Simulation Study for Optimized Demand Side Management in ......It is certiﬁed that Anzar Mahmood (Registration No. CIIT/SP 11-PEE-002/ISB) has carried out all the work related to

Simulation Study for Optimized Demand SideManagement in Smart Grid

ByAnzar Mahmood

CIIT/SP11-PEE-002/ISB

PhD Thesisin

Electrical Engineering

COMSATS Institute of Information TechnologyIslamabad - Pakistan

Spring 2015

COMSATS Institute of Information Technology


A Thesis Presented to

COMSATS Institute of Information Technology

In partial fulfillment

of the requirement for the degree of

PhD (Electrical Engineering)

By

Anzar Mahmood


Spring 2015

ii


A post graduate thesis submitted to the Department of Electrical Engineering as partial

fulfillment of the requirement for the award of degree of Ph.D(Electrical Engineering).

Name Registration Number

Anzar Mahmood CIIT/SP11-PEE-002/ISB

Supervisor

Dr. Nadeem Javaid

Associate Professor, Department of Computer Science,

COMSATS Institute of Information Technology,

Islamabad.

July, 2015

iii

Final Approval

This thesis titled


By

Anzar Mahmood


Has been approved

For the COMSATS Institute of Information Technology, Islamabad

External Examiner I:Prof. Dr. Muhammad Sher

Professor, Department of Computer Science & Software EngineeringInternational Islamic University, Islamabad

External Examiner II:Dr. Hasan Mahmood

Department of ElectronicsQuaid-i-Azam University, Islamabad

Supervisor:Dr. Nadeem Javaid

Department of Computer Science, Islamabad

HoD/Dean:Prof. Dr. Shahid A. Khan

Department of Electrical Engineering, Islamabad

Chairperson:Prof. Dr. Muhammad Junaid Mughal

Department of Electrical Engineering, Islamabad

iv

Declaration

I Anzar Mahmood (Registration No. CIIT/SP11-PEE-002/ISB)hereby declare that I have

produced the work presented in this thesis, during the scheduled period of study. I also declare

that I have not taken any material from any source except referred to wherever due that amount

of plagiarism is within acceptable range. If a violation of HEC rules on research has occurred

in this thesis, I shall be liable to punishable action under the plagiarism rules of the HEC.

Date: 6th July, 2015Anzar Mahmood


v

Certificate

It is certified that Anzar Mahmood (Registration No. CIIT/SP11-PEE-002/ISB)has carried

out all the work related to this thesis under my supervision at the Department of Electrical

Engineering, COMSATS Institute of Information Technology, Islamabad and the work fulfills

the requirement for award of PhD degree.

Date: 6th July, 2015

Supervisor:

Dr. Nadeem JavaidAssociate Professor, Department of Computer Science,

COMSATS Institute of Information Technology, Islamabad.

Head of Department:

Prof. Dr. Shahid A. KhanDepartment of Electrical EngineeringCOMSATS Institute of Information Technology, Islamabad.

vi

DEDICATION

Dedicated

to my parents, all the teachers, brothers, sisters, loving wife, children, all thefamily members

and to all those, known and unknown, who contributed in making me able toreach this prestigious level.

vii

Acknowledgements

After ALLAH Almighty, I would like to thank my supervisor, Dr. Nadeem Javaid, for his

continuous support and extended guidance. I felt very comfortable with him in discussions

about all the problems faced during the course and research work. Its an honour and pride for

me to be his student.

Special thanks to Dr. Sohail Razzaq, Naveed Ahmed Khan, Mohammad Asghar Khan, Faisal

Baig, Dr. Guftaar Ahmed, Dr. Raziq Yaqoob, Dr. Khurram Saleem Alimgeer, Adnan Za-

far, Sheeraz Ahmed, Ishfaq Ahmed, Kamran Latif, Mariam Akbar, Naeemullah, Mohammad

Babar, Khurram Shehzad, Saeed Ahmed, Zafar A. Khan, Ahsan Raza Khan, Rehman Zafar,

Hifsa Ashraf, Zain ul Abedin, Ahmad Hasan, and all those who helped me in one way or the

other.

I would also like to thank the COMSATS Institute of Information Technology for providing me

opportunity and support to conduct my research work at the department of Electrical Engineer-

ing Islamabad campus.

Anzar MahmoodCIIT/SP11-PEE-002/ISB

AbstractSimulation Study for Optimized Demand Side Management in Smart

Grid

Smart grid is envisioned to meet the 21st century energy requirements in a sophisticated manner

with real time approach by integrating the latest digital communications and advanced control

technologies to the existing power grid. It will dynamically connect all the stake holders of

smart grid through enhanced energy efficiency awareness corridor.

Smart Homes (SHs), Home Energy Management Systems (HEMS) and effect of home appli-

ances scheduling in smart grid are now familiar research topics in electrical engineering. Peak

load management and reduction of Peak to Average Ratio (PAR)and associated methods are

under focus of researchers since decades. These topics havegot new dimensions in smart grid

environment. This dissertation aims at simulation study for effective Demand Side Management

(DSM) in smart grid environment. This work is mainly focusedon optimal load scheduling for

energy cost minimization and peak load reduction.

This work comprehensively reviews the smart grid applications, communication technologies,

load management techniques, pricing schemes and related topics in order to provide an insight

to the environment required for dynamic DSM. Various network attributes such as Internet Pro-

tocol (IP) support, power usage, data rate etc. are considered to compare the communications

technologies in smart grid context. Techniques suitable for Home Area Networks (HANs) such

as ZigBee, Bluetooth, Wi-Fi, 6LoWPAN and Z-wave are discussed and compared in context of

consumer concerns and network attributes. A similar approach in context of utilities’ concerns

is adopted for wireless communications techniques for Neighborhood Area Networks (NANs),

which include WiMAX and GSM based cellular standards. Issues and challenges regarding

dynamic DSM in smart grid have been discussed briefly.

DSM is supposed to have a vital role in future energy management systems and is one of the

hot research areas. This study presents detailed review andanalytical comparison of DSM tech-

niques along with related technologies and implementationchallenges in smart grid. It also

covers consumers and utilities concerns in context of DSM toenhance the readers’ intuition

about the topic. Two major types of DSM schemes, incentive based and dynamic pricing based,

have been discussed and compared analytically. Dynamic pricing based HEMS are emphasized

as important tools for peak load reduction and consumers’ energy cost minimization. Dynamic

pricing based HEMS and their associated optimization techniques along with analytical com-

parison of the latest schemes have been described. Comparison of DSM techniques and study

of latest HEMS scheme provided the base for new ideas of partial baseline load and reserved

interrupting load to formulate two unique energy cost minimization problems. These models

resulted the following two solutions in which scheduling has been carried out through many

different algorithms to reduce peak load and consequently the PAR.

This work includes novel appliance scheduling solution named; Comprehensive Home Energy

Management Architecture (CHEMA), with multiple integrated scheduling options in smart grid

environment. Multiple layers of enhanced architecture aremodeled in Simulink with embed-

ded MATLAB code. Single Knapsack is used for scheduling and four different cases for cost

reduction are modeled. Fault identification and electricity theft control have also been added

along with the carbon foot prints reduction for environmental concerns. Simulation results have

shown the peak load reduction of 22.9% for unscheduled load with Persons Presence Controller

(PPC), 23.15% for scheduled load with PPC and 25.56% for flexible load scheduling. Simi-

larly total cost reduction of 23.11%, 24% and 25.7% has been observed, respectively. Smart

grid interface layer and load forecasting layers are not implemented in current work and will be

focused in future work.

Another novel comparative approach has also been proposed in this research, which investi-

gates the effect of multiple pricing schemes and optimization techniques for cost minimization

and peak load reduction. The proposed model uses multiple pricing schemes including Time

of Use (ToU), Real Time Pricing (RTP) day ahead case and Critical Peak Pricing (CPP). Pro-

posed optimization problem has been solved with multiple optimization techniques including

Knapsack, Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). Knapsack is used

with two options of limited slots scheduling and whole day scheduling. Comparative results

of the multiple pricing and optimization schemes have been discussed. Results show that the

best combination achieved with GA and CPP with 39.9223% costreduction. PSO showed the

43.73% cost reduction with all the pricing schemes.

The proposed schemes have many applications for peak load reduction and energy cost mini-

mization to benefit consumers and utilities. A user can schedule his load using one of the op-

tions provided in CHEMA according to his preferences. Similarly, maintenance activities can

be accommodated without disturbing the pre-defined schedule by using reserved interrupting

slots. In large buildings, reserved slots can be used to schedule heavy loads without generating

a peak.

x

List of ArticlesTotal No. of Articles = 36, Total Impact Factor = 37.072

List of Impact Factor Journal Publications

10. A. Mahmood, N. Javaid, S. Razzaq, “A Review of Wireless Communicationsfor Smart

Grid,” Renewable and Sustainable Energy Reviews, Vol.41, No. January 2015 pp. 248-

260, 2015.(Impact Factor: 5.901)

9. A. Mahmood, N. Javaid, A. Zafar, S. Ahmed, S. Razzaq, R. A. Riaz, “Pakistan’s Overall En-

ergy Potential Assessment, Comparison of LNG, TAPI and IPI Gas Projects,”Renewable

and Sustainable Energy Reviews, Vol.31, No. March 2014, pp. 182-193, 2014.(Impact

Factor: 5.901)

8. A. Mahmood, N. Javaid, S. Razzaq, M. A. Khan, “An Overview of Load Management

Techniques in Smart Grid,”International Journal of Energy Research(Impact Factor:

2.418)

7. M. A. Khan, N. Javaid,A. Mahmood, Z. A. Khan, N. Alrajeh, “A Generic Demand Side

Management Model for Smart Grid,”International Journal of Energy Research, 39(7),

pp. 954-964.(Impact Factor: 2.418)

6. A. Basit, G. A. S. Sidhu,A. Mahmood, “Efficient and Autonomous Energy Management

Techniques for the Future Smart Homes,”IEEE Transactions on Smart Grid,(Impact

Factor: 4.252)

5. A. R. Khan, A. Safdar,A. Mahmood, “Load Forecasting, Dynamic Pricing and DSM in

Smart Grid - A Review,”Renewable and Sustainable Energy Reviews, 54 (2016), 1311-

1322(Impact Factor: 5.901)

4. N. A. Khan, A. B. Awan,A. Mahmood, S. Razzaq, A. Zafar, G. A. S. Sidhu, “Combined

Emission Economic Dispatch of Power System Including SolarPhoto Voltaic Genera-

tion,” Energy Conversion and Management, 92(2015), pp. 82-91, 2015.(Impact Fac-

tor: 4.38)

xi

3. A. Mahmood, F. Baig, N. Javaid, “An Enhanced System Architecture with Flexible Load

Categorization for Optimized DSM in Smart Grid,”Under Revision in Applied Sciences,

(Impact Factor: 1.484)

2. A. Mahmood, N. Javaid, S. Razzaq, N. A. Khan “A Novel Comparative Approach for Home

Appliances Scheduling with Multiple Pricing and Optimization Schemes,” Submitted in

Energy(Impact Factor: 4.8)

1. R. Zafar, W. Ali, U. Naeem,A. Mahmood, K. Shehzad, N. Javaid, “Energy Management

through Social Networking in Smart Grid - A Review,” Submitted in Renewable and

Sustainable Energy Reviews.(Impact Factor: 5.901)

List of ISI Indexed Journal Publications5. F. Baig, A. Mahmood, N. Javaid, S. Razzaq, N. Khan, Z. Saleem, “Smart Home Energy

Management System for Monitoring and Scheduling of Home Appliances Using ZigBee,”

J. Basic Appl. Sci. Res.,3(5)880-891, 2013

4. S. I. Ali, M. Naeem,A. Mahmood, S. Razzaq, Z. Najam, S. Ahmed, S. H. Ahmed, “Methods

to Regulate Energy Consumption in Smart Homes,”J. Basic Appl. Sci. Res.,4(1)166-

172, 2014

3. M. N. Ullah, A. Mahmood, S. Razzaq, M. Ilahi, R. D. Khan, N. Javaid, “A Survey of Dif-

ferent Residential Energy Consumption Controlling Techniques for Autonomous DSM in

Future Smart Grid Communications,”J. Basic Appl. Sci. Res.,3(3)1207-1214, 2013

2. I. Khan, A. Mahmood, N. Javaid, S. Razzaq, R. D. Khan, M. Ilahi, “Home Energy Man-

agement Systems in Future Smart Grids, ”J. Basic Appl. Sci. Res.,3(3)1224-1231,

2013

1. M. U. Farooq,A. Mahmood, G. A. S. Sidhu, M. N. Ullah, Z. A. Khan, “Wind Power and

Smart Grid as an Environmental Obligation in Context of Energy Security for Pakistan,”

J. Basic Appl. Sci. Res.,3(9)518-527, 2013

xii

List of International Conference Proceed-ings

15. A. Mahmood, I. Khan, S. Razzaq, Z. Najam, N. A. Khan, M. A. Rehman, N. Javaid, “Home

Appliances Coordination Scheme for Energy Management (HoCS4EM) using Wireless

Sensor Networks in Smart Grids,”The 28th IEEE International Conference on Advanced

Information Networking and Applications, AINA- 2014, Victoria, Canada

14. A. Mahmood, M. N. Ullah, S. Razzaq, A. Basit, U. Mustafa, M. Naeem, N. Javaid, “A

New Scheme for Demand Side Management in Future Smart Grid Networks (SGNs),”

The 28th IEEE International Conference on Advanced Information Networking and Ap-

plications AINA-2014,Victoria, Canada

13. M. Anas, N. Javaid,A. Mahmood, “Minimizing Electricity Theft using Smart Meters

in AMI,” Seventh International Conference on P2P, Parallel, Grid, Cloud and Inter-

net Computing, 3PGCIC,pp.176-182, November 12-14, 2012, University of Victoria,

Victoria, Canada

12. N. Javaid, A. Sharif,A. Mahmood, “Monitoring and Controlling Power using ZigBee

Communications,”Seventh International Conference on Broadband, Wireless Comput-

ing, Communication and Applications, BWCCA,pp. 608-613, November 12-14, 2012,

University of Victoria, Victoria, Canada

11. M. N. Ullah, N. Javaid,A. Mahmood, “Residential Energy Consumption Controlling Tech-

niques to Enable Autonomous Demand Side Management in Future Smart Grid Com-

munications,” Eighth International Conference on Broadband and WirelessComputing,

Communication and Applications,2013, France

10. I. Khan, N. Javaid,A. Mahmood, “A Survey of Home Energy Management Systems in

Future Smart Grid Communications,”Eighth International Conference on Broadband

and Wireless Computing, Communication and Applications,2013, France

9. H. Ashraf, A. Hasan,A. Mahmood, Z. A. Khan, N. Javaid, “Peak Load Shaving Model

based on Individuals Habit,”IEEE 9th International Conference on Complex, Intelligent,

and Software Intensive Systems, CISIS-2015,Blumenau, Brazil

xiii

8. Z. U. Abedin, U. Shahid,A. Mahmood, N. Javaid “Application of PSO and GA for HEMS

in smart grid: A review,” IEEE 9th International Conference on Complex, Intelligent,

and Software Intensive Systems, CISIS-2015,Blumenau, Brazil

7. A. R. Khan, A. Safdar,A. Mahmood, “Load Forecasting and Dynamic Pricing based DSM

through HEMS- A Review,”IMTIC, 2015, Jamshor, Pakistan.

6. W. Ali, U. Naeem, R. Zafar,A. Mahmood, “Review of Social Networking based Energy

Management in Smart Grid,”IMTIC, 2015, Jamshoro, Pakistan

5. R. Zafar, U. Naeem, W. Ali,A. Mahmood, “Applications of ZigBee in Smart Grid: A

Review,” ICEET 2015, Superior University Lahore, Pakistan

4. Z. A. Khan,A. Mahmood, S. Razzaq, S. Ahmed, R. Nawaz, “Optimization based Individ-

ual and Cooperative DSM in Smart Grid Environment, ”PGSRET 2015, International

Islamic University, Islamabad, Pakistan

3. S. Ziauden,A. Mahmood, G. A. S. Sidhu, A. B. Awan, “A Simulink based Adaptive UFLS

Scheme,” IMTIC, 2015, Jamshoror, Pakistan

2. A. Arbaz, M. Waqas, K. Shehzad,A. Mahmood, “Home Energy Management and Knap-

sack Technique in Smart Grid Environment,”ICEET 2015, Superior University, Lahore,

Pakistan

1. M. Waqas,A. Mahmood, N. Javaid, S. Razzaq, “Optimized Energy Management System

using Electric Water Heater,”BWCCA, 2015, Poland

Additional Research Work6. A. Mahmood, H. Fakhar, N. Javaid “Analysis of Wireless Power Transmission,” IIISC

2014, International Industrial Information System Conference, IIISC, January 2014, Chi-

ang Mai, Thailand

5. S. Hayat, N. Javaid, A. Sharif,A. Mahmood, “Energy Efficient MAC Protocols,”IEEE 9th

International Conference on Embedded Software and Systems, HPCC-ICESS, pp. 1185-

1192, 2012, UK

xiv

4. A. Mahmood, A. Ismail, Z. Zaman, H. Fakhar, Z. Najam, M. S. Hasan, S. H. Ahmed,

“Comparative Study of Wireless Power Transmission Techniques,”J. Basic Appl. Sci.

Res.,4(1)321-326, 2014

3. A. Masood, Q. Hasan,A. Mahmood, “Flexible AC Transmission System Controllers: A

Review,” IMTIC, 2015, Jamshoro, Pakistan

2. M. S. Hasan,A. Mahmood, N. Khan, A. Munir, A. M. Fazal, Z. A. Khan, “Design and

Fabrication of Solar Thermal Battery using Molten Salt,”J. Basic. Appl. Sci. Res.,

3(6)1141-1150, 2013

1. M. A. Khan, A. Mahmood, M. Arif, M. N. Ullah, I. Khan, M. S. Hassan, A. Zafar, Z. A.

Khan, “Improvement in Perturb and Observe Method for Maximum Power Point Tracking

of Photovoltaic Panel,”J. Basic Appl. Sci. Res.,3(9)456-466, 2013

xv

Contents

Acknowledgements viii

Abstract ix

List of Articles xi

List of Figures xxi

List of Tables xxv

List of Symbols xxvii

1 Introduction 1

1.1 Energy Management and Electric Grid . . . . . . . . . . . . . . . . .. . . . . 1

1.2 Power Systems Basic Dynamics . . . . . . . . . . . . . . . . . . . . . . .. . 2

1.3 Smart Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4 Demand Side Management . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

1.4.1 Load Shedding Schemes . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.4.2 Incentive based DLC . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.4.3 Dynamic Pricing based HEMS . . . . . . . . . . . . . . . . . . . . . . 10

1.5 Smart Grid, DSM and Related Technologies . . . . . . . . . . . . .. . . . . . 11

xvi

1.6 The Research Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

1.7 Research Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 17

1.8 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 17

2 Related Work 19

2.1 DSM and Smart Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3 Energy and Electricity Situation in Pakistan as a Pretext for Smart Grid Imple-

mentation 28

3.1 Summary of the Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28

3.2 Current Energy Scenario of Pakistan . . . . . . . . . . . . . . . . .. . . . . . 29

3.3 Major Issues of the Energy Sector . . . . . . . . . . . . . . . . . . . .. . . . 31

3.3.1 Decreased Hydel Share in the Country’s Energy Mix . . . .. . . . . . 31

3.3.2 De-Rated Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.3.3 Circular Debt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.3.4 Management Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.4 Future Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 36

3.5 Energy Security and Import Options . . . . . . . . . . . . . . . . . .. . . . . 38

3.6 Energy Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 41

3.6.1 Hydro Power Potential . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.6.2 Solar Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.6.3 Wind Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.6.4 Bio Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3.6.5 Nuclear Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

3.6.6 Geo-Thermal Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.6.7 Hydrogen based Energy Resources . . . . . . . . . . . . . . . . . .. 47

3.6.8 Thar Coal and Coal Gasification . . . . . . . . . . . . . . . . . . . .. 48

xvii

3.6.9 Ocean Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.6.10 Analysis of Potential . . . . . . . . . . . . . . . . . . . . . . . . . .. 49

3.7 Smart Grid Implementation in Pakistan . . . . . . . . . . . . . . .. . . . . . 50

3.8 Conclusions of the Chapter . . . . . . . . . . . . . . . . . . . . . . . . .. . . 51

4 Smart Grid Applications and Technologies 53


4.2 Smart Grid Applications . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 54

4.2.1 AMI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

4.2.2 HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.2.3 DA and DERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

4.2.4 EVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

4.3 Wireless Communication Options for HANs . . . . . . . . . . . . .. . . . . . 61

4.3.1 ZigBee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

4.3.2 WLAN and Wi-Fi . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

4.3.3 Bluetooth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.3.4 6LoWPAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4.3.5 Z-Wave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4.3.6 Comparative Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.4 Wireless Communication Options for NANs . . . . . . . . . . . . .. . . . . . 67

4.4.1 WiMAX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4.4.2 Cellular Network Communications . . . . . . . . . . . . . . . . .. . 70

4.4.3 Comparison of Wireless NAN Technologies . . . . . . . . . . .. . . . 71

4.5 Dynamic Pricing Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . .. 73

4.5.1 ToU Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.5.2 CPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

xviii

4.5.3 RTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

4.6 Smart Grid and DSM Challenges . . . . . . . . . . . . . . . . . . . . . . .. . 76


5 DSM and Optimization based Appliances Scheduling 82


5.2 Optimization and Smart Appliances Scheduling . . . . . . . .. . . . . . . . . 83

5.2.1 Knapsack Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

5.2.2 PSO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

5.2.2.1 PSO Algorithm . . . . . . . . . . . . . . . . . . . . . . . . 85

5.2.3 GA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

5.3 Comparative Study of Optimization based ECS Schemes . . .. . . . . . . . . 89


6 Enhanced System Architecture for Optimized DSM in Smart Grid 96


6.2 Proposed CHEMA and its Implementation . . . . . . . . . . . . . . .. . . . . 97

6.2.1 Load Categorization . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

6.2.2 Energy Cost Optimization Model . . . . . . . . . . . . . . . . . . .. 101

6.2.3 Pricing Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

6.2.4 Space Heating Module . . . . . . . . . . . . . . . . . . . . . . . . . . 102

6.2.5 Water Heater and Refrigerator Modules . . . . . . . . . . . . .. . . . 104

6.2.6 TFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

6.2.7 GEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

6.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 109

6.3.1 Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

6.3.2 Case 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

xix

6.3.3 Case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.3.4 Case 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.3.5 Aggregated and Total Energy Consumption Results . . . .. . . . . . . 118


7 A Novel Approach for HEMS Scheduling with Multiple Pricing and Optimization

Techniques 125


7.2 The Proposed Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

7.3 Simulation Model: Results and Discussion . . . . . . . . . . . .. . . . . . . . 128

7.3.1 Knapsack Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

7.3.1.1 Knapsack Case One with ToU . . . . . . . . . . . . . . . . 129

7.3.1.2 Knapsack Case One with RTP . . . . . . . . . . . . . . . . 130

7.3.1.3 Knapsack Case One with CPP Results . . . . . . . . . . . . 132

7.3.2 Knapsack Case Two . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

7.3.2.1 Knapsack Case Two With ToU . . . . . . . . . . . . . . . . 133

7.3.2.2 Knapsack Case Two With RTP . . . . . . . . . . . . . . . . 134

7.3.2.3 Knapsack Case Two with CPP . . . . . . . . . . . . . . . . 135

7.3.3 PSO Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

7.3.3.1 PSO with ToU . . . . . . . . . . . . . . . . . . . . . . . . . 136

7.3.3.2 PSO with RTP . . . . . . . . . . . . . . . . . . . . . . . . . 138

7.3.3.3 PSO with CPP . . . . . . . . . . . . . . . . . . . . . . . . . 139

7.3.4 GA Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

7.3.5 Energy Consumption Results . . . . . . . . . . . . . . . . . . . . . .. 141


8 Conclusions and Future Work 151

xx

8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

8.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Bibliography 155

xxi

List of Figures

1.1 Multiple Interactions among Major Stakeholders of Smart Grid . . . . . . . . . 3

1.2 Breaker Interlocking Load Shedding Scheme. . . . . . . . . . .. . . . . . . . 7

1.3 Smart Grid Network Connecting Different Stakeholders .. . . . . . . . . . . . 14

1.4 Major Concerns of Electric Utility Companies . . . . . . . . .. . . . . . . . 14

1.5 Major Concerns of Electricity Consumers . . . . . . . . . . . . .. . . . . . . 15

3.1 Major Issues of Pakistan’s Energy Market . . . . . . . . . . . . .. . . . . . . 31

3.2 Reasons of Poor Revenue Collection . . . . . . . . . . . . . . . . . .. . . . . 34

3.3 Mechanism of Circular Debt . . . . . . . . . . . . . . . . . . . . . . . . .. . 35

3.4 Projected Population and Electricity . . . . . . . . . . . . . . .. . . . . . . . 38

3.5 Proposed Sites for Hydro Power Generation . . . . . . . . . . . .. . . . . . . 42

3.6 World Solar Insolation Map . . . . . . . . . . . . . . . . . . . . . . . . .. . . 44

3.7 Wind Map of Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.1 AMI Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

4.2 Interaction of HEMS with Various Smart Entities . . . . . . .. . . . . . . . . 57

4.3 Distribution Automation with Central Control and DERs .. . . . . . . . . . . 59

4.4 EVs Management System . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60

5.1 PSO Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

xxii

5.2 GA Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

6.1 Proposed CHEMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

6.2 CHEMA Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

6.3 ToU Pricing Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

6.4 Thermostat Subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 103

6.5 Heater Subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

6.6 Water Heater Complete Model . . . . . . . . . . . . . . . . . . . . . . . .. . 105

6.7 Water Heater Subsystem A . . . . . . . . . . . . . . . . . . . . . . . . . . .. 105

6.8 Refrigerator Complete Model in Simulink . . . . . . . . . . . . .. . . . . . . 106

6.9 Refrigerator Wall Temperature Subsystem . . . . . . . . . . . .. . . . . . . . 107

6.10 Fault Occurrence and Detection . . . . . . . . . . . . . . . . . . . .. . . . . 107

6.11 Current and Voltage Characteristics at TFD . . . . . . . . . .. . . . . . . . . 108

6.12 Pricing Scheme and Total Unscheduled Cost (Case 1) . . . .. . . . . . . . . . 111

6.13 HVAC Cost and Temperature (Case 1) . . . . . . . . . . . . . . . . . .. . . . 111

6.14 Refrigeration Cost, Power and Temperature Variation (Case 1) . . . . . . . . . 112

6.15 Water Heater Cost, Power and Temperature Variation (Case 1) . . . . . . . . . 112

6.16 Number of Lights and Cost Variation (Case 1) . . . . . . . . . .. . . . . . . . 113

6.17 PPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.18 Pricing Scheme and Total Cost (Case 2) . . . . . . . . . . . . . . .. . . . . . 114

6.19 HVAC Cost and Temperature Variation (Case 2) . . . . . . . . .. . . . . . . . 114

6.20 Lighting Cost (Case 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 115

6.21 Total Cost (Case 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 116

6.22 UCI Results (Case 4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 117

6.23 Refrigeration Cost (Case 4) . . . . . . . . . . . . . . . . . . . . . . .. . . . . 118

6.24 Aggregated Energy Consumption (Case 1) . . . . . . . . . . . . .. . . . . . . 119

xxiii




6.28 Comparison of Total Energy Consumption of Four Cases . .. . . . . . . . . . 122

6.29 Impact of Partial Baseline Load Inclusion in the Optimization Model . . . . . . 123

7.1 ToU Pricing Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

7.2 Comparison of Scheduled Load and Non-Scheduled Load with Cost . . . . . . 130

7.3 Generated Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 131

7.4 RTP Day Ahead Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

7.5 Comparison of Scheduled and Non-Scheduled Load with Cost . . . . . . . . . 132

7.6 CPP Used with Knapsack . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132

7.7 Comparison of Scheduled and Non-Scheduled Load with Cost for Knapsack

and CPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

7.8 Comparison of Scheduled and Non-Scheduled Load with ToUover 24 Hours . 133

7.9 Schedule Generated over 24 Hours with Knapsack and ToU . .. . . . . . . . . 134

7.10 Partial Baseline Load Reduction During Peak Hours . . . .. . . . . . . . . . . 134

7.11 Comparison of Scheduled and Non-Scheduled Load with RTP over 24 Hours . 135

7.12 Comparison of Scheduled and Non-Scheduled Load with CPP over 24 Hours . 135

7.13 Comparison of Scheduled and Non-Scheduled Load using PSO with ToU . . . 137

7.14 Schedule Generated with PSO and ToU . . . . . . . . . . . . . . . . .. . . . 137

7.15 RI Load Slots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

7.16 RI Load with Greater Number of Slots and Variation of Slot Numbers . . . . . 138

7.17 Constraint Violation in Case of PSO with RTP . . . . . . . . . .. . . . . . . . 139

7.18 Comparison of Scheduled and Non-Scheduled Load using PSO with CPP . . . 139

7.19 Cost Minimization using GA with ToU . . . . . . . . . . . . . . . . .. . . . . 140

7.20 Schedule Generated using GA with ToU . . . . . . . . . . . . . . . .. . . . . 140

xxiv

7.21 Cost Minimization using GA with RTP . . . . . . . . . . . . . . . . .. . . . . 141

7.22 Cost Minimization using GA with CPP . . . . . . . . . . . . . . . . .. . . . . 141

7.23 Slot-Wise Energy Consumption (Knapsack case one with ToU) . . . . . . . . . 144

7.24 Aggregated Energy Consumption (Knapsack case one withToU) . . . . . . . . 145

7.25 Slot-Wise Energy Consumption (Knapsack case two with RTP) . . . . . . . . . 145

7.26 Aggregated Energy Consumption (Knapsack case two withRTP) . . . . . . . . 146

7.27 Slot-Wise Energy Consumption (PSO with CPP) . . . . . . . . .. . . . . . . 146

7.28 Aggregated Energy Consumption (PSO with CPP) . . . . . . . .. . . . . . . 147

7.29 Cost Minimization using BIP with ToU . . . . . . . . . . . . . . . .. . . . . 147

7.30 Cost Minimization using BIP with RTP . . . . . . . . . . . . . . . .. . . . . 148

7.31 Cost Minimization using BIP with CPP . . . . . . . . . . . . . . . .. . . . . 148

7.32 Cost Minimization using WDO with ToU . . . . . . . . . . . . . . . .. . . . 149

7.33 Cost Minimization using WDO with RTP . . . . . . . . . . . . . . . .. . . . 149

7.34 Cost Minimization using WDO with CPP . . . . . . . . . . . . . . . .. . . . 150

xxv

List of Tables

1.1 Comparison of Dynamic Pricing based HEMS and Incentive based DLC . . . . 12

3.1 Comparison of Per Capita Electricity Consumption for Some Countries . . . . 36

3.2 Comparison of TAPI and IPI Gas Projects . . . . . . . . . . . . . . .. . . . . 40

3.3 Large Isolated Power Plants Operating on Solar Technologies . . . . . . . . . . 43

3.4 List of Countries with Top Ten Wind Power Capacities . . . .. . . . . . . . . 45

3.5 Comparison of Nuclear Power Generation for Some Countries . . . . . . . . . 46

3.6 Potential of Different Resources in Pakistan . . . . . . . . .. . . . . . . . . . 50

4.1 Comparison of HAN Technologies for Smart Grid . . . . . . . . .. . . . . . . 68

4.2 Comparison of NAN Technologies for Smart Grid . . . . . . . . .. . . . . . . 72

4.3 Smart Grid Challenges and Issues . . . . . . . . . . . . . . . . . . . .. . . . 80

5.1 Comparison of Different Schemes for HEMS . . . . . . . . . . . . .. . . . . 94

6.1 GHGs Emission and Electricity Consumption Equivalency. . . . . . . . . . . 110

6.2 Cost Comparison of Different Cases . . . . . . . . . . . . . . . . . .. . . . . 117

6.3 Carbon Emission Reduction of Different Cases . . . . . . . . .. . . . . . . . 118

6.4 Load Comparison of Different Cases . . . . . . . . . . . . . . . . . .. . . . . 121

7.1 Comparative Results of Different Techniques . . . . . . . . .. . . . . . . . . 150

xxvi

List of Abbreviations

ACS: Appliance Coordination System

ADR: Automated Demand Response

AMI: Advanced Metering Infrastructure

AMR: Automatic Meter Reading

CPP: Critical Peak Pricing

CHEMA: Comprehensive Home Eneregy Management Architecture

CPSO: Co-Evolutionary Particle Swarm Optimization

DA: Distribution Automation

DER: Distributed Energy Resources

DG: Distributed Generation

DLC: Direct Load Control

DR: Demand Response

DSM: Demand Side Management

DST: Decision Support Tool

ECC: Energy Consumption Controlling

ECS: Energy Consumption Scheduling

xxvii

EDGE: Enhanced Data Rates for GSM Evolution

EMS: Energy Management Systems

EMU: Energy Management Unit

EVs: Electric Vehicles

EPRI: Electric Power Research Institute

FFDs: Full Function Devices

GA: Genetic Algorithm

GHGs: Green House Gases

GPRS: General Packet Radio Service

GSM: Global System for Mobile Communications

GTS: Guaranteed Time Slots

HACS4EM: Home Appliances Coordination Scheme for Energy Management

HAN: Home Area Network

HEMS: Home Energy Management Systems

IBR: Inclining Block Rates

ICTs: Information and Communications Technologies

IETF: Internet Engineering Task Force

iHEM: In-Home Energy Management

ISM: Industrial, Scientific and Medical

ISO: Independent System Operators

LAN: Local Area Network

LP: Linear Programming

LTC: Load Tap Changer

xxviii

MAC: Medium Access Control

MIMO: Multiple-Input Multiple-Output

NAN: Neighborhood Area Network

NIST: National Institute of Standards and Technology

NTDC: National Transmission and Dispatch Company

OFDM: Orthogonal Frequency Division Multiplexing

OLM: Optimum Load Management

OREM: Optimization based Residential Energy Management

PAN: Personal Area Network

PAR: Peak to Average Ratio

PHEV: Plug in Hybrid Electric Vehicle

PHY: Physical

PSO: Particle Swarm Optimization

PV: Photo Voltaic

QoS: Quality of Service

RF: Radio Frequency

RFDs: Reduced Function Devices

RTP: Real Time Pricing

RTUs: Remote Terminal units

SC-FDMA: Single Carrier Frequency Division Multiple Access

SD: Super-frame Duration

SIM: Subscriber Identification Module

SIP: Session Initiation Protocol

xxix

ToU: Time of Use

UMTS: Universal Mobile Telecommunications System

V2I: Vehicle-to-Infrastructure

VANETs: Vehicle Ad Hoc Networks

VCG: Vickrey-Clarke-Groves

WAMR: Wireless Automatic Meter Reading

WAN: Wide Area Network

WCTs: Wireless Communications Technologies

WSNs: Wireless Sensor Networks

WLAN: Wireless Local Area Network

WMAN: Wireless Metropolitan Area Network

WPAN: Wireless Personal Area Network

WSHAN: Wireless Sensor Home Area Network

xxx

Chapter 1Introduction

1.1 Energy Management and Electric Grid

Energy is one of the most important components of human life, which is present ubiq-

uitously and can be rendered as soul of modern machine age. Energy management is

important and interesting focus of researchers since decades. According to the Inter-

national Energy Agency (IEA), the total primary energy supply of the world has been increased

to 13,541 Million Tons of Oil Equivalent (MTOE) in 2013 as compared to 6,106 MTOE in

1973 [1]. The largest amount of energy is consumed in buildings, which is estimated almost

40% of the world’s total consumption and has been doubled in 2010 as compared to 1971 [2].

It is the use of energy by buildings’ occupants, which is usually expressed as per capita energy

consumption and taken as an index of development and prosperity for a country [3].

Renewable energy resources are cleaner form of energy; however, the energy density and tech-

nological maturity of the fossil fuels dominate their use. It is estimated that the US energy

profile will constitute 33% of renewable resources by 2020 with high penetration of wind and

solar power plants [4]. Integration of the huge amount of Distributed Generation (DG), usually

consisting of renewable resources, to the power grid has raised many issues [5].

Services of the traditional grid have been used by the humanity since decades which consists

of four major parts: electricity generation, transmission, distribution and utilization. Global

population and the dependency level of humans on electricity are increasing exponentially [6].

The existing electric grid is of electro-mechanical natureand is characterized by one way com-

munication of data and power among users and utilities. The existing grid is not supposed to

1

1.2. Power Systems Basic Dynamics

meet the 21st century power quality and reliability requirements. Basicpower system dynamics

of the existing grid are described in the subsequent section.

1.2 Power Systems Basic Dynamics

Power system has two basic parameters: voltage level and frequency; these should be observed

continuously for system stability and reliability. Increasing gap between demand and supply

due to insufficient generation or transmission capability will cause overloading of system. Over-

loading will cause under frequency and under voltage conditions which will ultimately affect

the system’s stability.

There are two main mechanisms of power system control. Firstis the excitation system along

with automatic voltage regulator to control the generator reactive power and system’s voltage.

Second is the prime-mover control which is used for generator active power and frequency con-

trol. In addition to overloading disturbances; faults may also cause the transients and affect

system dynamics.

Usually the power system is controlled by central control station. All the generated power is

transmitted through different transmission lines of 500kV, 220 kV, 132 kV etc. The central

control station, with the help of regional control centers,controls the load shedding programs

using 132 kV substations. 132 kV substations step down the voltage at 11 kV. The distribu-

tion is completed through 11 kV feeders. Feeders consist of several 11 kV to low voltage step

down transformers which finally give the connections to consumers. In case of over loading, the

system is controlled through opening the breakers of 11 kV feeders manually from substation

which causes black out in whole area of feeder. The voltage levels described above are quite

general and may vary in different countries.

The present electricity infrastructure is a complex and aging system characterized by central-

ized power generation and distribution, one way power flow and lack of user-utility interaction

which leads to energy loss, overload conditions, power quality issues, poor peak load manage-

ment, lack of renewable energy usage, time wastage and manual operational processes. This

along with foreseen decline in fossil fuels availability, rise in fuel cost, related environmental

issues like global warming from greenhouse emissions and rising demand for electricity require

re-envisioning of the traditional electricity grid [7]. So, there exists a huge potential of research

and innovations for conversion of traditional grid into smart grid in order to meet the challenges

2

1.3. Smart Grid

Figure 1.1: Multiple Interactions among Major Stakeholders of Smart Grid

of 21st century [8].

1.3 Smart Grid

Smart Grid is an integration of the advanced Information andCommunication Technologies

(ICTs) to the existing electro-mechanical power systems [9]. The bi-directional flow of data

and power between utility and end users is one of the main characteristics of smart grid aimed

at demand management in an efficient and dynamic way [4]. In other words, smart grid intends

to put forward the existing power system infrastructure by exploiting todays state of the art

ICTs, digital networks and advanced control mechanisms. With the rapid growth in DG and

customer demands such as Plug-in Hybrid Electric Vehicles (PHEVs), several issues concern-

ing electric utilities and consumers have been raised including but not limited to power quality,

reliability, security, efficiency and power outages [10].

Rapid advancements in control, ICTs have allowed the conversion of traditional electricity grid

into smart grid that ensures productive interactions amongenergy providers (utilities), con-

sumers and other stakeholders [11]. These multiple and enhanced interactions, shown in Fig.1.1,

will help solving the issues raised in existing grid. Key components of smart grid are smart

meters, sensors, monitoring systems and data management systems that control the flow of

3

1.4. Demand Side Management

information among various stakeholders, making it a two waycommunications network, also

called AMI [12]. Other smart grid applications include EMS,Distributed Power Generation

(DPG) and its reliable integration to the system, equipmentdiagnostics, control, overall opti-

mized asset management etc. PHEVs and EVs have important effects regarding reliability of

grid and effective management of these vehicles needs intensive research. All of these applica-

tions strongly rely on communications infrastructure. HEMS require a short distance network,

called HAN. Communication between users and utilities needs a NAN and may also need a

Wide Area Network (WAN). Our objective is to explore smart grid fundamentals, applications,

and various short & long range communications technologiesthat can be applied to key smart

grid applications.

Real time monitoring and control will become possible in future smart grids by integration of

advanced ICTs. Provision of reliable energy services in context of increased environmental

concerns will become possible in future smart grids [13]. Various technologies are being devel-

oped by researchers for realization of smart grid including: AMI, HANs, DA, etc. [14].

Applications of smart grid include various areas like transmission and distribution automation,

optimized utilization, commodity trading of electricity in competitive markets etc. Our focus

is home energy management which can be optimized in smart grid environment [15]. Home

appliances equipped with sensors along with AMI make the real time energy monitoring possi-

ble which is a parameter of extreme importance for utilities. It will help in reduction of carbon

footprint caused by the excessive use of the peaker plants; abig source of greenhouse emis-

sions [16].

1.4 Demand Side Management

The need to control the demand in order to shape the load profile was first realized in1970s [17].

Now it has evolved to the concept of Demand Side Management (DSM) and is characterized

by utility operations and incentives for the consumers in order to bring power usage at desired

level at all times. Major objectives of DSM include: peak clipping, valley filling, peak shifting

and deploying new efficient uses [6]. DSM can help the consumers to lower their payments

and utility to minimize the need of peaking plants. Obviously, the utility desires the shape of

the load curve to be balanced with a reduced Peak to Average Ratio (PAR) for all the hours

4

1.4.1. Load Shedding Schemes

while consumers want reliable energy supplies at minimum cost. In literature, Load Manage-

ment (LM), Demand Response (DR) and DSM are found as overlapping concepts and are used

interchangeably [18].

There are three major types of DSM programs: load shedding, dynamic pricing based schemes

and incentive based schemes. Load shedding, also known as rolling blackout or feeder rotation,

is the process of disconnecting the load at feeder level in order to reduce the over load condition

on grid. Pricing based programs include Real Time Pricing (RTP), Time of Use (ToU) pricing,

Critical Peak Pricing (CPP) etc., whereas incentive based programs include Direct Load Control

(DLC), curtailable services, demand bidding etc. [19]. DLCacts only when the peak demand

crosses the certain limit whereas the dynamic pricing is an ongoing phenomenon. Since it is

difficult to respond the dynamic pricing schemes manually, the customers need Home Energy

Management Systems (HEMS) in order to automatically respond the price variations through

scheduling of their appliances for optimal cost. In contrast to above mentioned conventional

power system, modern dynamic EMS consists of the four major parts: smart end-use-devices,

smart distributed energy resources, modern building control systems and integration of the ad-

vanced communication infrastructure for bi-directional flow of data and power between users

and utility [9]. Major techniques of DSM are described in subsequent sections.

1.4.1 Load Shedding Schemes

Load shedding is the conventional DSM technique which is also known as rotational or rolling

blackout. It is the last measure taken by utility at feeder level to bridge the gap between supply

and demand and to maintain the minimum level of systems reliability and stability [20]. There

are two main problems associated with load shedding. First is the amount of load to be shed

and other is location of the load to be shed. Load shedding also helps in mitigating the after

effects of disturbances. Conventional method of load shedding consists of shedding of constant

amount of load according to pre-defined schedule which is made in light of demand curve and

available generation capacity. This practice is also applied in emergency conditions. However

modern adaptive load shedding schemes use different techniques to avoid under or over shed-

ding. Total amount of over load and the load to be shed can be measured by following formulae,

5


respectively [21].

L =Total Load−Total Generation

Total Generation(1.1)

LD =L

1+L −d(1− (Fmin/Fn))

1−d(1− (Fmin/Fn)), (1.2)

whereasLD is total amount of load to be shed againstL, d is the load reduction factor,Fmin is

the minimum allowable frequency andFn is the nominal frequency. In energy sector, dynamic

demand control is needed in order to save cost for utilities and the consumers. Reduced cost and

increased comfort level are the main incentives that motivate consumers to be part of effective

energy management.

Many schemes of load shedding are found in literature. Mustafa used the sum of squares of

the difference between the connected load and the supplied active and reactive power in [22]

for effective implementation of load shedding. The scheme included response based and event

based algorithms in order to introduce the adaptation; simulation results confirmed the suit-

ability of the scheme. A Programmable Logic Controller (PLC) based load control for energy

management is introduced in [21] which intelligently selects the loads to be shed with the help

of knowledge base attached with PLC. The fast response and adaptive nature is proved by ap-

plying the scheme on a mining power system as a case study. System losses are usually ignored

in the research schemes. In [23], authors propose a mathematical model which emphasizes the

minimization of system losses for optimal load shedding. The system is modeled in MATLAB

and simulation results show the effects of loss inclusion tothe overall load control mechanism

and system’s reliability. Idea of using voltage, frequencyand rate of change of these two factors

to calculate the amount of load to be disconnected has been presented in [24] for effective load

shedding.

In conventional load shedding schemes, amount and locationof the load to be shed is constant

which results in excessive or under shedding of the load. These schemes are also responsible

for blackouts. Many load shedding schemes have been reported in the literature. One of the

simplest schemes is known as breaker interlock load shedding scheme. In this simple method

a main breaker is inter locked with load breakers [25]. When the supply of the main breaker

fails, load breakers automatically open because of inter locking. The system is described in Fig.

1.2. No calculation of the load to be shed is required in this method as it is pre decided. Hence

this method is fast enough and the interlocking may be done using wires or remote signals.

Another method in conventional power grid is based on systemfrequency monitoring. When

6


UTILITY

MAIN

BREAKER

MAIN

TRANSFORMER

LOAD 3

LOAD BREAKER

LOAD 2

LOAD BREAKER

LOAD 1

LOAD BREAKER

BUS BAR

WIRED OR

WIRELESS

CONNECTION

Figure 1.2: Breaker Interlocking Load Shedding Scheme.

frequency gets lower than preset level, the system sheds thepre-defined load to restore the rated

frequency. This type of load shedding is known as under frequency load shedding. In under

frequency load shedding schemes, under frequency relays sense the fast frequency change or

slow decay of frequency at various critical points of the system. In case of reaching a preset

value of frequency, these relays delay for a pre-defined time. After this delay, first stage of

opening of interlocked breakers starts. If the system frequency is not restored then second

stage is initiated at second set point. This scheme has slow time response because of the delay

introduced and it may shed incorrect amount of load [24].

7

1.4.2. Incentive based DLC

1.4.2 Incentive based DLC

In beginning, DSM procedures were based on unidirectional communication between users and

utilities [26]. These procedures were mainly implemented by the utilities and the role of users

in DSM programs was negligible. However, integration of advanced communication infras-

tructure enables bi-directional flow of data and power amongdifferent stakeholders of power

system and hence allows more efficient DSM involving both utility and consumers [27]. Such

DSM programs are divided into two major categories: incentive based and dynamic pricing

based schemes. The most prominent incentive based pricing scheme is DLC which is discussed

in this section.

In DLC, the utility takes over the control and has an authority to shut down or cycle consumers

electrical appliances (depending on the contractual terms). Incentive based DSM programs like

DLC pay the incentive money to the consumer for the time whichthey are asked to reduce or

shutdown the load on short notice during peak period [28, 29]. In USA, one third electricity

is used by residential consumers of electricity and the important loads that contribute to the

higher peaks are thermostatically driven i.e. Air conditioners (ACs), space and water heaters

etc. [29, 30].

There are two types of loads: responsive and non-responsive. The responsive loads respond

to the signals of the utility while non-responsive loads do not communicate with utility [28].

Usually the air conditioners load is reduced by increasing the set point in a predefined manner

or by limiting the cycle run time. It has been observed that DLC of residential ACs reduce a

significant peak of the load [29]. However, a major problem with the DLC is the comfort level

of the consumer. It is very difficult to implement DLC withoutcausing inconvenience to the

consumer. Although consumers are paid an inconvenience cost but curtailment of a facility at

the exact moment when it is required the most (ACs in a sunny day), causes great discomfort

to the consumers [29–31]. Furthermore, after the peak time,another spike in the demand is

observed during such days as all the ACs that are turned off during the event are turned on

simultaneously to improve the temperature of the rooms. This phenomenon is called rebound

effect [29, 32].

Ripple control is a DLC technique and it has been applied in many countries like New Zealand,

Czech Republic, Germany, United Kingdom etc. [33]. In this method a high frequency (Usu-

ally 0.1 KHz to 1.5 KHz) signal is imposed on the standard 50 or60 Hz electricity signal.

8

1.4.2. Incentive based DLC

Receiver devices attached to the target loads shut down the loads upon receiving high frequency

signal. The load is kept shut until the signal is disabled. Ripple control technique was first

implemented in France in 1928 [18]. It was implemented as power line communication and

then evolved to radio ripple control.

In [28], a DLC based scheme is presented for DSM that includesan algorithm aimed at min-

imizing users’ energy cost as well as maximizing their comfort level. A controller is used to

execute the proposed algorithm in order to optimize customers energy price and comfort level

while taking into account current and near future tariffs and appliance constraints. Simulation

results in [28] prove that customers convenience is the mostprominent factor to decide their

involvement in DLC scheme.

Authors in [31] have proposed an interesting network for DLCbased load control which exe-

cutes an algorithm to maximize energy cost efficiency while meeting critical constraint on users

thermal comfort. The algorithm employs least enthalpy estimation based thermal comfort con-

trol of AC units in order to manage load efficiently while keeping thermal comfort level within

acceptable range. Furthermore, in [31] , the concept of group DLC is introduced in order to

reduce the effects of DLC constraints on air conditioning loads and to generate an optimized

schedule simultaneously.

[30] presents a stochastic optimization algorithm that controls the load at feeder level and min-

imizes the amount of load being controlled. It also ensures the minimum disruption in the

operation of electric appliances. A Monte Carlo simulationframework offers insight into the

benefits of employing the proposed algorithm. Results show the impact of different constraints

and parameters on effectiveness of the control mechanism.

Selective DLC for effective DSM is applied on a case study of 33 kV system of UK as reported

in [34]; the system is modeled in PSCAD, and magnitude of loadto be shed is calculated using

swing equation. A laboratory prototype for selective DLC isproposed in [35] which is based

on intelligent circuit breakers that implement selective load control. An innovative load control

is described in [36] with the pre specified timing of different breakers for different categories of

loads. Incentive based DLC schemes improve DSM and enhance the grid reliability. However,

there are certain disadvantages of these schemes such as privacy and security threats [27]. Also

DLC has less significance regarding small multiple residential loads because of large number of

devices to be controlled and low economic benefit achieved byheavy investment on controlling

mechanisms. Nevertheless, the technique could be relatively more effective for heavy industrial

9

1.4.3. Dynamic Pricing based HEMS

loads. In order to enhance the load management efficiency, need of the hour is to adopt an

autonomous and self-healing system which keeps check on utilities and consumers concerns si-

multaneously through online monitoring and dynamic response. One such mechanism, dynamic

pricing based HEMS, is elaborated in the subsequent section.

1.4.3 Dynamic Pricing based HEMS

This section explains the dynamic pricing based HEMS which are important tools of DSM and

can address the utilities’ concern of peak load management.HEMS automatically respond the

price variations by appliance scheduling in order to minimize the peak load and to optimize the

users total energy cost [26]. Authors in [29], suggest that the amount of energy utilized with

and without dynamic pricing schemes remains same, as purpose of dynamic pricing schemes is

load shifting and not the total load reduction. HEMS providean environment in which smart

household appliances such as air-conditioners, dishwashers, cloth dryers, washing machines,

etc. could talk to the grid and decide how to operate in the best possible way and automatically

schedule their activities at strategic timings based on available capacity [37, 38].

Awareness about real time energy usage with respect to utility point of concern and aware-

ness about RTP from consumers’ point of view are the fundamental blocks to construct and

implement effective DSM through HEMS. Sensors embedded in the appliances provide the ba-

sic information required for dynamic energy management. A home having smart and efficient

appliances with respect to energy consumption which operate on digital instructions is called

smart home [39].

Appliances having embedded sensors for information communication and auto control mecha-

nism are networked with main energy management system and collectively called HAN. Funda-

mental role is given to smart meter which can perform different energy management functions

for a long time without up gradation. HEMS with certain levelof intelligence can be upgraded

and modified independent of smart meters [32]. Smart meters along with data collection points

constitute the NAN. Integration of HAN and NAN on a single platform is given the name of

AMI. Certain security features can also be added to HEMS. This dynamic energy management

system is useful for customers regarding real time energy consumption monitoring and cost re-

duction.

Game-theoretic energy consumption scheduling for future smart grids is described in [40].

10

1.5. Smart Grid, DSM and Related Technologies

Simulation results confirm the effectiveness of game theoryoptimization technique in terms of

reduction in PAR and total energy charges. Comprehensive review of energy consumption in

buildings, DSM benefits and challenges has been carried out in [41, 42].

A comparison of incentive based DLC and dynamic pricing based HEMS with respect to differ-

ent aspects is presented in Table 1.1. It shows the suitability of DLC for heavy loads (industrial)

with greater potential of peak load reduction and greater user discomfort caused by DLC at

residential level. Dynamic pricing based HEMS cause less discomfort to the user and provide a

sophisticated approach to address users privacy concerns.

1.5 Smart Grid, DSM and Related Technologies

Various technologies have been developed and are being improved in order to achieve smart

grid objectives. These technologies encompass different parts of grid. For instance, better asset

management and DG within the stability and reliability limits is possible using smart grid net-

works [9]. On transmission side, Flexible Alternating Current Transmission System (FACTS)

enhances controllability and acts mainly as controllable voltage source [43]. High Voltage

Direct Current (HVDC) uses DC for high voltage transmissionand transfers power among vari-

ous sections of grid operating at different frequencies. SCADA controls various grid operations

in a computerized manner. Wide Area Monitoring Systems (WAMS), based on modern ac-

quisition techniques, monitor the grid at large area as wellas counteract dynamically when

faults/irregularities are detected [44].

As long as power distribution is concerned, there are substation and feeder automation tech-

nologies. The former processes operational and non-operational data being received from and

sent to other entities of the grid, while the later deals withvoltage and reactive power control as

well as improves reliability of the system by detecting faults and isolating faulty parts from the

grid [12].

Smart grid technologies directly related to end-users include smart meters, HEMS, EVs chargers

etc. [45]. Smart meters and HEMS are among the key equipmentsrequired for DSM imple-

mentation. DSM related smart grid technologies are being rapidly developed with the efforts

to modernize power grids in order to cope with the increasingenergy demands in future. High

speed bi-directional communication networks provide the framework for real-time monitoring

and control of transmission, distribution and end-user assets for effective coordination and uti-

11

1.5. Smart Grid, DSM and Related Technologies

Table 1.1: Comparison of Dynamic Pricing based HEMS and Incentive based DLC

Dynamic Pricing Based HEMS Incentive Based DLC

Switch is in users hands Switch is in utilitys hands

Every day phenomenon Mostly used in system emergencies

Lesser level of discomfort for users Higher Level of discomfort for consumers

Mitigation of discomfort by reducing cost of

energy during off-peak hours

Mitigation of discomfort by reducing cost

of energy, incentive payments for load shed-

ding and inconvenience cost payments.

Load control by load shifting i.e. more flex-

ible for users

Load control by load shedding i.e. more flex-

ible for utilities.

Requires active participation and awareness

of consumers

Do not requires active participation and

much awareness of consumers

Considered to be more effective for system

stability during everyday lifeConsidered to be effective during event days

Higher customer satisfaction Lesser customer satisfaction

Utilitys marginal profit varies for different

pricing schemes e.g. RTP and CPP

Utility marginal profit could be larger due to

greater peak load reduction potential

Users privacy is less affected More concerns with respect to users privacy

More security concerns during real time en-

ergy consumption monitoringLess vulnerable with respect to data theft

Relatively less potential of peak load reduc-

tion because of lighter loads.

It has more potential of peak load reduc-

tion when applied for industrial users (heavy

loads)

Involves more complexity because of differ-

ent priorities of different users

Relatively simple as utility priorities are

common for a large group of users.

Needs large no. of devices at residential levelNeeds relatively less no. of devices to con-

trol heavy loads

12

1.6. The Research Problem

lization of available energy resources. Furthermore, integration of computerized automation

at all levels of power system, especially at the distribution and consumer level, enables smart

grids to rapidly self regulate and heal. This in turn improves the entire system’s power quality,

reliability and stability.

DSM essentially depends on effective communication between utility and consumers as well

as on proper use of smart meters and HEMS. The communication network between utility and

consumers is called NAN while the communication among HEMS,smart meter and home ap-

pliances is referred to as HAN. Various long and short range communication protocols and tech-

nologies have been proposed for HANs and NANs, respectively. These include both wire-line

and wireless candidates, however use of wireless communications along with Wireless Sen-

sor Networks (WSNs) in smart grid has promised easily implementable solutions for classical

problems of optimal control and real time scheduling. WSNs also play a key role in appliances

coordination and DR management [8, 46, 47].

There are many stakeholders of smart grid including utilities, regulatory authorities, consumers,

market managers and independent system operators. All the stakeholders should be intercon-

nected through smart grid advanced communication infrastructure as depicted in Fig. 1.3. Tech-

nologies discussed in this are required to be embedded in thesystem for better asset manage-

ment, self-healing and optimized operation. WCTs requiredfor accomplishment of various

smart grid tasks and applications are described in subsequent section.

1.6 The Research Problem

Consumers around the globe need the continuous and reliableenergy supply in a cost effective

manner. Power quality and environmental concerns are important as well. Global consumers’

general concerns may vary according to the regional situations and their requirements. Smart

grid is envisioned to fully address these concerns in a sophisticated and dynamic way. Utilities’

major concerns and consumers’ main issues are highlighted in Figures. 1.4 and 1.5, respectively.

It is indicated in Fig. 1.5 that power quality and environmental effects should be at the top of

the priorities in order to ensure global sustainability. However, priority of the elements depends

heavily on individual consumer or set of consumers and may vary in different geographical

areas.

Peak demand is a major concern of the electric utilities as peak demand puts stress on sys-

13


Figure 1.3: Smart Grid Network Connecting Different Stakeholders

Figure 1.4: Major Concerns of Electric Utility Companies

14


Figure 1.5: Major Concerns of Electricity Consumers

tem stability, widens the supply-demand mismatch and causes adverse economic effects [48].

In smart grid, LM at user premises is one of the most importantissues related to enhancement

of grid efficiency as it results into peak load shaving and reduced probability of grid failure.

Various optimization techniques have been used for peak load shaving and cost minimization

based on the basic objective function given by equation 1.1.Major constraints to this objec-

tive function are the available capacity and fulfillment of total demand as defined in equation

1.2 [49].

T E Cmin=T

∑t=1

(ToU Pt ∗Pgrid,t) (1.3)

subject toT

∑t=1

Pi,t = PD (1.4)

whereT E C is total energy cost;ToU Pt is ToU price andPgrid,t is the power supply from grid,

Pi,t is power demand at timet, i shows the number of appliances andPD is total power demand.

The above mentioned optimization problem has been modified with inclusion of different pa-

rameters such as minimization of PAR, appliance waiting time, ensuring continuous number

of time slots for a particular load, etc. However, still there exists great potential for research

to identify and include more practical constraints to this problem in order to achieve the best

possible solutions.

15


There is an ongoing research in developing efficient HEMS architectures based on various op-

timization techniques. Intelligent cloud based HEMS is developed in [50] which allocates a

dynamic priority to appliances. Priority allocation depends on type and current status of the

appliance. The technique is successfully applied to a test bed and a reduction of 7.3 % in av-

erage power consumption has been achieved. Neural fuzzy logic based controller along with

MATLAB interface has been introduced in [51] that senses theactivities of the occupants inside

the home and facilitates them with required level of services. An Internet Protocol (IP) and

Java based network is presented in [52], which shows an experiment performed on 16 different

appliances in order to achieve effective scheduling.

In [53], a Linear Programming (LP) model has been presented that aims at electricity cost min-

imization by dividing a day in time slots of equal lengths with varying prices of electricity.

However, equal length time slots can be replaced with the varying length time slots according

to schedules of home occupants and dynamic prices. The objective function devised in [11]

minimizes the energy cost by scheduling the house loads in appropriate time slots. Inputs of LP

model are consumers’ requests and the model gives optimum appliance scheduling as output.

However, there still exists some potential for further sub division of load categories in order to

get optimized energy consumption.

A novel scheme, namely in Home Energy Management (iHEM), is presented in [54] for do-

mestic energy management. Unlike the LP model, consumers’ demands are processed in near

real time in the iHEM scheme. Consumers may turn on any appliance without any peak hours

concern; iHEM suggests a convenient start time. This start time cannot be changed without

user manual interruption. Enhanced load categorization and user comfort index may be added

in order to eliminate the user manual interruption.

In the light of above discussion; this research focuses at new enhanced energy management

model with new objective functions and practical constraints to cover the drawbacks mentioned

in different existing schemes. This work is based on furthersubdivision of load categories in

order to increase flexibility with cost minimization i.e. the idea of partial baseline load and

Reserved Interrupting(RI) load which will be discussed in detail in chapters six and seven. This

work also analyses the effects of multiple pricing schemes (ToU, RTP, CPP) in order to provide

more choices to the users and utilities in different situations. Another aspect of this model will

be the solution of energy management problem with various optimization techniques to pro-

vide the researchers a comparative insight regarding energy cost minimization. Furthermore,

16

1.7. Research Contributions

this model also includes the green effects regarding energyconsumption along with theft and

faults detection at domestic level. Inclusion of the above mentioned aspects make the proposed

scheme a novel and comprehensive energy management model. The problem has been imple-

mented in MATLAB using Simulink models.

1.7 Research Contributions

This thesis resulted in the following major contributions:

The work contributes the detailed reviews about: Pakistan’s energy market and major issues of

the energy sector, smart grid applications, DSM methods along with optimization techniques.

In the light of these reviews; the work presents design and simulation of Comprehensive Home

Energy Management Architecture (CHEMA). This is a six layered architecture implemented

in Simulink with embedded MATLAB code. Four out of six proposed layers have been im-

plemented in CHEMA. Second layer is characterized with multiple appliances scheduling and

cost minimization cases with enhancement of load categories. Idea of partial baseline load is

implemented in CHEMA for further energy cost minimization.

The thesis also includes a unique scheduling algorithm withenhanced load categorization i.e.

the inclusion of partial baseline load and RI load in the optimization model. Analyses of multi-

ple pricing schemes (ToU, RTP, CPP) along with multiple optimization techniques (Knapsack,

Particle Swarm Optimization (PSO) and Genetic Algorithm (GA)) has also been included in

this scheme. The novel idea of RI load ensures pre-defined slots for the operation of a specific

load.

1.8 Thesis Organization

The next chapter covers the related work regarding the energy management schemes. Chap-

ter Three encompasses Pakistan’s energy market and energy potential assessment in context of

smart grid and DSM implementation. This chapter covers the major issues of Pakistan’s en-

ergy sector and detailed potential assessment of differentenergy resources. The chapter also

includes future energy predictions according to the country’s current and world’s average per

capita energy consumption. Moreover, the chapter also takes account of energy import options

and indigenous energy potential assessment along with regional energy security situation. A

17

1.8. Thesis Organization

note on smart grid implementation has also been included in the chapter.

Chapter Four includes smart grid applications and technologies. Smart grid applications includ-

ing Advanced Metering Infrastructure (AMI), HEMS, Distributed Automation (DA), inclusion

of DER, Electrical Vehicles (EVs) have been covered in detail. A detailed comparative review

of Wireless Communication Technologies (WCTs) including the short range for Home Area

Networks (HANs) and wide range technologies for Neighbourhood Area Networks (NANs)

has been carried out in order to emphasize the significance ofadvanced communication infras-

tructure for dynamic DSM. The chapter also includes a note onvarious types of dynamic pricing

schemes including ToU, RTP and CPP. Furthermore, smart gridand DSM challenges have also

been briefed in Chapter Four.

The focus of Chapter Five is DSM and optimization based appliances scheduling; This chapter

describes optimization techniques employed for energy management as well as the role of pric-

ing schemes. It contains a detailed discussion on Knapsack,PSO and GA. The chapter includes

basic philosophy and inspiration of the mentioned techniques along with implementation flow

charts. A detailed comparative discussion regarding dynamic pricing based HEMS with the im-

plementation of various pricing schemes and their comparative analysis has also been included

in this chapter.

Chapter Six presents the CHEMA in detail. It covers the basicmotivation of the scheme, de-

tailed optimization problem and implementation of six layered architecture in Simulink. Second

layer of the proposed architecture contains four cases of energy cost minimization. Six layers

include: appliance interface, energy cost optimization, theft and faults detection, calculation of

green effects, load forecasting and dynamic pricing layers.

Chapter Seven takes account of the further enhancement in the energy optimization problem

proposed in chapter Six. It includes energy cost minimization model consisting of four differ-

ent types of loads. The problem has been solved with multiplepricing schemes and multiple

optimization techniques. Knapsack technique has been solved with two different cases. Op-

timization techniques including Knapsack, PSO and GA have been employed with ToU, RTP

day ahead case and CPP. Detailed comparative analysis has been included in the chapter.

Chapter Eight elaborates the conclusions and indicates thepotential future work.

18

Chapter 2Related Work

2.1 DSM and Smart Grid

Dynamic DSM in smart grid environment is supposed to provide the best solution of

energy management and peak load problems [55]. Energy cost minimization and

peak load reduction are the major issues under consideration. Many optimization

based algorithms are available for home appliances scheduling. PSO is one of the techniques

employed for HEMS scheduling [56, 57]. In [57], PSO based HMES has been explained in

context of future smart grid. The scheme employs Binary Particle Swarm Algorithm (BPSA)

for energy management in order to reduce total energy cost considering some constraints on

appliance usage like: limitations on power and personal life style of electricity consumers.

Through mathematical computations, a working time table isproposed for appliances to satisfy

both least tariffs and power limits from user end and power suppliers, respectively. Utility and

end users are inter-connected; utility observes the demandof users regularly and ask to turn off

or delay the appliances when demand increases the peak limit. When demand is low, utilities

ask the end users to use the appliances in order to shift the load. Operating time of the appli-

ances has been controlled and optimal time table is scheduled using BPSA. 100 end users have

been considered, with 11 appliances and having random operating time.

In [58], three GA applications in electrical distribution scenarios i.e. reconfiguring the network

for lost reductions, device placements of optimal safety and prioritizing venture in distributive

domains of network have been discussed. The paper describespreliminary results of prepara-

tory tests performed by using actual scenarios.

19

2.1. DSM and Smart Grid

Tabu search has been applied for the optimization of the household energy consumption which

anticipates the demand from previous profile and events [59]. The anticipation is used to opti-

mize the user comfort and cost. An information fusion based technique has been implemented

in [60] in which Bluetooth provides the infrastructure for appliances coordination. In [61],

Kasttner et.al. described an energy efficient smart yet comfortable home which is designed in

a user friendly manner. It makes the energy scheduling easy by adding learning capabilities

to appliances using artificial intelligence. Therefore, users are free from difficulty of changing

their preferences. A real-time, low-cost, low power and reliable system for monitoring home

appliances using wireless sensors is proposed in [16].

Various pricing strategies are used for users’ energy cost optimization. Authors have used day

ahead pricing scheme for home energy optimization in [53]. The objective of peak demand

optimization can be achieved by home appliances schedulingand/or use of DER. For this pur-

pose, various energy management schemes have been proposed. In [62], a Dynamic Demand

Response Controller (DDRC) to control the HVAC unit in accordance with RTP has been pre-

sented. The controller was designed considering thermal specifications of construction material

and other factors. The controller provides a threshold temperature to switch between heating

and cooling modes. This designed controller was tested on a single family house model which

resulted in peak load reduction and improved grid efficiency. DDRC is limited to the HVAC

only, we have presented a cost minimization model with various categories of loads.

In [63], Appliances Co-ordination (ACORD) scheme aims at ToU pricing benefits and de-

creased energy cost. Objective of ACORD scheme is to shift the consumer load to off-peak

periods. In-home WSNs are used for delivery of consumer requests to Energy Management

Unit (EMU). The work shows that the rate of consumer requestshas a sizeable effect on en-

ergy cost reduction. Energy consumption lowers significantly with an increase in request rates

from consumer side. This scheme only considers the scheduling of home appliances. Appli-

ances Co-ordination with Feed In (ACORD-FI) [64], is another energy management scheme for

energy-aware smart homes. In ACORD-FI both the home appliances and DERs are scheduled

with purpose of energy bill and GHG reduction. ACORD-FI performs scheduling on con-

sumer’s requests with consideration of peak hours, local energy generation as feed in and other

conflicting requests. ACORD-FI uses WSNs for communicationamong appliances, EMU and

smart meters. In [65], the authors have devised a Decision support Tool (DsT) for smart homes.

The DsT based scheme coordinates only the DERs. As a case study, a space heater, a PHEV, a

20


pool pump, a Photo Voltaic (PV) system and a water heater are scheduled based on ToU prices

by applying the Co-Evolutionary Particle Swarm Optimization (CPSO) technique. DsT is com-

posed of DER scheduling algorithm and an energy service model. Another scheme is presented

in [66], which ensures the net consumer benefit by maximized scheduling of the controllable

DERs. Using this scheme, the consumer electricity bill is reduced by 16-25 %.

A scheduling controller has been presented in [67] to schedule the residential electricity genera-

tion and storage keeping in view the dynamic pricing of grid.Mixed integer linear programming

with robust optimization approach was used to tackle stochastic input. The controller decides

on the basis of per unit price whether to sell or buy from the grid. This work is more focused

towards the energy storage and trading with grid. Our proposed model is more focused towards

the user cost minimization with multiple scheduling options.

A hardware demonstration of a proposed HEMS is presented in [68] which consists of a cen-

tral home energy management and controller units. The controller switch off/on the devices in

response of the utility price signals. This work is useful for cost reduction, however no opti-

mization technique has been implemented to get the best solution. In our model, we have used

Knapsack optimization to achieve the best possible scheduling.

A ZigBee and PLC based HEMS has been presented in [69] in whichuse of the renewable en-

ergy has been optimized with the development of consumptionand generation data base. The

work requires the development of data base about generationand consumption in order to get

optimized cost and only useful for the users which have renewable energy generation facility.

In our model, the user is independent of data base requirements.

A multi-objective mixed integer programming model has beendeveloped for smart home appli-

ances scheduling in [70]. The work considers the cost savingand user comfort simultaneously

and tries to create balance between cost and user comfort. This work considers the user ther-

mal comfort, however we have considered the person presencecontroller based lighting control

which makes our work more comprehensive.

An approach for the provision of reserve through DR in order to handle frequency response and

contingency management has been presented in [71]. The focus of the work is case study of 500

users to determine the DR effects at power system level. The work also discusses the behaviour

of the useful appliances for the provision of reserve with consideration of appliances thermal

characteristics and users comfort.

A smart switches based DSM model has been implemented in [72]which is based on priorities

21


assigned to the appliances. Appliances are handled automatically and their operation can be

controlled through mobile phone application via Global Positioning System (GPS). This tech-

nique is especially efficient for high energy consuming loads as simulation results proved a

reduction of 11.2 % in total power consumption.

Authors in [73] have considered a smart home which uses sensor network for HEMS. Smart

home maintains a balance among appliances usage and the human activities. Sensors perform

a number of tasks including occupants monitoring and their activities based future prediction

has been used for energy savings. Simulation results prove afair estimation of future electricity

consumption.

In [74], authors have proposed an efficient ZigBee and Infrared Remote (IR) based HEMS.

Authors consider an easily controllable room based on IRs. The system consists of: (i) auto-

matic standby power cut-off outlets; (ii) a light; and (iii)a ZigBee hub. ZigBee hub has an IR

code learning function and tells the IR remote control signal of a home device connected to the

power outlet. The ZigBee hub with learning function facilitates user to control the power outlet

and the light in order to reduce the wastage of standby power and to manage the home server

that displays the information about the power consumption.Results illustrate that the proposed

architecture causes more power reduction than common HEMS.

In [49], Cheah et al. have proposed ZigBee based HEMS, which has been developed with

LABVIEW software. The proposed HEMS consists of three main components: (i) the ZigBee

smart meter; (ii) Ethernet-ZigBee gateway; and (iii) HEMS SCADA. The ZigBee smart meter

communicates with the Ethernet ZigBee Gateway and other ZigBee enabled devices like Zig-

Bee Load Control Modules (LCM), ZigBee infrared controller, and different ZigBee sensors.

Information regarding real time total energy consumption,billing, customer and status of smart

meter, ToU pricing and maximum demand will be sent to the DataConcentrator Unit (DCU)

which is connected to the smart meter and the HEMS SCADA. ZigBee LCM controls and mon-

itors the home appliances remotely.

In [75], a Model Predictive Control (MPC) framework was proposed which was able to shave

peak curve by optimally shifting load in household scenarioto a large-capacity battery. Markov

process was utilized to predict user’s next location in order to estimate the future energy con-

sumption. An update algorithm was developed to update the probability of user location, energy

consumption and battery storage continuously. This model considers the single user location

for prediction of energy consumption. Our model considers the number of users present with

22


respect to time in order to control the lighting loads and hotwater consumption.

A system architecture consisting three layers has been presented in [26]. First layer is named as

Admission Controller (AC) and it uses the modified spring algorithm for management of appli-

ances’ requests. Second layer is used for optimization of home energy management and dened

as Load Balancer (LB). Mix integer programming is used at this layer to generate the optimized

schedule for appliances. Third layer consists of two modules and used for smart grid interface.

The third layer is not implemented in the scheme. We have proposed new CHEMA which

consists of six layers and multiple energy management options for user. CHEMA is based on

flexible user comfort index for energy management with the introduction of flexibility among

load categories. Unique features of CHEMA include: use of multiple cost minimization op-

tions with single Knapsack technique with inclusion of partial base line load into optimization

problem. Same model includes the theft and faults detectioncapability along with calculations

of green effects at residential level. Model also includes number of users present at home with

respect to time in order to control the partial base line loadand hot water consumption. Inclu-

sion of these features make CHEMA a unique comprehensive model.

In [11], an optimal and automatic residential energy consumption scheduler has been proposed

for a scenario where RTP is combined with IBR. Using the communication infrastructure of the

future smart grids, this work has proposed an optimal load management strategy for RTP. By

adopting this strategy, consumers may bring a balance between their energy bills and financial

situations. Optimum load management strategy consists of three steps i.e. forecasting, activity

scheduling by user and optimization problem. The proposed problem is nonlinear in nature

with a large number of possible combinations hence heuristic optimization techniques like GA

is used for the solution of this problem. This strategy reduces energy bill by 8-22 %.

A strategy to optimize the use of renewable energy sources inlight of weather parameters has

been presented in [76]. The system consists of hybrid energymanagement system including

Monte Carlo and PSO based optimization models. The work is focused in the direction of

DG optimization. In our model the work is focused in cost minimization through appliances

scheduling only.

Load shaping through quota from grid and inclusion of storage capability has been presented in

[77] keeping in view the effects of dynamic pricing. Beyond the quota limit, electricity pricing

is raised for the consumers. However, with the help of storage system, the consumer can get

benefit of selling during peak hours. A general HEMS model based on dynamic pricing signals

23


has been implemented in [78]. The unique feature is to control the user request in order to avoid

congestion on network. Pricing is made higher after a certain number of demand requests. This

work is useful in load shaping, however our model contains unique categorization of Reserved

Interrupting (RI) load, multiple pricing and optimizationtechniques.

In [79], 1 hour is divided into 5 slots of 12 minute and GA is used to obtain optimized solution.

The work is based on RTP with IBR model and using the 12 minutesslot, the average cost is

48.65 cents/day which reduces to 35.97 cents/day using GA. So after 3 months, consumer can

save 1166.6 cents.

Authors have used dynamic programming to minimize the totalenergy consumption cost in

[80]. Mixed integer LP has been used to obtain optimal solution of appliance scheduling with

consideration of renewable energy storage. The work also considers the thermal comfort of the

users.

In [81], load profile of 24 hours is considered using load factor. A battery model is used in

order to provide desired energy during peak hour for users. For a given time, at any instant of

time the energy required by consumer must be fulfilled by combined power supply from grid

as well as from battery. Battery in this case is charged during off load hours, discharge during

peak hours and is idle during middle load hours. Adaptive dynamic programming approach is

used for optimization in which savings of about 1257.9 centsper week is achieved. Where the

energy supplies from grid costs 4124.13 cents and with battery it costs 2866.64 cents during a

week period.

Authors in [82] have assumed ZigBee based smart home with a wireless sensor network. The

smart home have components like temperature and light sensors, heating, air conditioning, ven-

tilation system, smart appliances, thermostat and smart meter. All home components are con-

nected with a central computer system. Smart home components can be interacted remotely

by the web services that are implemented on central computerby using internet. Web services

give the information of energy consumption and the information can be used for energy man-

agement.

In [83], wide area smart grid ZigBee based communication network has been developed and a

new Demand Based Priority Queuing (DBPQ) technique to ensure the Quality of Service (QoS)

requirements of pilot protection traffic has been proposed.A proposed DBPQ technique is used

to provide reliable and high speed sending of the protectionsignals. For testing communication

model feasibility, a test bed has been developed. Other Machine to Machine (M2M) applica-

24


tions that differ in their performance needs can use the proposed technique.

In [84], authors proposed consumption scheduling mechanism using integer LP technique for

DSM in smart grid. The proposed mechanism also schedules both optimal power and opera-

tional time for power and time shift applications. Simulations for both the centralized and dis-

tributed cases have been performed and results show the effectiveness of the proposed model.

In [85], authors propose a model to minimize peak load for home demand side load manage-

ment in smart using integer LP. Authors applied LP for both the HANs and NANs applications

to demonstrate the proposed method suitability. In [86], energy optimization algorithms based

on mixed integer LP have been proposed in order to reduce costand save energy. For these

algorithms, end users or residents should be updated about billing cost so that they can choose

better option for them to reduce cost. Authors have discussed model for a time horizon and per-

formed simulations. In addition they have also discussed impacts on end users, sustainability,

smart grid and dynamic pricing policies. The proposed modelhelps users to reduce cost and

save energy.

In [87], LP technique is used for cost minimization. Authorshave assumed that price profile is

designed by utility to flatten global demand. To minimize thebilling cost, authors have designed

optimal energy flow between smart grid and Energy ManagementUnit (EMU) by solving model

through LP. This technique minimizes the daily billing costeffectively.

In [88], LP is used to optimize the energy requirements of Greece by using renewable energy

resources in different administrative and geographical regions. Authors have proposed a model

to cover the mismatch between the energy supply levels and the end consumptions by using

LP and the exergo-economic analysis for the Hellenic energysystem. They also define the re-

maining available space for energy recovery units from municipal solid waste in each region to

participate in the energy system. The proposed model helps to optimize renewable energy for

the Hellenic energy system and the use of m municipal solid waste in energy system.

In [26], Costanzo formulated an energy management problem by dividing the load into three

categories: base line load, regular load and burst load. Base line load is kept out of scheduling

optimization problem because it must be available at any time. Regular loads are operated ac-

cording to thermostat or temperature limits e. g. refrigerator, room air conditioner etc. Burst

loads may come on line at any time and need certain time slots for operation completion. There

are two types of burst loads which are preemptive and non-preemptive. The problem was solved

using binary integer programming. This approach has given auseful solution for house appli-

25


ances scheduling and cost minimization, however in order tocreate some flexibility and more

cost minimization, we have divided the load into four categories in which base line load is

partially included into optimization problem. In addition, a new category of RI loads has been

introduced which brings the flexibility in the users’ schedule. This category has been introduced

for those users who need a certain load in specific hours. These hours may vary user to user

based on their requirements. The mechanism can be implemented by specifying suitable time

slots using the proposed scheduling methodology.

In relation to previous chapter in which we implemented ToU pricing scheme at second layer,

here we are introducing multiple pricing schemes like: ToU,RTP day ahead case, and CPP.

Authors in [57] used PSO with ToU pricing scheme for home appliances scheduling and cost

minimization. However, we have used three different types of optimization techniques for solu-

tion of the optimization problem. i.e. Knapsack, PSO and GA,which provides implementation

flexibility as well as a comparison basis.

On the bases of above discussion, following problem has beenformulated.

In the proposed Costanzo architecture, there are total three layers, however there are certain

aspects regarding HEMS which should be added to make the architecture more comprehensive.

i.e. inclusion of green effects at residential level, detection of theft and faults at residential level

and incorporation of multiple pricing schemes in order to use it at different occasion. Inclusion

of partial baseline load into optimization problem for effective energy cost minimization.

Generally, three categories of appliances have been used inliterature: baseline load, regular

load and burst load. Base line load consists of essential lighting e.g. fans, lighting etc. This

type of load can be included into schedulable load on partialor half bases.

The definition of thermal comfort zone instead of sharp threshold to switch on/off the thermal

load is also helpful in cost minimization. Suitable comfortzone definition and its implementa-

tion is also necessary for comprehensive energy managementarchitecture.

Total cost minimization for single user have been done extensively in literature, however the

effect of the users present at different times effects the load consumption. The effect of the

number of the persons present for lighting loads and hot water consumption cause effective re-

duction in total cost.

On the bases of above discussion, we are proposing the six layered architecture which will be

discussed in Chpater Six. and at its second layer optimization problem of energy cost minimiza-

tion has been solved with multiple options. Inclusion of RI load in the energy cost minimization

26


problem for ensured operation of a particular load in the predefined slots generates another novel

HEMS scheme which will be discussed in Chapter Seven.

Pakistan’s energy market in context of smart grid and DSM implementation will follow this

discussion in next chapter.

27

Chapter 3Energy and Electricity Situation in

Pakistan as a Pretext for Smart Grid

Implementation

3.1 Summary of the Chapter

Pakistan is facing severe energy crisis in spite of the fact that nature has blessed her

with huge energy potential. Shortfall of electricity supply in the country is increasing

and has been recorded up to 4522 MW in 2010. This deficit reached to 7000 MW

in May, 2011. A comprehensive review of Pakistan’s energy sector in context of smart grid

implementation has been presented in this chapter. Energy potential, major issues of energy

sector and energy import options are discussed. Issues suchas poor management, combined

cycle capacity, low hydro power share, circular debt and energy security have been covered.

Energy potential assessment includes hydro, solar, wind, coal, nuclear, hydrogen cells, geo-

thermal, ocean resources and bio-mass. Future prediction calculations are based on country’s

current and world’s average per capita electricity consumption. Current oil and gas reserves of

the country contribute to only 5% and 48.8% of the energy mix and at the current rate will be

exhausted by 13 and 16 years, respectively. The overwhelming dependence of the energy sector

on imported fossil fuels may create a situation of energy security threat. However, dependence

upon the energy import options cannot be avoided in order to lessen the severity of energy cri-

sis in near future. Some of the energy import options are: Turkmanistan, Afghanistan, Pakistan

28

3.2. Current Energy Scenario of Pakistan

and India (TAPI); Iran, Pakistan and India (IPI) gas pipelines; Liqueed Natural Gas (LNG) from

Qatar etc. On the other hand, exploitation of vast renewablepotential such as hydro, solar and

wind requires serious attention. Exploitation of indigenous coal resources would also be a key

for solving energy crisis in the long run. Smart grid implementation and conversion of monopo-

listic market into competitive market for effective DSM have also been included in this chapter.

In summary, this chapter presents energy potential assessment in context of major issues, future

predictions and impact of energy import options. This in turn provides a big, clear and brighter

picture of the country’s energy sector keeping in view the smart grid and DSM options.

The energy cost minimization and peak load reduction solutions produced by this research work

show promising results and are general in nature. To apply such solutions in order to transform

the existing grid of our country, it is imperative to first understand the existing grid and its prob-

lems in detail. This chapter analyses the Pakistan’s energysector and electricity market in detail

in context of smart grid implementation.

3.2 Current Energy Scenario of Pakistan

Electricity requirement in Pakistan is increasing due to the rapid growth in population. The

available resources are not sufficient to meet the increasing demand which causes the acute

shortage. The inability of the energy supplies to cope with the demand of growing population

has a significant impact on the economic growth of the country. According to the state bank

of Pakistan, the electricity shortage alone resulted the loss of approximately 210 Billion Pak.

Rupees (PKR) i.e. 2% of GDP and unemployed 400,000 industrial workers [89]. The gap be-

tween demand and supply became an unsolved mystery for last two decades. In 2014, the total

install capacity in Pakistan was 21375 MW and the maximum generation during that period was

16302 MW [90]. There is approximate 5000 MW short fall which causes the load shedding of

8 to 10 hours in cities and 14 to 16 hours in urban areas.

Generally, there are two types of electricity markets i.e. monopolistic and competitive mar-

kets. In monopolistic approach single or a few companies provide services and have the power

to change the supply, price and quality. On the other hand thecompetitive markets have a

large number of producers which compete to facilitate the consumers. In Pakistan, the gen-

eration, transmission, distribution of the electricity isthe responsibility of two vertically inte-

29

3.2. Current Energy Scenario of Pakistan

grated public sector companies i.e. Water and Power Development Authority (WAPDA) and the

Karachi Electric Supply Corporation (KESC). WAPDA is responsible for providing services to

the whole country except Karachi which is served by KESC. Themonopolistic approach of the

two companies causes the inefficiency of electricity marketas a result prices and gap between

demand and supply is increasing day by day. Detailed analyses of Pakistan’s energy sector is

presented in subsequent sections in context of smart grid implementation.

The homeland Pakistan is a developing country of approximately 181.3 million people [91] and

is facing acute energy shortage. In 2010, it had to import 21.64 MTOE of energy in order to

fulfill its primary energy requirement of 63.09 MTOE [92]. Short fall of electricity supply in

the country is increasing with demand and has been recorded up to 4522 MW in 2010 for many

times of year. This deficit reached to 7000 MW in May, 2011 [93]. The Gross Domestic Prod-

uct (GDP) growth has fallen sharply from 3.8 % in 2010 to 2.4 % in 2011 whereas the inflation

rate in the country has risen from 10.1 % to 13.7 % during the same period [94]. Furthermore,

limited natural gas supplies to domestic and industrial users, has intensified the severity of the

energy crisis. In 2010, the indigenous oil and gas production was 64,948 Barrels/day and 4,063

Mcft/day respectively. According to Hydro Carbon Development Institute of Pakistan (HDIP),

the balance reserves of oil and gas by ending June 2010 were 41.13 MTOE and 498.70 MTOE

respectively. If the demand and supply remains unchanged, it can be calculated that the known

oil and gas reserves will exhaust in nearly 13 and 16 years, respectively as follows:

Oil production in2010= O= 3.180MTOE (3.1)

Gas production in2010= G= 30.812MTOE (3.2)

Balance Oil Reserves(30th June2010) = BO= 41.13MTOE (3.3)

Balance Gas Reserves(30th June2010) = BG= 498.7 MTOE (3.4)

Years f or exhaustion o f existing oil reserves=YO= BO/O= 13Years (3.5)

Years f or exhaustion o f existing gas reserves=YG= BG/G= 16Years (3.6)

This shows that under current demand and supply scenario theoil and gas reserves will exhaust

by 2020 and 2026, respectively. This clearly indicates thatPakistan desperately requires the

quest of other energy options for solution of its energy crisis on war-footing basis. Smarter

energy management through dynamic DSM in context of energy crisis and limited reserves is

also imperative. Major issues of Pakistan’s energy sector are described in detail in subsequent

section.

30

3.3. Major Issues of the Energy Sector

3.3 Major Issues of the Energy Sector

In 2010, Pakistan’s energy mix comprised of nearly 88.2% fossil fuels whereas hydro and nu-

clear constituted to 10.6% and 1.1% respectively. Similarly, the electricity generation is pri-

marily dependent on thermal power plants which contributedto nearly 67.3% of the electricity

generated. Hydro and nuclear power constituted to 29.4% and3.03% of the electricity gen-

erated while 0.26% was imported from neighboring countries. Contribution of hydro power

in production of electricity reduces to only 11% in winter due to seasonal variations in water

flow, forced and maintenance outages, etc. [95]. The major issues being faced by the country’s

energy sector are shown in Fig. 3.1.

Figure 3.1: Major Issues of Pakistan’s Energy Market

3.3.1 Decreased Hydel Share in the Country’s Energy Mix

In 1960, hydro power contributed to 64.9% of the electricitygenerated in the country but by

2010 its share reduced to only 29.4%. The growth of nuclear generation has been more or

less stagnant and only 787 MW could be installed up till now. The obvious option left to

meet growing demand was through expensive thermal generation. This shift of the principal

generation from hydro to thermal increased the cost of generation due to the hike in fossil fuel

prices. It also worsened the reliability of the electrical grid due to availability problem of fossil

31

3.3.2. De-Rated Capacity

fuels and initiated the issue of circular debt. This in turn created massive inflation and great

unrest amongst the general public. Pakistan has an agriculture based economy. Dams not only

supply water for irrigation but also provide protection against floods. Norway, Brazil, Iceland,

Canada and Austria produce 99%, 92%, 83%, 70% and 67% respectively of their electricity

through hydro power [96]. China has some 22,000 dams with storage capacity of 2280 Million

Acre Feet (MAF) approximately. USA has around 5,500 dams while India has 650 undertaken

dam projects with total of over 4000 dams [97, 98]. In contrast, Pakistans storage capacity is

only 8 % of its annual surface flow share of 142 MAF [99]. The last sizeable construction

was the 1450 MW Ghazi Brotha run-of-the-river project. The construction of Kalabagh dam is

unlikely to start due to severe controversy between political groups whereas the construction of

the proposed 4500 MW Diamer Basha dam has been delayed for more than ten years. Hydro

power is the most established form of renewable generation.Dams engage a huge capital but

on the other hand they also generate the lowest per unit cost of electricity.

3.3.2 De-Rated Capacity

The total installed capacity up till June, 2011 was 24,173 MWbut still extensive load shedding

was experienced throughout the country [100]. Whereas the company wise share of electricity

generation in the country was: Pakistan Electric Power Company (PEPCO), 50%; Independent

Power Producers (IPPs), 38.6%; Karachi Electric Supply Company (KESC), 8.4% and Pakistan

Atomic Energy Commission (PAEC), 3% [101]. Up till June 2011, the installed capacity of the

largest company in Pakistan i.e. WAPDA/PEPCO was 20,986 MW whereas the dependable ca-

pacity had de-rated to 18,734 MW. But maximum system capability averaged from the past four

years is 15,190 in summer and 13,815 MW in winters [102]. Thisis due to various reasons like

the problem of silt, decreased efficiencies of thermal plants, periodic lows in the flow of water,

problem of circular debt, fuel availability, forced and maintenance outages, auxiliary consump-

tion and transmission limitations etc. Electricity generation starts to de-rate immediately after

installation and requires scheduled maintenance. Thermalgeneration in the public sector was

not allocated suitable funds; it caused rapid decrease in generation capability.

32

3.3.3. Circular Debt

3.3.3 Circular Debt

Failure of PEPCO, central power purchasing company, in maintaining cash flows to its suppliers

is termed as circular debt of energy sector. Circular debt has decreased reliability of national

grid and is one of the major causes of load-shedding. Currentcircular debt of power sector has

reached 313 billion rupees [103]. Issue of circular debt arose due to imbalance in generation

cost and revenue collected by PEPCO. The reason for poor revenue collection and the pile

up of circular debt are shown in Fig. 3.2 and detailed below: First reason is promotion of

thermal production in the energy mix. Bulk share of thermal power projects in the energy mix

is responsible for high cost of electricity. In 1960s, contribution of the hydro power was almost

two third of the total supplies. But the role of international and national lobbies to guard their

vested interests in the promotion of thermal has contributed to the current high electricity prices,

inflation and decreased reliability of national grid. IPPs contribute 38.6 % of the generation and

require timely payments every month to meet their expenses,but the escalation in fuel prices

and delay in payments by huge Government institutions resulted in the situation of circular debt.

Second reason behind circular debt is poor revenue collection. Insufficient tariff, pilferage of

electricity, corruption, losses, dispute on tariff with FATA, AJK and KESC all contributed to the

cause. Delay in payments by federal and provincial institutions further elevated the situation.

Third reason is withdrawal of subsidies by Government of Pakistan under the pressure of IMF.

This resulted in increased tariff as PEPCO was already unable to meet its expenses due to

losses incurred. Per capita income of the country is USD 1258only [104]. Cost of electricity

was already high due to increased share of thermal generation; this in turn promoted the illegal

use of electricity resulting in further decrease in revenue. PEPCO contributes to almost 50 %

of the power generation in the country, but is itself dependent on the basic energy suppliers for

its operation. Basic energy suppliers include oil marketing companies and gas suppliers. In

June, 2006 PEPCO availed bank loans against government guarantees to overcome the effect of

withdrawal of subsidies. The situation was further exaggerated by the huge line losses and poor

revenue collection from Government offices and as a result PEPCO started delaying payments

to the basic energy suppliers and the IPPs in August, 2006 [103]. In this way the monetary

problem of PEPCO cascaded down to the basic energy suppliersand they were stressed to

engage banking sector to balance their cash inflows. The mechanism of the circular debt is

shown in Fig. 3.3.

33

3.3.3. Circular Debt

PEPCO Tariff Conflicts

Line Losses

Pilferage of Electricity

Delayed payment by Govt.

Institutions

Billing Irregularities

Corruption of Employees

Poor Revenue

Collection

Figure 3.2: Reasons of Poor Revenue Collection

34

3.3.4. Management Issues

Government Subsidies

Revenue Collection

PEPCOBASIC

ENERGY

SUPPLIERS

(a) Normal Cash Flow

Withdrawal of Subsidies

Poor Revenue Collection

PEPCODelayed

Payments

BASIC ENERGY

SUPPLIERS &

IPPs

Banking

Sector

To B

alance the

Cash

Flow

IPPs

(b) Mechanism of Circular Debt

Figure 3.3: Mechanism of Circular Debt

3.3.4 Management Issues

The policy frame work and its management are under strong political influence. Circular debt

is one of the many issues generated from poor management. Introduction of rental power plants

portrays the political influence on policy frame work of power sector. Ignoring merit, appoint-

ments of non-eligible employees on political basis etc. aresome examples of management

concern. Non technical losses or electricity theft is another aspect of poor management. A de-

tailed review of electricity theft identification and control has been presented in [105]. In 2011,

the technical losses of PEPCO were 20.85 % with 90,575 GWh generated and 18,877 GWh lost

at an average unit sale price of Pakistani Rs. 6.25 [100]. Since we know that Percentage Losses

( PL ) can be calculated by:

PL= (Energy Lost/Energy Generated)∗100 (3.7)

35

3.4. Future Predictions

Table 3.1: Comparison of Per Capita Electricity Consumption for Some Countries

CountryPer Capita Electricity

Consumption in kWh

USA 13,361

Australia 10,063

France 7,756

Germany 7,217

Saudi Arabia 7,967

UK 5,741

China 2,942

Turkey 2,474

India 644

Pakistan 457

Highest Per Capita:

Iceland51,447

World Average 2,892

A reduction of 1 % of this loss would result in Savings (SV) of approximately USD 62.9 million

and can be calculated as follows.

SV= 1%∗Energy Generated∗100 (3.8)

SV= 1%∗90575∗6.25= PKR6.038Billion (3.9)

SV=USD62.9 Million (1USD= 96PKR) (3.10)

3.4 Future Predictions

Pakistan’s per capita electricity consumption is 457 KWh which is around one sixth of world’s

average of 2892 KWh. The comparison of per capita electricity consumption with some other

countries is presented in Table 3.1. The population and therefore demand for energy is increas-

ing continuously. Current shortfall of electricity fluctuates between 4 to 7 GW and is expected

36

3.4. Future Predictions

to widen up to 14 GW by 2020 [106]. The current population of Pakistan is 181.3 million and is

growing at a rate of 2.69% per annum. At this rate the population and electricity requirements

by 2025 and 2050 can be estimated.

Pakistans projected population in year 2025 (PP25) can be estimated as follows:

PP25= 181.3∗ (1+0.0269)x (3.11)

PP25= 256Million , (x= 2025−2012) (3.12)

Similarly, the projected population in 2050 (PP50) can be estimated as follows:

PP50= 181.3∗ (1+0.0269)x (3.13)

PP25= 497.1 Million ,(x= 2050−2012) (3.14)

The per capita requirement is not same as per capita consumption due to a short fall of 4522

MW as recorded in 2010. The electricity requirement after including the load shedding factor

has been computed as 78.8 TWh approximately (consumption side). The per capita electricity

requirement has been computed as 640 kWh by NTDC. The per capita requirement is bound to

increase owing to improvement in lifestyle and infusion of electronic/digital devices in every

household. But even if the per capita requirement remains stagnant; the minimum electricity

requirement of the population in 2025 and 2050 can be estimated. Pakistans electricity require-

ments in 2025 (PER25) can be calculated as follows:

PER25= Percapita∗Population(2025) (3.15)

PER25= 640∗256= 163.8 TWh (3.16)

Similarly Pakistans electricity requirements in 2050 (PER50) can be calculated as follows:

PER50= Percapita∗Population(2050) (3.17)

PER50= 640∗497.1= 318.1 TWh (3.18)

These projected population and electricity requirements have been simulated in MATLAB to get

a clear picture and are shown in Fig. 3.4. The results clearlyshow that the countrys electricity

requirements are predicted to be more than twice current demand by 2025 and four times current

demand by 2050. But the progress in development of energy resources in the past does not

37

3.5. Energy Security and Import Options

Figure 3.4: Projected Population and Electricity

show a promising picture for future needs. Only 787 MW of nuclear capacity could be installed

between 1980 and 2011. The addition of only 7 GW of hydel capacity in the last 55 years

has raised serious doubts on the achievement of future targets [106]. To meet the tremendous

energy requirements, the country will have to explore the indigenous energy resources along

with import options which are elaborated in subsequent sections.

3.5 Energy Security and Import Options

The dependence of Pakistans energy sector on imported oil has created serious doubts on the

energy security of the country because international routes for shipment of the oil for Pakistan

may be battlefield of some international conflicts. Furthermore, in case of war the blockage

of the sea ports may halt the entire thermal electricity generation. The indigenous oil and gas

reserves will be depleted in 13 and 16 years respectively. Infuture, the wealthy nations will

outbid other countries in the run for energy import. Pakistan is in negotiation with its neigh-

boring countries to solve its energy crisis. The project of Iran gas pipeline, import of 10,000

MW electricity from Iran and import of LNG and LPG from Qatar is underway. Similarly, im-

port of 10,000 MW electricity from Kirghizstan and Tajikistan (CASA-1000), 1000 MW from

Uzbekistan is in the feasibility phase. India has also offered 10,000 MW of electricity export

38


to which Pakistan is seriously considering. But the political instability in Afghanistan and pre-

vailing gruesome situation of security in Sindh and Baluchistan has rendered the planned routes

vulnerable. The regional and international strategic political concerns have made the success of

these projects suspicious. The LNG import option from Qatarand comparison of TAPI and IPI

gas import projects are discussed in the following section.

Pakistan is generating 32 % of its electricity from natural gas and has 27.5 Trillion cubic feet

balance of recoverable gas reserves. As discussed in section 3, the current gas production is

around 4 Billion Cubic Feet per Day (BCFD) whereas the demandis approximately 6 BCFD.

The gas production is expected to fall to less than 01 BCFD by 2025 due to depletion of gas

reserves and demand will increase to 8 BCFD. Almost one thirdof the indigenous natural gas

is used for electricity generation (32 %) which causes a severe domestic and industrial gas load

shedding [106]. This has significantly damaged countrys export earnings and increased the

import bill. In view of the prevailing energy crises and exponentially decreasing gas reserves,

importing gas from neighboring gas rich countries has become the need of the hour.

The current prevailing geo-political scenario in the region has put a question mark on the launch

and completion of IPI gas pipeline proposed since 1993. As analternate to IPI, TAPI gas pipe-

line project is required to be materialized at a rapid pace inorder to meet the energy require-

ments of Pakistan in particular and the region in general. Itis worth mentioning here that India

has already withdrawn from the IPI project while Bangladeshhas shown its interest in the TAPI

project. Salient features of IPI and TAPI are summarized in Table 3.2. The proposed IPI project

would provide only 01 BCFD of gas at a cost of USD 1.25 billion.The proposed TAPI project

would provide 3.2 BCFD to 3 countries at a cost of USD 7.6 billion. If both IPI and TAPI

projects are completed, Pakistan will have secured around 3BCFD as compared to a demand

of 8 BCFD by 2025. Even then the demand supply gap will be around 5 BCFD because the

existing indigenous gas reserves of the country shall have exhausted by 2026 as estimated in

section 3.3. In the light of above facts: it will not be possible to feed gas based power plants in

future, contribute 32% of the power generation, if dependedonly on the above mentioned gas

pipe lines [104]. Both IPI and TAPI gas pipe line projects require security guarantees as well as

peace in the region.

While transportation of gas using long pipelines proves to be highly challenging and time con-

suming projects, LNG can prove to be an alternative means of long-distance transportation.

LNG is stored and transported in liquid form in tanker ships.After delivery to the market, it

39


Table 3.2: Comparison of TAPI and IPI Gas Projects

Detail IPI TAPI

Route Iran-Pakistan-IndiaTurkmenistan-Afghanistan-

Pakistan-India

Power Generation4,000 MW power generation

in Pakistan2500 MW

Gas (mmcfd) 60 to Pakistan and 90 to India38 to Pakistan and India each,

14 mmcfd to Afghanistan

Transporting CostLow processing and trans-

portation cost to Pakistan

New silk route between Cen-

tral and South Asia

Transit Fee 200 Million USD to Pakistan 217 Million USD to Pakistan

Economic EffectEconomic boost in pipe line

route

Economic boost in pipe line

route

Length 2,775 Km 1,680 km

Route

South Paras Field Iran -

Baluchistan - Multan (Pak-

istan) - India

Tolon Field Turkmenistan -

Afghanistan - Pakistan India

Total Cost USD 7.5 Billion USD 6 to 7 Billion

International SupportReservations of USA and EU

due to involvement of Iran

Supported by IMF, World

Bank and USA

is re-gasified and distributed via pipelines. In view of prevailing geopolitical and logistically

challenging circumstances in the region burdened with tensions, construction of risky and time

consuming pipeline projects would be less feasible. Takingthe complicated political dynamics

of the region into account, expanding the LNG trade and port capacity is preferable to devel-

oping potentially volatile pipelines. Prospects of bringing in LNG from Qatar, having the third

largest gas reservoirs and being the largest LNG exporter, are needed to be thoroughly analyzed.

Similarly agreements can be made to utilize the expertise from Australia, a major player in LNG

market. It is pertinent to declare here that our neighboringcountry is getting considerable en-

ergy benefits from Qatar and Australia. For development of LNG resources, Indian skills and

expertise may also be taken into account. LNG Plants are modular, compact and efficient. In

40

3.6. Energy Potential

addition to Qatar, India has also offered to export 200 Btu ofLNG to Pakistan. The price pro-

posed by India is USD 21 MMBTU including all the charges. Thisprice is higher than the USD

18 quoted by Qatar and much higher than the USD 11 proposed by Dutch firm 4Gas Mashal

LNG.

The IPI and TAPI Gas pipe line projects are highly dependent on the position of security and

peace within Afghanistan. Although Afghanistan has demanded only the transit fee from Pak-

istan and India and has left her gas share from TAPI gas pipe line project but in future her

requirements may need the establishment of gas infrastructures. It can be easily concluded

that the LNG import options must be prioritized to the gas pipeline projects in order to ensure

the availability of energy. Indigenous energy potential and its analysis has been presented in

subsequent sections.

3.6 Energy Potential

Pakistan has a huge potential of a variety of energy resources. The identification and quantifi-

cation of this potential is necessary to make the country self-dependent. The detailed study to

examine these resources is presented in following sub sections.

3.6.1 Hydro Power Potential

Hydra or water is one of the most significant components for human life. Per capita water

availability is a measure of relationship between water andhuman life. More than 1700m3

availability shows that country has no water scarcity whilebelow it shows stressed or seasonal

scarcity prevailing in any country. Lower than 1000m3 means severe shortage of water in a

nation; Pakistan is at the threshold of 1038m3. This is the result of delaying the development

of water reservoirs.

Dams are among world’s largest man-made structures and require huge capital and an extended

time period for completion. They not only help in generationof electricity by providing head but

also provide water for irrigation and protection against floods. Storage capacity of dams reduces

with time as they are filled with silt; but even when full of sedimentation, they can be operated

as run-of-the-river plants. Hydro power is the most established of all renewable technologies

and has a share of nearly 15.3 % in the global energy mix with total installed capacity up to

41

3.6.2. Solar Energy

Figure 3.5: Proposed Sites for Hydro Power Generation

970 GW at the end of 2011 [107]. At present, the countrys storage capacity is only 8% of its

annual surface flow share of 145 MAF. Terbela is the largest storage reservoir in the country

with storage capacity of 9.69 MAF which is now reduced to 70% because of silt accumulation.

Mangla has 5.05 MAF of storage capacity after upraising. Government has withdrawn from the

controversial Kalabagh Dam project and the completion of 4500 MW Diamer Basha Dam is

expected till 2022. The identified potential of large hydro power plants is 59 GW whereas the

potential of small hydro projects in Gilgit is around 2000 MW[108]. The proposed sites for

new hydro power plants are shown in the Fig. 3.5 [109].

3.6.2 Solar Energy

Solar PV and solar thermal technologies are now establishedtechnologies and huge projects

of bulk generation are underway in numerous countries. Solar generation is the most rapidly

growing of all the renewable resources. In 2011, the global solar PV cumulative capacity in-

creased by 74 % to almost 70 GW. The size of global solar PV industry is now over USD 100

Billion/year. More than 450 MW projects were added to increase the global concentrated solar

thermal power capacity by 35 % and enhance the total to 1,760 MW (mainly parabolic trough).

Similarly, solar thermal heating and cooling increased by 27 % to augment the global total

42

3.6.3. Wind Energy

Table 3.3: Large Isolated Power Plants Operating on Solar Technologies

Name Location Capacity Technology Type

Aqua Caliente Solar

ProjectArizona-USA 247 MWp Photovoltaic (PV)

Solar Energy Generat-

ing Systems

Mojave Desert,

California-USA354 MW Parabolic Trough

Martin Next Generation

Solar Energy Centre

Indiana Town,

Florida-USA75 MW

Steam Input into a

Combined Cycle

Puerto Errado Murica-Spain 31.4 MW Fresnel Reflector

Maricopa SolarPeoria, Arizona-

USA1.5 MW Dish Stirling

PS20 Solar Power

TowerSeville-Spain 20 MW Solar Power Tower

consumption to 232 GW. The solar thermal technology is more efficient and has the benefit of

comparatively easy storage and hybridization with other energy sources over solar PV. Stand

alone power plants of large capacities operating in the world based on solar technologies are

detailed in Table 3.3. According to world solar map, shown inFig. 3.6, prepared by NASA,

Pakistan lies in the region where the insolation is around 1900-2200 KWh per square meter,

which is the second highest region in the world [109]. The estimated potential of solar energy

in the country is approximately 2900 GWh. The 178 MW solar PV plants at Pakistan Engi-

neering Council and Planning Commission are the only on-grid solar pilot plants in the country.

Lack of infrastructure and non-existence of photovoltaic industry in the country are barriers to

solar energy development. The high per unit cost of solar energy is also a restriction to adopt

solar energy at lower level.

3.6.3 Wind Energy

According to the GWEC (Global Wind Energy Council) wind energy is now generated in over

75 countries with 21 countries having installed capacity above 1000 MW. In 2011, over 40

GW of capacity was installed increasing the world total to 238 GW approximately. A list of

countries with top ten installed wind power capacities is presented in Table 2.4 for comparison

43

3.6.4. Bio Mass

Figure 3.6: World Solar Insolation Map

[46]. Pakistan has a coastal line of around 1000 Km. Speed of wind reaches up to the 7-8

m/s in this region. Wind map of Pakistan is shown in Fig. 2.7 [109]. This speed is one of the

most suitable and exploitable wind speeds. Identified potential of Gharo, a coastal region, is

approximately 55 MW. Government has issued licenses to somecompanies. A 50 MW project

in Sindh has been instigated by a Turkish company. Government has also signed MoUs with

China for production of 2000 MW electricity generation. Government has a target to install 30

GW wind power by 2030 in the country.

3.6.4 Bio Mass

Bio mass provides almost 10% of the global primary energy supplies and is the 4th largest source

of energy. It can be in solid, liquid or gaseous form. It consists of the residues of different crops,

wood, animal waste, municipal solid waste, processed industrial waste, bio fuels and land fill

gases etc. In 2011, electricity generation from biomass increased by 9% and the total installed

global capacity is 72 GW. Pakistan has an agrarian economy and biomass has around 36%

share in the basic energy mix of the country. Unluckily a hugeamount of crops residue is being

burned aimlessly instead of generating energy. However, recently sugar mills announced to

produce 3000 MW from bagasse. Pakistan has a potential of 400,000 tons of bio diesel which

is environmental friendly and can be produced from non-edible oils. Castor bean is one of the

44

3.6.4. Bio Mass

Table 3.4: List of Countries with Top Ten Wind Power Capacities

Country Installed Capacities (MW)

China 62,634

USA 46,919

Germany 29,060

Spain 21,674

India 16,084

France 6,800

Italy 6,737

UK 6,540

Canada 5,265

Portugal 4,083

World Total 237.7 GW

Figure 3.7: Wind Map of Pakistan

45

3.6.5. Nuclear Energy

Table 3.5: Comparison of Nuclear Power Generation for Some Countries

Country Electricity Generation (%)

France 77

Slovakia 54

Belgium 54

Ukraine 47.2

Hungary 43.2

Slovenia 41.7

Switzerland 40.8

Sweden 39.6

South Korea 34.6

Armenia 33.2

Pakistan 7.6

heavily oil enriched seed which is self growing and found in arid areas of Pakistan. Castor oil;

extracted from castor bean; is used in many medicines and is an important export for Pakistan.

Its fascinating use is as a bio fuel. Processing of castor oilto convert into bio diesel is very

easy compared to other methods of producing bio diesel. Dissolving castor oil into alcohol will

convert it into bio diesel. Exploitation of biomass can helpto reduce the energy crisis in a great

deal.

3.6.5 Nuclear Energy

Nuclear is one of the most significant and controversial industries in the world. It plays its role

from the treatment of cancer, agriculture, strategic assetand concludes at provision of cheap

nuclear power. France is generating 77% of its electricity from nuclear power plants whereas

Pakistan is generating only 7.6% of its electricity from nuclear power. A list of some countries

producing nuclear power is presented in Table 3.5 for comparison purposes. Pakistan was the

first in South Asia in 1970s to enter in the era of nuclear powerby launching the Karachi

Nuclear Power Plant. The 137 MW was nearly 20 % of the requirement of Karachi at that time.

KANUPP is producing 80 MW now a day due to aging effects. KANUPP is based on the natural

46

3.6.6. Geo-Thermal Energy

uranium and heavy water technology known as the CANDU (Canada Uranium Deuterium).

Advancement in the nuclear power program is made on the basisof Pressurized Water Reactors

(PWRs). These reactors are based on natural water and enriched uranium. Two reactors C-1

and C-2, each of 350 MW, are operational at Chashma while two reactors (C-3 and C-4) of the

same rating are under construction. Pakistan Atomic EnergyCommission is trying to get 1000

MW nuclear power plants to install at Karachi in order to meetits target of generating 8800

MW by 2030.

3.6.6 Geo-Thermal Energy

According to some estimates, heat energy present in upper 10kilometers of the earth crust is

50,000 times the total oil and gas reserves of world [110]. Itvaries in ascending order from

north to south. There are two types of the systems used for thegeo thermal energy.

• Hydro-geothermal system

• Enhanced geothermal system

Hydro-geothermal energy systems deal with energy within 6 kilometers of the earth crust. Un-

derground water is converted into steam and then trapped in the earth rocks. Natural hot springs

and geysers are renowned hydro geothermal systems. Enhanced geothermal energy deals with

earth crust up to depth of 10 kilo meters. Geo-thermal energycan be used directly in the form

of heat or converted to electricity. Around 78 countries utilize direct geo-thermal heat energy

systems using ground source heat pumps which can provide heating and cooling. Geo-thermal

heating provides an estimated 58 GW and Geo-thermal power adds an estimated 11.2 GW to

meet global energy demand. Geo-thermal is not explored in Pakistan though it has promising

potential. Oil and gas companies have the records of their digging works which may prove

helpful in assessing the feasibility of geothermal energy systems.

3.6.7 Hydrogen based Energy Resources

World is moving towards hydrogen economy i.e. the conversion of internal combustion engines

to hydrogen fueled engines and development of the fuel cells. A fuel cell converts hydrogen and

47

3.6.8. Thar Coal and Coal Gasification

oxygen from air to electricity, water and heat. Fuel cells are categorized by the electrolyte used

and can be classified, on the basis of their operating temperatures, into two main categories:

1. Low temperature fuel cells (60-2500C)

2. High temperature fuel cells (600-10000C)

Low temperature fuel cells have been largely deployed in transport applications owing to their

quick start times, compact volume and lower weight comparedto high temperature fuel cells.

General types of low temperature fuel cells are proton exchange membrane fuel cells, phospho-

ric acid fuel cells, alkaline fuel cells, unitized regenerative fuel cells, direct methanol fuel cells

etc. High temperature fuel cells are more efficient in generating electricity than low tempera-

ture fuel cells. In addition, they provide high temperaturewaste heat, which is a problem for

transportation applications, but is a benefit in stationarygeneration applications. General types

of high temperature fuel cells are molten carbonate fuel cells and solid oxide electrolyte fuel

cells.

In Pakistan there is a need to realize this shift towards the hydrogen economy. In first stage,

hydrogen can be mixed along with natural gas to fuel the internal combustion engines. Develop-

ment of the hydrogen storage and distribution networks may be taken in the midterm planning.

It is expected that the world will shift to hydrogen economy up to the half of this century.

3.6.8 Thar Coal and Coal Gasification

Coal provides almost 27.3 % of the global primary energy supplies and is the second largest

source of energy. Burning of coal to convert it into coal gas under the earth in an environmen-

tal friendly manner is a well proven technology. There are following four techniques of coal

gasification.

• Sasol-Lurgi dry ash

• GE (originally developed by Texaco)

• Shell

• Conoco Phillips E-gas (originally developed by Dow)

48

3.6.9. Ocean Energy

In Pakistan the contribution of coal is 6 % in the energy mix and only 0.1 % in electricity genera-

tion. Only coal power plant in Pakistan is at Lakhra with operating capacity of 30 MW. Accord-

ing to HDIP, the country had to import around 3.064 MTOE of coal amounting to Rs.34,937

Million in 2010. Thar coal reserves are the 5th largest coal field in the world, with an esti-

mated 176 Billion tons of coal, and are considered sufficientfor the generation of 20,000 MW

of electricity for the next 40 years [111]. The exploitationof the coal reserves has not been

possible due to lack of mining and gasification technology. Survey of Pakistan has determined

the quality of coal using 2000 samples and showed that the coal is in range of Lignite B to

Sub-bituminous A with relatively low in sulfur and ash contents. Low quality coal is suitably

converted to coal gas which is mixture of carbon mono oxide, carbon dioxide ash and water

vapors through Sasol Lugri technique [111].

3.6.9 Ocean Energy

Energy can be captured from ocean waves, tides, salinity gradients and difference in ocean

temperatures. There is a lot of potential with respect to ocean related technologies such as tidal

waves, osmotic power and trapped methane sources under the surface of the ocean etc. The

worldwide installed capacity of ocean power is 527 MW, mostly in the form of pilot projects.

The tidal waves energy transformation technology is more mature than the osmotic and trapped

methane conversion technology. Pakistan has 1000 Km long coastal line. A lot of research

scope exists to quantify and assess the feasibility of this potential.

3.6.10 Analysis of Potential

At present, two ministries and more than a dozen affiliated organizations are responsible for

development of energy resources in Pakistan. It is proposedthat an independent energy author-

ity is required to study and formulate plans to meet the demand in the future. Pakistan has a

huge potential of different energy forms as shown in Table 3.6. However, proper planning is

required to harness these resources. These energy resources may be categorized on the basis of

their total potential, technological maturity, economic feasibility, seasonal reliability etc. and

strategies for short, medium and long term can be devised in order to meet the future energy

requirements of the country in the most feasible manner.

49

3.7. Smart Grid Implementation in Pakistan

Table 3.6: Potential of Different Resources in Pakistan

Sr. No. Energy Source Potential

1 Oil Reserves 498 MTOE approx

2 Gas Reserves 41.13 MTOE approx

3 Hydro Power 59 GW estimated

4 Solar Energy 2900 GW estimated

5 Wind Energy 55 GW estimated

7 Thar Coal 2700 Million Tons (measured)

8 Small Hydro 2 GW estimated

6 Nuclear Power Planned 8.8 GW up to 2030

9 Hydrogen based cells Still needed to be quantified

10 Geo-thermal potential -do-

11 Ocean related sources-do-

12 Bio mass -do-

3.7 Smart Grid Implementation in Pakistan

Various smart metering experimental projects have been commenced in different cities of Pak-

istan like Islamabad, Lahore, Gujranwala, Multan etc. The projects at Islamabad and Lahore

have been installed by M/S TelcoNet Services (Pvt.) LimitedIslamabad. Projects at Multan

and Gujranwala have been commissioned by MicroTech Industries (Pvt.) Limited Lahore. In

Islamabad Electric Supply Company (IESCO) and Lahore Electric Supply Company (LESCO)

experimental sites, mesh networks have been formed by connecting each meter to its neighbor-

hood using radio communication. Data could be intelligently sent from meter to meter to form

the most efficient path to the nearest gateway. The meter interface units have the capability to

form a very robust mesh network. The gateway supports a number of communications tech-

nologies for transporting data to the utility office. These technologies include: GPRS, ZigBee,

WiFi, Ethernet, DSL, etc. [112].

GSM based Automatic Meter Reading (AMR) system has been introduced by GEPCO for in-

dustrial meters installed on furnaces. The data connector is a communication server which is

installed on the transformer to collect data from these meters over Power Line Carrier (PLC).

50

3.8. Conclusions of the Chapter

Then collected data is forwarded to host server station via General Packet Radio Service (GPRS)

network, where data base is installed. There is no need of external modem or other hardware

for Global System for Mobiles (GSM) network. This system hasthe cheapest running cost as

Short Messaging Service (SMS) cost in GSM is very low. These meters are designed and built

to fulfill the requirements of industry as well as commercialconsumers. AMR metering offers

live monitoring of the electricity usage that aids to decrease operational cost, improves control

by giving real-time meter tampering alarms and improves theefficiency with the delivery of

different reports over the management system for the valuable analyses.

MicroTech is successfully running a experimental project at Multan Electric Power Company

(MEPCO) with its AMR Meters R411G which is also a GSM based AMRSystem [113]. The

government has decided to introduce prepaid smart meters inorder to control electricity theft

in Pakistan. These meters use prepaid cards for billing purpose. In future, electricity will be

provided using prepaid cards like in mobile phones in Pakistan. Users will be needed to credit

their electricity cards according to electricity needs every month [114].

3.8 Conclusions of the Chapter

Pakistan’s electricity requirements will triple by 2050. If sufcient resources are not allocated,

the energy crisis in the country will intensify. An independent energy authority at the national

level is required to make future plans for the development and utilization of indigenous re-

sources like hydro, coal, nuclear and renewables. It is alsorequired to analyze the available

options to import energy from neighboring countries in order to secure the future of the coun-

try. The comparison of TAPI and IPI with LNG import options reveals the suitability of LNG

over the gas pipelines. The optimum utilization of existinginstalled thermal generation and use

of combined cycle power plants is imperative. Pakistan has abundant potential of renewable

energy resources. The issue of circular debt can be minimized by introducing more renewable

energy in the national grid.

With the growing trend of smart grid application, Pakistan’s electricity producers are seeking

for the domestic and foreign investors. This will facilitate the utilities to overcome the elec-

tricity shortage. The current electricity market structure is not suitable for the investors due

to the monopolistic approach of WAPDA and KESC. The government should encourage the

private investors to contribute in smart grid. This change in policy will make the market more

51


competitive and consequently it will create an facilitative environment for smart grid and DSM

implementation and thus better customers satisfaction in the sense of cost minimization and

quality services. The smart grid implementation is much easier in competitive market as com-

pared to monopolistic market because the potential buyers have the flexibility of purchasing

electricity from different producers. In essence, the government should change policies to make

the electricity market competitive by encouraging the domestic and foreign investors.

The smart grid concepts of renewable energy resources integration, DSM, appliance scheduling

etc. are imperative in elimination of Pakistan’s current energy problems. This will become

clearer as we investigate smart grid technologies and applications in the following chapter.

52

Chapter 4Smart Grid Applications and Technologies


Smart grid is envisioned to meet the 21st century energy requirements in a sophisticated

manner with real time approach by integrating the latest digital communications and

advanced control technologies to the existing power grid. It will connect the global

users through energy efficiency and awareness corridor. This chapter presents smart grid funda-

mentals, applications, dynamic pricing and technologies required for smart grid implementation

in a systematic way. Smart grid applications including AMI,HEMS, DA, integration of EVs,

etc. are described in detail. Smart grid technologies with an emphasis on WCTs for HANs

and NANs are presented. Various network attributes like IP support, power usage, data rate etc.

are considered to compare the communications technologiesin smart grid context. Techniques

suitable for HANs like ZigBee, Bluetooth, Wi-Fi, 6LoWPAN and Z-Wave are discussed and

compared in context of consumer concerns and network attributes. A similar approach in con-

text of utilities concerns is adopted for NANs which includeWiMAX and GSM based cellular

standards. Dynamic pricing including ToU, RTP and CPP have been discussed in order to relate

with smart home appliances scheduling in following chapters. Smart grid issues and challenges

are elaborated at end of the chapter.

53

4.2. Smart Grid Applications

4.2 Smart Grid Applications

This section is dedicated for smart grid applications. Advanced ICT has converted the world

into a global village. It may be expected that implementation of the smart grid will ultimately

connect the global users through energy efficiency and awareness corridor. Smart grid has many

applications like AMI, DA, HEMS, EVs management, integration of DG from renewable energy

resources in an efficient and reliable way etc. Some of these applications are discussed in detail

in the following subsections.

4.2.1 AMI

A bi-directional communications network made by the integration of various technologies such

as smart meters, HANs, advanced sensors, control systems, standardized software interfaces

and information management systems to allow the gathering and dissemination of information

between user-end and utilities is known as AMI. These technologies are further integrated with

existing utility operations. AMI has potential to provide utilities with data related to energy

consumption for billing purposes, data on power-quality, voltage and load profiles. Obtaining

this data through AMI will eliminate the need for field trips of personnel for meter reading,

manual outage reporting and most restoration operations. In other words, it will provide remote

meter management (remote connect/disconnect) and outage detection. Via AMI, utilities and

customers will be in constant contact thus enabling utilities to send near-real time price infor-

mation so that consumers may consider energy conservation during peak hours to reduce bills

and control carbon dioxide emissions. Also AMI will enable demand management function-

ality in order to meet user demand in near real time. Moreovernear real time prices could be

communicated to user equipment via HEMS. These price signals would then be used to make

consumption adjustments based on learnt user preferences.Such an advanced control system

requires suitable communications networks both at user premises and to connect utilities and

consumers. Above mentioned applications are summarized inFig. 4.1. AMI can be thought

of as one of the modern grid milestones along with advanced distribution system, advanced

transmission system and assets integration. Smart meters in a neighborhood are linked to data

collectors which serve as central units. Together they forma NAN. In theory, these data collec-

tors can be either part of smart meters or separate devices. Every data collector is linked to an

54

4.2.1. AMI

Figure 4.1: AMI Features

AMI wide area network or backhaul through a collection point. All this data is provided to a

management application at a server at the utility end [115].Alternatively distributed manage-

ment application at multiple servers may be employed. A NAN has a coverage distance of a

few square miles. It requires bandwidth of the order 100500 Kbps (depending on desired data

services) and bi-directional communications. Latency requirements for NAN are between 1 and

15 sec [116].

Typically AMI requires infrequent uplink transmissions with small packet sizes leading to low

bandwidth requirements for individual consumers (overallrequirements increase considerably

due to a huge number of customers) and are latency tolerant. AMI systems typically use a

communications interval of 15 minutes to once per hour [116]. Broadcasting and multicas-

ting support helps to avoid sequential meter reading in AMI networks [117]. Potential pri-

vacy and security problems are associated with communications of metering data wirelessly.

Consumption information is a good indicator of consumers daily activities and presence or ab-

sence. Therefore like similar internetworking systems AMIis vulnerable to security threats [8]

and there is need for end-to-end encryption to provide confidentiality of metering data [118].

55

4.2.2. HEMS

Electricity theft and non-technical losses is another important issue of the power system distri-

bution. AMI can help in detection and reduction of electrical theft issues as well. Anas et al

[105] have reviewed electricity theft and detection issuesusing AMI and showed the reduction

in non-technical losses through MATLAB simulations. A comprehensive review of the issues,

challenges and advantages of smart meters and AMI has been carried out in [119] which covers

various technologies and features that can be included in smart meters. The authors have also

mentioned design, implementations and utilization challenges about AMI.

4.2.2 HEMS

In energy sector, dynamic demand control is needed in order to save cost for utilities and the

consumers. Moreover consumers have become more conscious about energy efficiency and

want to be aware of real time energy consumption and to be a part of its effective management.

Furthermore, increasing trend towards smart home appliances/devices necessitates reliable en-

ergy. At the same time, in many countries including Pakistan, there is a need of selective load

shedding or DLC upon the basis of the fair and implementable loads selection criteria in order

to solve the recent energy crisis. Implementation of this DLC mechanism is subjected to the

provision of HEMS to be placed in the user premises. The interaction of HEMS with various

smart entities is indicated in Fig. 4.2.

Basic theme is to be aware of energy usage at any time and to control energy flow in order to

achieve benefits such as saving money, automation, remote central control etc. The information

flows from appliances to HEMS through sensors and an efficientsensor network is required

to achieve high reliability and performance. HEMS benefit both the utility and consumer by

enabling energy supervision, monitoring and control. Its role is to intelligently monitor and

control energy usage by interfacing to smart devices, smartappliances, smart plugs and smart

meters as well as to provide peak load management.

The communications network connecting HEMS with the smart components at a certain premises

is the HAN. A smart meter is intended for the basic functions so that it can operate for longer

period of time and without any requirement of frequent software or hardware up gradation. For

the complex operations, a separate HEMS is employed. HEMS provide a certain level of in-

telligence implemented in software that can reside on a dedicated hardware. Upgrading and

installing new functions and protocols to an EMS is independent of smart meters. Advanced

56

4.2.2. HEMS

Figure 4.2: Interaction of HEMS with Various Smart Entities

HEMS can also include other aspects of HAN such as security systems. The consumers are

benefited heavily from such a system as they can monitor the real time energy used by the in-

dividual appliances. Moreover they can compare energy efficiency of different appliances and

control various appliances like air conditioning and ventilation systems, smart appliances and

smart plugs etc. On the other hand, utilities also get benefited from HEMS as the peak load

management can be employed using the available data. Dynamic DR management, which is

a desirable feature of smart grid system, can be integrated with individual household profiles

using HEMS. Baig et al have suggested a smart HEMS for smart grid using ZigBee sensors and

an interface designed in LABVIEW [120]. Khan et al [121], andNaeem et al. [122] have re-

viewed the HEMSs and optimization techniques proposed in context of smart grid for effective

DSM.

57

4.2.3. DA and DERs

4.2.3 DA and DERs

Reliability of traditional power distribution systems canbe significantly improved by employ-

ing real time monitoring and intelligent control systems. In such a smart system, sensors at

various distribution components would transmit data (status, current, voltage, frequency etc.)

via a suitable communication technology to a control system. The control signals by the con-

trol system would then control power supply and different quality parameters. Note that such

control system can be centralized or localized/distributed.

Moreover, the smart grid will enable large scale integration of DERs with the electricity grid.

The legacy power generation and transmission concept of large and far-off generation units pro-

viding power via lengthy cables, results in great losses in the form of heat. DERs can provide

power in a more cost-effective and efficient manner from sites nearer to the customers. Such an

integrated distribution system offers greater reliability and robustness as the backup is always

available [123], [8].

Integration of DERs with the electricity grid results in twoway power flow in contrast to the

traditional one way power flow. This creates some issues for the distribution network operators

since the legacy distribution networks were static requiring no significant control or reconfigura-

tion operations. However all DERs to be integrated into the smart grid would have an electronic

power processor and switching power interface for controlling the exchange of power and cur-

rents with the grid [8]. Thus in a smart grid the distributionnetworks will be in a constant

state of change depending on the amount and direction of power flow. This will require en-

ergy management system to embrace a more active approach rather than the traditional passive

one so that the distribution network can be reconfigured depending on the power flow changes.

Traditional control systems employed in industry like Supervisory Control and Data Acquisi-

tion (SCADA) can be modified to implement distribution management. Also, an active control

system needs access to control information from the distribution network. Therefore, a great

number of sensors will have to be deployed for monitoring of system conditions such as faults,

status of switches and circuit breakers, sectionalizers and reclosers, power flow direction and

magnitude. This would require a large transmission bandwidth and low latency so that control

information is provided to the controllers quickly [123].

Active control also symbolizes a move towards a more distributed control configuration. Dis-

tribution networks are divided into small and separate entities called micro-grids each with an

58

4.2.3. DA and DERs

Figure 4.3: Distribution Automation with Central Control and DERs

autonomous intelligent controller to manage it. These intelligent controllers receive sensor data

and actuate control operations on the micro-grid level, thereby reducing computational load at

the control room. The controllers of different micro-gridscommunicate among each other and

run collaboratively to achieve better management [123]. Load classification is also an impor-

tant aspect of the DSM and DA which is covered in [124]. A process model, consisting of five

stages, is adopted for load classification in smart grid. Clustering methods for load classification

have also been covered by the authors. Conceptual diagram for power flow, communication and

control network integration for automation of distribution is shown in Fig. 4.3.

59

4.2.4. EVs

Figure 4.4: EVs Management System

4.2.4 EVs

Current fossil fuel based transportation systems are a major source of greenhouse gas emissions.

Zero-carbon EVs are becoming commercially practical due torecent advancements in fuel cell

technologies [125]. EVs have significant potential to cut greenhouse gas emissions and fossil

fuel imports. These economic and environmental gains are accompanied with challenges like

enormous load growth and power quality issues such as frequency and voltage instabilities asso-

ciated with large-scale EV integration with the electricalgrid. For instance, if too many vehicles

start charging during off-peak periods, when electricity prices are low, then violations of feeder

thermal limits and distribution transformer capacity limits may take place. Similarly if many

vehicles with vehicle-to-ground capabilities simultaneously release energy back to the grid by

discharging their batteries (to take advantage of high prices) then the power system frequency

may become unstable.

A special purpose EV management system is required to properly coordinate the charging and

discharging process to avoid overload conditions of the grid. On-the-go communications be-

tween EVs and the grid is essential because of vehicular mobility. Moving EVs transmit pa-

rameters like location, battery status, and charging and discharging demand to the EV manage-

ment system which calculates real-time electricity price based on this information. Electricity

prices then regulate the charging/ discharging demand and power system stability is maintained

[118, 123]. Concept is elaborated pictorially in Fig.4.4.

Only wireless communications can provide such informationexchange. Existing cellular

60

4.3. Wireless Communication Options for HANs

communications networks already provide required coverage for mobile vehicles. Communi-

cations costs can be reduced by using Vehicle Ad Hoc Networks(VANETs) and Vehicle-to-

Infrastructure (V2I) communications related issues like handover and mobility support need to

be handled properly [118]. The communications infrastructure should be secure, scalable, and

reliable, have sufficient throughput with low latency for real-time behavior and be cost-effective.

A common standard regulating charging process related communications is also required. Ses-

sion Initiation Protocol (SIP) and the IEC 61850 protocol may be used for reliable, scalable and

secure communications between EVs and EV management system[117].

Realization of smart grid will be accompanied with a lot of benefits for mankind, however still

there are certain challenges and issues which are needed to be addressed. These challenges are

briefed in subsequent section. We need to compare some majoravailable WCTs for optimized

asset management in smart grid. There are two major areas of application for smart grid: Indoor

and outdoor. It should be noted here that HANs need short range, low data rate and low power

wireless technologies to be used within households while NANs require relatively long range,

high data rate and secure communication in order to exchangecontrol and data signals among

utilities and consumer premises. Smart grid, LM and relatedtechnologies have been discussed

in following section.

4.3 Wireless Communication Options for HANs

There are two different sets of communications technologies based on wired and wireless me-

dia. Each of these technologies has its own advantages and disadvantages that vary according

to nature of application. IEEE and many other regional and international bodies have identified

a number of wired as well as WCTs in smart grid applications. However, in many smart grid

applications the sheer number of communications links makes the use of wired solutions eco-

nomically and/or physically prohibitive. On the other handWireless technologies offer benefits

such as lower cost of equipment and installation, quick deployment, widespread access and

greater flexibility [126, 127].

This segment investigates suitability of different short range wireless communication proto-

cols for smart grid HANs implementation. ZigBee, Bluetooth, Wi-Fi, 6LoWPAN and Z-wave

are the technologies to be discussed. IEEE has set differentstandards for first three of these

technologies in which only the Physical (PHY) and Medium Access Control (MAC) layers are

61

4.3.1. ZigBee

defined: Bluetooth (802.15.1a), ZigBee (802.15.4b) and Wi-Fi (802.11g). Internet Engineering

Task Force (IETF) has introduced 6LoWPAN standard in order to achieve IPv6 enabled low

power communications. Z-Wave, a proprietary solution, is also considered for comparison pur-

pose. One of the major goals of this work is to provide a comparison basis for the main WCTs

to be used in HANs in the smart grid context. WCTs for HANs are discussed in detail in the

following subsections.

4.3.1 ZigBee

ZigBee [46, 47] is a wireless mesh network, built on the IEEE standard 802.15.4, is very effi-

cient and cost effective solution. However, it offers low data rate for Personal Area Networks

(PANs). This technology can be used broadly in device control, reliable messaging, home and

building automation, consumer electronics, remote monitoring, health care, and many other

areas. It is a low power network provided all devices are interconnected by IEEE 802.15.4

links with Direct Sequence Spread Spectrum (DSSS). Estimated data rates are 250 Kbps per

channel in the unlicensed 2.4 GHz band, 40 Kbps per channel inthe 915 MHz band and 20

Kbps per channel in the 868 MHz band. ZigBee supports 10-75 meter point to point, typically

30 m indoor and unlimited distance with mesh networking. In amesh network each node can

be reached by multiple links and connections are dynamically updated and optimized. Mesh

networks are de-centralized and each node can manage itselfin the changing conditions and

is able to dynamically self-route and connect with new nodesas needed. These features offer

scalability, greater stability and tolerance against node/link failures. This along with low power

utilization and low deployment cost makes ZigBee very attractive for the smart grid HAN ap-

plications.

ZigBee Alliance, an association of companies offering wireless solutions, has offered many

standards to suit different set of requirements. These standards include ZigBee Building Au-

tomation, ZigBee Remote Control, ZigBee Smart Energy, Smart Energy Profile 2, ZigBee

Health care, ZigBee Home Automation, ZigBee Input Device, ZigBee Light Link, ZigBee

Telecom Services and ZigBee Network Devices. Smart Energy Profile 2 (SEP2) offers IP func-

tionality which enables large scale multi-vendor deployment of smart meters, sensors, smart

appliances and energy displays etc. This will also allow communicating with IPv6 based nodes

using other network architectures like Wi-Fi, Ethernet etc. offering ever greater interoperability

62

4.3.2. WLAN and Wi-Fi

and compatibility.

ZigBee, a type of micro-power wireless communications technology, is supported by IEEE

802.15.4 standard that defines PHY and MAC layers. IEEE specifications are used by ZigBee

Alliance for structural expansion of network layer and application layer. A ZigBee network

uses three types of nodes: ZigBee Coordinator (Full Function Device, FFD), ZigBee Router

(FFD) and ZigBee End Device (Reduced Function Device, RFD) [128]. ZigBee Coordina-

tor initiates session and manages the security. Beacon and non-beacon enabled networks are

supported in ZigBee protocols. In beacon-enabled networks, the special network nodes (such

as ZigBee Routers) broadcast cyclic signals for verification of their existence to other nodes.

It helps receiving/transmitting data and increases efficiency of the network. In non-beacon-

enabled networks ZigBee Routers have their receivers continuously active, involving a more

vigorous power supply. Beacon intervals depend on data rate, at 250 Kbps 15.36 milliseconds

to 251.65824 seconds, at 40 Kbps 24 milliseconds to 393.216 seconds and at 20 Kbps 0.48 sec

to 786.432 seconds [115]. ZigBee uses access control list orAdvanced Encryption Standard

(AES-128) to guarantee a high-level security.

4.3.2 WLAN and Wi-Fi

WLAN, based on IEEE 802.11, employs the spread spectrum technology so that users can oc-

cupy the same frequency bands while causing minimal interference to each other. Networks

based on 802.11 are most popular for LAN usage with maximum data rates of 150 Mbps and

maximum coverage distances of 250 m [129]. WLAN also known asWireless Ethernet is able

to provide robust communications with low latency and capable of point-to-point as well as

point-to-multipoint transmissions. Wi-Fi (802.11b), operating on 2.4 GHz with DSSS modula-

tion, gives maximum data-rates of 11Mbps with a latency of 3.2 to 17 milliseconds [129]. Other

technologies like those based on IEEE 802.11a, which operates on 5.8 GHz using Orthogonal

Frequency Division Multiplexing (OFDM) and IEEE 802.11g (enhanced Wi-Fi), operating on

2.4GHz with DSSS modulation, increase obtainable data rateto 54 Mbps. Data rates of up to

600 Mbps can be obtained via 802.11n which uses Multiple Input Multiple Output (MIMO)

scheme. Security issues for WLANs are addressed in IEEE 802.11i (WPA-2) which uses the

advanced encryption standard [130].

Wi-Fi is a more popular name for certain 802.11 based technologies used in HANs, mobile

63

4.3.2. WLAN and Wi-Fi

phones, computers and many other electronic devices. Wi-Fitechnologies include 802.11n

(300 Mbps), 802.11b (11 Mbps), 802.11g (54 Mbps) and 802.11a(54 Mbps). Its main feature

is existing wide support; almost every new electronic device, be it a computer, laptop, game

console or a peripheral device comes with Wi-Fi technology.Wi-Fi is generally upper layer

protocol with IP being the most predominant protocol, allowing communications over the in-

ternet without needing a protocol translator. A non-profit international association of various

businesses/companies, Wi-Fi Association, formed in 1999,works to support and maintain Wi-

Fi technology.

Channel assignment for Wi-Fi is not consistent across the globe. For example, the number

of assigned channels is 11 in USA, 13 in Europe and 14 in Japan.Only a limited number of

channels may be used without any overlap which means a limited number of devices can be

connected in a Wi-Fi Wireless Local Area Network (WLAN) [131]. This interference along

with security issues are the big challenges in using Wi-Fi for the smart grid HAN applications.

However, advantage of Wi-Fi lies in high data rate, IP support, wide spread availability and

scalability [132].

WLAN/Wi-Fi is more suitable for applications with relatively smaller data rate requirements

and low interference environments. Ethernet based communications for interoperable sub-

station automation systems have been suggested by IEC 61850standard. IEC 61850 based

WLAN can boost the protection of distribution substations via intelligent monitoring and con-

trol using sensors and intelligent controllers with wireless interfaces. IEEE Power System Re-

laying Committee reports intra-substation applications like enhanced transformer differential

protection through the use of wireless Load Tap Changer (LTC) sensors. Communications

for Monitoring and Control at substations can be made more reliable via usage of redundant

wireless links alongside optical fibers. WLAN (Wi-Fi) with intelligent sensing and control

devices may also be used in differential protection applications. It is possible to develop a

dynamic self-organizing, self-healing network by combining WLAN with Wireless Mesh con-

cepts. Wireless-Mesh uses multi-hop routing for greater coverage distances with low trans-

mission powers [117]. Smart meters with Wi-Fi modules may beused for signal repetition

and addition of repeaters increases the coverage area and network capacity [128]. WLAN

can provide the aforementioned services for both Distributed Energy Resources (DERs) and

distribution substations. There are various challenges when dealing with WLANs such as:

• Electromagnetic interference: because high voltage electrical equipment can slow down

64

4.3.3. Bluetooth

communications or even can vanish out the signal entirely,

• Interference from other wireless equipment: will have unfavorable effects on communi-

cations,

• Wireless equipment for industrial environment is not readily available.

However advancements in acknowledgement protocols, algorithms for error correction and

buffering of data have increased the reliability of wireless communications. Developments like

Smart Antennas and waveguides have made it possible to combat radio frequency and electro-

magnetic interference for production of better WLAN devices [130].

4.3.3 Bluetooth

Bluetooth [133, 134] is another common wireless communications system used to exchange

data over short distances. It employs short wavelength radio transmission in the Industrial,

Scientific and Medical (ISM) band (2400-2480 MHz). Its main features are low power con-

sumption and fast data exchange as well as wide spread availability. Bluetooth technology was

developed by engineers at Ericsson in 1994. Later a group of companies started using Bluetooth

and made a special interest group (SIG) to maintain and enhance this technology. The IEEE

standard for Bluetooth is IEEE 802.15.1.

There are two topologies used in the Bluetooth i.e. Piconet and Scatternet. A Piconet is formed

by a Wireless Personal Area Network (WPAN) in which a mobile device is acting as a master

and other mobile devices will be serving as slaves. A Scatternet consists of two or more Pi-

conets. Bluetooth can be used for communications among smart home appliances, EMS and

the smart meter. It has maximum data rate is 1 Mbps, nominal range of 10 m, 79 RF channels,

1 MHz channel bandwidth and Max of 8 nodes. Bluetooth has three classes i.e., Class 1, Class

2 and Class 3 each having a different range.

Bluetooth has very short range which can be a problem when using this technology in smart

HANs, as longer distances may be involved. Moreover, it supports limited number of nodes

that can be a serious constraint in HANs. Bluetooth, like many other technologies is operated

at low power which means the strong noise can cause signals tobe lost or damaged. Moreover,

it works on 2.4 GHz and has interference issues with other wireless technologies like Wi-Fi,

ZigBee etc. with same system frequency. Furthermore, Bluetooth has some inherent security

65

4.3.4. 6LoWPAN

issues [135]. On the other hand, the most recent version of Bluetooth v4.0 has introduced

low energy technology. This low energy technology is being investigated by researchers for

IP support. See for instance [136] where the authors have implemented Bluetooth low energy

connectivity with IPv6 support.

4.3.4 6LoWPAN

6LoWPAN is the IETF working group formed in 2005. It enables IEEE 802.15.4 (IEEE sub-

committee for low rate WPAN) and IPv6 to work together in order to achieve IP enabled low

power networks of small devices including sensors, controllers etc. IETF RFC 4944 describes

the mechanism of combining IP and WPAN technologies. The Authors in [137] analyzed

6LoWPAN along with IPv6 for home automation networks. They used a web interface in order

to manipulate home appliances and described usefulness andchallenges of IPv6 and 6LoWPAN

in this regard. A smart grid specific discussion on IPv6 and 6LoWPAN application is presented

in [138] where the authors emphasized on memory management and portability challenges.

Various technology vendors are trying to adopt 6LoWPAN based protocols in order to achieve

IP functionality. For instance, ZigBee Alliance has developed an IP network specification, Zig-

Bee IP, which is based on IETF protocols including 6LoWPAN.

6LoWPAN uses mesh topology to support high scalability. Thescalability is also affected by

choice of routing protocols. For instance, Hierarchical routing is one of the routing proto-

cols used in 6LoWPAN to increase the network scalability [139]. Mesh topology also offers

the self-healing capability to the network as the traffic canbe re-routed in case of a broken

link. 6LoWPAN offers high interoperability as IP is supported in most of the modern technolo-

gies. An investigation of interoperability of 6LoWPAN-based web applications can be found

in [140]. Security of 6LoWPAN is one of the challenges faced by researchers and there are

ongoing efforts on improving security level. See for instance [141] where IPSec is used along

with IP to enhance 6LoWPAN security. The authors in [142] have compared link-layer security

and IPv6 security for 6LowPAN.

4.3.5 Z-Wave

A proprietary standard intended exclusively for remote control applications in residential and

business areas is given the name of Z-Wave. This protocol works at 868 MHz in Europe and

66

4.3.6. Comparative Study

908 MHz ISM band in USA. It has typically 30 m indoor range which extends up to 100 m

outdoor. Mesh networking is employed in Z-Wave which essentially means unlimited range.

The main advantages of this technology come from simple command structure, freedom from

house hold interference, low bandwidth control medium and IP support.

Z-Wave offered a low data rate of 9.6 Kbps; however it was extended to 40 Kbps later on

[143]. Z-Wave 400 series also supports 2.4 GHz band and 200 Kbps data rate [144]. Z-wave

automatically routes the message from one node to the other because of the routing capability

of all the nodes. Controllers and slaves are two types of devices defined by Z-wave. Controller

maintains the network topology. Slaves can also be used as routers and are useful for monitoring

of sensors.

4.3.6 Comparative Study

Comparative study of ZigBee, Wi-Fi, Bluetooth and 6LoWPAN reveals that Wi-Fi has advan-

tages of larger coverage and widespread availability, Bluetooth is easily accessible and provides

secure short range communications, 6LoWPAN adds IP functionality to WPANs and consumes

low power while ZigBee accommodates much more nodes, operates at low power and requires

low cost. ZigBee can be used in HANs, as well as for smart metering if used in a mesh struc-

ture. It can also provide remote monitoring of the smart meter and other devices. ZigBee has

reliable security and employs powerful encryption techniques. It has far superior networking

technique compared to other technologies which avoids the channel collision. On the other

hand, 6LoWPAN is suitable for IP enabled low power devices like sensors and controllers. Z-

Wave is characterized by the reliable transmission of shortmessages from controller to one or

more nodes. The main attributes of these technologies have been summarized in Table 4.1.

One of the major features of smart grid is bi-directional flowof information and power among

users and utilities. Achievement of this feature requires larger networks or NANS. Wireless

communications for NANs are discussed and compared in the following section 3.3.

4.4 Wireless Communication Options for NANs

Communications among energy utilities, smart meters, HANs, DERs and other possible smart

grid entities requires a large network with appropriate network architecture and communica-

67

4.4. Wireless Communication Options for NANs

Table 4.1: Comparison of HAN Technologies for Smart Grid

Aspect ZigBee Wi-Fi Bluetooth 6LoWPAN Z-Wave

Max. Speed

per Channel

250Kbps

(2.4GHz)

40Kbps

(915MHz)

11Mbps to

300MbpsMax. 1Mbps

250

(2.4GHz)

40Kbps

(915MHz)

20Kbps

(868MHz)

40Kbps

Reach 10-75m100m (in-

door)10m typical Up to 200m

30m in-

door, 100m

outdoor

StandardIEEE

802.15.4IEEE 802.11

IEEE

802.15.1

IETF RFC

4944Proprietary

IP SupportIPv6 only in

SEP2IPv6

Not

presently,

research

continues

Yes, IPv6 Yes

Adoption

Rate

Widely

adopted

Extremely

high

Extremely

highMedium Medium

Unique

Value

Low cost,

low power

usage, high

number of

nodes

High speed

mature

standards

Ease of

access,

no con-

figuration

require-

ment, secure

connection

Benefits of

both IP and

Bluetooth,

low power

consumption

No interfer-

ence from

household

devices

68

4.4.1. WiMAX

tions technology. Our objective in this section is to explore and compare communications tech-

nologies available to fulfill smart grid NAN requirements. Some of the NAN technologies are

discussed in the following subsections.

4.4.1 WiMAX

WiMAX, one of the standards from IEEE 802.16 series designedfor Wireless Metropolitan

Area Network (WMAN), aims at achieving worldwide interoperability for microwave access (2

to 66 GHz). WiMAX is characterized by low latency (¡100 ms forround trip), proper security

protocols (AES, AAA, etc.), lower deployment and operatingcosts, scalability [128] and avail-

ability of traffic management tools (traffic prioritization, Quality of service, etc.). Developing

WiMAX based architecture requires developing a utility-proprietary network having total con-

trol of traffic management and capable of coping with regularand emergency conditions [117].

The bandwidth and the range of WiMAX allow its use for smart grid NANs successfully [116].

For example, WiMAX can be used for:

• Wireless Automatic Meter Reading (WAMR): WiMAX technology is suitable for WAMR

as part of a utilitys AMI network as it offers efficient coverage and high data rates.

• RTP: WiMAX based AMI can easily be employed for provision ofreal time price signals

to the consumers based on their real time energy usage.

• Detection and Restoration of Outage: Using bi-directional WiMAX links, outages can be

quickly detected and power restored resulting in increasedreliability of power supply.

• Monitoring: Sensor data can be transmitted over WiMAX links for monitoring purposes.

There are certain challenges associated with WiMAX. For instance, WiMAX towers are based

on relatively costly radio equipment, care must be taken to ensure optimal locations are chosen

so that infrastructure expenses are reduced and QoS requirements are met. WiMAX frequencies

¿ 10 GHz are unable to pass through obstacles, hence lower frequencies are better suited for

Advanced metering applications particularly for urban areas. Since the lower frequency bands

are already licensed, leasing from third parties may be required [130].

69

4.4.2. Cellular Network Communications

4.4.2 Cellular Network Communications

2G, 2.5G, 3G and LTE (Long Term Evolution) are potential cellular communications technolo-

gies which can be used for smart grid communications. Data rates have evolved from 14.4 Kbps

for 2G-GSM to 56-171 Kbps for 2G-GPRS (General Packet Radio Service), to 2 Mbps in 3G

and recently to 50-100 Mbps for 4G-LTE [145]. Using pre-existing cellular networks saves

investments on utility dedicated communications infrastructure and allows rapid deployment

of applications. Sufficient bandwidth, High data rates, extensive coverage, lower maintenance

costs and strong security are enabling features of present day cellular networks.

An embedded network operator Subscriber Identification Module (SIM) or GPRS module in-

side a cellular radio unit integrated in smart meters can enable communications among smart

meters and utilities. Since the network operators GSM/GPRSnetwork can handle communi-

cations requirements for smart metering network, utilities can concentrate on applications and

services. GSM and GPRS provide users with anonymity and protection of their data along-

side authentication and signaling protection for security. Numerous cellular network operators

around the world have already approved to put GSM networks into service for AMI communi-

cations [128]. Developing Smart grid Communications network for EVs is possible by using

GSM network and SMS messages [45].

Based on GSM/EDGE (Enhanced Data Rates for GSM Evolution) and Universal Mobile Telecom-

munications System (UMTS) technologies, LTE is one of the latest developments in wireless

communications field. This wireless communications standard provides high speed data trans-

fer (up to 300 Mbps download and 75.4 Mbps upload depending ontechnology used) with low

latency. The first commercial deployment of LTE was experimented in Oslo and Stockholm

in December 2009. It uses OFDMA and Single Carrier FrequencyDomain Multiple Access

(SC-FDMA) in order to use minimum power and supports variouscell sizes (10m to 100Km).

LTE is fully compatible with other legacy standards like GSM/EDGE, UMTS and CDMA2000

etc. The latest form of LTE, known as LTE Advanced, offers approximately 3.3 Gbps download

speed. The LTE Advanced is also in commercial use since October 2012. 3G and 4G-LTE

cellular technologies operate on the licensed frequency range of 824- 894MHz/1900MHz [8].

Large network operators widely support LTE while chipset manufacturers are expected to re-

duce prices associated with LTE hardware [117]. Some mission-critical applications require

continuously available communications. Services of cellular companies are also used for cus-

70

4.4.3. Comparison of Wireless NAN Technologies

tomer markets which could result in congestion issues or reduced network efficiency for emer-

gency scenarios [128]. Emergency communications vehiclesacting as base stations can be

used in these situations. LTE can also fulfill the throughput, error-rate and latency requirements

for emergency situations [117].

LTE has two major application areas in smart grid. First it can be used for automated metering

and secondly for automating and controlling the distribution system. The authors in [146] an-

alyzed the application of LTE for automated electricity distribution systems and found that this

technology is feasible for NAN smart grid applications withhigh reliability and low latency.

LTE application for both automated metering and automated distribution is analyzed in [147].

Using third party networks means monthly recurring chargesfor every connection which trans-

lates into greater operating costs for the utility. To ensure quality and reliability of service and

avoid added operating costs, a utility may decide to build dedicated cellular networks [128].

Cellular networks can be used to:

• Provide communications for wide area smart applications like Automated Demand Re-

sponse (ADR), AMI and outage management [128].

• Provide communications between Remote Terminal units (RTUs) at substations and the

SCADA server at the utility [130].

• Provide communications to enable supervision of remote DERs. Non-critical informa-

tion may be communicated via SMS messages and DERs can be monitored with GPRS

systems [130].

4.4.3 Comparison of Wireless NAN Technologies

Bandwidth and the range of WiMAX make it appropriate for smart grid NANs applications.

However, pre-existing cellular networks with sufficient bandwidth, high data rates, extensive

coverage, lower maintenance costs and strong security can save capital investments on utility

dedicated communications infrastructure and allow rapid deployment of applications. How-

ever, use of pre-existing cellular networks will raise the operational cost in terms of monthly

recurring charges. A dedicated utility cellular network can ensure the high quality and reliabil-

ity with reduced operational cost. This comprehensive review has tabulated different wireless

71

4.4.3. Comparison of Wireless NAN Technologies

Table 4.2: Comparison of NAN Technologies for Smart Grid

Aspect WiMAX 2G-GSM 2.5G-GPRS

3G-UMTS-

CDMA2000-

EDGE

4G-LTE

Max. Speed

per Channel72 Mbps 14Kbps 171Kbps

More than

2Mbps

300Mbps,

3.3Gbps

for LTE

advanced

Reach 9km 10km 10km 10km 10km

Adoption

Rate

Widely

adopted

Extremely

high

Widely

adopted

Widely

adopted

Widely

adopted

Unique

Value

Low cost,

low latency

High Adop-

tion, ex-

tensive

coverage

Uses GSM

network

but sup-

ports data

transmission

High data

rate

Extremely

high data

rate

Applications AMI, ADRAMI, EVs,

ADR

AMI, EVs,

ADR

DERs, AMI,

EVs, ADR

DERs, AMI,

EVs, ADR

communications in a systematic way along with the challenges for smart grid implementation.

A Comparison of NANs wireless technologies is summarized inTable 4.2. Wireless communi-

cations discussed and compared in sections 3.4 and 3.5 can beemployed for various smart grid

applications. It is evident from sections 3.4 and 3.5 that HANs and NANs have very different

communication requirements. The operating environment, coverage range, data rate and secu-

rity requirements are particularly different. Wi-Fi, ZigBee, Bluetooth, 6LowPAN and Z-Wave

are quite suitable for HANs. ZigBee may be preferred becauseof supporting a large number of

nodes and low power consumption. 6LowPAN or Z-Wave may be employed in order to achieve

IP functionality. Wi-Fi may be considered due to its high data rate for some advanced appli-

cations (e.g. video monitoring). On the other hand, WiMAX and GSM-based networks offer

many benefits to NANs such as high data rate (to support bulk communication among utility

and a high number of consumers), greater coverage range and advanced security protocols. The

choice between a utility owned dedicated network and pre-existing cellular network depends on

72

4.5. Dynamic Pricing Schemes

a particular utilitys policies, budget etc.

4.5 Dynamic Pricing Schemes

Generation, transmission and distribution of electricitycosts a certain amount for each KWh of

energy consumed. There are mainly two types of costs involved in this process running/operational

costs and fixed costs. To recover these costs every consumer of the electricity is charged with

certain amount per KWh of electricity and this cost is known as tariff. The types of tariff con-

ventionally used for pricing of electricity are simple tariff, flat rate tariff, block rate tariff, two

part tariff, maximum demand tariff and power factor tariff [148]. But with the advancement of

technology and particularly with smart grids, these conventional tariffs are deficient in fulfilling

the requirements of fair pricing for electricity. With the introduction of DG in the grid system,

it has become very complex for these old tariff methods to comply with the requirements of

smart grids and Intelligent Electronic Devices (IEDs). In response of such needs, smart pric-

ing schemes have been devised to fulfill the requirements of modern systems. These pricing

schemes include RTP, TOU and CPP which are efficient time based pricing schemes. The smart

pricing plays a key role in DSM so that the system works efficiently [149]. DR is the response

of consumers to the changing electricity prices and helps the system to work reliably and ef-

ficiently. A successful DR for different countries depends on the time based pricing model

formulated by keeping in view the structure of the power industry of that country [150]. DSM

is accomplished using two types of programs. First one is incentive based program which is usu-

ally offered in whole sale electricity market in the form of contract. The incentive based DSM

program include DLC, interruptible/ curtailable service,demand bidding/ buy back, emergency

DR program, capacity market program and ancillary service markets. The second kind of DSM

program is time based program. The time based DSM program will make the users to choose

the time of usage of electricity keeping in view the prices ofelectricity [150]. The time based

DSM program involves the smart pricing schemes mentioned above. Detailed analysis of the

effects of different pricing schemes on DSM is discussed below;

73

4.5.1. ToU Pricing

4.5.1 ToU Pricing

This type of smart pricing scheme offers different prices atpeak time and off-peak time. During

the peak time, the prices are relatively higher than off-peak time. Some utilities divide the TOU

pricing scheme into three parts i.e. peak time, mid-peak time and off-peak time [151]. At the

peak time, rates are kept higher because of higher demand of energy, this higher demand will

result in utilization of power from more plants. Such power plants which are brought in the

system to meet the peak load are called peaking power plants and usually include plants which

are based on less efficient and expensive fuels such as diesel, petrol or natural gas as compared

to the base load or mid load plants which are hydro electric, solar thermal, geothermal, nuclear

etc. Furthermore, to meet the peak demand , expansion of existing facilities like installation of

new power plants, expansion of transmission and distribution networks will be required. Once

the supply of electricity increases, the system technical losses will also increase. So to curtail

the electricity usage during peak hours TOU pricing scheme introduces higher prices during

the peak hours. With higher price during the peak hours, users will be tempted to reduce their

electricity consumption during the peak hours [150].The smart pricing schemes are required to

be such that they are fair for utilities and consumers. Consumers can benefit from their own DG

such as a solar or wind generation system installed in their houses. They can defer the load to

the off-peak time and sell the electricity to the grid at a higher price and use electricity during

off-peak time at cheaper rates along with their own generation [152].Further complex forms

of TOU involve seasonal variations. Also peak and off-peak time can vary according to the

dynamic requirements of the system and weather conditions [153]. Prices in TOU scheme for

peak and off-peak hours are decided months earlier and henceprovide consumers a good time

to plan their electricity consumption schedule.

4.5.2 CPP

CPP is a modified form of TOU. This involves some time of year where the energy demand is

very high as compared to the rest of peak time during the year.TOU is unable to discriminate

the time of year when the cost of energy generated exceeds thepeak value set by in the TOU

pricing scheme. Sometimes it costs the utility a very high whole sale energy price which is out

of range of TOU peak price and consumer is charged with every day price. To keep the pricing

74

4.5.3. RTP

system fair, during normal days, the peak price charged to the consumer is same as TOU peak

price [151]. Torectify this economic problem and for peak shaving, CPP is declared. CPP

is declared only when the load forecast shows the load being very high i.e. making the day a

critical day. Usually the CPP day is announced a day ahead of the CPP. The CPP can be declared

for a certain number of days during a year, 15 as given by [152]and the pricing can be 15:1 for

peak to off-peak time [152]. Present situation when nuclearpower plants are being taken out of

service after the Fukushima Nuclear Power Plant accident, installation of new power generating

facilities can be avoided by using CPP [154]. CPP plays a verycritical role in stability of the

system because the price of electricity during the CPP peak duration is exorbitantly high and

naturally electricity consumer will prefer to reduce the load during this period. This will reduce

the peak load and system will be able to operate more flexibly.Also extension of existing and

installation of new facilities will be avoided which would have been inevitable without CPP.

4.5.3 RTP

RTP is considered to be the scheme that best reflects the cost incurred by the utility for the

electricity utilized by the consumer. In RTP scheme consumer is charged with a price nearest

to the real price of generation at that particular interval of time. The RTP scheme can be of

two types, hourly pricing and day ahead pricing. For hourly pricing, the price of electricity

for an hour is announced every single hour for the next hour. While for day ahead pricing of

RTP, the price of electricity for next 24 hours is announced beforehand which is selected by

predicting the load demand and viz-a-viz the generation cost. RTP requires the involvement

of consumers so that they can be provided the cheap electricity when it is produced at a lower

cost [150, 151]. Day ahead pricing can be more effective because consumers can get sufficient

time to plan their electricity consumption schedule, whilehourly pricing can be tedious for

consumers. RTP can be made affective only with the active participation of consumers and the

technological advancement achieved with IEDs can make it possible. Furthermore as discussed

in [11], load control schemes can be devised to benefit utilities and consumers. RTP signals

combined with automation at consumer level will benefit not only the consumer by reducing

load but it will help utility by system peak shaving and by reducing load through DSM in case

of capacity limitation of generation or distribution system. So a properly designed RTP scheme

increases the reliability of system, reduces the generation cost and lowers consumer’s electricity

75

4.6. Smart Grid and DSM Challenges

bill [155, 156]. The RTP can prove to be the most efficient pricing scheme, benefiting every

stake holder involved and optimized by using the automated control system for load.

4.6 Smart Grid and DSM Challenges

Conversion of conventional Power system into Smart grid is too much complex. Many kinds of

challenges are being faced in this transformation process [9]. Communications infrastructure,

required for applications of smart grid, is found among major and the most important chal-

lenges. Continuous research is required to increase understanding of problems associated with

smart grid communications requirements and to provide appropriate solutions. This section dis-

cusses the key issues and challenges to identify the main hurdles in implementation of smart

grid.

Complexity: Smart grid will be very complicated system [4]. Challenges like modeling,

analysis and design of a smart grid including communications infrastructure are required to be

addressed .The complexity of smart grids can be summarized in two main factors: the interde-

pendence between different infrastructures and the distributed nature of monitoring and control

functions.

The Fig. 3.5 illustrates the communications and control complexity in smart grid systems while

dealing with generation, transmission, distribution and consumer-end integration of all the sys-

tems in a common platform. Such a huge complexity is a major challenge in smart grid systems.

Efficiency: The smart grid system is supposed to be fully automated to getdesired efficiency

of the system. Infrastructure of smart grid is strengthenedby the communication protocols that

govern communications among various grid entities [8]. These protocols must be optimized

for various smart grid applications. Requirements to achieve high efficiency of the system are:

accurate time measurements, faster control messaging, integrated communication devices, en-

hanced computing and appropriate network topologies [128].

Consistency:A very common challenge faced in smart grid is consistency ofthe system. Most

of the smart grid applications depend heavily on the underlying communications infrastruc-

ture. A persistent communications system results into persistent grid actions. For instance,

the communications among consumer-end devices (EMS smart meter etc.) and utilities/energy

providers will result in consistent DR, quality monitoring, and disaster/outage management etc.

Moreover, a robust communications infrastructure is required to tackle disasters and power out-

76


ages etc. [8].

Security: One of the major challenges involved in smart grid is communications security. As

the smart grid is an interlinked network, any cyber securitybreach can potentially expose the

whole system. The system involves interconnection of various domains such as power plant,

consumer end, substation and information technology domains. Each domain individually as

well as whole system should be fully secured. Countless efforts are needed for provision of se-

cure interlinked smart grid communications network. Many of the researchers are in pursuance

to get enhanced security of smart grid as well as many organizations worldwide are working

on developing better solutions for smart grid security including IEEE, International Society

of Automation (ISA), North American Electrical Reliability Corporation-Critical Infrastructure

Protection and the US National Institute of Standards and Technology.

Standardization: Smart grid includes different standards like power generation and distribu-

tion, communications infrastructure, data management, power control and monitoring etc. IEEE

has defined the guidelines and standards for operation of smart grid using advanced technologies

in electrical engineering, ICT and power control. Communications standards are arguably most

important as most of the smart grid applications are communications dependent. Researchers

are working on communications standards related to power distribution systems DERs, trans-

mission substations, consumer requirements and network security. Work is also under way on

the operational necessities of interoperability, scalability and all other factors to set common

standards worldwide for smart grid. Other standardizationbodies include International Society

of Automation (ISA) and International Electro technical Commission (IEC). For instance, IEC

TC57 WG13 drafted the new international standards for enhanced grid reliability as well as for

cyber security [157].

Scalability: Smart grid communications infrastructure requires scalability of the system to ac-

commodate more and more devices in order to serve new end-users. A scalable communications

infrastructure for smart grid that uses one to many and many to many communication schemes

is presented in [158]. The authors evaluated the proposed schemes in terms of delays and

bandwidth usage. A scalable communications strategy has also been presented in [159] that

uses data-centric application platform for smart grid. Scalability is particularly important in the

AMI context as hundreds of thousands of meters need to communicate to energy provider and

the number of such meters keeps on increasing. Also with the passage of time, frequency of data

transmission between users and utilities will increase. Zhou et al [160] presented scalable com-

77


munications architectures for smart grid AMI and investigated the communications complexity

and scalability of the proposed system. The authors formulated optimization problems in order

to find minimum cost in terms of bandwidth and distance incorporating scalability of the system.

Interoperability: It refers to the ability of various systems or components to work/communicate

with each other in a smooth manner. Many organizations have worked on smart grid interop-

erability issues. For instance, the US National Institute of Standards and Technology (NIST)

has proposed a model of interoperability for smart grid in [161]. Different domains of smart

grid like generation, transmission, distribution, customers, operations, markets and Independent

System Operators (ISO) are needed to be interoperable and compatible from older to the newest

versions to ensure successful operations. This is very important aspect of the smart grid that

poses huge challenges [162, 163].

Self-Healing Actions: Handling a system during abnormal conditions or under faultis much

more complex as compared to the normal operation. That is whya key challenge for the im-

plementation of smart grid is encountered while defining thesystem under contingency. The

system needs to respond immediately to avoid system breakdown and must start self-healing ac-

tions within a small period of time after any contingency. Fast control signaling is a fundamental

requirement in such situations in order to control various actuators in the system. Continuous

monitoring through sensors and appropriate actions to prevent faulty conditions is an important

aspect of the smart grid. However, enabling system to self heal after occurrence of the fault

requires enormous automation, fast control and integration of the artificial intelligence at each

level. Isolation of the faulty part of the system to avoid spread of fault is fundamental require-

ment of the protection system. Reconfiguring the faulty partafter self healing in accordance to

the system conditions without manual interruption is a big challenge for smart grid [9].

Particular challenges regarding DSM include the lack of users’ awareness about the associated

advantages. Benefits and values offered by DSM are flexible and hence a better understanding

among users is required [8]. DSM based dynamic solutions areless competitive compared to

the traditional solutions because of the greater complexity and lack of standard business mod-

els. Existing market structure is not suitable for dynamic DSM solutions. DSM is supposed

to reduce the operational diversity of controlled appliances which causes a peak during load

recovery periods [41]. Development of appropriate techniques to reduce this peak is also an

important challenge for researchers. Electricity utilities and consumers may have conflicting

interests. For example, a utility wants to maximize its revenue while at the same time its con-

78


sumers want to minimize their bills. This is a particular challenge for DSM architectures and

programs to satisfy all the stakeholders [41]. Traditionalgame theory based approaches try to

meet this sort of challenges, but are time consuming and evenin some cases unsuccessful.

Incentive based DLC has its own set of challenges [41]. As utility controls the residential power

usage in a DLC scheme, the users may feel uncomfortable. Traditional DLC schemes may be

modified such that only selected appliances are controlled in such a way that users comfort

is maximized. The appliances controlled by DLC can also be divided into certain categories;

a certain load category may be activated in a certain scenario. This needs a comprehensive

policy making that would take into account interests of bothutilities and users. Real benefit

of DSM, however, comes from dynamic pricing based HEMS whichare dedicated devices for

controlling, scheduling and balancing the power consumed in a certain premises. There are

certain challenges regarding implementation of HEMS and their usefulness for DSM. One of

these issues is to introduce more efficient and comprehensive energy cost functions in order to

minimize energy consumption during peak hours as well as to minimize consumers’ bills and

maximize their comfort level [61]. Users can potentially fail DSM plans by using unfair means

like energy storage etc. This would result into unfair distribution of energy. Prevention of such

unfair distribution of energy among users during peak hoursis also an important challenge for

smart grid researchers and policy makers [55]. Various challenges and issues of smart grid with

important aspects are tabulated in Table 4.3.


Exploitation of enormous potential of smart grid, for wellbeing of mankind is dependent upon

the rapid development of advanced communications infrastructure and optimization of network

parameters. In this chapter, we have presented the smart grid applications, related WCTs for

HANs and NANs and dynamic pricing in detail in this chapter. We have compared WCTs for

HANs and NANs separately for them in context of consumer concerns and utility requirements.

Wi-Fi offers high data rate and larger range (100 m indoor) ascompared to Bluetooth short

range (10 m typical) secure communications, but ZigBee withlow cost, low power consump-

tion, reasonable indoor range (1075 m) and ability of accommodating a very large number of

79


Table 4.3: Smart Grid Challenges and Issues

Challenge/Issue Aspects

Complexity

Modeling, analysis and design of smart grid,

Interdependence between different infrastructures,

Distributed nature of monitoring and control functions

Efficiency

Optimization of network parameters,

Accurate time measurements and faster control messaging,

Integrated communications devices and enhanced computing,

Appropriate network topologies

Consistency

Consistent DR and quality monitoring,

Robust communications infrastructure required to tackle disasters and

power outages etc.

Security

Interconnection of various domains,

Ease of security breach in user premises,

Virus and hacking attacks

Standardization

Design, development and provision of common standards worldwide,

Uniformity among various standard organizations,

Regional and International concerns

ScalabilityAccommodation of more and more devices like smart meters,

Bandwidth adjustments according to additional users

Interoperability

Ability of various systems or components to work with each other in a

smooth manner,

Different domains of smart grid like generation, transmis-

sion,distribution, customers, operations, markets and independent sys-

tem operators are needed to be inter-operable and compatible from older

to the newest versions

Self Healing

To avoid system breakdown,

Must start self-healing actions within a small period of time after any

contingency,

80


nodes seems to be the best candidate for HANs. Low data rate ofZigBee (40250 kbps) is one

of the reasons it is a low power technology. Although ZigBee data rate is much less than Wi-Fi

(11300 Mbps), it is good enough for most of the usual HAN applications. Furthermore ZigBee

supports a high number of nodes (more than 64,000) which makes communications scalable

when new nodes enter the system. Z-Wave and 6LoWPAN are advantageous for IP enabled

low power devices. Bandwidth and the range of WiMAX make it appropriate for smart grid

NANs applications. WiMAX offers 72 Mbps speed which is 36 times greater than typical 3G

GSM speed. On the other hand, pre-existing cellular networks with sufcient bandwidth, high

data rates (more than 300 Mbps for 4G LTE), extensive coverage, lower maintenance costs and

strong security can save capital investments on utility dedicated communications infrastructure

and allow rapid deployment of applications. However, use ofpre-existing cellular networks will

raise the operational cost in terms of monthly recurring charges. A dedicated utility cellular net-

work can ensure the high quality and reliability with reduced operational cost. There exists a

clear trade-off when making a choice between dedicated WCTslike WiMAX and pre-existing

technologies like GSM cellular network. This choice essentially depends on a particular utility’s

budget and policies. This comprehensive review has tabulated different wireless communica-

tions in a systematic way along with the challenges for smartgrid implementation. Various

pricing schemes suitable with different scenarios should be used with the DSM applications.

The smart grid technologies highlighted here can be used to achieve the energy efficiency ob-

jectives. One of the key smart grid objectives is DSM throughoptimization based appliances

scheduling which is main topic of the following chapter.

81

Chapter 5DSM and Optimization based Appliances

Scheduling


DSM is supposed to have a vital role in future energy management systems. This

chapter presents detailed description and comparison of DSM techniques along with

optimization techniques required for smart home appliances scheduling. Dynamic

pricing based Energy Consumption Scheduling (ECS) schemes, featuring peak load reduction

and consumers’ energy cost minimization at residential level, are emphasized. Furthermore, the

chapter includes a description of dynamic pricing based home energy management and associ-

ated optimization techniques as well as comparative study of the latest schemes. Most of the

dynamic energy management solutions are based on the assumption of availability of advanced

information, communication and control infrastructure. However, realization of smart grid in

general and effective DSM in particular still faces many challenges.

The energy management task is naturally an optimization problem where energy consumption

and user comfort etc. are the objectives and energy availability and appliances’ specific require-

ments are usual constraints. The optimization related issues along with case studies of specific

optimization techniques are discussed in the following sections.

82

5.2. Optimization and Smart Appliances Scheduling

5.2 Optimization and Smart Appliances Scheduling

Optimization play important role in smart home scheduling in order to smoothen the load pro-

file and minimize the users’ cost. The energy management taskinvolves various objectives and

constraints. The researchers have highlighted many optimization techniques to solve such prob-

lems. A few relevant techniques have been discussed in the following sections followed by a

discussion on how these techniques can be effectively applied to energy management problems.

A detailed comparison of possible optimization techniquesis also presented at the end of this

chapter.

5.2.1 Knapsack Problem

Knapsack problem is a problem of combinatorial optimization which optimizes a given a set

of items each with a mass and value. It determines the number of each item to include in a

collection so that total weight is less than or equal to givenlimit and total value is large as

possible [164]. In other words knapsack problem is problem in which certain items of different

weights are given and we have to choose the one with weight notmore than some fixed weight

say W. There are two types of knapsack problems: 0-1 Knapsackproblem, fractional knapsack

problem.

In first case the item is either taken or not taken (accepted/rejected) there are no other possibil-

ities as in our case the appliance is OFF (0) or ON (1). In fractional knapsack case, items are

divisible and any fractional value of item can be considered. Mathematical formulation of this

problem is: Let there be n items, Zi to Zn where Zi has value Vi and weight Wi. Xi is number

of copies of item Zi which is mentioned above, must be zero or one. The maximum weight we

carry in bag is W. It is common to assume that all values must benon-negative. Mathematically,

maximizen

∑i=1

vixi (5.1)

Subject to

n

∑i=1

wixi ≤W (5.2)

Maximize the sum of values of items in the knapsack so that thesum of the weights must be

less than or equal to the knapsack capacity (W) [164]. There are many algorithms which can

83

5.2.2. PSO

be used to solve the 0-1 knapsack problem, for instance, brute force, dynamic programming,

memory functions, greedy algorithm, branch and bound, GA etc. [165]. Multiple knapsack

problem [166] may also be modeled by providing the 0-1 multiple knapsack problems with a

set of n items and set of m knapsacks(m≤ n), with Pj = profit of item j, Wi = weight of item

j and Ci =capacity of knapsack i. Select m disjoint subsets ofitems so that the total profit of

the selected items is a maximum and each subset can be assigned to a different knapsack whose

capacity is no less than the total weight of items in the subset. Formally [167]:

Maximizen

∑i=1

m

∑j=1

Pjxi j (5.3)

Subject to

n

∑j=1

wixi j ≤Ci ∈M = 1,m (5.4)

m

∑i=1

xi j ≤ 1, j ∈ N = 1,n (5.5)

xi j =

1, if item j is assigned to knapsack i

0,otherwise(5.6)

5.2.2 PSO

Optimization techniques are characterized by their ease ofuse, fast convergence properties and

their competency to solve non-linear and non-differentiable multi-optimization problems. A

large number of evolutionary techniques compete for the optimal solution to optimization prob-

lems. PSO is one of the robust techniques having a wide variety of applications [168–170].

PSO is inspired by bird flocking and was developed by Kennedy and Eberhart [171]. PSO

can be explained by considering a swarm of birds searching for food in a search space. During

search process, each bird has a position and velocity. All birds update their velocity and posi-

tion with the knowledge of their own position and the position of a bird nearest to food. PSO

technique makes the use of bird flocking scenario and considers each bird as a particle whose

position is a candidate solution in search space whereas thetotal number of particles designates

84

5.2.2. PSO

size of swarm (or population). In search space, each particle is characterized by its position,

velocity and fitness value.

5.2.2.1 PSO Algorithm

PSO is a computational methodology in which each particle isa solution belonging to a popula-

tion called swarm. All particles in the swarm move in the search space and each particle finds a

position according to its own experience and neighboring particles. The steps involved in PSO

algorithm are:

1. Starts with initial parameters like size of swarm, total number of iterations etc.

2. Particles are initialized randomly for their positions and velocities. The position and velocity

of ith particle in D dimensional space are represented by

Xi = [Xi1,Xi2,−−−−−−−−−−−−−−−−−−−−−XiD ] (5.7)

Vi = [Vi1,Vi2,−−−−−−−−−−−−−−−−−−−−ViD ] (5.8)

3. Fitness of objective function is evaluated. The objective function can be presented as:

Fi = F(Xi1,Xi2,−−−−−−−−−−−−−−−−−−−−−XiD ) (5.9)

4. Two best positions, known as global best position and personal best position, are obtained.

The personal best position is the best position obtained by aparticle itself up to current iteration

and is designated bypid . At first iteration,pid for each particle is same asXi, whereas for sub-

sequent iterations,pid of a particle is obtained by comparing its fitness value at current position

Xi with that of pid. If fitness value is better than that ofpid then set this value aspid ; whereas

pid remains unchanged otherwise. The global best position is the best position amongstpid and

is designated bypgbest.

5. Velocity of each particle is updated using velocity update equation [146].

Vt+1i = ωVt

i +c1∗ rand1(ptidi−Xt

i )+c2∗ rand2(ptgbest−Xt

i ) (5.10)

85

5.2.2. PSO

where,Vti is velocity ofith particle atith iteration,ω is inertia weight whilec1 andc2 are acceler-

ation coefficients.rand1 andrand2 are random numbers between (0,1).ptidi is the best position

of ith particle at iteration t.Xti is position ofith particle at iteration t.(pt

gbest is the global best

position at iteration t andVt+1i is updated velocity at iteration t + 1.

6. Position of each particle is updated using position update equation:

Xt+1i = Xt

i +Vt+1i (5.11)

7. Fitness of each particle is evaluated and global best and personal positions are obtained 8.

The process is repeated till convergence criterion is met. The pseudo code based on above

mentioned procedure for this algorithm is described here:

Start < PSO Algorithm>

Initialize parameters;

Initialize each particleXi ;

Initialize velocitiesVi;

Initialize best positionPid; (for each particle to current position)

Pgbest←minXiF(Xi);

While (iteration≤ iterationmax)

do

For each particleXi ;

For (d = 1 to D)

Initialize randomω in the interval (0, 1);

Accommodate velocity and position of a particle in d;

Vid ← ω∗ (vid +c1)∗ rand1∗ (Pid−xid)+c2∗ rand2∗ (Pgbest1dxid);

Velocity limit [vmin,vmax];

Update position;

xid ← xid +vid;

Particle, position limit[Xmin,Xmax];

End

Update best positionPi ;

if ( f (Xi)≤ f (Pi)) then

Pi ← Xi;

86

5.2.2. PSO

End

Update global positionPgbest;

If ( f (Pi)≤ f (Pgbest)) then

End

End

End

End

Pgbest← Pi

End

The above mentioned procedure is summarized in flow chart shown in Fig. 5.1.

PSO has wide range of applications. It has been utilized for speech reorganization in train-

ing Hidden Markov Models [172]. PSO has been used to find the optimal subset of image

and audio features, for speaker identification based on faceand voice reorganization in [173]

and [174], respectively. In [175], PSO algorithm has been employed for many image appli-

cations such as extracting useful objects from image, imagesegmentation etc. PSO has also

been successfully used in wireless networks for detecting the optimal path that an attacker may

use [176]. Furthermore, in the field of power engineering, PSO has wide variety of applica-

tions such as to solve load balancing, Economic Dispatch (ED), peak load management etc.

[168, 177].

In power systems, ED is one of optimization problems which has been solved frequently with

PSO for different situations. ED is the process of calculating the optimal output of all generat-

ing units, in order to meet the load demand at lowest cost, based on generation and transmission

constraints. Its objective is to minimize the generation cost. Considering the non-linearity of

ED problems, PSO algorithm provides the robust solution [178, 179]. In [180], the ED prob-

lem with Demand Participation (DP) representation is presented and solved by PSO. Equality

constraints create the major problem in PSO algorithm and are often handled by penalty func-

tion along with an enforcement mechanism. In [181], the dynamic economic load dispatch

problem is used to minimize the total operating cost by finding the optimal set of generators;

PSO algorithm is used to find the optimal schedule of generating units that can fulfil the load

demand at minimum cost while satisfying the constraints such as transmission losses, effect of

valve point and limits of ramp rate. The test system of five generators has been considered as

an example. The main objective of this problem is to find minimum cost to operate the generat-

87

5.2.2. PSO

Local best position = current position

Update velocities and positions of particles

Current position <

local best position

Local best position = current position

Current position

< global best

Global best position = current position

Stopping

Criteria met ?

start

STOP

Update velocities and positions of particles

Figure 5.1: PSO Flow Chart

88

5.2.3. GA

ing units while satisfying the load demand. For this purpose, the user load demand is properly

shared in all generating units.

5.2.3 GA

GA is an optimization approach, based on theoretical concept of natural evolutionary processes

such as mutation, inheritance, crossover and selection. GAcan easily handle non-linear prob-

lems. In GA, a population of chromosomes is initialized, andeach chromosome represents a

solution where the size of population depends on complexityof problem. The fitness value of

each individual of population is calculated by fitness function, relatively fit chromosomes are

selected to pass information to next generation and geneticprocedures such as mutation, selec-

tion and crossover are performed. Fitness of individuals increases as the number of generations

increases. This process continues until it converges to best set of chromosomes according to a

given criterion [182, 183]. The pseudo code for GA is:

Start < GA Algorithm>

Initialize population of individuals P (0);

While (generations≤max. generations)

do;

evaluation;

selection;

mutation;

generation ++;

display results;

End

Flow chart of GA is shown in Fig.5.2.

5.3 Comparative Study of Optimization based ECS Schemes

In smart grid, DSM at user premises is one of the most important issues related for enhance-

ment of grid efficiency as it results into peak load shaving and reduced probability of grid

89

5.3. Comparative Study of Optimization based ECS Schemes

Population initialization

Fitness Evaluation

Constraint satisfied

Survivor satisfied Varying individuals

output

YES

NO

NO

YES

Figure 5.2: GA Flow Chart

failure. Various optimization techniques have been used for peak load shaving and cost mini-

mization based upon the basic objective function given by equation 1.1. Major constraints to

this objective function are the available capacity and fulfillment of total demand as defined in

equation 1.2 [49]. There is an ongoing research in developing efficient HEMS architectures

based on various optimization techniques. Intelligent cloud based HEMS is developed in [50]

which allocates a dynamic priority to appliances. Priorityallocation depends on type and cur-

rent status of the appliance. The technique is successfullyapplied to a test bed and a reduction

of 7.3 %in average power consumption has been achieved. Neural fuzzy logic based controller

along with MATLAB interface has been introduced in [51] which senses the activities of the

occupants inside the home and facilitates them with required level of services. An IP, XML and

JAVA based network is presented in [52] which shows an experiment performed on 16 different

appliances in order to achieve effective scheduling.

A multi-scale stochastic optimization for HEMS is presented in [184] for Heating, Ventilation

and Air-Conditioning (HVAC) unit, charging of PHEV and scheduling of appliances whose op-

erations can be delayed up to a certain limit such as washing machine and dryer. Integration

90


of local renewable energy resources for appliance coordination is a cumbersome task and is

explained in [185]. Zhu et. al. used a layout of dynamic gamesfor DSM modelling which

consists of two layer optimization frameworks [14]. In [186], a heuristic optimization based

evolutionary method is used for handling large number of devices in dynamic DSM. Baig et al

suggested smart HEMS for smart grid using ZigBee sensors andan interface designed in LAB-

VIEW [120].

In smart grid, ECS on demand side has promising effects on peak load reduction and cost min-

imization. Total energy generation cost is reduced due to the minimal use of peaking plants.

Users are benefited by reduced energy bills due to schedulingof their peak loads to off-peak

hours. ECS is performed either in smart meters or dedicated HEMS. Using smart meters (con-

nected to home appliances and to utility through AMI) for ECSis economical as it avoids

necessity of a dedicated energy management device. Using dedicated HEMS on the other hand

is relatively costly but offers greater processing power, IP based solutions, and ease of opera-

tion, upgrading and maintenance.

Several research works are available in literature giving insight into ECS. For instance, authors

in [11] present a game theoretic approach to formulate an ECSscheme. In this scheme, the

users are the players while daily schedules of their home appliances are the strategies. Ad-

ditionally, an alternative pricing scheme is proposed. According to the scheme, incentives are

provided to all those users having ECS installed smart meters. The model presented in the paper

satisfies the billing mechanism criteria i.e. the total payments of registered users will be greater

than the total energy cost. An effective billing mechanism ensures the efficient utilization of

resources. Conventional billing mechanisms or pricing schemes include Inclining Block Rates

(IBR), TOU, CPP and RTP. These pricing schemes are also effective in terms of several energy

related objectives e.g., peak load reduction, load synchronization, etc. Bi-directional communi-

cation environment of smart grid provides more favorable environment for pricing.

It is difficult to optimize all the aspects in one scheme, so various schemes show optimality with

one aspect by trading off the other ones. Detailed comparison of ECS schemes is presented in

Table 4.2. Important aspects and conclusions of the Table 4.2 are summarized in the following

five points.

1. An important aspect that should be considered for effective DSM in smart grid is the fairness

among users. This issue arises from the fact that a single user or a group of users can possibly be

untruthful and may get extra benefits by deviating from the energy consumption schedule given

91


by the utility. Users can be enforced on truthfulness by installing an energy supervision and

control system to make sure that users do not violate the allowed limits of energy consumption

scheduling mechanism.

To ensure the user truthfulness, algorithms for energy scheduling controller must be designed so

as to tackle deviation from allowable limits of energy consumption. The algorithms presented

in [40, 53, 55, 187] are based on self-monitoring. In [40], itis worth noting that the given

algorithm ensures the self-monitoring as the presented billing mechanism directly relates each

users payoff to the energy cost. In this model, the individual and global cost minimizations are

closely related to each other. Whenever a user tries to get extra benefit by deviating from their

actual consumption schedule, the adopted algorithm will harm the cheating user in terms of

their individual payments. To tackle this problem more comprehensively, an alternative fairness

billing mechanism is presented in [55]. A fairness index hasbeen introduced corresponding to

the proposed billing mechanism. The fairness index can be defined as ”the vibrational distance

between normalized billing vector for billing mechanism and normalized billing vector”. This

shows that a lower fairness index points out a fairer billing. The simulation results of the pro-

posed algorithm show significant improvement in terms of fairness. It is inferred from Table

4.2 that scheme discussed in [55] is the best in assuring the fairness.

2. Billing system is an important aspect for implementationof DSM measures. In the recent

development of DSM realization, different billing mechanisms have been introduced for effec-

tive results in terms of various performance parameters. For example, RTP shows significant

impacts on PAR reduction and energy cost minimization [26, 66] and RTP with alternative

hour-by-hour billing mechanism shows good results in termsof improving Fairness (73%), re-

garding the users truthfulness and also operates near to optimality in terms of cost minimization

[53]. Among all these billing mechanisms summarized in Table 4.2, the less efficient systems

in terms of improving the total energy cost minimization as well as reduction in PAR are those

in which RTP is used along with IBR and those based on convex and increasing cost functions

[11, 13, 40, 53, 187].

3. In the existing grid infrastructure, prices are declaredaccording to the average load curve. In

certain peak hours, utilities run peaking plants to fill the gap between the maximum demand and

the total generation and hence the cost of energy rises. Therefore, PAR is taken as an important

parameter for comparison of load optimization techniques in Table 4.2. Various attempts to

reduce the total energy cost by minimizing the PAR have been made in [11, 13, 40, 53, 187],

92


results show significant PAR reduction and total energy costminimization.

4. Execution time of the algorithm is one of the important parameters needed to be optimized

for realization of modern ECS. The less execution time of thealgorithm shows good results

in terms of appliance waiting time. This problem is mainly highlighted in [66] in which the

execution time is based on both the preemptive and non-preemptive tasks as well as the number

of time slots. The execution time is reduced up to 2%, while getting the benefits of constrained

processing mechanism, a mechanism having search branches shortened when the partial peak

value exceeds the current best. The comparison between constrained and non-constrained be-

havior reveals that the constrained search has a less execution time and is more stable. The

search space size of the preemptive tasks has an adverse effect on the execution time. By re-

moving the useless search space traversal, the constrainedsearch can be performed in reduced

computation time. Table 5.1 gives relative comparison of execution time for various schemes.

5. The waiting time of appliances imposed by scheduler and the user’s cost minimization are

two opposite functions. For example, considering RTP, if someone needs to wash their dishes

during peak hours, then there will be two possible solutions, i.e., to wash their dishes during

peak hours at increased rates or to postpone their device operation till off-peak hours. To tackle

these conflicts, a trade-off can be made between these two objectives during mathematical mod-

eling [11, 187].


In this chapter, we have presented DSM techniques in smart grid including load shedding, in-

centive based DLC and dynamic pricing based ECS. Optimization technique including Knap-

sack, PSO and GA have been described in context of smart appliances scheduling. Compar-

ison of different dynamic pricing based ECS is presented incorporating various factors such

as billing mechanism, fairness among users, algorithm processing times etc. Ten recent and

prominent schemes of ECS have been compared analytically. The highest Fairness (73 %) is

ensured in [55] and the highest PAR reduction among the compared schemes is 38.1 %, pre-

sented in [11]. Various DLC schemes have been employed to control energy consumption.

These schemes are more useful for heavy loads with greater peak load reduction potential. ECS

provides more effective ways of DSM, especially for residential load, by employing efficient op-

timization techniques and ensuring the users’ privacy and comfort. Efficient DSM schemes are

93


Table 5.1: Comparison of Different Schemes for HEMS

EC

S

Schem

e

Optim

ization

Technique

Pricing

Schem

e

Users

Category

Fairness

Gain

Appli-

cation

Coverage

Area

Sim

ulation

Setup

PA

R

Reduction

Energy

Cost

Mini-

mization

Algorithm

Execution

Tim

e

Waiting

Tim

eof

Appliances

Utility

End-U

ser

Incentive

Based

ECS

Distributed

Algorithm

Based

Optimization

Convex and

Increasing

Cost

Functions

Residential NA 38.1% 37.8% Less More

Yes

Yes

Local

MAT

LAB

Game

Theoretic

Based ECS

Distributed

Algorithm

Based

Optimization

Convex and

Increasing

Cost

Functions

Residential Applicable 17% 19.6% Less More

Yes

Yes

Glo

bal

MAT

LAB

Heuristic

Optimiza-

tion

Based ECS

Heuristic

Based

Evolutionary

Algorithm

Day-Ahead

Load Shifting

Residential

Commercial

Industrial

NA

18.3%

18.3%

14.2%

5.0%

5.8%

10.0%

Relatively

More

Inversely

Related

to Delay

Yes

Yes

Glo

bal

MAT

LAB

,

RealG

rid

Netw

ork

Back-

tracking

Based ECS

Backtracking

Based ECS

RTP Residential NA 23.1% Not Men-

tioned

Less More

Yes

Yes

Local

Visu

alC+

+6

.0,

MS

Win

dow

GetT

rickC

ou

nt

Vickrey-

Clarke-

Groove

(VCG)

Based ECS

VCG Convex,

Differentiable

& Increasing

Function

Residential Applicable 19.3% 37.8% More More

Yes

Yes

Glo

bal

MAT

LAB

ORLC With

Price

Prediction

Weighted Aver-

age Filter Based

Price

Prediction

RTP

+

IBR

Residential NA 38% 25% NA Inversely

Related to

A.C.P &

Payments

Yes

Yes

Local

MAT

LAB

A Layered

Architecture

for DSM

Spring

Algorithm

ToU Residential NAAppli-

cable7.53% NA More

Yes

Yes

Local

MAT

LAB

,

Sim

ulin

k,

Stateflow

Load

Uncertainty

for ECS

Optimization

Based Schedul-

ing

RTP

+

IBR

Residential Applicable 25.6%-

28.9%

15.8%-

17.6%

NA More

Yes

Yes

Glo

bal

MO

SE

K

Optimality

and Fairness

Based ECS

Model

Game

Theory Based

Optimization

RTP

with Hour-

by-Hour

Billing Mecha-

nism

Residential 73% More

Efficient &

Inversely

Related to

Optimality

Not Men-

tioned

Not Men-

tioned

NA Not

Mentioned

Yes

Yes

Glo

bal

MAT

LAB

Game Theo-

retic

Centralized

Optimiza-

tion

Scheme

GTES

Algorithm

Dynamic pric-

ing

based on

Logarithmic

Function

Residential NA 18.6% and

directly

prop. to

no. of

users

7.5% and

directly

prop. to

no. of

users

More Not

Mentioned

Yes

Yes

Local

MAT

LAB

94


essential to control energy consumption. Various technologies are required in order to achieve

the benefits of DSM in smart grid, including ICTs and advancedcontrol mechanism. Find-

ing suitable communication and control infrastructure, making DSM policies and optimizing

energy consumption are the ongoing research areas related to efficient DSM in smart grid.

95

Chapter 6Enhanced System Architecture for

Optimized DSM in Smart Grid


DSM through optimization of home energy consumption in smartgrid environment is

now one of the well-known research areas. Scheduling has been done through many

different algorithms to reduce peak load and consequently the PAR. This chapter

presents CHEMA with integration of multiple appliances scheduling options and flexible load

categorization in smart grid environment. Multiple layersof enhanced architecture are mod-

eled in Simulink with embedded MATLAB code. Single Knapsackoptimization technique is

used for scheduling and four different cases for cost reduction are modeled. Fault identification

and electricity theft control is also added along with the carbon foot prints reduction for envi-

ronmental concerns. Simulation results prove the effectiveness of the proposed model. This

chapter is dedicated for description of our proposed energymanagement model for peak, PAR

reduction and overall energy cost minimization. Chapter covers the introduction and related

work along with the details of the proposed model. CHEMA, consisting of six layers has been

proposed and four out of six layers have been implemented in Simulink. Four different cases of

cost minimization are presented at second layer. Details are elaborated in subsequent sections.

96

6.2. Proposed CHEMA and its Implementation

6.2 Proposed CHEMA and its Implementation

DR and DSM based on bi-directional communication are enabled by smart grid in order to

smoothen the load curve of the traditional grid. DSM programs influence the behaviour of

the consumers regarding electricity consumption [9]. Although energy management/efficiency

can be achieved in several different ways, appliance scheduling is one promising method for

this purpose. A novel energy management mechanism, CHEMA, is presented in this chapter.

CHEMA has six layers out of which four layers have been implemented in Simulink. Four

different cases of cost minimization are integrated at second layer. Inclusion of partial base line

load into the cost minimization problem is one of the unique features of CHEMA. Other contri-

butions include six layered architecture, multiple cost minimization cases integrated at second

layer, faults and theft detection etc. The results and benefits of proposed scheme are discussed

at the end of chapter.

This section elaborates the details of the proposed architecture and its implementation. House-

hold appliances could be characterized on the bases of energy or power consumption. Major

portion of the house-hold appliances, almost 63%, consistsof energy based loads and is very

useful for DSM purposes [188]. These loads include water heaters, space heaters, air condition-

ers and refrigerators. We have added these basic loads in ourenergy management architecture

with least impact on user comfort. Our proposed model consists of six layers: Appliance In-

terface (AI), Optimized Energy Cost (OEC), Theft and FaultsDetection (TFD), Green Effects

(GEs), Demand Prediction (DP), and Dynamic Pricing or SmartGrid Interface (SGI). Proposed

architecture is depicted in Fig. 6.1.

In CHEMA, all the appliances are interfaced with first layer and basic parameters of all the

appliances are loaded into MATLAB workspace. In case of missing parameters while loading,

the model generates the error message and does not proceed tothe next layer. Second layer is

dedicated for energy cost optimization. Four energy management and cost minimization cases

are implemented at this layer: unscheduled energy cost model, unscheduled with Person Pres-

ence Controller (PPC), scheduled with PPC and scheduled with flexibility. Single Knapsack

optimization technique is used for scheduling options. Third layer is used for theft and faults

detection purposes and is based on malicious node detectionstrategy. Whenever fault (e.g.

phase to ground fault) occurs, the model detects it and a beepis generated for the user’s im-

mediate attention. Fourth layer is used for GEs. Remaining two layers are used for demand

97


Figure 6.1: Proposed CHEMA

prediction or load forecasting effects and dynamic pricing, respectively. We have used ToU

pricing scheme in this model. However various dynamic pricing models may be implemented

at this layer. Complete flow chart of enhanced energy management architecture is depicted in

Fig. 6.2. Four out of six proposed layers are implemented as dynamic pricing and demand

prediction have not been implemented in this work.

After the basic parameters initialization at AI, CHEMA algorithm checks for value of ‘S’

which stands for schedulability, if it is 1, algorithm goes to scheduling options, otherwise it

goes to unscheduled options. Furthermore, it checks for twoparameters: ‘P’ and ‘F’ which

stand for PPC and flexibility, respectively. If ‘P’=1, CHEMAactivates the unscheduled with

PPC energy cost minimization option, otherwise it is confined to unscheduled option. If ‘F’ is

one, the flexibility option is actuated. Whichever option isactuated the power consumption,

temperature variations and total cost results are displayed. Ii represents the appliances current

98


Start

Initialize energy

management

architecture

Check Readiness

of all the

appliances

and modules

Actuate energy

optimization layerIf S=1

No

Yes

Yes

If F=1If P=1

No

Calculate carbon

foot prints

Publish overall

resultsEnd

Initilize

Unscheduled

without PPC

module

Initilize

Unscheduled with

PPC module

Initilize scheduled

with PPC module

Initialize scheduled

with flexibility

module (UCI)

Publish the cost, load

and temperature resultsIf Ii ≠ Is

Declare fault and

generate beep

Initilize green

effects layer

Yes

No

Yes

No

Yes

Figure 6.2: CHEMA Flow Chart

99

6.2.1. Load Categorization

andIs represent the source current. Now, the running or operational appliances are interfaced

simultaneously with TFD and GEs layers. Faults and theft aredetected at third layer and user

is informed about abnormal conditions by beep generation. Carbon foot prints are calculated

at fourth layer in order to make the user more conscious aboutenvironmental concerns. Run-

time observation of cost and carbon foot prints surely affects the users’ energy consumption

behaviour. Various aspects and modules of CHEMA are explained in subsequent sections.

6.2.1 Load Categorization

Home appliances are categorized with respect to different aspects in literature [189]; however,

there are three major categories of home appliances: essential load or baseline load, regular

or continuously running load and burst or schedulable loads. We have divided the load into

following categories and subcategories. This model deals with the basic categories.

First category is of essential load which includes: lighting, fans and communication equip-

ment only. Whereas, lighting can be divided into sub categories of off-peak lighting in such a

way that peak lighting is one half of the off-peak lighting.

χ =

12 χOP f or Peak Hours

χOP Otherwise(6.1)

whereχ shows the total lighting load andχOP denotes off-peak lighting load. This load is

partially added in the scheduling problem. Inclusion of essential loads on half bases as indi-

cated in equation 6.1, such as lighting load, is the enhancement in the optimization problem

presented in [26]. It will minimize the cost more effectively with the control of the lighting in

peak hours. Therefore base line load is partially added intoscheduling problem. In addition,

in off-peak hours, the lighting is controlled on the bases ofno. of persons present. If set of

time slots is presented byT = 1,2,3, .............T and T is the total no. of time slots and set of

persons is presented asP= p1, p2, p3, ............P , and P shows the maximum no. of persons

in the house, then

100

6.2.2. Energy Cost Optimization Model

P=

p1 f or ∀ t1≤ t ≤ t2

p2 f or ∀ t2≤ t ≤ t3

p3 f or ∀ t3≤ t ≤ t4

(6.2)

where, t1, t2, ..... belongs to T. Second major category is ofregular or continuous running loads

which include: refrigerator, central heating/cooling, water heater etc. Third major category is of

schedulable loads which include: water pump, electric iron, washing machine and dishwasher

etc. We are introducing flexible load categorization which is explained with the help of follow-

ing scenarios.

• HVAC temperatures are given a tolerance band yet maintaining reasonable comfort of the

user.

• Hot water consumption is controlled with PPC i.e. on the bases of personnel presence in

the house.

6.2.2 Energy Cost Optimization Model

In the light of above discussion, the scheduling problem hasbeen formulated for a set of ap-

pliancesAPP= 1,2,3, ............N in a horizon of 24 equal time slots which are represented

by a of time slotsT = 1,2,3, ., ., ., .,TN. Appliances power rating and per unit energy cost

are used for calculating the cost for each category of the load. If Ai j , Ri j andBi j represent the

burst, regular and partial baseline loads’ costs andai j , r i j andbi j represent the on and off states

of corresponding category with 1 and 0 values, respectively. The optimization problem will be

as follows:

Minimize∑i, j

Ai, j ∗ai j +∑i, j

Ri, j ∗ r i j +∑i, j

χi, j ∗bi j (6.3)

subject to:

∑i

CA+∑i

CR+∑i

Cχ ≤Cj ,∀ j ∈ T, (6.2a)

ai j , r i j ,bi j = [0,1], (6.2b)

101

6.2.3. Pricing Scheme

Figure 6.3: ToU Pricing Scheme

where,CA, CR andCχ denote the capacity of burst loads, regular loads and partial base line

loads, respectively. Time slot is denoted by j and T denotes the total time slots. Total available

capacity in jth slot is denoted byCj . Binary optimization problem described in equations 6.2

and 6.2 (a, b) is a linear problem and is solved for different scheduling options using single

Knapsack optimization technique. Implementation of various Simulink modules is described in

subsequent sections.

6.2.3 Pricing Scheme

There are various ways of pricing strategies in dynamic energy management modeling like:

ToU, RTP, CPP, etc. In our model, we have adopted ToU pricing scheme and cost is taken as

PKR 10.0/kWh for off-peak hours, PKR 12.0/kWh for mid peak and PKR 15.0/kWh for peak

hours. The Scheme is implemented in Simulink with clock and embedded code as shown in Fig.

6.3. Rates of this module are called at various places to compute the energy cost of different

appliances.

6.2.4 Space Heating Module

Space heating module calculates the heating cost on the bases of modeled environment, thermal

parameters of premises and heating system. Initial room temperature is taken as 20C or 68

degree Fahrenheit. The model is based on the following differential equations [190, 191]:

102

6.2.4. Space Heating Module

Figure 6.4: Thermostat Subsystem

Figure 6.5: Heater Subsystem

dQdt

= (Th−Tr)∗mdot∗c, (6.4)

where,dQdt represents the flow of heat from heater into the room,Th is heater temperature,Tr is

room temperature,mdot is the mass of air, and c is the specific heat capacity of the air. Room

temperature variations are calculated on the basis of the net heat flow considering the losses of

the heat into environment as:

(dQdt

)losses=Tr −Tout

Req, (6.5)

dTr

dt=

1mair ∗c

(dQh

dt−

dQlosses

dt), (6.6)

where,Req represents the equivalent thermal resistance of the room. “Thermostat” block is

used to switch on and off the heater on the variations of five degree farenheit. Thermostat block

is shown in Fig. 6.4 in whichTdi f f shows the temperature difference between current room

temperature and set point. Heater subsystem is shown in the Fig 6.5.

103

6.2.5. Water Heater and Refrigerator Modules

6.2.5 Water Heater and Refrigerator Modules

Water heater is another important load, with thermal storage capability, used in this architecture

for DSM purposes. Water heaters are considered the second largest load among house hold

appliances and consume about 30% of the total energy consumed [192]. Electrical energy used

in water heater consists of two parts: one is used for heatingof the cold water and other is used

for compensation of losses from tank to ambient. There are two ways of heat loss in the water

heater: one is the conduction losses through walls of the heater and other is temperature drop by

water drawn for usage and entry of colder water for compensation. Energy flow analysis of the

water heater provides the following differential equationto determine the inside temperature of

water. On the bases of in and out heat flow the water heater model is given by the following

equation [192].

TH(t) = TH(τ)e−(1

R′C)(t−τ)+GR′Tout+BR′T in+QR′∗

[1−e−(1

R′C)(t−τ)] (6.7)

whereτ: initial time (hr)

TH(τ): initial temperature (F)

Tin : incoming water temperature(F)

Tout : ambient air temperature outside tank(F)

TH(t) : temperature of water in tank at time t (F)

Q : energy input rate as function of time(W)

R : tank thermal resistance(m2.F/W)

SA : surface area of tank(m2)

G= SA/R (W/F)

WD : water demand (L/hr)

Cp : specific heat of water (W/(F.kg))

D : density of water = 1 kg/L

B : D.WD.Cp (W/ F)

C : (volume of tank).(density of water).Cp (W/ F)

R′ = 1/(B+G)(W/F)

104

6.2.5. Water Heater and Refrigerator Modules

Figure 6.6: Water Heater Complete Model

Figure 6.7: Water Heater Subsystem A

Water temperature inside the heater is represented at the left hand side of the equation and

ambient losses, energy required for water heating and totalinput are taken at the right hand side

of the equation. Various modules of above equation are implemented as subsystems such as

water demand term implementation is shown in Fig. 6.6 in which input port 2 is showing the

water consumption which is based on Electric Power ResearchInstitute (EPRI) forecast model

[193]. The overall implementation of the equation is given in Fig. 6.7 in which subsystem A is

used to implement equation 6.7 and subsystem B is reserved for the thermostat based decision

about heater on/off by comparison of set point and inside water temperature. Subsystem C

converts the joules/hr to watts and consequently the cost iscalculated. The values for cost,

temperature and power consumed by water heater are stored and displayed.

Refrigerator modeling is based on equation 6.14 as follows.The refrigerator system con-

105

6.2.6. TFD

Figure 6.8: Refrigerator Complete Model in Simulink

sists of three subsystems. First subsystem or module is usedto implement the wall temperature

and second is used to implement the indoor temperature whilethird compares the indoor tem-

perature and set point to turn on or off the refrigerator. Thesame module is used to convert the

joules into watts and to perform subsequent cost calculations according to given rates. Refrig-

erator model is shown in Fig. 6.8 and wall temperature subsystem is shown in Fig. 6.9. The

refrigerator wall temperature is given as:

Tw =∫(To∗2.9863)dt+

∫(Tr ∗2.9863)dt

−∫(Tw∗5.9726)dt, (6.8)

where,To, Tr andTw represent the outdoor temperature, the room temperature and wall temper-

ature, respectively. This equation gives the approximate temperature of refrigerator wall.

6.2.6 TFD

Electricity theft and faults identification are important concerns of users which are addressed in

this architecture. There are many ways of theft detection and minimization [105]. One simple

approach is, whenever a line to ground fault occurs or some theft attempt is made, there is a

change of total current of house loads. This fact is used in the development of faults and theft

106

6.2.6. TFD

Figure 6.9: Refrigerator Wall Temperature Subsystem

Figure 6.10: Fault Occurrence and Detection

identification module. IfIs is the total current required by all the on appliances andI1, I2,.....In

represent the ’n’ appliances current then,

n

∑i=1

Ii =

Is, Normal Operation

otherwise, Declare Faulty Condition,(6.9)

where, Ii represents theith appliance current andIs shows the source current. Figures 6.10

and 6.11 show the output of TFD. It can be seen in Fig. 6.10 thatfault is detected at the time

of occurrence. In addition to the visual display, a beep is also generated at the time of fault

occurrence for the user attention. Figure 6.11 shows the voltage current characteristics at TFD.

107

6.2.7. GEs

Figure 6.11: Current and Voltage Characteristics at TFD

6.2.7 GEs

Another layer of environmental concerns or green effects has been introduced in order to relate

the residential energy consumption to the Green House Gases(GHGs) emission or carbon foot-

print. Carbon footprints are defined as “the total sets of greenhouse gas emission caused by an

organization, event, product or person” [194]. GHGs in the earths atmosphere are water va-

por, carbon dioxide, methane, nitrous oxide and ozone. These gases absorb and emit radiation

within thermal infrared range and cause the greenhouse effect.

In order to make the calculations simple, Wright, Kemp, and Williams have suggested the

definition as: “the measure of total amount ofCO2 and MethaneCO2 emission of a defined

population, system or activity, considering all relevant sources, sinks storage within the spatial

and temporal boundaries of population, system or activity of interest” [194]. The reduction

of carbon footprints by developing alternative projects such as solar, wind energy is called car-

bon offsetting [195]. GHGs can be measured by monitoring theemission continuously or by

estimating the data related to emission i.e. (amount of fuel) and applying relevant conversion

factor. Some of the conversion factors are calorific values,emission factors, and oxidation fac-

tors. These conversion factors enable us to convert activity data into fuel liters, miles driven,

tons of waste into kilograms of carbon dioxide equivalent (CO2e). CO2e is a universal unit for

108

6.3. Results and Discussion

measuring the global warming potential. There are two categories of GHGs emission. First

is direct GHGs emission which is the emission at the point of use of a fuel or energy carrier

(or in the case of electricity, prior to the point of generation). In other words, the emission

from sources which are owned or controlled by reporting entity. Second type is indirect GHGs

emission which is the emission emitted by the use of a fuel or energy carrier i.e. as a result of

extracting and transforming energy sources. Indirect emission i.e. purchased electricity, extrac-

tion and production of purchased fuels and transport related activities in vehicles that are not

controlled by reporting entity. There are various parameters included in the GHGs conversion

factors for electricity consumption [196]. Calculations of carbon emissions are performed as

follows:

If electricity consumption = 20,000 kWh,

Emission factor [22] = 0.000689551 metric tonsCO2/kWh, then

Amount o f carbon dioxide= EC∗EF, (6.10)

where, EC shows the electricity consumed in kWh and EF represents the emission factor in

Kg/kWh

Amount of carbon dioxide = 13791 Kg. (for above values of EC and EF)

These results are compared with emission of carbon from different sources in Table 6.1 [196].

Relative reduction of carbon emission with different casesof our proposed architecture is pro-

vided in the results section.

6.3 Results and Discussion

In CHEMA, we have implemented four layers out of six proposedlayers. We have used three

regular loads consisting of room space heater, refrigerator and electric water heater; six burst

loads of different ratings: two of 200 W, and one of 265, 300, 400 and 1000 W each. Five lights

of 20 W each are used in lighting modules. PPC senses the presence of the persons and controls

the heating and lighting of the home accordingly. ToU pricing scheme is used with the 24 timing

slots of one hour each. Simulation were performed on intel core i3, 2.4 GHZ machine with 6

GB RAM. Results for four major categories of energy management architecture are elaborated

in subsequent subsections.

109

6.3.1. Case 1

Table 6.1: GHGs Emission and Electricity Consumption Equivalency

Annual GHGs Emission From CO2 Emission From Carbon Seized By

2.9 passenger vehicles1552 gallons of gasoline con-

sumed

354 tree seedlings grown for

10 years

32,836 miles/year driven by

an average passenger vehicle

14, 813 pounds of coal

burned

11.3 acres of US forests in

one year

4.9 tons of waste sent to the

land fill

0.183 tanker trucks worth of

gasoline

0.106 acres of US forests

preserved from conversion to

cropland in one year

0.707 Garbage trucks of

waste recycled instead of

landfilled

32.1 barrels of oil consumed NA

6.3.1 Case 1

In case 1, baseline, regular and burst loads for proposed system are run without any control or

scheduling mechanism in order to find the total load for a day.Fig. 6.12 shows the total cost of

unscheduled load and pricing scheme used in which per unit prices of PKR 10/kWh (off-peak),

PKR 12/kWh (mid. peak) and PKR 15/kWh (peak) are used. Figures 6.13 to 6.15 show cost,

power and temperature variation of HVAC, refrigerator, water heater and lighting respectively.

Total cost in this case is PKR 1692.3.

6.3.2 Case 2

In case 2, we implemented a strategy to control the devices whose consumption is dependent on

the presence of persons in home. In this case PPC senses the presence of person at home and

then controls home appliances accordingly. Lighting and hot water consumption are controlled

with the help of PPC. Number of persons vary from 3 to 5 in different time slots. This technique

is useful to minimize the cost and to enhance the systems reliability. Figures 6.16 to 6.19 show

the PPC, pricing scheme, HVAC cost and lighting cost, respectively. It is evident that usage of

PPC has reduced HVAC cost to PKR 1301.1. This provides a 22.9%savings in total cost.

110

6.3.2. Case 2

Figure 6.12: Pricing Scheme and Total Unscheduled Cost (Case 1)

Figure 6.13: HVAC Cost and Temperature (Case 1)

111

6.3.2. Case 2

Figure 6.14: Refrigeration Cost, Power and Temperature Variation (Case1)

Figure 6.15: Water Heater Cost, Power and Temperature Variation (Case 1)

112

6.3.2. Case 2

Figure 6.16: Number of Lights and Cost Variation (Case 1)

Figure 6.17: PPC

113

6.3.2. Case 2

Figure 6.18: Pricing Scheme and Total Cost (Case 2)

Figure 6.19: HVAC Cost and Temperature Variation (Case 2)

114

6.3.3. Case 3

Figure 6.20: Lighting Cost (Case 2)

6.3.3 Case 3

In case 3, we run PPC and single Knapsack scheduling algorithm together to further reduce the

cost of consumption for proposed model. Figure 6.20 shows the total consumption cost for case

3 which has been reduced to PKR 1283.3.

6.3.4 Case 4

The focus of case 4 is to develop a system which schedules the user loads in context of user

comfort. Burst loads can only be operated at scheduled time otherwise scheduler becomes

ineffective and the net load consumption cost will not be reduced as expected by the use of

scheduler. The concept of User Comfort Index (UCI) is introduced to measure and ensure the

user comfort. In UCI, user has been provided with the flexibility to anytime switch on the

desired burst load at his requirement, however, UCI asks user to provide some input about

comfort level. The example of UCI is given below

• Season = winter

• Space Heating = HVAC

115

6.3.4. Case 4

Figure 6.21: Total Cost (Case 3)

• Comfort level is given as:

UCI =

Tmin = 23C and Tmax= 27C f or HVAC

Tmin = 5C and Tmax= 10C f or Re f rigrator

Tmin = 45C and Tmax= 55C f or Water Heater,

(6.11)

by default the settings for appliances are as: HVAC 27C, refrigerator 5C and water heater

is 55C. When scheduler is used to switch on burst load, no change inthese thresholds is

required. When using the option of flexible load categorization, it is needed to provide the

required changings in threshold levels for ensuring the operation of burst load which now has

changed its category and has become essential load. There are some fixed slots in which the

thresholds are relaxed automatically in order to implementthe concept of flexible load catego-

rization defined in section 3. Now these appliances go to standby mode and new settings are

updated for them which are as: HVAC 23C, refrigerator 10C and water heater 45C. These

appliances will now be on only if the desired comfort is breached otherwise they will remain

in standby mode till the burst load is switched off by the user. Figure 6.21 shows the usage of

UCI.

116

6.3.4. Case 4

Figure 6.22: UCI Results (Case 4)

When the burst load is activated under flexible load category, the threshold of every appli-

ance changes and it goes to standby mode. In Fig. 6.22 above, ‘1’ stands for active and ‘0’

stands for standby mode. The burst load is switched on between 8:00 to 12:00 am. As specified

in Fig. 6.23, when burst load is activated the threshold changes. It shows that before 08:00 am

the threshold of refrigerator was 5C but after the activation of burst load threshold changes.

Switching the device to standby mode will provide user an advantage as in standby mode. These

appliances consume power almost equal to zero watts as shownin Fig. 6.24, power consumed

by refrigerator during this period is almost negligible. Same happens with water heater and

HVAC.

Table 6.2 shows total cost for load consumed for one day whilecarbon foot print results are

summarized in Table 6.3.

Table 6.2: Cost Comparison of Different Cases

NO Case Total Cost (KWh/day)(Rs)

1 Unscheduled without PPC1692.3

2 Unscheduled with PPC 1301.1

3 Scheduled with PPC 1286.3

4 Load Flexibility (UCI) 1257.1

117

6.3.5. Aggregated and Total Energy Consumption Results

Figure 6.23: Refrigeration Cost (Case 4)

Table 6.3: Carbon Emission Reduction of Different Cases

NO Case Total Load (KWh/day) CO2 Emission (Kg/day)

1 Unscheduled without PPC188.4860 126.2856

2 Unscheduled with PPC 145.1734 97.0429

3 Scheduled with PPC 144.8401 97.0429

4 Load Flexibility (UCI) 140.2925 93.9960

6.3.5 Aggregated and Total Energy Consumption Results

This section is dedicated for discussion of aggregated energy consumption of five major mod-

ules used in four different cases of CHEMA. The section also includes a comparison of total

energy consumption of the four cases.

As discussed earlier, five major modules include: space heating, refrigerator, water heater, light-

ing and burst load/scheduling module. Case one shows the operation of these modules without

any scheduling or controlling technique except the usage ofthermostats with regular loads with

predefined temperature settings. Aggregated energy consumption of five major modules for

case one is shown in Fig. 6.25. In figure 6.25 and similar figures of following cases, modules

1 to 5 correspond to HVAC, refrigerator, water heater, lighting and burst load/scheduling mod-

118


1 2 3 4 50

20

40

60

80

100

120

140Total Energy Consumed by Each Module (Case 1)

En

erg

y C

on

su

me

d (

kW

h)

Modules (No.)

Figure 6.24: Aggregated Energy Consumption (Case 1)

ules, respectively. It is clearly seen that the highest consumption of energy is made by space

heating module (121.76 kWh) and because of continuous usageand temperature limits. The

minimum energy consumption corresponds to the lighting load which is 1.92 kWh while total

energy consumption for case one is calculated as 188.486 kWh.

The second case of the CHEMA is unscheduled load with PPC controller, the strategy of

PPC consists of two parts: one is the control of lighting loadaccording to the presence of

persons in the home and second is the control of regular loadsaccording to the forecasted

consumption of water demand. The forecasting is based on theEPRI model. The basic purpose

of the EPRI model is to control the consumption of energy according to the demand and save

the wastage the energy on the bases of forecasted demand. Loads with thermal storage capacity

are ideal for DSM and application of EPRI model on regular load is helpful in saving total

energy consumption and ultimately a reasonable reduction in energy cost. EPRI model has been

designed on the bases of data collected by eleven different companies. Parameters recorded

during collection of data include latitude, air temperature, water temperature etc. The model is

based on the strong statistical analysis and it considers various categories of users along with

the demographic and climatic factors. Initially 16 variables were considered for the model but

finally the interdependent were dropped off and only seven independent variables were included

119


1 2 3 4 50

20

40

60

80

100

120Total Energy Consumed by Each Module Case 2

En

erg

y C

on

su

me

d (

kW

h)

Modules (No.)


in the final design of the model. Results of the aggregated energy consumption of second case

are depicted in Fig 6.26.

Fig. 6.26 reveals a reasonable reduction in consumption of the HVAC, water heater and lighting

load which is 19.17 kWh, 23.08 kWh and 0.78 kWh, respectively. Second case corresponds to

total energy consumption of 145.17 kWh.

Third case of CHEMA includes scheduling of burst loads. The scheduling is based on the

single Knapsack algorithm which is inspired from an exampleof thief with a sack in a shop

to decide optimally the weights and values of the items constrained to the total capacity of

the sack. Results of the case three are shown in Fig. 5.27 and there is a minor reduction in

aggregated energy consumption and cost has been observed. This is because of the fact that the

total capacity of the burst load is limited to 4.28 kWh which is merely 2% of the total regular

load. Therefore this case shows a reduction of 43.64 kWh energy consumption as compared to

the case one and a further decrease of 0.33 kWh as compared to case two.

Case four of CHEMA has been designed in order to integrate theeffects of the user comfort

in the energy cost minimization model. The basic idea is to relax the temperature thresh hold

limits of the regular loads as described in equation 5.11. The changing of the thresh hold limits

is subject to the activation of the burst load. Results of thecase four are mentioned in Fig. 6.28.

This case further reduces the aggregated energy consumption by 4.55 kWh by taking advantage

120


1 2 3 4 50

20

40

60

80

100


En

erg

y C

on

su

me

d (

kW

h)

Modules (No.)


of the regular load temperature limits relaxation.

Fig. 6.29 shows the comparison of the four cases regarding total energy consumption. Total

energy consumption of the case 1 to 4 is 188.486 kWh, 145.173 kWh, 144.84 kWh and 140.29

kWh, respectively. It is clearly seen that there is a visibleand reasonable reduction in case 2 to

case 4 as compared to case 1 while a little bit further reduction is also caused among cases 2 to

4 by application of different strategies. In summary, case 2corresponds to a 23.11%, case 3 to

24% and case 4 to 25.7% energy cost reduction as compared to case 1.

Table 6.4 shows the total energy consumption results of fourcases.

Table 6.4: Load Comparison of Different Cases

NO Case Total Load (KWh/day)

1 Unscheduled without PPC188.4860

2 Unscheduled with PPC 145.1734

3 Scheduled with PPC 144.8401

4 Load Flexibility (UCI) 140.2925

Our proposed model consists of six layers and has many uniquefeatures including the partial

121


1 2 3 4 50

20

40

60

80

100


En

erg

y C

on

su

me

d (

kW

h)

Modules (No.)


1 2 3 40

20

40

60

80

100

120

140

160

180

200Comparason of Four Cases for Total Energy Consumed

En

erg

y C

on

su

me

d (

kW

h)

Case (No.)

Figure 6.28: Comparison of Total Energy Consumption of Four Cases

122


0 5 10 15 20 250

5

10

15

20S

che

du

al L

oa

d c

ost

PK

R/K

WH

Time in hr

Cost 1Cost 2ToU Cost

Figure 6.29: Impact of Partial Baseline Load Inclusion in the Optimization Model

baseline load concept. Regarding the energy cost minimization, the effect of the partial baseline

load is compared with the reference paper [26]. Fig. 6.30 shows the ToU pricing scheme, cost

of the reference paper (Cost 1) and our scheme’s cost (Cost 2). The appliances’ ratings are taken

same as the reference paper and the baseline load is assumed 1000 W which is dealt according

to equation 6.1. It is clearly seen that in peak hours the costof our proposed model is less than

the reference paper.


In this chapter, we have proposed CHEMA for home energy management with multiple ap-

pliances scheduling options and flexible load categoriz ation. Six layers architecture has been

proposed and four layers have been implemented in Simulink with embedded MATLAB code.

Simulation results have shown the peak load reduction of 22.9% for unscheduled load with PPC,

23.15% for scheduled load with PPC and 25.56% for flexible load categorization. Similarly total

cost reduction of 23.11%, 24% and 25.7% has been observed respectively. Aggregated energy

consumption of various modules used in four cases has also been investigated in this chapter.

Results of energy consumption show total energy consumption of the case 1 to 4 is 188.486

kWh, 145.173 kWh, 144.84 kWh and 140.29 kWh, respectively. Inclusion of partial baseline

123


load into optimization problem caused cost reduction during peak hours. Smart grid interface

layer and load forecasting layers are not implemented in this chapter and effect of the multiple

dynamic pricing is investigated in the next chapter.

124

Chapter 7A Novel Approach for HEMS Scheduling

with Multiple Pricing and Optimization

Techniques


HEMS and effect of home appliances scheduling in smart grid are now familiar topics

of research in electrical engineering. Peak load management, PAR reduction and

associated methods are under focus of researchers since decades. These topics have

got new dimensions in smart grid environment. In this chapter, we have presented a novel

energy management model with enhanced load categorizationand multiple pricing schemes.

Novel idea of RI load is used to ensure the operation of a particular load in the pre-defined slots.

The proposed model uses multiple pricing schemes includingToU, RTP day ahead case and it

is solved with multiple optimization techniques includingKnapsack, PSO and GA. Knapsack

is used with two options of limited slots scheduling and whole day scheduling. Comparative

results of various schemes, generated from combination of pricing and optimization techniques,

have been discussed.

125

7.2. The Proposed Model

7.2 The Proposed Model

Dynamic energy management in smart grid environment with the help of appliances schedul-

ing is one of the major tools for effective DSM in smart grid. There are many approaches for

dynamic energy management using appliances scheduling. Inprevious chapter, we took mo-

tivation from Costanzo energy optimization problem to include the partial baseline load into

CHEMA at second layer [26]. In this chapter, we are enhancingthe scheme with respect to two

aspects: one is the enhancement of the load categorization and other is inclusion of multiple

pricing schemes into energy optimization problem. Effect of multiple dynamic pricing has been

proposed at sixth layer of CHEMA and is being investigated asseparate model in this chapter.

In continuation to previous chapter work and with referenceto Costanzo model, total load has

been divided into three categories: baseline load, regularload and burst load. We included par-

tial baseline load in optimization problem in previous work. This approach has given a useful

solution for house appliances scheduling and cost minimization, however, in order to create

some flexibility and more cost minimization, there is a need of creating new category of appli-

ances which may be used with predefined user slots in order to avoid the peak load or to avoid

the schedule disturbance. To fulfil this requirement, we areintroducing a new category of RI

loads which brings the flexibility in the user schedule. Thiscategory has been introduced for

those users who need a certain load in specic hours. These hours may vary user to user based on

their requirements. The mechanism can be implemented by specifying suitable time slots using

the proposed scheduling methodology.

Another aspect of energy optimization is use of multiple pricing schemes at different occasions

in order to equip the user with multiple pricing schemes, we have introduced multiple pricing

schemes such as ToU, RTP day ahead case and CPP. Furthermore,the solution of a problem

using different techniques provides insight to the technical depth of the problem. We have used

three different types of optimization techniques for solution of the optimization problem. i.e.

Knapsack, PSO and GA, which provides implementation flexibility as well as a comparison

basis. In the proposed model, the lighting load has been included in the scheduled and non-

scheduled baseline load with respect to peak hours as described in equation 5.1. This load has

been added into our basic problem of cost minimization with the assumption that base load

lighting will reduce to half in peak hours. Here in this chapter we are enhancing our proposed

scheme and model by introducing the RI load with the assumption that a facility of plugging a

126

7.2. The Proposed Model

load less than or equal to 1000 W is provided to user in such a way that its priority becomes

highest whenever plugged in the predefined time slot, i.e. the reserved time slot. This is the

reason the load is given the name of RI load. Need of this type of load arises from the scenario

that a user may compromise on baseline lighting to reduce it up to half. The reason of such

compromise may be necessity for pressing clothes by use of anelectric iron in peak hours or,

when a user needs operation of water pump in peak hours due to water shortage etc. Another ex-

ample of RI loads may be of external loads such as drill machine or welding plants, when used

for maintenance purpose could disturb the already running schedule. In order to avoid schedule

disturbance, user may set his maintenance activity accordingly using the RI slots. These loads

are normally schedulable, but under the above mentioned circumstances such loads should get

the highest priority. Mathematically,

ξ =

1000W f or Pre De f ined Slots

0 Otherwise(7.1)

whereξ denotes the total RI load. A set of all appliances is defined asA= 1,2, ...,N, and all

time slots are defined as a set ofT = 1,2, ...,TN. We define a binary variable to represent the

ON and OFF state of the appliances such that:

xi, j =

= 1 i f i th appliance is ON in jth time slot

0 I f appliace is OFF(7.2)

Total number of appliances which are ON atjth time slot can be computed as:

λON = ∑i, j

xi, j (7.3)

If ζi represents the power rating ofith appliance andΩ j denotes the per unit price ofjth time

slot, the cost (σ) of the appliance may be obtained as:

σ = ∑i, j(ζi ∗Ω j) (7.4)

In context of the above discussion, an objective function ofcost minimization has been formu-

lated which includes partial baseline load and RI load into main function. It is essentially an

127

7.3. Simulation Model: Results and Discussion

integrated optimization problem for home energy management. Mathematically,

Minimize∑i, j

Ai, j ∗ai j +∑i, j

Ri, j ∗ r i j +∑i, j

χi, j ∗bi j (7.5)

+∑i, j

ξi, j ∗ ri i j

subject to:

∑i

CA+∑i

CR+∑i

Cχ +∑i

Cξ ≤Cj , ∀ j ∈ T, (7.5a)

ai j , r i j ,bi j ∈ 0,1, (7.5b)

∑i

r i, j = k, (7.5c)

r1+ r2+ r3+ ....+ rn = F, (7.5d)

r1r2+ r2r3+ ....+ rn−1rn = F−1, (7.5e)

where,Ai j , Ri j , χi j andξi j represent the cost of burst, regular, partial baseline and RI loads,

respectively which are computed according to equation 7.4.Notationsai j , r i j , bi j andri i j rep-

resent the ON and OFF states of corresponding category according to equation 7.3. These

variables are binary in nature and can only have 0 or 1 value.CA, CR, Cχ andCξ represent the

capacity of burst, regular, partial baseline and RI loads, respectively in the scheduling problem.

Time slot is indexed by j and T denotes the total number of timeslots. Total available capacity

in jth slot is denoted byCj and k is a number that shows users frequency of scheduling a regular

load. F represents the no. of consecutive slots required. Constraints 7.5(d) and 7.5(e) ensure

the consecutive slots for regular loads. Also

ai j ≤ r it , t = j, j +1, .... j +β−1,

whereβ = Total time slots required for completion of appliance operation and defined as:

β = Total energy required per unit time/Appliance power rating.

We solved this problem by three different techniques with multiple pricing schemes and com-

pared the results.

7.3 Simulation Model: Results and Discussion

For simulations, six appliances have been included in the simulation model. Baseline load has

been partially included into scheduled loads and first appliance with half of the baseline load

128

7.3.1. Knapsack Results

is fixed for baseline scheduling which is restricted not to bescheduled in peak hours. We have

used sixth appliance for RI load, however, user may decide about the slot number to be fixed

for RI. In simulation model, we fixed peak hour time slots 12 to14 for RI loads. We have used

three different pricing schemes for simulation purposes i.e. ToU, RTP day ahead case and CPP.

The model has been analysed with different optimization andpricing schemes with four major

cases. Simulation results are elaborated in subsequent sections.

7.3.1 Knapsack Results

Knapsack problem is a problem of combinatorial optimization which optimizes the number of

various items having certain mass and value. It determines the number of each item to include

in a collection so that total weight is less than or equal to given limit and total value is as large

as possible [164]. In other words, knapsack problem is a problem in which certain items of

different weights and values are given and we have to choose the ones with total weight not

more than some fixed value say W and having maximum collectivevalue among all feasible

combinations. IfCi andpi represent the cost of the energy usage and power, respectively for ith

appliance and ,Pj represent the total power capacity ofjth time slot then,

minimizen

∑i=1

Cixi (7.6)

Subject to

n

∑i=1

pixi ≤ Pj (7.7)

7.3.1.1 Knapsack Case One with ToU

First of all the problem has been solved with single binary knapsack optimization technique

with multiple pricing schemes including ToU, RTP and CPP. InToU pricing, three different

pricing rates are used for off, mid, and full peak hours as: PKR.8/kWh, PKR 15/kWh and PKR

16/kWh, respectively. Loads of 200, 200, 300, 500, 1000, and265 Watt ratings have been

included with following no. of slots required for operationcompletion: 12, 10, 8, 8, 4, and 12

129


0 5 10 15 20 250

2

4

6

8

10

12

14

16

18

PK

R/K

WH

Time (Hrs)

ToU Pricing Scheme

Figure 7.1: ToU Pricing Scheme

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18Scheduled Load with Cost

Time (Hrs)

Pric

e an

d Lo

ad

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18Non−Scheduled Load with Cost

Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.2: Comparison of Scheduled Load and Non-Scheduled Load with Cost

hours, respectively. In first case the strategy adopted is tochoose the slots with minimum price

for scheduling of appliances. The results are shown in Figures 7.1 to 7.3.

It is seen in this instance, the algorithm selects a total of fourteen time slots which corre-

spond to minimum price and is able to schedule the 13.5 kW loadout of 17.98 kW total load.

This is because there is not enough capacity to schedule all the appliances, for instance sixth

load is operated for only one slot instead of 12 slots due to insufficient capacity. It generates a

cost with scheduled load of PKR 191.72 against total load of 21.6 kW. The total unscheduled

cost is PKR 265.64, comparison of scheduled and non-scheduled costs reveals a reduction of

27.8 %. The algorithm reduced the PAR to 32.63 % and took 16.22seconds for 600 iterations

execution on intel core i3, 2.4 GHZ machine with 6 GB RAM.

7.3.1.2 Knapsack Case One with RTP

In RTP, the energy price is set dynamically after each hour according to the dynamic demand

on the grid. There ar two types of RTP: one is RTP in which prices are set after each hour

130


1 2 3 4 5 6 7 8 9 10 11 12 13 140

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Schedule Generated

Time Slots

App

lianc

e O

pera

tion

Load 1Load 2Load 3Load 4Load 5Load 6

Figure 7.3: Generated Schedule

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18

20

22

PK

R/K

WH

Time (Hrs)

RTP (Day Ahead Case)

Figure 7.4: RTP Day Ahead Case

dynamically and other is RTP day ahead case in which prices are decided one day ahead for

next 24 hours. We have used RTP day ahead case in our model. In contrast to ToU, we have

used more variations in prices to implement RTP in realisticmanner. In RTP, it is difficult for

users to set their priorities manually, as price is changingin real time, so the automatic model

of appliance scheduling helps the users to get maximum cost benefit. Figures 7.4 to 7.6 show

the results of applying Knapsack case one with RTP, and final results of all the cases are given

in Table 7.1.

It is seen in Fig. 7.4 that there are more variations as compared to ToU pricing, as it is

decided on the basis of load forecasting. In this pricing we took minimum price as PKR 6/kWh

and maximum as PKR 20/kWh. The scheduler is executed under the condition of putting all the

load in the slots with less than or equal to PKR 12/kWh. Figure7.5 shows the scheduled and

131


0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Pric

e an

d Lo

ad

Time (Hrs)0 5 10 15 20 25

0

2

4

6

8

10

12

14

16

18


Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.5: Comparison of Scheduled and Non-Scheduled Load with Cost

0 5 10 15 20 250

5

10

15

20

25

30

35

40

45

PK

R/K

WH

Time (Hr)

CPP

Figure 7.6: CPP Used with Knapsack

unscheduled comparison and it is clear form that scheduler is not able to execute all the load. It

succeeded in scheduling of 10. 55 kW load out of 17.98 kW. It caused a reduction of 27.9% in

cost and 41.47 % reduction in PAR. It took time of 13.89 secends.

7.3.1.3 Knapsack Case One with CPP Results

In CPP, the difference of the off peak pricing and peak pricing is too much larger than the

normal ToU pricing scheme. This type of scheme may be employed at some critical day or

event. By applying this scheme we want to analyse the effectsof scheduling on critical events

as compared to normal scheduling. In our simulation model, minimum price for CPP is taken

PKR 10/kWh for off peak hours and PKR 30/kWh for mid peak and PKR 40/kWh. for peak

hours. Results are depicted in Figures. 7.7 to 7.9.

Fig. 7.6 shows the CPP used and Fig. 7.7 shows the comparison of scheduled and unsched-

uled load with cost according to the generated schedule. It may be noted that scheduler can

schedule only 12.41 kw of load out of 17.98 kW. Time elapsed is14.94 seconds.

132

7.3.2. Knapsack Case Two

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Pric

e an

d Lo

ad

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.7: Comparison of Scheduled and Non-Scheduled Load with Cost for Knapsack and CPP

0 5 10 15 20 250

2

4

6

8

10

12

14


Time (Hrs)

Pric

e an

d Lo

ad

0 5 10 15 20 250

2

4

6

8

10

12

14


Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.8: Comparison of Scheduled and Non-Scheduled Load with ToU over 24 Hours

7.3.2 Knapsack Case Two

7.3.2.1 Knapsack Case Two With ToU

In Knapsack case one, we limited the scheduler to the slots capable of meeting the certain crite-

ria regarding minimum cost or cost less than or equal to a specific value. The scheduler had to

put the loads against limited capacity and slots. In second case, moving towards more complex

scenario, partial baseline load has been implemented whileemploying scheduler for 24 hours.

Pricing and loads are same as case one. Figure 7.8 shows the comparison of scheduled and

non-scheduled loads with cost while Fig. 7.9 shows the implementation of peak load halving

during peak hours.

It is clearly seen in Fig. 7.10 that partial baseline load which was included in the scheduling

133

7.3.2. Knapsack Case Two

0 5 10 15 20 250

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Schedule Generated

Time Slots

Ap

plia

nce

s O

pe

ratio

n


Figure 7.9: Schedule Generated over 24 Hours with Knapsack and ToU

0 5 10 15 20 250

20

40

60

80

100

120

140

160

180

200Partial Baseline Load Restriction during Peak Hours

Time Slots

Par

tial B

asel

ine

Load

Figure 7.10: Partial Baseline Load Reduction During Peak Hours

problem is restricted for peak hours. This mechanism resulted in 27.19% PAR and 12.3% cost

reduction and algorithm took 14.94 seconds to execute.

7.3.2.2 Knapsack Case Two With RTP

Results of Knapsack case two with RTP are shown in Fig. 7.11. In this case, partial baseline

load is restricted for only two extreme peak hours and it resulted in increase of scheduled cost.

Remaining results are almost similar to that of Knapsack case two and ToU.

134

7.3.3. PSO Results

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Pric

e an

d Lo

ad

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.11: Comparison of Scheduled and Non-Scheduled Load with RTP over 24 Hours

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Pric

e an

d Lo

ad

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Pric

e an

d Lo

ad

PriceLoad

PriceLoad

Figure 7.12: Comparison of Scheduled and Non-Scheduled Load with CPP over 24 Hours

7.3.2.3 Knapsack Case Two with CPP

In this case, 24 hours schedule is implemented with CPP in which the pricing difference is

larger as compared to normal ToU. Results are depicted in Fig. 7.12.

In continuation of discussion of partial peak load reduction, it is restricted for four slots.

This system resulted in 18.83% cost reduction and 27.19% PARreduction.

7.3.3 PSO Results

Optimization techniques are characterized by ease of use, fast convergence properties and

competency to solve a variety of problems including non-linear and non-differentiable multi-

optimization problems. A large number of evolutionary techniques compete for the optimal

135

7.3.3. PSO Results

solution to optimization problems. PSO is one of the robust techniques having a wide variety of

applications [168–170]. PSO is inspired by bird flocking andwas developed by Kennedy and

Eberhart [171]. PSO can be explained by considering a swarm of birds searching for food in

a search space. During search process, each bird has a position and velocity. All birds update

their velocity and position with the knowledge of their own position and the position of a bird

nearest to food. PSO technique makes use of bird flocking scenario and considers each bird as a

particle whose position is a candidate solution in search space whereas the total number of par-

ticles designates size of swarm (or population). In search space, each particle is characterized

by its position, velocity and fitness value.

BPSO is a type of PSO in which decision variables are binary i.e. 0 or 1. The algorithm of

BPSO is same as that of PSO except the following differences:BPSO has binary variables so

particles are initialized randomly for binary positions.

Xi = [Xi1,Xi2..Xin] ∀ Xi1,Xi2..Xin ε 0,1 (7.8)

Binary value having probability of 0.5 is assigned to each dimension of each particle.

xid = f (x) =

1 if rand≥ 0, (7.9)

0 otherwise. (7.10)

Where d=1, 2,..N

Position of each particle is updated by following equations

xk+1id = f (x) =

1 if rand≥ 0, (7.11)

0 rand< sigmoid(Vk+1id ), (7.12)

where sigmoid function is calculated as :

sigmoidk+1id =

1

1+exp−Vk+1id

(7.13)

7.3.3.1 PSO with ToU

In our model, we used PSO with following parameters.

Maximum velocity =Vmax = 4;

Minimum velocity =Vmin = -4;

Total Particles = swarm = 10;

No. of iterations =nitra = 600;

136

7.3.3. PSO Results

0 5 10 15 20 250

2

4

6

8

10

12

14

16

Pric

e an

d Lo

ad

Time (Hrs)

Non−Schedualed Load with Cost

0 5 10 15 20 250

2

4

6

8

10

12

14

16P

rice

and

Load

Time (Hrs)

Schedualed Load with Cost

PriceLoad

PriceLoad

Figure 7.13: Comparison of Scheduled and Non-Scheduled Load using PSO with ToU

0 5 10 15 20 250

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

App

lianc

e O

pera

tion

Time Slots

Schedule Generated with PSO and ToU

Load1Load 2Load 3Load 4Load 5Load 6

Figure 7.14: Schedule Generated with PSO and ToU

Coefficient of social component =c1 = 2.0;

Coefficient of acceleration =c2 = 2.0;

Results are shown in Figures. 7.13 to 7.16 Fig. 7.13 shows thecomparison of scheduled and

non-scheduled load with PSO and ToU pricing scheme. Figure 7.14 shows the generated sched-

ule whereas figures 7.15 and 7.16 depicts the RI load slots.

It may be seen that PSO generates a better schedule than the Knapsack when it is operated

for 600 iterations. It takes 11.76 seconds to execute and results in 43.15% PAR reduction along

with 11.76% cost reduction. Figure 7.16 represents the RI load for 12 slots including the mid

peak and peak slots; this case shows the 34.29% PAR reduction. It clearly shows that increasing

the number of RI slots and the choice of slots will affect the scheduler ability of cost and PAR

reduction. If the slots include the mid peak and peak slots then the effect will be severe as

compared to off peak slots. Another fact about the RI load is of time elapsed, if the scheduler is

137

7.3.3. PSO Results

0 5 10 15 20 250

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

RI L

oa

d

Time Slots

RI Load Specified Slots

Figure 7.15: RI Load Slots

0 5 10 15 20 250

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

RI L

oad

Time (Hrs)

RI Load Specified Slots

Figure 7.16: RI Load with Greater Number of Slots and Variation of Slot Numbers

run without RI load it takes 11.95 seconds as compared to 12.07 seconds of RI case. It means

that inclusion of RI load into scheduling problem causes a very slight increase in the execution

time.

7.3.3.2 PSO with RTP

This combination generates almost the same schedule as ToU,however the percent cost reduc-

tion is reduced to 13.86% as it was expected due to less numberof minimum pricing slots in

RTP day ahead case. This combination generates 56.8% PAR reduction. Constraint violation

results are shown in Fig. 7.17.

138

7.3.4. GA Results

0 5 10 15 20 25−1000

−900

−800

−700

−600

−500

−400

−300

−200

−100

0

Time (Hr)

Con

stra

int V

iola

tion

(Ava

ilabl

e C

apac

ity)

Constraint Violation

Figure 7.17: Constraint Violation in Case of PSO with RTP

0 5 10 15 20 250

5

10

15

20

25

30

35

40

Pric

e an

d Lo

ad

Time (Hrs)

Non−Scheduled Load with Cost

0 5 10 15 20 250

5

10

15

20

25

30

35

40

Pric

e an

d Lo

ad

Time (Hrs)

Scheduled Load with Cost

PriceLoad

PriceLoad

Figure 7.18: Comparison of Scheduled and Non-Scheduled Load using PSO with CPP

7.3.3.3 PSO with CPP

This combination generates almost the same schedule as thatof PSO with RTP, however, the

cost reduction is reached upto 28.52 % as expected due to larger difference of the off peak and

peak hour prices in CPP. Figure 7.18 shows the comparison of scheduled and unscheduled load

with cost when PSO is employed with CPP.

7.3.4 GA Results

GA is an optimization approach, based on theoretical concept of natural evolutionary processes

such as mutation, inheritance, crossover and selection. GAcan easily handle non-linear prob-

lems. In GA, a population of chromosomes is initialized, andeach chromosome represents a

solution where the size of population depends on complexityof the problem in hand. The fit-

139

7.3.4. GA Results

0 5 10 15 20 250

2

4

6

8

10

12

14

16

Pric

e an

d Lo

ad

Time (Hrs)

Non−Schedualed Load with Cost

0 5 10 15 20 250

2

4

6

8

10

12

14

16P

rice

and

Load

Time (Hrs)

Schedualed Load with Cost

PriceLoad

PriceLoad

Figure 7.19: Cost Minimization using GA with ToU

0 5 10 15 20 250

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time (Hr)

App

lianc

e O

pera

tion

Scheudle Generated with GA and ToU


Figure 7.20: Schedule Generated using GA with ToU

ness value of each individual element of population is calculated by fitness function, relatively

fit chromosomes are selected to pass information to next generation (i.e. population for next

iteration) and genetic procedures such as mutation, selection and crossover are performed. Fit-

ness of individual elements increases as the number of generations (iterations) increases. This

process continues until it converges to the best set of chromosomes according to a given crite-

rion [182, 183].

Results of GA with ToU, RTP and CPP are shown in Figs. 7.19 to 7.22. GA simulation

is run for 600 iterations. The results show the early convergence after almost 100 iterations,

however, the simulation is run for full 600 iterations in order to make it comparable with other

techniques used in the model.

It is clear from GA results that it looks the slots and large no. of variables simultaneously

and takes more time than PSO and Knapsack algorithm for 600 iterations. However, GA reduces

140

7.3.5. Energy Consumption Results

0 5 10 15 20 250

5

10

15

20

25

30

35

40P

rice

and

Load

Time (Hrs)


0 5 10 15 20 250

5

10

15

20

25

30

35

40

Pric

e an

d Lo

ad

Time (Hrs)


PriceLoad

PriceLoad

Figure 7.21: Cost Minimization using GA with RTP

0 5 10 15 20 250

5

10

15

20

25

30

35

40

PK

R/K

WH

Time (Hrs)


0 5 10 15 20 250

5

10

15

20

25

30

35

40

PK

R/K

WH

Time (Hrs)


CostLoad

CostLoad

Figure 7.22: Cost Minimization using GA with CPP

the cost more effectively.

7.3.5 Energy Consumption Results

Total energy consumption of an appliance is considered an important parameter in energy man-

agement models as it is helpful to determine the total numberto time slots required to complete

appliance operation. Mathematically:

βai =Eai

Pai(7.14)

Where,βai represents the total time slots required for appliance operation,Eai denotes appliance

total energy consumption andPai shows the appliance power rating. Energy consumed by an

appliance over 24 hours scheduling horizon can be represented mathematically as:

eai = [e1ai...................e

Tai] (7.15)

141


Where,eai representsith appliance energy consumption over scheduling horizon ande1ai de-

notes energy the first slot energy consumption whileT is used to show the length of 24 hours

scheduling horizon. Energy consumed by the appliances can be aggregated as:

Ei =T

∑t=1

eai,t ∀ i = 1.................n (7.16)

Where,eai,t shows theith appliance energy consumption in time slot ‘t’ andEi shows theith

appliance aggregated energy consumption over twenty four hours. This section describes the

aggregated and slot wise energy consumption of the appliances. Energy consumption results

help us in analysis and monitoring of appliances operation.In continuation of our results dis-

cussion of four different cases, we have examined various cases regarding appliances energy

consumption.

In Knapsack first case, model is executed in such a way that thescheduler scans the pricing

scheme in order to pick the minimum pricing slots. After selection of minimum pricing hori-

zon, the scheduler puts the loads into minimum pricing slotsto fulfil the cost minimization

objective. Results of Knapsack first case with ToU, slot-wise energy consumption is shown in

Fig. 7.23 while the aggregated energy consumption results are shown in figures 7.24.

Fig. 7.23 shows a scheduling horizon of 14 slots. The scheduling horizon is reduced to 14 slots

instead of 24 because of the application of minimum price strategy. The scheduling horizon

consisting of 14 slots is in accordance with the ToU pricing slots as shown in Fig. 7.1 which

shows the 14 slots against PKR 8/KWhr. Minimum price strategy definitely affects the total

load operation and the scheduler is not able to schedule all the loads. The effect of reduced

capacity is clearly seen in Fig. 7.33, where the aggregated energy consumption for load 5 is

only 1 kWhr instead of desired 4 kWhr.

In case of Knapsack with RTP case, the scheduler is limited tothe 12 slots scheduling hori-

zon because there are more price variations in case of RTP. Load 5 has not been scheduled by

scheduler because of the limited capacity and minimum pricestrategy. Therefore, energy con-

sumption of load 5 zero. CPP pricing scheme is applied on events and critical days in which the

difference between off-peak and peak pricing is almost three times greater as that of ToU and

RTP schemes. Energy consumption results of Knapsack case one with CPP schemes reveal a

scheduling horizon of 12 slots only because the greater difference of off-peak and peak prices.

Similar to the case of ToU, the scheduler is failed to schedule the load 5 and aggregated en-

ergy consumption for load 5 is zero. In summary of Knapsack case on, the scheduler is failed to

schedule the load 5, therefore the aggregated energy is calculated as zero. The minimum pricing

142


strategy reduces the scheduler capacity to run all the appliances according to desired number of

slots.

Cost minimization is very attractive option for users, however, the full appliances operation and

required energy consumption in order to meet the comfort of the users is also necessary to make

the scheduling and HEMS a popular option. In order to enhancethe scheduling capacity and

full operation of load 5, in Knapsack case two, the scheduleris freed of minimum price strategy

in order to expand the scheduling horizon over 24 slots. The results of Knapsack case two with

RTP for slot-wise and aggregated consumption are shown in Figures 7.25 and 7.26.

It is seen clearly in Figs. 7.25 to 7.26 that load has been executed for desired number of slots and

therefore the aggregated energy consumption of load 5 is same in all the three cases. However,

the effect of cost reduction in three cases is different because of different pricing schemes used.

Cost reduction has been observed 12.3% in case of Knapsack case two with ToU and 18.83%

in case of CPP pricing scheme.

PSO technique is one of the heuristic techniques used for solution of complex optimization

problems and is inspired from food searching technique of bird flocks. It is an easy, user friendly

and fast converging optimization technique. PSO results for slot wise and aggregated energy

consumption have been shown in Figures 7.27 to 7.28. Examination of PSO generated slot wise

energy consumption results reveal that the load profile is more flattened than the Knapsack re-

sults and continuous slots appliances operation and bettercost minimization and PAR reduction

results. The aggregated energy consumption of load 6 is reduced because of using it as RI load

and limiting it to three slots only.

GA is another heuristic optimization algorithm which is characterized by crossover and muta-

tion effects. Distribution of load 5 in GA results is different in three cases and the schedule

generated is more flattened load profile as compared to Knapsack and PSO because of the cross

over and mutation properties of GA.

In summary, results show that although Knapsack case one with CPP provides 49.2% cost re-

duction, it cannot schedule the full load. Also it is for CPP which is only useful for critical days

or events. Overall better cost minimization has been achieved with GA and PAR reduction has

been better achieved with PSO. PSO showed the 43.73% PAR reduction with all the pricing

schemes.

Figures 7.29 to 7.31 show the comparison of scheduled and unscheduled load with cost

for Binary Integer Programming (BIP) with ToU, RTP and CPP, respectively. BIP has been

143


1 2 3 4 5 6 7 8 9 10 11 12 13 140

100

200

300

400

500

600

700

800

900

1000Appliances Energy Consumption

Load

Time (Hrs)

Load1Load 2Load 3Load 4Load 5Load 6

Figure 7.23: Slot-Wise Energy Consumption (Knapsack case one with ToU)

used in [197] for cost minimization. Comparison of BIP shows27%, 31% and 33% cost

reduction in three cases. Comparing these values with our results, it is clear that our model is

superior with respect to cost minimization as our schemes ofGA with three pricing schemes

have achieved 32.46%, 39.17% and 39.92% cost reduction. Another technique named Wind

Driven Optimization (WDO) has been used in [198] for energy cost minimization. Figures

7.32 to 7.34 show the scheduled and unscheduled cost comparison for WDO with ToU, RTP

and CPP, respectively. WDO has shown 21.11%, 31.24% and 32.84% cost reduction for three

cases. Comparing our schemes with this technique we, can observe that our schemes reduce

more cost as compared to WDO.


In this chapter, we have presented a novel home energy management model that includes par-

tial baseline and RI loads into cost optimization problem. Also a comparative study has been

presented in order to compare multiple dynamic pricing and optimization techniques for home

appliances scheduling. Three pricing schemes namely ToU, RTP day ahead case and CPP have

been analyzed employing three different optimization techniques i. e. Knapsack, PSO and GA.

Knapsack has been used with two different options. Results show that although Knapsack case

one with CPP provides 49.2% cost reduction, it can not schedule the full load. Also it is for

CPP which is only useful for critical days or events. Overallbetter cost minimization has been

144


1 2 3 4 5 60

500

1000

1500

2000

2500

3000Total Energy Consumed by Each Appliance

Ene

rgy

Con

sum

ed

Appliances (No.)

Figure 7.24: Aggregated Energy Consumption (Knapsack case one with ToU)

0 5 10 15 20 250

100

200

300

400

500

600

700

800

900


Load

Time (Hrs)


Figure 7.25: Slot-Wise Energy Consumption (Knapsack case two with RTP)

achieved with GA and PAR reduction has been better achieved with PSO. PSO showed the

43.73% PAR reduction with all the pricing schemes. Cost comparison of Knapsack with BIP

shows 10.1% more reduction in case of Knapsack over 24 hours.This comparison also revealed

that BIP algorithm takes 38.9 seconds for execution which isgreater than the Knapsack case

with ToU pricing and it cases 31.23% PAR reduction. Effect ofthe RI load duration and slot

numbers has also been studied in this scheme which shows thatincreasing the duration of RI

adversly affects the scheduler’ capability of cost and PAR reductio. The proposed model has

diverse application for consumers and utilities. It is worth noting that appliances scheduling

can severely affect the users comfort. Integration of user comfort level into cost optimization

problem is planned in our future work. Suggestion of new pricing scheme with inclusion of

145


1 2 3 4 5 60

500

1000

1500

2000

2500

3000

3500


Ene

rgy

Con

sum

ed

Appliances (No.)

Figure 7.26: Aggregated Energy Consumption (Knapsack case two with RTP)

0 5 10 15 20 250

100

200

300

400

500

600

700

800

900


Load

Time (Hrs)


Figure 7.27: Slot-Wise Energy Consumption (PSO with CPP)

already existing schemes characteristics in a single day inorder to get better cost minimization

solutions is also a part of our future work.

146


1 2 3 4 5 60

500

1000

1500

2000

2500

3000

3500


Ene

rgy

Con

sum

ed

Appliances (No.)

Figure 7.28: Aggregated Energy Consumption (PSO with CPP)

0 5 10 15 20 250

2

4

6

8

10

12

14


Time (Hrs)

Price

an

d L

oa

d

0 5 10 15 20 250

2

4

6

8

10

12

14


Time (Hrs)

Price

an

d L

oa

d

PriceLoad

PriceLoad

Figure 7.29: Cost Minimization using BIP with ToU

147


0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Price

an

d L

oa

d

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Price

an

d L

oa

d

PriceLoad

PriceLoad

Figure 7.30: Cost Minimization using BIP with RTP

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Price

an

d L

oa

d

0 5 10 15 20 250

5

10

15

20

25

30

35


Time (Hrs)

Price

an

d L

oa

d

PriceLoad

PriceLoad

Figure 7.31: Cost Minimization using BIP with CPP

148


0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Price

an

d L

oa

d

0 5 10 15 20 250

2

4

6

8

10

12

14

16

18


Time (Hrs)

Price

an

d L

oa

d

PriceLoad

PriceLoad

Figure 7.32: Cost Minimization using WDO with ToU

0 5 10 15 20 250

5

10

15

20

25

30

35

40

Price

an

d L

oa

d

Time (Hrs)


0 5 10 15 20 250

5

10

15

20

25

30

35

40

Price

an

d L

oa

d

Time (Hrs)


PriceLoad

PriceLoad

Figure 7.33: Cost Minimization using WDO with RTP

149


0 5 10 15 20 250

5

10

15

20

25

30

35

40

Price

an

d L

oa

d

Time (Hrs)


0 5 10 15 20 250

5

10

15

20

25

30

35

40

Price

an

d L

oa

d

Time (Hrs)


PriceLoad

PriceLoad

Figure 7.34: Cost Minimization using WDO with CPP

Table 7.1: Comparative Results of Different Techniques

Knapsack Case 1 Knapsack Case 2 PSO GA

ToU

% Cost Reduction 27.82 12.3 21.63 32.46

Time Elapsed 16.22 14.94 11.8 142.89

% PAR Reduction 32.63 27.19 43.73 36.38

RTP

% Cost Reduction 27.9 7 13.5 39.17

Time Elapsed 13.89 21.87 12.07 143.47

% PAR Reduction 41.47 27.19 43.73 36.38

CPP

% Cost Reduction 49.2 18.83 28.52 39.92

Time Elapsed 14.94 22.28 11.8 145.46

% PAR Reduction 33.36 27.19 43.73 36.38

150

Chapter 8Conclusions and Future Work

In this chapter, overall conclusions of the study are presented that highlight the need of proposed

schemes. CHEMA is a comprehensive architecture which covers multiple dimensions of the

energy management issue. Comparative approach provides anintuition about the peak load

shaving, cost minimization and flexible load scheduling in depth.

8.1 Conclusions

This research work provides following important conclusions:

Pakistan’s electricity requirements will triple by 2050. If sufficient resources are not allocated,

the energy crisis in the country will intensify. An independent energy authority at the national

level is required to make future plans for the development and utilization of indigenous re-

sources like hydro, coal, nuclear and renewables. It is alsorequired to analyze the available

options to import energy from neighboring countries in order to secure the future of the coun-

try. The comparison of TAPI and IPI with LNG import options reveals the suitability of LNG

over the gas pipelines. The optimum utilization of existinginstalled thermal generation and use

of combined cycle power plants is imperative. Pakistan has abundant potential of renewable

energy resources. The issue of circular debt can be minimized by introducing more renewable

energy in the national grid. Smart grid implementation is necessary for effective energy man-

agement.

Exploitation of enormous potential of smart grid, for wellbeing of mankind, is dependent upon

the rapid development of advanced communications infrastructure and optimization of network

151

8.1. Conclusions

parameters. We have reviewed the WCTs for HANs and NANs separately and compared them

in context of consumer concerns and utility requirements. Wi-Fi offers high data rate and larger

range (100 m indoor) as compared to Bluetooth short range (10m typical) secure communica-

tions, but ZigBee with low cost, low power consumption, reasonable indoor range (10-75 m)

and ability of accommodating a very large number of nodes seems to be the best candidate for

HANs. Low data rate of ZigBee (40-250 Kbps) is one of the reasons it is a low power tech-

nology. Although ZigBee data rate is much less than Wi-Fi (11-300 Mbps), it is good enough

for most of the usual HAN applications. Furthermore ZigBee supports a high number of nodes

(more than 64,000) which makes communications scalable when new nodes enter the system.

Z-Wave and 6LoWPAN are advantageous for IP enabled low powerdevices. Bandwidth and

the range of WiMAX make it appropriate for smart grid NANs applications. WiMAX offers

72 Mbps speed which is 36 times greater than typical 3G GSM speed. On the other hand,

pre-existing cellular networks with sufficient bandwidth,high data rates (more than 300 Mbps

for 4G LTE), extensive coverage, lower maintenance costs and strong security can save capital

investments on utility dedicated communications infrastructure and allow rapid deployment of

applications. However, use of pre-existing cellular networks will raise the operational cost in

terms of monthly recurring charges. A dedicated utility cellular network can ensure the high

quality and reliability with reduced operational cost. There exists a clear trade-off when mak-

ing a choice between dedicated WCTs like WiMAX and pre-existing technologies like GSM

cellular network. This choice essentially depends on a particular utilitys budget and policies.

This comprehensive work has tabulated different wireless communications in a systematic way

along with the challenges for smart grid implementation.

In this work, we have presented a detailed review of DSM techniques in smart grid including

incentive based DLC and dynamic pricing based HEMS. DSM related smart grid technologies

and challenges have also been discussed. A comparison of different dynamic pricing based

HEMS is presented incorporating various factors such as billing mechanism, fairness among

users, algorithm processing times etc. Ten recent and prominent schemes of HEMS have been

compared analytically. The highest Fairness (73%) is ensured in [55] and the highest PAR

reduction is 38.1 % [11] among the compared schemes. VariousDLC schemes have been em-

ployed to control energy consumption; more useful with greater peak load reduction potential

for heavy loads. HEMS provide more effective ways of LM, especially for residential load, by

employing efficient optimization techniques and ensuring the users privacy and comfort. Effi-

152

8.1. Conclusions

cient DSM schemes are essential in order to control energy consumption; various technologies

are required in order to achieve the benefits of DSM in smart grid, including ICTs and advanced

control mechanism. Finding suitable communication and control infrastructure, making LM

policies and optimizing energy consumption are the ongoingresearch areas related to efficient

DSM in smart grid.

In this work, we have proposed CHEMA with flexible load categorization. Multi layered archi-

tecture has been implemented in Simulink with embedded MATLAB code. Simulation results

have shown the peak load reduction of 22.9% for unscheduled load with PPC, 23.15% for sched-

uled load with PPC and 25.56% for flexible load categorization. Similarly total cost reduction

of 23.11%, 24% and 25.7% has been observed, respectively. Aggregated energy consumption

of various modules used in four cases has also been investigated in CHEMA. Results of energy

consumption show total energy consumption of the case 1 to 4 is 188.486 kWh, 145.173 kWh,

144.84 kWh and 140.29 kWh, respectively. Smart grid interface layer and load forecasting lay-

ers are not implemented in current work and will be focused infuture work.

This work also presented a novel home energy management model that includes partial base

line and RI loads into cost optimization problem. Also a comparative study has been pre-

sented in order to compare multiple dynamic pricing and optimization techniques for home

appliances scheduling. Three pricing schemes namely ToU, RTP day ahead case and CPP have

been analyzed employing three different optimization techniques i. e. Knapsack, PSO and GA.

Knapsack has been used with two different options. Results show that although Knapsack case

one with CPP provides 49.2% cost reduction, it can not schedule the full load. Also it is for

CPP which is only useful for critical days or events. Overallbetter cost minimization has been

achieved with GA and PAR reduction has been better achieved with PSO. PSO showed the

43.73 % PAR reduction with all the pricing schemes. Cost comparison of Knapsack with BIP

shows 10.1% more reduction in case of Knapsack over 24 hours.This comparison also revealed

that BIP algorithm takes 38.9 seconds for execution which isgreater than the Knapsack case

with ToU pricing and it cases 31.23% PAR reduction. Effect ofthe RI load duration and slot

numbers has also been studied in this scheme which shows thatincreasing the duration of RI

adversly affects the scheduler’ capability of cost and PAR reductio.The proposed model has

diverse application for consumers and utilities. It is worth noting that appliances scheduling

can severely affect the users comfort. Integration of user comfort level into cost optimization

problem is planned in our future work. Suggestion of new pricing scheme with inclusion of

153

8.2. Future Work

already existing schemes characteristics in a single day inorder to get better cost minimization

solutions is also a part of our future work.

8.2 Future Work

Enhancement of this work with remaining layers implementation is aimed in future work. Fol-

lowing are the future work directions:

• Individual habit based modeling of energy controlling module implementation

• Pattern recognition based persons presence controller for energy management

• Scalability Analysis of the proposed schemes with respectto number of devices and users

• Development of new pricing scheme for better DSM in smart grid environment in context

of appliance scheduling

• Co-operative DSM schemes and design of implementation strategy for Pakistan

• Short term load forecasting based appliance scheduling

In addition, development of smart grid labs, energy management test beds equipped with various

communications technologies including wired and wirelesssensor networks for implementation

of proposed architecture is also targeted in future work. Development of co-operative DSM

schemes and design of implementation strategy for Pakistanis also one of the future plans.

154

Bibliography

[1] Alternate Energy Development Board, “Mowp gop pakistan,” (htt p :

//www.aedb.org/parl.htm);2013, accessed May 2013.

[2] International Energy Agency (IEA), “Policy path ways modernising building energy

codes 2012,” 2013.

[3] A. Mahmood, N. Javaid, A. Zafar, A. R. Raja, S. Ahmed, and S. Razzaq, “Pakistan’s

overall energy potential assessment, comparison of LNG, TAPI and IPI gas projects,”

Renewable and Sustainable Energy Reviews, vol. 31, pp. 182–193, 2014.

[4] M. Alizadeh, X. Li, Z. Wang, A. Scaglione, and R. Melton, “Demand-side management

in the smart grid: Information processing for the power switch,” Signal Processing Mag-

azine, IEEE, vol. 29, no. 5, pp. 55–67, 2012.

[5] Q. Zhang, K. N. Ishihara, B. C. Mclellan, and T. Tezuka, “An analysis methodology for

integrating renewable and nuclear energy into future smartelectricity systems,”Interna-

tional Journal of Energy Research, vol. 36, no. 15, pp. 1416–1431, 2012.

[6] H. Fathabadi, “Ultra high benefits system for electric energy saving and management of

lighting energy in buildings,”Energy Conversion and Management, vol. 80, pp. 543–549,

2014.

[7] M. Isaac and D. P. Van Vuuren, “Modeling global residential sector energy demand for

heating and air conditioning in the context of climate change,” Energy Policy, vol. 37,

no. 2, pp. 507–521, 2009.

155

BIBLIOGRAPHY

[8] Y. Yan, Y. Qian, H. Sharif, and D. Tipper, “A survey on smart grid communication in-

frastructures: Motivations, requirements and challenges,” Communications Surveys &

Tutorials, IEEE, vol. 15, no. 1, pp. 5–20, 2013.

[9] C. W. Gellings, “The smart grid: enabling energy efficiency and demand response”.The

Fairmont Press, Inc., 2009.

[10] J. Lindgren, R. Niemi, and P. D. Lund, “Effectiveness ofsmart charging of electric vehi-

cles under power limitations,”International Journal of Energy Research, vol. 38, no. 3,

pp. 404–414, 2014.

[11] A.-H. Mohsenian-Rad, V. W. Wong, J. Jatskevich, and R. Schober, “Optimal and au-

tonomous incentive-based energy consumption scheduling algorithm for smart grid,” in

Innovative Smart Grid Technologies (ISGT), 2010, pp. 1–6, IEEE, 2010.

[12] R. E. Brown, “Impact of smart grid on distribution system design,” inPower and En-

ergy Society General Meeting-Conversion and Delivery of Electrical Energy in the 21st

Century, 2008 IEEE, pp. 1–4, IEEE, 2008.

[13] A.-H. Mohsenian-Rad and A. Leon-Garcia, “Optimal residential load control with price

prediction in real-time electricity pricing environments,” Smart Grid, IEEE Transactions

on, vol. 1, no. 2, pp. 120–133, 2010.

[14] Q. Zhu, Z. Han, and T. Basar, “A differential game approach to distributed demand side

management in smart grid,” inCommunications (ICC), 2012 IEEE International Confer-

ence on, pp. 3345–3350, IEEE, 2012.

[15] M. Ullah, A. Mahmood, S. Razzaq, M. Ilahi, R. Khan, and N.Javaid, “A survey of

different residential energy consumption controlling techniques for autonomous dsm in

future smart grid communications,”arXiv preprint arXiv:1306.1134, 2013.

[16] J.-J. Wang and S. Wang, “Wireless sensor networks for home appliance energy man-

agement based on zigbee technology,” inMachine Learning and Cybernetics (ICMLC),

2010 International Conference on, vol. 2, pp. 1041–1046, IEEE, 2010.

[17] N. S. Rau, P. Eng, K. Rose, K. W. Costello, and Y. Hegazy,“ Methods to Quantify Energy

Savings from Demand-side Management Programs: A TechnicalReview,” . National

Regulatory Research Institute, 1991.

156

BIBLIOGRAPHY

[18] K. Kostkova, L. Omelina, P. Kycina, and P. Jamrich, “An introduction to load manage-

ment,”Electric Power Systems Research, vol. 95, pp. 184–191, 2013.

[19] J. Osborne and D. Warrier, “A primer on demand response–the power grid: Evolving

from a dumb network to a smart grid,”Thomas Weisel Partners, Tech. Rep, 2007.

[20] R. Petrichenko and V. Chuvychin, “Smart load shedding system,” inPower Electronics

for Distributed Generation Systems (PEDG), 2012 3rd IEEE International Symposium

on, pp. 64–71, IEEE, 2012.

[21] H. Seyedi and M. Sanaye-Pasand, “New centralised adaptive load-shedding algorithms to

mitigate power system blackouts,”Generation, Transmission & Distribution, IET, vol. 3,

no. 1, pp. 99–114, 2009.

[22] M. Mostafa, M. El-Hawary, G. Mbamalu, M. Mansour, K. El-Nagar, and A. El-Arabaty,

“Steady-state load shedding schemes: a performance comparison,” Electric power sys-

tems research, vol. 38, no. 2, pp. 105–112, 1996.

[23] D. Hazarika and A. Sinha, “Method for optimal load shedding in case of generation defi-

ciency in a power system,”International Journal of Electrical Power & Energy Systems,

vol. 20, no. 6, pp. 411–420, 1998.

[24] A. A. Girgis and S. Mathure, “Application of active power sensitivity to frequency and

voltage variations on load shedding,”Electric Power Systems Research, vol. 80, no. 3,

pp. 306–310, 2010.

[25] F. Shokooh, J. Dai, S. Shokooh, J. Taster, H. Castro, T. Khandelwal, and G. Donner, “An

intelligent load shedding (ils) system application in a large industrial facility,” inIndustry

Applications Conference, 2005. Fourtieth IAS Annual Meeting. Conference Record of the

2005, vol. 1, pp. 417–425, IEEE, 2005.

[26] G. T. Costanzo, G. Zhu, M. F. Anjos, and G. Savard, “A system architecture for au-

tonomous demand side load management in smart buildings,”Smart Grid, IEEE Trans-

actions on, vol. 3, no. 4, pp. 2157–2165, 2012.

[27] A. Ipakchi and F. Albuyeh, “Grid of the future,”Power and Energy Magazine, IEEE,

vol. 7, no. 2, pp. 52–62, 2009.

157

BIBLIOGRAPHY

[28] M. Rastegar, M. Fotuhi-Firuzabad, and M. Moeini-Aghtaie, “Improving direct load con-

trol implementation by an inititative load control method,” in Electrical Power Distribu-

tion Networks (EPDC), 2013 18th Conference on, pp. 1–5, IEEE, 2013.

[29] G. R. Newsham and B. G. Bowker, “The effect of utility time-varying pricing and load

control strategies on residential summer peak electricityuse: a review,”Energy Policy,

vol. 38, no. 7, pp. 3289–3296, 2010.

[30] B. Ramanathan and V. Vittal, “A framework for evaluation of advanced direct load con-

trol with minimum disruption,”Power Systems, IEEE Transactions on, vol. 23, no. 4,

pp. 1681–1688, 2008.

[31] C.-M. Chu and T.-L. Jong, “A novel direct air-conditioning load control method,”Power

Systems, IEEE Transactions on, vol. 23, no. 3, pp. 1356–1363, 2008.

[32] P. Palensky and D. Dietrich, “Demand side management: Demand response, intelligent

energy systems, and smart loads,”Industrial Informatics, IEEE Transactions on, vol. 7,

no. 3, pp. 381–388, 2011.

[33] A. Neuberg, “Ripple control in the czech republic and demand side management,” in

Electricity Distribution-Part 1, 2009. CIRED 2009. 20th International Conference and

Exhibition on, pp. 1–5, IET, 2009.

[34] R. Tyagi and J. W. Black, “Emergency demand response fordistribution system contin-

gencies,” inTransmission and Distribution Conference and Exposition,2010 IEEE PES,

pp. 1–4, IEEE, 2010.

[35] P. Supriya, T. Nambiar, R. Charu, A. Tyagi, V. Nagadharni, and M. Deepika, “A labo-

ratory prototype of a smart grid based demand side management,” in Innovative Smart

Grid Technologies-India (ISGT India), 2011 IEEE PES, pp. 28–31, IEEE, 2011.

[36] J. A. Qureshi, M. Gul, and W. A. Qureshi, “Demand side management through innovative

load control,” inTENCON 2010-2010 IEEE Region 10 Conference, pp. 580–585, IEEE,

2010.

[37] V. Ricquebourg, D. Menga, D. Durand, B. Marhic, L. Delahoche, and C. Loge, “The

smart home concept: our immediate future,” inE-Learning in Industrial Electronics,

2006 1ST IEEE International Conference on, pp. 23–28, IEEE, 2006.

158

BIBLIOGRAPHY

[38] S. T. Mak, “Knowledge based architecture serving as a rigid framework for smart grid

applications,” inInnovative Smart Grid Technologies (ISGT), 2010, pp. 1–8, IEEE, 2010.

[39] M. A. Sheikh, “Renewable energy resource potential in pakistan,”Renewable and Sus-

tainable Energy Reviews, vol. 13, no. 9, pp. 2696–2702, 2009.

[40] A.-H. Mohsenian-Rad, V. W. Wong, J. Jatskevich, R. Schober, and A. Leon-Garcia,

“Autonomous demand-side management based on game-theoretic energy consumption

scheduling for the future smart grid,”Smart Grid, IEEE Transactions on, vol. 1, no. 3,

pp. 320–331, 2010.

[41] G. Strbac, “Demand side management: Benefits and challenges,”Energy Policy, vol. 36,

no. 12, pp. 4419–4426, 2008.

[42] H.-C. Sun and Y.-C. Huang, “Optimization of power scheduling for energy management

in smart homes,”Procedia Engineering, vol. 38, pp. 1822–1827, 2012.

[43] G. Fandi, Z. Muller, L. Straka, and J. Svec, “Facts devices influence on power losses in

transmission systems,” inElectric Power Engineering (EPE), Proccedings of the 2014

15th International Scientific Conference on, pp. 29–33, IEEE, 2014.

[44] X. Yuan, Z. Qian, Y. Zhou, Y. Wang, and M. Yan, “Discussion on the development trend

of smart grid and its key technology,” inElectricity Distribution (CICED), 2012 China

International Conference on, pp. 1–8, IEEE, 2012.

[45] C. Hochgraf, R. Tripathi, and S. Herzberg, “Smart grid charger for electric vehicles us-

ing existing cellular networks and sms text messages,” inSmart Grid Communications

(SmartGridComm), 2010 First IEEE International Conference on, pp. 167–172, IEEE,

2010.

[46] J.-S. Lee, C.-C. Chuang, and C.-C. Shen, “Applicationsof short-range wireless tech-

nologies to industrial automation: a zigbee approach,” inTelecommunications, 2009.

AICT’09. Fifth Advanced International Conference on, pp. 15–20, IEEE, 2009.

[47] K. Tsang, W. Lee, K. Lam, H. Tung, and K. Xuan, “An integrated zigbee automation

system: An energy saving solution,” inMechatronics and Machine Vision in Practice,

2007. M2VIP 2007. 14th International Conference on, pp. 252–258, IEEE, 2007.

159

BIBLIOGRAPHY

[48] H. Kumar Nunna and S. Doolla, “Energy management in microgrids using demand re-

sponse and distributed storage a multiagent approach,”Power Delivery, IEEE Transac-

tions on, vol. 28, no. 2, pp. 939–947, 2013.

[49] P. Cheah, R. Zhang, H. Gooi, H. Yu, and M. Foo, “Consumer energy portal and home

energy management system for smart grid applications,” inIPEC, 2012 Conference on

Power & Energy, pp. 407–411, IEEE, 2012.

[50] J. Byun, I. Hong, and S. Park, “Intelligent cloud home energy management system using

household appliance priority based scheduling based on prediction of renewable energy

capability,”Consumer Electronics, IEEE Transactions on, vol. 58, no. 4, pp. 1194–1201,

2012.

[51] H. J. Zainzinger, “An artificial intelligence based tool for home automation using mat-

lab,” in Tools with Artificial Intelligence, 1998. Proceedings. Tenth IEEE International

Conference on, pp. 256–261, IEEE, 1998.

[52] T. Tajikawa, H. Yoshino, T. Tabaru, and S. Shin, “The energy conservation by informa-

tion appliance,” inSICE 2002. Proceedings of the 41st SICE Annual Conference, vol. 5,

pp. 3127–3130, IEEE, 2002.

[53] P. Samadi, H. Mohsenian-Rad, V. W. Wong, and R. Schober,“Tackling the load uncer-

tainty challenges for energy consumption scheduling in smart grid,” Smart Grid, IEEE

Transactions on, vol. 4, no. 2, pp. 1007–1016, 2013.

[54] M. Erol-Kantarci and H. T. Mouftah, “Wireless sensor networks for cost-efficient resi-

dential energy management in the smart grid,”Smart Grid, IEEE Transactions on, vol. 2,

no. 2, pp. 314–325, 2011.

[55] Z. Baharlouei, M. Hashemi, H. Narimani, and H. Mohsenian-Rad, “Achieving optimality

and fairness in autonomous demand response: Benchmarks andbilling mechanisms,”

2013.

[56] N. Gudi, L. Wang, and V. Devabhaktuni, “A demand side management based simulation

platform incorporating heuristic optimization for management of household appliances,”

International Journal of Electrical Power & Energy Systems, vol. 43, no. 1, pp. 185–193,

2012.

160

BIBLIOGRAPHY

[57] Y. Zhou, Y. Chen, G. Xu, Q. Zhang, and L. Krundel, “Home energy management with pso

in smart grid,” inIndustrial Electronics (ISIE), 2014 IEEE 23rd International Symposium

on, pp. 1666–1670, IEEE, 2014.

[58] D. M. Falcao, “Genetic algorithms applications in electrical distribution systems,” in

Computational Intelligence, Proceedings of the World on Congress on, vol. 2, pp. 1063–

1068, IEEE, 2002.

[59] L. D. Ha, S. Ploix, E. Zamai, and M. Jacomino, “Tabu search for the optimization of

household energy consumption,” inInformation Reuse and Integration, 2006 IEEE In-

ternational Conference on, pp. 86–92, IEEE, 2006.

[60] L. Zhang, H. Leung, and K. C. Chan, “Information fusion based smart home control

system and its application,”Consumer Electronics, IEEE Transactions on, vol. 54, no. 3,

pp. 1157–1165, 2008.

[61] W. Kastner, M. J. Kofler, and C. Reinisch, “Using ai to realize energy efficient yet com-

fortable smart homes,” inFactory Communication Systems (WFCS), 2010 8th IEEE In-

ternational Workshop on, pp. 169–172, IEEE, 2010.

[62] J. H. Yoon, R. Baldick, and A. Novoselac, “Dynamic demand response controller based

on real-time retail price for residential buildings,”Smart Grid, IEEE Transactions on,

vol. 5, no. 1, pp. 121–129, 2014.

[63] M. Erol-Kantarci and H. T. Mouftah, “Wireless sensor networks for domestic energy

management in smart grids,” inCommunications (QBSC), 2010 25th Biennial Sympo-

sium on, pp. 63–66, IEEE, 2010.

[64] M. Erol-Kantarci and H. T. Mouftah, “Using wireless sensor networks for energy-aware

homes in smart grids,” inComputers and Communications (ISCC), 2010 IEEE Sympo-

sium on, pp. 456–458, IEEE, 2010.

[65] M. A. A. Pedrasa, T. D. Spooner, and I. F. MacGill, “Coordinated scheduling of residen-

tial distributed energy resources to optimize smart home energy services,”Smart Grid,

IEEE Transactions on, vol. 1, no. 2, pp. 134–143, 2010.

161

BIBLIOGRAPHY

[66] J. Lee, H.-J. Kim, G.-L. Park, and M. Kang, “Energy consumption scheduler for demand

response systems in the smart grid.,”Journal of Information Science & Engineering,

vol. 28, no. 5, 2012.

[67] T. Hubert and S. Grijalva, “Modeling for residential electricity optimization in dynamic

pricing environments,”Smart Grid, IEEE Transactions on, vol. 3, no. 4, pp. 2224–2231,

2012.

[68] M. Kuzlu, M. Pipattanasomporn, and S. Rahman, “Hardware demonstration of a home

energy management system for demand response applications,” Smart Grid, IEEE Trans-


[69] J. Han, C.-S. Choi, W.-K. Park, I. Lee, and S.-H. Kim, “Smart home energy management

system including renewable energy based on zigbee and plc,”Consumer Electronics,

IEEE Transactions on, vol. 60, no. 2, pp. 198–202, 2014.

[70] A. Anvari-Moghaddam, H. Monsef, and A. Rahimi-Kian, “Optimal smart home energy

management considering energy saving and a comfortable lifestyle,”Smart Grid, IEEE


[71] M. Heleno, M. Matos, and J. Lopes, “Availability and flexibility of loads for the provision

of reserve,”Smart Grid, IEEE Transactions on, vol. 6, no. 2, pp. 667–674, 2015.

[72] J. Moenik, M. Gornik, B. Murovec, and A. Zemva, “A concept to optimize power con-

sumption in smart homes based on demand-side management andusing smart switches,”

Electrotechnical Review, vol. 80, no. 5, pp. 217–221, 2013.

[73] H.-A. Ibrahim-Attia, “Energy saving through smart home,” The Online Journal on Power

and Energy Engineering (OJPEE), vol. 2, no. 3, pp. 223–227.

[74] J. Han, C.-S. Choi, and I. Lee, “More efficient home energy management system based

on zigbee communication and infrared remote controls,”Consumer Electronics, IEEE


[75] K. Long and Z. Yang, “Model predictive control for household energy management based

on individual habit,” inControl and Decision Conference (CCDC), 2013 25th Chinese,

pp. 3676–3681, IEEE, 2013.

162

BIBLIOGRAPHY

[76] S. Kahrobaee, S. Asgarpoor, and W. Qiao, “Optimum sizing of distributed generation and

storage capacity in smart households,”Smart Grid, IEEE Transactions on, vol. 4, no. 4,

pp. 1791–1801, 2013.

[77] T. Jiang, Y. Cao, L. Yu, and Z. Wang, “Load shaping strategy based on energy storage

and dynamic pricing in smart grid,”Smart Grid, IEEE Transactions on, vol. 5, no. 6,

pp. 2868–2876, 2014.

[78] S. Althaher, P. Mancarella, and J. Mutale, “Automated demand response from home en-

ergy management system under dynamic pricing and power and comfort constraints,”

[79] Z. Zhao, W. C. Lee, Y. Shin, and K.-B. Song, “An optimal power scheduling method for

demand response in home energy management system,”Smart Grid, IEEE Transactions

on, vol. 4, no. 3, pp. 1391–1400, 2013.

[80] F. De Angelis, M. Boaro, D. Fuselli, S. Squartini, F. Piazza, and Q. Wei, “Optimal home

energy management under dynamic electrical and thermal constraints,”Industrial Infor-

matics, IEEE Transactions on, vol. 9, no. 3, pp. 1518–1527, 2013.

[81] T. Huang and D. Liu, “Residential energy system controland management using adap-

tive dynamic programming,” inNeural Networks (IJCNN), The 2011 International Joint


[82] A. A. Khan and H. T. Mouftah, “Web services for indoor energy management in a smart

grid environment,” inPersonal Indoor and Mobile Radio Communications (PIMRC),

2011 IEEE 22nd International Symposium on, pp. 1036–1040, IEEE, 2011.

[83] N. S. Nafi, K. Ahmed, M. Datta, and M. Gregory, “A novel zigbee based pilot protection

scheme for smart distribution grid,” inTelecommunication Networks and Applications

Conference (ATNAC), 2014 Australasian, pp. 146–151, IEEE, 2014.

[84] Z. Zhu, J. Tang, S. Lambotharan, W. H. Chin, and Z. Fan, “An integer linear programming

and game theory based optimization for demand-side management in smart grid,” in

GLOBECOM Workshops (GC Wkshps), 2011 IEEE, pp. 1205–1210, IEEE, 2011.

[85] Z. Zhu, J. Tang, S. Lambotharan, W. H. Chin, and Z. Fan, “An integer linear program-

ming based optimization for home demand-side management insmart grid,” inInnova-

tive Smart Grid Technologies (ISGT), 2012 IEEE PES, pp. 1–5, IEEE, 2012.

163

BIBLIOGRAPHY

[86] T. Hubert and S. Grijalva, “Realizing smart grid benefits requires energy optimization al-

gorithms at residential level,” inInnovative Smart Grid Technologies (ISGT), 2011 IEEE

PES, pp. 1–8, IEEE, 2011.

[87] J. Leithon, S. Sun, and T. J. Lim, “Energy management strategies for base stations pow-

ered by the smart grid,” inGlobal Communications Conference (GLOBECOM), 2013

IEEE, pp. 2635–2640, IEEE, 2013.

[88] G. Xydis and C. Koroneos, “A linear programming approach for the optimal planning of

a future energy system. potential contribution of energy recovery from municipal solid

wastes,”Renewable and Sustainable Energy Reviews, vol. 16, no. 1, pp. 369–378, 2012.

[89] Institute of Pubilc Polices, “State of the economy: Emerging from the crisis economic,

second annual report,”Beconhouse National Universty, Lahore, 2009.

[90] Power Wing, Water and Power Development Authority, “Power index,” htt p :

//www.wapda.gov.pk/htmls/power− index.html. (Last accessed in Dec. 2014), 2014.

[91] Population Clock at Population Census Organization, GoP, “Population census,”

http://www.census.gov.pk/index.php, [accessed Nov 2012].

[92] Hydrocarbon Development Institute of Pakistan (HDIP), “Pakistan energy yearbook,”

Ministry of Petroleum and Natural Resources, Government ofPakistan, 2012.

[93] A. Malik, “Power crisis in pakistan: A crisis in governance?,” tech. rep., Pakistan Insti-

tute of Development Economics, 2012.

[94] Ministry of Finance GoP, “Executive summary,” htt p :

//www. f inance.gov.pk/survey/chapter12/ExecutiveSummary.pd f, (Last accessed

May 2013).

[95] N. Kalair, “Water conflicts and hydroelectricity in south asia: The indus water treaty, a

review article,” Global Research, March 21, 2012.

[96] National Transmission and Despatch Company Ltd.(NTDC), Pakistan, “Electric-

ity demand forecast based on multiple regression analysis,2011 to 2035,” htt p :

//www.ntdc.com.pk/EDF20112035.pd f.

164

BIBLIOGRAPHY

[97] National Transmission and Despatch Company Ltd.(NTDC), Pakistan, “Power system

statistics,”htt p : //www.ppib.gov.pk/Power%20System%20Statistics.pd f.

[98] M. M. Chaudhry, S. Mohammad, and A. B. Sufi, “Role of waterresources development in

the economy of pakistan,”http://pecongress.org.pk/images/upload/books/Paper658.pdf,

[accessed May 2013].

[99] Water Resource Development Council Pakistan,www.wrdc.com.pk, (Last accessed May

2013).

[100] Private Power and Infrastructure Board Pakistan, “Hydro power resources of pakistan,”

htt p : //www.ppib.gov.pk/HYDRO.pd f, (Last accessed May 2013).

[101] State Bank of Pakistan, “Annual report 2010-2011,thestate of pakistan’s economy,”

[102] National Transmission and Despatch Company Ltd.(NTDC) Pakistan, “Data reference

book, 4th issue 2011,” 2011.

[103] S. S. Ali and S. Badar, “Dynamics of circular debt in pakistan and its resolution,”The

Lahore Journal of Economics, vol. 15, no. 2010, pp. 61–74.

[104] A. Cohen, L. Curtis, and O. Graham, “The proposed iran-pakistan-india gas pipeline: An

unacceptable risk to regional security,”Heritage Foundation, Washington, DC, vol. 30,

2008.

[105] M. Anas, N. Javaid, A. Mahmood, S. Raza, U. Qasim, and Z.A. Khan, “Minimizing

electricity theft using smart meters in ami,” inP2P, Parallel, Grid, Cloud and Internet

Computing (3PGCIC), 2012 Seventh International Conference on, pp. 176–182, IEEE,

2012.

[106] M. A. Choudhary, N. Khan, A. Ali, and A. Abbas, “Achievability of pakistan’s 2030 elec-

tricity generation goals established under medium term development framework (mtdf):

Validation using time series models and error decomposition technique,” inEnergy 2030

Conference, 2008. ENERGY 2008. IEEE, pp. 1–10, IEEE, 2008.

[107] REN21 Secretariat Paris, “Renewables 2012 global status report,” htt p :

//www.ren21.net/Portals/0/documents/Resources/%20GSR2012%20highres.pd f,

accessed May 2013.

165

BIBLIOGRAPHY

[108] Energy Expert Group, “Economic advisory council, ministry of finance pakistan,”Inte-

grated energy plan 2009-2022.

[109] M. Asif, “Sustainable energy options for pakistan,”Renewable and Sustainable Energy

Reviews, vol. 13, no. 4, pp. 903–909, 2009.

[110] N. A. Zaigham and Z. A. Nayyar, “Renewable hot dry rock geothermal energy source

and its potential in pakistan,”Renewable and Sustainable Energy Reviews, vol. 14, no. 3,

pp. 1124–1129, 2010.

[111] A. W. Bhutto and S. Karim, “Coal gasification for sustainable development of the energy

sector in pakistan,”Energy for Sustainable Development, vol. 9, no. 4, pp. 60–67, 2005.

[112] A. Zafar, “Study of smart grid system operation in pakistan,”MS Dessertation in Elec-

trical Engineering at COMSATS Institute of Information Technology Islamabad, 2012.

[113] Micro Tech, “pilot project of amr meters at mepco,” htt p :

//www.mtilimited.com/newsroom/mepco−meters.php(accessedinDec.2014), 2014.

[114] Govt. of Pakistan, “Prepaid metering network,” htt p :

//dunyanews.tv/index.php/en/Pakistan/259547−Electricity− to−be− provided−

like− prepaid−mobile− pho(accessedinDec.2014), 2014.

[115] Q. Zhang, Y. Sun, and Z. Cui, “Application and analysisof zigbee technology for smart

grid,” in Computer and Information Application (ICCIA), 2010 International Conference

on, pp. 171–174, IEEE, 2010.

[116] P. Rengaraju, C.-H. Lung, and A. Srinivasan, “Communication requirements and analysis

of distribution networks using wimax technology for smart grids,” in Wireless Communi-

cations and Mobile Computing Conference (IWCMC), 2012 8th International, pp. 666–

670, IEEE, 2012.

[117] A. Patel, J. Aparicio, N. Tas, M. Loiacono, and J. Rosca, “Assessing communications

technology options for smart grid applications,” inSmart Grid Communications (Smart-

GridComm), 2011 IEEE International Conference on, pp. 126–131, IEEE, 2011.

[118] X. S. Shen, “Empowering the smart grid with wireless technologies [editor’s note],”Net-

work, IEEE, vol. 26, no. 3, pp. 2–3, 2012.

166

BIBLIOGRAPHY

[119] S. S. S. R. Depuru, L. Wang, and V. Devabhaktuni, “Smartmeters for power grid:

Challenges, issues, advantages and status,”Renewable and sustainable energy reviews,

vol. 15, no. 6, pp. 2736–2742, 2011.

[120] F. Baig, A. Mahmood, N. Javaid, S. Razzaq, N. Khan, and Z. Saleem, “Smart home

energy management system for monitoring and scheduling of home appliances using

zigbee,”J. Basic. Appl. Sci. Res, vol. 3, no. 5, pp. 880–891, 2013.

[121]

[122] M. Ullah, N. Javaid, I. Khan, A. Mahmood, and M. Farooq,“Residential energy con-

sumption controlling techniques to enable autonomous demand side management in fu-

ture smart grid communications,” inBroadband and Wireless Computing, Communica-

tion and Applications (BWCCA), 2013 Eighth International Conference on, pp. 545–550,

IEEE, 2013.

[123] F. Bouhafs, M. Mackay, and M. Merabti, “Links to the future: communication require-

ments and challenges in the smart grid,”Power and Energy Magazine, IEEE, vol. 10,

no. 1, pp. 24–32, 2012.

[124] K.-l. Zhou, S.-l. Yang, and C. Shen, “A review of electric load classification in smart grid

environment,”Renewable and Sustainable Energy Reviews, vol. 24, pp. 103–110, 2013.

[125] H. Su, M. Qiu, and H. Wang, “Secure wireless communication system for smart grid

with rechargeable electric vehicles,”Communications Magazine, IEEE, vol. 50, no. 8,

pp. 62–68, 2012.

[126] J.-S. Lee, Y.-W. Su, and C.-C. Shen, “A comparative study of wireless protocols: Blue-

tooth, uwb, zigbee, and wi-fi,” inIndustrial Electronics Society, 2007. IECON 2007. 33rd

Annual Conference of the IEEE, pp. 46–51, IEEE, 2007.

[127] M. Rafiei, S. M. Elmi, and A. Zare, “Wireless communication protocols for smart meter-

ing applications in power distribution networks,” inElectrical Power Distribution Net-

works (EPDC), 2012 Proceedings of 17th Conference on, pp. 1–5, IEEE, 2012.

[128] V. C. Gungor, D. Sahin, T. Kocak, S. Ergut, C. Buccella,C. Cecati, and G. P. Hancke,

“Smart grid technologies: communication technologies andstandards,”Industrial infor-

matics, IEEE transactions on, vol. 7, no. 4, pp. 529–539, 2011.

167

BIBLIOGRAPHY

[129] W. Wang, Y. Xu, and M. Khanna, “A survey on the communication architectures in smart

grid,” Computer Networks, vol. 55, no. 15, pp. 3604–3629, 2011.

[130] P. P. Parikh, M. G. Kanabar, and T. S. Sidhu, “Opportunities and challenges of wireless

communication technologies for smart grid applications,”in Power and Energy Society

General Meeting, 2010 IEEE, pp. 1–7, IEEE, 2010.

[131] R. Riggio, T. Rasheed, S. Testi, F. Granelli, and I. Chlamtac, “Interference and traffic

aware channel assignment in wifi-based wireless mesh networks,” Ad Hoc Networks,

vol. 9, no. 5, pp. 864–875, 2011.

[132] L. Li, H. Xiaoguang, C. Ke, and H. Ketai, “The applications of wifi-based wireless sensor

network in internet of things and smart grid,” inIndustrial Electronics and Applications

(ICIEA), 2011 6th IEEE Conference on, pp. 789–793, IEEE, 2011.

[133] A. C. Davies, “An overview of bluetooth wireless technology tm and some competing lan

standards,” inCircuits and Systems for Communications, 2002. Proceedings. ICCSC’02.

1st IEEE International Conference on, pp. 206–211, IEEE, 2002.

[134] C. Bisdikian, “An overview of the bluetooth wireless technology,”IEEE Commun Mag,

vol. 39, no. 12, pp. 86–94, 2001.

[135] C. Gehrmann, J. Persson, and B. Smeets,Bluetooth security. Artech House,

htt p : //www.artechhouse.com/Main/BillingCountry.aspx?ahbRedirect =

1&pageurl = %2f International%2f Books%2f Bluetooth− Security− 869.aspx,

2004.

[136] H. Wang, M. Xi, J. Liu, and C. Chen, “Transmitting ipv6 packets over bluetooth low

energy based on bluez,” inAdvanced Communication Technology (ICACT), 2013 15th


[137] B. Dorge and T. Scheffler, “Using ipv6 and 6lowpan for home automation networks,” in

Consumer Electronics-Berlin (ICCE-Berlin), 2011 IEEE International Conference on,

pp. 44–47, IEEE, 2011.

[138] A. P. Castellani, G. Ministeri, M. Rotoloni, L. Vangelista, and M. Zorzi, “Interoperable

and globally interconnected smart grid using ipv6 and 6lowpan,” in Communications

(ICC), 2012 IEEE International Conference on, pp. 6473–6478, IEEE, 2012.

168

BIBLIOGRAPHY

[139] G. K. Ee, C. K. Ng, N. K. Noordin, and B. M. Ali, “A review of 6lowpan routing proto-

cols,” Proceedings of the Asia-Pacific Advanced Network, vol. 30, pp. 71–81, 2010.

[140] J. He and X. Huang, “Increased interoperability: Evolution of 6lowpan-based web appli-

cation,” inBroadband Network and Multimedia Technology (IC-BNMT), 2011 4th IEEE


[141] S. Raza, S. Duquennoy, T. Chung, T. Voigt, and U. Roedig, “Securing communication

in 6lowpan with compressed ipsec,” inDistributed Computing in Sensor Systems and

Workshops (DCOSS), 2011 International Conference on, pp. 1–8, IEEE, 2011.

[142] S. Raza, S. Duquennoy, J. Hoglund, U. Roedig, and T. Voigt, “Secure communication for

the internet of thingsa comparison of link-layer security and ipsec for 6lowpan,”Security

and Communication Networks, 2012.

[143] Z. Pei, Z. Deng, B. Yang, and X. Cheng, “Application-oriented wireless sensor network

communication protocols and hardware platforms: A survey,” in Industrial Technology,

2008. ICIT 2008. IEEE International Conference on, pp. 1–6, IEEE, 2008.

[144] C. Gomez and J. Paradells, “Wireless home automation networks: A survey of architec-

tures and technologies,”IEEE Communications Magazine, vol. 48, no. 6, pp. 92–101,

2010.

[145] Z. Feng and Z. Yuexia, “Study on smart grid communications system based on new

generation wireless technology,” inElectronics, Communications and Control (ICECC),

2011 International Conference on, pp. 1673–1678, IEEE, 2011.

[146] P. Cheng, L. Wang, B. Zhen, and S. Wang, “Feasibility study of applying lte to smart

grid,” in Smart Grid Modeling and Simulation (SGMS), 2011 IEEE First International

Workshop on, pp. 108–113, IEEE, 2011.

[147] J. Du and M. Qian, “Research and application on lte technology in smart grids,” inCom-

munications and Networking in China (CHINACOM), 2012 7th International ICST Con-

ference on, pp. 76–80, IEEE, 2012.

[148]

169

BIBLIOGRAPHY

[149] S. Chakraborty, T. Ito, and T. Senjyu, “Smart pricing scheme: A multi-layered scoring

rule application,”Expert Systems with Applications, vol. 41, no. 8, pp. 3726–3735, 2014.

[150] N. Nazar, M. Abdullah, M. Hassan, and F. Hussin, “Time-based electricity pricing for

demand response implementation in monopolized electricity market,” inResearch and

Development (SCOReD), 2012 IEEE Student Conference on, pp. 178–181, IEEE, 2012.

[151] S. Braithwait, D. Hansen, and M. OSheasy, “Retail electricity pricing and rate design in

evolving markets,”Edison Electric Institute, pp. 1–57, 2007.

[152] A. A. Jose and A. Pahwa, “Economic evaluation of small wind generation ownership un-

der different electricity pricing scenarios,” inNorth American Power Symposium (NAPS),

2010, pp. 1–4, IEEE, 2010.

[153] S. Shao, T. Zhang, M. Pipattanasomporn, and S. Rahman,“Impact of tou rates on distri-

bution load shapes in a smart grid with phev penetration,” inTransmission and Distribu-

tion Conference and Exposition, 2010 IEEE PES, pp. 1–6, IEEE, 2010.

[154] M. Kii, K. Sakamoto, Y. Hangai, and K. Doi, “The effectsof critical peak pricing for elec-

tricity demand management on home-based trip generation,”IATSS Research, vol. 37,

no. 2, pp. 89–97, 2014.

[155] S. Widergren, C. Marinovici, T. Berliner, and A. Graves, “Real-time pricing demand

response in operations,” inPower and Energy Society General Meeting, 2012 IEEE,

pp. 1–5, IEEE, 2012.

[156] L. P. Qian, Y. J. A. Zhang, J. Huang, and Y. Wu, “Demand response management via

real-time electricity price control in smart grids,”Selected Areas in Communications,

IEEE Journal on, vol. 31, no. 7, pp. 1268–1280, 2013.

[157] D. Becker and T. Saxton, “Cim standard for model exchange between planning and op-

erations,” inPower and Energy Society General Meeting-Conversion and Delivery of

Electrical Energy in the 21st Century, 2008 IEEE, pp. 1–5, IEEE, 2008.

[158] S. Meiling, T. Steinbach, T. C. Schmidt, and M. Wahlisch, “A scalable communica-

tion infrastructure for smart grid applications using multicast over public networks,” in

Proceedings of the 28th Annual ACM Symposium on Applied Computing, pp. 690–694,

ACM, 2013.

170

BIBLIOGRAPHY

[159] Y.-J. Kim, J. Lee, G. Atkinson, H. Kim, and M. Thottan, “Sedax: A scalable, resilient,

and secure platform for smart grid communications,”Selected Areas in Communications,

IEEE Journal on, vol. 30, no. 6, pp. 1119–1136, 2012.

[160] Z. Li, Q. Liang, X. Cheng, and T. Znati, “Outage performance evaluation over rayleigh

fading channels of wifi direct networks in smart grid application,” in Innovative Smart

Grid Technologies-Asia (ISGT Asia), 2012 IEEE, pp. 1–5, IEEE, 2012.

[161] D. Von Dollen, “Report to nist on the smart grid interoperability standards roadmap,”

Electric Power Research Institute (EPRI) and National Institute of Standards and Tech-

nology, 2009.

[162] L. Berrio and C. Zuluaga, “Concepts, standards and communication technologies in

smart grid,” inCircuits and Systems (CWCAS), 2012 IEEE 4th Colombian Workshop

on, pp. 1–6, IEEE, 2012.

[163] E. Ferro and F. Potorti, “Bluetooth and wi-fi wireless protocols: a survey and a compari-

son,”Wireless Communications, IEEE, vol. 12, no. 1, pp. 12–26, 2005.

[164] “Knapsack problem,”htt p: //en.wikipedia.org/wiki/Knapsackproblem, (Last accessed

December 2014).

[165] “Knapsack problem,”htt p : //micsymposium.org/mics2005/papers/paper102.pd f,

(Last accessed December 2014).

[166] N. Kumaraguruparan, H. Sivaramakrishnan, and S. S. Sapatnekar, “Residential task

scheduling under dynamic pricing using the multiple knapsack method,” inInnovative

Smart Grid Technologies (ISGT), 2012 IEEE PES, pp. 1–6, IEEE, 2012.

[167] “Knapsack problem,”htt p : //www.or.deis.unibo.it/kp/Chapter6.pd f, (Last accessed

December 2014).

[168] H. M. Lugo-Cordero, A. Fuentes-Rivera, R. K. Guha, andE. I. Ortiz-Rivera, “Particle

swarm optimization for load balancing in green smart homes,” in Evolutionary Compu-

tation (CEC), 2011 IEEE Congress on, pp. 715–720, IEEE, 2011.

[169] Y. Shi and R. Eberhart, “A modified particle swarm optimizer,” in Evolutionary Com-

putation Proceedings, 1998. IEEE World Congress on Computational Intelligence., The

1998 IEEE International Conference on, pp. 69–73, IEEE, 1998.

171

BIBLIOGRAPHY

[170] T. Gonsalves and A. Egashira, “Parallel swarms oriented particle swarm optimization,”

Applied Computational Intelligence and Soft Computing, vol. 2013, p. 14, 2013.

[171] J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Neural Networks, 1995.

Proceedings., IEEE International Conference on, vol. 4, pp. 1942–1948, IEEE, 1995.

[172] L. Xue, J. Yin, Z. Ji, and L. Jiang, “A particle swarm optimization for hidden markov

model training,” inSignal Processing, 2006 8th International Conference on, vol. 1,

IEEE, 2006.

[173] R. M. Ramadan and R. F. Abdel-Kader, “Face recognitionusing particle swarm

optimization-based selected features,”International Journal of Signal Processing, Im-

age Processing and Pattern Recognition, vol. 2, no. 2, pp. 51–65, 2009.

[174] W. Yang, X. Yu, J. Liu, C. Li, and W. Wan, “Audio feature optimization based on the

pso and attribute importance,” inAudio Language and Image Processing (ICALIP), 2010


[175] H. Gao, W. Xu, J. Sun, and Y. Tang, “Multilevel thresholding for image segmentation

through an improved quantum-behaved particle swarm algorithm,” Instrumentation and

Measurement, IEEE Transactions on, vol. 59, no. 4, pp. 934–946, 2010.

[176] N. Ghosh, S. Nanda, and S. Ghosh, “A quantitative approach towards detection of an

optimal attack path in a wireless network using modified pso technique,” inCommunica-

tion Systems and Networks and Workshops, 2009. COMSNETS 2009. First International,

pp. 1–10, IEEE, 2009.

[177] Z.-L. Gaing, “Particle swarm optimization to solvingthe economic dispatch consider-

ing the generator constraints,”Power Systems, IEEE Transactions on, vol. 18, no. 3,

pp. 1187–1195, 2003.

[178] V. Karthikeyan, S. Senthilkumar, and V. Vijayalakshmi, “A new approach to the solu-

tion of economic dispatch using particle swarm optimization with simulated annealing,”

International Journal on Computational Sciences & Applications (IJCSA) Vol.3, No.3,

June 2013, 2013.

[179] R. Rahmani, M. F. Othman, R. Yusof, and M. Khalid, “Solving economic dispatch

problem using particle swarm optimization by an evolutionary technique for initializing

172

BIBLIOGRAPHY

particles,”Journal of Theoretical and Applied Information Technology, vol. 46, no. 2,

pp. 526–536, 2012.

[180] T. de Fatima Araujo and W. Uturbey, “Generation and demand management via particle

swarm optimization,” inEnvironment and Electrical Engineering (EEEIC), 2011 10th


[181] F. Farheen, M. Ansari, and N. Kardam, “Implementationof particle swarm optimization

for dynamic economic load dispatch problem,” inEnergy Efficient Technologies for Sus-

tainability (ICEETS), 2013 International Conference on, pp. 1273–1278, IEEE, 2013.

[182] G. Rashed, H. Shaheen, and S. Cheng, “Optimal locationand parameter settings of mul-

tiple tcscs for increasing power system loadability based on ga and pso techniques,”

in Natural Computation, 2007. ICNC 2007. Third InternationalConference on, vol. 4,

pp. 335–344, IEEE, 2007.

[183] Q. Xie and Z. Liu, “Efficiency evaluation for collaborative design based on ga-bp algo-

rithm,” in Computer Supported Cooperative Work in Design, 2008. CSCWD2008. 12th


[184] L. Jia, Z. Yu, M. C. Murphy-Hoye, A. Pratt, E. G. Piccioli, and L. Tong, “Multi-scale

stochastic optimization for home energy management,” inComputational Advances in

Multi-Sensor Adaptive Processing (CAMSAP), 2011 4th IEEE International Workshop

on, pp. 113–116, IEEE, 2011.

[185] T. Dang and K. Ringland, “Optimal load scheduling for residential renewable energy

integration,” inSmart Grid Communications (SmartGridComm), 2012 IEEE Third Inter-

national Conference on, pp. 516–521, IEEE, 2012.

[186] T. Logenthiran, D. Srinivasan, and T. Z. Shun, “Demandside management in smart grid

using heuristic optimization,”Smart Grid, IEEE Transactions on, vol. 3, no. 3, pp. 1244–

1252, 2012.

[187] P. Samadi, H. Mohsenian-Rad, R. Schober, and V. W. Wong, “Advanced demand side

management for the future smart grid using mechanism design,” Smart Grid, IEEE Trans-


173

BIBLIOGRAPHY

[188] R. Pratt, C. Conner, E. Richman, K. Ritland, W. Sandusky, and M. Taylor, “Description

of electric energy use in single-family residences in the pacific northwest,”DOE/BP,

vol. 13795, p. 21, 1989.

[189] M. Erol-Kantarci and H. T. Mouftah, “Tou-aware energymanagement and wireless sen-

sor networks for reducing peak load in smart grids,” inVehicular Technology Conference

Fall (VTC 2010-Fall), 2010 IEEE 72nd, pp. 1–5, IEEE, 2010.

[190] R. J. Gran,Numerical computing with simulink, Volume 1: Creating simulations, vol. 1.

SIAM, 2007.

[191] Mathworks, “House thermal model,”htt p : //www.mathworks.com/help/simulink/

examples/thermal−model−o f −a−house.html, (Last accessed August 2014).

[192] L. Paull, D. MacKay, H. Li, and L. Chang, “Awater heatermodel for increased power sys-

tem efficiency,” inElectrical and Computer Engineering, 2009. CCECE’09. Canadian


[193] J. D. Lutz, X. Liu, J. E. McMahon, C. Dunham, L. J. Shown,and Q. T. McCure, “Mod-

eling patterns of hot water use in households,”Lawrence Berkeley National Laboratory,

1996.

[194] Wikipedia, “Carbon print,” htt p : //en.wikipedia.org/wiki/Carbonfoot print2014),

(Last accessed August 2014).

[195] Wikipedia, “Carbon off-setting,” htt p : //www.carbonneutral.com/resource−

hub/carbon−o f f setting−explained, (Last accessed August 2014).

[196] Wikipedia, “Green power,”htt p : //www.epa.gov/greenpower/pubs/calcmeth.htm,

(Last accessed August 2014).

[197] R. Mohammad, F.-F. Mahmud, and Z. Hamidreza, “Home energy management incor-

porating operational priority of appliances,”Electrical Power and Energy Systems 74

(2016), pp. 286–292, 2016.

[198] B. R. Mohammad, J. Nadeem, and A. Ashfaq, “An efficient power scheduling scheme

for residential load management in smart homes,”Applied Sciences 74 (2016), pp. 1134–

1163, 2015.

174

Documents

Simulation Study for Optimized Demand Side Management in ......It is certiﬁed that Anzar Mahmood (Registration No. CIIT/SP 11-PEE-002/ISB) has carried out all the work related to