Predictive Modelling for Energy Management and Power Systems Engineering

By Pijush Samui

Predictive Modeling for Energy Management and Power Systems Engineering introduces readers to the cutting-edge use of big data and large computational infrastructures in energy demand estimation and power management systems. The book supports engineers and scientists who seek to become familiar with advanced optimization techniques for power systems designs, optimization techniques and algorithms for consumer power management, and potential applications of machine learning and artificial intelligence in this field. The book provides modeling theory in an easy-to-read format, verified with on-site models and case studies for specific geographic regions and complex consumer markets.

• Presents advanced optimization techniques to improve existing energy demand system
• Provides data-analytic models and their practical relevance in proven case studies
• Explores novel developments in machine-learning and artificial intelligence applied in energy management
• Provides modeling theory in an easy-to-read format

• Language: English
• Publisher: Elsevier Science
• Release date: Sep 30, 2020
• ISBN: 9780128177730

Book preview

Predictive Modelling for Energy Management and Power Systems Engineering

Edited by

Ravinesh Deo

School of Sciences, University of Southern Queensland, QLD, Australia

Pijush Samui

Department of Civil Engineering, NIT Patna, Patna, Bihar, India

Sanjiban Sekhar Roy

School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India

Table of Contents

Cover image

Title page

Copyright

List of Contributors

About the editors

Foreword

Preface

What problem does this book solve?

Why would readers choose this book?

Rigor

Chapter 1. A Multiobjective optimal VAR dispatch using FACTS devices considering voltage stability and contingency analysis

Abstract

1.1 Introduction

1.2 Problem formulation

1.3 A proposed hybrid particle swarm optimization and gravitational search algorithm

1.4 Stability index

1.5 Flexible alternating current transmission systems modeling

1.6 Simulation results

1.7 Conclusion

References

Appendix

Chapter 2. Photovoltaic panels life span increase by control

Abstract

Acronyms

Chapter outcome

2.1 Introduction

2.2 Degradation modes of photovoltaic panels

2.3 Real-time simulation model

2.4 Thermal model of a photovoltaic panel

2.5 Mitigation of degradation via control

2.6 Conclusion

Acknowledgments

References

Chapter 3. Community-scale rural energy systems: General planning algorithms and methods for developing countries

Abstract

List of Acronyms

3.1 Introduction

3.2 Conclusion

Acknowledgments

References

Chapter 4. Proven energy storage system applications for power systems stability and transition issues

Abstract

4.1 Introduction

4.2 Proven energy storage for increased service provision

4.3 Grid functions for energy storage system

4.4 Energy storage characterization for digital inertia

4.5 Test model of the transmission system

4.6 Future implications of hybridized scheme to transition issues

4.7 Chapter summary

References

Chapter 5. Design and performance of two decomposition paradigms in forecasting daily solar radiation with evolutionary polynomial regression: wavelet transform versus ensemble empirical mode decomposition

Abstract

5.1 Introduction

5.2 Study area and evaluation criterion

5.3 Methodology

5.4 Models implementation and application

5.5 Results and discussions

5.6 Conclusions

Appendix

References

Chapter 6. Development of data-driven models for wind speed forecasting in Australia

Abstract

6.1 Introduction

6.2 Materials and methods

6.3 Results of short-term wind speed forecasting

6.4 Results of daily wind speed forecasting

6.5 Summary

References

Chapter 7. Hybrid multilayer perceptron-firefly optimizer algorithm for modelling photosynthetic active solar radiation for biofuel energy exploration

Abstract

Acronyms

7.1 Introduction

7.2 Chapter background review

7.3 Materials and methodology

7.4 Application results and analysis

7.5 Discussion

7.6 Conclusion

References

Further reading

Chapter 8. Predictive modeling of oscillating plasma energy release for clean combustion engines

Abstract

8.1 Introduction

8.2 Challenges of plasma discharge under engine conditions

8.3 Experimental setup and methodology

8.4 Predictive modeling of oscillating plasma discharge

8.5 Conclusions

References

Chapter 9. Nowcasting solar irradiance for effective solar power plants operation and smart grid management

Abstract

9.1 Introduction

9.2 Solar irradiance

9.3 Statistical models for short-time solar irradiance

9.4 Performance of the solar irradiance forecast

9.5 Conclusions

References

Chapter 10. Short-term electrical energy demand prediction under heat island effects using emotional neural network integrated with genetic algorithm

Abstract

10.1 Introduction

10.2 Theoretical overview

10.3 Study area and data

10.4 Predictive model development

10.5 Results and discussion

10.6 Conclusions and remarks

10.7 Limitations and further research

References

Chapter 11. Artificial neural networks and adaptive neuro-fuzzy inference system in energy modeling of agricultural products

Abstract

11.1 Introduction

11.2 Data collection and energy calculation

11.3 Artificial neural network

11.4 Adaptive neuro-fuzzy inference system

11.5 Validation of artificial neural network and adaptive neuro-fuzzy inference system model

11.6 Other models of machine learning

11.7 Interpretation of results

11.8 Conclusion

Acknowledgment

References

Chapter 12. Support vector machine model for multistep wind speed forecasting

Abstract

12.1 Introduction

12.2 Literature review

12.3 Materials and method

12.4 Results and discussion

12.5 Conclusion

References

Appendix

Chapter 13. MARS model for prediction of short- and long-term global solar radiation

Abstract

13.1 Introduction

13.2 Literature review

13.3 Materials and methodology

13.4 Results and discussion

13.5 Conclusion

References

Chapter 14. Wind speed forecasting in Nepal using self-organizing map-based online sequential extreme learning machine

Abstract

14.1 Introduction

14.2 Literature review

14.3 Materials and methods

14.4 Short-term forecasting

14.5 Daily forecasting model

14.6 Monthly forecasting model

14.7 Conclusion

References

Chapter 15. Potential growth in small-scale distributed generation systems in Brazilian capitals

Abstract

15.1 Introduction

15.2 Distributed generation in Brazil

15.3 Measurement method

15.4 Results

15.5 Conclusion

Acknowledgments

References

Chapter 16. Trend of energy consumption in developing nations in the last two decades: a case study from a statistical perspective

Abstract

16.1 Introduction

16.2 Related work

16.3 Implementation

16.4 Conclusion

References

Index

Copyright

Elsevier

Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

Copyright © 2021 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-12-817772-3

For Information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Candice Janco

Acquisitions Editor: Graham Nisbet

Editorial Project Manager: Leticia M. Lima

Production Project Manager: Kamesh Ramajogi

Cover Designer: Greg Harris

Typeset by MPS Limited, Chennai, India

List of Contributors

Tilda Akiki,     Holy Spirit University of Kaslik, USEK, Jounieh, Lebanon

Mumtaz Ali,     Deakin-SWU Joint Research Center on Big Data, School of Information Technology, Deakin University, VIC, Australia

Youssouf Amrane,     Laboratory of Electrical and Industrial Systems, University of Sciences and Technology Houari Boumediene, Algiers, Algeria

Viorel Badescu,     Candida Oancea Institute, Polytechnic University of Bucharest, Bucharest, Romania

Dilki T. Balalla,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Kwok-wing Chau,     Department of Civil and Environmental Engineering, Hong Kong Polytechnic University, Kowloon, Hong Kong

Ravinesh Deo,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Alireza Ghaemi,     Department of Civil Engineering, Graduate University of Advanced Technology, Kerman, Iran

Harshna Gounder,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Fatemeh Hosseini-Fashami,     Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran

Sagthitharan Karalasingham,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Sungwon Kim,     Department of Railroad Construction and Safety Engineering, Dongyang University, Yeongju, South Korea

Anshuman Dey Kirty,     Vellore Institute of Technology, Vellore, India

Nour EL Yakine Kouba,     Laboratory of Electrical and Industrial Systems, University of Sciences and Technology Houari Boumediene, Algiers, Algeria

Timothy Littler,     Department of Energy, Power and Intelligent Control (EPIC), IEEE and EEECS Research Society, Queen’s Belfast University, Belfast, Northern Ireland

Alejandro López-González

Institute of Industrial and Control Engineering, Universitat Politècnica de Catalunya—BarcelonaTech, Barcelona, Spain

Department of Electrical Engineering—Campus Terrassa (ESEIAAT)—BarcelonaTech, Tarrassa, Spain

Socioeconomic Centre of Petroleum and Alternative Energies, Universidad del Zulia, Maracaibo, Venezuela

Nacer K M’Sirdi,     Aix Marseille Université, Marseille, France

Aziz Naamane,     Aix Marseille Université, Marseille, France

Ashkan Nabavi-Pelesaraei

Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran

Management of Fruit and Vegetables Organizations, Tehran Municipality, Tehran, Iran

Bechara Nehme,     Holy Spirit University of Kaslik, USEK, Jounieh, Lebanon

Ananta Neupane,     School of Sciences, University of Southern Queensland, Toowoomba, QLD, Australia

Thong Nguyen-Huy

School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Centre for Applied Climate Sciences, University of Southern Queensland, Toowoomba, QLD, Australia

Vietnam National Space Center, Vietnam Academy of Science and Technology, Hanoi, Vietnam

Eugenia Paulescu,     Faculty of Physics, West University of Timisoara, Timisoara, Romania

Marius Paulescu,     Faculty of Physics, West University of Timisoara, Timisoara, Romania

Ramendra Prasad,     Department of Science, School of Science and Technology, The University of Fiji, Saweni, Lautoka, Fiji

Shobna Mohini Mala Prasad,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Shahin Rafiee,     Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran

Nawin Raj,     School of Sciences, University of Southern Queensland, Springfield, QLD, Australia

Mohammad Rezaie-Balf,     Department of Civil Engineering, Graduate University of Advanced Technology, Kerman, Iran

Paula D. Rigo,     Post-Graduate Program in Production Engineering, Federal University of Santa Maria (UFSM), Santa Maria, Brazil

Carmen B. Rosa,     Post-Graduate Program in Production Engineering, Federal University of Santa Maria (UFSM), Santa Maria, Brazil

Neelesh Sharma,     University of Southern Queensland, Springfield, Springfield, QLD, Australia

Julio Cezar M. Siluk,     Post-Graduate Program in Production Engineering, Federal University of Santa Maria (UFSM), Santa Maria, Brazil

Qingyuan Tan,     Department of Mechanical, Automotive & Materials Engineering, University of Windsor, Windsor, ON, Canada

Jean Ubertalli,     IEEE PES member, Queen’s Belfast University (QUB), Belfast, Northern Ireland

Linyan Wang,     Department of Mechanical, Automotive & Materials Engineering, University of Windsor, Windsor, ON, Canada

Meiping Wang,     Department of Mechanical, Automotive & Materials Engineering, University of Windsor, Windsor, ON, Canada

Zaher Mundher Yaseen,     Sustainable Developments in Civil Engineering Research Group, Faculty of Civil Engineering, Ton Duc Thang University, Ho Chi Minh City, Vietnam

Xiao Yu,     Department of Mechanical, Automotive & Materials Engineering, University of Windsor, Windsor, ON, Canada

Barbar Zeghondy,     Holy Spirit University of Kaslik, USEK, Jounieh, Lebanon

Ming Zheng,     Department of Mechanical, Automotive & Materials Engineering, University of Windsor, Windsor, ON, Canada

About the editors

Professor Ravinesh Deo is an Associate Professor at the University of Southern Queensland, Australia, the Program Director for the Postgraduate Science Program, and Research Leader in Artificial Intelligence. He also serves as an Associate Editor for two international journals: Stochastic Environmental Research and Risk Assessment and the ASCE Journal Hydrologic Engineering journal (USA). As an Applied Data Scientist with proven leadership in artificial intelligence, his research develops decision systems with machine learning, heuristic, and metaheuristic algorithms to improve real-life predictive systems especially using deep learning explainable AI, convolutional neural networks, and long short-term memory networks. He was awarded internationally competitive fellowships including the Queensland Government US Smithsonian Fellowship, Australia–India Strategic Fellowship, Australia–China Young Scientist Exchange Award, Japan Society for Promotion of Science Fellowship, Chinese Academy of Science Presidential International Fellowship and Endeavour Fellowship. He is a member of scientific bodies, and has won Publication Excellence Awards, Head of Department Research Award, Dean’s Commendation for Postgraduate Supervision, BSc Gold Medal for Academic Excellence, and he was the Dux of Fiji in Year 13 examinations. Professor Deo has held visiting positions at the United Stations Tropical Research Institute, Chinese Academy of Science, Peking University, Northwest Normal University, University of Tokyo, Kyoto and Kyushu University, University of Alcala Spain, McGill University, and National University of Singapore. He has undertaken knowledge exchange programs in Singapore, Japan, Europe, China, the United States, and Canada and secured international standing by researching innovative problems with global researchers. He has published books with Springer Nature, Elsevier, and IGI and over 190 publications of which over 140 are Q1 including refereed conferences, edited books, and book chapters. Professor Deo’s papers have been cited over 4000 times with a Google Scholar H-Index of 36 and a Field Weighted Citation Index exceeding 3.5.

Professor Pijush Samui is currently an Associate Professor at the National Institute of Technology, Patna, India. He is an established researcher in the application of Artificial Intelligence (AI) for solving different problems in engineering. He has developed a new method for prediction of response of soil during an earthquake. He has produced charts for the prediction of the response of soil during an earthquake and has developed equations for the prediction of lateral spreading of soil due to an earthquake. He has developed equations for the determination of the seismic liquefaction potential of soil based on strain energy and prediction of seismic attenuation. He has developed efficient models for the prediction of the magnitude of reservoir-induced earthquakes. He has developed models for the determination of medical waste generation in hospitals with equations used for practical purpose. The developed models can be used for the Clean India project. He has determined frequency effects on liquefaction by using the Shake Table. He has applied AI techniques for the determination of bearing capacity and settlement of foundation and equations for the determination of bearing capacity and settlement of shallow foundations. He also developed equations for the determination of compression index and angle of shearing resistance of soil. He has developed equations for the prediction of uplift capacity of suction caisson. He also developed equations for the determination of fracture parameters of concrete. His active research activity is evident from his extensive citation of publications in Google Scholar (total frequency of 1280) with an H-Index of 22. Dr. Samui has published journal papers, books/book chapters, and peer reviewed conference papers with coauthors from Australia, India, Korea, and several other nations. He also holds the position of Visiting Professor at the Far East Federal University (Russia).

Sanjiban Sekhar Roy is an Associate Professor in the School of Computer Science and Engineering, Vellore Institute of Technology. He joined VIT in the year 2009 as an Asst. Professor. His research interests include deep learning and advanced machine learning. He has published around 50 articles in reputed international journals (with SCI impact factors) and conferences. He also is an editorial board member for a handful of international journals and a reviewer for many highly reputed journals such as Neural Processing Letters (Springer), IEEE Access: The Multidisciplinary Open Access Journal, Computers & Security (Elsevier), International Journal of Advanced Intelligence Paradigms (Inderscience International Publishers), International Journal of Artificial Intelligence and Soft Computing (Inderscience International Publishers), Ad Hoc Networks (Elsevier), Evolutionary Intelligence (Springer), Journal of Ambient Intelligence and Humanized Computing (Springer), Iranian Journal of Science and Technology, Transactions of Electrical Engineering (Springer). He uses deep learning and machine learning techniques to solve many complex engineering problems, especially those related to imagery. He is specialized in deep convolutional neural networks and generative adversarial networks. Dr. Roy also has edited many books with reputed international publishers such as Elsevier, Springer, and IGI Global. Very recently, the Ministry of National Education, Romania in collaboration with Aurel Vlaicu University Arad Faculty of Engineers, Romania has awarded Dr. Roy with a Diploma of Excellence as a sign of appreciation for the special achievements obtained in scientific research in 2019.

Foreword

Anirudh Singh Professor , Lautoka, Fiji Islands

The demand for electrical power has been rising globally, both at national and community scales, and there is a need to find more sustainable and newer forms of electrical power generation resources. The world is concurrently faced with the challenge of mitigating climate change, a large portion of which is due to the emission of greenhouse gases arising from the use of fossil fuels.

Renewable energy is in the unique position of addressing both these issues simultaneously. The inclusion of renewable energy technologies (RETs) such as hydropower, wind, solar, and biomass to the generation mix of power grid supplies is routine practice. Such technologies currently supply some 26% of the global electrical power generation. As they displace almost the same fraction of fossil fuel power from the generation mix, these RETs reduce global greenhouse gas emissions by a comparable proportion.

Renewable energy generation consists of dispatchable (synchronous) power such as hydropower and biomass, and variable (or asynchronous) generation such as wind and solar. While synchronous generation may be added seamlessly to the generation mix, the inclusion of asynchronous generation requires more care. The variable nature of such renewable sources makes the total output of the grid supply unpredictable, and their integration into the system leads to system instabilities. These two issues necessitate, amongst other things, the predictive modeling of variable renewable energy resources as well the use of new methodologies for enhancing system strength.

This Edited Book considers the development of computational tools for prediction and optimization of energy production for power systems using computer-aided algorithms and energy management methodologies. The choice of the chapter contributions has been meticulously executed by the Editors. They consist of a wide range of topics specific to energy optimization and forecasting, and include the forecasting and nowcasting of wind and solar energy resources, enhancing the system viability and strength via digital inertia in the form of battery storage and providing algorithms for the management of community-scale rural energy systems.

Amongst the expected ultimate outcomes of this publication is the improvement of power grid system efficiency and its performance. This will have immediate consequences on efficiency of energy distribution at the national and community levels, and make a positive impact on countries’ emissions reduction programs.

The publication of this book comes in the wake of the launch of Sustainable Development Goals (SDGs) and the Paris Agreement in 2015. It synergizes well with Goal 7 of SDG, which is to ensure access to affordable, reliable, sustainable, and modern energy for all. The substantive agreement reached in Paris with regard to climate change mitigation was the undertaking by all Parties to Nationally Determined Contributions (NDCs) to greenhouse gas reductions. Following the release of the IPCC Special Report on Global Warming of 1.5°C in October 2018 and the subsequent Climate Action Summit of September 23, 2019, there has been vigorous debate regarding the adequacy of the NDCs in achieving the agreed goal of net zero emissions by 2050. The outcome of these deliberations was the realization that much greater reductions in emissions were necessary than those proposed originally in the Paris Agreement. The present book will go a long way toward facilitating the design and implementation of power grid systems that improve national and community-scale distribution as well as reducing overall national GHG emissions.

This book is of great relevance to two of the major ongoing discussions on global issues, and is an immensely valuable and timely addition to the scientific literature on energy modeling and management.

22 March 2020

Preface

Today climate advocates are advising energy industries to embed renewable energies into power grids; the role of artificial intelligence in demand side management remains paramount but the success of this vision is at the heart of latest modeling or optimization techniques. National electricity markets, energy management experts, electronic, electrical, and mechatronic engineers should be familiar with advanced optimization techniques that can be used to improve existing energy demand systems, and also to integrate renewable energy into real power grids.

This book provides ideas on optimization techniques as an interdisciplinary concept. The book acts as a common platform required by practitioners to become familiar with the latest developments of energy optimization techniques based on artificial intelligence. It is written to provide a one-spot collective resource for practitioners to learn about predictive models in the energy sector, their practical applications, and case studies.

The purpose is to provide the modeling theory in an easy-to-read format verified with onsite models (i.e., case studies) for specific geographic regions and scenarios. There is a need for this sort of text because we currently have several models in isolated contexts. Putting the theory of energy simulation models and applying those optimization techniques in a single bound book will help novice readers to grasp the concepts more easily than highly technical publications.

What problem does this book solve?

Currently, technical papers and books present materials in a way such that both a beginner reader and energy experts find it too hard to grasp the ideas. A requirement for postgraduate, early and mid-career researchers is to read and understand energy modeling in a way that they can quickly relate the theory and practical applications. This book will provide such a platform whereby readers will appreciate both the theory and practical applications, and also see the comparison of different energy management and optimizations in different chapters.

Why would readers choose this book?

Readers will choose this book because it contains both theory and practice related to energy demand management in a single document, has several optimization models in this area, provides easy-to-understand chapters, and supports people new to the field. For experts, the book will be appealing as it gives first-hand experience about artificial intelligence models—an area that is growing in the current phase.

The book is written as a practical guide focused for postgraduate teaching (case studies, modeling, and simulations), early and mid-career research and teaching scholars, academics, renewable energy practitioners, electrical and electronic engineers, climate energy scientists, and future energy policy makers. It will serve as a highly summarized text on latest developments in energy, consumer energy simulations, and energy demand side management.

The book provides the latest approaches for energy exploration, advanced predictive models, and case studies in geographically diverse locations, modern techniques, and demonstrations to apply artificial intelligence in decision-making for the renewable and conventional energy sector. It will therefore be a useful resource for the energy industry—particularly for engineering and energy management experts.

Rigor

This book is compiled carefully with highly focused chapters that will present to the readers the modern-day optimization techniques in energy exploration (particularly a balanced account of theory and case studies) applied in the energy demand side and real-life power management system. It will make a significant contribution to the development of mathematical tools and data simulation models, and their relevance to different geographic power distributions and case studies that will support modern-day energy engineering applications.

The text will be a useful resource for power systems engineering and the design of energy management platforms in complex consumer markets, for scientific application of real-time energy prediction and management systems, and for integrating artificial intelligence tools for real-time adaptive systems incorporated in energy predictions and management environments. The book will assist modern-day engineers and scientists to become familiar with advanced optimization techniques for better power systems designs, optimization techniques, and different algorithms for consumer power management.

It is our hope that all readers will benefit significantly in learning about the state-of-the-art machine learning models and decision support systems, including energy management science and energy policy perspectives.

Happy reading and learning!

Ravinesh Deo

University of Southern Queensland, Towoomba, QLD, Australia

June 24, 2020, Email: [e-mail]

Chapter 1

A Multiobjective optimal VAR dispatch using FACTS devices considering voltage stability and contingency analysis

Youssouf Amrane and Nour EL Yakine Kouba,    Laboratory of Electrical and Industrial Systems, University of Sciences and Technology Houari Boumediene, Algiers, Algeria

Abstract

An effective allocation of the reactive power in an electrical network aims generally to improve the voltage profile, to minimize transmission losses, and\or to maximize the network voltage stability margin. To solve this kind of problem a hybrid technique combining particle swarm optimization and gravitational search algorithm (PSO-GSA) is proposed. The essential goal of this study is to ensure the feasibility of the power system in the state of contingencies. Therefore we will consider unfavorable cases to prepare, prevent, and plan the system to deal with any incidents. The proposed program will provide a solution to any variation occurring in the transport of energy or suggested to be studied. For this purpose two critical situations are simulated and studied. Also, this study considers the installation of two different flexible alternating current transmission systems devices, namely, the static volt ampere reactive compensator and the thyristor controlled series compensator. To identify the location of the latter, two stability index methods are used, namely, fast voltage stability index and line stability index. The proposed method is applied on the equivalent Algerian electric power system 114-bus. The obtained results are compared with PSO and GSA separately. The results obtained by the proposed method show its effectiveness for improving the reactive power planning problem.

Keywords

FACTS devices; optimal reactive power planning study; multiobjective optimal VAR dispatch; hybrid PSO-GSA; voltage stability index; contingency analysis; equivalent Algerian electric power system

1.1 Introduction

Numerous complex power system planning and operations optimization problems have to be solved by the power system engineers and researchers. Optimal reactive power planning (ORPP) is one example of an optimization problem which is concerned with the security and economy of a power system operation. The ORPP is one of the most complex problems of power systems since it requires the simultaneous minimization of two objective functions. The first one deals with the minimization of operation cost by reducing real power loss and improving the voltage profile. The second objective is to minimize the allocation cost of additional reactive power sources (capacitive or inductive banks, FACTS devices, etc.). Also, the ORPP problem must satisfy a number of physical and operational limitations constraints. The latter include the load flow equations, real and reactive power generator, lower and upper limits of the tap ratios of transformers, shunt capacitor or reactor outputs, and generator voltages (Amrane et al., 2014).

The ORPP is modeled as a large-scale nonlinear programming problem (NLP). To solve the ORPP problem many conventional and intelligent optimization algorithms have been proposed, such as quadratic and sequential quadratic programming (QP/SQP) (Grudinin, 1998), interior point method (IPM) (Amrane et al., 2014; Oliveira et al., 2015), particle swarm optimization (PSO) (Amrane and Boudour, 2015; Pourjafari and Mojallali, 2011), differential evolution algorithm (DEA) (Amrane et al., 2015), and bacterial foraging algorithm (BFA) (BelwinEdward et al., 2013).

In this paper, a hybrid PSO and gravitational search algorithm (PSO-GSA) (Mirjalili and Hashim, 2010) is proposed to solve the ORPP problem. The PSO-GSA has been found to be robust and flexible in solving the complex optimization problem. To validate the robustness of this method 23 benchmark functions have been used to validate the performance of the PSO-GSA algorithm and compare it with standard PSO and GSA (Mirjalili and Hashim, 2010). The obtained results show that the number of functions which the PSO-GSA performs well is nearly twice that of PSO and GSA, which shows it is robust and effective. Lenin et al. (2014) applied the PSO-GSA method to solve the optimal reactive power dispatch (ORPD) problem for real power loss and the minimization of bus voltage deviations. Mangaiyarkarasi and Raja (2014) showed another use of PSO-GSA, which cartels the exploiting and exploring features of the PSO and GSA to achieve the objective of determining the optimal location and optimal size of the static volt ampere reactive compensator (SVC), and thereby minimize the voltage deviation from the nominal value. The hybrid PSO-GSA algorithm also has been applied on the economic load dispatch problem (ELD) problem considering transmission loss, prohibited zones and ramp rate limits (Ashouri and Hosseini, 2013; Hardiansyah, 2013; Jiang et al., 2014). A state-of-the-art of the proposed method in several electrical engineering domains is presented in the appendix section.

The voltage instability study is considered as one of the critical issues in the electric power system. In this chapter, the voltage instability study is based on two different stability indexes. Namely fast voltage stability index (FVSI) (Amrane et al., 2014; Musinin and Abdul Rahman, 2002) and line stability index (Lmn) (Moghavemmi and Omar, 1998) are studied and used to identify the weakest bus and the most critical line in the system. The proposed approach has been tested on ORPP problems using SVCs and TCSCs devices for the equivalent Algerian electric power system 114-bus. Two stability indexes, FVSI and Lmn, are used to identify the weakest buses and lines to install the SVC and TCSC devices.

1.2 Problem formulation

In this chapter, the global objective function of the ORPP problem is to minimize three objective functions that represent: (1) the investment cost of FACTS devices (SVC and TCSC) (BelwinEdward et al., 2013); (2) transmission real power losses; and (3) voltage stability, while satisfying several equality and inequality constraints.

1.2.1 Objectives functions

1.2.1.1 Minimizing the investment cost of SVC and TCSC devices

(1.1)

(1.2)

1.2.1.2 Minimizing the transmission real power losses

(1.3)

1.2.1.3 Minimizing the voltage stability

(1.4)

where fFACTS is the objective function of the FACTS devices investment cost; Gc and Hc represent equality and inequality constraints of the system; U is the vector of controls variables; and X is the vector of state variables. fSVC is the cost function of SVC; fTCSC is the cost function of TCSCis the TCSC reactive power; fPloss is the objective function of real power losses problem; Vi,Vj are the voltage magnitudes; Gk, is the conductance of branch k; θi,θj are the voltage angel at buses i and j; NLi is the number of transmission lines; fSTA is the objective function of voltage stability; Zijline is the line impedance; and Xijline is the line reactance connecting bus i and bus j, while Qjr is the reactive power at the receiving end and Vis is the sending end voltage.

1.2.2 Equality and inequality constraints

1.2.2.1 Equality constraints (the load flow equations)

(1.5)

(1.6)

1.2.2.2 Inequality constraints (technical limitations)

1. Generator constraints:

(1.7)

(1.8)

2. FACTS device constraints:

(1.9)

(1.10)

3. Transformer constraints:

(1.11)

4. Security constraints:

(1.12)

(1.13)

with

(1.14)

(1.15)

where PDi, QDi are real and reactive power at bus i; PGi, QGi are real and reactive powers of the ith generator; Vi is the voltage magnitude at bus i; NBus is the number of buses; Gij and Bij are the conductance and susceptance between i and j; θij is the phase angle difference between the voltages at i and j;θji is the phase angle difference between the voltages at j and i; Nbus is the number of buses; VG is the generator voltages; QG is the reactive power outputs; NG is the TCSC susceptance; NSVC and NTCSC is the transformer tap settings; NT is the number of transformers; VL is the voltage at load bus and {Stfrom, Stto} is the transmission line loading; Nload is the number of load buses and NLi the number of transmission lines.

The equality constraints given by Eqs. (1.5) and (1.6) are satisfied by running the power flow Newton-–Raphson algorithm. The control variables presented in (1.7), (1.9), (1.10), and (1.11) are self-controlled, and the dependent variables are added in the quadratic penalty terms to the objective function in order to keep their final value close to their operating limits.

The objectives functions are standardized in a comparative manner with the base case, but by considering the fact that the objective of active power losses and the voltage stability problem are more important than the FACTS devices costs. For this reason, we use different coefficients for each objective (Amrane et al., 2013).

(1.16)

.

In the above objective function Vilim, QGilim, and Silim are defined in the following equations.

(1.17)

(1.18)

(1.19)

, and λs are the penalty factors which can be increased in the optimization procedure; Vilim, QGilim, and Stlim are defined in the following equations. Fig. 1.1 shows the global problem formulation.

Figure 1.1 Global formulation.

1.3 A proposed hybrid particle swarm optimization and gravitational search algorithm

1.3.1 Particle swarm optimization

PSO is a swarm intelligence method inspired by the social behavior of bird flocking or fish schooling, and developed for global optimization algorithm by J. Kennedy and R. Eberhart in 1995 (Kennedy, 1995). It has become one of the most popular techniques applied in various optimization problems, due to its ease and capability to find near-optimal solutions. The PSO uses a number of particles that constitute a swarm. Each particle traverses the search space looking for the global optima (minimum or maximum). The particles that constitute the PSO system fly around in a multidimensional search space; during this flight each particle adjusts its position according to its own experience, and the experience of the neighboring particles, making use of the best position encountered by itself and its neighbors. The swarm direction of a particle is defined by the set of neighboring particles and its history experience (Soliman and Mantawy, 2012).

The flowchart of PSO is shown in Fig. 1.2.

Figure 1.2 Flowchart of the PSO algorithm.

1.3.2 Gravitational search algorithm

The GSA is a novel metaheuristic searching algorithm which was proposed by E. Rashedi et al. in 2009 (Rashedi et al., 2009). The basic physical theory, from which this algorithm is inspired, is based on Newton’s theory (law of gravity and of motion). In Newton’s theory every particle in the universe attracts every other particle with a force that is directly proportional to the product of masses and inversely proportional to the square of the distance between them (Lenin et al., 2014; Rashedi et al., 2009).

The flowchart of PSO is shown in Fig. 1.3.

Figure 1.3 Flowchart of the GSA algorithm.

1.3.3 A hybrid particle swarm optimization gravitational search algorithm

The hybrid algorithm proposed in this study is a combination of a PSO algorithm and a GS algorithm. The PSO-GSA is a hybrid method which was proposed by S. Mirjalili et al. in 2010 (Lenin et al., 2014). Its basic idea is to combine the ability of social thinking (gbest) in PSO (Soliman and Mantawy, 2012) with the local search capability of GSA (Rashedi et al., 2009). The flowchart of PSO-GSA is shown in Fig. 1.4.

Figure 1.4 Flowchart of Hybrid PSO-GSA algorithm.

The details of the PSO-GSA-based optimization algorithm are as follows:

Step 1: A set of initial populations are created randomly within the minimum and maximum limits of the control variables and chosen as a parent population.

Step 2: The objective function for each particle in the initial population is evaluated.

Step 3: Calculation of the gravitational force, gravitational constant, and resultant forces among particles using (1.20), (1.21), and (1.22) respectively:

Gravitational force

(1.20)

Gravitational constant

(1.21)

Resultant forces

(1.22)

Step 4: The calculation of M acceleration for all particles of particles as defined in (1.23).

M acceleration

(1.23)

with

(1.24)

Step 5: The calculation of velocities of all particles using (1.26).

Velocities of all particles

(1.25)

with

(1.26)

Step 6: Update position of each particle according to (1.27).

Updating position

(1.27)

Step 7: The objective function of the new searching points and the evaluation values are calculated. The process of updating velocities and positions will be stopped by meeting the end criterion.

Step 8: If the stopping criterion is met (maximum number of generations is reached or the optimal point is achieved), then print the results. Otherwise, go to Step 2.

Step 9: Return the best solution.

is the gravitational forces from particle j on particle i at a specific time t; G is the total force acting on particle i in a dimension d; t is a specific time and Mi is the mass of object Iis the acceleration of all particles; pbest is the valuation of fitness function and gbest is the best particle among all particles. X is the particle coordinates; V is the velocity; W is the inertia weight factor; worst and best are respectively the worst and best fitness.

1.4 Stability index

In this section, the FVSI and Lmn are reviewed and used to estimate the maximum loadability and to identify the critical lines and buses to install the FACTS controllers.

1.4.1 Fast voltage stability index

The FVSI was proposed by I. Murisin et al. (Musinin and Abdul Rahman, 2002), and is based on the concept of power flow through a single line.

Taking the symbol i as the sending bus and j as the receiving bus, the FVSI can be defined by:

(1.28)

1.4.2 Lmn

The Lmn index was proposed by Moghavemmi and Omar (1998), and is formulated on the base of a power transmission concept in a single line.

The Lmn can be reproduced as:

(1.29)

It is important to note that the value of FVSI and Lmn must be kept lower than 1.00 in aim to preserve a stable system.

The steps implemented for identifying the critical buses and lines (Amrane et al., 2014) are described below.

1.5 Flexible alternating current transmission systems modeling

In this chapter, two typical FACTS devices have been used: SVC and TCSC.

1.5.1 Thyristor controlled series compensator model

The basic idea behind the power flow control with the TCSC is to decrease or increase the overall lines’ effective series transmission impedance, by adding a capacitance or inductance correspondingly (Amrane et al., 2014).

This model represents the TCSC by a variable reactance XTCSC. The active and reactive power injected to the nodes are represented by the following equations:

(1.30)

(1.31)

with

(1.32)

To avoid overcompensation, the working range of the TCSC should be limited between (Amrane et al., 2014):

(1.33)

1.5.2 Static volt ampere reactive compensator model

In this chapter, the SVC is modeled as a variable shunt reactive susceptance jbsvc installed at the node i (Fig. 1.6). In this case, only one term of the nodal admittances matrix is modified, corresponding to the node where the SVC is connected (Amrane et al., 2014).

The current generated or absorbed by SVC is represented based on the total susceptance by the following equation:

(1.34)

Reactive power injected by the SVC is presented as follows:

(1.35)

1.6 Simulation results

In order to verify the effectiveness of the proposed approach, the hybrid PSO and gravitational search algorithm (PSO-SGA) has been tested for the equivalent Algerian electric power system 114-bus (220/60 kV). For comparison purposes, two other algorithms are also implemented for solving the problem, namely PSO and GSA. Table 1.1 shows the parameters, number of iterations, and population size of these algorithms. The penalty factors in (1.15) are listed in Table 1.2. The programs have been written in MATLAB-7 language and executed on a 2.91 GHz CPU dual-core with 4 GO RAM.

Table 1.1

Table 1.2

To validate the effectiveness of the proposed approach, three case studies are considered:

• Case 1: Base case (nominal point).

• Case 2: Heavy case (20% of the base case).

• Case 3: Increased reactive power at critical nodes (10% of QMax FVSI).

1.6.1 Description of the test system and simulation results

To prove the robustness of the proposed technique in solving larger power systems, the equivalent Algerian electric power system is considered (Amrane et al., 2015). The equivalent network consists of 175 transmission lines, 15 generator-buses, 99 load buses, and 17 tap changer transformers. The total system’s real and reactive power demands are 3146.2 MW and 1799.4 MVar. The top 15 weakest buses and lines for the equivalent Algerian electric power system 114-bus are listed in Table 1.3. The chosen buses and lines which will receive the FACTS devices (SVCs and TCSCs devices) are listed in Table 1.4. The control variable limits and the description of the test systems are listed respectively in Tables 1.5 and 1.6. The lower and upper limits of the load-bus voltages were set to 0.90 and 1.1 p.u., respectively.

Table 1.3

Table 1.4

Table 1.5

Table 1.6

Tables 1.7–1.11 show the optimal settings of control variables obtained with the proposed method, PSO and GSA, using two types of FACTS devices (SVC and TCSC) for the three proposed case studies. From these tables we can see that all the control variables obtained by the proposed method and the two compared methods are within the safe range. Thus the load voltages obtained by the three algorithms using SVC and TCSC are given in Figs. 1.5–1.9. Also the load voltages obtained are within the permissible limits (0.9 and 1.1 p.u.). These results are encouraging and show the effectiveness of the proposed approach (Tables 1.12–1.15).

Table 1.7

Table 1.8

Table 1.9

Table 1.10

Table 1.11

Figure 1.5 Load voltage of Case 1 with SVC.

Figure 1.6 Load voltage of Case 1 with TCSC.

Figure 1.7 Load voltage of Case 2 with SVC.

Figure 1.8 Load voltage of Case 2 with TCSC.

Figure 1.9 Load voltage of Case 3 with SVC.

Table 1.12

Table 1.13

Table 1.14

Table 1.15

The minimum active power losses, voltage stability, and investment cost of FACTS devices obtained by the application of the proposed method, PSO, and GSA for the three case studies using the SVC and TCSC are listed in Tables 1.12 and 1.16. The results show that in all case studies, the minimum active losses found by the installation of SVC device are greater than those achieved by the installation of TCSC.

Table 1.16

In Case 1, the investments on SVCs devices of 1.96, 2.34, and 2.00 p.u. are respectively obtained by the application of PSO-GSA methods, PSO, and GSA, where the active power losses are reduced respectively by 15.80%, 7.616%, and 15.92%. The power losses are reduced by 3.03%, 5.77%, and 11.51% respectively by installing 0.69, 0.62, and 0.60 p.u. of TCSC.

Also we can show from the obtained results that the minimum voltage stability obtained by the application of the TCSC is greater than that reached by SVC. In Case 2, investments of 0.67, 0.70, and 0.65 p.u. of TCSC, respectively, are obtained by the application of PSO-GSA, PSO, and GSA. The minimum voltage stability is reduced to 0.40, 0.41, and 0.37 p.u., and by using SVC devices it is reduced to 0.69, 0.70, and 0.65 p.u.

The results obtained by increasing the reactive power at critical buses (Case 3), resulted in an increase of the installed reactive power. By comparing the results of Case 3 with the Case 1, we notice a net increase of reactive power. In Case 1, the total reactive energy invested by the proposed method is 1.96 p.u. and reactive power installed in Case 3 is 3.15 p.u. which shows the effectiveness and robustness of the proposed algorithm. The convergence characteristics of the proposed method, PSO, and GSA by the SVC and TCSC are respectively presented in Figs. 1.10–1.14.

Figure 1.10 Objective function convergence characteristic for Case 1 using SVC.

Figure 1.11 Objective function convergence characteristic for Case 1 using TCSC.

Figure 1.12 Objective function convergence characteristic for Case 2 using SVC.

Figure 1.13 Objective function convergence characteristic for Case 2 using TCSC.

Figure 1.14 Objective function convergence characteristic for Case 3 using SVC.

1.7 Conclusion

The hybrid PSO-GSA has been used for solving the ORPP problem using two kinds of the FACTS (SVC and TCSC) devices. Two stability indexes are used to determine the optimal placement of FACTS devices.

The obtained results show that the installation of SVC devices reduces the system active power losses more than by the installation of TCSC devices. On the other hand, to obtain good minimum voltage stability, TCSC is better than SVC devices.

The simulation results show the performance of the hybrid PSO-SGA algorithm which minimizes the FACTS devices amount, transmission power losses, and the system voltage stability, even in the most critical cases.

References

1. Amrane, Y., Boudour, M., 2015. Particle swarm optimization based reactive power planning for line stability improvement. In: 3rd International Conference on Control, Engineering & Information Technology, May, pp. 1–7.

2. Amrane, Y., Boudour, M. Belazzoug, M., 2013. Optimal reactive power planning based on particle swarm applied to the Algerian electrical power system, systems and control (ICSC). In: 2013 3rd International Conference on, pp. 804–809.

3. Amrane, Y., Boudour, M., Belazzoug, M., 2014. A New Hybrid Technique for Power Systems Multi-Facts Optimization Design. International Transactions on Electrical Energy Systems.

4. Amrane Y, Boudour M, Ladjici AA, Elmaouhab A. Optimal VAR control for real power loss minimization using differential evolution algorithm. Electr Power Energy Syst. 2015;66:262–271.

5. Ashouri M, Hosseini SM. Application of new hybrid particle swarm optimization and gravitational search algorithm for non convex economic load dispatch problem. J Adv Comput Res. 2013;4(2):41–51.

6. BelwinEdward J, Rajasekar N, Sathiyasekar K, Senthilnathan N, Sarjila R. An enhanced bacterial foraging algorithm approach for optimal power flow problem including FACTS devices considering system loadability. ISA Trans. 2013;52:622–628.

7. Grudinin N. Reactive power optimization using successive quadratic programming method Power systems. IEEE Trans on. 1998;13(4):1219–1225.

8. Hardiansyah. A novel hybrid PSO-GSA method for non-convex economic dispatch problems. I.J Inf Eng Electron Bus. 2013;5:1–9.

9. Jiang S, Ji Z, Shen Y. A novel hybrid particle swarm optimization and gravitational search algorithm for solving economic emission load dispatch problems with various practical constraints. Int J Electr Power Energy Syst. 2014;55:628–644.

10. Kennedy, J., 1995. Particle swarm optimization. In: Proceedings of the IEEE International Conference on Neural Networks, vol. IV, pp. 1942–1948.

11. Lenin K, Ravindranath Reddy B, Surya Kalavathi M. A new hybrid PSOGSA algorithm for solving optimal reactive power dispatch problem. Int J Mechatron Electr Comput Technol. 2014;4(10):111–125 ISSN: 2305-0543.

12. Mallick S, Ghoshal SP, Acharjee P, Thakur SS. Optimal static state estimation using improved particle swarm optimization and gravitational search algorithm. Int J Electr Power Energy Syst. 2013;52:254–265.

13. Mangaiyarkarasi SP, Raja TSR. Optimal location and sizing of multiple static VAR compensators for voltage risk assessment using hybrid PSO-GSA algorithm. Arab J Sci Eng. 2014;39(11):7967–7980.

14. Mirjalili, S., Hashim, S.Z.M., 2010, A new hybrid PSOGSA algorithm for function optimization. In: IEEE International Conference on Computer and Information Application, ICCIA 2010, China, pp. 374–377.

15. Moghavemmi M, Omar FM. Technique for contingency monitoring and voltage collapse prediction. EE Proc Gener Transm Distrib. 1998;145(6):634–640.

16. Musinin, I., Abdul Rahman T.K., 2002. Novel fast voltage stability index (FVSI) for voltage stability analysis in power transmission system. In: IEEE Student Conference Research and Development, Shah Alam, Malaysia, pp. 7803–7565.

17. Oliveira EJ, et al. An optimal power flow based on safety barrier interior point method. Electr Power Energy Syst. 2015;64:977–985.

18. Pourjafari E, Mojallali H. Predictive control for voltage collapse avoidance using a modified discrete multi-valued PSO algorithm. ISA Trans. 2011;50:195–200.

19. Precup, R.E., David, R.C., Stinean, A.I., Radac, M.B., Petriu, E.M., 2014. Adaptive hybrid particle swarm optimization-gravitational search algorithm for fuzzy controller tuning. In: Innovations in Intelligent Systems and Applications (INISTA) Proceedings, 2014 IEEE International Symposium on. IEEE, pp. 14–20.

20. Purcaru, C., Precup, R.-E., Iercan, D., et al., 2013. Hybrid PSO-GSA robot path planning algorithm in static environments with danger zones. In: System Theory, Control and Computing (ICSTCC), 2013 17th International Conference. IEEE, pp. 434–439.

21. Rashedi E, Nezamabadi-pour H, Saryazdi S. GSA: a gravitational search algorithm. Inf Sci. 2009;179:2232–2248.

22. Soliman, S.A.-H., Mantawy, A.-A.H., 2012. Modern Optimization Techniques With Applications in Electric Power Systems. Springer, Energy Systems.

Appendix

This section contains a state-of-the-art of the hybrid PSO-GSA technique that has arisen in the recent state-of-the-art literature (Table 1.A1).

Table 1.A1