Lecture Notes in Electrical Engineering 604

Axaykumar Mehta
Abhishek Rawat
Priyesh Chauhan Editors

Advances
in Control
Systems and its
Infrastructure
Proceedings of ICPCCI 2019
Lecture Notes in Electrical Engineering

Volume 604

Series Editors

Leopoldo Angrisani, Department of Electrical and Information Technologies Engineering, University of Napoli
Federico II, Naples, Italy
Marco Arteaga, Departament de Control y Robótica, Universidad Nacional Autónoma de México, Coyoacán,
Mexico
Bijaya Ketan Panigrahi, Electrical Engineering, Indian Institute of Technology Delhi, New Delhi, Delhi, India
Samarjit Chakraborty, Fakultät für Elektrotechnik und Informationstechnik, TU München, Munich, Germany
Jiming Chen, Zhejiang University, Hangzhou, Zhejiang, China
Shanben Chen, Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai, China
Tan Kay Chen, Department of Electrical and Computer Engineering, National University of Singapore,
Singapore, Singapore
Rüdiger Dillmann, Humanoids and Intelligent Systems Lab, Karlsruhe Institute for Technology, Karlsruhe,
Baden-Württemberg, Germany
Haibin Duan, Beijing University of Aeronautics and Astronautics, Beijing, China
Gianluigi Ferrari, Università di Parma, Parma, Italy
Manuel Ferre, Centre for Automation and Robotics CAR (UPM-CSIC), Universidad Politécnica de Madrid,
Madrid, Spain
Sandra Hirche, Department of Electrical Engineering and Information Science, Technische Universität
München, Munich, Germany
Faryar Jabbari, Department of Mechanical and Aerospace Engineering, University of California, Irvine, CA,
USA
Limin Jia, State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China
Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland
Alaa Khamis, German University in Egypt El Tagamoa El Khames, New Cairo City, Egypt
Torsten Kroeger, Stanford University, Stanford, CA, USA
Qilian Liang, Department of Electrical Engineering, University of Texas at Arlington, Arlington, TX, USA
Ferran Martin, Departament d’Enginyeria Electrònica, Universitat Autònoma de Barcelona, Bellaterra,
Barcelona, Spain
Tan Cher Ming, College of Engineering, Nanyang Technological University, Singapore, Singapore
Wolfgang Minker, Institute of Information Technology, University of Ulm, Ulm, Germany
Pradeep Misra, Department of Electrical Engineering, Wright State University, Dayton, OH, USA
Sebastian Möller, Quality and Usability Lab, TU Berlin, Berlin, Germany
Subhas Mukhopadhyay, School of Engineering & Advanced Technology, Massey University, Palmerston
North, Manawatu-Wanganui, New Zealand
Cun-Zheng Ning, Electrical Engineering, Arizona State University, Tempe, AZ, USA
Toyoaki Nishida, Graduate School of Informatics, Kyoto University, Kyoto, Japan
Federica Pascucci, Dipartimento di Ingegneria, Università degli Studi “Roma Tre”, Rome, Italy
Yong Qin, State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China
Gan Woon Seng, School of Electrical & Electronic Engineering, Nanyang Technological University,
Singapore, Singapore
Joachim Speidel, Institute of Telecommunications, Universität Stuttgart, Stuttgart, Baden-Württemberg,
Germany
Germano Veiga, Campus da FEUP, INESC Porto, Porto, Portugal
Haitao Wu, Academy of Opto-electronics, Chinese Academy of Sciences, Beijing, China
Junjie James Zhang, Charlotte, NC, USA
The book series Lecture Notes in Electrical Engineering (LNEE) publishes the latest developments in
Electrical Engineering - quickly, informally and in high quality. While original research reported in
proceedings and monographs has traditionally formed the core of LNEE, we also encourage authors to
submit books devoted to supporting student education and professional training in the various fields and
applications areas of electrical engineering. The series cover classical and emerging topics concerning:

• Communication Engineering, Information Theory and Networks

• Electronics Engineering and Microelectronics
• Signal, Image and Speech Processing
• Wireless and Mobile Communication
• Circuits and Systems
• Energy Systems, Power Electronics and Electrical Machines
• Electro-optical Engineering
• Instrumentation Engineering
• Avionics Engineering
• Control Systems
• Internet-of-Things and Cybersecurity
• Biomedical Devices, MEMS and NEMS
For general information about this book series, comments or suggestions, please contact leontina.
[email protected].
To submit a proposal or request further information, please contact the Publishing Editor in your
country:

China
Jasmine Dou, Associate Editor ([email protected])

India
Aninda Bose, Senior Editor ([email protected])

Japan
Takeyuki Yonezawa, Editorial Director ([email protected])

South Korea
Smith (Ahram) Chae, Editor ([email protected])

Southeast Asia
Ramesh Nath Premnath, Editor ([email protected])

USA, Canada:
Michael Luby, Senior Editor ([email protected])

All other Countries:

Leontina Di Cecco, Senior Editor ([email protected])

** Indexing: The books of this series are submitted to ISI Proceedings, EI-Compendex, SCOPUS,
MetaPress, Web of Science and Springerlink **

More information about this series at http://www.springer.com/series/7818

Axaykumar Mehta Abhishek Rawat
• •

Priyesh Chauhan
Editors

Advances in Control Systems

and its Infrastructure
Proceedings of ICPCCI 2019

123
Editors
Axaykumar Mehta Abhishek Rawat
Department of Electrical Engineering Department of Electrical Engineering
Institute of Infrastructure Technology Institute of Infrastructure Technology
Research and Management (IITRAM) Research and Management (IITRAM)
Ahmedabad, Gujarat, India Ahmedabad, Gujarat, India

Priyesh Chauhan
Department of Electrical Engineering
Institute of Infrastructure Technology
Research and Management (IITRAM)
Ahmedabad, Gujarat, India

ISSN 1876-1100 ISSN 1876-1119 (electronic)

Lecture Notes in Electrical Engineering
ISBN 978-981-15-0225-5 ISBN 978-981-15-0226-2 (eBook)
https://doi.org/10.1007/978-981-15-0226-2
© Springer Nature Singapore Pte Ltd. 2020
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

This book presents a part of the proceedings of the International Conference on

Power, Control and Communication Infrastructure (ICPCCI 2019) held on 4–5 July
2019, at the Institute of Infrastructure Technology, Research and Management
(IITRAM), Ahmedabad. The conference was technically and financially supported by
Space Application Centre (SAC)—Indian Space Research Organization (ISRO);
Institution of Engineers India (IEI); Gujarat Council on Science and Technology
(GUJCOST), Government of India; and Science and Engineering Research Board
(SERB), Government of India. The conference aimed to be a melting pot of
researchers from academia and R&D organizations, and practicing engineers from the
utilities, industries and service providers to discuss, share and promote the ideas
related to state-of-the-art innovations in technologies for the power–control–com-
munication sectors. The conference included four keynote speeches: 1. C-band SAR
payload based on active antenna concept by Mr. Tapan Mishra, Distinguished
Scientist, Senior Advisor to Secretary, DOS/Chairman, ISRO; 2. Intelligent
hand-held robotics for surgery by Dr. Kalyana Veluvolu, Associate Professor,
Kyungpook National University, South Korea; 3. Transformers for modern power
grids by Dr. Parag Upadhyay, Principal R&D Scientist, US Corporate Research
Center of ABB, Raleigh, NC, USA; and 4. Importance of ethics in research pub-
lishing by Mr. Aninda Bose, Senior Editor, Springer. There was also an industry talk
on role of energy storage in Indian power system by Mr. J. D. Trivedi, Engineer,
State Load Dispatch Centre, Gujarat Energy Transmission Corporation Limited,
Vadodara. The papers were presented in eight technical sessions, viz. 1.
Power System Protection and Planning, 2. Communication Systems‐I, 3. Control
and Microgrids, 4. Computational Intelligence in Electrical System, 5. Integration of
Renewable Energy Sources and Electric Vehicles, 6. Communication Systems‐II, 7.
Power Converters and Control and 8. Condition Monitoring and Energy Economics.
The chapters included in this volume are on advances, latest trends, challenges
and developments of the control system technology and its infrastructure. The book
addresses a range of solutions to the problems faced by engineers and researchers to
design and develop controllers for emerging areas like smart grid, integration of
renewable energy, automated highway systems, haptics, unmanned aerial vehicles,

v
vi Preface

sensor networks, robotics, formation control and many more. The solutions
discussed in this book will encourage and inspire researchers, industry professionals
and policymakers to put these methods into practice.
We, the editors of this book, are thankful to all the contributors who have
submitted their research papers to ICPCCI 2019 to keep the standard high. The
editors would also like to acknowledge the reviewers from various domains for
their prompt and constructive assessments. We would also like to thank the
International Advisory Committee and the chairpersons of different committees.
Our special thanks to Series Editor, Lecture Notes in Electrical Engineering,
Springer, for giving us the opportunity to publish this edited volume in the series.
We are indeed grateful to Mr. Aninda Bose, Senior Editor, Hard Sciences, Springer
Nature India, for the cooperation and support to bring out this volume.

Ahmedabad, India Axaykumar Mehta

Abhishek Rawat
Priyesh Chauhan
Contents

Discrete Higher Order Sliding Mode Protocol for Leader-Following

Consensus of Heterogeneous Discrete Multi-Agent System . . . . . . . . . . . 1
Keyurkumar Patel and Axaykumar Mehta
Design of Sliding Mode Control for LNG Storage Tank System . . . . . . 11
Dashrath S. Panchal, Devesh P. Soni and Dipesh H. Shah
Distributed Event-Triggered Sliding Mode Control for Voltage
Synchronization of DC Microgrid Using Leader–Follower
Consensus Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Jatin Savaliya, Keyurkumar Patel and Axaykumar Mehta
A Fractional and Integer Order PID Controller for Nonlinear System:
Two Non-interacting Conical Tank Process Case Study . . . . . . . . . . . . . 37
Himanshukumar R. Patel and Vipul A. Shah
Modified Droop Control Strategy for Load Sharing and Circulating
Current Minimization in Low-Voltage Standalone DC Microgrid . . . . . 57
Rakesh Sukhadiaa and Saurabh Pandya
Primary and Secondary Droop Control Method for Islanded
Microgrid with Voltage Regulation and Current Sharing . . . . . . . . . . . 75
Hitesh M. Karkar and Indrajit N. Trivedi
Equal Load Sharing in DC Microgrid Using Line Resistance
Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Shivam Chaturvedi and Deepak M. Fulwani
Enhanced Control and Dynamic Power Management of Li-Ion
Battery and Ultracapacitor for Hybrid Energy Storage System . . . . . . . 97
Pranav Mohod, Pradyumn Chaturvedi and Shubham Gajbhiye
Active and Reactive Power Control of Photovoltaic Power Plant
Under Normal and Abnormal Grid Conditions . . . . . . . . . . . . . . . . . . . 109
Veera Joshi, Bhinal Mehta and Siddharth Joshi

vii
viii Contents

CPG-UVT-Based Grid-Connected PV System . . . . . . . . . . . . . . . . . . . . 123

Ravindra B. Kuhada, Bharti B. Parmar and Mahesh H. Pandya
Grid Integration of Solar PV System . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Vyas Dhruv, Chudasama Richa, Ambaleeya Afasana, Bosamiya Swati,
Gajjar Rital and Pandya Rajen
Comparative Analysis for INC and P&O MPPT Based Photovoltaic
Energy Conversion System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Anal Patel, Siddharth Joshi and Bhinal Mehta
Generation and Regulation of DC High-Voltage Using Modified
Technique with Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Varun H. Desai, Abdeali A. Challawala, Kevin B. Ray
and Mulav P. Rathod
Energy Regeneration in Direct Torque Control Drive
of Induction Motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Pravinkumar D. Patel and Saurabh N. Pandya
Emerging Heuristic Optimization Algorithms for Expansion
Planning and Flexibility Optimization in Sustainable Electrical
Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Jigar Sarda, Kartik Pandya and Margi Shah

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

About the Editors

Axaykumar Mehta received his Bachelor of Engineering in Electrical

Engineering (1996), M.Tech. (2002) and Ph.D. (2009) degrees from Gujarat
University, Ahmedabad, Indian Institute of Technology Kharagpur, and Indian
Institute of Technology Bombay, respectively. He is currently an Associate
Professor at the Institute of Infrastructure Technology Research and Management
(IITRAM), Ahmedabad, Gujarat, India. He has more than 20 years of teaching
experience at the undergraduate and graduate levels at various premier institutions.
He has published more than 75 research papers, book chapters in national/
international conference proceedings, books and journals. He has authored three
books with Springer Nature Singapore and also published five patents at the Indian
Patent Office, Mumbai. His research interests include sliding mode controls, net-
worked control systems, multi-agent systems and control of smart grids/microgrids.
Dr. Mehta received the Pedagogical Innovation Award from Gujarat Technological
University (GTU) in 2014, Dewang Mehta National Education Award (Regional)
for Best Professor in Electrical Engineering in 2018. He is a senior member of the
IEEE, member of the IEEE Industrial Electronics Society (IES) and Control System
Society (CSS) and life member of the Indian Society for Technical Education
(ISTE), Institute of Engineers India (IEI), Systems Society of India (SSI) and
Society of Power Engineers (SPE).

Abhishek Rawat received his Bachelor of Engineering in Electronics and

Communication Engineering (2001), from Rajiv Gandhi Technological University,
Bhopal. He received his Master of Technology (2006) and Ph.D. (2012) from
Maulana Azad National Institute of Technology, Bhopal, India. He is currently an
Assistant Professor at the Institute of Infrastructure Technology Research and
Management (IITRAM), Ahmedabad, Gujarat, India. He has 16 years of research,
academic and professional experience in the different premier institutions.
Dr. Rawat is also a senior member IEEE and published more than 50 articles in

ix
x About the Editors

international journals, book chapters and national and international conference

proceedings. He received the Young Scientist Award from MPCOST, Bhopal, in
2007, involved in the field trials of the IRNSS receiver and published four Indian
patents. His research interests include Navigation systems, Satellite Communication
and Peripheral security.

Priyesh Chauhan is an Assistant Professor at the Department of Electrical

Engineering, Institute of Infrastructure, Technology, Research & Management,
Ahmedabad, India. He completed his Ph.D. in Electrical Engineering (Power
Electronics & Electric Drives) at IIT Delhi, New Delhi, in 2014. With 17 years of
research, academic and professional experience, he has published 4 papers in ref-
ereed international journals and 18 papers in international conferences. He has filed
1 Indian patent application. He has also contributed to the preparation of 5 suc-
cessful project proposals with total approved grants of SGD 5 million. His research
interests are renewable energy source-based off-grid and microgrid-interactive
distributed generation, ship power systems architectures and optimization, and
energy storage integration, transitions and power management system.
Discrete Higher Order Sliding Mode
Protocol for Leader-Following Consensus
of Heterogeneous Discrete Multi-Agent
System

Keyurkumar Patel and Axaykumar Mehta

Abstract In this paper, we proposed higher order discrete sliding mode

control protocol using discrete-time supertwisting algorithm for the consensus of
leader-following discrete heterogeneous linear multi-agent system (DHLMAS). The
DHLMAS is represented using fixed, directed interaction graph topology. The pro-
posed consensus protocol has the ability to synchronize the heterogeneous follower
agents with leader agent in a finite time steps. The 2-DOF flexible joint and the
2-DOF flexible link robotic arms are considered as the leader and follower agents
respectively for the simulation and experimental study. Further, the robustness of
proposed consensus protocol is assessed by giving the matched disturbance.

Keywords Discrete-time sliding mode control (DSMC) · Discrete heterogeneous

linear multi-agent system (DHLMAS) · Discrete-time supertwisting algorithm
(DSTA)

1 Introduction

In last decade, distributed cooperative control among multiple systems known as

Multi-Agent System (MAS) has become popular among the academicians and the
researchers due to large range applications in different applied research domains,
such as power systems, under water vehicle, vehicle urban traffic management, and
many more undiscovered area of applications [1–3]. The main functionality in coop-
erative control of MAS is to develop the distributed local controllers by collecting
information from the neighbouring agent to achieve predefined task. To perform this
task, different algorithms of cooperative control, such as rendezvous, and flocking,
are proposed in literature [4, 5]. Among all algorithms, the consensus algorithm
has been most popular with different control solutions in recent years [6]. Consen-

K. Patel (B) · A. Mehta

Department of Electrical Engineering, Institute of Infrastructure Technology Research
and Management, Ahmedabad 380026, Gujarat, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 1

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_1
2 K. Patel and A. Mehta

sus means to reach a common agreement by interacting with each other. Mainly
consensus of MAS is divided into two parts [7]: first part is average consensus in
which individual agent update its information with neighbouring agent and agreed
upon average value amongst them. The second group is leader-following consen-
sus in which all the follower agents continuously monitor the leader information
and try to adjust with each other. The existing leader-following consensus protocols
appeared in the literature mostly have same leader and agent dynamics. However the
problem is more challenging when leader-following agents’ dynamics are different
in nature or heterogeneous.
To solve this kind of heterogeneous leader-following consensus problem many
academician and researchers have proposed different protocol. Zhang et al. proposed
leader-following synchronization of heterogeneous MAS using H∞ base consensus
protocol [8]. Hua et al. proposed a protocol of leader-following output consensus for
heterogeneous nonlinear MAS using state feedback method [9].
Simultaneously, different researchers around the world have also proposed a
leader-following consensus protocol for heterogeneous Discrete multi-agent system
(DMAS) in discrete domain which is reliable for digital implementation. Wang et
al. proposed leader-following consensus protocol of discrete heterogeneous multi-
agent system based on Lyapunov function [10]. Zhao et al. proposed protocol for the
consensus discrete heterogeneous MAS using algebraic graph theory [11].

2 Preliminaries of Discrete-Time Heterogeneous

Multi-Agent System

2.1 Graph Theory

Consider a weighted digraph G = (V, E, A) finite set of N nodes. V = {v1 , v2 , . . . ,

v N } is vertex, a set of edges E ⊂ V × V and the edges between vertex among all
the network graph defined as adjacency matrix A = [ai j ] ∈ R N ×N . An edge started
at particular node j and ended at node i is given by (v j , vi ), which means that the
information forwarded from node j to node i. For this the value of ai j for edge
(v j , vi ) is nonnegative, i.e., ai j > 0 if (v j , vi ) ∈ E, otherwise, ai j = 0. In this paper,
it is assumed that there is no any kind of self loops, i.e., aii = 0, ∀i ∈ N . Now let
us define the in or out-degree matrix as D = diag{di } ∈ R N ×N with di = j∈Ni ai j
and the Laplacian matrix as L = D − A. If information forwarded from j node to i
node than graph is said to be a directed graph (i, j) ∈ E. The positive numbers ai0,
i = 1, . . . , N is used to elaborate the communications between follower agents and
leader agent. If ith follower agent received the information from the leader agent than
ai0 > 0 otherwise ai0 = 0. The diagonal B = diag(a10, . . . , a N 0) matrix infer the
communication between follower agents and the leader agent.
Lemma 1 [16]. If digraph G has a rooted spanning tree, then the matrix (L + B) is
invertible.
Discrete Higher Order Sliding Mode Protocol … 3

2.2 Representation of Leader-Following Discrete-Time

Heterogeneous Linear Multi-Agent System (DHLMAS)

Discrete heterogeneous multi-agent system is defined as

xi (k + 1) = Fi xi (k) + G i (u i (k) + Di (k)) ∀i ∈ N, (1)

where i = 1, . . . , N , Fi ∈ Rn×n and G i ∈ Rn×m are the system matrix and input
matrix of ith system respectively. State vector xi (k) ∈ Rn and the input vector u i (k)
∈ Rm , Di ∈ Rm is matched disturbance acting on ith system.

Assumption 1 The matrix pair (F, G) for ith system in (1) is controllable.
The global DHLMAS from (1) is defined as

X (k + 1) = (I N ⊗ Fi )X (k) + (I N ⊗ G i )(u(k) + D(k)), (2)

xi (k) = [x1 (k), x2 (k), . . . , x N (k)]T ∈ Rn N and the input vector u i (k) = [u 1 (k),
u 2 (k), . . . , u N (k)]T ∈ Rm N , Di (k) = [D1 (k), D2 (k), D3 (k), . . . , D N (k)]T ∈ Rm N
matched disturbance vector acting on ith system. The disturbance is assumed to
be bounded for the ith agent, where by Di (k) ≤ i with a known upper bound
i > 0, ⊗ denotes the kronecker product.
Equation (2) may rewritten as

X (k + 1) = F̃ X (k) + G̃(u(k) + D(k)), (3)

where F̃ = (I N ⊗ Fi ) , G̃ = (I N ⊗ G i ).
The leader dynamics are defined as

x0 (k + 1) = F x0 (k) + Gu 0 (k), (4)

where x0 (k) ∈ Rn is the state vector of the leader.

Definition 1 The system defined in (3) of DHLMAS is achieve the consensus in

finite time step k ∗ ∈ [0, ∞) such that

lim X i (k) − x0 (k) = 0. (5)

k→k ∗

Problem identification: To develop a robust DSMC-based higher order consensus

protocol using DSTA for global DHLMAS such that all the follower agents (3) follow
the leader trajectory for fixed, undirected graph topology.
The local neighbouring error of leader-follower network is defined as [13]

ei (k) = ai j (k)[xi (k) − x j (k)] + ai0 (k)[xi (k) − x0 (k)]. (6)
j∈N
4 K. Patel and A. Mehta

Using (6) and according to Lemma 1, the graph theory based global consensus error
can be derived as
 
ē(k) = (I N + D + B)−1 )(L + B)) ⊗ In x̃, (7)

The weighted matrix Γ¯ = ((I N + D + B)−1 )(L + B)) eigenvalues are obtained
using Gersgorin circle criteria [14], which is inside the unit circle. And x̃ =
X (k) − 1 N ⊗ x0 (k)
we can express (7) as
ē(k) = (Γ¯ ⊗ In )x̃. (8)

Consider (Γ¯ ⊗ In ) = γ̄ and substituting the value of x̃

ē(k) = γ̄ (X (k) − 1 N ⊗ x0 (k)) , (9)

ē(k + 1) = γ̄(X (k + 1) − 1 N ⊗ x0 (k + 1)), (10)

3 Higher Order DSMC for the Consensus of DHLMAS

Using DSTA

In this section, higher order DSMC consensus protocol using DSTA for the global
DHLMAS is derived in the form of theorem (1) as given below.
Theorem 1

u(k) = −(γ̄ G̃)−1 γ̄ F̃ X (k) + γ̄(−1 N ⊗ x0 (k + 1)) − s̃(k)
 
+h̄ 1 T |s̃(k)| sign(s̃(k)) − Z (k) − D(k). (11)

Proof sliding surface for individual ith agent is defined using (6) as

s̃i (k) = ei (k). (12)

The global sliding surface is rewritten for leader-follower network as

s̃(k) = ē(k). (13)

Now using (9)

s̃(k + 1) = γ̄ (X (k + 1) − 1 N ⊗ x0 (k + 1)) . (14)

Further,

s̃(k + 1) = γ̄ F̃ X (k) + G̃(u(k) + D(k)) − 1 N ⊗ x0 (k + 1) . (15)

Discrete Higher Order Sliding Mode Protocol … 5

Now motivated from the DSTA [15], let us define higher order DSMC using DSTA
for ith agent as

s̃i (k + 1) = s̃i (k) − h̄ 1i T |s̃i (k)|sign(s̃(k)) + z i (k)
z i (k + 1) = z i (k) − h̄ 2i T sign(s̃i (k)) (16)

where h̄ 1i = [h̄ 11 , h̄ 12 , h̄ 13 , . . . h̄ 1 N ] ∈ Rn N , h̄ 2i = [h̄ 21 , h̄ 22 , h̄ 23 , . . . h̄ 2 N ] ∈ Rn N are

gain parameters, T is the sampling time.
Then DSTA for global consensus of global DHLMAS can be defined as

s̃(k + 1) = s̃(k) − h̄ 1 T |s̃(k)|sign(s̃(k)) + Z (k)
Z (k + 1) = Z (k) − h̄ 2 T sign(s̃(k)) (17)

Comparing (15) and (17), we may write


s̃(k) − h̄ 1 T s̃(k)sign(s̃(k)) + Z (k) = γ̄ F̃ X (k) + G̃(u(k)
+D(k)) − 1 N ⊗ x0 (k + 1)) (18)

From (18), we may derive


u(k) = −(γ̄ G̃)−1 γ̄ F̃ X (k) + γ̄(−1 N ⊗ x0 (k + 1)) − s̃(k)
 
+h̄ 1 T |s̃(k)|sign(s̃(k)) − Z (k) − D(k). (19)

This completes the proof.

4 Results Discussion

4.1 Simulation and Experimental Results Discussion

In this section, we consider total four number of agents, among them three follower
agents and one leader agent with different dynamics for leader-following consensus.
In this study, 2-DOF flexible joint robotic arm and 2-DOF flexible link robotic arm
position(θ11 ) of stage-1 and position (θ21 ) of stage-2 are to be consider for consensus.
The simulation and experimental study are executed using Matlab R15 interface with
QUARC software . The discrete model of 2-DOF serial flexible joint robotic arms
for stage-1 and stage-2 are discretized at sampling rate T = 0.002 and is obtained
as
6 K. Patel and A. Mehta
⎡ ⎤
0.9997 0.0002696 0.001865 1.818 × 10−7
⎢7.81 × 10−5 0.9999 5.028 × 10−8 0.001999 ⎥
Fi1 = ⎢
⎣ −0.2634
⎥,
0.2634 0.868 0.0002695 ⎦
0.07809 −0.07809 7.455 × 10−5 0.9993
 
G i1 = 0.0002673 3.545 × 10 0.2612 7.042e × 10−6 ,
−9
 
Ci1 = 1 0 0 0 . (20)

Similarly, for stage-2

⎡ ⎤
0.9979 0.002078 0.001739 1.416 × 10−6
⎢0.0007443 0.9993 4.63 × 10−7 0.001994 ⎥
Fi2 = ⎢
⎣ −1.983
⎥,
1.983 0.7499 0.002075 ⎦
0.7433 −0.7433 0.0006783 0.994
 −8

G i2 = 0.0004521 5.827 × 10 0.4315 0.0001149 ,
 
Ci2 = 1 0 0 0 . (21)

Similarly, The state space model for 2-DOF serial flexible link robotic arms for
stage-1 and stage-2 is obtained as
⎡ ⎤
1 0.001206 0.001879 8.126 × 10−5
⎢0 0.9983 0.0001208 0.001999 ⎥
Fi1 = ⎢
⎣0 1.181
⎥,
0.8817 0.001206 ⎦
0 −1.65 0.1182 0.9983
 
G i1 = 0.0002695 −0.0002694 0.2639 −0.2638 ,
 
Ci1 = 1 0 0 0 . (22)

⎡ ⎤
1 0.003261 0.00128 4.306 × 10−5
⎢0 0.9947 0.0007132 0.00193 ⎥
Fi2 = ⎢
⎣0 2.763
⎥,
0.3857 0.03775 ⎦
0 −4.818 0.6043 0.9345
 
G i2 = 0.0004178 −0.0004137 0.3563 −0.3505 ,
 
Ci2 = 1 0 0 0 . (23)

Figure 1 represent the leader-following DHLMAS system where the leader nota-
tion is given as 0 and remaining follower agents notation is 1, 2, 3, respectively. A
square signal with amplitude 20 (◦ ) with frequency 0.1 Hz is given to the leader agent
for both the stage of robotic arms. The various informative matrix for communica-
tion between agents such as adjacency matrix A, diagonal matrix D, pinning gains
matrix B and Laplacian matrix L are defined as
Discrete Higher Order Sliding Mode Protocol … 7
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
100 000 1 0 0
D = ⎣0 2 0 ⎦ , A = ⎣1 0 0 ⎦, B = diag{1, 0, 0}, L = ⎣−1 2 0 ⎦ . (24)
001 010 0 −1 1

For checking the robustness of proposed consensus protocol, a matched distur-

bance is given to the each DHLMAS with magnitude Di (k) = 0.002 ∗ cos(0.01k) to
the follower agent. Gain for each ith follower agent h̄ 1i , h̄ 2i are considered as 0.15
and 0.0002 for DSTA.
The proposed consensus protocol defined in (19) is applied to the system deifned
in (3). Figure 2 shows the simulation result of position consensus of 2-DDF robotic
heterogeneous system stage-1 using DSTA and find that all the follower agents follow
the leader trajectory in a finite time.
Similarly, Fig. 3 shows the simulation result of position consensus of 2-DOF
robotic heterogeneous system stage-2.
Figure 4 shows the experimental setup available in our laboratory.
Figures 5 and 6 show experimental validation of the position consensus of 2-DOF
robotic heterogeneous system for stage-1 and stage-2 using DSTA protocol. It is
observed that all the follower agents follow the leader trajectory in finite time.

Fig. 1 The communication topology for leader-following consensus

8 K. Patel and A. Mehta

Fig. 2 Position (θ11 ) consensus of 2DOF robotic system stage-1

Fig. 3 Position (θ21 ) consensus of 2DOF robotic system stage-2

Fig. 4 Experimental setup for leader-follower consensus

Discrete Higher Order Sliding Mode Protocol … 9

Consensus of heterogeneous system ( Experimental Result- stage-1)

40
Leader position
Position consensus stage-1 of robotic arm

Position of follower agent-1

30 Position of follower agent-2
Position of follower agent-3

20

10

-10

-20

-30

-40
0 5 10 15 20 25 30

Time(seconds)

Fig. 5 Position (θ11 ) consensus of 2DOF robotic system stage-1

Consensus of heterogeneous system ( Experimental Result - stgae-2)

25
Leader position
Position consensus stage-2 of robotic arm

Position of follower agent-1

20
Position of follower agent-2
Position of follower agent-3
15

10

-5

-10

-15

-20

-25
0 5 10 15 20 25 30

Time(seconds)

Fig. 6 Position (θ21 ) consensus of 2DOF robotic system stage-2

5 Conclusion

In this paper, a higher order DSMC-based DSTA consensus protocol is proposed for
leader-following network of DHLMAS using graph theory. Proposed protocol attain
the consensus of leader-follower network in finite time steps. The proposed consensus
protocol is validated in simulation as well as experimentally using 2-DOF serial
flexible joint and 2-DOF flexible link robotic arms. Moreover, it is inferred from the
results that this consensus protocol globally applies to the actual application. Finally,
robustness property is checked by applying matched disturbance to the individual
follower agents. In future, the DSTA consensus protocol for consensus of DHLMAS
with switching topology, bipartite network shall be explored.
10 K. Patel and A. Mehta

References

1. Huang C, Feng C, CAO J, (2019) Consensus of cyber-physical power systems based on multi-
agent systems with communication constraints. J Modern Power Syst Clean Energy. https://
doi.org/10.1007/s40565-018-0491-4
2. Chen S, Ho D (2016) Consensus control for multiple AUVs under imperfect information caused
by communication faults. Inf Sci 370–371:565–577. https://doi.org/10.1016/j.ins.2016.04.037
3. Zhonghe H, Yangzhou C, Jianjun S et al (2013) Consensus based approach to the signal control
of Urban traffic networks. Procedia Soc Behav Sci 96:2511–2522. https://doi.org/10.1016/j.
sbspro.2013.08.281
4. Bhowmick C, Behera L, Shukla A, Karki H (2016) Flocking control of multi-agent system
with leader-follower architecture using consensus based estimated flocking center. In: IECON
2016—42nd annual conference of the IEEE industrial electronics society. https://doi.org/10.
1109/iecon.2016.7793149
5. Liu W, Huang J (2017) Adaptive leader-following rendezvous and flocking for a class of
uncertain second-order nonlinear multi-agent systems. Control Theory Technol 15:354–363.
https://doi.org/10.1007/s11768-017-7083-0
6. Qin J, Ma Q, Shi Y, Wang L (2017) Recent advances in consensus of multi-agent systems: a brief
survey. IEEE Trans Ind Electron 64:4972–4983. https://doi.org/10.1109/tie.2016.2636810
7. Fan Y, Yang J (2016) Average consensus of multi-agent systems with self-triggered controllers.
Neurocomputing 177:33–39. https://doi.org/10.1016/j.neucom.2015.10.117
8. Liu K, Duan P, Duan Z et al (2018) Leader-following consensus of multi-agent systems with
switching networks and event-triggered control. IEEE Trans Circ Syst I: Reg Papers 65:1696–
1706. https://doi.org/10.1109/tcsi.2017.2762420
9. Zhang D, Xu Z, Karimi H et al (2018) Distributed H∞ output-feedback control for consensus of
heterogeneous linear multiagent systems with aperiodic sampled-data communications. IEEE
Trans Ind Electron 65:4145–4155. https://doi.org/10.1109/tie.2017.2772196
10. Hua C, Li K, Guan X (2019) Leader-following output consensus for high-order nonlinear
multiagent systems. IEEE Trans Autom Control 64:1156–1161. https://doi.org/10.1109/tac.
2018.2836024
11. Wang D, Yu M (2016) Leader-following consensus for heterogeneous multi-agent systems
with bounded communication delays. In: 14th international conference on control, automation,
robotics and vision (ICARCV). https://doi.org/10.1109/icarcv.2016.7838628
12. Zhao H, Fei S (2018) The consensus for discrete-time heterogeneous networked systems con-
sisting of second-order agents and third-order agents. IEEE Access 6:14204–14211. https://
doi.org/10.1109/access.2018.2804929
13. Patel K, Mehta A (2018) Discrete-time sliding mode control for leader following discrete-time
multi-agent system. In: IECON 2018-44th annual conference of the IEEE industrial electronics
society. https://doi.org/10.1109/iecon.2018.8591273
14. Varga R (2004) Gergorin and his circles. Springer Ser Comput Math. https://doi.org/10.1007/
978-3-642-17798-9
15. da Silva M, de Camargo R, Pinheiro J et al (2014) Discrete-time super-twisting sliding-mode
control applied to a DSTATCOM-based voltage regulator for a self excited induction generator.
In: IECON 2014-40th annual conference of the IEEE industrial electronics society. https://doi.
org/10.1109/iecon.2014.7049145
16. Ren C, Chen C (2015) Sliding mode leader-following consensus controllers for second-order
non-linear multi-agent systems. IET Control Theory Appl 9:1544–1552. https://doi.org/10.
1049/iet-cta.2014.0523
Design of Sliding Mode Control for LNG
Storage Tank System

Dashrath S. Panchal, Devesh P. Soni and Dipesh H. Shah

Abstract This paper presents the seismic control problem of a broad Liquefied Nat-
ural Gas (LNG) storage tank using Sliding Mode Control (SMC) with base isolation
technique. To reduce seismic forces, the LNG tank is isolated from base by lam-
inated isolation bearing made from intrinsic rubber. In order to control excessive
displacement of LNG tank at isolation level, magneto-rheological (MR) dampers
are installed. The Magneto-rheological (MR) dampers are admired by robust slid-
ing mode control designed using constant rate reaching law. The stability condition
for closed-loop system is derived using Lyapunov Approach. The robustness of the
designed controller is proved and compared with uncontrolled system for different
seismic ground motion. The simulation results substantiated that the sliding mode
control (SMC) strategy is more effective in reducing the structural responses as
compared to uncontrolled system.

Keywords Sliding mode control · Clipped optimal controller ·

Magneto-rheological damper · Stability

D. S. Panchal (B)
Sardar Vallabhbhai Patel Institute of Technology, Vasad, India
e-mail: [email protected]
D. P. Soni
Department of Civil Engineering, Sardar Vallabhbhai Patel Institute of Technology, Vasad, India
e-mail: [email protected]
D. H. Shah
Department of Instrumentation and Control Engineering, Sardar Vallabhbhai Patel Institute of
Technology, Vasad, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 11

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_2
12 D. S. Panchal et al.

1 Introduction

Our industrialized world increases energy demand day by day. With the increase in
energy demand more LNG storage tanks are required. Nowadays, to control liquid
storage tank is come into the picture and to establish with control this type of struc-
ture with seismic control strategy. Seismic isolation works by horizontally installing
flexible bearings under a structure which serves to extend its vibration period beyond
the predominant vibration period of the ground motion and thereby increasing its
energy dissipation capability. The volumes of these tanks are very large and have
capacities of about 160,000 m3 . A full containment tank as shown in Fig. 1a is con-
tented of an inside tank with steel material and an outer tank with concrete material.
Figure 1b shows a typical LNG chain procedure of extraction, transportation, and
regasification. Natural gas is a fossil fuel content of majorly methane (90%) and
lesser compounds of nitrogen, oxygen, carbon dioxide, and sulfur. With liquefaction
procedure that produces LNG removes any oxygen, carbon dioxide, sulfur content,
and water. At standard temperature is approximately about −161.52 °C. LNG also
provides terminal facilities with a pipeline construction.
LNG storage tanks are strategically very important structures; hence, effective
protection against seismic excitations is required. Failures of storage tanks not only
instantly disrupts essential infrastructure but can also cause fires or environmental
contaminations when flammable materials or hazardous chemicals leak. Hence to
protect such structures from the seismic ground motions, implementation of appro-
priate base isolation system proves to be more effective as compared to any other
technique. Also, the design, construction, and analysis of LNG storage facility need
highly advanced technology compared to the general structures due to hazardous
nature of LNG.

Fig. 1 a View of modern LNG tank. b LNG chain: extraction, transportation and regasification
Design of Sliding Mode Control for LNG Storage Tank System 13

Several analytical and parametric studies were carried out in the past to demon-
strate the effectiveness of seismic isolation for earthquake resistant design of LNG
storage tanks. Malhotra et al. [2] provided the theoretical background of a simpli-
fied seismic design procedure for cylindrical ground-supported tanks by considering
impulsive and convective (sloshing) actions of the liquid in flexible steel or concrete
tanks fixed to rigid foundations. Jansen and Dyke [3] presented the evaluated of dif-
ferent types of semi-active control strategy used with number of MR damper. Iemura
et al. [4] presented the study of cylindrical supported tanks with different strategies
likewise passive, semi-active, and active strategy. Jin et al. [5] applied axisymmetric
finite element mesh to LNG storage tanks. For a finite analysis of tank to find out the
different parts of the structure like as inner tank, outer tank, and convective height of
the liquid. Douglas et al. [6] presented the extra-large tank used with full containment.
Different isolator may be used with appropriate suitable strategy likewise Friction
pendulum bearing and high damping rubber bearing. Dotoli et al. [7] simulated the
seismic behavior of an LNG tank during an earthquake. Christovasilis and Whittaker
[8] computed the earthquake response under isolated vertical, cylindrical tank with
a mechanical analog. Lee et al. [9] addressed that the study with fixed and isolator
base LNG storage tank and also computed base shear, displacement, and sloshing
action should be studied. Bharti et al. [10] studied the effectiveness of MR damper for
seismic response mitigation of adjacent multistory buildings using Lyapunov direct
approach control scheme, involving passive-off, passive-on, and semi-active control
strategies. Based on the results of the numerical study, it has been observed that the
MR damper is an effective device to control the response of both the buildings for
a wide range of ground motion. Panchal and Soni [12] studied the seismic response
of LNG storage tank likewise isolated, vertical tank using MFPS system. Shah and
Mehta [14] presented a sliding mode controller is developed for a class of nonlinear
multi-input and multi-output (MIMO) systems. With a MIMO system to control with
Lyapunov Scheme for disturbances.
From all above-mentioned literature survey [1–14] the authors concluded that
there is need for developing the robust control technology for the LNG Tank system
that nullifies the effect of disturbances which is one of the major concerns in LNG
storage tank system. This motivates the authors to develop a robust control strategy for
base-isolated LNG storage tank equipped with magneto-rheological damper system
in the presence of matched uncertainty. Recently, SMC has received much attention
in controlling structures due to its robustness properties and simple configuration.
The controller has ability to reject the effect of matched and unmatched uncertainty
that exists in real time systems which results in improvement of performance of the
system.
The structure of the paper is as follows: The mathematical model of controlled
LNG tank system and problem statement is mentioned in Sect. 2. Section 3 describes
the design of sliding mode control for LNG tank system using constant rate reaching
law. The stability analysis of closed-loop system is presented in Sect. 4. The sim-
ulation results and comparative analysis of controlled and uncontrolled system are
mentioned in Sect. 5 followed by conclusion in Sect. 6.
14 D. S. Panchal et al.

2 Modeling of LNG Tank

The liquefied natural gas storage tank is structurally constructed and modeled into
two layers. Schematically these layers are constructed with two appropriate models,
one is outer portion of the tank designed as Dunkerly model and second one is the
inner tank which is designed by Malhotra’s model. Figure 2 shows the diagram of
model of LNG storage tank.
The procedure of structural model suggested by Malhotra et al. [2] was based on
the work of Veleto’s with some modifications that included the following:
• The higher impulsive modal mass can be combined with the first impulsive mode
whereas the higher convective modal mass can be combined with the first convec-
tive mode.
• Modal heights should be modified which accounts the contribution of higher modes
to the base overturning moment.
• Generalizing the formula for the impulsive period so that it could be applied to
steel and concrete tanks of various wall thicknesses.
As shown in Fig. 2, it can be noticed that the signal generated from the MR damper
will be applied to the controller. The controller would compute the measured signal
and appropriate control actions would be generated to control the damper force in
the presence of matched uncertainty that is applied at the input side of the system.
The governing equations of motion of the LNG tank are expressed as

m c (ü c + ü b ) + cc u̇ c + kc u c = F − m c ü g , (1)

m i (ü i + ü b ) + ci u̇ i + ki u i = F − m i ü g , (2)

m c ü c + m i ü i + m it ü b + ce u̇ b + ke u b = F − m it ü g , (3)

Fig. 2 Schematic of
simplified model of LNG
tank
Design of Sliding Mode Control for LNG Storage Tank System 15

m ot (ü ot + ü b ) + cot u̇ ot + kot u ot = F − m ot ü g , (4)

m it = m c + m i + m iw + m b , (5)

where uc is the displacement relative to the bearing displacement in the horizontal

direction for the convective mass, ui and uot is that for the impulsive and outer tank
mass, ub is the displacement of the bearings relative to the ground in the horizontal
direction, cc and ci are the damping coefficient of the convective and impulsive
mass, cot is that of the outer tank mass, ce is the equivalent damping of the isolation
devices, k c , k i and mass, k e is the equivalent stiffness of the isolation devices, u¨g is
the earthquake ground acceleration, F is the resisting force, miw is the mass of the
inner tank wall, mot is the total mass of the outer tank, mb is the mass of the bottom
plate and foam glass, and mit is the total mass of the inner tank at bottom plate.
Thus, the state-space model of the above system [14] is given by

ẋ(t) = Ax(t) + B(u(t) + D d(t)). (6)

y(t) = C x(t). (7)

where
       
0 1 0 −M −1 K −M −1 C M −1 
A= ,B = ,C = ,D = .
−M −1 K −M −1 C M −1  100 000 0

x(t) ∈ R n×1 is system state vector, u(t) ∈ R m×1 is control input vector in terms
of voltage, y(t) ∈ R r ×1 is system output vector, A ∈ R n×n , B ∈ R n×m , C ∈ R r ×n ,
D ∈ R p×n are the matrices of appropriate dimensions, d(t) ∈ R p×1 is matched
bounded disturbance vector applied at input side of the channel with |d(t)| ≤ dmax .
Problem Statement: The main objective is to design and compare the robustness
of sliding mode control derived using constant rate reaching law with uncontrolled
system for LNG storage tank system (6), (7) in the presence of matched disturbances
applied at the ground level.

3 Design of SMC for LNG Tank System

In this section, sliding mode controller is designed for controlled for LNG tank
system (6), (7) in the presence of matched uncertainty. The sliding mode control
involves (i) design of sliding surface and (ii) reaching law.
Theorem 1 The controller designed for system (6), (7) would generate the proper
control force to MR damper system in the presence of matched uncertainty provided
the control law is designed as
16 D. S. Panchal et al.

u(t) = −(Cs B)−1 [Cs Ax(t) + qsgn(s(t))] − Dd(t). (8)

Proof Consider the sliding surface as

s(t) = Cs x(t), (9)

where s(t) is the sliding surface and Cs is the sliding gain vector designed using LQR
approach.
The sliding surface in (9) should satisfy the ‘’ reaching condition that ensures
the finite-time convergence to s = 0.

ṡs < −|s|. > 0, ∀t. (10)

Thus in order to satisfy condition (10) the constant reaching law proposed in [11]
is given by

ṡ = −qsgn(s(t)). (11)

where q > 0.
Referring to Eq. (9), (11) can be written as

Cs ẋ(t) = −qsgn(s(t)). (12)

Substituting the value of ẋ(t) in Eq. (12) we have

Cs [Ax(t) + B(u(t) + Dd(t))] = −qsgn(s(t)). (13)

Further simplification gives

Cs Ax(t) + Cs B(u(t) + D d(t))] = −qsgn(s(t)). (14)

Further solving above Eq. (14), the control law can be expressed as

u(t) = −(Cs B)−1 [Cs Ax(t) + qsgn(s(t))] − Dd(t). (15)

This completes the Proof.

4 Stability Analysis

In this section, the stability analysis of the closed system (6), (7) is proved for the
designed controller (15) in the presence of matched uncertainty and sliding surface
(9).
Design of Sliding Mode Control for LNG Storage Tank System 17

Consider the quadratic Lyapunov function as

1 T
Vs (t) = s (t)s(t). (16)
2
Taking time derivative of Eq. (16) we have

V̇s (t) = s T (t)ṡ(t), (17)

Referring Eq. (11), Eq. (17) can be written as

V̇s (t) = s T (t)[Cs ẋ(t)], (18)

Substituting the value of ẋ(t) we have

V̇s (t) = s T (t)[Cs [Ax(t) + B(u(t) + Dd(t))]], (19)

Substituting the value of u(t) from Eq. (18) we have

   
V̇s (t) = s T (t) Cs Ax(t) + B −(Cs B)−1 [Cs Ax(t) + qsgn(s(t))] − Dd(t) + Dd(t) ,
(20)

Further simplification gives

V̇s (t) = −s T (t)[qsign(s(t))] < 0. (21)

The term V̇s (t) is negative as q is strictly positive integer. So the closed-loop
system is asymptotically stable in the presence of matched uncertainty.

5 Numerical Study

In the present study, the following parameters of LNG storage tank are taken from
the paper of Panchal and Soni [12] and that of MR damper are taken from the paper
of Bharti et al. [10] The LNG storage tank is composed of outer concrete and inner
steel tank. The inner steel tank has a radius r of 40 m and total height of 35 m which
is fully anchored to a concrete slab. The tank is filled to a liquid height H of 33 m.
The inner tank contains LNG having density ρl as 480 kg/m3 . The total mass of LNG
ml is 7.96 × 107 kg. The tank wall is made of three courses, the lower course is
25 mm thick, the middle course is 18 mm, and the upper course is 12 mm thick. The
total mass of the inner tank wall miw is 1.21 × 106 kg, and the height of its center of
gravity hiw is 14.59 m. For steel, Es is 2 × 1011 N/m2 , ρ = 7.9×103 kg/m3 . The total
mass of the inner tank at bottom plate mit is 1.82 × 107 kg. The outer tank wall is
18 D. S. Panchal et al.

Fig. 3 Clipped optimal

control forces

made up of concrete where height L is 40 m, the medium radius of the outer tank Dc
is 41 m, the wall thickness of the outer tank tc is 0.9 m, the density of the concrete
ρc is 2500 kg/m3 , the modulus of elasticity of the concrete Ec is 3 × 1010 N/m2 , the
Poisson’s ratio of the concrete wall υc is 0.3, the dome mass md is 2.93 × 107 kg,
and the total mass of the outer tank wall mot is 2.32 × 107 kg.
The MR damper parameters have been suitably scaled to suit the damper defor-
mation behavior and the values of which are as follows: α0a = 8.70 kN/m/V, γ =
496 m-2, α0b = 6.40 kN/m/V, β = 496 m-2, c0a = 50.30 kN s/m, η = 195 sec-1, c0b
= 48.70 kN s/m/V, k0 = 0.0054 kN/m, c1a = 8106.20 kN s/m, k1 = 0.0087 kN/m,
c1b = 7807.90 kN s/m/V, x0 = 0.18 m, Ad = 810.50, n = 2. Based on the above
parameters for LNG storage tank as well as for the MR damper the simulation has
been generated in MATLAB using SIMULINK Tool. The simulation program has
been developed for uncontrolled, base isolated and controlled systems under sliding
mode control.
To induce MR damper to generate approximately the corresponding desired opti-
mal control force f ci , the command signal is Vi is selected as follows. The voltage
force relationship and the same can be graphically represented as shown in Fig. 3.
The effectiveness of MR damper is determined for the sliding mode control strat-
egy. Under sliding mode control the damper command voltage is governed by the
control law. Clipped optimal control law with sliding mode control is utilized to
control the damper forces and state-space approach is used to solve the dynamic
equations, both for uncontrolled and controlled system. The results of the study are
evaluated for the past three earthquake ground motion history namely, Imperial Val-
ley ground motion of 1940 with PGA = 0.313 g, Imperial valley California ground
motion of 1979 with PGA = 0.358 g, and Loma Prieta 1989 ground motion with
PGA = 0.512 g.
The sliding gain Cs is computed using LQR method by proper selec-
tion of Q = diag(1000, 0) and R = diag(1, 0) matrices having Cs =
[26.685534.04918.1024 − 3.03550.78606.2791 − 12.8164 − 28.2493]. The sys-
tem shows the stable and satisfactory response for q = 1. State-space matrices
are obtained
Design of Sliding Mode Control for LNG Storage Tank System 19

Matrix A =
⎡ ⎤
0 0 0 0 1 0 0 0
⎢ 0 0 0 0 0 1 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 1 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 0 1 ⎥
⎢ ⎥;
⎢ −491.1092 −44.5066 0.1119 9.8696 −2.5715 −0.1326 −0.0018 0.6283 ⎥
⎢ ⎥
⎢ −126.3822 −224.7526 −0.1119 9.8696 −0.6618 −0.6696 −0.0018 0.6286 ⎥
⎢ ⎥
⎣ −126.3822 −44.5056 −0.5139 9.8696 −0.6618 −0.1326 −0.0081 0.6283 ⎦
126.3822 44.5056 0.1119 −9.8696 0.6618 0.1326 0.0018 −0.6283
⎡ ⎤
0
⎢ 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 ⎥
Matrix B = ⎢ ⎥;
⎢ 0.66 ⎥
⎢ ⎥
⎢ 0.66 ⎥
⎢ ⎥
⎣ 0.66 ⎦
−0.66
Matrix C =
⎡ ⎤
1 0 0 0 0 0 0 0
⎢ 0 1 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 1 0 0 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 1 0 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 1 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 1 0 0 ⎥
⎢ ⎥;
⎢ ⎥
⎢ 0 0 0 0 0 0 1 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 0 1 ⎥
⎢ ⎥
⎢ −491.1092 −44.5056 −0.1119 9.8696 −2.5715 −0.1326 −0.0018 0.6283 ⎥
⎢ ⎥
⎢ −126.3822 −224.7526 −0.1119 9.8696 −0.6618 −0.6696 −0.0018 0.6283 ⎥
⎢ ⎥
⎣ −126.3822 −44.5056 −0.5139 9.8696 −0.6618 −0.1326 −0.0081 0.6283 ⎦
126.3822 44.5056 0.1119 −9.8696 0.6618 0.1326 0.0018 −0.6283
20 D. S. Panchal et al.
⎡ ⎤
0
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢
Matrix D = ⎢ ⎥;
0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ ⎥
⎢ 6.60022440762986e − 09 ⎥
⎢ ⎥
⎢ 6.60022440762986e − 09 ⎥
⎢ ⎥
⎣ 6.60022440762986e − 09 ⎦
−6.60022440762986e − 09

The response parameters of interest for the study are outer tank displacement,
impulsive displacement, convective displacement, and outer absolute acceleration.
The hysteresis behavior of MR damper is also studied. The peak response quantities
for each earthquake are shown in Table 1.
The graphical representation of displacement and acceleration history under earth-
quakes is depicted in Figs. 4, 5, 6. Hysteresis behavior of MR damper for the three
earthquakes is shown in Fig. 7. It is noted from Table 1 that percentage reductions
in outer tank displacement for the controlled system are approximately about 85, 75
and 65% under Imperial Valley (1940), Loma Prieta (1980) and Imperial Valley Cal-
ifornia (1979) earthquake, respectively. Similarly, the impulsive displacement and
outer absolute acceleration reduce in a range of about 85% whereas the convective
displacement reduces by about 5–25%.

6 Conclusion

In this paper, a robust sliding mode controller is designed for LNG storage tank
system in the presence of matched uncertainty. The control algorithm is derived
using constant rate reaching law drives the system variables onto the switching plane
at a constant rate. The stability of closed-loop MIMO system is assured through
Lyapunov approach. The performance of the resulting control algorithm is compared
to uncontrolled system through simulation for the selected three earthquake ground
motion data. The following remarks can be concluded from the simulation results
obtained from the numerical study as follows:
• The value of outer tank displacement, Impulsive displacement, and convective
displacement reduces considerably for the controller system as compared to the
uncontrolled one under wide range of earthquake ground motions.
• The controlled system results in a reduction in the value of displacement with the
decrease in outer absolute acceleration.
Table 1 Peak response quantitues of the tank in the presence of various earthquakes
Earthquake 1940 Imperial Valley 1980 Loma Prieta 1979 Imperial Valley California
Control Uncontrolled Base isolated Controlled Uncontrolled Base isolated Controlled Uncontrolled Base isolated Controlled
strategy
Outer 0.0183 0.004 0.003 0.0234 0.0021 0.002 0.033 0.0114 0.008
displacement
(m)
Impulsive 0.0681 0.009 0.008 0.0396 0.0058 0.004 0.0708 0.0247 0.018
displacement
(m)
Convective 0.4563 0.312 0.305 0.26 0.2672 0.24 1.68 1.67 1.572
displacement
(m)
Design of Sliding Mode Control for LNG Storage Tank System

Isolator – 0.234 0.178 – 0.1233 0.089 – 0.6380 0.44

displacement
(m)
Outer 6.19 0.53 0.29 8.56 0.61 0.58 10.7 0.51 0.44
acceleration
(m/s2 )
21
22 D. S. Panchal et al.

0.02
displacement (m) 0.0041
Controlled Uncontrolled Base isolated
Outer tank

0.0030

-0.02
0 5 10 15 20 25 30
Time (sec)
0.1
displacement (m)

Controlled Uncontrolled Base isolated

0.0099
Impulsive

0.0085

-0.1
0 5 10 15 20 25 30
Time (sec)
displacement (m)

1
Convective

Controlled Uncontrolled Base isolated

0.3121
0
0.3058

-1
0 5 10 15 20 25 30
Time (sec)

Fig. 4 Outer tank, impulsive, and convective displacement under 1940 Imperial Valley earthquake

0.5
displacment(m)

0.2344 Controlled Base isolated

Isolator

0.1789

-0.5
0 5 10 15 20 25 30
Time(sec)

Fig. 5 Isolator displacement under 1940 Imperial Valley earthquake

10
Base Isolated Uncontrolled Controlled
Acceleration(m/s2)

0.53
Outer absolute

-5
0.29

-10
0 5 10 15 20 25 30
Time(sec)

Fig. 6 Outer absolute acceleration under 1940 Imperial Valley earthquake

Design of Sliding Mode Control for LNG Storage Tank System 23

0.05
damper force(kN)
Normalized

-0.05
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
Time(sec)

Fig. 7 MR damper hysteresis behavior history under 1940 Imperial Valley earthquake

References

1. Dunkerley S (1894) On the whirling and vibration of shaft. Philos Trans R Soc Lond 185:279–
360
2. Malhotra PK, Wenk T, Wieland M (2000) Simple procedure for seismic analysis of liquid
storage tanks. Struct Eng Int
3. Jansen M, Dyke J (2000) Semi-active control strategies for MR dampers: comparative study. J
Eng Mech ASCE 126(8)
4. Iemura H, Igarashi A, Kalantari A (2004) Enhancing dynamic performance of liquid storage
tanks by semi-active controlled dampers. In: 13th world conference on earthquake engineering,
paper no. 773
5. Jin B, Jeon S, Kim S, Kim Y, Chung C (2004) Earthquake response analysis of LNG storage
tank by axisymmetric finite element model and comparison to the results of the simple model.
In: 13th world conference on earthquake engineering, paper no. 394
6. Douglas H, Rotzer J, Maurer H (2005) Hazard and safety investigations for LNG tanks. LNG
J: 23–24
7. Dotoli R, Lisi D, Bardaro D (2007) Sloshing response of LNG storage tank subjected to seismic
loading. In: 6th European LS-DYNA users’ conference
8. Christovasilis P, Whittaker A (2008) Seismic analysis of conventional and isolated LNG tanks
using mechanical analogs. Earthq Eng Res Inst 24(3):599–616
9. Lee K, Kim J, Seo H (2010) Seismic response of LNG storage tank under different base
conditions and liquid height. In: The international society of offshore and polar engineers
(ISOPE), ISBN 978
10. Bharti SD, Dumne SM, Shrimali MK (2010) Seismic response analysis of adjacent buildings
connected with MR dampers. Eng Struct 32:2122–2133
11. Ruifu Z, Dagen W, Xiaosong R (2011) Seismic analysis of a LNG storage tank isolated by a
multiple friction pendulum system. Earthq Eng Eng Vib 10(2):253–262
12. Panchal VR, Soni DP (2013) Seismic behaviour of isolated fluid storage tanks: a-state-of the-art
review. Korean Soc Civ Eng (KSCE) J Civ Eng 18(4):1097–1104
13. Shah DH, Mehta AJ (2017) Fractional delay compensated discrete-time SMC for networked
control system. Digit Commun Netw 3:112–117
14. Shah DH, Mehta AJ (2018) Discrete-time sliding mode control for networked control system.
Springer, Singapore
Distributed Event-Triggered Sliding
Mode Control for Voltage
Synchronization of DC Microgrid Using
Leader–Follower Consensus Protocol

Jatin Savaliya, Keyurkumar Patel and Axaykumar Mehta

Abstract This paper presents a secondary voltage control with the help of a
distributed event-triggered sliding mode control for DC microgrid along with primary
droop control mechanism. The proposed leader-following protocol with event trig-
gered mechanism synchronizes the voltage magnitude of each follower Distributed
Generation units (DGUs) with the leader agent (GDU) in a DC microgrid. The pro-
posed distributed event-triggered sliding mode control reduces information exchange
among distributed generator agents without compromising desired performance.
Finally the effectiveness of the designed control mechanism is validated through
simulation of DC microgrid in MATLAB/Simulink.

Keywords Event-trigger · Multi-agent system (MAS) · Leader–follower

consensus · DC microgrid · Secondary voltage synchronization

1 Introduction

In last few decades, in order to confront the prevailing issues of energy crisis and
global warming, energy industry has encouraged to integrate the distributed gen-
erators (DGs) in microgrid system [2]. Penetration of renewable energy sources in
DC microgrid system proffers high electrical efficiency, reliability, and scalability
[7, 16]. Easy integration of renewable sources due to their DC nature, lower energy
losses due to reduced number of power conversion stages, absence of skin effect,
corona loss, and also reduction the additional exertion of synchronization, control of

J. Savaliya · K. Patel (B) · A. Mehta

Institute of Infrastructure Technology Research and Management (IITRAM),
Ahmedabad, India
e-mail: [email protected]; [email protected]
J. Savaliya
e-mail: [email protected]
A. Mehta
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 25

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_3
26 J. Savaliya et al.

reactive power are the advantages of DC microgrid which make it more prominent
and convenient in practice than AC [4, 12]. Hierarchical control architectures have
been developed to exploit advantages of microgrid [5]. Due to the limitations of the
primary control level of the hierarchical structure, introduction of secondary control
level becomes inevitable which compensate for errors added by the primary level and
maintain the desired voltage level of the system [5, 13]. Tertiary control schedules
economic power exchange between microgrid and main-grid [1, 8].
These control layers communicate through centralized or decentralized communi-
cation topologies for desired operations [6, 9]. The centralized control strategies are
prone to the single point of failure and consequently controller issues [17]. Moreover,
it requires complex communication network which costs uneconomical infrastruc-
ture [17]. In contrast to the centralized architecture, distributed schemes exchange
the information of local measurement with only neighboring units which reduces
the dependency of the control schemes on communication network, infrastructure
cost, channel bandwidth, and consequently enhance reliability of system in case
of communication failure [3, 10]. In practice, each unit is allocated with limited
bandwidth for real time transmission of data packets, a novel scheme to reduce the
data exchange from sensors to controller without compromising system performance
would be more applaudable.
Many approaches has been adopted by many researchers to reduce communication
traffic in system network. An aperiodic sampling concept in network control system
(NCS) is proposed in [15, 18] which transmit the information only in occurrence of
disturbances. Compared to the periodic sampling, aperiodic sampling requires lower
rate to update controller which results in low bandwidth requirement. However,
between two consecutive triggering events, signal information is being held by zero
order hold which can be resembled as open loop operation of the system may become
critical for the system performance.
In this paper, a model-based event-triggered secondary control has been adopted
for multi-agent DC microgrid. Multi-agent leader-following consensus protocol with
distributed robust sliding mode control structure adds the feature of cooperation
among each agent of DC microgrid which is novel approach compared to the average
voltage regulation as per the authors’ best knowledge. In addition to communication
less primary control, proposed reaching law for sliding mode control provides fast,
scalable, and robust voltage stabilization. Designed architecture is distributed type
with model-based event-triggered mechanism which makes the system less depen-
dent on communicated information and reduces data packet transmission in channels.
Moreover, global stability of Multi-agent microgrid system is also proved.

1.1 Paper Organization

The remaining paper is organized as follows: Sect. 2 formulates the problem associ-
ated with DC microgrid due to primary droop control, with the help of preliminaries of
graph theory for multi-agent DC microgrid system. Designed procedure and stability
Distributed Event-Triggered Sliding Mode Control … 27

of distributed event-triggered sliding mode controller is briefed in Sect. 3. Section 4

validates the system performance through simulation results and paper is concluded
in Sect. 5

2 Preliminaries of Graph Theory and Problem Formulation

The interaction among the agents can be represented by undirected connected graph
G = (V, E, A), where V is the non-empty set of nodes or vertices, E ⊆ V × V rep-
resents the set of undirected edges, and A = (ai j ) is the weighted adjacency matrix
with nonnegative entries. An edge ei j is defined by an unordered pair of nodes
(μi , μ j ) and ei j ∈ E if and only if ai j > 0. The Laplacian matrix L = (li j ) N ×N of
the graph G can be defined as L = D − A. Undirected graphs always has symmetric
Laplacian matrix from definition of adjacency matrix. The Laplacian matrix L has
raw sum equals to zero so that all graphs have the first eigenvalue at zero. Further,
D = diag(d1 , d2 , . . . , d N ) is known as diagonal degree matrix. Degree matrix pro-
vides information of the number of edges connected to each node. Degree also shows
the number of neighbors of a vertex. The element of degree matrix is presented by
as D = [di j ], where i and j are the vertices of graph.
Furthermore, digraph g describes interconnection topology access to the refer-
ence. Pinning gain matrix is diagonal matrix G = diag {g1 , . . . , gn }, defined as a

gi > 0 if Vr e f is pinned to DG i
G = diag (gi ) : (1)
gi = 0 otherwise

Generally, reference is defined at node 0. Therefore, the edge (0, i) ∈ E exists with
the weight gi > 0. Here, gi is known as pinning gain. If gi > 0, means node i is
pinned and reference is available to it.

2.1 Multi-agent DC Microgrid

DC microgrid consists of N parallel-connected DGs to the common load through

connecting cable is shown in Fig. 1. Each DG is represented by a DC voltage source
controlled by the Buck converter, where DC voltage source represents renewable
energy source supplying to the constant power load connected at the point of common
coupling (PCC). Here, in microgrid model load is considered as disturbance.
For the design of distributed secondary controller, the concept of multi-agent
system (MAS) is implemented for DC microgrid where each DG acts as the agent and
exchange the information with neighboring agents. The communication of whole DC
microgrid network can be represented by undirected graph G = (V, E). Nonempty
set V = {1, 2, . . . , N } indicates DGU nodes. Due to undirected graph the flow of
information among DGU nodes is bidirectional. Inspiring from [14] linear dynamics
28 J. Savaliya et al.

LOAD

DG1 DG2 DG3 DGN

Fig. 1 Multi-agent DC microgrid architecture

of ith DG agent for single bus DC microgrid is given by:

ẋi = Ai xi + Bi u i + Mi w (2)

In above Eq. (2), Ai , Bi , and Mi are the system matrices which can be written as:
⎡ R 1 ⎤ ⎡ ⎤ ⎡ ⎤
di
− − 1 0
⎢ L di L di ⎥ −
⎣ L di ⎦, and M1 = ⎣ PC P L ⎦.
A1 = ⎣ 1 1 ⎦, B1 = −
− 0 Rci
Ci Rci

2.2 Primary Droop Control

In DC microgrid, power sharing for load demand and power generation mismatch
is achieved by primary droop controller which mimics the operation of parallel
synchronous generator. Primary droop control determines the output current and
power of each converter in microgrid based on locally available information of DG.
Power in DC microgrid is proportional to the current, therefore, the implementation
of droop is achieved by both voltage and current. Droop gain is decided by rated
power and voltage. Droop control is formulated as follows :

vi = vui − di i i (3)

where Vdi , Ii , and di are output voltage, current, and droop gain for the ith DG. V0i is
the nominal output voltage for the DG which is modified by the secondary controller
Distributed Event-Triggered Sliding Mode Control … 29

to achieve desired reference. Droop generated output voltage vi is fed to the inner
voltage and current control loop which calculates the duty cycle for DC–DC buck
converter switch. Droop provides approximate power sharing and introduces voltage
offsets. Due to this trade-off between voltage regulation and current sharing, choice
of droop constant becomes important. The following equation decides the droop
constant:
Δv
di = max (4)
ii

where Δv is the allowable voltage deviation in microgrid and i imax is the maximum
current output of ith DG. Limitation of primary controller motivates to design sec-
ondary controller to restore the voltage.

3 Voltage Restoration with Distributed Secondary

Event-Triggered Mechanism

Event-trigger mechanism reduces the communication cost and energy between differ-
ent agents, a distributed event-triggered sliding mode control with leader-following
consensus protocol is proposed in this section. The proposed distributed controller
not only synchronizes the voltage magnitude of all DGs to the desired value Vr e f but
also excludes Zeno behavior. In leader–follower consensus ensures the accurate tra-
jectory tracking of leader agent by other agents in finite time. With the help of local
communication, the tracking consensus aims to maintain the coordination among
leader and followers and follower update their states with leader’s states. However,
leader can only try to follow only given reference without being influenced by other
agents.

3.1 Consensus Algorithm and Controller Synthesis

In DC microgrid, the secondary voltage controller chooses appropriate input Viu for
the droop characteristic in Eq. (3). This droop characteristic for multi-agent system
can be written as:
Vd = Vu − D I (5)

where Vd = [v1 , v2 , . . . , v N ]T , Vu = [vu1 , vu2 , . . . , vu N ]T are the vectors, and D =

diag(d1 , d2 , . . . , d N ) matrix.
Periodic sampling can be replaced by event-triggered-based sampling, in which
next sample instant dependants on trigger of event. Hence, the control law is derived
for ∀t ∈ t k , t k+1 for event-triggered mechanism.
Figure 2 shows distributed secondary control mechanism. Considering the effec-
tiveness of event-triggered control algorithm, voltage tracking error for ith agent can
be defined as :
30 J. Savaliya et al.

Microgrid Bus
LC Filter
DG2
DG3
Vi V2
V3

ri Communication Network
PWM
Vn
V4
Inner Current Inner Voltage
Controller Controller
DGn DG4

Distributed Secondary
Controller

Fig. 2 Distributed secondary control algorithm for DC microgrid

vi (t) = ai j vi (tkii ) − v j (tki j ) + gi vi (tkii ) − vr e f (6)

j∈Ni

where ai j , gi are the elements of adjacent and pinning gain matrix, whereas tk j is the
latest triggering instant of the jth DG at the instant t.
Event-triggered mechanism introduces errors due to discretization of sensor
signals through event-triggered mechanism. Now defining di (t) = vi (t) − vr e f ,
d̂i (t) = vi (tki ) − vr e f , t ∈ tkii , tki+1
i
and measurement error with respect to volt-
age output of ith DG.

ei (t) = vi (t) − vi (tkii ) (7)

The consensus tracking error in (6) can be modified as:

vi (t) = ai j (vi (t) − v j (t)) − (ei (t) − e j (t)) + gi dˆi (t) (8)
j∈Ni

The error candidate with the help of graph theory in global form can be represented
by:
εv (t) = (L + G)(Vd (t) − (ei (t) − e j (t)) − 1 N ⊗ d̂(t)) (9)

where εv (t) = [v1 , v2 , . . . vn ]T , d̂ = 1 N ⊗ d̂i (t), ei (t), and e j (t) are error vectors.
Here, ⊗ and 1 N define kronecker product and vector of ones of length N . Taking
(L + G) = H and 1 N ⊗ d̂ = V0 , the above Eq. (9) can be modified as:

εv (t) = H(Vd (t) − (ei (t) − e j (t)) − V0 (t)) (10)

Distributed Event-Triggered Sliding Mode Control … 31

Let the sliding surface for the ith agent to design robust sliding mode controller
is given by 
σi (t) = vi (t) + λi vi (t)dt (11)

where λi is the adjustable gain which can be tuned as per needs of design. From (11)
sliding manifold for the global system is written as:

σ(t) = εv (t) + λ εv (t)dt (12)

Taking In ⊗ λi = λ and with the help of (9), (12) further can be expanded as follows:

σ(t) = εv (t) + λ H(Vd (t) − (ei (t) − e j (t)) − V0 (t))dt (13)

From droop characteristic

σ̇(t) = ε̇v (t) + λH(Vu (t) − D I (t) − (ei (t) − e j (t)) − V0 (t)) (14)

Using reaching law in [11], the consensus reaching law for ith agent is defined as

σ̇i (t) = −qi σi (t) − ki sgn(σi (t)) (15)

Global reaching law for ki instant in global form:

σ̇(t) = −Qσ(tkii ) − K sgn(σ(tkii )) (16)

Comparing (14) and (16), consensus protocol for voltage synchronization in micro-
grid is written as

Vu (t) = λ−1 H−1 (−Qσ(tki

i ) − K sgn(σ(t i )) − ε̇ (t)) + D I (t) + (e (t) − e (t)) + V (t)
ki v i j 0
(17)

3.2 Stability of DC Microgrid System

Let consider Lypunov candidate such that

1 T
V = σ (t)σ(t) (18)
2

Taking the time derivative of this candidate given in (18) for t ∈ t k , t k+1 ] yields
32 J. Savaliya et al.

V̇ = σ T (t)σ̇(t) (19)
= σ (t)(ε̇v (t) + λH(Vu (t) − D I (t) − v0 )
T

= σ T (t)(−ε̇v (t) + λH(λ−1 H−1 (Qσ(tkii ) − K sgn(σ(tkii )) − ε̇v (t))

+D I (t) + (ei (t) − e j (t)) + V0 (t) − D I (t) − V0 (t)) (20)
≤σ T
(t)(−Q|σ(tkii )|) −K sgn(σ(tkii )) + ||ei (t) − e j (t)||)

where sign (σi (t)) = sign σi t k is full-filled for ∀t ∈ tkii , t ki+1 when σi (t) > 0
or σi (t) < 0

V̇ ≤ −||σ(t)||(Q|σ(tkii )|) + K sgn(σ(tkii )) − ||ei (t) − e j (t)||) (21)

⇒ V̇ ≤ −ζ σi (t) (22)

where ζ = Q|σ(tkii )|) + K sgn(σ(tkii )) − ||ei (t) − e j (t)|| is positive definite as the
Q and K are the positive definite.

4 Results and Discussion

In order to show the effectiveness of designed event-triggered concept, a DC micro-

grid consists of four parallel-connected DGs to common load is shown in Fig. 3. The
parameters of all the DGs and transmission lines are kept same which are listed in
Table 1. The required bidirectional communication for distributed control approach
is shown through the communication topology in which DG1 is considered as leader
and remaining three DGs act as followers as shown in Fig. 4.

Fig. 3 Simulated microgrid

configuration DG 1

DG 2

L
DG 3 O
A
D
DG 4
Distributed Event-Triggered Sliding Mode Control … 33

Table 1 DG parameters of Parameters Symbol Values

microgrid
Converter parameters
DC power supply Vi 600 V
Converter capacitance Ci 4000 µF
Converter inductance Li 100 µH
Switch+ inductor resistance ri 0.025 
Switching frequency f 10 KHz
Line parameters
Line Inductance L ci 1.8e-3
Line Resistance Rci 0.05
Droop control parameters
Droop gain di 0.04

Fig. 4 Microgrid Leader 0 Follower 1

communication topology
DG 1 DG 2

DG 3 DG 4

Follower 2 Follower 3

The simulated DC microgrid supplies the constant power load of 20 kW and

reference value of 400 V for voltage magnitude synchronization of all DGs is pinned
to the DG1. It can be observed from Fig. 5a that distributed event-triggered controller
synchronize the voltage level. and it can be observed that load increments are applied
at t = 1.5 s and t = 2.5 s of 4 kW and 10 kW, respectively, which can be observed
in Fig. 5b. The broadcast periods of all four DGs are shown in Fig. 6 gives event-
triggered time instant and release intervals of all agents.

5 Conclusion

The event triggered mechanism with the proposed distributed secondary con- troller
for DC microgrid fulfilled the idea of voltage restoration and synchronize with
reduced communication. It makes the control action aperiodic which help to reduce
the control effect and minimal usage of energy without compromising system per-
formance.
34 J. Savaliya et al.

(a) 420
410

400
Voltage (V)

390

380

370
V1 V2 V3 V4
360
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (Sec.)

(b) 10

8
Power (kW)

P P P P
1 2 3 4
2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (Sec.)

Fig. 5 Result of event-trigger control a output voltages, b power sharing

Fig. 6 Broadcast period for a DG1, b DG2, c DG3, and d DG4

Distributed Event-Triggered Sliding Mode Control … 35

References

1. Bidram A, Davoudi A (2012) Hierarchical structure of microgrids control system. IEEE Trans
Smart Grid 3(4):1963–1976. https://doi.org/10.1109/tsg.2012.2197425
2. Che L, Shahidehpour M, Alabdulwahab A, Al-Turki Y (2015) Hierarchical coordination of a
community microgrid with AC and DC microgrids. IEEE Trans Smart Grid 6(6):3042–3051.
https://doi.org/10.1109/tsg.2015.2398853
3. Ensermu G, Bhattacharya A, Panigrahy N (2019) Real-time simulation of smart DC microgrid
with decentralized control system under source disturbances. Arab J Sci Eng. https://doi.org/
10.1007/s13369-019-03842-6
4. Fregosi D, Ravula S, Brhlik D, Saussele J, Frank S, Bonnema E, Scheib J, Wilson E: A com-
parative study of DC and AC microgrids in commercial buildings across different climates and
operating profiles. In: 2015 IEEE first international conference on DC Microgrids (ICDCM).
IEEE (2015). https://doi.org/10.1109/icdcm.2015.7152031
5. Guerrero JM, Chandorkar M, Lee TL, Loh PC (2013) Advanced control architectures for
intelligent microgrids—part i: Decentralized and hierarchical control. IEEE Trans Ind Electro
60(4):1254–1262. https://doi.org/10.1109/tie.2012.2194969
6. Hatziargyriou N, Dimeas A, Tsikalakis A (2005) Centralized and decentralized control of
microgrids. Int J Distrib Energy Resour 1:197–212
7. Hirsch A, Parag Y, Guerrero J (2018) Microgrids: a review of technologies, key drivers, and out-
standing issues. Renew Sustain Energy Rev 90:402–411. https://doi.org/10.1016/j.rser.2018.
03.040
8. Hu HY, Peng YG, Xia YH, Wang XM, Wei W, Yu M (2017) Hierarchical control for parallel
bidirectional power converters of a grid-connected DC microgrid. Front Inf Technol Electron
Eng 18(12):2046–2057. https://doi.org/10.1631/fitee.1601497
9. Hwang CS, Kim ES, Kim YS (2016) A decentralized control method for distributed generations
in an islanded DC microgrid considering voltage drop compensation and durable state of charge.
Energies 9(12):1070. https://doi.org/10.3390/en9121070
10. Liu W, Gu W, Xu Y, Xue S, Chen M, Zhao B, Fan M (2015) Improved average consensus
algorithm based distributed cost optimization for loading shedding of autonomous microgrids.
Int J Electr Power Energy Syst 73:89–96. https://doi.org/10.1016/j.ijepes.2015.04.006
11. Mehta A, Bandyopadhyay B (2015) Preliminaries of sliding mode control. In: SpringerBriefs
in applied sciences and technology, pp. 9–25. Springer, India. https://doi.org/10.1007/978-81-
322-2238-5_2
12. Peyghami S, Mokhtari H, Blaabjerg F (2017) Hierarchical power sharing control in DC
microgrids. In: Microgrid, pp. 63–100. Elsevier. https://doi.org/10.1016/b978-0-08-101753-
1.00003-6
13. Shafiee Q, Guerrero JM, Vasquez JC (2014) Distributed secondary control for islanded
microgrids—a novel approach. IEEE Trans Power Electron 29(2):1018–1031. https://doi.org/
10.1109/tpel.2013.2259506
14. Tucci M, Riverso S, Vasquez JC, Guerrero JM, Ferrari-Trecate G (2016) A decentralized scal-
able approach to voltage control of dc islanded microgrids. IEEE Trans Control Syst Technol
24(6):1965–1979. https://doi.org/10.1109/TCST.2016.2525001
15. Wang X, Lemmon MD (2011) Event-triggering in distributed networked control systems. IEEE
Trans Autom Control 56(3):586–601. https://doi.org/10.1109/tac.2010.2057951
16. Xie Y, Ning J, Huang Y, Jia J, Jian Z (2013) A review of DC micro-grid protection. In: Advances
in brain inspired cognitive systems, pp. 338–347. Springer, Berlin. https://doi.org/10.1007/978-
3-642-38786-9_38
17. Yazdanian M, Mehrizi-Sani A (2014) Distributed control techniques in microgrids. IEEE Trans
Smart Grid 5(6):2901–2909. https://doi.org/10.1109/tsg.2014.2337838
18. Yue D, Tian E, Han QL (2013) A delay system method for designing event-triggered controllers
of networked control systems. IEEE Trans Autom Control 58(2):475–481. https://doi.org/10.
1109/tac.2012.2206694
A Fractional and Integer Order PID
Controller for Nonlinear System: Two
Non-interacting Conical Tank Process
Case Study

Himanshukumar R. Patel and Vipul A. Shah

Abstract The mathematical model for two non-interacting conical tank level
(TNICTL) process is proposed. The control of TNICTL process is difficult due
to the nonlinearity. Hence, the nonlinear process is linearised around proper operat-
ing points. Then, the fractional-order and integer order PID controller are designed
and the control parameters are tuned using Genetic Algorithm (GA) and Ziegler–
Nichols (ZN) tuning methods. Also, system component (leak) and actuator faults
are introduced into TNICTL and validate the proposed controller tolerance capacity.
The closed-loop controller performances are simulated and compared in terms of
settling time, rise time and integral error criteria. It is found that fractional order
PID controller has better servo and regulatory response than the integer order PID
controller. The liquid inflow rates are optimally manipulated by the fractional-order
PID controller to track the set point variation and to reject the faults effectively than
integer order PID controller.

Keywords Actuator fault · Fractional PID controller · Genetic algorithm ·

Non-interacting process · Nonlinear · System component fault · Ziegler–Nichols

1 Introduction

A primary problem in process/chemical/food industries is control of liquid level and

flow in process tank [1–3]. Conical tanks are extensively used in process indus-
tries, petrochemical industries, food process industries, and wastewater treatment

Supported by Dharmsinh Desai University.

H. R. Patel (B) · V. A. Shah

Department of Instrumentation and Control Engineering, Dharmsinh Desai University, Nadiad
387001, Gujarat, India
e-mail: [email protected]
V. A. Shah
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 37

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_4
38 H. R. Patel and V. A. Shah

industries [4–6]. Control of conical tank is a challenging problem because of its

nonlinearity and constantly changing in area of cross section [7, 8]. Hence for these
reasons, the conical tank process is taken here. Conventional PID controllers are
simple, robust provided the system is linear [9, 10]. But the process considered here
has nonlinear characteristics which are represented as piecewise linearized models
[8]. Multiple linear models of conical tank with many PI controllers were imple-
mented. Many researchers have been carried out in the level control of the conical
tank process [4, 5, 7, 8, 11–13]. To enhance the performance of the conical tank
level control system some advanced fuzzy logic base PID, IMC controller is applied
on different applications [4–8, 11–14]. Also for nonlinear system, Takagi–Sugeno
fuzzy modeling is used to linearized model and control in the operating range of the
system discussed in [15, 16], but the limitation is mathematical model of the system
is required at different operating ranges.
Any feedback control system maintaining the stability and control performance
are the very curtail criterion to get optimum output. To ensure this criterion as we
know, two control scheme presented in classical control theory, state feedback and is
output feedback are two important techniques in system control [17, 18]. Specially,
the PID controller is by far the most dominating form of feedback in use today. Due to
its functional simplicity and performance robustness, the PID controllers are still used
for many industrial applications such as process controls, motor drivers, flight control,
instrumentation, etc [6, 13, 17]. But at the same time, conventional PID controller will
not give optimum performance when malfunction/fault occurs in the system which
causes the performance degradation and system instability [9, 17]. To overcome
this limitation fractional calculus has the potential to accomplish what integer order
calculus cannot [18]. Fractional-order dynamic systems and controllers have been
increasing in interest in many areas of science and engineering in the last few years. In
most cases, our objective of using fractional calculus is to apply the fractional-order
controller to enhance the system control performance. Fractional calculus (FC) is a
mathematical topic with more than 300 years old history, but its application to physics
and engineering has been reported only in recent years. The theory of FC is a well-
adapted tool to the modeling of many physical phenomena, allowing the description
to take into account same peculiarities that classical integer order models simply
neglect [19, 20]. The fractional integrodifferential operators are a generalization of
integration and derivation to noninteger order (fractional) operators. It is remarkable
the increasing number of studies related with the application of fractional controllers
in many areas of science and engineering [18, 21]. The fractional PIλ D μ controller,
[19], the CRONE controllers [22] and the fractional leadlag compensator, [23] are
some of the well-known fractional-order controllers. In some of these works, it
is verified that the fractionalorder controllers can have better disturbance rejection
ratios and less sensitivity to plant parameter variations compared to the traditional
controllers. Some fractional-order PID controller for nonlinear level control system
presented in recent years [6, 13, 14, 24, 25].
In this work, two non-interacting conical tank level (TNICTL) subject to actuator
and system component (leak) faults is addressed from simulated point of view. To
this extent, a nonlinear model representing the plant behavior in an accurate way
A Fractional and Integer Order PID Controller for Nonlinear System … 39

is first developed. Based on this model representation of the plant, a classic integer
order PID control strategy is developed whose parameters are tuned using the Genetic
Algorithm (GA) and Ziegler and Nichols (ZN) method and considering a linearized
model at three different operating points, similar to what was done in some prelim-
inary works, both, from simulation and experimental [6] viewpoints. Then, integer
order PID control strategies and their fractional-order versions, fractional-order PID
(FOPID) is designed, where the controller parameters are tuned using ZN, and GA.
This paper organized as follows. The mathematical model for proposed TNICTL
process is briefly explained in Sect. 2. The basics of fractional-order system intro-
duced in Sect. 3. The fractional and integer order PID controller design procedure
using GA and ZN is given in Sect. 4. The simulation results of controller are analyzed
in Sect. 5. Finally, the conclusion of proposed work is highlighted in Sect. 6.

2 Process Description

In TNICTL, the process contains two identical conical tanks in the series manner
shown in Fig. 1 [4, 5]; the one identical pumps deliver the liquid inflow Fin and one
outlet flow Fout in tank 1, the outlet flow Fout of tank 1 is inlet flow off to the tank
2 and outlet flow is Fo , respectively, through the two control valves V1 and V2 . The
two tanks under consideration should have the same cross-sectional area. The main
objective in mathematical modeling is to find the transfer function for the system
that is a relation between output variable h 2 (s) (liquid height in Tank 2) and input
variable Fin (s) (Inlet flow to tank 1). In this case, the manipulated variable is Fin (s)
and controlled variable is h 2 (s).

Fig. 1 Prototype structure

of two-tank canonical
non-interacting level control
system (TTCNILCS) [4, 5]
40 H. R. Patel and V. A. Shah

2.1 Mathematical Modeling of TNICTL

Prototype structure of TNICTL is shown in Fig. 1 [4], where the inlet flow rate
(Fin ) which is manipulated variable to control the controlled variable tank 2 height
(h2 ) via proposed hybrid controller. In the given system two control valve (V1 ) and
(V2 ) provides constant flow rate, and third control valve (C V1 ) manipulating (Fin )
and archiving the controlled variable (h2 ) via proposed controller. Parameters of
prototype structure are mentioned in Table 1.
Mathematical modeling of the liquid tank system is derived using the total mass
balance equation. According to that principle described in [4, 26],
Rate of Accumulation = inflow outflow
The mathematical model is derived for both the conical tanks separately as follows.
From Fig. 2. Single conical tank system is given and can be written as following
equation:
A = πr 2 (1)

Table 1 Physical parameters of TNICTL [4]

Parameter Symbol Value
Total height of the tank H 60 cm
Top radius of the Tank R 15 cm
M V1 valve coefficient β1 5 cm2 /s
M V2 valve coefficient β2 2 cm2 /s
Gravitational constant g 9.82 m/s2
Actuator fault coefficient α1 0–10 %
System component fault F1 0–10 %

Fig. 2 Single conical tank

[4]
A Fractional and Integer Order PID Controller for Nonlinear System … 41

From Fig. 1.
R
tan θ = (2)
H
At any height (h 1 ) of Tank 1
r R
tan θ = = (3)
h1 H

Simplified mass balance equation for tank 1 of TNICTL can be written by (1) as
per [4, 27],  
dA(h1 ) √
dh1 Fin − 1
3
h1 dt
− β 1 2gh 1
=   (4)
dt h2
1
3
π R 2 H12

Area of the canonical Tank 1 at height (h 1 ) and Tank 2 at any height (h 2 )

π R 2 h12 π R 2 h22
A1 = 2
, A2 = (5)
H H2
Similarly, for Tank 2 the mass balance equation is given by following relation [4]
[27],  √ 
2)
√
dh2 β1 2gh1 − 13 h2 dA(hdt
− β2 2gh2
=  2
 (6)
dt 1
π R 2 h2
3 H 2

where outlet flow of Tank 1 and Tank 2 given by

 
Fout = β1 2gh1 , Fo = β2 2gh2 (7)

Now, mathematical model of TNICTL with actuator and system component (leak)
faults given as follows [4]:
 √ 
dA(h1 )
dh1 α1 Fin − 1
3
h1 dt
− β1 2gh1
=  2
 (8)
dt 1
π R 2 h1
3 H2

 √ √ 
dA(h2 )
dh2 α1 β1 2gh1 − 1
3
h2 dt
− β2 2gh2 − F1
=   (9)
dt h2
1
3
π R 2 H22

where α1 denotes actuator fault and F1 denotes system component (leak) fault.
For the simplification take both the tank identical in size so area of both the tank
is same.
Combining (1) and (2) would give the nonlinear transfer function of TNICTL (6)
42 H. R. Patel and V. A. Shah

K1 K1
GP = e−td s (10)
(τ1 s + 1) (τ2 s + 1)

By considering the prototype structure parameters, the transfer function of the

proposed system is expressed as follows:
First, a fractional-order LTI model is identified from the step experiment. The
identification data is recorded from the real plant and is transformed such that both
input and output signals are in range u, y ∈ [0, 1]. The obtained model is as follows:

1.2393
G(s) = (11)
40038.51s 2 + 1135.1s + 1

It can be seen, that this model tends to an integer order FOPDT model. The latter
can be approximated using iopid_tune tool [29] in FOMCON as

4.4589 −0.208s
G(s) = e (12)
61.51s + 1

3 Fractional-Order Systems

Fractional calculus is a generalization of ordinary calculus. The main idea is to

develop a functional operator D, associated to an order n not restricted to integer
numbers, that generalizes the usual notions of derivatives (for a positive n) and
integrals (for a negative n). Just as there are several alternative definitions of (usual,
integer) integrals (due to Riemann, Lebesgue, Steltjes, etc.), so there are several
alternative definitions of fractional derivatives that are not exactly equivalent. The
most usual definition is due to Riemann and Liouville and generalizes two equalities
easily proved for integer orders [20]:

Dx−n
x
(x − t)n−1
c f (x) = f (t)dt, n ∈ N, (13)
c (n − 1)!

D n D m f (x) = D n+m f (x), m ∈ Z−

0 ∨ n, m ∈ N0 . (14)

The full definition of D becomes

⎧ x (x−ζ )−ν−1
⎨ c Γ (−ν)
f (ζ )dζ if ν < 0,
ν
c Dx f (x) = f (x) if ν = 0, (15)
⎩ ν−n
D n [c Dx f (x)] if ν > 0,

where n = mink∈ N : k > ν

It is worth noticing that, when ν is positive but noninteger, operator D still needs
integration limits c and x; in other words, D is a local operator for natural values of
A Fractional and Integer Order PID Controller for Nonlinear System … 43

ν (usual derivatives) only. The Laplace transform of D follows rules rather similar
to the usual ones:

ν S ν F(s) if ν ≤ 0,
L[0 Dx f (x)] = ν k Dxν−k−1 (16)
S F(s) − n−1
k=0 S 0 f (0) if n − 1 < ν < n ∈ N.

This means that, if zero initial conditions are assumed, systems with a dynamic
behavior described by differential equations involving fractional derivatives give rise
to transfer functions with fractional powers of s.
Even though n may assume both rational and irrational values in (4), the names
fractional calculus and fractional-order systems are commonly used for purely his-
torical reasons. Some authors replace fractional with noninteger or generalized, how-
ever. Thorough expositions of these subjects may be found in [19, 20, 28].
The control law of the PIλ D μ controller, where the integral component is of order
λ and the derivative component is of order μ, has the following form in the time
domain
μ
u(t) = K p · e(t) + K i · i tλ e(t) + K d · Dt e(t) (17)

where e(t) is the error signal and the fractional-order integral is defined as i tλ x(t) =
μ
Dt x(t). In the Laplace domain, the transfer function corresponding to the parallel
form of the fractional-order PID controller is the following:

Ki
C(s) = K p + + Kd · Sμ (18)
Sλ
It can be easily seen, that in the frequency domain this controller has obvious
advantages over the classical one. By definition, s α = ( jw)α and thus more freedom
in changing the shape of the response is achieved. This is a very important controller
design method that is frequently used.

4 Fractional and Integer Order Controllers Tuning

4.1 Tuning Using Genetic Algorithm

Genetic algorithm is a random search method that can be used to solve nonlinear sys-
tem of equations and optimize complex problems. GA uses probabilistic transition
rules instead of deterministic rules and handles a population of potential solutions
known as individuals or chromosomes that evolve iteratively. Each iteration of the
algorithm is termed a generation. The evolution of solutions is simulated through a
fitness function and genetic operators such as reproduction, crossover, and mutation
[30]. Genetic algorithm as illustrated in Fig. 3 is typically initialized with a random
population. This population (mating pool) is usually represented by a real-valued
number or a binary string called a chromosome. The performance of the individual
44 H. R. Patel and V. A. Shah

Fig. 3 Flow diagram of

genetic algorithm

is measured and assessed by the objective function, which assigns each individ-
ual a corresponding number called its fitness. The fitness of each chromosome is
assessed and a survival of the fittest strategy is applied. In this work, the error value
is used to assess the fitness of each chromosome. There are three main operations
in a genetic algorithm: reproduction, crossover, and mutation. The sequences of
operations involved in GA are described in Fig. 3.
Genetic Algorithm Steps
Step 1. Initialize the parameter with a population of random solutions, such as
crossover rate, mutation rate, number of clusters, and number of generations. Deter-
mine the coding mode.
Step 2. Compute and evaluate the value of the fitness function.
Step 3. Proceed with crossover and mutation operation and make up the new cluster.
Step 4. Repeat step 2, till the best value is obtained.
According to control objectives, five parameters, K p , K i , K d , λ and μ of a frac-
tional PID controller are required to be designed in these settings. In this study, it is
introduced next optimality criterion which involves besides steady state error e, i.e
IAE, Peak overshoot M p , as well as settling time Ts .

 
J =  M p  + Ts + |e| dt → min (19)

Fitness function is designed as:

f g = Jmax + Jmin − Jg (20)

A Fractional and Integer Order PID Controller for Nonlinear System … 45

Table 2 Parameters used in genetic algorithm

Parameter Type/value
Maximum generations 250
Population size N = 100
Crossover probability Pc = 0.75
Mutation probability Pm0 = 0.1
Initial mutation probability l = 25
Generation gap gr = 0.35
Encoding Binary
Selection Uniform
Crossover Single point crossover
Mutation Uniform

Table 3 Controller parameters values tuned by GA

Controller Kp Ki Kd λ μ
PID 0.9628 0.1369 0.9247 – –
FOPID 0.8553 0.1458 0.5687 0.9394 0.3466

where are Jmax , Jmin the largest value and the smallest value of J , respectively,
observed thus far, as well as Jg value of criterion for the current population
(Tables 2 and 3).

4.2 Ziegler–Nichols Method

The problem of tuning parameters of PID controllers is when the parameters defining
the dynamics of the system are unknown. To deal with this problem, several strategies
have been developed. The most common is the Ziegler and Nichols tuning method,
which has also been extended to the case of tuning FOPID controllers. There are
two variants of the method, which are based on a particular dynamic behavior of
the plant under specific conditions. This information can be obtained either from the
mathematical model and/or experimentally.
Integer Order PID Controller Tuning The step response of the open-loop TF is
first obtained and this method is based on the assumption that the plant can be suitably
represented by a TF of first order plus a time delay (FOTFPTD), described as

K
G(s) = e−τd s (21)
1 + Ts
46 H. R. Patel and V. A. Shah

Fig. 4 S-shaped unit-step response [16]

Table 4 Tuning rules for ZN step response method for IOPID

Controller Kp Ti Td
P T/Td ∞ 0
PI 0.9*(T/Td ) (Td /0.3) 0
PID 1.2*(T/Td ) 2*Td 0.5*Td

The system response when it is excited with a step input is assumed to have a
similar response as shown in Fig. 4.
From the information contained in Fig. 4, parameters K, τd and T are obtained.
Then the parameters of the PID controller can be determined as indicated in Table 4.
Fractional-Order PID Controller Tuning
Tuning by Minimisation
Monje et al. [31] proposed that fractional PIDs be tuned by requiring them to
satisfy the following five conditions (C being the controller and G the plant):
(1) The gain-crossover frequency wcg is to have some specified value:
 
C( jw cg )G( jw cg ) = 0 d B (22)

(2) The phase margin φm is to have some specified value:

− π + φm = arg[C( jw cg )G( jwcg )] (23)

(3) So as to reject high-frequency noise, the closed-loop transfer function must

have a small magnitude at high frequencies; thus it is required that at some specified
frequency oh its magnitude be less than some specified gain:
 
 C( jwcg )G( jwcg ) 
 <H (24)
 1 + C( jw )G( jw )
cg cg
A Fractional and Integer Order PID Controller for Nonlinear System … 47

(4) So as to reject output disturbances and closely follow references, the sensitivity
function must have a small magnitude at low frequencies; thus it is required that at
some specified frequency w1 its magnitude be less than some specified gain:
 
 1 
 
 1 + C( jw )G( jw )  < N (25)
cg cg

(5) So as to be robust in face of gain variations of the plant, the phase of the open-
loop transfer function must be (at least roughly) constant around the gain-crossover
frequency: 
d
dw
arg[C( jwcg )G( jwcg )] wwcg = 0 (26)

Conditions are five because five are the parameters to tune. To satisfy them all the
authors proposed the use of numerical optimisation algorithms, namely of Nelder-
Meads simplex method as implemented in Matlabs function fmincon (the condition
in (20) is assumed as the condition to minimize; conditions in (21)–(24) are assumed
as constraints). This is effective but allows local minima to be obtained.
The minimisation tuning method presented above in subsection was applied to
plants given by (19) for several values of L and T , with K = 1. The parameters of
fractional PIDs thus obtained vary in a regular manner with L and T . Using the least-
squares method, it is possible to translate that regularity into polynomial formulas to
find acceptable values of the parameters from the values of L and T .
A set of rules is given in Table 5 for the conditions given in Eq. 25. This set of
rules may be applied if
0.1 ≤ T ≤ 50 and L ≤ 0.5 (27)

Only one set of parameters is needed in this case because the range of values of
L with which these rules cope with is more reduced. They were designed for the
following specifications:

Table 5 Parameters for the set of tuning rules [21]

Controller P I D λ μ
1 1.0574 0.6014 0.8793 1.1851 0.2778
L 24.5420 0.4025 15.0846 0.3464 2.1522
T 0.3544 0.7921 0.0771 0.0492 0.0675
L2 46.7325 0.4508 28.0388 1.7317 2.4387
T2 0.0021 0.0018 0.0000 0.0006 0.0013
LT −0.3106 −1.2050 1.6711 0.0380 0.0021
48 H. R. Patel and V. A. Shah

Table 6 Controller parameters values tuned by ZN

Controller Kp Ki Kd λ μ
PID 354.861 147.622 36.905 – –
FOPID 1.143 3.471 6.894 0.763 0.189

Fig. 5 Proposed control scheme using GA

wcg = 0.5 rad/s

φm = 1 rad ≈ 57◦
wh = 10 rad/s
(28)
w1 = 0.01 rad/s
H = −20 dB
N = −20 dB

Of course, sets of rules other than those two above might have been found in the
same way as these were for different sets of specifications (Table 6).
The control scheme of FOPID/IOPID using GA is depicted in Fig. 5.

5 Simulation Results

Simulation using genetic algorithm-based FOPID controller for a TNICTL system,

including performance indices such as ITAE separately, is carried out for both servo
and regulatory cases. In all simulation runs, the TNICTL system is simulated using
the nonlinear first principles model given by (4) and (6). First regulatory and servo
response of FOPID using GA without fault compared with other controller FOPID
using ZN, PID using GA, and PID using ZN. The variation in the controller outputs is
presented in Figs. 6 and 7. The performance measures like percentage settling time Ts ,
rise time Tr and peak overshoot M p are calculated for various controllers as shown
in Table 7. Further, it is evident that the ITAE indices are compared and clearly
presented in Table 8, which clearly says that FOPOID using GA is superior then
A Fractional and Integer Order PID Controller for Nonlinear System … 49

Fig. 6 Regulatory response without faults

Fig. 7 Servo response without faults

Table 7 Comparison of specifications for different controllers

System FOPID by GA FOPID by ZN PID by GA PID by ZN
specification
Ts (s) 14.31 16.07 17.01 20.63
Tr (s) 8.9 7.96 10.94 9.89
M p (%) 0 0.54 0.11 0.88

other compared controllers. Also FOPID using GA found to be smoother response

as compared to all other controllers (refer to Figs. 6 and 7).
Simulation studies have been carried out to demonstrate the fault tolerance capa-
bility of the proposed controller on TNICTL process at two type of faults with two
different nature (i.e. abrupt and incipient). The system component (leak) fault in the
bottom of tank 2 with magnitude (from 0 to 10 %) has been introduced at the 50th
50 H. R. Patel and V. A. Shah

Table 8 ITAE value Comparison for Controllers

Controller FOPID by GA FOPID by ZN PID by GA PID by ZN
Regulatory 6.137 10.601 22.984 40.423
response
Servo response 8.102 14.218 30.282 46.248

Fig. 8 ITAE performance comparisons for regulatory and servo responses of TNICTL without
fault

Fig. 9 Regulatory response with abrupt system component fault

sampling instant and maintained at up to 100th sampling instant (refer to Figs. 8 and
9). At the instance of leak fault introduce in to TNICTL the FOPID using GA con-
troller is able to quickly tolerate the faults as compared to other described controller
(refer to Figs. 8 and 9). The ITAE error caparison also validate the same which is
presented in Table 9.
A Fractional and Integer Order PID Controller for Nonlinear System … 51

Table 9 ITAE value comparison for Controllers subject to system component fault
Process with leak FOPID by GA FOPID by ZN PID by GA PID by ZN
fault
Abrupt 7.792 12.945 25.104 45.136
Incipient 8.029 13.918 28.142 49.084

Fig. 10 Regulatory response with incipient system component fault

Fig. 11 Regulatory response with abrupt actuator fault

The proposed controller is tests on TNICTL process subject to actuator fault in

two different nature (Actuator 1 f 1 fault, refer Fig. 1). Figures. 10, 11, and 12 give
similar observation. The proposed controller will give superior response as compared
to other presented controllers, which is an evident by ITAE error indices comparison
shown in Table 10.
The fault tolerance capability is calculated in terms of fault recover time Fr t , the
proposed controller FOPID using GA is compared with other controller presented
52 H. R. Patel and V. A. Shah

Fig. 12 Regulatory response with incipient actuator fault

Table 10 ITAE value comparison for Controllers subject to actuator fault

Process with FOPID by GA FOPID by ZN PID by GA PID by ZN
actuator fault
Abrupt 8.216 13.270 27.802 49.017
Incipient 8.982 14.539 30.360 53.248

Table 11 Fault recovery time from system component faults

Process with leak FOPID by GA (s) FOPID by ZN (s) PID by GA (s) PID by ZN (s)
fault
Abrupt 7.72 8.12 11.94 11.06
Incipient 27.93 29.19 32.43 44.57

Table 12 Fault recovery time from actuator faults

Process with FOPID by GA (s) FOPID by ZN (s) PID by GA (s) PID by ZN (s)
actuator fault
Abrupt 11.99 13.68 19.53 20.49
Incipient 33.23 34.14 43.40 43.62

in Tables 11 and 12. The fault recovery time of FOPID using GA is superior in both
the fault case and two types of fault nature (Figs. 13 and 14).
A Fractional and Integer Order PID Controller for Nonlinear System … 53

Fig. 13 ITAE performance comparison a TNICTL system with system component (leak) faults
and b TNICTL system with actuator faults

Fig. 14 Fault recover time T f r comparison a fault recover time from system component (leak)
fault and b fault recover time from actuator fault

6 Conclusion

In this paper, GA-based FOPID/IOPID controller tuning has been designed and
applied to TNICTL process subject to actuator and system component faults. From
the simulation studies, we infer that optimized controller parameters obtained by
implementing this algorithm with a IAE, settling time and peak overshoot as cost
indices have achieved satisfactory set point tracking and fault tolerance in the oper-
ating range of the TNICTL process. The same magnitude of an actuator fault is more
critical then system component faults, and incipient nature fault will degrade severely
then abrupt nature of leak and actuator faults.

References

1. Himanshukumar R, Vipul S (2018) Fault tolerant control systems: a passive approaches for
single tank level control system. J Instrum Control Eng 6(1):11–18. https://doi.org/10.26634/
jic.6.1.13934
2. Himanshukumar R, Vipul S (2018) A framework for fault-tolerant control for an interacting
and non-interacting level control system using AI. In: Proceedings of the 15th international
conference on informatics in control, automation and robotics, vol 1. ICINCO, pp 180–190.
https://doi.org/10.5220/0006862001900200
3. Himanshukumar R, Vipul S (2019) Stable fault tolerant controller design for TakagiSugeno
fuzzy model-based control systems via linear matrix inequalities: three conical tank case study.
Energies 12(11):2221. https://doi.org/10.3390/en12112221
4. Himanshukumar R, Vipul S (2019) A passive fault-tolerant control strategy for a non-linear
system: an application to the two tank conical non-interacting level control system. MASKAY
9(1):1–8. https://doi.org/10.24133/maskay.v9i1.1094
54 H. R. Patel and V. A. Shah

5. Himanshukumar R, Vipul S (2018) A fault-tolerant control strategy for non-linear system:

an application to the two tank Conical non interacting level control system. In: 2nd inter-
national conference on distributed computing, VLSI, electrical circuits and robotics (IEEE
DISCOVER), 13–14 Aug. IEEE Xplore, Manglore, India, pp 64–70. https://doi.org/10.1109/
DISCOVER.2018.8674119
6. Jauregui C, Duarte M, Orostica R, Travieso Torres J (2016) Conical tank level control with
fractional PID. IEEE Latin Am Trans 14(6):2598–2604. https://doi.org/10.1109/TLA.2016.
7555225
7. Himanshukumar R, Vipul S (2019) Fault tolerant controller using interval type-2 TSK logic
control systems: application to three interconnected conical tank system. In: Kearfott R, Batyr-
shin I, Reformat M, Ceberio M, Kreinovich V (eds) Fuzzy techniques: theory and applica-
tions. IFSA/NAFIPS 2019. Advances in intelligent systems and computing, vol 1000. Springer,
Cham. https://doi.org/10.1007/978-3-030-21920-8_42
8. Himanshukumar R, Vipul S (2019) Passive fault tolerant controller using hybrid controller for
effective control of two tank canonical non-interacting system. J Math Stat Sci Publ (in press)
9. Himanshukumar R, Vipul S (2018) Fuzzy logic based passive fault tolerant control strategy for a
single-tank system with system fault and process disturbances. In: 5th international conference
on electrical and electronics engineering (ICEEE). IEEE Xplore, Istanbul, Turkey, pp 257–262.
https://doi.org/10.1109/ICEEE2.2018.8391342
10. Himanshukumar R, Vipul S (2019) Performance comparison of passive fault tolerant control
strategy with PI and fuzzy control of single-tank level process with sensor and system fault.
Am J Eng Appl Sci 12(2):236–246. https://doi.org/10.3844/ajeassp.2019.236.246
11. Rajesh T, Arun Jayakar S, Siddharth SG (2014) Design and implementation of IMC based PID
controller for conical tank level control process. Int J Innovative Res Electr Electron Instrum
Control Eng 2(9):2041–2045
12. Ganesh Ram A, Abraham Lincoln S (2013) A model reference-based fuzzy adaptive PI con-
troller for non-linear level process system. Int J Res Rev Appl Sci 4(2):477–486
13. Dinesh C, Manikanta VV, Rohini HS, Prabhu KR (2015) Real time level control of conical
tank and comparison of fuzzy and classical PID controller. Indian J Sci Technol 8(S2):40–44.
https://doi.org/10.17485/ijst/2015/v8iS2/58407
14. Gireesh N, Sreenivasulu G (2014) Comparison of PI controller performances for a conical
tank process using different tuning methods. In: International conference on advances in elec-
trical engineering. IEEE Xplore, Vellore, India, pp 1–4. https://doi.org/10.1109/ICAEE.2014.
6838426
15. Himanshukumar, R., Vipul S.: Fault Tolerant Control Using Interval Type-2 Takagi-Sugeno
Fuzzy Controller for Nonlinear System. In: Abraham A., Cherukuri A., Melin P., Gandhi N.
(eds.) Intelligent Systems Design and Applications 2018, Advances in Intelligent Systems and
Computing, vol 941,Springer, Cham. 10.1007/978-3-030-16660-1_15
16. Himanshukumar R, Vipul S (2019) Actuator and system component fault tolerant control using
interval type-2 Takagi-Sugeno fuzzy controller for hybrid nonlinear process. Int J Hybrid Intell
Syst 15(3):143–153. https://doi.org/10.3233/HIS-190267
17. Himanshukumar R, Vipul S (2018) Fault detection and diagnosis methods in power generation
plants-Indian power generation sector perspective: an introductory review. J Energy Manage
(JEM) 2(2):31–50
18. Oldham KB, Spanier J (1974) The fractional calculus: theory and applications of differentiation
and Integration to arbitrary order. Academic Press, New York, NY, USA
19. Podlubny I (1999) Fractional differential equations: an introduction to fractional derivatives,
fractional differential equations to methods of their solution and some of their applications.
Academic Press, San Diego
20. Miller KS, Ross B (1993) An introduction to the fractional calculus and fractional differential
equations. Wiley, New York
21. Valrio D, Costa JS (2006) Tuning of fractional PID controllers with Ziegler–Nichols-type rules.
Signal Process 86(10):2771–2784. https://doi.org/10.1016/j.sigpro.2006.02.020
A Fractional and Integer Order PID Controller for Nonlinear System … 55

22. Oustaloup A, Sabatier J, Lanusse P (1999) From fractal robustness to CRONE control. Frac-
tional Calc Appl Anal 2(1):1–30
23. Monje CA, Feliu V (2004) The fractional-order lead compensator. In: IEEE international con-
ference on computational cybernetics (ICCC), 30 Aug–1 Sept. IEEE Xplore, Vienna, Austria,
pp 347–352. https://doi.org/10.1109/ICCCYB.2004.1437746
24. Kala H, Aravind P, Valluvan M (2013) Comparative analysis of different controller for nonlinear
level control process. In: IEEE conference on information and communication technologies.
IEEE Xplore, Tamil Nadu, India, pp 724–729. https://doi.org/10.1109/CICT.2013.6558189
25. Loverro A (2004) Fractional calculus: history, definitions and applications for the engineer.
University of Notre Dame, Notre Dame, U.S.A
26. Bequette BW (2003) Process control: modeling, design, and simulation. Prentice Hall, USA
27. Isermann R, Ball P (1997) Trends in the application of model-based fault detection and diagno-
sis of technical processes. Control Theory Pract 5(5):709–719. https://doi.org/10.1016/S0967-
0661(97)00053-1
28. Samko SG, Kilbas A, Marichev OI (1993) Fractional integrals and derivatives. Gordon and
Breach, Yverdon
29. Tepljakov A, Petlenkov E, Belikov J (2011) FOMCON: a MATLAB toolbox for fractional-
order system identification and control. Int J Microelectron Comput Sci 2(2):51–62
30. Man KF, Tang KS, Kwong S (1996) Genetic algorithms: concepts and applications. IEEE Trans
Ind Electron 43(5):519–534. https://doi.org/10.1109/41.538609
31. Monje CA, Vinagre BM, Chen YQ, Feliu V, Lanusse PJ (2004) Sabatier, proposals for fractional
P I λ D μ tuning. In: Fractional Differentiation and its Applications, Bordeaux
Modified Droop Control Strategy
for Load Sharing and Circulating
Current Minimization in Low-Voltage
Standalone DC Microgrid

Rakesh Sukhadiaa and Saurabh Pandya

Abstract This Paper focuses on issues developed in parallel DC–DC converter in

low-voltage isolated DC Microgrid. Two main issues are faced by DC Microgrid
when two or more than two converters are connected in parallel, one is unequal
load sharing and second circulating current. This occurs due to the output voltage
of converter should be different. The main reason for the output voltage of both the
converters is different due to change in load, input power, and also feedback voltage
and current. If two or more than two sources are connected in parallel and there is any
mismatch in output voltage will initiate circulating current and difference in current
sharing which cause overload to the converters with variation in power sharing.
Droop control is the popular method for load sharing in DC Microgrid but due to
its disadvantages poor circulating current and voltage droop arise due to mismatch
in the converters output voltages. Another method is virtual resistance introduced
which gives the relationship between virtual resistance and current sharing as well
as virtual resistance and bus voltage in DC microgrid. This proposed modified droop
control technique try to minimize circulating current and losses in the output side of
converters.

Keywords PV sources · DC–DC boost converter · Load sharing · DC microgrid ·

Parallel converters · Without droop control · Virtual resistance (Conventional
droop) method and modified droop control using MOSFET

1 Introduction

In renewable energy sources, PV power generation system has gained growing atten-
tion in recent years. A photovoltaic system is made from several photovoltaic solar
cells. An individual small cell is capable of generating about 1 or 2 W of power

R. Sukhadiaa (B)
KSV University, Gandhinagar 382015, Gujarat, India
e-mail: [email protected]
S. Pandya
EE Department, LE College, Morbi, India
© Springer Nature Singapore Pte Ltd. 2020 57
A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_5
58 R. Sukhadiaa and S. Pandya

which approximately depends on the type of material used. For higher output, PV
cell can be connected together to form higher power modules. Apart from the solar
PV module, the system components comprise a battery charge controller, inverter,
and MPPT controller. A balanced system includes components and equipment that
convert DC supply from the solar PV module to AC grid supply [1]. Commercial PV
control algorithm is available and research activity is focused [2] on the electronics
applied to PV modules as a measure to improve the PV systems behavior and to
reduce the losses in PV systems [3]. In [4] PV installations should be increased as
rooftop system within distributed power system which is very popular today. The PV
semi-isolated system are shown and discussed in [5] and verify the technical feasi-
bility of suggested system controls. Investigate some methods [6] that can be used
to reduce the fluctuations in the power generated from a large customer-owned PV
system, in the order of megawatts. In [7] focuses on inverter technologies for single
phase grid connected PV module which are classified and various inverter topologies
are presented. They [8] covered some standards that inverter for PV and grid applica-
tion must fulfill, which focus on power quality, injection of DC current into the grid,
detection of islanding operation, and system grounding. An important result is that
the amplitude of the ripple across a PV module should not exceed 3.0 V in order to
have a utilization efficiency of 98% at full generation. Renewable energy technologies
and distributed energy resources developed and encourage technological innovation
to reduce greenhouse gas emission and achieve energy security and independence to
meet the growing electricity demand. DC transmission power is 10% more than AC
transmission power compared to system efficiency. DC microgrid instability depends
on voltage fluctuation, power quality and suddenly change the load needed to inves-
tigate and resolved the effective use of DC microgrid generation. In this system [9]
the peak power of the PV modules and feed or absorbs power from the utility auto-
matically without any operation mode control. In [10] discuss the state of the art for
grid tied PV inverters at low and medium power level, mainly intended for rooftop
applications. The grid connected PV system is presented in [11] with a Le-Blanc
transformer is connected with multilevel inverter. Then the technology turns optimal
design of transformer less photovoltaic PV inverters in [12]. The common problems
encountered in PV connected grid with specially focus on voltage profile at the point
of common coupling (PCC) between PV and grid in [13]. In [14] gives importance of
grid connected PV system with high conversion efficiency and power factor exceed-
ing 90% as well as THD less than 5%. In [15] developed maximum power point
tracking method and output current shaping function into a buck-boost derived con-
verter and then inverts the shaped current though grid frequency bridge to the grid.
The performance checked in [16] of the photovoltaic module in non-ideal condi-
tions and proposes topologies to minimize the degradation of performance caused
by such type of non-ideal conditions. The design of one-cycle controller of a single-
stage inverter for photovoltaic applications given in [17] which allows the design
of maximum power point tracking perturb and observe controller that improves the
inverter performance. The prototype version of the control strategy of permanent
magnet synchronous generator for MPPT and compares with the results provide in
[18]. Single phase Cuk converter topology of grid connected PV system developed
Modified Droop Control Strategy for Load Sharing … 59

in [19] which has a wide voltage range for PV array voltage. In [20] gives a review on
the key technologies for microgirds, including new power electronic, protection and
communication technologies. In addition, it possesses the features of easy control as
the two-stage system, high efficiency as the one-stage system, and the most important
is that the load voltage is regulated. Simple moving voltage average technique used
to reduce the fluctuations [21] in solar PV output voltage under non-uniform solar
irradiance. These techniques with fixed step direct control incremental conductance
method is very useful to improve the performance of PV system by maximum power
point tracking system when varying fast change in solar irradiations. The simple
moving voltage average technique to reduce the fluctuations in solar PV output volt-
age under non-uniform solar irradiances. The maximum power point tracking part
of the control system of the PV generator dominates the dynamic behavior of the
system. DC microgrid different control techniques are reviewed in [22] and their
problems are discussed. The output current of the converters are accurately propor-
tional to the rated current of converters, and output voltage of converters are regulated
close to the reference value. The task of Energy management system given [23] to
manage the energy between source and load with simple model of dc microgrid by
MATLAB/Simulink containing photovoltaic array, wind turbine generator, battery
storage, grid and energy management block. The need of renewable energy based
DC microgrid given in [24] with stringent control for harnessing maximum available
energy, energy storage system demands efficient management and DC-link voltage
must be maintained constant. The small-signal model of the suggested control sys-
tem for a simplified DC microgrid is obtained and its stability is analyzed in order
to design the control parameters [25]. Microgrid can be classified as AC microgrid
and DC microgrid depending on the nature of bus voltage. DC microgrid has several
advantages over AC microgird. Many of the renewable energy sources like solar
photovoltaic panels and fuel cells produce DC power and many of the existing loads
like computers, servers, LED lamps, and electric vehicles work on DC power [26].
New Decentralized power management and load sharing technique given in [27]
with PV system and battery conditions to share the load. Supervisory power man-
agement system for a grid interactive microgrid with hybrid energy storage system
given in [28] to reduce number of sensors required. To regulate the voltage a new
control strategy is proposed in [29] for control the power sharing between batteries
and Super capacitors to match the generation and demand. The effect of changes in
temperature as well as climate changes which fluctuated output of photovoltaic sys-
tem, so improved control strategy presented in [30] to control grid connected battery
energy system for mitigate PV fluctuations. Hybrid-PV systems for the provision of
electricity and hot water supply for the domestic sector in UK [31]. Prioritized rule
based load management technique presented in [32] to provide of optimum energy
gleaning and management in stand-alone PV-battery system with effective results.
The control of AC microgrid deals with power flow, load sharing, voltage regulation,
and mitigation of various kinds of power quality issues [33], whereas in DC micro-
grid, power quality issues such as skin effect and reactive power are not present.
Therefore, compared to AC, DC microgrids are highly efficient, reliable, easy to
control, and economic [34]. There are several control issue related to the microgrid,
60 R. Sukhadiaa and S. Pandya

including interconnection schemes between distributed generators and common DC

grid, voltage control among the parallel generators with proper load sharing and min-
imize circulating current passing between converters [35, 36]. In novel droop control
method for the converter parallel operation which controls the reference voltage of
each module is to improve the output voltage regulation and the current sharing [37].
Equal current sharing was achieved along with low-voltage regulation in [38] which
is verified through detailed simulation study. Adaptive droop resistance technique
was implemented for the compensation voltage in DC–DC boost converter in [39].
In this paralleling power converters method classified and focused [40] on the cur-
rent sharing that is approached by simulation of two parallel buck converter system.
Design and calculation of the DC–DC boost converter implemented with all details
[41].

1.1 Design of DC–DC Boost Converter

As the name, the step-up or boost converter gives as the output voltage always greater
than the input voltages. The boost topology is given in Fig. 1 the boost converter
consists of a SPDT switch, an inductor that is connected to the pole of SPDT switch,
and a capacitor is connected to the throws of the SPDT switch. Inductor is connected
between the pole and the input source. As shown in Fig. 2, the operative circuit

P RL
SOLAR C
PANEL

Fig. 1 DC–DC boost converter with SPDT switch [41]. Vin = 72 V, Vout = 110 V, V o = Vi
1−D ,
IO
1 − D = 0.65, D = 0.35, Vin Iin = VO IO , Iin = 1−D , IO = 4.54 A

S1
L
P
RL
SOLAR C
PANEL
S2

Vin D
Fig. 2 DC–DC boost converter with SPDT switch [41]. L = i L f s , i L = 13% o f i L , L =
5.58 × 10−3 H, C= IO D
VO f s , VO = 8% o f VO , C = 3.46 × 10−3
Modified Droop Control Strategy for Load Sharing … 61

Table 1 DC–DC boost

Parameters Values
converter parameters
Output power 490 W
Output voltage 110 V
Inductor 5.58E–03
Input capacitor 4e–3F
Nominal switching frequency 5 kHz
Output capacitor 3.46e–4F
Type of cable Polycab
Internal resistance 2 m/m

during the period of DTS of a switching period S2 is ON and S1 is OFF, and during
the operative circuit during the period of (1 − D)TS of the switching period S1 is ON
and S2 is OFF. The inductor current cannot change instantaneously. The inductor
current starts to decrease. This negative ddit of the inductor current develops sufficient
voltage L ddit with a polarity such as to drive the inductor current through S1 to charge
the output capacitor [41].

1.2 System Configuration

There are two PV sources (250 W) connected in parallel with 110 V DC bus-line
to the variety of electrical load (25 ) another load (25 ) connected in parallel
when the breaker is operated after 0.11 s. Table 1 gives the details of two same PV
sources with rated output power 500 W, with output constant voltage at DC grid
110 V, calculated values of inductor and capacitor. As per given below, consider
the polycab cable used in PV system which has internal resistance 0.002 /m, here
considered 50 m long cable with 0.1  resistance.

1.3 DC–DC Boost Converter Closed Loop

A DC–DC boost converter consists of an inductor, diode, MOSFET, output filter

capacitor, and resistive load. When the switch is closed and supply voltage is given,
inductor current increases. When the switch is opened, both inductor and supply
voltage gets discharged through the load. The higher voltage at the output is obtained
when the given input voltage is low. PV input is variable which is based on irradiance
and temperature of atmosphere. PV input voltage is variable 48–72 V which is given
to DC–DC boost converter then the output 110 V fixed at load (Fig. 3).
62 R. Sukhadiaa and S. Pandya

Cable
L Resistance
Load 25 Ω
0.5 Ω

PV PANEL 110V
C1
48-72V C2 OUTPUT
VOLTAGE

Fig. 3 Simulation of DC–DC boost converter with fixed 110 V output

2 Control Voltage Technique for Parallel-Connected

Converters

The DC–DC boost converters are used in applications where the required output
voltage is needed to be higher than the source voltage. When converters are connected
in parallel and if there is change in voltage or load, then this will cause mismatching
in converter output voltage which will cause circulating current. Circulating current
will increase the flow current through the switches which will increase the power
electronic switch ratings and losses, cause overload and poor load sharing among
the parallel-connected converters. There are different types of methods to control the
voltage at DC microgrid.
There are different cases considered to control voltage of DC microgrid.
I. No Droop Technique (fixed cable resistor)
II. Virtual output Resistance method (droop gain)
III. Modified Droop control method (with using MOSFET)
Load sharing and circulating current issues for parallel DC–DC converters con-
nected to a low-voltage DC microgird shown in Fig. 4. Two parallel DC–DC con-
verters are interfaced with PV arrays and DC grid. PV source 1 and PV source 2
have voltage, current, and fixed cable resistance represented by V1 , V2 , I1 , I2 , R1 ,
and R2 , respectively. Load is connected to 25 . The breaker is operated after 0.11 s
and load is connected in 12.5  parallel, so the total load is 25  after 0.11 s. IC12
Circulating current passing from PV source 1 to PV source 2 if V1 > V2 . Otherwise,
IC21 Circulating current passing from PV source 2 to PV source 1 if V2 > V1 .

2.1 No Droop Technique (Fixed Cable Resistor)

Distributed generators are connected in parallel to the common DC bus using DC–DC
converters. Control strategies are applied [37–39] to control the distributed gener-
ators. It consists of three levels of control—primary, secondary, and tertiary levels.
The objective of primary control is to ensure proper load sharing between the dis-
tributed generators. The objective of the secondary control level is to regulate the
Modified Droop Control Strategy for Load Sharing … 63

Fig. 4 Two parallel DC CABLE

source to low-voltage DC RESITANCE
microgrid [33]
0.1 Ω

CABLE i1
RESITANCE

0.1 Ω

i2

LOAD 25 Ω

LOAD 25 Ω

BREAKER OPERATED AFTER

0.11 SECOND

voltage deviations and the tertiary control level regulates the flow of power between
the grid and the microgrid. This method is popular decentralized control method
used to share load between parallel converters. The advantages of these methods do
not require any communication channel between the sources. This method has been
proposed in [33] (Fig. 5).
By applying KVL law, the output current of the converter can be determined from
the equation and circulating current can be calculated from

V1 − I1 R1 − I L R L = 0 (1)

V2 − I2 R2 − I L R L = 0 (2)

I L = I1 + I2 (3)

V1 = I1 R1 + (I1 + I2 )R L = 0 (4)

V2 = I2 R2 + (I1 + I2 )R L = 0 (5)

after solving Eqs. (4) and (5)

64 R. Sukhadiaa and S. Pandya

Cable
Resistance
L 0.1 OHM

110V 25
OUTPUT OHM
PULSE VOLTAGE
PV PANEL C1 C2
48-72V

Cable
L Resistance
0.1 OHM
110V
OUTPUT
VOLTAGE
PV PANEL C1 C2
PULSE
48-72V

Fig. 5 Two parallel DC–DC boost converter with same cable resistance no droop technique. V 1 —
Output voltage of PV Source 1, V 2 —Output voltage of PV Source 2, I 1 —Output curent of PV
Source 1, I 2 —Output current of PV Source 2, R1 —Cable Resistance of PV Source 1, R2 —Cable
Resistance of PV Source 2, I C12 —circulating current from convt. 1 to convt. 2, I L —Load current,
RL —Load resistance

V1 (R2 + R L ) − V2 R L
I1 = (6)
R1 R2 + R1 R L + R2 R L
V2 (R1 + R L ) − V1 R L
I2 = (7)
R1 R2 + R1 R L + R2 R L

Circulating current passing between converter 1 to converter 2, then

V1 − V2
IC12 = (8)
R1 + R2
I1 R1 − I2 R2
IC12 = (9)
R1 + R2

If both the resistance are equal and then R1 = R2 .

I1 − I2
IC12 = (10)
2
The drawback of No-droop control method in dc microgrid are poor voltage
regulation. In this method use only cable resistance considered, to overcome the
limitations, an instantaneous droop calculation method variable resistance droop
control (conventional droop control) method used [11]. If all converters have same
output resistance and same output voltage then the circulating current should be zero.
Modified Droop Control Strategy for Load Sharing … 65

But here, the consideration for the PV source 1 and PV source 2 has same 0.1 
cable resistance and circulating current is calculated from the 1000 R and 25 T as
per standard condition for the PV system.

V1 = 117.3 V, V2 = 118.2 V, VL = 101.8 V and R1 = 0.1 , R2 = 0.1 ,

RL = 12.5 , I1 = 4.42 A I2 = 4.39 A, And IL = 8.81 A.

The circulating current passing through the system is (I1 − I2 ) when breaker
changed from 25 to 12.5 , circulating current flow is I C = 5.2 A. The major
drawbacks of this method used in DC microgrid are poor voltage regulation, use only
cable resistance considered, to overcome these limitations, an instantaneous droop
calculation method variable resistance droop control (conventional droop control)
method used.

2.2 Virtual Output Resistance Method (Conventional Droop

Gain)

When two DC–DC converters are connected in parallel with same nominal output
voltage, there is no chance for flow circulating current from them. But if there in case
of minor voltage difference between Dc source, or if two converters are at different
locations the output resistance of each converter may not be same, at that time
circulating current flows between the sources. A virtual output resistance method tries
to reduce circulating current as well as proper load sharing between the converters.
This adjustment should be done by the reference voltage given to the inner voltage
and current loop. In this method, to keep the output voltage constant, it uses the
feedback and compares with reference voltage which is required. Then the generated
error should be given to PI controller. The product of output current and droop gain
are also compared with the output of PI controller, then error generated should be
given to another PI controller and given the pulse to the gate of MOSFET of boost
converter via PWM generator to control the output voltage. In this method, voltage is
sensed across the series resistor and used to produce a drop in the output voltage that
is proportional to the output current of the supply. The parallel DC–DC converter
control using virtual resistance method is given in Fig. 6

VL = VREF − RD IL (11)

where VL = Output Voltage at Load, IL = Load Current, and

RD = Droop Resistance (virtual output Resistance) VREF =
No Load output reference voltage or nominal voltage

VN = Nominal output voltage = 110 V

EV = Maximum allowed votlage deviation 5%
66 R. Sukhadiaa and S. Pandya

Cable
Resistance
PV DC-DC 0.1 OHM
PANEL BOOST
48-72V CONV. I

DRIVER &
PULSE GEN.

CURRENT VOLTAGE
LOOP GAIN
LOOP

VREF.

25 OHM
Cable LOAD
Resistance
PV DC-DC 0.1 OHM
PANEL BOOST
48-72V CONV. I

DRIVER &
PULSE GEN.

CURRENT VOLTAGE
LOOP GAIN
LOOP

VREF.

Fig. 6 Two PV sources with same cable resistance using virtual resistance method

EV
VREF = VN −
2
5% O F 110
VREF = 110 −
2
5.5
VREF = 110 −
2
VREF = 110 − 2.75
VREF = 107.25 V (12)

EV
RD =
IL
5.5
RD =
8.81
RD = 0.624 (13)

Simple calculation is given for the system 1000 R and 25 T as per standard
condition and the results are given below for the calculation of virtual resistance
Modified Droop Control Strategy for Load Sharing … 67

VL = 101.8 V, VREF = 107.25 V, EV = 5.5 V, IL = 8.31 A

VL = VREF − RD IL , get RD = 0.624

Then the circulating current flowing through the p converters is I C = 0.4 A.

This method considers only output current then multiplies with the RD droop
resistance and then it is compared with the reference voltage and generates the error
signal.
In this control strategy, DC source consists of one or more switches which are
driven by a PWM block. The control strategy consists of two loops one is voltage
loop and another current loop to allow the load sharing in parallel operation and
improves the dynamic performance of output voltage. One of the disadvantages of
this method is voltage deviation that is caused due to changing in load and this is
removed from the modified droop control method using MOSFET.

2.3 Modified Droop Control Method (Using MOSFET)

MOSFET can operate only in a switching mode, acting like either an open or closed
switch. When MOSFET is connected to DC circuit or DC supply, the MOSFET
can be used as DC switch to control larger DC current and loads. One of the main
advantages of using a MOSFET as a switch in a DC circuit is that it has a very high
current gain. MOSFET is voltage controlled device when the device is fully ON
VGS(TH) > VGS otherwise OFF. No gate current flow during ON to OFF state. The
switching time is very short in the range of few nano seconds depending on device.
The ON-state resistance RDS(ON) of the MOSFET between drain and source increase
rapidly with the device blocking voltage rating BVDSS that can be expressed as

RDS(ON) = K BVDSS
2.5−2.7
(14)

where K is constant and average power switching loss

1  
PS = Vd IO f s tc(ON) + tc(OFF) (15)
2
MOSFET are positive temperature coefficient so it is easily paralleled. Modified
droop control method by using MOSFET at the point of common coupling. It is
used in parallel with cable resistance to reduce circulating current and proper power
sharing. MOSFET located at the point of common coupling (PCC) which is used a
high-level system control scheme to maintain compliance. In this application, two PV
systems connected in parallel with using MOSFET as a switch to increase reliability,
power quality, and minimization circulating current at the point of common coupling.
Using the concept of MOSFET switched the intentional to control the power flow
between two DC–DC converters, and DC microgrid with using the cable resistance.
The system becomes more fast, smart, and reliable.
68 R. Sukhadiaa and S. Pandya

V1 = 109.2 V, V2 = 110.7 V, VL = 108.5 V

R1 = 0.1 , R2 = 0.1 , RL = 12.5 ,
I1 = 4.42 I2 = 4.39 A, And IL = 8.81 A.

The circulating current passing through the system is (I1 − I2 ).

The voltage difference from the output side is taken as feedback and compared
with the reference voltage then the generated error signal given to PI controller with
the value of P is 0.009 and I is 160. This signal given to the gate of MOSFET with
R O N = 0.40 , VGS = 10 V, VDS = 200 V IFR630 Power MOSFET.

Req = (1 − D)RDROOP1 (16)

Req = (1 − D)RDROOP2 (17)

D is the duty cycle change with the change in output voltage, the switch connected
in parallel with the resistance RDROOP which is changed. This method is most suitable
than the virtual droop control method (conventional) because in this method consider
RD droop resistance is constant for all the time but in this method is come into the
circuit when the output voltage is change otherwise the switch is bypass the resistor.
This method gives the circulating current flowing through the parallel converters
I C = 0.395 A (Fig. 7).

3 Compare All the Cases as Per the Given Data

There are different cases considered for two different parameter irradiance and tem-
perature. Table 2 gives the data with irradiation 1000 W/m2 and three temperatures
25 °C which is considered with same cable resistance 0.1  for both PV-1 and PV-2.
All are three methods are simulated. This observation table gives the details about
input and output power, current, and also power with power losses and circulating
current flow between them. In all the cases that is the case III which has minimum
power losses and minimum circulating current (Fig. 8).
There are different cases considered as per table.
I. Conventional technique is considered as case-I
II. Virtual output Resistance method (droop gain) is considered as case-II
III. Modified Droop control method (with using MOSFET) is considered as case-III
(Fig. 9)
As shown in figure when no droop is applied with the DC microgrid system,
mostly the two PV sources are equal voltage so the circulating current from these
two sources are I C = 5.2 A at the time of breaker ON but after some time the value of
current is 0.3 A compared to total current passing from the I L = 8.81 A (Fig. 10).
Modified Droop Control Strategy for Load Sharing … 69

DRIVER &
PULSE GEN.

V OUT
VREF.

PV DC-DC
PANEL BOOST 25
48-72V CONV. I OHM
Cable
Resistance LOAD
DRIVER & 0.1 OHM
PULSE GEN.

CURRENT VOLTAGE
LOOP GAIN
LOOP

VREF.
DRIVER &
PULSE GEN.

VREF.

PV DC-DC
PANEL BOOST
48-72V CONV. I
Cable
DRIVER & Resistance
PULSE GEN. 0.1 OHM V OUT

CURRENT VOLTAGE
LOOP GAIN
LOOP

VREF.

Fig. 7 Two parallel DC–DC boost converter with same cable resistance modified droop using
MOSFET

Table 2 Losses and circulating current as per irradiation and temperature for different cases
Radiation Temp. in Different Input Output Efficiency Losses Circulating
W/m2 deg. cases power power in current
Watt flow in
Amp
1000 25 Case-I 490 480 97.95918 10 5.2
Case-II 490 473 96.53061 17 0.4
Case-III 490 481 98.16327 9 0.395
All data are taken when both the PV sources connected in parallel with considering same cable
resistance 0.1  for both PV-1 and PV-2
70 R. Sukhadiaa and S. Pandya

POWER LOSS
20 CIRCULATING CURRENT

10
0 CIRCULATING CURRENT
POWER LOSS
CASE- CASE-
1 CASE-
2 3

Fig. 8 Column chart from Table 2 gives the comparison between circulating current and power
losses

Fig. 9 Circulating current flow when load changes and current stable after 0.11 s when conventional
droop control method (only cable resistance considered case-I)

Fig. 10 Circulating current flow when load changes and current stable after 0.22 s when virtual
resistance method (Case-II)

When conventional droop (Virtual Resistance) method is applied with the DC

microgrid system, mostly the two PV sources are of equal voltages, circulating current
of the sources are I C = 0.4 A compared to total current passing from the I L = 8.81 A
(Fig. 11).
Modified droop control method applied with the DC microgrid system, mostly
the two PV sources are equal voltage, so the circulating current of all the sources are
I C = 0.395 A compared to total current passing from the I L = 8.81 A (Figs. 12,
13, and 14).
Modified Droop Control Strategy for Load Sharing … 71

Fig. 11 Circulating current flow when load changes and current stable after 0.22 s when modified
droop control method (using MOSFET Case-III)

Fig. 12 Output power and power losses due to circulating current flow after 0.11 s when load is
changed when conventional droop control method (only cable resistance considered Case-I)

Fig. 13 Output power and power losses due to circulating current flow after 0.11 s when load is
changed by virtual resistance method (Case-II)

Fig. 14 Output power and power losses due to circulating current flow after 0.11 s when load is
changed by when modified droop control method (using MOSFET Case-III)
72 R. Sukhadiaa and S. Pandya

4 Conclusion

From the above results three different voltage control techniques are compared with
focus on circulating current and power losses at DC microgrid. A parallel DC–
DC converter system with same cable resistance primary or no droop has more
circulating current with power losses. Then use virtual resistance (conventional)
method which has gain with feedback to input but not got accurate result from the
droop gain which is fixed value. The third method is modified droop control using
MOSFET parallel with resistance in which easily vary the resistance. Among three
techniques compared with different temperature and irradiance that modified droop
control technique is best which has lowest circulating current as well as low power
loss. This technique also gives good voltage regulation. The future scope of the work
includes the consideration of microgrid system with different two or three sources
like wind, solar, etc., compare all the methods with different ratings of distributed
sources connected in parallel then try to find the best method among these.

References

1. Kumar LA, Surekha P (2015) Solar PV and wind energy conversion systems
2. Tan YT, Kirschen DS, Jenkins N (2004) A model of PV generation suitable for stability analysis.
IEEE Trans Energy Convers 19(4):748–755
3. Roman E, Alonso R, Ibañez P, Elorduizapatarietxe S, Goitia D (2006) Intelligent PV module
for grid-connected PV systems. IEEE Trans Ind Electron 53(4):1066–1073
4. De la Cruz J, Gómez P (2010) Real problems in utility high voltage network due to grid
connected photovoltaic power generation. The experience of Endesa. In: Proceedings of 2010
international conference on renewable energies and power quality, Granada, Spain, pp 1–6
5. Locment F, Sechilariu M, Houssamo I (2012) DC load and batteries control limitations for
photovoltaic systems. Experimental validation. IEEE Trans Power Electron 27(9):4030–4038
6. Omran WA, Kazerani M, Salama MMA (2011) Investigation of methods for reduction of power
fluctuations generated from large grid-connected photovoltaic systems. IEEE Trans Energy
Convers 26(1):318–327
7. Kjaer SB, Pedersen JK, Blaabjerg F (2005) A review of single-phase grid-connected inverters
for photovoltaic modules. IEEE Trans Ind Appl 41(5):1292–1306
8. Chang JM, Chang WN, Chung SJ (2006) Single-phase grid-connected PV system using three-
arm rectifier-inverter. IEEE Trans Aerosp Electron Syst 42(1):211–219
9. Carrasco JM, Franquelo LG, Bialasiewicz JT, Galván E, PortilloGuisado RC, Prats MM,
Moreno-Alfonso N (2006) Power-electronic systems for the grid integration of renewable
energy sources: a survey. IEEE Trans Industr Electron 53(4):1002–1016
10. Schimpf F, Norum LE (2008) Grid connected converters for photovoltaic, state of the art, ideas
for improvement of transformerless inverters. In: Nordic workshop on power and industrial
electronics (NORPIE/2008), 9–11 June 2008, Espoo, Finland. Helsinki University of Technol-
ogy
11. Pires V, Martins JF, Foito D, Hão C (2012) A grid connected photovoltaic system with a
multilevel inverter and a Le-Blanc transformer. Int J Renew Energy Res (IJRER) 2(1):84–91
12. Koutroulis E, Blaabjerg F (2013) Design optimization of transformer less grid-connected PV
inverters including reliability. IEEE Trans Power Electron 28(1):325–335
Modified Droop Control Strategy for Load Sharing … 73

13. El-Amin IM, Ahmed MK (2014) Impact of a PV system on a power grid. In: 2014 international
symposium on power electronics, electrical drives, automation and motion (SPEEDAM), pp
966–970. IEEE
14. Eltawil MA, Zhao Z (2010) Grid-connected photovoltaic power systems: technical and potential
problems—a review. Renew Sustain Energy Rev 14(1):112–129
15. Ho BM, Chung SH, Hui SYR (2004) An integrated inverter with maximum power tracking for
grid-connected PV systems. In: Nineteenth annual IEEE applied power electronics conference
and exposition, 2004. APEC’04, vol 3, pp 1559–1565. IEEE
16. Xiao W, Ozog N, Dunford WG (2007) Topology study of photovoltaic interface for maximum
power point tracking. IEEE Trans Ind Electron 54(3):1696–1704
17. Fortunato M, Giustiniani A, Petrone G, Spagnuolo G, Vitelli M (2008) Maximum power point
tracking in a one-cycle-controlled single-stage photovoltaic inverter. IEEE Trans Industr Elec-
tron 55(7):2684–2693
18. Tan K, Islam S (2004) Optimum control strategies in energy conversion of PMSG wind turbine
system without mechanical sensors. IEEE Trans Energy Convers 19(2):392–399
19. Yang Y, Zhao FP (2011) Adaptive perturb and observe MPPT technique for grid-connected
photovoltaic inverters. Procedia Eng 23:468–473
20. Ding M, Zhang Y, Mao M (2009) Key technologies for microgrids-a review. In: International
conference on sustainable power generation and supply, 2009, SUPERGEN’09, pp 1–5. IEEE
21. Ali A, Wang Y, Li W, He X (2015) Implementation of simple moving voltage average technique
with direct control incremental conductance method to optimize the efficiency of DC microgrid.
In: 2015 international conference on emerging technologies (ICET), pp 1–5. IEEE
22. Ali A, Farooq A, MuhRenewable Energyad Z, Habib F, Malik SA (2015) A review: DC micro-
grid control and energy management system. Energy sources (RES) 7(10)
23. Bhavsar YS, Joshi PV, Akolkar SM (2015) Energy management in DC microgrid. In: 2015
international conference on energy systems and applications, pp 209–213. IEEE
24. Tank I, Mali S (2015) Renewable based dc microgrid with energy management system. In: 2015
IEEE international conference on signal processing, informatics, communication and energy
systems (SPICES), pp 1–5. IEEE
25. Peyghami S, Mokhtari H, Blaabjerg F (2017) Decentralized load sharing in a low-voltage direct
current microgrid with an adaptive droop approach based on a superimposed frequency. IEEE
J Emerg Sel Top Power Electron 5(3):1205–1215
26. Xia Y, Peng Y, Yang P, Yu M, Wei W (2017) Distributed coordination control for multiple
bidirectional power converters in a Microgrid AC/DC microgrid. IEEE Trans Power Electron
32(6):4949–4959
27. Karimi Y, Oraee H, Golsorkhi MS, Guerrero JM (2017) Decentralized method for load sharing
and power management in a PV/battery Microgrid source islanded microgrid. IEEE Trans
Power Electron 32(5):3525–3535
28. Kotra S, Mishra MK (2017) A supervisory power management system for a microgrid with
HESS. IEEE Trans Ind Electron 64(5):3640–3649
29. Kollimalla SK, Mishra MK, Ukil A, Gooi HB (2017) DC grid voltage regulation using new
HESS control strategy. IEEE Trans Sustain Energy 8(2):772–781
30. Daud MZ, Mohamed A, Hannan MA (2013) An optimal state of charge feedback control
strategy for battery energy storage in hourly dispatch of PV sources. Procedia Technol 11:24–
31
31. Herrando M, Markides CN, Hellgardt K (2014) A UK-based assessment of Microgrid PV
and solar-thermal systems for domestic heating and power: system performance. Appl Energy
122:288–309
32. Ayodele TR, Ogunjuyigbe ASO, Akpeji KO, Akinola OO (2017) Prioritized rule based load
management technique for residential building powered by PV/battery system. Eng Sci Technol
Int J 20(3):859–873
33. Augustine S, Mishra MK, Lakshminarasamma N (2015) Adaptive droop control strategy for
load sharing and circulating current minimization in low-voltage standalone DC microgrid.
IEEE Trans Sustain Energy 6(1):132–141
74 R. Sukhadiaa and S. Pandya

34. Thomas RM, Jose D (2015) Droop control method for parallel dc converters used in standalone
pv-wind power generating system
35. Batarseh I, Siri K, Lee H (1994) Investigation of the output droop characteristics of parallel-
connected DC-DCconverters. In: 25th annual IEEE power electronics specialists conference,
PESC’94 record, vol 2, pp 1342–1351. IEEE
36. Irving BT, Jovanovic MM (2000) Analysis, design, and performance evaluation of droop
current-sharing method. In: Fifteenth annual IEEE applied power electronics conference and
exposition, 2000, APEC 2000, vol 1, pp 235–241. IEEE
37. Kim JW, Choi HS, Cho BH (2002) A novel droop method for converter parallel operation.
IEEE Trans Power Electron 17(1):25–32
38. Anand S, Fernandes BG (2012) Modified droop controller for paralleling of dc–dc converters
in standalone dc system. IET Power Electron 5(6):782–789
39. Huang HH, Hsieh CY, Liao JY, Chen KH (2011) Adaptive droop resistance technique for adap-
tive voltage positioning in boost DC–DC converters. IEEE Trans Power Electron 26(7):1920–
1932
40. Luo S, Ye Z, Lin RL, Lee FC (1999) A classification and evaluation of paralleling methods for
power supply modules. In: 30th annual IEEE power electronics specialists conference, 1999,
PESC 99, vol 2, pp 901–908. IEEE
41. Umanad L. Power electronics—essential and applications. Willy publication
Primary and Secondary Droop Control
Method for Islanded Microgrid
with Voltage Regulation and Current
Sharing

Hitesh M. Karkar and Indrajit N. Trivedi

Abstract This paper presents primary and secondary control of islanded microgrid.
In conventional primary droop control, there is a problem of poor voltage regula-
tion of microgrid. The drawback of conventional primary droop control is improved
by voltage shifting based primary control. Voltage shifting based primary control
is not completely satisfied by the voltage regulation of microgrid. Also, there is
current sharing problem in both types of primary control. So, in secondary con-
trol, both current sharing and voltage regulation are achieved by slope adjusting and
voltage shifting simultaneously. In this secondary control method, average value of
current, voltage, and droop resistance of two neighboring converters is calculated
then controlled by additional layer of distributed secondary control over both pri-
mary controls. By adjusting droop coefficient of different converters making same
value of output impedance of converter, current sharing and voltage regulation are
achieved. Also, in transient state, current sharing is achieved by using this secondary
control technique. The performance of proposed microgrid is verified by MATLAB
simulation.

Keywords Current sharing · Droop control · Microgrid · Voltage regulation

1 Introduction

Distributed generation (DG) has an important role for increasing stability [1], effi-
ciency [2, 3], and reliability [4] in comparison with conventional power system.
DG unit, storage device, and load are widely integrated with microgrid. Nowadays,
microgrid becomes more popular due to increasing penetration of renewable energy
sources (like solar, wind) and energy storage devices like battery. In DC, microgrid

H. M. Karkar (B)
Gujarat Technological University, Ahmedabad, Gujarat, India
e-mail: [email protected]
I. N. Trivedi
G.E.C, Gandhinagar, Gujarat, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 75

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_6
76 H. M. Karkar and I. N. Trivedi

there is no problem of synchronizing and frequency regulation as compared to AC

microgrid. So, its stability is improved [4]. Microgrid has various sources and loads.
When inverter is on and exchanges the power from AC grid as per source generation,
it can be operated with grid-connected mode. When inverter is off then it is oper-
ated in islanding mode. In islanding mode, DG unit and storage device maintain the
stability of microgrid.
There are three levels of hierarchical control in microgrid. Primary level is respon-
sible for load sharing in DG and inner control of converter like voltage stability.
Secondary control is responsible for compensating voltage fluctuation produced by
primary layer. There are several centralized, decentralized, and distributed control
for maintaining stability of microgrid [4–11]. DG units of microgrid are connected
in parallel to common bus, current sharing has become one issue. In [12], current
sharing method master-slave, average current, and droop control are mentioned. The
droop control method is mostly used in microgrid for the purpose of load sharing
between converters [6–11]. It is a virtual control of impedance of converter. The
shared load is inversely proportional to output impedance of converter. Advantage of
droop control is high reliability and simplicity in DC microgrid [13, 14]. Advantage
and limitation of droop control are given [1, 15–18].
The drawback of conventional primary droop control is that output voltage of com-
mon bus linearly reduces with increasing load. The voltage shifting based proposed
primary control is used to solve this problem. In proposed primary control, voltage
regulation is improved but it does not exactly maintain rated voltage of common bus.
And also, there is a problem of current sharing.
So, to solve the above problem, secondary control is used [17–22]. The secondary
control is used to compensate the deviation due to primary control [17–21]. There
are two methods for secondary control in DC microgrid: centralized control and
distributed control. There is one microgrid central controller (MGCC) in centralized
secondary control method [17]. It generates voltage restoration value which is com-
monly applied to reference voltage of primary control. But if there is single-point
failure in MGCC, then system is failing. So in distributed secondary control [18] in
which local controller is used, only larger droop coefficient is required to achieve
current sharing. In [19–21], distributed secondary control is used with two controllers
for mean value of DC current and voltage to solve the problem of voltage regulation
and current sharing. But in transient state, the microgrid’s performance is very poor.
In this paper, the conventional primary, voltage shifting based primary and sec-
ondary control technique are mentioned. There are three PI controllers used in sec-
ondary control technique. In secondary control, its performance is good in dynamic
condition under fast-changing load. In secondary control, first controller is used for
restoration of voltage deviation in each converter produced proposed primary control
and other two controllers work together and regulate droop coefficients separately.
So, output impedance of each converter would become same. Average value of volt-
age, current, and droop coefficient is controlled to its reference value, respectively,
to achieve accurately current sharing and voltage regulation. This secondary control
is enhanced dynamic behavior under variable load condition.
Primary and Secondary Droop Control Method … 77

2 Equivalent Circuit for Microgrid with Primary

and Secondary Control

The microgrid is modeled with two nodes. There are two DC sources connected
in parallel with load. Each method with two converters is analyzed by Thevenin’s
equivalent circuit model.

2.1 Conventional Primary Droop Control

As per conventional droop control method, Eq. (1) is expressed.

∗
vi = vdc − i dci rdi (1)

∗
where vdc = Reference voltage of converter, rdi = Droop resistance, i dci = Output
current, and vi = Output voltage of converter, where i = 1,2.
Equations (2) and (3) can be derived from Fig. 1.
∗
vbus = vdc − i dc1 rd1 − i 1rc1 (2)

∗
vbus = vdc − i dc2 rd2 − i 2 rc2 (3)

Equation (4) can be derived from Eqs. (2) and (3)

i dc1 rc2 − (rd2 /rd1 ) · rc1 rd2

= + (4)
i dc2 rc1 + rd1 rd1

- +

Load
+

Fig. 1 Microgrid with conventional primary droop control

78 H. M. Karkar and I. N. Trivedi

As per conventional droop control, output current is inverse to its virtual resistance
in each converter. In Eq. (4), for analysis purpose, if capacity of two sources is same,
then rd1 = rd2 .
For different exist between distribution sources and line parameter
rd2 rc2
= (5)
rd1 rc1

If the following condition is not satisfied, current sharing error can’t be removed.
rd2 rc2
= (6)
rd1 rc1

Assumed that normally in dc microgrid system is not very large so the resistance of
cable is very small as compared to droop resistance. Since rd2  rc2 and rd1  rc1 .
So, expression of Eqs. (2) and (3) is as follows:

i dc1 rd2 + rc2 rd2

= ≈ (7)
i dc2 rd1 + rc1 rd1

The different current sources are responsible to produce the circulating current in
microgrid.

2.2 Voltage Shifting Based Primary Droop Control

The poor voltage regulation is the drawback of conventional primary droop control.
So in this method of voltage shifting based primary droop control, v is added with
the reference voltage of converter to compensate the voltage deviation.
Equivalent circuit of voltage shifting based primary droop control is as shown
in Fig. 2. v1 and v2 are two voltage shifting values. This value is added in
conventional primary droop control equation. So, Eq. (8) is expressed as follows:
∗
vi = vdc − i dci rdi + vi (8)
  ∗    
where vi = k pv + ksiv vdc − v̄dci − k pc + ksic i dci − ī dci
The voltage deviation is varying with load current. Its value does not exist beyond
its maximum limit. So, the value of droop coefficient in (9) is

vmax
rdi ≤ (9)
i dci(max)

where i dci(max) is the maximum value of converter output current

From circuit as per Fig. 2, the Eq. (10) can be expressed as
Primary and Secondary Droop Control Method … 79

- +

Fig. 2 Microgrid with voltage shifting based primary droop control

G1 G2
vbus = v∗ + v∗ (10)
G 1 + G 2 + G load dc1 G 1 + G 2 + G load dc2

where G 1 = 1/(rc1 + rd1 ), G 2 = 1/(rc2 + rd2 ), G load = 1/rload ,

∗ ∗ ∗ ∗
vdc1 = vdc + v1 , vdc2 = vdc + v2

The equation of output current (11) and (12) can be derived as

∗ ∗
G 1 [(G 2 + G load ).vdc1 − G 2 .vdc2 ]
i dc1 = (11)
G 1 + G 2 + G load
∗ ∗
G 2 [(G 1 + G load ).vdc2 − G 1 .vdc1 ]
i dc2 = (12)
G 1 + G 2 + G load

The difference of two output current in Eq. (13) is

i dc = i dc1 − i dc2 (13)

Substituting the value of i dc1 and i dc2 in Eq. (13) and i dc can be expressed as
 ∗ ∗
  ∗ ∗

2 vdc1 − vdc2 rload + (rd2 + rc2 )vdc1 − (rd1 + rc1 )vdc2
i dc = (14)
(rc1 + rc2 + rd1 + rd2 )rload + (rd2 + rc2 )(rd1 + rc1 )
di dc
The derivation of drload
is derived in Eq. (15)
 ∗ ∗

di dc (rd1 + rc1 − rd2 − rc2 ) (rd2 + rc2 )vdc1 + (rd1 + rc1 )vdc2
= (15)
drload (rc1 + rc2 + rd1 + rd2 )rload + (rd2 + rc2 )(rd1 + rc1 )
80 H. M. Karkar and I. N. Trivedi

From this above Eq. (15), it is concluded that if rc1 + rd1 = rc2 + rd2 , then the
output impedance of each converter would be equal and the effect of i dc will be
minimized by load variation [23]. So, the current sharing error is eliminated, when
equivalent output impedance is kept equal.
The drawback of this method is that here fixed value of droop resistance is used
in each converter. So total impedance of converter would be unequal. So dynamic
performance under fast-changing load current is poor.

2.3 Secondary Droop Control

In this method, two droop resistance rd1 and rd2 are not fixed but adapting together.
Here, current sharing accuracy and voltage regulation are fulfilled by adjusting slope
of droop curve and shifting droop curve. Here, PI controller is used for average cur-
rent and average droop coefficient to maintain the steady-state relationship between
rd1 , rd2 , i dc1 , and i dc2 as per the following Eqs. (16) and (17) [23].

i dc1 ≈ i dc2 (16)

rd1 + rd2
≈ r∗ (17)
2
Equation (18) is derived as per the circuit in Fig. 3
 ∗   ∗ 
vdc + v − (rd1 + rc1 )i dc1 = vdc + v − (rd2 + rc2 )i dc2 (18)

Substituting the value of i dc1 ≈ i dc2 in Eq. (19)

- +

Load
+

Fig. 3 Microgrid with secondary droop control

Primary and Secondary Droop Control Method … 81

rd1 + rc1 ≈ rd2 + rc2 (19)

As per the above Eq. (19), equivalent output impedance of each converter would
be same after implementing secondary controller with three PI controllers. Equiv-
alent output impedance keeps the current sharing accuracy with variable load. It is
calculated from Eqs. (16) and (17) as

rc2 − rc1 rc1 − rc2

rd1 ≈ r ∗ + rd2 ≈ r ∗ + (20)
2 2
These Eq. (20) are used to show the corresponding change in droop coefficient
and line impedance [23].
When equivalent line impedance of two converters is same, droop coefficient
would be equal to setting value. If line impedance of two converters is different,
droop coefficients of the converter are adjusted adaptively to make equal equivalent
output impedance.

3 Circuit Diagram of Microgrid with Control Scheme

Figure 4 shows the control circuit diagram of microgrid with primary, voltage shifting
based primary and secondary control scheme. There are two boost converters (dc to
dc) connected in parallel with common load bus. In primary control scheme, voltage
loop, current loop, and droop coefficient loop are used. Equation (1) is applied in
primary control scheme. In voltage shifting based primary control scheme, v1 and
v2 are added in each converter over conventional primary control.

DC/DC DC Bus
+ -
Converter-1

+
DG
UNIT#1 -

Secondary Control Conventional Primary Control

+ + + +
+ -
PI + + Voltage Current
Loop Loop PWM
-
Voltage - -
Shifting
Based
+ +
+ -
PI
Primary
+ Control Load
-
+ DC/DC
+ Converter-2
+ +
+ -
PI +
DG
UNIT#2 -

+ + + +
+ -
PI + + Voltage Current
Loop Loop PWM
-
- -

+ +
+ -
PI
+
- +
+
+ +
+ -
PI

Fig. 4 Circuit diagram of microgrid with control technique

82 H. M. Karkar and I. N. Trivedi

In secondary control, average value calculation of voltage, current, and droop

coefficient of neighboring converter is calculated by three PI controllers. The average
voltage controller is compensating the voltage deviation over voltage shifting based
primary control by producing the voltage shifting value. So, it regulated the output
voltage of converter. Average current and droop coefficient controllers are used for
droop curve adjusting by adaptively controlling the local droop coefficient of each
converter. By using these two current compensating and droop coefficient controllers,
controlled output impedance of two converters is same. So, we can enhance the
current sharing accuracy. In this secondary control scheme, voltage shifting and
slope adjustment is done simultaneously

4 Result and Discussion

This proposed islanding microgrid of 400 V DC bus voltage is modeled with pro-
posed primary and secondary control scheme. To conform feasibility and working of
proposed microgrid scheme, test bench is created by MATLAB/Simulink with two
parallel DG Unit of 10 kW and 8 kW capacity, respectively.
Figure 5 shows the waveform of total load current of microgrid. The load current
is increasing at 0.3 and 0.7 s due step-up load. There are different voltage V1 and
V2 of converter 1 and converter 2, respectively, in Fig. 6a. Figure 6b shows the load
voltage of microgrid in conventional primary droop control. In conventional primary
droop control, the voltage is drooped at 0.3 and 0.7 s due to increasing load. So,
in conventional primary droop control method, there is a increase in current caused
higher voltage droop.
Figure 7 shows the waveform voltage for voltage shifting based primary droop
control method. In this method, less voltage droop occurs as compared to conven-
tional method. The waveform of voltage for secondary control is shown in Fig. 8.
After applying secondary control, voltage of microgrid is regulated within 4 V even
when load is increasing at 0.3 and 0.7 s.

30
Current (A)

20

10
Load Current
0
0 0.25 0.5 0.75 1
Time (s)

Fig. 5 Load current waveform

Primary and Secondary Droop Control Method … 83

420 420
V1 V Load
400 Converter 1 V2
400
Voltage (V)

Voltage (V)
380 380
Converter 2
360 360

340 340
0.05 0.25 0.45 0.65 0.85 0.05 0.25 0.45 0.65 0.85
Time (s) Time (s)
(a) Converter Voltage (b) Load Voltage

Fig. 6 Performance study of conventional primary droop control method

420 420
Converter 1
400 400
Voltage (V)

Voltage (V)
380 Converter 2 380

360 V1 360
V2 V load
340 340
0.05 0.25 0.45 0.65 0.85 0.05 0.25 0.45 0.65 0.85
Time (S) Time (s)
(a) Converter Voltage (b) Load Voltage

Fig. 7 Performance study of voltage shifting based primary droop control method

420 420
Converter 1
400 400
Voltage (V)
Voltage (V)

Converter 2
380 380

360 360
V1 V Load
V2
340 340
0.05 0.25 0.45 0.65 0.85 0.05 0.25 0.45 0.65 0.85
Time (S)
Time (S)
(a) Converter Voltage (b) Load Voltage

Fig. 8 Performance study of secondary control method

Also, transient response during secondary control is shown in Figs. 9 and 10.
When a sudden step up a load at 2 s, there is no current sharing error of converter.
It is shown in Fig. 9a. And at same time, voltage response of converter is also good.
The voltage of each converter is maintained nearly 400 V of dc microgrid voltage.
In Fig. 10, same transient resource performance is achieved for current sharing and
voltage restoration in secondary control during suddenly step-down load.
84 H. M. Karkar and I. N. Trivedi

6 440
Converter 1 Converter 1
4 400
Current (A)

Votage (V)
Converter 2 Converter 2
2 360
I1 V1
I2 V2
0 320
0.5 1.5 2.5 3.5 0.5 1.5 2.5 3.5
Time (S) Time (S)
(a) Current sharing (b) Voltage response

Fig. 9 Performance study of secondary control for transient response during step-up load

6 440
I1 V1
I2 Converter 1 V2

4 400

Voltage (V)
Current (A)

Converter 1 Converter 2
2 360
Converter 2
0 320
0.5 1.5 2.5 3.5 0.5 1.5 2.5 3.5
Time (S) Time (S)

(a) Current sharing (b) Voltage response

Fig. 10 Performance study of secondary control for transient response during step-down load

5 Conclusion

This paper presents primary and secondary droop control strategy of islanded micro-
grid. Droop control scheme is used for proportional load sharing between parallel
converters in microgrid. In conventional primary droop control scheme, bus voltage
is degraded during increasing load current. So, voltage shifting based primary con-
trol method is used over conventional primary control. But dynamic performance
of microgrid is very poor during fast-changing load current. Also, the voltage reg-
ulation of microgrid is slightly poor. So, distributed secondary control scheme is
used with three compensating controllers over primary control. First average voltage
controller restores the average value of output voltage over proposed primary con-
troller. Second and third controller adaptively adjusts the local droop coefficient to
make same equivalent impedance. So current sharing accuracy is exactly reached.
Also, by combining both average current and droop coefficient controllers, current
sharing and voltage regulation is good under fast-changing load current. As per sim-
ulated waveform, we can conclude that by using this secondary control scheme, both
current sharing accuracy and voltage regulation is achieved by secondary control
scheme over primary control.
Primary and Secondary Droop Control Method … 85

References

1. Kakigano H, Miura Y, Ise T (2010) Low-voltage bipolar-type dc Microgrid for super high-
quality distribution. IEEE Trans Power Electron 25(12):3066–3075
2. Chen YK, Wu YC, Song CC, Chen YS (2013) Design and implementation of energy man-
agement system with fuzzy control for dc Microgrid systems. IEEE Trans Power Electron
28(4):1563–1570
3. Zhang L, Wang Y, Li H, Sun P (2012) Hierarchical coordinated control of dc Microgrid with
wind turbines. In: Proceedings of 38th annual conference on IEEE industrial electronics society
(IECON), 2012, pp 3547–3552
4. Salomonsson D, Söder L, Sannino A (2008) An adaptive control system for a dc Microgrid for
data centers. IEEE Trans Ind Appl 44(6):1910–1917
5. Guerrero JM, Vasquez JC, Matas J, de Vicu˜na LG, Castilla M (2011) Hierarchical control of
droop-controlled AC and DC microgrids—a general approach toward standardization. IEEE
Trans Ind Electron 58(1):158–172
6. Lee J, Han B, Cha H (2012) Development of hardware simulator for DC micro-grid operation
analysis. In: Proceedings of IEEE power and energy society general meeting, 2012, pp 1–8;
Khorsandi A, Ashourloo M, Mokhtari H (2014) A decentralized control method for a low-
voltage DC microgrid. IEEE Trans Energy Convers 793–801
7. Dragičević T, Lu X, Vasquez JC, Guerrero JM (2015) DC microgrids–part I: a review of control
strategies and stabilization techniques. IEEE Trans Power Electron
8. Chen D, Xu L, Yao L (2013) DC voltage variation based autonomous control of dc microgrids.
IEEE Trans Power Del 28(2):637–648
9. Che L, Shahidehpour M (2014) DC microgrids: economic operation and enhancement of
resilience by hierarchical control. IEEE Trans Smart Grid 5(5):2517–2526
10. Shafiee Q, Dragicevic T, Vasquez JC, Guerrero JM (2014) Hierarchical control for multiple
DC-microgrids clusters. IEEE Trans Energy Convers 29(4):922–933
11. Lin C, Wang P, Xiao J, Tang Y, Choo FH (2014) Implementation of hierarchical control in dc
microgrids. IEEE Trans Ind Electron 61(8):4032–4042
12. Moussaoui Z, Batarseh I, Lee H, Kennedy C (1996) An overview of the control scheme for
distributed power systems. In: Proceedings of Southcon conference, 1996, pp 584–591
13. Dragicevic T, Guerrero J, Vasquez J, Skrlec D (2014) Supervisory control of an adaptive-
droop regulated dc microgrid with battery management capability. IEEE Trans Power Electron
29(2):695–706
14. Huang P, Xiao W, Moursi MS (2013) A practical load sharing control strategy for DC microgrids
and DC supplied houses. In: Proceedings of 39th annual conference IEEE industrial electronics
society 2013, pp 7122–7126
15. Meng L, Dragicevic T, Guerrero JM, Vasquez JC (2014) Dynamic consensus algorithm based
distributed global efficiency optimization of a droop-controlled DC microgrid. In: IEEE inter-
national requirements engineering conference, pp 1276–1283
16. Beerten J, Belmans R (2013) Analysis of power sharing and voltage deviations in droop-
controlled DC grids. IEEE Trans Power Syst 28(4):4588–4597
17. Anand S, Fernandes BG, Guerrero JM (2013) Distributed control to ensure proportional load
sharing and improve voltage regulation in low voltage DC microgrids. IEEE Trans Power
Electron 28(4):1900–1913
18. Lu X, Guerrero JM, Sun K, Vasquez J (2014) An improved droop control method for dc micro-
grids based on low bandwidth communication with dc bus voltage restoration and enhanced
current sharing accuracy. IEEE Trans Power Electron 29(4):1800–1812
19. Nasirian V, Moayedi S, Davoudi A, Lewis FL (2015) Distributed cooperative control of DC
microgrids. IEEE Trans Power Electron 30(4):2288–2303
20. Nasirian V, Davoudi A, Lewis FL, Guerrero JM (2014) Distributed adaptive droop control for
DC distribution systems. IEEE Trans Energy Convers 29(4):944–956
21. Shafiee Q, Guerrero JM, Vasquez JC (2014) Distributed secondary control for islanded micro-
grids—a novel approach. IEEE Trans Power Electron 29(2):1018–1031
86 H. M. Karkar and I. N. Trivedi

22. Yu X, She X, Huang A (2013) Hierarchical power management for DC microgrid in islanding
mode and solid-state transformer enabled mode. In: Proceedings of 39th annual conference of
the IEEE industrial electronics society, 2013, pp 1656–1661
23. Wang P, Lu X, Yang X, Wang W, Xu D (2015) An improved distributed secondary control
method for DC microgrids with enhanced dynamic current sharing performance. IEEE Trans
Power Electron
Equal Load Sharing in DC Microgrid
Using Line Resistance Estimation

Shivam Chaturvedi and Deepak M. Fulwani

Abstract The objective of this work is to achieve equal load sharing among parallel-
connected converters by estimating the line resistance. Difference in line resistance
leads to unequal power sharing in parallel-connected converters. Unequal load shar-
ing is eliminated by droop control method. However, voltage regulation at the DC
bus puts a limit on value of droop constant. To eliminate unequal power sharing esti-
mation of the value of resistance of interconnecting cable is done by local voltage and
current values of a converter. The estimated value is used to adjust the voltage ref-
erence so as to compensate the voltage drop on the connecting cable. This improves
voltage at the DC bus and also leads to equal load sharing among converters. The
proposed method has been validated through simulation of three parallel-connected
boost converters.

Keywords DC microgrid · Load sharing · Droop control · Resistance estimation

1 Introduction

Depletion of conventional sources of energy and increasing energy demand is leading

to increased installations of DC microgrids and DC nanogrids all over the world. DC
microgrids are made by interconnection of several DC–DC Boost converters or Buck
converters. A typical configuration of DC microgrid is shown in Fig. 1. The loads
are connected directly to the DC microgrid or by a DC–Dc converter so as to draw
constant power (CPLs) from microgrid. Interconnection of DC–Dc converters and
loads may lead to unstability of microgrid [1–3]. It is desirable that the voltage DC
microgrid should be constant irrespective of the load changes. Also, the converters
at the source end should share equal load.

S. Chaturvedi (B) · D. M. Fulwani

Indian Institute of Technology Jodhpur, Jodhpur 342037, India
e-mail: [email protected]
D. M. Fulwani
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 87

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_7
88 S. Chaturvedi and D. M. Fulwani

Fig. 1 A typical DC microgrid

Droop control is widely used control method that ensures equal power sharing,
at the cost of voltage regulation for higher values of droop constants. The control of
DC microgrid can be central in nature [4, 5]. In central controller receives informa-
tion from other distributed controllers and provides voltage and current references
to them. This method has a drawback that there can be a single point of failure. To
overcome this drawback, various distributed communication-based methods have
been proposed [6–9]. Various schemes such as cooperative control [10], dynamic
consensus control is proposed in [11]. However, such communication-based sys-
tems are complex and it may lead to cost increments of a microgrid. The main cause
of unequal power sharing is the cable resistance. Estimation of cable resistance is
helpful to achieve equal loading. In [12] authors propose an active resistance identifi-
cation method, however, it requires DC bus voltage as well. In some cases, resistance
estimating devices are used; however, this adds to increase in number of components
[13]. All the above methods either require an extra component thereby increasing
device count or require communication requirements so that each converter gets to
know the state of every other converters. However, if the rating of converters is same
then equal load sharing can be achieved without communication. In this paper, a
line resistance estimation methodology is proposed such that the voltage reference
is increased by an amount equal to the voltage drop in line resistance which in turn
results in equal load sharing and good voltage regulation.
In Sect. 2, the method to estimate line resistance using local parameters is dis-
cussed. The designing of controllers for voltage regulation is discussed in Sect. 3.
Equal Load Sharing in DC Microgrid Using Line Resistance Estimation 89

(a) Boost converter control (b) Open loop characteristic

Fig. 2 Converter physical diagram and characteristic bode plot

The small signal analysis for a boost converter DC bus is done in Sect. 4. Section 5
consists of validation of proposed method by simulation of three parallel-connected
boost converters. The system performance is evaluated by increasing and decreasing
the load demand.

2 Estimation of Line Resistance

The DC bus can be considered to have high capacitance, due to parallel-connected

converters, thereby maintaining the voltage of DC bus, irrespective of instantaneous
changes in load or source. Resistance of connecting line can be estimated by finding
out the ratio of small variation in voltage to small variation in current. This value of
estimated resistance is used to increase the reference value, thereby compensating
the voltage loss in the lines connecting the converter to the bus. Once the resistance
is estimated then equal load sharing could be obtained. Analysis of the proposed
method is done in the section below. The proposed configuration that can be used to
estimate the value is shown in Fig. 2.

3 Control Methodology

The Proportional Integral (PI) control has been used to control the converters. Two-
loop current and voltage control is implemented Fig. 3. In such type of control, outer
voltage controller compares the output voltage value to the desired reference value
and provides reference current value to the inner current control [15]. Inner control
compared the reference value to the inductor current value and accordingly generates
the signals to be compared by the PWM block which in turn, provides the on–off
signals to the converter’s switch at a required duty cycle.
90 S. Chaturvedi and D. M. Fulwani

Fig. 3 Two-loop voltage and current control

(a) Frequency Response: Vo /il (b) Frequency Response: IL /d

Fig. 4 Closed loop frequency responses

To design a two-loop control, first the frequency response of Vo /d is analysed

Fig. 4. From this response we obtain the gain crossover frequency. The current control
is designed to have 1/10th gain crossover frequency of Vo /d and voltage controller
is designed to have 1/10th the gain crossover frequency of Vo /il frequency response.
Here, Vo is converter’s output voltage, I L is inductor current and d is duty cycle.
The phase margin should be atleast π /3 at the gain crossover frequency. This
could be observed from above waveforms. The blue waveform is response of plant
without controller and the orange waveform is response of the plant with controller.
Hence, the current control is made to operate faster than the voltage controller. Also,
the controllers have the dynamics faster than the plant. The frequency response
of output voltage variation with the duty cycle is shown in Fig. 2b. The frequency
response of output voltage variation with the inductor current is shown in Fig. 4a.
The frequency response of inductor current variation with the duty cycle is shown in
Fig. 4b.
Equal Load Sharing in DC Microgrid Using Line Resistance Estimation 91

(a) Boost converter: Switch on (b) Boost converter: Switch off

Fig. 5 Averaged modelling

4 Analysis

The averaged model of boost converter connected to a DC bus is done in two time
instances as shown in Fig. 5. First, when the switch is on and the other is when the
switch is off. Then by small signal analysis the transfer function of the change of
output voltage with respect to the duty cycle, and output current with respect to the
duty cycle is obtained.
When the switch S is on:
di L dvo Vo
L = Vs C = (1)
dt dt RLine + Ct ||RLoad

When the switch is off :

di L dvo Vo
L = Vs − Vo C = iL − (2)
dt dt RLine + Ct ||RLoad

In above equations, Ct is the capacitor used to maintain the DC bus voltage, R Line
is the line resistance, R Load is the equivalent of all loads, Vs is the source voltage and
Vo is the voltage at capacitor C terminals, d is the duty cycle. Averaging the state
over one switching cycle

di L dvo Vo
L = Vs − (1 − d)Vo C = (1 − d)i L − (3)
dt dt RLine + Ct ||RLoad

Introducing small perturbations : i L =I L + iˆL , vo =Vo + vˆo , Vs =Vs + vˆs , d=D + d̂

and write above equation in matrix form :
       
sL (1 − D) iˆL (s) Vo 1
(1 − D) sC + 1 = d̂(s) + vˆ (s)
RLoad
RLine + sCt R vˆo (s) IL 0 s
Load +1

Let 1
= sC + 1
R for simplicity, also neglect small variations in input
Req RLine + sCt RLoad +1
Load
voltage Vs ;
92 S. Chaturvedi and D. M. Fulwani
    
iˆL (s) −(1 − D) Vo
1
1
= Req d̂(s) (4)
vˆo (s) sL
Req
− (1 − D)2 −(1 − D) sL IL

From (4) the transfer functions with respect to duty cycle variations irrespective
of other variations can be found out as under:

iˆL (s)
Vo
Req
− (1 − D)I L iˆc (s) vˆo (s)
= = sC (5)
d̂(s) sL
Req
− (1 − D)2 d̂(s) d̂(s)

iˆc (s) s 2 C L I L − (1 − D)sVo C iˆo (s) iˆL (s) iˆc (s)

= = − (6)
d̂(s) sL
Req
− (1 − D)2 d̂(s) d̂(s) d̂(s)

Substitute values and simplify (6) :

iˆo (s)
Vo
Req
− (1 − D)I L + (1 − D)sVo C − s 2 C L I L
= (7)
d̂(s) sL
Req
− (1 − D)2

From (5) and (7) :

vˆo (s) vˆo (s) d̂(s)
= .
iˆo (s) d̂(s) iˆo (s)

vˆo (s) s L I L − (1 − D)Vo

= (8)
iˆo (s) Vo
Req
− (1 − D)I L + (1 − D)sVo C − s 2 C L I L

ˆ
Hence, the transfer functions vd̂(s)
ˆo (s) i o (s)
, d̂(s) and viˆˆo(s)
(s)
have been derived. These will be
o
used for analysis of behaviour of small change in voltage to small change in output
current.A Dc bus can be considered to be a DC capacitor of large value which keeps
the voltage at DC bus constant irrespective of small load changes or small transients
in the converters connected to it. This assumption is necessary for further analysis.
From Fig. 6b change in bus voltage with respect to output voltage of a converter,
and change in output voltage with respect to change in reference voltage can be found
out by voltage and current division rules.

ˆ (s)
vbus RLoad
=
vˆo (s) RLine + RLoad + sCbus RLoad RLine

iˆo (s) sCbus RLoad + 1

=
vˆo (s) RLine + RLoad + sCbus RLoad RLine

vbus
ˆ (s) vbus
ˆ (s) vˆo (s) iˆo (s) iˆo (s) vˆo (s)
= . = . (9)
vref
ˆ (s) vˆo (s) vref
ˆ (s) vref
ˆ (s) vˆo (s) vref
ˆ (s)
Equal Load Sharing in DC Microgrid Using Line Resistance Estimation 93

vˆo (s) ˆ
(a) Boost converter connected to a DC bus (b) Step response vref
, iˆo (s)
ˆ (s) vref (s)
and vvref
ˆ (s)
bus
ˆ (s)
with high Cbus

Fig. 6 Physical interfacing to DC bus and step responses

The transfer function vvref ˆo (s)

ˆ (s)
can be obtained by block reduction of two loop control
voltage and current technique shown in Fig. 3. The response of transfer functions
vbus
ˆ (s) ˆ
vref
ˆ (s)
, vvref
ˆo (s)
, iˆo (s)
ˆ (s) vref (s)
is shown in Fig. 6b.
By applying Kirchoff’s Voltage law :

Vo (t) − Io (t)RLine = Vbus (t) (10)

After small time interval δt :

Vo (t + δt) − Io (t + δt)RLine = Vbus (t + δt) (11)

Vo (t + δt) = Vo + vo Io (t + δt) = Io + i o (12)

However, due to high capacitance of DC bus Vbus does not changes instantaneously
compared to converter’s voltage and current,

Vbus (t + δt) = Vbus

Substract (11) and (10) we obtain :

vo − i o RLine = 0

Therefore,
vo
RLine =
i o

Hence, if small changes in output voltage and current of a converter connected to a

DC line is known, then the resistance of connecting line can be found out.
94 S. Chaturvedi and D. M. Fulwani

Table 1 Simulation parameters

Parameters Values
Vin 150 V
Vout 380 V
L, C, Cbus 100 μH, 50 μF, 2000 μH
RLine1 , RLine2 , RLine3 0.5 , 1 , 1.5 
RLoad1 , RLoad1 50 , 25 C

5 Validation of Proposed Methodology

To validate the above analysis, a simulation was done using Simulink in Matlab 2013
software. The simulation parameters are given in Table 1. Three boost converters were
connected in parallel as shown in Fig. 7. Each converter is controlled by PI controller.
All the converters have the same rating and the feed a resistive load. As each converter
has same rating there should be equal power sharing among them. However, this is
not the case. The converter’s share of load is affected by the resistance of connecting
cable. Converter having higher cable resistance shares least load while the one having
lowest cable resistance shares maximum load.
The waveforms of current shows that initially the power was not equally shared.
Initially, the resistance estimation method was not active upto 0.25 s. The load is
doubled at 0.12 s and then reduced back to previous value at 0.23 s. Uptill 0.25 s

Fig. 7 Simulated three parallel-connected boost converters with resistance estimation

Equal Load Sharing in DC Microgrid Using Line Resistance Estimation 95

(a) DC bus voltage (b) Current sharing in per unit

(c) Estimated line resistance (d) Filtered estimated value

Fig. 8 Simulation results

the effect of cable resistance on the load sharing can be observed see Fig. 8b. The
estimation block was activated after 0.25 s. The estimation block induces small
step perturbations in reference voltage and calculated the ratio of voltage to current
change. Thereafter, it could be observed that the load is shared equally among all
converters. The load is again doubled at 0.32 s and then reduced to original value at
0.42 s. Even if the load is increased or decreased, the converters continue to share
equal loads Fig. 8d. The overall voltage of the DC bus also increases at 0.25 s as
the drops in the connecting cables is compensated by increasing the reference of
the converters, respectively, Fig. 8a. The estimated values are close to the actual line
resistances Fig. 8b.

6 Conclusion

It can be concluded from simulation results that load demand is shared equally after
the estimation methodology is implemented. Hence, small perturbations in voltage
and current has been used to estimate line resistance value. The voltage reference
is adjusted to compensate voltage drop in connecting cable, thereby resulting in
improved voltage regulation. The estimated values are found to be very close to the
actual values of line resistances. Hence, the resistance estimation methodology can
be used to implement equal load sharing and to improve the DC bus voltage.
96 S. Chaturvedi and D. M. Fulwani

References

1. Kwasinski A, Onwuchekwa CN, Member S (2011) Dynamic behavior and stabilization of DC

micro-grids with instantaneous constant-power loads 26(3):822–834
2. Singh S, Fulwani D, Kumar V (2015) Robust sliding-mode control of dc/dc boost converter
feeding a constant power load. IET Power Electron 8(7):1230–1237
3. Huddy SR, Skufca JD (2013) Amplitude death solutions for stabilization of DC microgrids
with instantaneous constant-power loads. IEEE Trans Power Electron 28(1):247, 253
4. Bidram A, Davoudi A (2012) Hierarchical structure of microgrids control system. IEEE Trans
Smart Grid 3(4):1963–1976
5. Tsikalakis AG, Hatziargyriou ND (2008) Centralized control for optimizing microgrids oper-
ation. IEEE Trans Energy Convers 23(1):241–248
6. Tucci M, Riverso S, Vasquez JC, Guerrero JM, Ferrari-Trecate G (2015) A decentralized
scalable approach to voltage control of DC islanded microgrids. Dipartimento di Ingegneria
Industriale edell’Informazione, Universit‘a degli Studi di Pavia, Pavia. Technical report, Italy
7. Anand S, Fernandes B (2012) Reduced order model and stability analysis of low voltage DC
microgrid. IEEE Trans Ind Electron 1
8. Zhao J, Dorfler F (2014) Distributed control and optimization in DC microgrids
9. Lu X, Guerrero JM, Sun K, Vasquez JC (2014) An improved droop control method for DC
microgrids based on low bandwidth communication with DC bus voltage restoration and
enhanced current sharing accuracy. IEEE Trans Power Electron 29(4):1800–1812
10. Moayedi S, Nasirian V, Lewis FL et al (2015) Team-oriented load sharing in parallel DC-DC
converters. IEEE Trans Ind Appl 51(1):479–490
11. Nasirian V, Moayedi S, Davoudi A, Lewis F (2015) Distributed cooperative control of DC
microgrids. IEEE Trans Power Electr 30:2288–2303
12. Liu C, Zhao J, Wang S, Lu W, Qu K (2018) Active identification method for line resistance in
DC microgrid based on single pulse injection. IEEE Trans Power Electron 33(7):5561–5564
13. Khorsandi A, Ashourloo M, Mokhtari H, Iravani R (2016) Automatic droop control for a low
voltage DC microgrid. IET Gener Transm Distrib 10:41–47
14. Sira-Ramirez H, Perez-Moreno R, Ortega R, Garcia-Esteban M (1997) Passivity-based con-
trollers for the stabilization of DC-to-DC power converters. Automatica 33:499–513
15. Prasanna CUR, Rathore AK (2014) Small signal analysis of current-fed full-bridge isolated
DC/DC converter with active-clamp and control system implementation using PSoC. IEEE
Trans Ind Electron 61(3):1253–1261
Enhanced Control and Dynamic Power
Management of Li-Ion Battery
and Ultracapacitor for Hybrid Energy
Storage System

Pranav Mohod, Pradyumn Chaturvedi and Shubham Gajbhiye

Abstract A major problem to tackle in applications of the electric field is energy

storage. Various systems have been proposed for energy storage based upon power
and energy densities. This paper deals with one of the mixed energy storage systems
known as hybrid energy storage system (HESS). The hybrid energy storage system
consists of a battery as a supply in steady state and an ultracapacitor as supply during
transient state. To make optimum use of a hybrid energy storage system, there must
be power sharing from source to load as well as between the battery and the ultra-
capacitor. In this paper, energy and power management between the battery and the
ultracapacitor in hybrid energy storage system with a DC-to-DC bidirectional con-
verter consisting of two-switch topology and a control strategy based upon charging
and discharging of energy storing devices is presented. Typical configuration used for
the battery and the ultracapacitor along with different power management scenarios
are studied. The performance of the system is validated by simulation results.

Keywords Battery · Hybrid energy storage system (HESS) · Power management ·

Bidirectional converter · Ultracapacitor (UC)

1 Introduction

The increase in demand for electricity has led to more generation causing more
impact on the environment as the conventional thermal power plants have high carbon
emission rates resulting in carbon footprint. The major cause of carbon emissions are
due to generation of electricity and passenger transport system. This impact has been
driving the society to make use of renewable energy sources for power generation and
to transcend in the use of electric vehicles (EV) rather than conventional diesel fuel

P. Mohod · P. Chaturvedi (B) · S. Gajbhiye

Department of Electrical Engineering, Visvesvaraya National Institute of Technology, Nagpur,
India
e-mail: [email protected]
P. Mohod
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 97

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_8
98 P. Mohod et al.

vehicles. These problems can be solved by using solar energy as it does not emit any
carbon during operation but has carbon footprint only during the construction and
maintenance phase. Use of both solar power and electric vehicles need energy storage
system to operate without any interruption. Earlier, only the battery was used as an
energy storage device terming it as Battery Energy Storage System (BESS) [1, 2].
There are certain problems associated with BESS such as incapability during transient
operating conditions. In solar power, the fluctuations are caused by temperature
variations, intermittent solar radiation, and partial shading effect whereas in electric
vehicles, acceleration and deceleration cause power fluctuations.
These conditions cannot be handled by a battery alone, hence, the combined
energy storage of a battery and an ultracapacitor is proposed known as hybrid energy
storage system [3, 4]. Ultracapacitors are employed with batteries due to their quick
response to fluctuating demands as they have high power density; however, the bat-
tery is having high energy density. Figure 1 shows the equivalent circuit of the Li-ion
battery [5] used in the system. Ro is the ohmic resistance, Rp and Cp are the polar-
ization resistance and the capacitance, respectively, for transient response during
charging and discharging. V oc is the open circuit voltage. Figure 2 shows the equiv-
alent circuit diagram of an ultracapacitor with V uc as the ultracapacitor voltage and
RESR as its equivalent series resistance. Table 1 shows the Li-ion battery parameters,
similarly, Table 2 shows the ultracapacitor parameters.
Various control technique methods were used for energy management of hybrid
energy storage system. Fuzzy logic controller technique used for energy management
of hybrid energy storage system had an advantage of an unprecise modeling of the
plant depending on the designer’s knowledge. This approach gave a satisfactory
solution but not the optimal results for every condition [6]; moreover, it did not take

Fig. 1 Equivalent circuit diagram of Li-ion battery

Enhanced Control and Dynamic Power Management of Li-Ion Battery … 99

Fig. 2 Equivalent circuit diagram of ultracapacitor

Table 1 Lithium-ion battery

Battery parameters Value
specifications
Nominal voltage (V) 26.40
Rated capacity (Ah) 6.60
Initial state-of-charge (%) 100
Fully charged voltage (V) 30.72
Number of cells in parallel 6

Table 2 Ultracapacitor
Parameters Value
specifications
Rated voltage (V) 16
Initial voltage (V) 16
Operating temperature (Celsius) 25
Rated capacitance (F) 500

care of charging rate limitations of the battery and lacked the ability to satisfy the
pulsed peak loads. To completely regulate the power injection in the system based
on the load requirement, a bidirectional power electronic converter has to be used.
A buck–boost converter can do both step-up operation and step-down operation,
which eliminate the use of two different dedicated converters, buck converter and
boost converter, to be employed in the system; but a buck–boost converter possesses
the problem of right-half plane zero and higher voltage stress on the switch, hence
a two-switch buck–boost converter has been proposed in [7, 8]. Even though it
overcomes the problems faced in single-switch buck–boost converter, it has higher
100 P. Mohod et al.

power loss than a single-switch buck–boost converter owing to extra elements used
in the circuit. This two-switch buck–boost bidirectional converter as shown in Fig. 3,
used in this paper, contains only two MOSFETs with body diodes that are known as
active synchronous rectification.
Active synchronous rectification improves the efficiency of rectification by replac-
ing the diode with power MOSFET or power BJT in a power electronic converter.
The low on-resistance property of MOSFET reduces ohmic losses compared to diode
rectifier. Another option is to use the Schottky diode [9, 10] which has 0.3 V drop
across it unlike 0.7–1 V drop for diodes but even they are lossy when voltage is low
and current is high in circuit 1; such an example is the two-switch buck boost bidi-
rectional converter used for ultracapacitor which has to take higher currents, hence
this reduces the possibility of using Schottky diodes in the circuit for reducing the
losses.

Fig. 3 Circuit diagram of hybrid energy storage system

Enhanced Control and Dynamic Power Management of Li-Ion Battery … 101

2 System Description

2.1 Configuration

The topology used in this paper for two-switch buck–boost converter [8, 11] has two
MOSFETs with body diodes, and no extra switch is considered apart from those.
The system configuration used in this paper as shown in Fig. 4 consists of a battery
with its corresponding two-switch buck–boost bidirectional converter having switch
S1 and S2 with their body diodes. The value of the inductor used is L_batt = 1.2
mH. An ultracapacitor is connected in parallel [12, 13] with the battery having its
own two-switch buck boost bidirectional converter using switch S3 and S4 with their
body diodes with inductor value L_uc = 2 mH. These two energy storing elements,
i.e., the battery and the ultracapacitor in parallel form a hybrid energy storage system
supplying pulsed peak load that can be an electric vehicle which causes fluctuations
during acceleration and deceleration or it may be a DC grid which has fluctuation
from source of PV panel. The load has a characteristic to supply back to source that
is regenerative braking in case of electric vehicles and surplus power to DC grid in
case of supply from solar PV panels.
Furthermore, the behavior of HESS for stepped change load requirement is stud-
ied. The specifications of energy storing elements are the same as in Tables 1 and
2. The resistive load is switched at an interval of 1 s so that the power demand is
increased gradually for a cycle period of 3 s. Load requirement cycle is repeated from
t = 3 s, hence there is a drop in the power requirement at this instant, thus covering
both the requirements of increase in load and decrease in load power.

2.2 Control Strategy

Figure 4 shows the control strategy for the battery and ultracapacitor, respectively.
Current passing through the converter is sensed and is compared with the reference
current which is decided according to the load requirements. The error from the
comparison is passed to PI controller whose output is a signal which is further
compared with a triangular wave to obtain a duty ratio used for firing the switches.
The PI controller parameters are obtained by the heuristic approach. The values are
K p = 1.5 and K i = 1. The duty ratio is given to switches S1 and S3 when power
required by the load is more than the power obtained from source, i.e., Pload > Psource .

Fig. 4 Block diagram of control strategy for ultracapacitor and battery

102 P. Mohod et al.

This is during the discharge cycle of energy storing elements whereas the duty ratio
is given to switches S2 and S4 when the power required by the load is less than the
power obtained from source, i.e., Pload < Psource . This is during the charging cycle of
energy storing elements generally when the power is given back through regenerative
braking/surplus power from PV panel to DC grid diverted to energy storing devices.

2.3 Battery and Ultracapacitor Sizing

To select the proper UC size [14], peak current is an important parameter to be

considered

E = V.I.t (1)

where
V is the grid voltage for DC grid application (48 V)
I is the load current (35 A)
t is the operating time for UC (10 s)
The stored energy in the ultracapacitor can be calculated as follows:

E = 0.5C V 2 (2)

The maximum exchanged energy of UC is calculated as follows:

 
E max−ex = 0.5C Vmax
2
−Vmin
2
(3)

where V max = 17 V and V min = 15 V are maximum and minimum allowable voltages
of UC
E max-ex is the maximum allowable exchanged energy
Equating (1) and (3), we get

C = 480 F (4)

As we have not considered the losses due to ESR, the approximate value chosen
for C = 500 F.
Justification for battery specification chosen

I = 35 A(maximum load current)

V = 26.4 V(Nominal voltage)
Enhanced Control and Dynamic Power Management of Li-Ion Battery … 103

Now,

A − h × number of cells = Total current output for 1 hour:

(5)
6.6 × 6 = 39.6 A
Total current output for 1 hr × Nominal voltage = Power output:
(6)
39.6 × 26.4 = 1045.44 W

The power output is just above the maximum steady-state power required by the
load (1000 W).

3 Simulation Results

The simulation for the configuration along with the proposed control strategy used
was carried out in MATLAB. The pulsed peak load considered can be an EV during
acceleration and deceleration or the fluctuating power from a solar array. Power
required by the load is chosen such that the HESS has to meet the peak load at the
beginning, i.e., at t = 0 s. A steady load requirement is taken care of during t = 3 s
to t = 6 s and finally after t = 8 s, there is no power required by the load to see how
much time the energy-storing devices of HESS take to settle down and how they
share the power between them when the system has to settle down.
Figure 5a shows the power obtained from HESS and Fig. 5b shows the power
required by the load. This shows that the wave shape obtained from HESS is exactly
the same as required with some ripples as discussed further. Figure 6 shows that

Fig. 5 Comparison between a output power obtained from hybrid energy storage system and
b output power required by the pulsed peak load
104 P. Mohod et al.

Fig. 6 Ultracapacitor and

battery power of HESS

the entire peak load required initially which is supplied by the UC alone gradually
transferring the load to the Li-ion battery until the power required by the load is in
steady state thereafter only the battery handles the entire load requirement. This is
due to the inherent property of the UC to discharge quickly owing to its high power
density. Whereas the battery discharges slowly and in the meantime, the pulsed power
requirement is thereby met by only UC. After t = 6 s, the total power required can
be supplied by the battery alone but the UC has to act as it is a sudden change
at this instant; hence, the UC starts charging by taking power from the battery.
Simultaneously, the battery power recedes gradually and linearly according to the
load requirement. The rate of charging of UC increases till t = 8 s after this instant
UC is still charging but the rate of charge starts decreasing as the power is no more
required by the load. This is that span of time where the battery power is taken by the
UC so that power from the battery does not fall to zero immediately which causes
stress on the battery and thereby reduces the battery life.
Figure 7 shows that the output voltage is regulated at 48 V during the entire
simulation period; Fig. 8 shows the state of charge (SOC) of the Li-ion battery used
for HESS. The SOC of the battery drops from 100 to 99.1% smoothly in a time span
of 8 s. SOC of the battery falls considerably from t = 3 s to t = 6 s; this is because
the power required by the load is not fluctuating which causes the battery to act alone
thereby reducing its SOC. When the power required by the load is in steady state,

Fig. 7 Output voltage

Enhanced Control and Dynamic Power Management of Li-Ion Battery … 105

Fig. 8 SOC of Li-ion

battery

Fig. 9 Current of Li-ion

battery

only the Li-ion battery responds by supplying the entire load requirement. Figure 9
shows that the battery current takes short time after t = 3 s to a reach steady-state
value of 35 A; this is because the UC ensures that there is no sudden change in the
power immediately after t = 3 s. When the power required by the load is zero from t
= 8 s to t = 10 s, the current and SOC of the battery settle down. Figure 10 shows the
output result obtained for the sudden switching in resistive load at regular intervals
for a cycle period of 3 s with a simulation time of 5 s. Figure 11 shows that the entire
peak load required initially is supplied by the UC alone gradually transferring the
load to the Li-ion battery until the power required by the load is in steady state; only

Fig. 10 Comparison between a output power obtained from hybrid energy storage system and
b output power required by the load
106 P. Mohod et al.

Fig. 11 a Ultracapacitor
power and b battery power

thereafter the battery is able to handle the entire load requirement which can be seen
from the interval t = 0.5 s to t = 1 s. At t = 1 s, there is a sudden change in the load
demand causing a peak pulse. This peak pulse is entirely met by the UC but even
after that, the UC takes some time to settle down which is the cause of the ripples in
the results obtained. Only when the UC settles down, the complete load is handled
by the battery; meanwhile, the battery supports the UC to meet the load requirement
accordingly, hence it has some slope. This process repeats at t = 2 s After t = 3 s, the
total power required can be supplied by the battery alone but the UC has to act as it is
a sudden change at this instant, hence, the UC starts charging by taking power from
the battery. Simultaneously, the battery power recedes linearly according to the load
requirement; this is due to the cycle repetition and drop in load requirement. The
rate of charging of UC increases till t = 3.1 s; after this, instant UC is still charging
but the rate of charge starts decreasing as the power output by the HESS takes some
time to act so that it can be met by the battery alone. This is that span of time where
the battery power is taken by the UC so that power from the battery does not fall to
zero immediately which causes stress on the battery and thereby reduces the battery
life. Figure 12 shows that the output voltage is regulated at 48 V during the entire
simulation period with some spikes during sudden switching.

Fig. 12 Regulated output

voltage
Enhanced Control and Dynamic Power Management of Li-Ion Battery … 107

4 Conclusion

This paper presents a configuration of a hybrid energy storage system using DC-to-
DC bidirectional converters of two switches, also known as synchronous rectification,
with a control strategy based upon the charging and discharging requirements of
the energy storing devices. The proposed control strategy can regulate the output
voltage along with the power requirements of load by sharing the power between the
energy storing elements, i.e., the battery and ultracapacitor. In addition, the proposed
control strategy with the corresponding configuration efficiently handles the pulsed
peak loads with the help of the ultracapacitor which reduces the stress on the battery
thereby increasing the battery life and hence the reliability of hybrid energy storage
system increases. It lowers the ratings of energy storing elements used in the hybrid
energy storage system thereby reducing the cost, weight, and space occupied by
the battery and the ultracapacitor altogether. The presented system is best suited
for electric vehicle applications with the optimized operation of the battery and
ultracapacitor.

References

1. Shimizukawa J, Iba K, Hida Y, Yokoyama R (2010) Mitigation of intermittency of wind power

generation using battery energy storage system. In: Proceedings of 45th international confer-
ence universities power engineering, pp 1–4
2. Serban I, Marinescu C (2014) Battery energy storage system for frequency support in microgrids
and with enhanced control features for uninterruptible supply of local loads. Int J Electr Power
Energy Syst 54:432–441
3. Miller JR (1995) Battery-capacitor power source for digital communication applications: simu-
lations using advanced electrochemical capacitors. In: Proceedings of electrochemical society,
vol 95–29, Oct 1995, pp 246–254
4. Brandhorst Jr HW, Chen Z (2001) Achieving a high pulse power system through engineer-
ing the battery-capacitor combination. In: Proceedings of 16th annual battery conference on
applications and advances, Jan 2001, pp 153–156
5. Mesbahi T, Khenfri F, Rizoug N et al (2016) Dynamical modeling of Li-ion batteries for electric
vehicle applications based on hybrid particle swarm—Nelder-Mead (PSO–NM) optimization
algorithm. Electr Power Syst Res 131:195–204
6. Ferreira AA, Pomilio JA, Spiazzi G, Silva LA (2008) Energy management fuzzy logic super-
visory for electric vehicle power supplies system. Trans Power Electron 23:107–115
7. Ahmad AA, Abrishamifar A (2007) A simple current mode controller for two switches buck-
boost converter for fuel cells. In: Proceedings of IEEE EPC, Oct 2007, pp 363–366
8. Son H-S, Kim JK, Lee J-B, Moon S-S, Park J-H, Lee S-H (2017) A new buck-boost con-
verter with low voltage stress and reduced conducting components. IEEE Trans Ind Electron
64(9):7030–7038
9. Emadi A (2009) Integrated power electronic converters and digital control. CRC Press, pp
145–146
10. Abdullah M, Yatim A, Tan C, Samosir A (2013) Control of a bidirectional converter to interface
ultracapacitor with renewable energy sources. In: Proceedings of IEEE international conference
on industrial technology, pp 673–678
108 P. Mohod et al.

11. Zhang J, Lai J-S, Yu W (2008) Bidirectional dc-dc converter modeling and unified controller
with digital implementation. In: Proceedings of 23rd annual IEEE application power electronics
conference Expo, pp 1747–1753
12. Ortuzar M, Moreno J, Dixon J (2007) Ultracapacitor-based auxiliary energy system for an
electric vehicle: implementation and evaluation. IEEE Trans Ind Electron 54(4):2147–2156
13. Farhadi M, Mohammed O (2015) Performance enhancement of actively controlled hybrid dc
microgrid incorporating pulsed load. IEEE Trans Ind Appl 51(5):3570–3578
14. Fenol SD, Caluyo FS, Lorenzo JL (2017) Simulation and modeling of charging and discharging
of supercapacitors. In: International conference on circuits, system and simulation
Active and Reactive Power Control
of Photovoltaic Power Plant Under
Normal and Abnormal Grid Conditions

Veera Joshi, Bhinal Mehta and Siddharth Joshi

Abstract This paper presents the implementation of adaptive control strategies in

photovoltaic (PV) power plants for control of active and reactive power under various
conditions. The enhanced performance of the PV system under different environmen-
tal conditions and under distorted grid conditions is shown. The PV system behaves
according to the stringent grid code and does not get disconnected during a few
unhealthy grid conditions. The active power injection to the grid is tracked and it
delivers the maximum power according to Maximum Power Point Tracking (MPPT)
following the input conditions. The reactive power injection is also controlled and is
set according to the need of different grid conditions. The performance analysis of
three phase solar-PV system is carried out in the Matlab/Simulink platform.

Keywords Active power control · Reactive power control · MPPT · DC link

regulation

1 Introduction

The worldwide acceptance of the renewable energy sources (RES) like solar and
wind due to its unlimited availability is well known and is increasing day-by-day.
The severe impact of the greenhouse gas emissions is noticed highly in recent times
and looking at the concern of global warming the use of renewable energy sources
is inevitable. Hence the focus is shifting towards harnessing more energy from the
renewable energy sources particularly wind and solar [1]. The energy demand is
increasing due to the growth in developing countries, modernization, industrialization

V. Joshi (B) · B. Mehta · S. Joshi

Electrical Engineering Department, Pandit Deendayal Petroleum University, Gandhinagar
382007, Gujarat, India
e-mail: [email protected]
B. Mehta
e-mail: [email protected]
S. Joshi
e-mail: [email protected]
© Springer Nature Singapore Pte Ltd. 2020 109
A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_9
110 V. Joshi et al.

and improving life style of people. There is urgent need to create the balance in supply
and demand with more and more penetration of RES looking at the limited stock of
oil, coal and gas as well as environmental aspects.
Looking at the above aspects the huge capacity solar parks is upcoming. The
reliability of solar PV system is at least around 15–20 years. The grid integration
of PV has many advantages including the cost of battery and energy management.
As the penetration of renewable energy is increasing the more and more stringent
grid codes are enforcing the technology advancement and optimized performance
of RES. The grid integrated PV system needs to behave in a similar manner as
synchronous generator during normal and abnormal grid behavior. The active power
control, reactive power control, frequency control, inertial control, voltage control,
fault ride through capability, power quality etc. are being the recent requirement of
grid codes and needs to be satisfied by RES looking at the challenges of its intermittent
nature [2].
The development and deployment of modern power electronics devices, its appli-
cation and control has made the PV system more promising and efficient. For PV
to be able to deliver continuous, reliable and stable maximum power a hybrid sys-
tem including a battery bank is connected to PV. But the use of battery has its own
advantages and disadvantages like huge weight, high costs, performance with time
and pollution [3]. For the large scale integration of RES, the behavior of PV during
abnormal conditions such as faults is crucial and thus for grid connected PV power
plants, the focus is laid on fault-ride-through (FRT) capability [4].
The active and reactive power injection control using various complex strategies
are reported by many researchers till date [5–9]. The dynamic behavior or charac-
teristics of grid connected PV power generation system is analyzed and the dc bus
voltage regulation has been studied in [10]. DC link capacitor is one of the cru-
cial factors to be considered because it affects the life of dc-ac converter i.e., grid
tied inverter. To reduce the cost of the dc link capacitor, its size is kept small [11].
The multi-mode operation of PV system is provided during sag in [12]. The active
power is still supplied during voltage sag. Also in [13], an interactive PV system is
developed such that each module is acting independently according to the ambient
conditions and output power detection.
Of the all other disturbances in grid, voltage sags are more frequent and have
adverse impact on converter’s operation. The reasons for the voltage sag may be
switching of loads, short-circuit faults, load unbalance etc. [14]. With respect to the
limits of current flowing in the duration of unbalance voltage sags, the active and
reactive power production is addressed in [15]. The active and reactive power control
simultaneously is presented in [16]. The control strategy applied in [16] works to
maximize the power produced and addresses different strategies to produce power
during disturbances in grid. In [17], FRT capability is given by either supplying no
active power during voltage sag in two methods and by changing the duty of the
dc-dc converter to have new current reference during sag so that lesser active power
is supplied along with droop control for reactive power.
In this paper simple but accurate adaptive control strategy is proposed for the
active and reactive power injection by PV to the grid as per the requirement under
Active and Reactive Power Control … 111

healthy and distorted grid conditions. This paper also deals with the performance
and analysis of three phase PV system during change in input i, e irradiance and
temperature and the control of active and reactive power injection during unhealthy
grid conditions like voltage sag, voltage swell and harmonics.

2 Integrated Solar PV Grid Tied System with Adaptive

Control

See Figs. 1, 2 and Tables 1, 2.

Fig. 1 Block diagram of PV plant integrated with grid

112 V. Joshi et al.

Fig. 2 IV–PV and MPPT curve (change in irradiance and temperature)

Table 1 PV panel parameters

PV module specifications
Maximum operating voltage (Vmpp ) 54.67
Maximum operating current (Impp ) 5.5
Number of modules in parallel 54
Number of modules in series 5

Table 2 Inverter
Inverter specifications
specifications
Input dc link voltage 500 V
Output ac voltage peak 366.6 V
R 3 mfi
L 0.5 mH

3 Control Strategy

The control strategy of the dc-dc boost converter is decided by MPPT algorithm
based on incremental conductance algorithm [18]. Vector control strategy is used to
decouple the active power and the reactive power control independently via control
of voltage source inverter. The VSC control system consists of two control loops.
One external control loop regulates the dc link voltage and hence maintain the dc link
midpoint voltage constant. By maintaining the dc link voltage constant, it is ensured
to control the current components id and iq which ultimately controls the active power
and reactive power, respectively. The reference signal of active and reactive power
Pref and Qref is respectively generated independently by idref and iqref . The output of
Active and Reactive Power Control … 113

dc link controller used to maintain voltage at dc link constant is taken as idref /Pref (in
pu). The Qref /iqref is given independently according to grid requirements of reactive
power. For unity power factor operation Qref is set to zero.
The other control loop regulate grid currents id and iq (i.e., active and reactive cur-
rent components) i.e., ultimately active and reactive power components. For main-
taining unity power factor, iqref can be set to zero. But under some abnormal conditions
when sensed the reference is changed from zero. Thus reactive power output from
inverter becomes non-zero and it supports the grid by providing reactive power under
some abnormal grid conditions. For the currents id and iq to be positive, it implies
that the converter generates real power and absorbs reactive power respectively. The
voltages are sensed at PCC and then are converted from abc to dq0 reference. The
output of the current controller which is Vd and Vq is then again changed to abc i.e.,
modulating signals which is used further by PWM generator as shown in Fig. 3.

Fig. 3 Control strategy

114 V. Joshi et al.

4 Mathematical Modeling of Inverter and Adaptive Control

The output voltage equations for the mathematical modeling of grid tied inverters
are as follows:

Vd = L(pId ) + R(Id ) + wL(−Iq ) + Vdg (1)

Vq = L(pIq ) + R(Iq ) + wL(−Id ) + Vqg (2)

where,
p = d/dt
Vdg , Vqg = d and q axis components for grid side voltages, respectively. Vd , Vq = d
and q axis components of inverter side voltages, respectively. L, R = inductance and
resistance between inverter and grid respectively. Id , Iq = d and q axis components
of grid side currents, respectively.
The calculations for active and reactive power can be obtained from below equa-
tions

P = Vdg ∗ Id (3)

Q = −Vdg ∗ Iq (4)

The control strategy as mentioned is unity power factor control strategy under
healthy grid conditions. But it changes to adaptive control strategy during abnormal
conditions; when there is a voltage sag or a voltage swell greater than 0.1 pu, the
PV system supports the grid requirements of reactive power. The condition can be
sensed by change in dc link voltage which leads to a larger change in idref . Hence
when this condition occurs, the idref is changed to ensure LVRT operation and the
iqref is set non zero. The flowchart for the adaptive control is shown in Fig. 4.

4.1 Grid Synchronization

Synchronizing with the utility grid is important for connecting the PV with the grid
so that the inverter output voltages are in sync with the grid voltages and is shown
in Fig. 3. PLL is used for grid synchronization of the PV with that of grid. In PLL,
initially the transformation is carried out from abc to dq reference frame of the grid
voltages. Followed by that is PI controller, which is adopted so that the Vq is made
zero. By doing this the voltage is aligned to the d-axis. The PI controller output along
with the integrator then produces the grid voltage phase angle, which is then given
back to the Park transformation. Hence synchronous reference frame (SRF) PLL is
a closed loop control which obtains the phase angle of grid.
Active and Reactive Power Control … 115

Fig. 4 Flowchart for

adaptive control

4.2 DC Link Voltage Regulation

A dc link voltage controller is used to maintain Vdc at the output of the boost converter
at the dc link capacitor to the reference voltage set. The
√ reference
√ voltage for the dc
link capacitor is set to larger than Vmin which is (2 2V/ 3), where V is the rms
voltage value at the output of inverter at grid side. The balance of real power of
system which is PV-grid connection is also taken care by the dc voltage regulator.
Thus the proper tuned output of this controller would represent the current reference
idref or the real power reference Pref .

4.3 Current Regulator and Vabc Reference Generation

In this block the reference currents idref , iqref are compared with measured values of
id and iq, respectively. The error is tuned with the help of PI controller. Id current
reference is the output of the capacitor DC link voltage controller. Iq current reference
is set independently. The reference reactive component is set to zero under healthy
116 V. Joshi et al.

conditions of grid. This shall be known as unity power factor control strategy. But
this reference is changed under unhealthy grid condition to support the grid during
abnormal conditions. Hence the strategy is changed to adaptive control strategy. Here
the current id if positive then inverter generates active power and if iq is positive then
inverter absorbs reactive power. Vd and Vq voltage outputs of the current controller
are converted to three modulating signals Vabc _ref used by the PWM Generator.

5 Results and Discussion

The solar three phase PV system is simulated at various types of conditions. The sim-
ulations are carried out at standard test condition (STC) and ambient environmental
conditions i.e., at 1000 W/m2 and 25 °C. The following case studies are considered
in order to evaluate the performance
– Change in irradiance
– Change in temperature
– Presence of harmonics
– During voltage sag
– During voltage swell.

5.1 Change in Irradiance

The performance of system was observed with irradiance variation. An abrupt change
in irradiation at 0.6 to 1 s from 1000 to 700 W/m2 is simulated. The other change
in irradiance is from 700 to 800 W/m2 and 800 to 900 W/m2 followed by 900 to
1000 W/m2 . The variation of irradiance is demonstrated in Fig. 5. The PV side voltage
is approximately constant and active power Pref changes according to the irradiance.
This ensures the MPPT operation as the voltage and current are following the tracking
algorithm. This depicts the conditions such as partial shading and/or faulty PV/some
panel from array is faulty in real case scenario.

5.2 Changes in Temperature

The results with changes in temperature are shown in Fig. 6. The output power of PV
module, and the voltage at PV panel are also shown in Fig. 6 ensuring the required
output as per the I–V P–V characteristics with change in temperature. The panel
tracks the maximum power point at the different temperature conditions.
Active and Reactive Power Control … 117

Fig. 5 Changes in
irradiation

Fig. 6 Changes in
temperature
118 V. Joshi et al.

Fig. 7 Performance with harmonics

5.3 Presence of Harmonics

The performance of the test system is evaluated and analyzed with the presence of
harmonics in grid voltage. The harmonics of order third and fifth with amplitude
and phase 0.2, −25° and 0.15, 35° respectively are inserted at 1 s. The impact of
presence of harmonics in three phase voltage, three phase current and active and
reactive mean power is depicted in Fig. 7. The currents at the grid side are observed
which are almost sinusoidal and hence ensures the operation as per the adaptive
control strategy proposed. Also the real power and reactive power are maintained
constant.

5.4 Performance with Voltage Sag

At time t = 1 s, a voltage sag of 0.15 pu is applied for 0.15 s and thus the voltage
becomes 0.85 pu from 1 to 1.15 s as shown in Fig. 8. Under this conditions as the
voltage is less than 0.9 pu, the control strategy changes from unity power factor
control strategy to adaptive control strategy according to Fig. 4. Hence the iqref and
reactive power is no more equal to zero, but it is non-zero according to the grid
conditions. This ensures the grid code regulations. Also the currents try to increase,
Active and Reactive Power Control … 119

Fig. 8 Performance under voltage sag

as a result of increase in idref due to the voltage sag. This is due to the fact that it
tries to maintain the real power output constant. But idref cannot be given as it is
(the increased value) because due to increased ac current at the output of inverter, at
the transformer and grid side, it might make the over-current protection active and
disconnect the PV system. Hence a limiter is used that does not allow the current to
increase beyond a pre-defined limit.
The PV module is operating at normal conditions at MPPT for 1000 W/m2 and
25 °C. The variation in three phase voltage, three phase current and active and reactive
mean power is depicted in Fig. 8. Also, as seen in Fig. 8 reactive power regulation is
achieved and a reduced active power is still supplied to the grid. Hence this allows
the PV system to remain connected to the grid and does not allow over-current
protection to trip and simultaneously enhance reactive power support enduring the
LVRT operation. The operation without adaptive control does not limit the current
and hence the active power is not reduced and reactive power support is not enhanced.
The results without adaptive control are compared with adaptive control in Fig. 8
for active and reactive power. Therefore, better results are obtained with adaptive
control.
120 V. Joshi et al.

Fig. 9 Performance with voltage swell

5.5 Performance with Voltage Swell

The voltage swell is applied in two phases of magnitude 0.2 pu, during the time
interval of 1.05 to 1.2 s as shown in Fig. 9. The PV module is operating at MPPT
at 25 °C and 1000 W/m2 . The performance is analyzed by observing the variation
in three phase voltage, three phase current and active and reactive mean power as
depicted in Fig. 9. The PV system continues to supply active power during the voltage
swell. Also as the voltages increase beyond 0.1 pu, the reactive power support is given
to the grid and control changes to the adaptive control and hence as shown in Fig. 9
the reactive power regulation is attained.

6 Conclusion

This paper presents the active and reactive power control of three phase grid con-
nected PV system under abnormal and normal conditions. The simple yet accurate
adaptive control is implemented and demonstrated successfully. The system is sim-
ulated and results are shown under different conditions like PV shading, voltage
harmonics, voltage sag, voltage swell. The PV module is able to remain connected
to the grid under all the given conditions and track the power as per the conditions
Active and Reactive Power Control … 121

thus follows the MPPT operation and also ensure the FRT capability. The limiter is
applied for current so that under fault conditions that may lead to sag, the PV system
is able to inject lesser amount of active power. In addition, the adaptive control for
reactive power ensures that the PV system supports for reactive power regulation.

References

1. Solar Power+Management Magazine. https://solarpowermanagement.net/home. Accessed 27

Apr 2019
2. Craciun B-I, Kerekes T, Sĕra D, Teodorescu R (2012) Overview of recent grid codes for PV
power integration. In: 2012 13th international conference on optimization of electrical and
electronic equipment (OPTIM). IEEE, pp 959–965
3. Xiong Xiaoling, Tse Chi K (2013) Bifurcation analysis of standalone photovoltaic battery
hybrid power system. IEEE Trans Circuits Syst I 60(5):1354–1365
4. Mashaly S, Abdallah MH (2016) Low voltage ride through and fault ride through capability
of 40kw PV model grid connected. Saudi Arabia Smart Grid (SASG), pp 1–6
5. Wang Y, Yang P, Yin X, Ma Y (2015) Evaluation of low-voltage ride through capability of a two-
stage grid-connected three-level photovoltaic inverter. In: 2014 17th International Conference
on Electrical Machines and Systems ICEMS 2014. pp 822–828
6. Fawzy T, Premm D, Bletterie B, Goršek A (2011) Active contribution of PV inverters to voltage
control—from a smart grid vision to full-scale implementation. E & I Electr Eng Inf Technol
128(4):110–115
7. Viswadev R, Ramana VV, Venkatesaperumal B, Mishra S (2018) Real and reactive power con-
trol of solar grid-tie inverter under distorted grid conditions. In: 2018 International conference
on power, instrumentation, control and computing (PICC). pp 1–6
8. Dötter G, Ackermann F, Bihler N, Grab R, Rogalla S, Singer R (2014) Stable operation of PV
plants to achieve fault ride through capability—evaluation in field and laboratory tests. IEEE
5th international symposium on power electronics for distributed generation systems (PEDG).
pp 1–8
9. Kabiri R, Holmes DG, McGrath BP (2014) The influence of PV inverter reactive power injection
on grid voltage regulation. In: IEEE 5th international symposium on power electronics for
distributed generation systems. pp 1–8
10. Chen Y-M, Wu H-C, Chen Y-C (2008) Dynamic characteristics of the grid- connected PV
power generation system. In: IEEE international conference on industrial technology. pp 1–6
11. Hur N, Jung J, Nam K (2001) A fast dynamic DC-link power-balancing scheme for a PWM
converter-inverter system. IEEE Trans Ind Electron 48:794–803
12. Tang Cheng-Yu, Chen Yao-Ting, Chen Yaow-Ming (2015) PV power system with multi-mode
operation and low-voltage ride-through capability. IEEE Trans Ind Electron 62(12):7524–7533
13. Shimizu Toshihisa, Hashimoto Osamu, Kimura Gunji (2003) A novel high-performance utility-
interactive photovoltaic inverter system. IEEE Trans Power Electron 18(2):704–711
14. Chen T-H, Yang C-H, Yang N-C (2013) Examination of the definitions of voltage unbalance.
Int J Electr Power Energy Syst Jul 2013(49):380–385
15. Todorovis I, Grabis S, Ivanovis Z (2018) Grid-connected converter active and reactive power
production maximization with respect to current limitations during grid fault. Department of
Power, Electronic and Telecommunication Engineering, Faculty of Technical Sciences, Trg
Dositeja Obradovica 6, Novi Sad, Serbia, pp 311–322
16. Ebrahimi E, Sanjari MJ, Gharehpetian GB (2014) Control of three-phase inverter-based DG
system during fault condition without changing protection coordination. Int J Electr Power
Energy Syst Dec 2014(63):814–823
122 V. Joshi et al.

17. Mirhosseini Mitra, Pou Josep, Agelidis Vassilios G (2014) Single- and two-stage inverter-
based grid-connected photovoltaic power plants with ride-through capability under grid faults.
IEEE Trans Sustain Energy 6(3):1150–1159
18. Putri RI, Wibowo S, Rifa’i M (2015) Maximum power point tracking for photovoltaic using
incremental conductance method. Energy Proc 68:22–30
CPG-UVT-Based Grid-Connected PV
System

Ravindra B. Kuhada, Bharti B. Parmar and Mahesh H. Pandya

Abstract The paper presents a grid-connected PV system with the unit vector tem-
plate method which is used to control the PV inverter. A control strategy, constant
power generation is used to control the power of the PV system with a DC–DC
converter while high PV penetration is being issued. Therefore, PV penetration is
limited with the CPG control technique without using any extra requirement, e.g.,
extra grid or battery storage. With MATLAB/Simulink, simulation is carried out for
a grid-connected PV system by constant power generation (active power control)
and the unit vector template method is used as an inverter control technique for grid
synchronizing.

Keywords PV (Photovoltaic) · MPPT (Maximum power point tracking) · CPG

(Constant power generation) · P&O (Perturb and Observe) · UVT (Unit vector
template) · PCC (Point of common coupling)

1 Introduction

The universal demand for electrical energy is on the rise due to continuously increas-
ing demand by industrial as well as domestic users. Increased energy consumption
raises serious environmental protection issues which lead us to use renewable energy
as an alternative solution. Generation of electrical power at a small scale, ranging
from 3 kW to 10 MW, by use of renewable energy is a concept of Distributed Gen-
eration (DG). It is normally installed in a distribution system close to the end user
to provide power demand [1]. The most commonly used DG technologies are pho-
tovoltaic cells, small hydropower, wind power, fuel cells, and biomass as these are

R. B. Kuhada (B) · B. B. Parmar · M. H. Pandya

Lukhdhrji Engineering College, Morbi 363642, India
e-mail: [email protected]
B. B. Parmar
e-mail: [email protected]
M. H. Pandya
e-mail: [email protected]
© Springer Nature Singapore Pte Ltd. 2020 123
A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_10
124 R. B. Kuhada et al.

Installed renewable sources with Grid interactive in

India as of May 31, 2018

Wind Power 36.62

Solar Power 28.1
Bio Mass Power 8.7
Small Hydro Power 4.46
Waste to Power 0.13

0 10 20 30 40
in 1000 MW
Installed renewable sources with Grid interactive in India as of May 31,2018

Fig. 1 Installed renewable sources with grid interaction in India as of May 31, 2018

inexhaustible and are renewed by nature itself. The photovoltaic system is widely
used to generate electricity because it is pollution-free, does not need any fossil fuels,
and has no moving part. It is renewed by itself by the sun, and no harmful water or
air pollution is created.
Figure 1 shows that in India, electricity generation is produced maximum by
wind energy sources. The most reliable source of renewable source of energy is
PV generation which is the most promising renewable energy technologies and is
foreseen to replace all conventional sources by 2040.
In the renewable energy sector, India is the fourth most renewable energy market
in the world [2] and its total installed solar power capacity has reached by 28 GW by
the end of 2018, which is almost 5.5% of the global installed solar power capacity,
and would further increase to 8% by 2035. Moreover, the installation level of single-
phase rooftop PVs is also increasing along with the installation of a PV system in
the distribution system.

2 Maximum Power Point Tracking (MPPT)

Solar power generation is based on solar irradiance and atmospheric temperature.

Therefore, when the level of irradiance will increase, the output of the PV power will
be increased and when the temperature will increase, the output of the PV power will
be decreased which are shown in Figs. 2 and 3. The PV module has a unique operating
point to obtain maximum power from PV that is recognized as the maximum power
point (MPP). To track this power from the sun, numerous MPPT algorithms are
implemented which are based on the theory of maximum power transfer which
states that “Maximum Power transfer from the supply side to the load side when
CPG-UVT-Based Grid-Connected PV System 125

P(W)
100

T=0O C MPPT
80 O
T=25 C
T=50O C
60

40
T=75O C

20

0
0 5 10 15 20 25
V (V)

Fig. 2 Effect of different temperature changes on PV curves

P(W)
100
MPPT
80 G=1000 W/m2

G=750 W/m2
60

G=500 W/m2
40

20 G=250 W/m2

0
0 5 10 15 20 25
V (V)

Fig. 3 Effect of different irradiance changes on PV curves

126 R. B. Kuhada et al.

Table 1 Cases of PV generation and load demand

Sr. No. Power generation versus load demand Load supply
1 PPV = Pload By only PV generation
2 PPV < Pload By both PV and grid
3 PPV > Pload By PV generation and excess power feed
into the grid

supply-side impedance match the load side impedance”. In this paper, Perturb and
Observe algorithm is used for MPPT.
P&O algorithm is the most popular MPPT algorithm technique which is also
recognized as the Hill climbing method. This algorithm is widely used among all the
available algorithms as it is the simplest one and also gives quite an accurate result.
This algorithm senses the PV array voltage and current and from it and calculates
the actual available power and varies the duty cycle of the DC–DC converter until
the maximum power is obtained from it [3, 4].
As the PV generation increases continuously, it results in various high PV penetra-
tion issues like Overvoltage, Frequency Fluctuation, Voltage Imbalance, and Reverse
Power flow. The photovoltaic system is connected to the distribution grid majorly.
With enough capacity of the PV system in the distribution system, the PV generation
may match the demand or may not match the load demand. If the PV system has
generation capacity which is equal to the connected load, the load will be supplied by
the PV system, only then no requirement of power is supplied by the grid. If the PV
system has generation capacity less than the connected load, the part of the load will
be supplied by the grid also. And when the PV generation is more than the connected
load, excess generation will be given into the grid as shown in Table 1.
On a sunny clear day, the grid is affected by overloading of the high PV penetration
with the third case [5]. So, the power grid cannot sustain more power than its capacity.
That is why, to reduce this effect of high PV penetration on the grid, advanced control
technique is implemented.

3 Constant Power Generation Control Technique

To deal with these high PV penetration issues, the Distribution System Operator
(DSO) has to increase the grid line. But it imposes an extra cost on the system.
Therefore, it is not an economical option. Another alternative is to connect the energy
storage device or battery with the PV system to deal with the fluctuating PV gen-
eration. When the generated PV power is more than the required load demand, the
battery will be charged by the excess energy which is used to supply the load during
low irradiation condition. However, it is not economical as its cost is high, so it
increases the total installation cost of the PV system due to its limited life time.
CPG-UVT-Based Grid-Connected PV System 127

Fig. 4 Power versus Time

Curve

So, the distributed system operator has to reduce the PV installation. But it is also
not a viable option because if the PV installation is reducing, then at lower irradiation
conditions the PV generation cannot fully supply the load.
The low voltage grid has low X/R ratio resulting in more effective active power
control than the reactive power control. This active power control is also recognized
as the CPG control or it can be called constant power generation (CPG) control [7].
CPG control is used to give a maximum limit in power to the grid, i.e., it curtails
the excess power feed into the grid. This control can be achieved by modifying
the MPPT algorithm. In the DC–DC converter, regulation of the duty cycle can be
achieved by the topology discussed in this paper which is a non-isolated DC–DC
step-up converter. This system contains two modes of operation (i) based on the PV
output power and (ii) set maximum feed-in power limit (Plimit ), either it can be in
MPPT or CPG mode [6].
According to this maximum power point tracking with the active power control,
if the available PV output power PPV is less than the set Plimit , then the PV system
operates in the MPPT mode [8]. If the available PV output power PPV is greater than
the set Plimit , then the PV system operates in the CPG mode to limit the maximum
feed-in power to the grid to Plimit which is described in Fig. 4.
So, the PV output power is given by

PPV (t), W hen PPV (t) < Plimit
Po = (1)
Plimit , W hen PPV (t) ≥ Plimit

Figure 4 shows the five different operating regions of the PV system during the
day. During operating regions 1, 3, and 5, the PV system operates in the MPPT mode
and during operating regions 2 and 4, it operates in the CPG mode [6]. There is a
Perturb and Observe method applied for CPG.
Perturb and Observe (P&O)-Based CPG Algorithm:
In this, the constant power generation control is achieved by modifying the Perturb
and Observe (P&O) algorithm. The operating voltage of the PV cell is being changed
128 R. B. Kuhada et al.

with the P&O method in either direction and are compared against Plimit of the actual
PV output power.
If the actual power of PV is to be less than the Plimit , it will match the actual power
of PV with the previous value of the PV power. If the difference between them is to
be greater than zero, algorithm P&O will move in the same direction otherwise, it
will go into the opposite direction. If the actual PV power is to be more than Plimit ,
it will match the actual power of PV with the previous value of the PV power. If the
difference is to be greater than zero, then algorithm P&O will move in the opposite
direction otherwise, it will go in the same direction.

4 Unit Template Method for the PV Inverter Control

Majorly, the PV system is connected with the grid and for grid-connected PV systems,
synchronizing between the grid and PV must be required. In this paper, the unit
template method is used as a grid integration technique.
With the grid-connected PV system, the supply voltages are distorted. Hence,
inverter output voltages and grid voltages must be in the same phase and same
magnitude at the point of common coupling (PCC) which is required for grid syn-
chronization [11].
In this technique, unit vector templates are extracted from the voltages at PCC.
The distorted input source voltages contain harmonic components in addition to the
fundamental component. For extraction of these unit vectors, the voltages are first
measured and gain (1/V M ) is applied to them where V M is being the peak value of
the voltages at PCC. With the DC link voltage and the reference voltage, the error
is calculated by the PI regulator which is taken as the amplitude of reference supply
currents [9]. Then, three-phase reference supply currents are derived using their
amplitude and unit vectors of voltages which are in phase with the supply currents
and send into a hysteresis band for generating the gate pulses for the PV inverter [10]
which is described in Figs. 5 and 6. The UVT method is basically used to find out
the amplitude of the source currents and it requires lesser number of sensors counts
and computational delay.

5 Simulation Results

See Fig. 7.
In this simulation, irradiance is varied between 1000–500 W/m2 and
500–1000 W/m2 according to the irradiance duty cycle which is also varied. At
t = 0.4 s, the irradiance is varied to its maximum capacity 1000 W/m2 . So, the
PV system generates 3.3 kW power according to with its maximum capacity and
1.6 kW power generating with its half capacity because irradiance is reduced to half
CPG-UVT-Based Grid-Connected PV System 129

VDC*

-+
VDC PI
regulator
iS*

va isa*
vb Peak Value Estimator
vc (VM) isb*

ua isc*
ub
uc

Fig. 5 Block diagram of the unit template method

Isa* +
_
Ia

NOT

GATE PULSES

Isb* +
_
PV
INVERTER
Ib

NOT

GATE PULSES

Isc* +
_
Reference
Currents
Ic

NOT

Actual Currents

Fig. 6 Block diagram of hysteresis current controller

of 1000 W/m2 during 0.6–1.1 s, and again increased 3.3 kW PV power after t = 1.1 s
(Figs. 8 and 9).
But, this PV system is working in two modes, the MPPT mode (1.65 kW) and the
CPG mode (2 kW), and the voltage is decreased while the CPG is used and current is
decreased while irradiance is decreased. In this simulation, power is limited to 2 kW
by the CPG control method. So, at t = 0.4 s, 3.3 kW generated power is controlled to
2 kW when PV is working in the CPG mode. After t = 0.6 s, in the MPPT working
mode, PV power is generating 1.6 kW at its maximum capacity because irradiance
130 R. B. Kuhada et al.

Fig. 7 Line diagram of grid-connected PV system

Fig. 8 Change in Irradiance (W/m2 ) versus Time (s)

Fig. 9 PV Power (W) versus Time (s)

is reduced to half as described in Fig. 9 and after t = 1.1 s, the PV output power is
controlled to 2 kW (Figs. 10 and 11).
As shown in Fig. 12, 1 kW resistive load is taken. The grid is synchronized by
using the unit template method and the pulse is generating by hysteresis current
control method for the PV inverter control.
By using the unit vector template method, the inverter gives the output of almost
of 1.9 kW out of 3.3 kW power when CPG control is at t = 0.4 s. When MPPT
control is at t = 0.6 s than the inverter gives the output of 1.5 kW power of maximum
CPG-UVT-Based Grid-Connected PV System 131

Fig. 10 PV Voltage (V) versus Time (s)

Fig. 11 PV Current (A) versus Time (s)

at 500 W/m2 irradiance occur. When t = 1.1 s and CPG control is used than same
output is achieved as 1.9 kW power given at point of common coupling, as shown in
Fig. 13.
With the synchronized grid and 1 kW load, the first load is supplied with 1 kW
PV power, after that excess power of PV is transferred to the grid which is shown in
Fig. 14 as a negative sign because PV is supplied to the grid.
1 kW three-phase resistive load is supplied with its demand for power which has
230 V phase voltages and 1.44 A phase currents which are shown in Figs. 15 and 16.

Fig. 12 Load Power (W) versus Time (s)

132 R. B. Kuhada et al.

Fig. 13 Power at PCC (W) versus Time (s)

Fig. 14 Power at PCC (W) versus Time (s)

Fig. 15 Three-Phase Load Voltages (V) versus Time (s)

6 Conclusion

In this presented paper, an advanced control technique (CPG) is implemented to

control the active power flow by limiting the maximum feed of the PV power into
the grid which is used to reduce the high PV penetration issue. Simulation is carried
out for 3.3 kW PV power generation with the lower grid capacity. The grid-connected
PV system is implemented by the unit vector template method. Load requires the
power of 1 kW while all the excessive PV generation power is fed through the grid
and the PV generation curtails 2 kW by the CPG control technique, Hence, excess
power is not supplied to the grid by PV.
CPG-UVT-Based Grid-Connected PV System 133

Fig. 16 Three-Phase Load Currents (A) versus Time (s)

References

1. Ackermann T, Andersson G, Söder L (2001) Distributed generation: a definition. Electr Power

Syst Res 57:195–204
2. Renewable Energy industry in India. https://www.ibef.org/industry/renewable-energy.aspx
3. Abdul-Kalaam R, Muyeen S, Al-Durra A (2016) Review of maximum power point tracking
techniques for photovoltaic system. Glob J Control Eng Technol 2:8–18
4. De Brito MAG, Galotto L, Sampaio LP, e Melo GDA, Canesin CA (2013) Evaluation of the
main MPPT techniques for photovoltaic applications. IEEE Trans Ind Electron 60:1156–1167
5. Yongheng Y, Blaabjerg F, Wang H (2014) Constant power generation of photovoltaic sys-
tems considering the distributed grid capacity. In: Applied power electronics conference and
exposition (APEC), 2014 Twenty-Ninth Annual IEEE, pp 379–385. IEEE
6. Sangwongwanich A, Yang Y, Blaabjerg F (2016) High-performance constant power generation
in grid-connected PV systems. IEEE Trans Power Electron 31(3):1822–1825
7. Sangwongwanich A, Yang Y, Blaabjerg F, Wang H (2018) Benchmarking of constant power
generation strategies for single-phase grid-connected photovoltaic systems. IEEE Trans Ind
Appl 54(1):447–457
8. Sangwongwanich, A, Yang Y, Blaabjerg F (2016) Sensorless reserved power control strategy
for two-stage grid-connected photovoltaic systems. In: 2016 IEEE 7th international symposium
on power electronics for distributed generation systems (PEDG), pp 1–8. IEEE
9. Singh B, Chandra A, Al-Haddad K (2014). Power quality: problems and mitigation techniques.
Wiley
10. Prathyusha D, Venkatesh P (2015) Power quality improvement of a three phase four wire
system using UPQC. Int Res J Eng Technol (IRJET), 02(04)
11. Patel A, Chaturvedi P (2016) Performance of SRF-UVTG based UPQC DG for integration of
solar PV with non-linear loads. In: 2016 IEEE international conference on power electronics,
drives and energy systems (PEDES), pp 1–5. IEEE
Grid Integration of Solar PV System

Vyas Dhruv, Chudasama Richa, Ambaleeya Afasana, Bosamiya Swati,

Gajjar Rital and Pandya Rajen

Abstract This paper introduces a simulation of three-phase grid-connected solar

photovoltaic system. The objective of this paper is to represent a grid synchroniza-
tion strategy by using the current control method. The focus is to design an inverter
control topology which will enable inverter to synchronize with the grid. The theo-
retical concept of the grid synchronization is described in detail here. The MATLAB
simulation results for grid integration of solar PV system agrees with the theoretical
concept.

Keywords Phase-locked loop (PLL) · Maximum power point tracking (MPPT) ·

Park transmission · Clarke transmission · Voltage source control (VSC)

V. Dhruv (B) · A. Afasana · B. Swati · G. Rital

Department of Electrical Engineering, Parul Institute of Technology, Parul University, Vadodara,
India
e-mail: [email protected]
A. Afasana
e-mail: [email protected]
B. Swati
e-mail: [email protected]
G. Rital
e-mail: [email protected]
C. Richa
Department of Electrical Engineering, Parul Institute of Technology, Parul University, Junagadh,
India
e-mail: [email protected]
P. Rajen
L&T Power Training Institute, L&T Knowledge City, Vadodara 391760, Gujarat, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 135

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_11
136 V. Dhruv et al.

1 Introduction

In the present day scenario, the consumption of energy is increasing, and the rate of
increment of this electrical energy is less as compared to its increment in utilization.
Recent trends have shown the increase in usage of renewable energy sources like
water, air, wind, sun, etc. [1].
Integrating this renewable energy generation system to the grid has one major
problem. As these renewable energy sources produce fluctuating power, the equip-
ment used is DC to DC converter and DC to AC inverter for changing the nature of
the output of renewable energy sources. In India, solar and wind energy is widely
available. Out of above, solar energy has become a more popular choice.
The author in [2], implemented the three-phase grid-tied inverter, but it is with only
battery as energy source for the inverter. In [3] single-phase rooftop-grid-connected
PV system which is capable of improving power factor and low THD is introduced. In
[5] the transformation of three-phase quantities in rotating reference frame is shown
with control design for both voltage and current and [5, 9], shows single-stage and
two-stage grid-connected solar PV system. As in [10] shows suppression of grid
harmonics and reactive current.
The proposed paper describes a control strategy which is used to generate the duty
cycle of an inverter. To achieve this an “Inverter Control Topology” is constructed.
This topology includes three key concepts: “Phase Locked Loop (PLL)” [2, 4], “Park
and Clarke Transformation” [5, 6], and “Maximum Power Point Tracking (MPPT)”
[7, 8]. The PLL drives the necessary angle (ωt) for Park and Clarke Transforma-
tion. The MPPT tracks the maximum power point which is then compared with the
reference values which ultimately generates the duty cycle for inverter. This allows
inverter switching in such a way that the power produced by the solar PV system
synchronizes with the grid.

2 Grid Connection Principle

Electrical grid is an interconnected network to deliver electricity from generation

end to demand end. Grids supply voltages at largely constant amplitudes. Hence it
is said that the network works as a voltage source. Photovoltaic system converts the
sun’s radiation in the form of light into usable electricity. Hence, the solar PV system
also works as a voltage source.
The two voltage sources of different values cannot be connected in parallel. To join
the two voltage sources in parallel, impedance is introduced between them, which is
a non-dissipating energy storage device as shown in Fig. 1 [3].
The inductor and the one-phase solar PV system together work as a current source.
The current injected into the grid should be in phase with the grid voltage as shown
in Fig. 2. The value of grid current and voltage is as follows:
Grid Integration of Solar PV System 137

Fig. 1 Grid connection

principle

Fig. 2 Gird voltage and

current waveforms

Vg = Vm sin ωt (1)

Ig = Im sin ωt (2)

The voltage across the output of the inverter (Vi ) should be

Vi = Vm sin ωt + ωL Im cos ωt. (3)

Current across the inductor is

1
Ig = ∫(V i − V g)dt. (4)
L
A current control strategy determines the value of current across the inductor,
which will appropriately set the duty cycle of inverter value of I g is achieved an
overall pictorial representation of current control is shown in Fig. 3.

3 Phase-Locked Loop

Phase-locked loop or phase lock loop is a nonlinear feedback system which tracks
the phase of the input signal, and the output signal is maintained and synchronized
with frequency and phase of the reference input signal [1].
One of the simplest methods of phase lock loop is an electronic circuit consisting
of a variable frequency oscillator or VFO and a phase detector in a feedback loop [4].
Figure 4a shows block diagram representation of MATLAB simulation and Fig. 4b
shows actual MATLAB simulation.
138 V. Dhruv et al.

Fig. 3 Block diagram of control topology

Fig. 4 a Block diagram for PLL topology. b Simulation for PLL topology
Grid Integration of Solar PV System 139

The mathematics behind ABC to alpha–beta (Clarke transformation) is shown

below [5].
Suppose the resultant space vector for these three vectors (ABC) is ‘R.’
Therefore,
2π 4π
R = ra e j0 + rb e j 3 + rc e j 3
   
2π 4π 2π 4π
= ra + rb cos + rc cos + j rb sin + rc sin
3 3 3 3
 √ √ 
 rb rc  3 3
r∝ + jrβ = ra − − + j rb − rc
2 2 2 2
⎡ ⎤
 r
r∝ 1 −√21 − 21√ ⎣ a ⎦
R∝β = = rb
rβ 0 − 23 − 23
rc

The alpha–beta to dq0 (Park transformation) is shown below [6]

Rdq = Rαβ e− jρ
 
Rdq = r∝ + jrβ (cos ρ − j sin ρ)
   
rd + jrq = r∝ cos ρ + rβ cos ρ + j −r∝ sin ρ + rβ sin ρ
  
rd cos ρ sin ρ r∝
=
rq − sin ρ cos ρ rβ

4 Maximum Power Point Tracking

Maximum power point tracking is a concept widely used for increasing the efficiency
of a solar PV system. There are many different methods to track the point at which
the PV module gives maximum power output [7]. Such methods are as follows:
(1) Hill climbing method
• P & O (Perturb and Observe)
• Inc. Cond. (Incremental and Conductance)
(2) Power slope method
(3) Differential method.
140 V. Dhruv et al.

The simulation uses the Incremental Conductance method. The algorithm works
as follows:
• First the value of difference of two consecutive voltages (dv) and two successive
currents (di) is measured by subtracting V (k) [ the voltage at particular instant of
time] and V (k-1) [the voltage before V (k)] and I(k) [ the current at a particular
instant of time] and I(k-1) [ the current before I(k)].
• If the value of dv = 0, then the value of di is measured. If the value of di = 0, then
there will be no change in the duty cycle. If di is not equal to 0 then if di > 0, the
duty cycle is decreased otherwise increased.
• If the value of dv is not equal to 0, then the value of di/dv is measured. If the value
of di/dv = −i/v then there is no change in the duty cycle. If the value of di/dv > −
i/v, then the value of the duty cycle is decreased otherwise increased.
• Figure 5a shows flow chart of Inc Con method. Figure 5b shows PV characteristics
and IV characteristics of the PV module. From Fig. 5b it can be said that,
di t
dvt
+ it
vt
= 0, the operating point is at a maximum power point
di t
dvt
+ it
vt
> 0, the operating point is on the left of maximum power point
di t
dvt
+ it
vt
< 0, the operating point is on the right of maximum power point
To get the reference grid current, the MPPT algorithm is used to generate a duty
cycle for the buck-boost converter. The input impedance of the buck-boost converter
shows that buck-boost converter can track the operating point on the whole PV and IV
curve. The derivation of the input impedance equation for the buck-boost converter
is as follows:

5 Inverter Control

After deriving the values of d-axis component of current (Id) and q-axis component
of current (Iq) from phase lock loop (PLL) and reference grid current value from
MPPT, the value of d-axis current (Id) and current at maximum power point (Impp)
are compared, and the error signal is fed to the PI controller. Note that Id is called
the active component and Iq as the reactive component of the current.
To make the active component in phase with the d-axis the reactive component
should be made zero. The vector representation is shown below:
Therefore, the Iq is compared with the Iq* and the value of Iq* is kept zero to
make Id in phase with d-axis as shown in Fig. 6a.
The simulation block diagram for inverter voltage control is as follows:
Simulation:
Figure 7a represents MATLAB simulation of MPPT using Buck-Boost Converter.
Figure 7b represents two-stage grid-connected solar PV system
Grid Integration of Solar PV System 141

Fig. 5 a Flow chart of incremental conductance method. b IV and PV curve of the solar PV cell

Figure 7c shows waveforms of active power which is distorted in the beginning

for 0.1 s.
Here Fig. 7d shows reactive power waveform which is also distorted for 0.1 s.
And Fig. 7e shows voltage and current waveforms. As it is seen the current wave-
forms are distorted because of filter absence.
142 V. Dhruv et al.

Fig. 6 a Vector representation of Alpha-Beta to dq0 transformation. b Simulation block diagram

of inverter control topology

6 Conclusion

In this work, a current control method is introduced for grid integration of solar PV
system. The overall simulation of the system is performed in MATLAB. As shown in
Fig. 7b two-stage grid-connected solar PV system is used. The control strategy highly
depends on the MPPT as it provides the reference current value which will ultimately
flow into the grid. The three-phase to two-phase alpha–beta transformation and vice
versa is achieved by Park and Clarke Transformation and the data obtained from it is
used for analyses and PLL is used to for deriving angle. The simulation results show
that active power which is fed to the grid is around 65–70 KW and reactive power is
±13.49 KVAR.
Grid Integration of Solar PV System 143

Fig. 7 a MPPT simulation in MATLAB. b Main simulation diagram. c Active power. d Reactive
power. e Grid voltage and current
144 V. Dhruv et al.

Fig. 7 (continued)

Acknowledgements Authors would like to express their gratitude to Royal Academy of Engi-
neering (RAE), Parul University (PU), Larsen & Toubro (L&T), and University of Surrey as this
research paper is out of research work carried out as part of Industry Defined Research Project enti-
tled “Green Refrigeration Systems Using Solar Energy” funded by Royal Academy of Engineering,
UK under Newton Bhabha Fund under the Higher Education Partnership—India Program.

References

1. Magdam P, Joshi PN, Yadav S (2017) Simulation of three-phase grid tie inverter. In: IEEE 2nd
international conference for convergence in technology (IZCT)
2. Abdalrahman A, Zekry A, Alshazly A (2012) Simulation and Implementation of grid-connected
inverters. Int J Comput Appl
Grid Integration of Solar PV System 145

3. Beena V, Jayuraju M, Sebis Davis K (2018) Active and reactive power control of signal phase
transformerless grid-connected inverter for distributed generation system. Int J Appl Eng Res
4. Hsieh G-C, Hung JC (1996) Phase locked loop techniques- a survey. In: IEEE Transactions of
Industrial Electronics
5. Sunny AC, Das B, Kasari PR, Sarkar A, Bhattacharya S, Chakrabarty A (2016) SVPWM based
decoupled control of active and reactive power for single stage grid connected solar PV system.
In: IEEE 7th power india international conference (PIICON) 2016
6. Gonzalez M, Cardenas V, Pazos F (2004) DQ transformation development for single-phase
system to compensate harmonic distortion and reactive power. In: 9th IEEE international power
electronic congress
7. Ahmed AS, Abdullah BA, Abdelaal WGA (2017) MPPT algorithm performance and evolution.
In: 11th international conference on computer engineering system (ICCES) IEEE, 19th January
2017
8. Gomathi B, Sivakami P (2016) An incremental conductance algorithm based solar maximum
power point tracking system. Int J Electr Eng
9. Altin N, Ozdemir S, Komurcugil H, Sefa I, Biricik S (2018) Two-stage grid-connected inverter
for PV systems. In: IEEE 12th international conference on compatibility, power electronics
and power engineering
10. Biricik1 S, Komurcugil H, Basu M (2016) Photovoltaic supplied grid-tie three-phase inverter
with active power injection and reactive harmonic current compensation capability. In: IECON
- 42nd annual conference of the IEEE industrial electronics society 2016
Comparative Analysis for INC and P&O
MPPT Based Photovoltaic Energy
Conversion System

Anal Patel, Siddharth Joshi and Bhinal Mehta

Abstract This paper presents the comparative study between two MPPT algorithm
techniques: perturb and observe algorithm and incremental conductance algorithm.
These two methods are tested for 6 kW PV array. A DC–DC boost converter is used to
interface the PV array and DC load. The DC–DC converter works on INC and P&O
MPPT methods. Different atmospheric conditions are applied to the PV panel and
analyzed in which MPPT method gives the better result. The analysis and simulation
shall be done in MATLAB/Simulink platform.

Keywords Incremental conductance · Maximum power point tracking · Perturb

and observe · PV array

1 Introduction

Recently, the environmental concern has become a big issue, this is because of the
carbon emission due to the use of the fossil fuels. Now a days there is a trend to
promote renewable energy sources, which are free of cost and pollution free. The
most popular renewable resource is solar energy because it is pollution-free and
maintains free abundant energy. Nowadays, solar energy is used as electrical energy
and thermal energy. Worldwide, solar installed capacity in the years 2015, 2016, and
2017 was, respectively, 256 GW, 306 GW, and 401 GW [1]. In India, solar installed
capacity reached 28.18 GW as of March 31, 2019 [2]. In India, solar installed capacity
in the year 2015, 2016, 2017, and 2018 was, respectively, 3 GW, 6 GW, 12 GW, and
21 GW [2]. The power output from PV varies with the change in insolation and

A. Patel (B) · S. Joshi · B. Mehta

Department of Electrical Engineering, School of Technology, Pandit Deendayal Petroleum
University, Gandhinagar 382007, India
e-mail: [email protected]
S. Joshi
e-mail: [email protected]
B. Mehta
e-mail: [email protected]
© Springer Nature Singapore Pte Ltd. 2020 147
A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_12
148 A. Patel et al.

temperature. The maximum power point tracking (MPPT) technique tries to extract
maximum power from the PV panel for varying conditions. The intersection between
load line and I-V curve of the photovoltaic module is known as operating point of the
PV at particular climatic conditions. The operating point of the PV will change with
temperature and irradiance [3]. A DC–DC converter with a control circuit ensures the
correct switching of the converter to extract the optimal power from the PV source.
The duty cycle of the DC–DC converter changes according to the control algorithm
implemented for optimum operation of the PV panel and for tracking of MPP power at
different conditions. Various maximum power point algorithms have been proposed
till date with their adaptive control [4]. Different atmospheric conditions are applied
to the PV panel and analyzed in which the MPPT method gives a better result. The
incremental conductance (INC) MPPT method is used to overcome the disadvantages
of the fact that the P&O algorithm is subjected to sudden changing environmental
conditions [5]. This article presents in detail, a comparative study between the two
most popular algorithm techniques, namely, INC algorithm and P&O algorithm. In
[6] change in the irradiation and temperature occurs several times. This has done by
sensing the parameters such as voltage and current. The Incremental conductance
among other methods is the one of the best method to track maximum power point
with sudden change in climatic conditions.

2 Photovoltaic System and Its Modeling

In a photovoltaic system, electrical energy is generated from solar energy. There are
two parts of a PV system (1) Standalone PV system (2) Grid-connected PV system.
A standalone PV system is used in remote and isolated areas where electricity is
not easily available, also transportation cost of electricity is more. Generally, the
PV system is connected with other elements like a DC–DC converter and inverter
for power conditioning; such assembly is known as balance of systems (BoS). The
power output from PV varies with the change in insolation and temperature. The
maximum power point tracking (MPPT) technique tries to extract maximum power
from the PV panel for varying conditions. The intersection between load line and
I-V curve of the photovoltaic module is known as operating point of PV at particular
climatic conditions. The operating point of the PV will change with temperature
and irradiance. Hence, the duty ratio varies such that the panel operates at maximum
power point with MPPT. The PV array with P&O [6, 7] and INC [3, 4] MPPT method
tries to operate at MPP point. The P&O method fails to track maximum power point
under sudden change in climatic conditions hence the P&O method is not efficient
for sudden change in climatic conditions. The incremental conductance (INC) MPPT
method is used to overcome the disadvantages of the fact that the P&O algorithm is
subjected to sudden changing environmental conditions [8, 9]. The comparison can
be done based on solar output power, current, voltage, and response time to reach in
their steady-state value. In this paper, the comparative analysis of these two methods
Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 149

Solar PV
modules
Current Sensor Boost Converter
IPV Isensed

DC Load
VPV

Voltage
sensor
VPV
Comparator
Iref Error
IPV P&O + PI Controller
MPPT Algorithm Limiter +
VPV -
Isensed -
Triangular Carrier Waveform

Fig. 1 Block diagram of the system

has been done in MATLAB/Simulink and the results are analyzed. The block diagram
of the system is shown in Fig. 1.

2.1 Modeling of PV System

The equivalent electrical circuit of a PV cell can be shown in Fig. 2. The output current
I from PV terminals is equal to the photo-generated current I ph , the diode current I d,
and the shunt current I sh . The series resistance Rs describes the internal resistance to
current flow. This depends on the depth of the impurities in the PN junction and the
contact resistance. The shunt resistance represented by Rsh is inversely proportional
to the leakage current to the ground for the ideal photovoltaic cell Rs = 0 (no series
loss) and Rsh = ∞ (no ground fault). The efficiency of the photovoltaic cell depends
on the small variation Rs and does not depend on the variation of Rsh. A slight

Rse
S Id
+
Rsh

Vpv
Ip

Fig. 2 Equivalent circuit of a PV cell [8]

150 A. Patel et al.

I
Id +
+ L L
O Isc O
PV V A V A
Cell - D D

-
Fig. 3 Simple model of a PV cell [10]

increase in the value of Rs reduces the efficiency of the photovoltaic cell.

q(Vp +Ip Rs ) Vp + Ip Rs
IPVES = Ip − Irs (e AkT − 1) − (1)
Rsh

The simplified circuit of the photovoltaic cell is as shown in Fig. 2. A simple

equivalent circuit for a PV cell comprises the current source which is driven by
sunlight connected in parallel with a real diode. The main important parameters for
PV are open-circuit voltage and short-circuit current. When the output terminals of
the equivalent circuit for the PV cell are shorted, no current flows in the diode since V
= 0, so the maximum current (theoretically all the current generated from PV) from
the ideal source flows through the shorted terminals. The obtained current is known
as short-circuit current (I SC ). Similarly, open-circuit voltage (V OC ) is obtained by
making these terminals open. Figure 3 shows PV cell/module connected to the DC
load (resistive load in this case).
By applying KCL in Fig. 2, one can obtain current I flowing through the load,

I = ISC − Id (2)

q Vd
Id = I0 (e kT − 1) (3)

q Vd
I = ISC − I0 (e kT − 1) (4)

At the open circuit condition the load current is zero i.e. I = 0, so open-circuit
voltage V oc is obtained,
 
kT ISC
VOC = ln +1 (5)
q I0
 
ISC
VOC = 0.0257 ln +1 (6)
I0
Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 151

The two main points of the photovoltaic panel comprise open-circuit voltage and
short-circuit current as mentioned above. At V OC, the output power is zero, and at
I SC also the output power is zero. The other two important points of the PV panel
present in the I-V curve is I R or I MPP , i.e., rated current or current at maximum power
point and V R or V MPP , i.e., rated voltage or voltage at the maximum power point. The
PV module is operated in the constant current region and constant voltage region.
From V OC to around MPP is constant voltage region, and after MPP to I SC point the
characteristics lies under constant current region. Equation (2) to (6) are showing
modeled equations for the PV module which is used in PVES for modeling of the
PV cell or module.
The PV cell converts solar light directly into electricity. To obtain high power,
the PV cell is connected in series and parallel combinations to generate the required
current and voltage. However, the efficiency of energy conversion is determined by
the efficiency of the PV system that produces power. The efficiency of the PV system
varies with change in atmospheric conditions such as varying solar temperature and
irradiance. These factors also impact the I-V and PV characteristics of a PV panel.
The PV array has an optimum operating point which is called maximum power
point (MPP), which varies with the solar temperature and irradiance. The maximum
power point tracker extracts the maximum power from the PV panel in changing
solar irradiance and temperature. So it provides high efficiency of the PV system.
The module and its parameters can be shown in Table 1. All the parameters are given
according to STC (T STC = 25 °C, S STC = 1000 W/m2 , 1.5 AM) [5, 8].
Figures 4 and 5 show the impact of the solar irradiation and temperature on the
PV panel characteristics, respectively. The solar irradiation on the earth ground is
1000 W/m2 is considered as a reference called at standard test conditions (STC),
i.e., measured at 25 °C. Solar irradiance depends on many factors, such as cloud
covering, earth movement, gases, and humidity. Solar irradiation decreases due to all
these factors hence the output current of the solar panel decreases. But the variation
in voltage is much lower as shown in Figs. 4 and 5. At higher solar irradiation,
the output current of the PV panel is more. The open-circuit voltage depends on the

Table 1 PV panel and its

Parameters Values
parameters
Panel name SunPower
SPR-305E-WHT-D
Open-circuit voltage (V oc ) 64.2
Short-circuit current (I sc ) 5.96
Series resistance (Rs ) 0.37152
Shunt resistance (Rsh ) 269.5
Ideality factor (A) 0.945
Temperature coefficient of I sc (K i ) 0.061745
Series module (N s ) 5
Parallel module (N p ) 4
152 A. Patel et al.

Fig. 4 Impact of solar irradiation on I-V and PV curves of a solar array for 6.1 kW modules

Fig. 5 Impact of solar temperature on I-V and PV curves of a solar array for 6.1 kW modules
Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 153

temperature of the PV module [6]. When the junction temperature of PV increases the
open circuit voltage of the module decreases and hence the power output decreases
[6].

3 MPPT Techniques for PV Systems

The MPPT is an instrument that extracts the maximum power available from the PV
array at any given instant. The power available from a PV module passes through a
particular operating point and this peak also changes with atmospheric conditions.
This point environmental on the knee of the PV curve of the solar panel. The power
of the PV panel is high at this point. By change the solar irradiance and temperature,
the PV panel extracts the maximum power by using the maximum power point
tracking algorithm. Hence, the changing maximum power needs to be continuously
forced to operate at the voltage corresponding to the MPP. The P&O and INC MPPT
algorithms for the boost converter help to extract maximum power from the PV array
in a standalone PV system. Under constant atmospheric conditions, the maximum
power point or operating point of a PV array is determined by the intersection of the
I pv − V pv characteristic and the load line is shown in Fig. 6. This is usually done
by introducing a DC–DC converter between the PV source and the load. The duty
cycle of the switching signal to the converter is controlled by pulse width modulation
to optimize the current drawn from the PV array. The correct amount of current, in
turn, forces the PV array to work at the optimum power point. Thus, this converter
basically regulates the input voltage to the optimum value continually irrespective
of the variations in the cell temperature and the insolation.
The following two methods are widely used for MPPT in PV systems:
1. Perturb and observe method
2. Incremental conductance method
P&O method: The most important and conventional algorithm implemented in
the MPPT is the P&O method which consists of changing the operating point and
measuring the power supplied by the photovoltaic array. This process is followed
by the comparison of the power values in two consecutive moments that lead to the

Fig. 6 Concept of MPP [6]

154 A. Patel et al.

decision of the direction of the perturbation during the next sampling cycle. This
type of action leads to escalation as the PV curve reaches the MPP and oscillates
until the PV curve changes due to the next state change in the atmosphere.
Incremental conductance (INC): As is obvious from the PV curve of solar mod-
ules, the slope is positive, negative, and zero on the left, right, and peak point,
respectively. This observation is utilized while implementing the INC algorithm.
Observations from Eqs. (7) and (8) reveal that only the voltage and current values
are used to carry out tracking, thereby avoiding the need to calculate the power.
Since it involves division operations, its hardware and software implementation is
relatively complex. The value of voltage and current multiplication gives the value
of power in the PV panel. Power from the PV panel is the multiplication of voltage
and current. The concept of the algorithm is elaborated in Eqs. (8) and (9) (Figs. 7
and 8).
Differentiating the equation of power with respect to voltage:

Fig. 7 Flowchart diagram of the P&O method [6]

Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 155

Fig. 8 Flowchart diagram of the INC method [3]

     
dP d(I ∗ V ) dV dI dI
= = I ∗ + V ∗ = I +V (7)
dV dV dV dV dV

dP
When slope dV = 0, the PV panel operating point reaches MPP. Thus the condition
would be:
dP
=0
dV  
dI
I+ V∗ =0 (8)
dV
 
dI I
=− (9)
dV V
156 A. Patel et al.

4 Simulation and Its Analysis

Case I: Boost Converter Simulation Using P&O MPPT Method: The simulation
analysis is done using the P&O method using 6 kW for the PV array. The PV panel
consists of five modules in series, and four modules are connected in parallel. The
maximum power point is tracked by the P&O algorithm. The results are also obtained
for different irradiation and temperature applied to the PV array. The value of voltage,
current, and power from the PV panel and the DC load are shown in a tabulated format.
The duty cycle based P&O MPPT method is used for generating gating pulse of the
boost converter. One can see from Table 2 that the maximum power point is being
tracked by the MPPT controller based on the P&O algorithm. The response time is
fast because of steady-state input quantities.
Case II: Boost Converter Simulation Using INC MPPT Method: The simulation
analysis is done using the INC method using 6 kW for the PV array. The PV panel
consists of five modules in series, and four modules are connected in parallel. The
maximum power point is tracked by the INC algorithm. The results are also obtained
for different irradiation and temperature applied to the PV array. The value of voltage,
current, and power from the PV panel and the DC load are shown in a tabulated format.
The duty cycle based INC MPPT method is used for generating gating pulse of the
boost converter. One can see from Table 3 that the maximum power point is being
tracked by the MPPT controller based on the INC algorithm. The response time is
fast because of steady-state input quantities.

Table 2 Simulation result of current, voltage, and power of the PV panel and load using the P&O
MPPT method
S (W/m2 ) T (°C) I pv (A) V pv (V) Ppv (W) I load (A) V load (V) Pload (W)
1000 25 22.3 272 6081 12.18 498 6074
1000 30 22.3 268 5993 12.09 495 5983
1000 35 22.3 263 5903 12 491 5894
1000 40 22.4 259 5813 11.91 487 5804
1000 45 22.7 251 5712 11.8 483 5704
800 25 17.8 272 4858 10.88 445 4850
800 30 17.7 269 4786 10.8 442 4778
800 35 18.07 260 4707 10.72 439 4707
800 40 18.13 255 4633 10.64 435 4634
800 45 18.07 252 4565 10.55 432 4560
600 25 14.01 220 3091 8.68 355 3085
600 30 14.05 221 3107 8.7 356 3101
600 35 14 221 3123 8.7 357 3117
600 40 14.11 222 3135 8.7 358 3130
600 45 14.13 222 3145 8.7 358 3139
Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 157

Table 3 Simulation result of current, voltage, and power of the PV panel and load using the INC
MPPT method
S (W/m2 ) T (°C) I pv (A) V pv (V) Ppv (W) I load (A) V load (V) Pload (W)
1000 25 21.9 276 6072 12.1 498 6070
1000 30 21.9 272 5982 12 494 5980
1000 35 22 267 5892 11.9 491 5890
1000 40 22 263 5801 11.9 487 5799
1000 45 21 260 5690 11.7 482 5688
800 25 18.1 267 4837 10.8 444 4834
800 30 18.1 263 4769 10.7 441 4766
800 35 18.1 258 4700 10.7 438 4697
800 40 18.1 254 4630 10.6 435 4627
800 45 18.1 250 4559 10.5 431 4556
600 25 13.9 242 3377 9.01 368 3324
600 30 13.9 240 3351 9.01 368 3324
600 35 14 237 3324 8.99 367 3306
600 40 14 234 3291 8.95 366 3279
600 45 14 231 3255 8.9 364 3246

Case III: Comparison of P&O and INC MPPT Method with Sudden Change
in Irradiation: To check the effectiveness of the MPPT algorithm, the simulation
studies are carried out for sudden change in radiation and variable change in radiation.
Initially, the sudden change in solar irradiation is applied and temperature of the PV
module remains constant for the PV system with duty cycle based P&O and INC
method, individually. The value of voltage, current, and power from the PV panel is
shown in a tabulated format. The same response is checked for the INC method with
the same input conditions. The system is tested for sudden change in radiation level
varying between 60 and 1000 W/m2 . The operating point of the PV will change with
sudden change in irradiance. Hence, the duty ratio varies such that the panel operates
at maximum power point with MPPT and provides high conversion efficiency. The
PV array with P&O [6, 7] and INC MPPT [3, 4] methods tries to operate at MPP point.
With sudden change in the solar irradiance, the incremental conductance algorithm
in tracking the MPP is better than the P&O algorithm. The INC algorithm is more
efficient for tracking an MPP and it also tracks MPP quickly as compared to the P
& O algorithm so the INC algorithm is used in this work [8, 9] (Table 4).

Table 4 The value of voltage, current, and power from the PV panel using the P&O and INC MPPT
methods
Method I pv (A) V pv (V) Ppv (W)
P&O MPPT 13.9 219.7 3051
INC MPPT 14 236.2 3293
158 A. Patel et al.

Table 5 The value of voltage, current, and power from the PV panel using the P&O and INC MPPT
methods
Method I pv (A) V pv (V) Ppv (W)
P&O 22.35 274.4 6084
INC 21.17 281.4 6010

Case IV: Comparison of P&O and INC MPPT Method with Steady Change in
Irradiation: Initially, the steady change in solar irradiation is applied and temperature
of the PV module remains constant for the PV system with duty cycle based INC
and P&O method, individually. The value of voltage, current, and power from the
PV panel is shown in a tabulated format. The same response is checked for the P&O
method with the same input conditions. The input is applied with a steady change
in radiation, i.e., radiation of 700–1000 W/m2 and temperature 25 °C. Both the
MPPT methods extract maximum power from the PV array. But for steady change in
atmospheric conditions, the P&O MPPT method extracts maximum power compared
to the INC MPPT method. So, the P&O algorithm is better than the INC algorithm.
This algorithm gives high accuracy and efficiency especially under steady change in
solar irradiance (Table 5).

5 Conclusion

The MPPT algorithm assures maximum power point from the PV array during vari-
ation in climatic conditions (temperature, solar irradiation) to increase the overall
effectiveness of the system. The effectiveness of the algorithm may vary with change
in the above-said conditions. The effectiveness of the MPPT algorithm varies from
96 to 99% in both cases. The system is tested for various permutations and combi-
nations of solar radiations and temperature of PV. The INC algorithm tracks MPP
with greater accuracy with sudden change in radiation and the P&O algorithm tracks
MPP with steady change in radiation. The power delivered to load varies from 6 to
3.2 kW with change in radiation from 1000 to 600 W/m2 . Both the MPPT algorithms
have advantages for particular climatic conditions and these algorithms improve the
effectiveness of the system.

References

1. https://en.wikipedia.org/wiki/Solar_power_by_country#cite_note-ren21-gsr-2017-36
2. https://en.wikipedia.org/wiki/Solar_power_in_India
3. Koizumi H, Kurokawa K (2005) A novel maximum power point tracking method for PV module
integrated converter. In: IEEE 36th power electronics specialists conference, pp 2081–2086
Comparative Analysis for INC and P&O MPPT Based Photovoltaic … 159

4. Abdulkadir M, Samosir AS, Yatim AHM (2012) Modelling and simulation of maximum power
point tracking of photovoltaic system in Simulink model. In: 2012 IEEE international confer-
ence on power and energy (PECon), Kota Kinabalu Sabah, Malaysia, 2–5 Dec 2012
5. Garg R, Singh A, Gupta S (2014) PV cell models and dynamic simulation of MPPT trackers in
MATLAB. In: 2014 international conference on computing for sustainable global development
(INDIACom), pp 6–12. https://doi.org/10.1109/indiacom.2014.6828003
6. Hairul NZ, Saad M (2010) Comparison study of maximum power point tracker techniques for
PV systems. In: Proceedings of the 14th international Middle East power systems conference
(MEPCON’10), Cairo University, Egypt, 19–21 Dec 2010
7. Bhatnagar P, Nema RK (2013) Maximum power point tracking control techniques: state-of-
the-art in photovoltaic applications. Renew Sustain Energy Rev 23:224–241
8. Verma AK, Singh B, Kaushik SC (2010) An isolated solar power generation using boost
converter and boost inverter. In: Proceedings of the national conference on recent advances in
computational technique in electrical engineering, SLITE, Longowal, India, paper 3011, pp
1–8, 19–20 Mar 2010
9. Kalirasu A, Dash SS (2010) Simulation of closed loop controlled boost converter for solar
installation. Serbian J Electr Eng 7(1):121–130
10. Jiyong L, Honghua W (2009) A novel stand-alone PV generation system based on variable
step size INC MPPT and SVPWM control. In: IEEE-IPEMC, pp 2155–2160
11. Hassaine L, Olias E, Quantero J, Salas V (2014) Overview of power inverter topologies and
control structures for grid connected photovoltaic system. Renew Sustain Energy Rev 30:796–
807
12. Yang L-S, Liang T-J, Senior Member, IEEE. Analysis and implementation of a novel bidirec-
tional DC–DC converter
13. Shu-Hung CH, Tse KK, Ron Hui SY, Mok CM, Ho MT (2003) A novel maximum power
point tracking technique for solar panels using a SEPIC or Cuk converter. IEEE Trans Power
Electron 18:717–24
14. Ishaque K, Salam Z (2013) A review of maximum power point tracking techniques of PV system
for uniform insolation and partial shading condition. Renew Sustain Energy Rev 39:475–488
15. Love GN, Wood AR (2007) Small signal model of a power electronic converter. In: 2007
international power engineering conference, IPEC 2007, pp 636–664
Generation and Regulation of DC
High-Voltage Using Modified Technique
with Application

Varun H. Desai, Abdeali A. Challawala, Kevin B. Ray and Mulav P. Rathod

Abstract In conventional methods of high-voltage DC generation, there is a limi-

tation of withstanding the short-circuit current. The output voltage of conventional
methods is twice that of the input voltage for a single stage. This voltage increases
as the stage increases; therefore by using hybrid topology, the voltage level for a
single stage can be increased which is higher than the conventional rectifier circuit
and also the limitation of withstanding the short-circuit current can be overcome. In
a conventional method, the output voltage is fixed with a fixed input voltage, so by
using the above hybrid topology the output voltage is regulated by keeping the input
voltage fixed.

Keywords DC generation · Voltage regulation · Power devices

1 Introduction

Normally, for the generation of DC voltage of up to 100 kV, conventional methods

are used and the output currents are about 100 mA. In the HVDC generation system,
there are different conventional methods for generating High Voltage DC. Rectifier
circuits for producing high DC voltage from AC sources may be (a) Half-wave rec-
tifier, (b) Full-wave rectifier, (c) Voltage doubler circuit, and (d) Cascaded voltage
doubler. The number of diode and capacitor combinations required to give the high
voltage increases proportionally with the ratio of output to input voltage. The volt-
age multiplier is obtained by cascading the clamping circuit and half-wave rectifier
circuits [1, 2]. A good method to select the capacitor is based on the voltage rating
which should be more than the actual peak voltage applied [3]. In the rectifier circuit,
whether it is half or full wave, the output DC is less than the A.C. maximum voltage.
Hence, to overcome the limitation of the rectifier circuit, Cascaded voltage multiplier
circuit is used. Thus, in all the above conventional methods in order to change the
output voltage, we have to change the input voltage; it is not possible to get variable

V. H. Desai (B) · A. A. Challawala · K. B. Ray · M. P. Rathod

Sardar Vallabhbhai Patel Institute of Technology, Vasad, Anand, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 161

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_13
162 V. H. Desai et al.

output voltage by giving fixed input voltage also in all the above configurations the
output voltage per stages is of 2 Vmax so it can be possible to reduce the stage with
higher voltage. Cockcroft–Walton multiplier circuits which are applied lessen the
number of stages without using a transformer. By the implementation of this tech-
nique, the volume can be minimized and also associated power loss can be reduced.
By omitting the transformer method, cost and size for making Cockcroft–Walton
Voltage Multiplier are reduced [4].
The conventional methods discussed above are usually low-current DC voltage
sources but provide a higher voltage. The voltage can be stepped up by using a
number of diodes in series in a suitable manner, and capacitors without using bigger
transformer for rectification. Though the multistage Half Wave Series Cockcroft–
Walton Voltage Multiplier circuit has been widely used for High Voltage generator
applications, the detailed steady-state analysis and reverse recovery problems of
the successive operating stages of the Half Wave Series Cockcroft–Walton Voltage
Multiplier circuit at high switching frequency have not yet been found in the state-
of-the-art works. The existing analysis for the multistage HW Series CW Voltage
Multiplier circuit is based on the ideal diode model [5, 6]. The silicon carbide Schottky
diode without reverse recovery is used to replace the conventional ultrafast silicon
diode in all voltage multiplier stages to aggravate the reverse recovery problems in
greater frequency operations [7–11]. However, replacing the silicon diodes in all
voltage multiplier stages with power electronics devices will significantly increase
the regulation of voltage multiplier for the HV generator and also it gives ease in
controlling the voltage by the user.
Hence, after defining the problem summary, we came to know about the limitations
of conventional methods so we are conducting this project mainly to overcome the
short circuit current. Also, we can get a regulated output. In this project, we are aiming
to make a HYBRID (MODIFIED) TOPOLOGY to fulfill all the above requirements.
In this hybrid topology, we are using the concept of chopper for regulating the
output by using the power electronic devices which is more efficient, cheap, and
quick responsive than the existing devices used in conventional methods. The block
diagram which gives a brief summary is shown in Fig. 1.

Fig. 1 Adopted method block diagram

Generation and Regulation of DC High-Voltage … 163

2 Design, Analysis, and Implementation of Proposed

Hybrid Topology

Our aim is to build the new combined HYBRID (modified) TOPOLOGY circuit
in the cascaded voltage doubler. Also by replacing uncontrolled devices with power
electronics devices, we can generate high DC voltage and regulate the output voltage,
which includes the cascaded voltage doubler and the multiplier circuit.

2.1 Implementation of Control Circuit

(1) Integrator and Comparator circuit: Designing the control circuit for giving gate
pulse to IGBT is done by using the concept of zero crossing detector. First, the
supply is stepped down to the desired value and then it is given to a zero crossing
detector (ZCD). The output is a square wave which is given as an input to the
generator for generating the triangular wave. The generated triangular wave is
given to an op-amp circuit which has unity gain that inverts the input, i.e., here,
as the input is the triangular wave (Fig. 2).
Hence the op-amp circuit will just invert the triangular wave. Therefore, at the
output side, the inverted triangular wave will be achieved. The two triangular
waves which are generated in the first phase, i.e., triangular wave and inverted
triangular wave, are given to the positive terminal and variable DC supply to the
negative terminal of comparator 1 and comparator 2, respectively. The function
of the comparator is to compare two waves; here, whenever the triangular wave
is greater than the DC supply, the comparator will give the output, i.e., whenever
the positive of op-amp is more than the negative terminal.
(2) Differentiator and Zero Crossing Detector circuit: The output of stepped-down
transformer is given to the differentiator circuit. The function of the differentiator
is to differentiate the input signal; the input signal is a sine wave, and thus the
output of differentiator is cosine wave. This is then given to ZCD which will
give the square wave of cosine (Fig. 3).
(3) Generation of gate pulse for odd devices: The pulse for odd devices is generated
using the logic gates, i.e., AND and NOT. The output of ZCD sine and ZCD
cosine are given to a logic gate AND1. The output of comparator 1 and the
output of AND1 are given to another logic gate AND (odd) which result in the
pulse for odd device (Fig. 4).
(4) Generation of gate pulse for even devices: The output of ZCD sine and ZCD
cosine are given first to NOT gate, respectively. The output of this NOT gate is
given to the logic AND2 gate. The output of comparator 2 and AND2 is given to
the other logic gate, i.e., AND (even) which results in the pulse for even device
(Fig. 5).
Regulation: Now by keeping the input fixed, we can change the output voltage.
This can be done by changing the variable DC supply of the comparator 1 and
164 V. H. Desai et al.

Fig. 2 a Integrator and Comparator 1. b Zero crossing detector sine. c Comparator 1 output wave-
form. d Comparator 2 output waveform
Generation and Regulation of DC High-Voltage … 165

Fig. 2 (continued)

comparator 2, which result in the variable duty pulse of AND (odd) and AND (even)
gate. These pulses are given to the power devices.
The summary of gate pulse generation is mentioned below:
The input sine wave is converted into a square wave with the help of an integrator
and the output of it is compared with the DC signal and a pulse is formed; this pulse
is then given to the logic devices which give the final gate pulse (Fig. 6).

2.2 Implementation of Power Circuit

In the circuit during the positive half cycle of supply voltage, the first IGBT is
forward biased and the second one is reverse biased, and variable gate pulse is given
to IGBT1. Therefore, the flow of current charges the capacitor to positive peak value
and capacitor C1 is charged through IGBT1 to Vin. Similarly, for negative half cycle
of the input voltage, the first IGBT is reverse biased and the second one is forward
biased and variable gate pulse is given to IGBT2. Therefore, the potential of capacitor
166 V. H. Desai et al.

Fig. 3 a Differentiator and

zero crossing detector.
b Zero crossing detector of
cosine wave
Generation and Regulation of DC High-Voltage … 167

Fig. 3 (continued)

C1 adds to that of the source, thus charging C2 to 2 Vin through IGBT2. The second
unit has the same operation as that of the first unit which is connected in the series
through an isolation transformer and the output is taken through the terminal as
shown in Fig. 7.

3 Testing and Control of Power Circuit

3.1 Mounting of Control Circuit

See Fig. 8.

3.2 Mounting of Supply Circuit

See Fig. 9.

3.3 Mounting of Isolation and Power Circuit

See Fig. 10.

168 V. H. Desai et al.

Fig. 4 a Circuit for gate pulse for odd device. b Gate pulse for odd devices

3.4 Mounting of Full Circuit (Front View)

See Fig. 11.

Generation and Regulation of DC High-Voltage … 169

Fig. 5 a Circuit for gate pulse for even devices. b Gate pulse for even devices 1
170 V. H. Desai et al.

Fig. 6 Summary
Generation and Regulation of DC High-Voltage … 171

Fig. 7 a Power circuit. b Output voltage according to different duty cycles. c Output voltage with
25% duty. d Output voltage with 6.8% duty
172 V. H. Desai et al.

Fig. 7 (continued)

Fig. 8 Mounting of control circuit

Generation and Regulation of DC High-Voltage … 173

Fig. 9 Mounting of supply circuit

Fig. 10 Mounting of isolation and power circuit

174 V. H. Desai et al.

Fig. 11 Full circuit (front view)

4 Output Voltage and Waveforms of Circuit

See Figs. 12, 13, 14, 15, and 16.

Fig. 12 Output voltage when circuit is off

Generation and Regulation of DC High-Voltage … 175

Fig. 13 Output voltage when circuit is on

Fig. 14 Gate pulse of odd devices

5 Conclusion

From the vast research and implementation, we were able to understand the voltage
multiplier theory. Through numerous references and citations, we were able to decide
the required components for designing a modified technique for HVDC generation;
for high-voltage testing also the concept and technique adopted for the voltage regu-
lation. In addition to that, we were able to simulate the adopted topology and obtained
the output voltage with and without regulation. During the material failure, there will
be a heavy flow of short-circuit current which can be limited by designing the iso-
lation transformer having inductance up to 5–6% and the other way is to connect a
176 V. H. Desai et al.

Fig. 15 Gate pulse of even devices

Fig. 16 Combined gate pulse

water column resistor in series with the medium or material to be tested. From the
above shown images, depicting the working of our model, we are able to observe
the variable gate pulse and the generated high voltage in the measuring instruments.
We were able to design the suitable control circuit which can generate the desired
gate pulse for the power device; due to the replacement of diode with the controlled
device, it became easy and very efficient to vary the output voltage according to
requirement. This model can produce a suitable voltage for the desired output for
suitable applications.
Generation and Regulation of DC High-Voltage … 177

References

1. Thakare AR, Urkude SB, Argelwar RP (2015) Analysis of Cockroft–Walton voltage multiplier.
Int J Sci Res Publ 5(03):1–3
2. Senthil Raaj GS, Sundar Rajan GT (2013) Simulation and implementation of single-phase
single-stage high step-up AC–DC matrix converter based of Cockroft–Walton voltage multi-
plier. In: International conference on innovations in intelligence instrumentation optimization
and signal processing
3. Beck JM (2008) Using rectifier in voltage multiplier circuits. Vishay general semiconductor
application note, pp 1477–1480
4. Azmi NA, Ismail RC, Jamuar SS, Murad SAZ, Isa MNM, Lim WY, Zulkifeli MA (2017)
Design of DC high voltage and low current power supply using Cockroft–Walton (C–W)
voltage multiplier. IEEE
5. Shenkman A, Berkovich Y, Axelrod B (2004) Novel AC–DC and DC–DC converters with a
diode-capacitor multiplier. IEEE Trans Aerosp Electron Syst 40(4):1286–1293
6. Sun J, Ding X, Nakaoka M, Takano H (2000) Series resonant ZCS-PFM DC–DC converter
with multistage rectified voltage multiplier and dual-mode PFM control scheme for medical-use
high-voltage X-ray power generator. Electr Power Appl Proc 147(6):527–534
7. Mao S, Popovic J, Ferreira JA (2018) Diode reverse recovery process and reduction of half-
wave series Cockcroft–Walton voltage multiplier for high frequency high voltage generator
applications
8. Iqbal S, Singh GK, Besar R (2008) A dual-mode input voltage modulation control scheme for
voltage multiplier based X-ray power supply. IEEE Trans Power Electron 23(2):1003–1008
9. Mao S, Wu T, Lu X, Popovic J, Ferreira JA (2016) High frequency high voltage power con-
version with silicon carbide power semiconductor devices. In: Proceedings of IEEE electronic
system-integration technology conference (ESTC), pp 1–5
10. Lürkens P, Guimaraes P, Godignon P, Millán J (2012) High voltage SiC Schottky diodes in
rectifiers for X-ray generators. Mater Sci Forum 717–720:1245–1248
11. Katzir L, Shmilovitz D (2016) A 1-MHz 5-kV power supply applying SiC diodes and GaN
HEMT cascode MOSFETs in soft switching. IEEE J Emerg Select Top Power Electron
4(4):1474–1482
Energy Regeneration in Direct Torque
Control Drive of Induction Motor

Pravinkumar D. Patel and Saurabh N. Pandya

Abstract This paper discusses kinetic energy recovery into electrical energy during
deceleration of an industrial three-phase induction motor. The energy recovery helps
to improve the efficiency of the induction motor drive. Simulation of direct torque
control (DTC) drive for induction motor with load is carried out, and energy recovery
results are shown. Improving energy utilization of a DTC-based induction motor
drive for traction and industrial drives is discussed. Energy-saving opportunity is
more where rapid deceleration/acceleration occurs. High inertia load like elevators,
escalators, high-inertia rotary saws, rock crushers and grinders with flywheels, sheet
saws driven by high-inertia wheels, centrifuges, and flywheel presses have more
chance to get energy regeneration during deceleration. Induction motor drives which
use braking resistors need to be removed, and energy storage devices need to be
charged during deceleration of induction motor through the bidirectional converter
for energy recovery.

Keywords AC drive · Direct torque control · Energy recovery · Energy

regeneration · Induction motor drives

1 Introduction

The three-phase squirrel cage induction motor (IM) is the most widely used motor in
the industrial world. The diverse range of induction motors takes 65% of the entire
electrical energy available. Various speed control strategies of induction motors are
classified in the literature and are widely recognized [1–3]. Nowadays, AC drives are
used in almost all industrial applications like transport vehicles, marine propulsion,
high-power fan, pumps, cranes, conveyors, lift, and compressors. High-performance

P. D. Patel (B)
Electrical Engineering, Gujarat Technological University, Ahmedabad, Gujarat, India
e-mail: [email protected]
S. N. Pandya
L.E. College, Morbi, Gujarat, India
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 179

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_14
180 P. D. Patel and S. N. Pandya

adjustable speed drives are also increasingly used in electrical traction and other
electric vehicles. Direct torque control is a method to control speed and torque by
varying the inverter voltage space vector magnitude and direction through a lookup
table. DTC was originally introduced in the middle of the 1980s [4, 5] by Takahashi
Isao and Noguchi. The DTC technique has a limitation of making some drift in
extremely low-frequency operations which can, however, be compensated easily
and automatically to minimize the effect of variation of machine constant [4]. The
basic structure of a direct torque control VSI-fed induction motor is shown in Fig. 1.
The stator-oriented vector plane is divided into six vectors. By applying a switch-
ing table of DTC inverter voltage vector, one can increase/decrease the angle between
stator flux and rotor flux to control the torque of the induction motor. The DTC tech-
nique is based on decoupled control of stator flux and torque. In classical DTC works,
during torque dynamic, the optimum switching state vector is selected based on the
value of torque error status, stator flux error status, and orientation of stator flux for
rotation of the shaft [6]. In the induction motor drive, during deceleration or stop
period, usually, energy stored in the inertia of the system is not recovered but wasted

Fig. 1 Simplified structure of DTC of an induction motor [1]

Energy Regeneration in Direct Torque Control Drive … 181

in a braking resistor. In this paper, energy recovery during braking or deceleration

period is proposed.

1.1 Literature Review

F. J. T. E. Ferraira et al. found 174.6e/year of annual saving for 200 kw, 2/4 induc-
tion motor [7]. The outcome shows by pole amplitude modulation (PAM) method
changing the synchronous speed half by change 2 poles to 4 poles. It is found that
around 70 kJ energy recovery in each stop, considering 15 stops per hour and total
6000 h/year operating time [7].
Direct torque control for BLDC motor is projected to regenerate electrical energy
from the kinetic energy and bringing it back to the batteries. S. Geraeea et al. calcu-
lated the state of charge of the battery in common direct torque control against the
modified one which shows 0.6% improvement through 0.9 s simulation time [8].
K. Itani et al. show results of energy recovery efficiency changing from 3.7% for
high-friction road type to 11.2% for medium friction road, for 60 kW permanent
magnet synchronous motors (PMSM) [9].
To find possibilities of energy saving from electric braking in the transportation
vehicle, L. H. Bjornsson et al. found recoverable energy (E recoverable /E traction ) = 26%
and reusable energy (E reuse /E traction ) = 17%. Recoverable energy is the energy which
is fed back of the available kinetic energy to the battery after losses like air drag,
rolling resistance. Reusable energy is the percentage of the recoverable energy which
could be reutilized again to the motor after losses of the inverter and other losses
[10].
C. L. Chua et al. found that the electric vehicle drive dynamic testing system
simulates the full-range speed and torque output to save 65–70% energy. The inverter
drives the three-phase induction motor with the torque and speed control, and this
three-phase induction motor operates in the regenerative braking mode to further
feed the power back to the utility system through the power regenerative inverter
with a unit power factor and low harmonics sine wave [11].
The basic idea of this paper is to improve the energy recovery of a DTC-based
induction motor drive during deceleration of high inertia load. The simulation study
of the given block diagram as in Fig. 2 is shown in Sect. 3. In this paper, a novel
strategy for energy recovery during deceleration or braking period based on change on
frequency is discussed. For variable frequency drive operations, when the induction
motor is commanded to decelerate from higher speed (N) to lower speed (N  ), its
synchronous speed also transits from Ns to Ns . Hence, during the transition, the
actual speed of the induction motor is N and the new synchronous speed Ns , as Ns
< N; hence, regenerative action occurs for a short period of time during which energy
regeneration is possible [12].
182 P. D. Patel and S. N. Pandya

Fig. 2 Block diagram for energy recovery for a 5-HP induction motor drive

1.2 Energy Regeneration

Typically, the motor braking energy is dissipated in a dynamic braking resistor in

the DC link. It works as a pulsed resistance and protects the DC capacitor against
overvoltage during motor braking. Rapid speed reduction results in a negative slip
command, and the motor goes into generator braking. The scope for regenerated
energy must then be returned to the line by the feedback converter or stored in the
energy storage devices like a capacitor bank or battery [13].
High inertia load such as electric vehicles, winders, centrifuges, pumps, and
grinders are more difficult to accelerate and decelerate. The total mass moment
of inertia referring to the motor shaft can be compounded from the kinetic energy of
the drive [14].
Let a motor with the rotor inertia of I m drive a load with the moment of inertia I L
through the transmission of gear ratio N.
The kinetic energy K L of the load rotating with angular velocity ωL is

IL ωL2
KL =
2
while the motor kinetic energy is, K m its rotor velocity is ωm ,

Im ωm
2
Km =
2
The total kinetic energy can be expressed as
Energy Regeneration in Direct Torque Control Drive … 183
   2
ωL2 ω
KT = IL + Im m
ωm
2 2
ωm 2
K T = IT
2
 2  
ωL
where IT = IL + Im
ωm
2

The difference between the motor torque T m and load torque T l is dynamic torque
Td

IT dωL
Td = Tm − Tl =
dt
Hence from the above Dynamic torque equation, high moment of inertia makes a
sluggish response. Therefore high dynamic torque required for fast deceleration and
acceleration both [14].

2 Strategy for Energy Regeneration

The bidirectional DC–DC converter is proposed by Xinxiang Yan et al. for energy
recovery [15]. Figure 2 shows the suggested power topology for a DTC-based three-
phase induction motor drive which is simulated using MATLAB/Simulink. The direct
torque control method is used to control the inverter, and the buck–boost bidirectional
converter is controlled by a different control strategy which is discussed in[16]. In this
paper, the inner current control, outer voltage control strategy is utilized. The control
strategy is discussed by S. Kim et al. [17]. During acceleration, the capacitor bank
(SC) and battery need to be fed to the DC link using a boost converter. During the
transient period, suddenly increased load transients are supplied by a capacitor bank.
During the deceleration of motor and load, DC link voltage increases, so recovery
of energy possible in a capacitor through the buck-boost converter. Here, the buck
converter is used to charge the capacitor bank. Energy recovery is possible frequently
in electric traction, lift, textile mills, paper mills, etc.

3 Simulation Results and Discussion

The proposed block diagram for energy recovery, which consists of a direct torque
control based induction motor drive and a DC–DC bidirectional converter with con-
stant current constant voltage (CCCV) control strategy, has been simulated in MAT-
LAB/Simulink.
184 P. D. Patel and S. N. Pandya

Table 1 Three-phase
Parameters Ratings
induction motor parameters
Rated power 5.4 HP
Frequency 50 Hz
Rated voltage 400 V
Rated speed 1440 rpm
Pole pairs 2
Stator resistance 1.405 
Rotor resistance 1.395 
Stator leakage inductance 5.83 mH
Rotor leakage inductance 5.83 mH
Mutual inductance 0.1722 H
Rotor inertia (J) 0.0131 kg m2
Friction factor (F) 0.002985 N m s

Here, the bidirectional converter is controlled by the CCCV control strategy [17].
During deceleration, mechanical energy which is stored as inertia energy will get
released and the DC bus voltage increase. The DC bus sensor senses the voltage, and
if the voltage is higher than 620 V, the buck action is needed. Therefore the buck
PWM should be generated for switch S1, whereas S2 remains off as shown in Fig. 2.
The bidirectional converter works as a buck converter which transfers power and
charges the capacitor bank. So when the energy flow is reverse, stored energy in the
capacitor is fed back to the DC bus, which helps during acceleration of the motor and
transient heavy load conditions. The DC–DC bidirectional converter used to control
charging and discharging of capacitor. The DC bus voltage and DC voltage of the
capacitor bank are to be sensed and the following action occurs. Here, the capacitor
voltage capacity of 450 V is assumed (Table 1).
V cap > 150 V, V dc < 580 V, boost action, capacitor bank discharge.
V cap < 400 V, V dc > 620 V, buck action, capacitor bank charge.
Figure 3 shows the motor stator current with respect to time. The motor stator
current variation with frequency according to rotor speed is observed. Figure 4 rep-
resents rotor speed in rpm with time. The reference speed of the motor is changed
to 500 rpm at time 0 s, 1440 rpm at 3 s, 0 rpm at 5 s, and 1400 rpm at 8 s with 10 N
m load applied all the time. Electromagnetic torque is observed in Fig. 5. During

Table 2 Simulation
Sr. No. Time (s) Speed (rpm) Torque (N m)
conditions for speed and
torque change with time 1 0 500 10
2 3 1440 –
3 5 0 –
4 8 1400 –
Energy Regeneration in Direct Torque Control Drive … 185

Stator Current vs Time

40

30
Stator Current (A)

20

10

-10

-20
0 1 2 3 4 5 6 7 8 9 10
Time (sec)

Fig. 3 The induction motor stator current (A) for different conditions shown in Table 2

1500

1000
Rotor Speed(rpm)

500

-500
0 1 2 3 4 5 6 7 8 9 10
Time(sec)

Fig. 4 Rotor speed (rpm) with respect to time (s)

Electromagnetic Torque (N.m) vs Time (sec)

80

60
Electromagnetic Torque (N.m)

40

20

-20

-40

-60
0 1 2 3 4 5 6 7 8 9 10
Time(sec)

Fig. 5 Electromagnetic torque (N m) with time (s)

186 P. D. Patel and S. N. Pandya

acceleration at 3 and 8 s, the torque requirement increases up to 30–40 N m. Capaci-

tor bank discharge shows negative current during acceleration and voltage decreases
from 290 to 240 V during acceleration as illustrated in Figs. 7 and 8. The DC bus
voltage demonstrated in Fig. 6 rises from 580 to 640 V during deceleration and is
to be maintained at 600 V ±10%. Figure 7 shows the capacitor voltage and capaci-
tor current concerning time. The capacitor bank voltage increase from 90 to 225 V
during deceleration from 5 s. The capacitor bank is 30% charged during this decel-
eration. The capacitor bank charging current is shown +70 A during deceleration
at 5 s. It is found −10 A discharging current which is fed back during acceleration
at 8 s as represented in Fig. 8. Capacitor discharges 7% during this deceleration.
To trace the working of the drive, a combined view for rotor speed, electromagnetic
torque (N m), capacitor voltage, and DC bus voltage with time are shown in Fig. 9.
Bidirectional converter works on buck PWM for switch S1 during deceleration and

DC Bus Voltage
700 640V
580V 580V
600
DC Bus Voltage V

500

400

300

200

100

0
0 1 2 3 4 5 6 7 8 9 10
Time S

Fig. 6 DC bus voltage with time (s)

DC Vand current at cap strage bank terminal vs Time

300
DC Vand current at cap strage bank

250

200
terminal

150

100

50

-50
0 1 2 3 4 5 6 7 8 9 10
Time Sec

Fig. 7 (Upper trace) Capacitor (as energy storage device) voltage and (Lower trace) capacitor cur-
rent with respect to time (s); the capacitor bank voltage increase from 90 to 225 V when deceleration
occurs at 5 s
Energy Regeneration in Direct Torque Control Drive … 187

Storage Capacitor Bank Current vs Time

80
Storage Capacitor Bank Current A

60

40

20
1.08 A

-10.56 A
-20
0 1 2 3 4 5 6 7 8 9 10
Time Sec

Fig. 8 Storage capacitor bank current (A) with time (s), the capacitor bank charging current (+ve
for charging) 70 A, charging during deceleration at 5 s and current (−ve for discharging) during
acceleration at 3 and 8 s

Fig. 9 Combined view for a rotor speed (rpm), b electromagnetic torque (N m), c capacitor (as
energy storage device) voltage, d DC bus voltage with time (s)

boost PWM for switch S2 during acceleration. Figure 10 shows the PWM pulses,
which are generated by the CCCV control strategy. Recovered energy in the capac-
itor during deceleration cycle at 5 s is 945 J, which shows 0.00026 unit saving per
deceleration. The energy saving depends on the no. of decelerations that occur, the
inertia of mechanical system, and deceleration time. The applications like grinders
with flywheels, sheet saws driven by high-inertia wheels, centrifuges, and flywheel
presses may have a great opportunity of energy saving through regeneration.
188 P. D. Patel and S. N. Pandya

Fig. 10 Bidirectional converter—Buck PWM during deceleration (upper trace), boost PWM (sec-
ond trace) with DC bus voltage (third trace), and rotor speed (rpm) (lower trace) diagram

4 Conclusion

Energy regeneration during deceleration of the induction motor by direct torque

control drive is presented. Capacitor bank as an energy storage device is charging
during deceleration as shown by the graph and the voltage is increasing across it.
Simulation using MATLAB/Simulink is shown with results, which helps to further
implement energy recovery drive of the induction motor. Simulation results and
discussion show that the proposed strategy for energy recovery has a significant
energy-saving potential during deceleration of the three-phase induction motor drive.

References

1. Bose BK (2001) Modern power electronics and AC drives. Prentice Hall PTR, pp 408–413
2. Reza CMFS, Islam MD, Mekhilef S (2014) A review of reliable and energy efficient direct
torque controlled induction motor drives. Renew Sustain Energy (Elsevier) 37
3. Buja G, Kazmierkowski M (2004) Direct torque control of PWM inverter-fed AC motors—a
survey. IEEE Trans Ind Electron 51(4):744–757
4. Isao T, Toshihiko N (1986) A new quick-response and high-efficiency control strategy of an
induction motor. IEEE Trans Ind Appl 22:820–827
5. Depenbrock (1988) Direct self-control (DSC) of inverter-fed induction machine. IEEE Trans
Power Electron 3:420–429
6. Pandya SN, Chatterjee JK (2008) Torque ripple minimization in direct torque control based
IM drive Part-I: single-rate control strategy. In: Joint international conference on power system
technology and IEEE power India conference, New Delhi, pp 1–8
7. Ferreira FJTE, Cistelecan MV, de Almeida AT, Baoming G (2008) Simple strategy to recover
energy during stopping period in large high-inertia line-fed induction motor driven systems.
In: 2008 18th international conference on electrical machines, Vilamoura, pp 1–6
Energy Regeneration in Direct Torque Control Drive … 189

8. Geraeea S et al (2018) Regenerative braking of electric vehicle using a modified direct torque
control and adaptive control theory. Comput Electr Eng (Elsevier)
9. Itani K et al (2016) Comparison between two braking control methods integrating energy
recovery for a two-wheel front driven electric vehicle. Energy Convers Manag (Elsevier)
10. Bjornsson L-H, Karlsson S (2016) The potential for brake energy regeneration under Swedish
conditions. Appl Energy (Elsevier)
11. Chua CL et al (2011) Study of an electric vehicle drive dynamic testing system with energy
recovery. In: 2011 International conference on power electronics and engineering application,
Taiwan. Procedia Eng (Elsevier)
12. Patel PD, Pandya SN (2019) Energy regeneration during deceleration of direct torque control
of induction motor drive for electric vehicles. In: Third IEEE international conference on
electrical, computer and communication technologies (ICECCT 2019), Coimbtore, India, 20–
22 Feb 2019
13. Kazmierkowski MP, Tunia H (1994) Automatic control of converter-fed drives. Elsevier, Ams-
terdam
14. Trzynadlowski AM (2001) Control of induction motors. Academic press, USA
15. Yan X, Patterson D (1999) Improvement of drive range, acceleration and deceleration per-
formance in an electric vehicle propulsion system. In: 30th annual IEEE power electronics
specialists conference. Record. Charleston, SC, vol 2, pp 638–643
16. Saleh M, Esa Y, Mhandi Y, Brandauer W, Mohamed A (2016) Design and implementation of
CCNY DC microgrid testbed. In: IEEE industry applications society annual meeting, Portland,
OR, pp 1–7
17. Kim S, Jeon J, Cho C, Ahn J, Kwon S (2008) Dynamic modeling and control of a grid-
connected hybrid generation system with versatile power transfer. IEEE Trans Ind Electron
55(4):1677–1688. https://doi.org/10.1109/TIE.2007.907662
Emerging Heuristic Optimization
Algorithms for Expansion Planning
and Flexibility Optimization
in Sustainable Electrical Power Systems

Jigar Sarda , Kartik Pandya and Margi Shah

Abstract The expansion planning and flexibility optimization of sustainable elec-

trical power systems are facing higher complexity introduced by massive integration
of variable renewable generation, the increasing need of facts and HVDC devices
for flexibility in highly interactive energy markets, responsive demand and multi-
energy sector coupling. Therefore, the expansion and flexibility management prob-
lems involved in investments decision-making and operational planning need con-
sideration of more accurate models such as non-linear models, probabilistic models
and a large number of decision variables. As the problem is difficult to tackle using
classical optimisation tools, metaheuristic methods are depicted to solve it. The paper
is based on three benchmark systems to evaluate the feasibility and effectiveness of
optimization algorithms in systems of different characteristics and size. Also, the
paper presents the results and statistical comparative evaluation of the performance
of different emerging heuristic optimization algorithms.

Keywords Mixed-integer linear programming · Power flow control · Transmission

expansion planning

J. Sarda (B) · K. Pandya · M. Shah

CHARUSAT University, 388421 Anand, GJ, India
e-mail: [email protected]
K. Pandya
e-mail: [email protected]
M. Shah
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2020 191

A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2_15
192 J. Sarda et al.

Calculation of Load
Line Data Flow analysis

Mathematical Model of
Bus Data
Transmission Expansion (AC
model)

Meta-heuristic
Optimisation
Specification & Output
Techniques
Economic Pa- (Minimum
rameters of TEP Cost)
Transmission
lines

1 Introduction

Transmission system expansion planning gives us effective ways to add new trans-
mission circuits into the system. To have economically viable power systems with
a restructured environment, it is necessary to have an organised system. To reduce
system operational and planning cost while satisfying technical and economic con-
straints, is the overall goal of transmission system expansion [1–3]. Generally, AC
and DC model of power flow are used to solve the expansion problem. With fewer
control possibilities and increasing complexity of problems due to incorrectly esti-
mated costs, the DC model has its own limitation though a large amount of work has
been reported using this model [4, 5]. In the literature, work has been done regard-
ing the AC model to solve the expansion problem. Constructive heuristic algorithm
(CHA) [6, 7] and black-box optimisations are some of the methods stated in the lit-
erature. Also, for more complex systems, CHA is unable to provide good results and
the black-box algorithm cannot handle dispatchable generation. In [8], Continuously
Variable Series Reactor (CVSR) to the transmission expansion planning (TEP) prob-
lem is introduced. Compared to the DC model, the AC model can inculcate planning
of reactive power in the expansion problem. Seeing the complexity and non-linearity
of the TEP problem using the AC model, it can be said that optimal solutions using
any of the optimisation techniques cannot be assured [9, 10]. Therefore, to overcome
the above limitation, different heuristic methods are used.
In this paper, transmission system expansion planning is done using the AC model.
Attention is drawn towards the performance of different heuristic computational opti-
misation techniques which can give reliable suboptimal solutions with less consump-
tion of time for planning of addition of transmission circuits. The organisation of the
paper is as follows: The problem definition used to solve the TEP problem is shown
Emerging Heuristic Optimization Algorithms … 193

in Sect. 2. Section 3 shows Cross-Entropy Covariance Matrix Adaption Evolution-

ary Strategy method of optimization which is a combination of two metaheuristic
algorithms. Three well known test systems are used for analysing the performance
of computational intelligence methods for transmission system planning applica-
tion and Sect. 4 shows the statistical analysis and expansion plan of the benchmark
systems. The conclusion is discussed in Sect. 5.

2 Problem Formulation

The objective function defined in Eq. (1) shows the minimum transmission system
expansion planning cost [11]. In this TEP problem, a number of different optimization
techniques are used for a single transmission line topology.

min v = cab n ab (1)
(a,b)∈

Subject to

0 ≤ tn ≤ m; tn integer (2)

where
v cost due to addition of new circuits.
cab cost of a circuit that can be added between the buses a–b.
n ab number of circuits between the buses a–b.
tn total no. of circuits.
m maximum no. of circuits added in any transmission topology.

3 Cross-Entropy Covariance Matrix Adaption

Evolutionary Strategy (CE_CMAES)

Cross-Entropy Covariance Matrix Adaption Evolutionary Strategy method of opti-

mization is a combination of two metaheuristic algorithms. One is the Cross-Entropy
method and the other is the Covariance Matrix Adaption Evolutionary Strategy
method.
Cross-entropy is used to solve the objective function for finding the optimal value,
denoted by γ ∗ , of the decision variables, thus
 
γ ∗ = S x ∗ = min
x∈ S(x) (3)

where
194 J. Sarda et al.

x generic point in space .

S performance function.
Then CE identifies the associated problem. Thus (3) is modified as the approxi-
mation of the probability Pu {}, for a constant parameter vector u ∈ l,

l(γ ) = Pu {S(X ) ≤ γ } = E u I{S(X )≤γ } (4)

where
X random vector.
E u expectation operator.
I{.} indicator function.
f (.;v) is taken in the form of a Gaussian distribution parameterized by the mean
vector, μ, and variance vector, σ 2 . This adaptive algorithm will initialize a number
of densities f (.; u), f (.; v1), f (.; v2), …, f (.; vk ) that are navigated in the direction of
the theoretically optimal f (.; v*).
CE algorithm generates at each iteration k, a set of N samples of X, i.e., X 1 , …,
X N , using the distribution f(.; vk−1 ), and selects the level γk as the (1–ρ) quantile of
the performance values S(X 1 )… S(X N ) [12]. The next step is to calculate the new
vector parameter vk which is calculated using the smoothing parameter α,

vknew = αvk + (1 − α)vk−1 (5)

This iterative process is stopped when γk ∼

= γ ∗.
Covariance Matrix Adaptation Evolution Strategy (CMAES) is a stochastic,
derivative-free method for numerical optimization of non-linear or non-convex con-
tinuous optimization problems [13].
There are five state variables used for search space n and iteration step k in
CMAES.
• m k ∈ R n , mean distribution.
• σk > 0, step size.
• Ck , covariance matrix.
• pσ ∈ R n , evolution paths.
• pc ∈ R n , evolution paths.
The candidate solutions xi ∈ R n are calculated using the following equations:
 
xi ∼ N m k , σk2 ∗ Ck (6)

∼ m k + σk ∗ N (0, Ck ) (7)

The updated mean value is calculated using Eq. (8),

Emerging Heuristic Optimization Algorithms … 195

μ

m k+1 = m k + wi (xi=1:λ − m k ) (8)
i=1

where sum of the weights w1 ≥ w2 ≥ · · · ≥ wμ > 0 is equal to one.

The new value of the evolution path pσ is calculated using the following equation:
 √ −1/2 m k+1 − m k
pσ = (1 − cσ ) ∗ pσ + 1 − (1 − cσ )2 ∗ μw ∗ Ck ∗ (9)
σk

The step size σk is updated using the following equation:

  
cσ  pσ 
σk+1 = σk ∗ exp −1 (10)
dσ E ∗ N (0, I )

where
 −1

μ
μw = wi2 variance effective solution mass.
i=1
−1/2
Ck = Ck−1 unique symmetric square root of the inverse of Ck .
dσ parameter of damping.
If the value of  pσ  is larger than the predicted value, step size σk is increased;
otherwise, it is decreased.
    
√  n+1 √ 1 1
EN (0, I ) = 2 ∗  n  ≈ n ∗ 1 −
2
+ ∗ n2 (11)
 2 4n 21

Covariance matrix is updated using Eq. (13)

 √ m k+1 − m k
pc = (1 − cc ) ∗ pc + 1[0,α√n ] ∗ ( pσ ) 1 − (1 − cσ )2 ∗ μw ∗
σk
(12)
 xi:λ − m k μ  
  xi:λ − m k T
Ck+1 = 1 − c1 − cμ + cs ∗ Ck + c1 ∗ pc ∗ pcT + cμ ∗ wi
σk σk
i=1
(13)

where
T Transpose.
α 1.5.
c1 ≈ 2/n 2 learning rate for the rank-one update of the covariance matrix.
cμ ≈ μw /n 2 learning rate for the rank-μ update of the covariance matrix.
196 J. Sarda et al.

4 Test Results

The different heuristic optimization techniques were tested on the well known test
systems such as Garver’s 6-bus, IEEE 24-bus and IEEE 118-bus [11]. These algo-
rithms were executed in MATLAB, Intel core i3 3220 CPU with 8 GB RAM. The
open-source tool MATPOWER [14] was also used to manage the function evaluation
Emerging Heuristic Optimization Algorithms … 197

Table 1 Expansion plans and performance comparison among heuristics methods for Garver’s
scenario (a)
Lines CE_CMAES CE_CS CMAES CE
1–2 0 0 1 0
2–3 1 4 0 0
2–6 2 0 0 3
3–5 2 0 1 1
4–6 0 1 1 0
Best (US$×1000) 110 110 110 110
Worst (US$×1000) 130 130 130 130
Average (US$×1000) 120.32 120.32 120.96 122.90
% success 48.387 48.387 45.161 35.483
Time (s) 4818.3 6429.4 6610.9 5766.4

process by using the optimal power flow model. TSPOPF which implements a spe-
cial interior-point technique was used to solve the AC optimal power flow [15]. For
comparing the performance of different heuristic optimization techniques for TEP,
31 trials for each system have been taken.

4.1 Garver’s 6-Bus System

This system consists of six buses and 15 rights-of-way. The maximum number of
lines allowed per right-of-way is five in this case. For those conditions, the search
space is composed of 615 transmission line topologies. For this test system, two
expansion scenarios were analysed as follows: (a) allowing dispatchable generation
and (b) allowing non-dispatchable generation. The solution was obtained using 50
particles and 100 iterations. Tables 1 and 2 show the results of different heuristic
optimization techniques for dispatchable generation and non-dispatchable genera-
tion, respectively. The results show that the novel hybrid method CE_CMAES gives
better performance in terms of average cost, percentage success and time.

4.2 IEEE 24-Bus System

The system has 24 buses and 41 candidate branches. The maximum allowed lines
per right-of-way are five and the number of topologies is 641 . The results were
obtained using 120 particles and 100 iterations. Table 3 shows the result of different
heuristic optimization techniques for IEEE 24-bus system. The result shows that the
198 J. Sarda et al.

Table 2 Expansion plans and performance comparison among heuristics methods for Garver’s
scenario (b)
Lines CE_CMAES CE_CS CMAES CE
1–2 3 4 1 2
1–5 2 0 1 0
1–6 1 0 0 1
2–4 0 0 0 1
2–6 2 0 2 1
3–5 0 3 2 1
3–6 0 1 1 0
4–6 0 0 2 1
Best (US$×1000) 268 268 268 268
Worst (US$×1000) 268 268 408 268
Average (US$×1000) 268 268 303.16 268
% success 100 100 32.25 100
Time (s) 5803.2 6820.9 6923.3 5978.9

Table 3 Expansion plans and performance comparison among heuristics methods for IEEE 24-bus
system
Lines CE_CMAES CE_CS CMAES CE
6–10 1 1 1 1
14–16 1 1 1 1
Best (US$×1000) 70 70 70 70
Worst (US$×1000) 70 70 81 70
Average (US$×1000) 70 70 70.45 70
% success 100 100 93.54 100
Time (s) 17530.64 23965.9 22704.7 19666.7

novel hybrid method CE_CMAES gives better performance in terms of average cost,
percentage success and time.

4.3 IEEE 118-Bus System

This system has 118 buses and 186 rights-of-way. The maximum number of lines
allowed per right-of-way is eight in this case. For those conditions, the search space
is composed of 9186 transmission line topologies. Table 4 shows the result of different
heuristic optimization techniques for the IEEE 118-bus system. The result shows that
Emerging Heuristic Optimization Algorithms … 199

Table 4 Expansion plans and performance comparison among heuristics methods for IEEE 118-bus
system
Lines CE_CMAES CE_CS CMAES CE
8–9 1 1 2 2
8–5 1 1 1 1
9–10 1 2 2 1
38–37 2 2 1 2
108–109 2 1 1 1
Best (US$×1000) 396 396 396 396
Worst (US$×1000) 396 522 872 522
Average (US$×1000) 396 400.06 623.16 407.64
% success 100 96.77 70.96 87.09
Time (s) 493148.7 627567.5 617184.4 592666.8

Total Cost of TEP

1200
1117.741935

1000
854.3225806 858.3870968 868.5483871

800

600

400

200

0
CE_CMAES CE_CS CE CMAES

Fig. 1 Total cost for TEP of different heuristics methods

the novel hybrid method CE_CMAES gives better performance in terms of average
cost, percentage success and time.
The total transmission expansion planning cost comparison of all the system for
different heuristics methods is shown in the figure (Fig. 1).

5 Conclusions

In order to compensate the complications of the power system, the AC model neces-
sitates its importance to solve the transmission system expansion problem. In this
paper, to solve the expansion problem, various metaheuristics techniques like Cross-
Entropy Covariance Matrix Adaption Evolutionary Strategy, Cross Entropy Cuckoo
Search, Covariance Matrix Adaption Evolutionary Strategy and Cross Entropy are
200 J. Sarda et al.

implemented. Results clearly depict that, for the given test systems and expansion
problem, CE_CMAES gives better results in terms of transmission system expansion
planning cost and execution time. Hence, the AC model shows its significance for
the implemented method to solve the TEP problem.

References

1. Latorre G, Cruz RD, Areiza JM, Villegas A (2003) Classification of publications and models
on transmission expansion planning. IEEE Trans Power Syst 18(12):938–946
2. Romero R, Monticelli A, Garcia A, Haffner S (2002) Test systems and mathematical models
for transmission network expansion planning. IEE Proc Gener Transm Distrib 149(1):27–36
3. Hemmati R, Hooshmand R-A, Khodabakhshian A (2013) Comprehensive review of generation
and transmission expansion planning. IET Gener Transm Distrib 7(9):955–964
4. Kezunovic M, McCalley JD, Overbye TJ (2012) Smart grids and beyond: achieving the full
potential of electricity systems. In: Proceedings of IEEE, Special Centennial Issue, pp 1329–
1341
5. Bent R, Toole GL, Berscheid A (2012) Transmission network expansion planning with complex
power flow models. IEEE Trans Power Syst 27(2):904–912
6. Romero R, Rocha C, Mantovani JRS, Sanchez IG (2005) Constructive heuristic algorithm for
the DC model in network transmission expansion planning. IEE Proc Gener Transm Distrib
152(2):277–282
7. Rider MJ, Garcia AV, Romero R (2007) Power system transmission network expansion planning
using AC model. Proc IET Gener Transm Distrib 1(5):731–742
8. Zhang X, Tomsovic K, Dimitrovski A (2017) Security constrained multi-stage transmission
expansion planning considering a continuously variable reactor. IEEE Trans Power Syst
9. Rider MJ, Garcia AV, Romero R (2008) Transmission system expansion planning by a branch-
and-bound algorithm. IET Gener Transm Distrib 2(1):90–99
10. Al-Hamouz ZM, Al-Faraj AS (2002) Transmission expansion planning using nonlinear pro-
gramming. In: Proceedings of the IEEE power engineering society transmission and distribution
conference, vol 1, pp 50–55
11. Torres SP, Castro CA (2014) Expansion planning for smart transmission grids using AC model
and shunt compensation. IET Gener Transm Distrib 8(5):966–975
12. Rubinstein JRY, Krose DP (2004) The cross-entropy method. A unified approach to combina-
torial optimization, Monte- Carlo simulation, and machine learning. Springer, New York
13. Hansen N The CMA evolution strategy: a comparing review. In: Towards a new evolutionary
computation, advances on estimation of distribution algorithms. Springer, Berlin, pp 1769–
1776
14. Zimmerman RD, Murillo-Sánchez CE, Thomas RJ (2011) MATPOWER: steady-state oper-
ations, planning, and analysis tools for power systems research and education. IEEE Trans
Power Syst 26(1):12–19
15. Wang H, Murillo-Sanchez CE, Zimmerman RD, Thomas RJ (2007) On computational issues
of market-based optimal power flow. IEEE Trans Power Syst 22(3):1185–1193
Author Index

A Mohod, Pranav, 97
Afasana, Ambaleeya, 135

P
C Panchal, Dashrath S., 11
Challawala, Abdeali A., 161 Pandya, Kartik, 191
Chaturvedi, Pradyumn, 97 Pandya, Mahesh H., 123
Chaturvedi, Shivam, 87 Pandya, Saurabh N., 57, 179
Parmar, Bharti B., 123
Patel, Anal, 147
D Patel, Himanshukumar R., 37
Desai, Varun H., 161 Patel, Keyurkumar, 1, 25
Dhruv, Vyas, 135 Patel, Pravinkumar D., 179

F R
Fulwani, Deepak M., 87 Rajen, Pandya, 135
Rathod, Mulav P., 161
Ray, Kevin B., 161
G Richa, Chudasama, 135
Gajbhiye, Shubham, 97 Rital, Gajjar, 135

J S
Joshi, Siddharth, 109, 147 Sarda, Jigar, 191
Joshi, Veera, 109 Savaliya, Jatin, 25
Shah, Dipesh H., 11
Shah, Margi, 191
K Shah, Vipul A., 37
Karkar, Hitesh M., 75 Soni, Devesh P., 11
Kuhada, Ravindra B., 123 Sukhadiaa, Rakesh, 57
Swati, Bosamiya, 135

M
Mehta, Axaykumar, 1, 25 T
Mehta, Bhinal, 109, 147 Trivedi, Indrajit N., 75
© Springer Nature Singapore Pte Ltd. 2020 201
A. Mehta et al. (eds.), Advances in Control Systems and its Infrastructure,
Lecture Notes in Electrical Engineering 604,
https://doi.org/10.1007/978-981-15-0226-2