t = 1.2 s. result of this large operating range variation, the traditional PI
Other parameters, at the same time, remain the same. Owing to control scheme cannot deliver sufficient damping torque into
the changes in the power of the lumped CPL, data as recorded networks, consequently, both the main bus voltage and the
in Fig. 5 (a), the dynamic and transient stability of the voltage input current are unstable, as shown in Figs. 5 (b) and 5 (c),
at the main DC-bus and input current are changed as indicated respectively. Meanwhile, as shown in Fig. 5 (b) and Fig. 5
in Fig. 5 (b) and Fig. 5 (c), respectively. It is clear from these (c), the developed BSTSMC controller and the existing
responses that proposed BSTSMC and existing nonlinear con- nonlinear controllers can assure stability. Although all
trollers are capable of maintaining stability whereas the stan- nonlinear controllers can keep an output voltage tracking error
dard fine-tuning PI controller fails to maintain the global of zero, the designed controller can settle the response faster
stability of the distribution network. In comparison to the with less overshoot/undershoot.
SMC, BSMC and PI, the designed controller can provide fas- Based on the above simulation findings, it can be inferred
ter dynamic reactions with reduced undershoot/overshoot. that the developed controller can offer a DC-DC boost
Again, the CPL power is reduced from 4 kW to 0.5 kW at t converter supplying power to a CPL with quick dynamic
= 2.6 s, which is seen as a significant system disruption. As a convergence speed, precise DC-bus voltage tracking, and high
542 Z. Alam et al.

Fig. 5 Demonstration of the dynamic stability of the (b) main DC-bus voltage and (c) input inductor current under wide variations in (a)
power of the CPL.

large disturbance stability. Furthermore, even after substantial sources are connected in series to form a 12 V–115 V variable
changes in CPL power, the DC-bus voltage maintains its sta- DC input voltage for the boost converter. The control algo-
bility. As a consequence, the simulation outcome agrees with rithm, along with the voltage and current approximation non-
the theoretical analysis. linear functions, is deployed to the Delfino F28335 controller
using MATLAB/Simulink’s ”deploy to hardware” feature.
4.1. Experimental validation The F28335’s analog-to-digital converter can only measure
up to 3 V, so a proper voltage designer circuit is used to mea-
The effectiveness of the designed controller is verified using a sure the input (Vin ) and output voltage (Vout ) for feedback,
prototype boost converter, as shown in Fig. 6. The experimen- along with a nonlinear predictor function to accurately
tal setup, presented in Fig. 6, includes a DC power supply, approximate the input and output voltage from the voltage
boost converter, DC-bus capacitor, and CPL loads. The pro- divider network output value. The inductor current (IL ) is
posed control structure for the CPL-based DC system is shown measured using the hall-effect current sensor ACS712, but to
in Fig. 2 and implemented on the Delfino-F28335 controller, accurately measure the current, the sensor’s output signal is
which generates the control signal for the prototype boost con- amplified using a negative feedback op-amp amplifier with a
verter feeding the CPL. In the experimental setup, three DC nonlinear predictor function, allowing measurement in the

Fig. 6 Experimental setup for the DC-DC boost converter powered CPL.
Robust hybrid nonlinear control approach for stability enhancement 543

lA range. All three state variables are fed back to the Delfino responses are also recorded in Fig. 7. As illustrated in Fig. 7,
F28335 controller after measurement. A 12 V DC-DC isola- the designed nonlinear controller can provide the accurate
tion converter is used for IGBT gate driving purposes as a con- and fast-tracking performance of the output voltage and cur-
trol voltage source between the gate and source terminals of rent while ensuring the system’s stability.
the IGBT, and opto-couplers are used to turn the IGBTs on Case II: Controller performance evaluation under the DC-
and off. To create a 0.94 mH inductor, four series-connected bus voltage variation.
0.47 mH inductors are used as a 1.88 mH inductor. A The variation of the DC-bus voltage has a destabilizing
programmable-speed DC motor is used as the CPL, and the impact on the DC-DC boost converter-powered CPL. Hence,
speed of the motor is kept constant using a voltage sensor to examine the destabilization impact on the system, at t = 0.4
and ATmega8A-16PU microcontroller through PWM, provid- s, the DC bus voltage is decreased by 33.33% of its nominal
ing a constant power load to the boost converter. During value (48 V) to 32 V while the input voltage and CPL are kept
implementation of the designed robust controller, the technical at their rated values i.e., 24 V and 40 W, respectively. The sud-
specifications of the prototype boost converter are listed in den drop of the DC bus voltage increases the load current from
Table 2. Three different case studies are performed on an 0.872 A to 1.180 A since the load draws constant power from
experimental platform to verify the global stability of the pro- the DC supply, which can be examined in Fig. 8. From Fig. 8,
posed control algorithm under pure CPL loads, whose it is seen that the designed control scheme provides faster
dynamic performances are demonstrated below. tracking (less than 40 ms) and less overshoot of the load volt-
Case I: Controller performance evaluation under the varia- age and current when the DC–bus voltage is abruptly declined.
tion of the DC supply. Now, the DC-bus voltage is increased from 32 V to 40 V and
In this case, the dynamic performance of the proposed com- 48 V at t = 1 s and 1.4 s, respectively, in which the load cur-
posite controller is tested when the variation of the DC supply rent is decreased and recorded as 1.07 A and 0.872 A, respec-
voltage is maintained at 50% of its nominal value while con- tively, as shown in Fig. 8. From Fig. 8, it is observed that
sidering the pure CPL is rated at 40 W and other parameters during these transients, the designed composite controller
are kept constant as mentioned in Table 2. At t = 0.4 s, a can accurately provide faster tracking and less overshoot while
large variation in the input voltage is made in which it is handling the overall stability of the whole system.
reduced from 24 V to 12 V. Under this large disturbance in Case III: Controller performance evaluation under the CPL
the input supply voltage, the dynamic response of the DC- power variation.
bus voltage, output current, and CPL power are illustrated From 1–3, it is seen that the variation in CPL power has the
in Fig. 7. From Fig. 7, it can be observed that the output volt- highest destabilizing impact on the DC-bus voltage and on an
age and current can be accurately tracked to their desired val- overall stability of the power converter. Therefore, to verify
ues while maintaining a less overshoot and faster dynamic the effectiveness of the proposed controller, the change in
stability with the designed robust hybrid control scheme. CPL power has been considered in two scenarios whereas
Again, another variation in the input voltage is made at t the input voltage and output reference voltage are kept the
= 1.44 s in which the DC supply voltage is increased from same as 24 V and 48 V, respectively as shown in Fig. 9. In case
12 V to its rated value of 24 V and corresponding dynamic of first scenario that is at t = 0.4 s, the CPL power is increased
from 25 W to 50 W in which the load current is measured as
1.06 A after increasing this load power as shown in Fig. 9.
In this case, the proposed robust nonlinear control approach
Table 2 System parameters for experimental setup. can maintain the overall system stability with faster dynamic
Parameters of DC system Value response. At t = 1.44 s, another scenario has been produced
DC supply voltage, Vg 24=12=24 V
by decreasing the CPL power from 50 W to 25 W. From
Main DC-bus voltage, vdc 48=32=40=48 V Fig. 9, it is seen that the proposed controller has the ability
Constant power load, PCPL 40 W to maintain the global DC-bus voltage stability by enhancing
Switching frequency of the converter, (fsw ) 50 kHz a faster dynamic and transient response under the CPL load
Inductance of the boost converter, Lb 1:88 mH power variation.
Main DC-bus capacitance, Cdc 1000 lF Based on the experimental results, it can be concluded that
the proposed composite controller ensures faster transient
response, accurate tracking of the system responses, and over-

Fig. 7 Experimental results when a sudden variation happens to the input voltage.
544 Z. Alam et al.

Fig. 8 Experimental results when a sudden variation happens to the reference voltage.

Fig. 9 Experimental results when a sudden variation happens to the CPL power.

all system stability under different operating conditions.
Under any operating condition, the steady-state error is
Hence, experimental results are also maintained in consistency considerably lower as compared to existing controllers.
with the theoretical theorem and simulation results. However,
it is seen that the experimental results are a bit slower than the A robust adaptive hybrid controller will be designed in the
simulation results. This is because the practical results suffer future, utilizing the same reaching rule to handle both interior
from real-life system conditions such as noise, electromagnetic and external uncertainties.
interference (EMI), equivalent series resistance (ESR), and
delays, which are not considered during the simulation. Declaration of Competing Interest

5. Conclusion The authors declare that they have no known competing

financial interests or personal relationships that could have
This paper looked at the negative incremental impedance- appeared to influence the work reported in this paper.
induced instability of a boost converter with CPLs. To allevi-
ate this instability problem, a hybrid nonlinear controller Acknowledgement
incorporating a recursive backstepping control theory and a
super-twisting reaching law-based sliding mode control theory This project was funded by the Deanship of Scientific Research
is described in this study. To develop the presented controller, (DSR), King Abdulaziz University, Jeddah, under Grant No.
the nonlinear system is first transformed into linear canonical (DF-487–135-1441). The authors, therefore, gratefully
form using the total stored energy as the output function using acknowledge DSR technical and financial support.
the exact feedback linearization methodology. Under various
conditions, the results of simulations and experiments are uti-
lized to validate the efficacy and theoretical analysis of the References
designed controller. When compared to three existing con-
trollers, the newly designed controller outperforms them in [1] B. Liu, F. Zhuo, Y. Zhu, H. Yi, System operation and energy
management of a renewable energy-based dc micro-grid for high
every scenario. The following are the key findings based on
penetration depth application, IEEE Trans. Smart Grid 6 (3)
the analysis presented in this paper:
(2014) 1147–1155.
[2] T.K. Roy, M.A. Mahmud, Active power control of three-phase

Both simulation and experimental results show that the grid-connected solar PV systems using a robust nonlinear
designed controller can stabilize the boost converter with adaptive backstepping approach, Sol. Energy 153 (2017) 64–76.
faster transient and less overshoot with substantial fluctua- [3] A.M. Rahimi, A. Emadi, An analytical investigation of DC/DC
tions in input voltage, output voltage and CPL power. This power electronic converters with constant power loads in
performance, however, is not achievable with the existing vehicular power systems, IEEE Trans. Vehic. Technol. 58 (6)
nonlinear method or typical PI controllers. (2008) 2689–2702.
Robust hybrid nonlinear control approach for stability enhancement 545

[4] Z. Jin, G. Sulligoi, R. Cuzner, L. Meng, J.C. Vasquez, J.M. [21] Y.-C. Jeung, D.-C. Lee, T. Dragičević, F. Blaabjerg, Design of
Guerrero, Next-generation shipboard DC power system: passivity-based damping controller for suppressing power
Introduction smart grid and DC microgrid technologies into oscillations in DC microgrids, IEEE Trans. Power Electron. 36
maritime electrical netowrks, IEEE Electrif. Magaz. 4 (2) (2016) (4) (2020) 4016–4028.
45–57. [22] J. Liu, Z. Liu, W. Chen, H. Su, Passivity-based control for
[5] T. Adefarati, R. Bansal, Integration of renewable distributed interleaved double dual boost converters in DC microgrids
generators into the distribution system: a review, IET Renew. supplying constant power loads, IEEE Trans. Transport.
Power Gener. 10 (7) (2016) 873–884. Electrif. (2021).
[6] H. Wu, V. Pickert, M. Ma, B. Ji, C. Zhang, Stability study and [23] M.A. Hassan, E.-P. Li, X. Li, T. Li, C. Duan, S. Chi, Adaptive
nonlinear analysis of DC–DC power converters with constant passivity-based control of DC–DC buck power converter with
power loads at the fast timescale, IEEE J. Emerg. Select. Top. constant power load in DC microgrid systems, IEEE J. Emerg.
Power Electron. 8 (4) (2020) 3225–3236. Select. Top. Power Electron. 7 (3) (2018) 2029–2040.
[7] S. Singh, V. Kumar, D. Fulwani, Mitigation of destabilising [24] T.K. Roy, M.A. Mahmud, W. Shen, A.M. Oo, An adaptive
effect of CPLs in island DC micro-grid using non-linear control, partial feedback linearizing control scheme: An application to a
IET Power Electron. 10 (3) (2017) 387–397. single machine infinite bus system, IEEE Trans. Circuits Syst. II
[8] W. Du, J. Zhang, Y. Zhang, Z. Qian, Stability criterion for Express Briefs 67 (11) (2019) 2557–2561.
cascaded system with constant power load, IEEE Trans. Power [25] G. Sulligoi, D. Bosich, G. Giadrossi, L. Zhu, M. Cupelli, A.
Electron. 28 (4) (2012) 1843–1851. Monti, Multiconverter medium voltage DC power systems on
[9] M. Gheisarnejad, H. Farsizadeh, M.-R. Tavana, M.H. ships: Constant-power loads instability solution using
Khooban, A novel deep learning controller for DC–DC buck– linearization via state feedback control, IEEE Trans. Smart
boost converters in wireless power transfer feeding CPLs, IEEE Grid 5 (5) (2014) 2543–2552.
Trans. Industr. Electron. 68 (7) (2020) 6379–6384. [26] S. Arora, P. Balsara, D. Bhatia, Input–output linearization of a
[10] A.B. Jusoh, The instability effect of constant power loads, in: boost converter with mixed load (constant voltage load and
PECon 2004. Proceedings. National Power and Energy constant power load), IEEE Trans. Power Electron. 34 (1)
Conference, 2004, IEEE, 2004, pp. 175–179. (2018) 815–825.
[11] M. Cespedes, L. Xing, J. Sun, Constant-power load system [27] T.K. Roy, M.A. Mahmud, Fault current compensations in
stabilization by passive damping, IEEE Trans. Power Electron. resonant grounded distribution systems to mitigate powerline
26 (7) (2011) 1832–1836. bushfires using a nonsingular terminal sliding model controller,
[12] A. Hosseinipour, H. Hojabri, Small-signal stability analysis and IET Generation, Transmission & Distribution (2021).
active damping control of DC microgrids integrated with [28] S. Singh, D. Fulwani, V. Kumar, Robust sliding-mode control
distributed electric springs, IEEE Trans. Smart Grid 11 (5) of DC/DC boost converter feeding a constant power load, IET
(2020) 3737–3747. Power Electronics 8 (7) (2015) 1230–1237.
[13] A.M. Rahimi, A. Emadi, Active damping in DC/DC power [29] W. Jiang, X. Zhang, P. Lin, X. Zhang, H.H.C. Iu, T. Fernando,
electronic converters: A novel method to overcome the problems Combined sliding-mode control for the IFDBC interfaced DC
of constant power loads, IEEE Trans. Industr. Electron. 56 (5) microgrids with power electronic loads, IEEE J. Emerg. Select.
(2009) 1428–1439. Top. Power Electron. 8 (4) (2020) 3396–3410.
[14] N. Ghanbari, S. Bhattacharya, Constant power load challenges [30] C. Zheng, T. Dragičević, J. Zhang, R. Chen, F. Blaabjerg,
in droop controlled DC microgrids, IECON 2019–45th Annual Composite robust quasi-sliding mode control of DC–DC buck
Conference of the IEEE Industrial Electronics Society, Vol. 1, converter with constant power loads, IEEE J. Emerg. Select.
IEEE, 2019, pp. 3871–3876. Top. Power Electron. 9 (2) (2020) 1455–1464.
[15] S. Augustine, M.K. Mishra, N. Lakshminarasamma, Adaptive [31] B.A. Martinez-Treviño, A. El Aroudi, E. Vidal-Idiarte, A. Cid-
droop control strategy for load sharing and circulating current Pastor, L. Martinez-Salamero, Sliding-mode control of a boost
minimization in low-voltage standalone DC microgrid, IEEE converter under constant power loading conditions, IET Power
Trans. Sustain. Energy 6 (1) (2014) 132–141. Electronics 12 (3) (2019) 521–529.
[16] S.K. Ghosh, T.K. Roy, M. Pramanik, A. Hanif, M. Mahmud, [32] Q. Xu, W. Jiang, F. Blaabjerg, C. Zhang, X. Zhang, T.
et al, Design of nonlinear backstepping double-integral sliding Fernando, Backstepping control for large signal stability of
mode controllers to stabilize the dc-bus voltage for DC–DC high boost ratio interleaved converter interfaced DC microgrids
converters feeding CPLs, Energies 14 (20) (2021) 6753. with constant power loads, IEEE Trans. Power Electron. 35 (5)
[17] Z. Karami, Q. Shafiee, Y. Khayat, M. Yaribeygi, T. Dragicevic, (2019) 5397–5407.
H. Bevrani, Decentralized model predictive control of DC [33] Q. Xu, C. Zhang, C. Wen, P. Wang, A novel composite
microgrids with constant power load, IEEE J. Emerg. Select. nonlinear controller for stabilization of constant power load in
Top. Power Electron. (2019). DC microgrid, IEEE Trans. Smart Grid 10 (1) (2017) 752–761.
[18] N. Vafamand, S. Yousefizadeh, M.H. Khooban, J.D. Bendtsen, [34] X. Li, X. Zhang, W. Jiang, J. Wang, P. Wang, X. Wu, A novel
T. Dragičević, Adaptive TS fuzzy-based mpc for DC microgrids assorted nonlinear stabilizer for DC–DC multilevel boost
with dynamic CPLs: Nonlinear power observer approach, IEEE converter with constant power load in DC microgrid, IEEE
Syst. J. 13 (3) (2018) 3203–3210. Trans. Power Electron. 35 (10) (2020) 11181–11192.
[19] M. Gheisarnejad, A. Mohammadzadeh, M.-H. Khooban, [35] J. Wu, Y. Lu, Adaptive backstepping sliding mode control for
Model Predictive Control Based Type-3 Fuzzy Estimator for boost converter with constant power load, IEEE Access 7 (2019)
Voltage Stabilization of DC Power Converters, IEEE Trans. 50797–50807.
Industr. Electron. 69 (12) (2021) 13849–13858. [36] S. Bacha, I. Munteanu, A.I. Bratcu, et al, Power electronic
[20] Z. Karami, Q. Shafiee, S. Sahoo, M. Yaribeygi, H. Bevrani, T. converters modeling and control, Advanced textbooks in control
Dragicevic, Hybrid model predictive control of DC–DC boost and signal processing 454 (454) (2014).
converters with constant power load, IEEE Trans. Energy
Convers. 36 (2) (2020) 1347–1356.