An Innovative Heuristic Algorithm for IoT-enabled

An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing Countries

By

Bilal Hussain CIIT/FA/ISB

PhD Thesis In Electrical Engineering

COMSATS University Islamabad, Islamabad, Pakistan

Fall, 2018

COMSATS University Islamabad

An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing Countries

A Thesis Presented to

COMSATS University Islamabad, Islamabad, Pakistan

In partial fulfillment of the requirement for the degree of

PhD (Electrical Engineering)

By


Fall, 2018 ii

An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing

A Post Graduate Thesis submitted to the Department of Computer Science as partial fulfilment of the requirement for the award of Degree of PhD (Electrical Engineering).

Name

Registration Number


Supervisor:

Dr. Qadeer Ul Hassan, Chief Engineer/Associate Head, Department of Electrical Engineering, COMSATS University, Islamabad, Islamabad Campus.

Co-Supervisor:

Dr. Nadeem Javaid, Associate Professor, Department of Computer Science, COMSATS University, Islamabad, Islamabad Campus. iii

Certificate of Approval This is to certify that the research work presented in this thesis, entitled “An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing Countries” was conducted by Mr. Bilal Hussain under the supervision of Dr. Qadeer Ul Hassan. No part of this thesis has been submitted anywhere else for any other degree. This thesis is submitted to the Department of Computer Science, COMSATS University Islamabad, Islamabad, in the partial fulfillment of the requirement for the degree of Doctor of Philosophy in the field of Electrical Engineering.

Bilal Hussain

Signature:

Examinations Committe:

................................................... External Examiner 1:

..................................................... External Examinar 2:

(Designation and Office Address)

(Designation and Office Address)

................................................... Dr. Qadeer Ul Hassan, Supervisor, Department of Electrical Engineering, CUI, Islamabad.

..................................................... Dr. Nadeem Javaid, Co-Supervisor, Department of Computer Science, CUI, Islamabad.

................................................... Prof. Dr. M. Junaid Mughal, Chairperson/HoD, Department of Electrical Engineering, CUI, Islamabad.

..................................................... Prof. Dr. Shahid A. Khan Dean, Department of Electrical Engineering, CUI, Islamabad.

iv

Author’s Declaration I Bilal Hussain, CIIT/FA/ISB hereby state that my PhD thesis titled “An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing Countries” is my own work and has not been submitted previously by me for taking any degree from this University i.e., COMSATS University Islamabad or anywhere else in the country/ world. At any time if my statement is found to be incorrect even after I graduate the University has the right to withdraw my PhD degree.

Date:

Signature of the student:


v

Plagiarism Undertaking I solemnly declare that research work presented in this thesis titled, “An Innovative Heuristic Algorithm for IoT-enabled Smart Homes for Developing Countries” is solely my research work with no significant contribution from any other person. Small contribution/ help wherever taken has been duly acknowledged and that complete thesis has been written by me. I understand the zero tolarnace policy of HEC and COMSATS University Islamabad towards plagiarism. Therefore, I as an author of the above titled thesis declare that no portion of my thesis has been plagiarized and any material used as reference is properly referred/ cited. I undertake if I am found guilty of any formal plagiarism in the above titled thesis even after award of PhD degree, the University reserves the right to withdraw/ revoke my PhD degree and that HEC and the university has the right to publish my name on the HEC/ university website on which names of students are placed who submitted plagiarized thesis.

Signature of the student:

Date:


vi

Certificate It is certified that Bilal Hussain, CIIT/FA15/ISB has carried out all the work related to this thesis under my supervision at the Department of Computer Science, COMSATS University Islamabad, Islamabad and the work fulfills the requirement for award of PhD degree.

Supervisor:

Date:

Dr. Qadeer Ul Hassan, Chief Engineer/Associate Head, Department of Electrical Engineering, COMSATS University Islamabad, Islamabad. Co-Supervisor:

Dr. Nadeem Javaid, Associate Professor, Department of Computer Science, COMSATS University Islamabad, Islamabad.

Head of Department:

Prof. Dr. M. Junaid Mughal, Professor, Department of Electrical Engineering, COMSATS University Islamabad, Islamabad. vii

DEDICATION

Dedicated to my teachers

viii

ACKNOWLEDGEMENT -

ix

ABSTRACT This dissertation explores and identifies that home energy management systems (HEMSs) are used to implement demand side management in homes. Based on integration of renewable energy sources (RESs) and energy storage systems (ESSs), HEMS operation (HEMO) is classified into demand response (DR) and DR synergized with RESs and ESSs optimal dispatch (DRSREOD). DR-based HEMO depends on shifting of the consumer load towards off-peak times. DRSREOD-based HEMS benefits the consumer and the utility by reducing the cost of generation, reducing energy bills, minimizing green house gas (GHG) emissions, achieving overall energy savings and increasing energy sustainability. The contributions in this dissertation are three fold. First, this dissertation reviews the most recent literature on various models for DRSREOD-based HEMO. The reviewed models for HEMO are classified into dichotomous approaches as DR versus DRSREOD-based individual versus coordinated, deterministic versus stochastic, single-objective versus multi-objective and conventional techniques versus advanced heuristics-based. In addition, the tradeoffs among the dichotomous approaches, challenges pertinent to coordination and eminent issues related to standardization requirements for modeling home appliances (HAs) are investigated. Second, an improved algorithm for a DRSREOD-based HEMS is then proposed in this dissertation. This heuristic-based algorithm considers DR, photovoltaic (PV) availability, the state of charge and charge/discharge rates of the storage battery and the sharing-based parallel operation of more than one power sources to supply the required load. The HEMS problem has been solved to minimize the cost of energy (CE) and time-based discomfort (T BD) with conflicting tradeoffs. The mixed scheduling of appliances (delayed scheduling for some appliances and advanced scheduling for others) is introduced to improve the CE and T BD performance parameters using an inclining block rate (IBR) pricing scheme. A set of optimized tradeoffs between CE and T BD has been computed to address multiobjectivity using a multi-objective genetic algorithm with pareto optimization to perform the tradeoff analysis and to enable consumers to select the most feasible solution. Third, a drastically rising demand of electricity has forced a number of utilities in developing countries to impose large-scale load shedding (LS). A HEMS based on DRSREOD integrated with an LS-compensating dispatchable generator (LDG) (DRSREODLDG) ensures an uninterrupted supply of power for the consumers subjected to LS. The LDG operation to compensate the interrupted supply of x

power during the LS hours; however, accompanies the release of GHGs emissions as well that need to be minimized to conserve the environment. A 3-step simulation based posteriori method is proposed to develop a scheme for eco-efficient operation of DRSREODLDG-based HEMS. The method provides the tradeoffs between the net cost of energy (CEnet) to be paid by the consumer, the T BD due to shifting of HAs to participate in the HEMS operation and minimal emissions (T EM iss) from the local LDG. At step-1, primary tradeoffs for CEnet, T BD and T EM iss are generated through a heuristic that takes into account PVs availability, the state of charge and the related rates for the storage system, mixed shifting of HAs, IBR, the sharing-based parallel operation of power sources, and selling of the renewable energy to the utility. At step-2, a constraint filter based on the average value of T EM iss is used to filter out the tradeoffs with extremely high values of T EM iss. At step-3, a constraint filter made up of an average surface fit for T EM iss is applied to screen out the tradeoffs with marginally high values of T EM iss. The selected solutions are classified for critical tradeoff analysis to enable the consumer by choosing the best option from a diverse set of eco-efficient tradeoffs between CEnet, T BD and T EM iss. Finally, this thesis focuses on decomposed-weighted-sum particle swarm optimization (DWS-PSO) approach which is proposed for optimal operations of price-driven DR (PDDR) and PDDR- synergized with the renewable and energy storage dispatch (PDDR-RED) based HEMSs. Simulation results show the effectiveness of all the proposed schemes in comparison to the previous schemes.

xi

Journal Publications 1 Hussain, Bilal, Qadeer Ul Hasan, Nadeem Javaid, Mohsen Guizani, Ahmad Almogren, and Atif Alamri. “An Innovative Heuristic Algorithm for IoT-Enabled Smart Homes for Developing Countries.” IEEE Access 6 (2018): 15550-15575. (IF=3.557). Download. 2 Hussain, Bilal, Nadeem Javaid, Qadeer Hasan, Sakeena Javaid, Asif Khan, and Shahzad Malik. “An Inventive Method for Eco-Efficient Operation of Home Energy Management Systems.” Energies 11, no. 11 (2018): 3091. (IF=2.676). Download. 3 Hussain, Bilal, Asif Khan, Nadeem Javaid, Qadeer-ul-Hasan, Shahzad A. Malik, Omar Ahmad, Amir Hanif Dar and Ahmad Kazmi. “A Weighted-sum PSO Algorithm for HEMS: A New Approach for the Design and Diversified Performance Analysis.” Submitted in Electronics (2018). (IF=2.110).

xii

Conference Proceedings 1 Bilal, Hussain, Nadeem Javaid, Qadeer-ul-Hasan and Asma Rafique. “An inventive method for eco-efficient operation of home energy management system.” Accepted in International Conference on Cyber Security and Computer Science (ICONCS), Karabuk University (KBU), Turkey, 2018. Download.

xiii

Contents Dedication

viii

Acknowledgements

ix

Abstract

x

Journal Publications

xii

Conference Proceedings

xiii

List of Abbreviations

xxvi

List of Symbols

xxxi

1 Introduction 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Requirement for DRSREOD-based HEMS . . . . . . . . . . . . . 1.2.1 Home appliances . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Renewable energy sources . . . . . . . . . . . . . . . . . . 1.2.3 Energy storage systems . . . . . . . . . . . . . . . . . . . . 1.2.4 HEMS controller . . . . . . . . . . . . . . . . . . . . . . . 1.2.5 Home area network . . . . . . . . . . . . . . . . . . . . . 1.2.6 Smart meter . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Models for HEMO . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 Demand response versus DRSREOD-based HEMO models 1.3.2 DA versus RT HEMO models . . . . . . . . . . . . . . . . 1.3.3 Deterministic versus stochastic HEMO models . . . . . . . 1.3.4 Single versus multi objective HEMO models . . . . . . . . 1.3.5 Individual versus co-ordinated HEMO models . . . . . . . 1.3.6 Conventional versus advanced heuristics-based techniques . 1.4 Energy challenges and opportunities . . . . . . . . . . . . . . . . . 1.4.1 Power shortfalls and load shedding . . . . . . . . . . . . . 1.4.2 Green house gas emissions . . . . . . . . . . . . . . . . . . xiv

. . . . . . . . . . . . . . . . . .

1 2 3 3 4 4 5 5 6 6 7 9 9 10 10 10 11 11 12

1.5

1.6

Dissertation problem statement, focus, and 1.5.1 Thesis problem statement . . . . . 1.5.2 Thesis focus . . . . . . . . . . . . . 1.5.3 Thesis contributions . . . . . . . . Thesis organization . . . . . . . . . . . . .

contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

2 Related work 2.1 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Components for HEMO . . . . . . . . . . . . . . . . . . . 2.2 Categories of HEMSs . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 DR-based HEMSs . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 DRSR-based HEMSs . . . . . . . . . . . . . . . . . . . . . 2.2.3 DRSREOD-based HEMSs . . . . . . . . . . . . . . . . . . 2.2.4 DRSREOD-based HEMSs with optimally sized DGs to cope with LS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Emissions reduction . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Emissions reduction using DR-based HEMSs . . . . . . . . 2.3.2 Emissions reduction using DRSREOD-based HEMSs . . . 2.3.3 Emissions reduction in MGs . . . . . . . . . . . . . . . . . 2.3.4 Emissions reduction in DRSREODLDG-based HEMS . . . 2.4 Conclusion of the Chapter . . . . . . . . . . . . . . . . . . . . . . 3 Survey 3.1 Overview of HEMO modeling . . . . . . . . . . . . 3.1.1 HAs . . . . . . . . . . . . . . . . . . . . . . 3.1.2 RESs . . . . . . . . . . . . . . . . . . . . . . 3.1.3 ESSs . . . . . . . . . . . . . . . . . . . . . . 3.1.4 Fossil-based DGs . . . . . . . . . . . . . . . 3.1.5 Electric Tariffs . . . . . . . . . . . . . . . . 3.1.6 Main Objectives for HEMO Problem . . . . 3.1.6.1 Minimization of CE . . . . . . . . 3.1.6.2 Minimization of Discomfort Level . 3.1.6.3 Maximization of the Usage of RESs 3.1.6.4 Minimization of P AR . . . . . . . 3.1.6.5 Minimization of GHG emission . . 3.1.7 Constraints for HEMO Problem . . . . . . 3.1.7.1 HA Constraints . . . . . . . . . . . 3.1.7.2 Tariff Constraints . . . . . . . . . . 3.1.7.3 ESS Constraints . . . . . . . . . . xv

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . .

13 13 14 17 19

. . . . . .

21 22 22 32 32 33 34

. . . . . . .

36 37 40 41 42 42 43

. . . . . . . . . . . . . . . .

45 46 47 48 48 49 49 50 50 51 51 52 52 52 52 53 53

3.2 3.3

3.1.7.4 RDESs Constraints . . . . . . . . . . . . . . . . . 3.1.7.5 Constraint for the Balance of Energy . . . . . . . 3.1.7.6 Other Constraints . . . . . . . . . . . . . . . . . 3.1.8 Scheduling Resolution . . . . . . . . . . . . . . . . . . . . 3.1.9 Modeling Uncertainty of Data . . . . . . . . . . . . . . . 3.1.10 Coordination in HEMO . . . . . . . . . . . . . . . . . . . 3.1.11 Techniques to Solve Problem for HEMO . . . . . . . . . . 3.1.11.1 Conventional Techniques . . . . . . . . . . . . . . 3.1.11.2 Advanced Meta-heuristic Techniques . . . . . . . 3.1.11.3 Expert Systems . . . . . . . . . . . . . . . . . . . State of the Art in HEMO modeling . . . . . . . . . . . . . . . . Salient Issues and Challenges in HEMO . . . . . . . . . . . . . . . 3.3.1 Standardization Requirements for Response Classes . . . . 3.3.2 Formalizing Requirements to Handle Diversification of Control Parameters and Optimization Approaches . . . . . . . 3.3.3 Coordinated Approaches and Configurations for HEMO . . 3.3.4 Handling Multi-objectivity . . . . . . . . . . . . . . . . . 3.3.4.1 WSM . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4.2 Bounded Objective Method (BOM) . . . . . . . . 3.3.4.3 Physical Programming (PP) . . . . . . . . . . . . 3.3.4.4 Pareto Optimization (PO) . . . . . . . . . . . . . 3.3.4.5 Utility Function (UF) . . . . . . . . . . . . . . . 3.3.5 Modeling Uncertainty of Data . . . . . . . . . . . . . . . . 3.3.5.1 SOP . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5.2 RO . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5.3 CCO . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5.4 SDP . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5.5 Stochastic Fuzzy Optimization . . . . . . . . . . 3.3.5.6 MPC . . . . . . . . . . . . . . . . . . . . . . . . 3.3.6 Optimal Computational Techniques for HEMO Models . . 3.3.7 Computational Burden and Problem Complexity . . . . .

4 Models, problem formulation and proposed solutions 4.1 Proposed system model . . . . . . . . . . . . . . . . . . 4.1.1 HAs . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2 Electricity tariffs . . . . . . . . . . . . . . . . . 4.1.3 RESs . . . . . . . . . . . . . . . . . . . . . . . . 4.1.4 ESS . . . . . . . . . . . . . . . . . . . . . . . . xvi

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . .

53 53 53 53 54 54 55 55 57 59 59 98 99

. . . . . . . . . . . . . . . . .

99 103 105 106 107 107 107 108 108 109 109 110 110 110 111 112 114

. . . . .

119 120 121 122 124 125

4.2

4.3

4.4

4.5

4.6

4.1.5 LDG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Formulating the HEMS optimization problem . . . . . . . . . . . . 4.2.1 Objectives for the HEMS problem . . . . . . . . . . . . . . . 4.2.1.1 Minimization of CE/ CEnet . . . . . . . . . . . . . 4.2.1.2 Minimization of T BD for the consumer . . . . . . 4.2.1.3 Minimization of the peak load . . . . . . . . . . . . 4.2.1.4 Minimal size of the generator required to cope with LS . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1.5 Minimization of emissions . . . . . . . . . . . . . . 4.2.2 Constraints for the HEMS Problem . . . . . . . . . . . . . . 4.2.2.1 HA constraints . . . . . . . . . . . . . . . . . . . . 4.2.2.2 Tariff constraints . . . . . . . . . . . . . . . . . . . 4.2.2.3 SB constraints . . . . . . . . . . . . . . . . . . . . 4.2.2.4 Energy balance constraint . . . . . . . . . . . . . . 4.2.3 Optimization techniques for solving the HEMS problem . . 4.2.3.1 Conventional techniques . . . . . . . . . . . . . . . 4.2.3.2 Advanced heuristic techniques . . . . . . . . . . . . 4.2.4 Techniques for handling multi-objectivity in the HEMS problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Algorithms for a DR-based HEMS, a DRSREOD-based HEMS and optimal DG sizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Algorithm 1 for a DR-based HEMS with DS or MS . . . . . 4.3.2 Algorithm 2 for a DRSREOD-based HEMS with DS or MS . 4.3.3 Algorithm 3 for optimal DG sizing to cope with LS in a DRSREOD-based HEMS with MS . . . . . . . . . . . . . . DRSREODLDG-based HEMS . . . . . . . . . . . . . . . . . . . . . 4.4.1 Constrained filtration of tradeoffs to HEMS optimization . . 4.4.2 Regression based surface fitting techniques to develop ASCF Algorithms for eco-efficient TOs for DRSREODLDG-based HEMS . 4.5.1 Algorithm 4 to generate operating schemes and the primary TOs for DRSREODLDG-based HEMS (Step-1) . . . . . . . 4.5.2 Algorithm 5 for filtration mechanism to harness eco-efficient TOs for DRSREODLDG-based HEMS (Step-2 and Step-3) . DWS-PSO- based system model . . . . . . . . . . . . . . . . . . . . 4.6.1 HAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.2 DPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.2.1 ToUP . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.2.2 DA-RTP . . . . . . . . . . . . . . . . . . . . . . . . xvii

125 127 128 129 130 133 134 134 134 134 135 135 136 137 137 138 138 140 140 141 144 146 147 147 150 152 154 155 156 156 157 159

4.7

4.8

4.6.2.3 CPP . . . . . . . . . . . . . . . . . . . . . . . . . . 159 4.6.2.4 IBR . . . . . . . . . . . . . . . . . . . . . . . . . . 160 4.6.3 RESs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 4.6.4 SBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Formulating the HEMS optimization problem . . . . . . . . . . . . 162 4.7.1 Objectives for optimal HEMS operation . . . . . . . . . . . 164 4.7.1.1 Minimization of CE . . . . . . . . . . . . . . . . . 164 4.7.1.2 Minimization of T BD . . . . . . . . . . . . . . . . 165 4.7.1.3 Minimization of Ppeak . . . . . . . . . . . . . . . . 166 4.7.2 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 4.7.3 Meta-heuristic techniques to solve energy management problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 4.7.4 Handling of multi-objectivity in energy managment problems 168 Algorithms for a PDDR- and PDDR-RED- based HEMSs using DWS-PSO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 4.8.1 Algorithm 6 for a PDDR-based HEMS using DWS-PSO . . 169 4.8.2 Algorithm 7 for a PDDR-RED-based HEMS using DWS-PSO171

5 Simulation results and discussion 173 5.1 Simulations for the optimal operation of DR- and DRSREOD-based HEMSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 5.1.1 Simulation of a DR-based HEMS with DS . . . . . . . . . . 175 5.1.2 Simulation of a DR-based HEMS with MS . . . . . . . . . . 176 5.1.3 Simulation of a DRSREOD-based HEMS with DS . . . . . . 178 5.1.4 Simulation of a DRSREOD-based HEMS with MS . . . . . . 181 5.1.5 Critical analysis of HEMS scheduling (A-D) . . . . . . . . . 183 5.2 Simulations for DG sizing to cope with LS in a DRSREOD-based HEMS with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 5.2.1 Critical analysis of DG sizing to cope with LS in a DRSREODbased HEMS with MS . . . . . . . . . . . . . . . . . . . . . 186 5.3 DRSREODLDG based HEMS operation and the filtration mechanism192 5.3.1 Simulations for DRSREODLDG-based HEMS operation to compute primary TOs using algorithm 1 . . . . . . . . . . . 192 5.3.1.1 Trends for CEnet . . . . . . . . . . . . . . . . . . 196 5.3.1.2 Trends for T BD . . . . . . . . . . . . . . . . . . . 200 5.3.1.3 Trends for T EM iss . . . . . . . . . . . . . . . . . 202 5.3.1.4 Critical analysis for T EM iss and TOs for CEnet and T EM iss . . . . . . . . . . . . . . . . . . . . . 204 xviii

5.3.2

5.4

Simulations for filtration mechanism to harness eco-efficient TOs using algorithm 2 . . . . . . . . . . . . . . . . . . . . . 207 5.3.2.1 Simulation for filtration using AVCF (step-2) . . . 207 5.3.2.2 Simulation for filtration using ASCF (step-3) . . . 208 5.3.3 Critical tradeoff analysis of solutions for eco-efficient DRSREODLDGbased HEMS operation . . . . . . . . . . . . . . . . . . . . . 212 Simulations for PDDR- and PDDR-RED- based HEMS using DWSPSO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 5.4.1 Simulations for PDDR- based HEMS using DWS-PSO . . . 216 5.4.2 Simulation results discussion for PDDR- based HEMS . . . . 218 5.4.3 A DPA of DWS-PSO algorithm for PDDR- based HEMS . . 226 5.4.4 Simulations for PDDR-RED based HEMS using DWS-PSO . 229 5.4.5 Simulation results discussion for PDDR-RED based HEMS 232 5.4.6 A DPA of DWS-PSO algorithm for PDDR-RED- based HEMS242 5.4.7 Summary of DWS-PSO results . . . . . . . . . . . . . . . . 245

6 Conclusion and Future Work 249 6.1 Conclusions, Future Work . . . . . . . . . . . . . . . . . . . . . . . 250 6.1.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 6.1.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 7 References

259

xix

List of Figures 1.1 1.2

HEMS architecture for a smart home . . . . . . . . . . . . . . . . . 4 Un-even trends for T EM iss as related to CEnet and T BD . . . . 16

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10

Time range available for appliance operation in DS . . . . . . Maximum and actual values of the time delay in DS . . . . . . Time range available for appliance operation in AS . . . . . . Maximum and actual values of the advance-completion time in Fixed load profile for a smart home . . . . . . . . . . . . . . . 2S-ToUP scheme . . . . . . . . . . . . . . . . . . . . . . . . . 3S-ToUP scheme . . . . . . . . . . . . . . . . . . . . . . . . . DA-RTP scheme . . . . . . . . . . . . . . . . . . . . . . . . . CPP scheme (2-stage) . . . . . . . . . . . . . . . . . . . . . . PV energy profile . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

131 131 133 133 157 158 158 159 160 161

Two-stage ToU tariff scheme for August-September 2016 . . . . . . Tradeoff between CE and T BD(D) for a DR-based HEMS with DS Scheduled loads for a DR-based HEMS with DS . . . . . . . . . . . Tradeoff between CE and T BD(M ) for a DR-based HEMS with MS Scheduled loads for a DR-based HEMS with MS . . . . . . . . . . Tradeoff between CE and T BD(D) for a DRSREOD-based HEMS with DS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 Scheduled loads for a DRSREOD-based HEMS with DS . . . . . . 5.8 Load, PV, SB and energy parameters for a DRSREOD-based HEMS with DS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.9 Power constraints for P schd, P chg and P sold for a DRSREODbased HEMS with DS . . . . . . . . . . . . . . . . . . . . . . . . . 5.10 Power constraints for P grid, P pv and P dis for a DRSREOD-based HEMS with DS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.11 Tradeoff between CE and T BD(M ) for a DRSREOD-based HEMS with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.12 Scheduled loads for a DRSREOD-based HEMS with MS . . . . . .

175 176 176 177 177

5.1 5.2 5.3 5.4 5.5 5.6

xx

. . . . . . AS . . . . . . . . . . . .

178 179 179 180 180 181 181

5.13 Load, PV, SB and energy parameters for a DRSREOD-based HEMS with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 5.14 Power constraints for P schd, P chg and P sold for a DRSREODbased HEMS with MS . . . . . . . . . . . . . . . . . . . . . . . . . 182 5.15 Power constraints for P grid, P pv and P dis for a DRSREOD-based HEMS with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 5.16 Power supplied from the grid with and without a DRSREOD-based HEMS with an additional DG to cope with LS . . . . . . . . . . . . 185 5.17 Load, P pv, P chg, P dis, P gen and energy parameters for a DRSREODbased HEMS with a DG to cope with LS . . . . . . . . . . . . . . . 186 5.18 Power constraints for P schd, P chg, P sold and P dl for a DRSREODbased HEMS with an LS-compensating generator . . . . . . . . . . 186 5.19 Power constraints for P grid, P pv, P dis and P gen for a DRSREODbased HEMS with an LS-compensating generator . . . . . . . . . . 187 5.20 Generator size classification based on CE/T BD(M ) TOs . . . . . . 187 5.21 Generator size classification based on T BD(M )/CE TOs . . . . . . 191 5.22 Two-stage ToU tariff scheme . . . . . . . . . . . . . . . . . . . . . . 193 5.23 Primary tradeoff solutions with un-even surface for T EM iss generated through Algorithm 1. . . . . . . . . . . . . . . . . . . . . . . 194 5.24 Relations among primary TOs for CEnet, TBD and TEMiss using algorithm 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 5.25 Relation between CEnet and P dl for DRSREODLDG-based HEMS 198 5.26 Power and emission profiles for DRSREODLDG-based HEMS operation for solution-1 . . . . . . . . . . . . . . . . . . . . . . . . . . 198 5.27 Power and emission profiles for DRSREODLDG-based HEMS operation for solution-100 . . . . . . . . . . . . . . . . . . . . . . . . . 200 5.28 Relation between CEnet and T BD for a DRSREODLDG-based HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 5.29 Relation between T BD and T EM iss for a DRSREODLDG-based HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 5.30 Power and emission profiles for DRSREODLDG-based HEMS operation for solution-23 . . . . . . . . . . . . . . . . . . . . . . . . . 202 5.31 Power and emission profiles for DRSREODLDG-based HEMS operation for solution-27 . . . . . . . . . . . . . . . . . . . . . . . . . 203 5.32 Power and emission profiles for DRSREODLDG-based HEMS operation for solution-73 . . . . . . . . . . . . . . . . . . . . . . . . . 204 5.33 Variations in T EM iss with the related TOs for CEnet and T BD . 205

xxi

5.34 Application of AVCF to screen out the TOs with larger T EM iss values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.35 Eco-efficient solutions selected using average surface filtration based on algorithm 2 (Step-3) . . . . . . . . . . . . . . . . . . . . . . . . . 5.36 Relation between % Reduction in CEnet, TBD and TEMiss for eco-efficient TOs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.37 DWS-PSO for PDDR- based HEMS: A comparative performance for maximal reduction in CE using weights of (1,0) . . . . . . . . . 5.38 Load profile with maximal reduction in CE for 2S-ToUP (DS) . . . 5.39 Load profile with maximal reduction in CE for 2S-ToUP (MS) . . . 5.40 Load profile with maximal reduction in CE for 3S-ToUP (DS) . . . 5.41 Load profile with maximal reduction in CE for 3S-ToUP (MS) . . . 5.42 Load profile with maximal reduction in CE for DA-RTP (DS) . . . 5.43 Load profile with maximal reduction in CE for DA-RTP (MS) . . . 5.44 Load profile with maximal reduction in CE for CPP (DS) . . . . . . 5.45 Load profile with maximal reduction in CE for CPP (MS) . . . . . 5.46 Load profile with minimal value of TBD with weights = (0, 1) . . . 5.47 Tradeoffs between CE and TBD for PDDR-based HEMS . . . . . . 5.48 DWS-PSO for PDDR-RED- based HEMS: A comparative performance for maximal reduction in CEnet using weights of (1,0) . . . . 5.49 DWS-PSO for RED- based HEMS: A comparative performance at minimal TBD using weights of (0,1) . . . . . . . . . . . . . . . . . . 5.50 Load profile for minimal TBD using 2S-ToUP (Non-shifted) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . 5.51 Load profile for maximal reduction in CE using 2S-ToUP (DS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . 5.52 Load profile for maximal reduction in CE using 2S-ToUP (MS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . 5.53 Load profile for minimal TBD using CPP (Non-shifted) for PDDRRED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.54 Load profile for maximal reduction in CE using CPP(DS) for PDDRRED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.55 Load profile for maximal reduction in CE using CPP(MS) for PDDRRED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.56 Load profile for minimal TBD using 3S-ToUP (Non-shifted) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . 5.57 Load profile for maximal reduction in CE using 3S-ToUP (DS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . . xxii

208 212 214 217 219 220 221 222 223 224 224 225 227 227 230 231 233 234 235 236 237 237 238 239

5.58 Load profile for maximal reduction in CE using 3S-ToUP (MS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . 5.59 Load profile for minimal TBD using DA-RTP (Non-shifted) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . 5.60 Load profile for maximal reduction in CE using DA-RTP (DS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . 5.61 Load profile for maximal reduction in CE using DA-RTP (MS) for PDDR-RED- based HEMS . . . . . . . . . . . . . . . . . . . . . . 5.62 Tradeoffs between CEnet and TBD for PDDR-RED- based HEMS

xxiii

. 240 . 240 . 241 . 242 . 243

List of Tables 2.1 2.2 3.1 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17

Related work relevant to the proposed HEMS algorithms for DR, DRSREOD and generator sizing . . . . . . . . . . . . . . . . . . . . 38 Related work continued. . . . . . . . . . . . . . . . . . . . . . . . . 39 Dynamic Tariffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Deterministic) . . . . . . . . . . . . . Models for DR-based HEMO (Stochastic) . . . . . . . . . . . . . . Models for DRSREOD-based HEMO (Deterministic) . . . . . . . . Models for DRSREOD-based HEMO (Stochastic) . . . . . . . . . . Response Classifications of HDs . . . . . . . . . . . . . . . . . . . . Reduction in CE and P AR for DR-based HEMO . . . . . . . . . . Reduction in CE and P AR for DR-based HEMO (Not-incorporating NCAs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reduction in CE and P AR for DRSREOD-based HEMO . . . . . . Reduction in CE for I-HEMO versus COHEMO . . . . . . . . . . . Objectives Focused in HEMO . . . . . . . . . . . . . . . . . . . . . Methods to Handle Multi-objectivity in HEMO . . . . . . . . . . . Methods to Handle Stochasticity in Solving HEMO Problem . . . . Computational Techniques to Solve HEMO Problems . . . . . . . . Hybrid Computational Techniques to Solve HEMO Problems . . . . Computing Time with Related Parameters for Deterministic HEMO Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Computing Time with Related Parameters for Stochastic HEMO Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiv

50 60 61 62 63 64 65 66 67 67 72 85 100 101 102 104 105 106 108 111 113 114 116 117

4.1 4.2 4.3 4.4 4.5 4.6 4.7

SHAs and scheduling specifications for the DS and MS scenarios NSHAs considered for scheduling . . . . . . . . . . . . . . . . . PV system specifications . . . . . . . . . . . . . . . . . . . . . . SB and inverter specifications . . . . . . . . . . . . . . . . . . . LDG specifications . . . . . . . . . . . . . . . . . . . . . . . . . SHAs and scheduling specifications for the DS and MS scenarios PV unit, SB and inverter specifications [7] . . . . . . . . . . . .

5.1

. . . . . . .

. . . . . . .

122 123 125 126 126 157 162

Comparison of the maximum reductions in CE, the net bill, and the peak load for different HEMS categories . . . . . . . . . . . . . 183 5.2 Generator sizing based on CE/T BD(M ) TOs . . . . . . . . . . . . 188 5.3 PRIMARY TOs FOR DRSREODLDG-BASED HEMS USING ALGORITHM 1 (STEP-1) . . . . . . . . . . . . . . . . . . . . . . . . 194 5.4 TRADEOFFS ACHIEVED AFTER APPLYING AVCF BASED ON ALGORITHM 2 (STEP-2) . . . . . . . . . . . . . . . . . . . . 206 5.5 A COMPARISON OF PERFORMANCE PARAMETERS FOR POLYNOMIAL BASED ASCF . . . . . . . . . . . . . . . . . . . . . . . . 210 5.6 ECO-EFFICIENT SOLUTIONS FOR DRSREODLDG-BASED HEMS USING ALGORITHM 2 (STEP-3) . . . . . . . . . . . . . . . . . . 211 5.7 CRITICAL TRADEOFF ANALYSIS FOR ECO-EFFICIENT DRSREODLDGBASED HEMS OPERATION . . . . . . . . . . . . . . . . . . . . . 214 5.8 DWS-PSO FOR PDDR- BASED HEMS: REDUCTIONS IN CE, PEAK LOAD AND TBD FOR A DIVERSIFIED SET OF TPs . . 218 5.9 PERFORMANCE METRICS OF DWS-PSO- BASED ALGORITHM FOR PDDR-BASED HEMS . . . . . . . . . . . . . . . . . . . . . . 228 5.10 DWS-PSO FOR PDDR-RED- BASED HEMS: MAXIMUM REDUCTIONS IN CE, PEAK LOAD AND TBD FOR DIVERSIFIED TPs USING WEIGHTS (CE, TBD) = 1,0 . . . . . . . . . . . . . . 230 5.11 DWS-PSO FOR RED- BASED HEMS: REDUCTIONS IN CE AND PEAK LOAD FOR DIVERSIFIED TPs USING WEIGHTS (CE, TBD) = (0, 1) FOR MINIMAL TBD . . . . . . . . . . . . . . . . . 232 5.12 PERFORMANCE METRICS OF DWS-PSO FOR PDDR-REDBASED HEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

xxv

List of Abbreviations A AS Advanced scheduling AVCFAverage-value-based constraint filter ASCF Average-surface-based constraint filter ANN Artificial neural network ACO Ant colony optimization AC Air conditioner AS Advanced scheduling/advanced scheduled B BPSO Binary particle swarm optimzation BCO Bee colony optimization BB Branch-and-bound BOM Bounded Objective Method C CE Cost of energy purchased from grid CoHEMO Co-ordinated HEMO CPP Critical peak pricing CCO Chance-constrained optimization CC Centralized coordination CO Central operator COP Conventional optimization CP Convex programming CTP Cutting plane CCA Colonial competitive algorithm D DSM Demand-side management DRSREOD DR synergized with RESs and ESSs optimal dispatch DRSREODLDG - DRSREOD integrated with load shedding-compensating DG DA Day-ahead DAP DA pricing xxvi

DGs DS DP DE DC DPr DRSR DS DPA DPS DWS-PSO E EM ESSs EVs EFTs EP EA ES F FCs G GT GHG GA H HEMS HDs HEMO HAs HEMC HAN HESs HVAC I I-HEMO IBR IPC -

Dispatchable generators Delay scheduling Dynamic programming Differential evolution Decentralized coordinations Demand profiling DR synergized with renewables Delayed scheduling Diversified performance analysis Dynamic pricing scheme / signal Decomposed-weighted-sum PSO Energy management Energy storage systems Electric vehicles Emission factors Electricity price Evolutionary algorithm Expert Systems Fuel cells Game-theory Green house gas Genetic algorithm Home energy management system HEMS devices Home energy management system optimization Home appliances HEMS controller Home area network Hybrid energy systems Heating, ventilation and air-conditioning Individual HEMO Inclined block rate Illinois Power Company xxvii

IP L LP LDGs LSD LS LoT LMs LMP M MO MOO MGs MS MOGA/PO MPC MASs MILP MINLP MBAs MIL MDP MS MOGA N CEnet NCAs NAN NLP NYISO NSHA O OBJFs OF P PV PEMS PSO -

Integer programming Linear programming Load shedding-compensating dispatchable generators Load shedding Large-scale load shedding Length of operation time Lagrange multipliers Locational marginal price Multi-objective/ Multi-objectivity Multi objective optimization Microgrids Mixed scheduling Multi-objective genetic algorithm or pareto optimization Model predictive control Multi-agent systems Mixed-integer linear programming Mixed integer non-linear programming Model-based appliances Mixed integer linear Markov decision process Mixed scheduling/mixed scheduled Multi-objective GA Net cost of energy non-controllable appliances Neighborhood area network Non linear programming New York independent system operator Non-shiftable home appliance Objective functions Objective function Photovoltaic Prosumers based EM and sharing Particle swarm optimization xxviii

POS PFs P AR PP PO POS PDDR PDDR-RED Q QP R RESs RT RE RTP RDESs RO RM S SB SHAs SMs SOO SBs SOC SOP SDP SA 2S-ToUP 3S-ToUP T T BD ToU T EM iss TS TP TO U

Pareto optimal set Pareto fronts Peak-to-average ratio Physical Programming Pareto Optimization Pareto-optimal set Price-driven demand response PDDR synergized with RESs and ESS optimal dispatch Quadratic programming Renewable energy sources Real-time Renewable energy RT pricing Renewable-based distributed energy sources Robust optimization Resource management Storage battery Shiftable home appliances Smart meters Single objective optimization Storage batteries State of charge Stochastic optimization Stochastic decision process Simulated annealing 2-stage time-of-use pricing 3-stage time-of-use pricing Time-based discomfort Time of use T BD and minimal GHG emissions Tabu search Test problem Tradeoff

xxix

UF W WTs WSM Z ZEB -

Utility Function Wind turbines Weighted sum method Zero energy building

xxx

List of Symbols CE EP IBR N KSOC T BD - T st A&BAlpha Beta CE CEsold EP F IT IBR Iter N kLOT N g max NP V C N T bill P app P chg P chg max P dis P dis max -

Cost of energy purchased from grid Rate of electric energy from the grid (tariff) Inclining block rate Number of slots in a scheduling horizon Number of SHAs State of charge Time-based discomfort Decision vector for starting times of SHAs Vector for numbering SHAs Vector of the starting slots of the SHAs’ operating time intervals Vector of the ending slots of the SHAs’ operating time intervals Cost of energy purchased from the grid Cost of energy sold to the grid Vector of electricity prices Vector of feed-in tariffs Inclining block rate Number of iterations Number of slots in the scheduling horizon Number of SHAs Vector of the lengths of the SHA operating times Maximum number of generations for the GA Net present value in cash (total present cost of HES minus reduction in cost of energy usage after HES installation) Net bill to be paid to the utility Vector of per-slot power values for the SHAs Vector of SB charging power values Maximum SB charge rate Vector of SB discharging power values Maximum SB discharge rate xxxi

P gds P grid P gen P gsize P pv P schd P sold PT P dl SOC SOC max SOC min SOC(0) Stype T BD T BD(A) T BD(D) T BD(M ) Ts Xa -

Power grid status Vector of values representing power from the grid Vector of values representing power supplied by the DG DG size required to cope with LS Vector of PV power values Vector of scheduled loads Vector of values representing energy sold to the grid Power threshold for IBR application Vector representing power supplied to a dummy load during LS Vector of states of charge Maximum SOC limit Minimum SOC limit Initial SOC at the start of the scheduling horizon Vector of the scheduling types for the SHAs Time-based discomfort due to scheduling Average time-based discomfort due to AS Average time-based discomfort due to DS Average time-based discomfort due to MS Decision vector for the start times of the SHAs Power vector based on T s for the ath SHA

xxxii

Chapter 1 Introduction

1

Chapter 1

1.1

1.1. INTRODUCTION

Introduction

Over the past few decades, demand for energy has increased at a drastic pace, while energy generation capabilities have not been upgraded at a sufficient rate to reach out the rising demand. A balance between demand and generation of energy is the pivotal requirement for stable operation of a power system. The said balance in the past has maintained by the utilities while upgrading their centrally located generation capacities using an approach known as supply-side management. Over the last decade; however, demand-side management (DSM) has emerged as an alternative method for energy management (EM) in order to maintain the said balance while focusing on the consumer side. DSM function in a home is implemented by home EM system (HEMS) through optimal operations of HEMS devices (HDs) [1]-[6]. The methodology to compute the optimal schemes for the said operations is called home EM system operations (HEMO). For the integration of renewable energy (RE) sources (RESs) and energy storage systems (ESSs) models for HEMO are classified as: (a) Price-based demand response (DR) and (b) DR synergized with RESs and ESSs optimal dispatch (DRSREOD) [7]. DR-based HEMO model includes a method to compute the optimal schedules of home appliances (HAs) that enables shifting of the consumer’s load from peak to off-peak periods. Such shifting of the loads results in a smoother demand profile that benefits the utility by reducing the cost of generation. Utilities encourage consumers to participate in DR-based activities by offering lower energy prices during off-peak hours. Consumers benefit through the reduction in cost of energy CE obtained by shifting their loads towards off-peak hours. Research on various models for DR-based HEMO has been carried out over the last decade. Environmental concerns, reports on approaching the exhaustion of fossil reserves and rapidly reducing prices of RESs and ESSs have incentivized the consumers to install DRSREOD-based HEMS at their homes in order to supply a portion of their

2

PhD thesis by: Bilal Hussain

Chapter 1

1.2. REQUIREMENT FOR DRSREOD-BASED HEMS

loads from RESs and inject the excess energy back into the grid for monetary and environmental benefits [8]. Accordingly, photovoltaic (PV) technology and wind turbines (WTs) are the fastest growing sources of green energy throughout the world over the past few years [9]. The aforementioned scenario has led to an active pursuance of research on optimal operations of DRSREOD-based HEMS. Such operations combine optimal schedules of shiftable HAs (SHAs) with the dispatch schemes of RESs, ESSs and the grid.

1.2

Requirement for DRSREOD-based HEMS

Since, the power supplied by RESs is intermittent in nature; therefore, an ESS is also must be integrated into such a HEMS to introduce dispatchability. Such a DRSREOD-based HEMS provides additional benefits to the consumer (and the utility) by reducing energy bills, reducing peak demands, achieving overall energy savings and enabling the sale of surplus energy to the utility. A DRSREOD-based HEMS mainly consists of HAs, RESs, an ESS, a HEMS controller (HEMC), a home area network (HAN) for local communication and smart meters (SMs) for two-way communication between the consumer and the utility for the exchange of pricing and consumption information. A general HEMS architecture is shown in Fig. 1.1. The main components of DRSREOD-based HEMS include the following:

1.2.1

Home appliances

The HAs installed in a home are shifted in time or curtailed in power in order to participate in DR. Broadly, these HAs can be classified as: non-shiftable and SHAs [10]. Non-shiftable HAs consist of those appliances that have a rigid start and end time requirement. The operation of such appliances must be completed within the given length of operation time. Thus, these appliances cannot be shifted to off-peak time slots due to their non-flexible nature. Such appliances may consist of 3


Chapter 1


Remote control

Internet

RESs

Local DG

Smart meter

Ulity company

Figure 1.1: HEMS architecture for a smart home

a refrigerator, TV, electric heater, etc. SHAs like washing machines, dishwasher, iron, etc. can be shifted to other time slots to minimize the peak load and also save consumer’s CE [11]-[12]. Further, the power of some appliances like AC’s, fans, lights, etc. can be interrupted and curtailed based on DR signals.

1.2.2

Renewable energy sources

The electricity production, transportation, and distribution via a centralized system are expensive. Therefore, RESs are emerged to supply power in a cost-effective and distributed way. RESs supply green energy to the local loads and the grid. Among the other RESs, PV and wind are the most suitable and widely available energy sources [13].

1.2.3

Energy storage systems

Since the energy produced by RES is intermittent by nature and also prone to fluctuations, therefore, ESS is used in HEMS. The ESSs store extra energy from 4


Chapter 1


RESs and cheap energy from the grid during off-peak hours that can be consumed by the HAs or transmitted to the grid during peak hours.

1.2.4

HEMS controller

The HEMC is a vital component that embeds all the intelligence required to carry out HEMO function. It is based on a control algorithm that carries all of the computational intelligence required for optimal HEMS operation. For a DR-based HEMS, the controller computes the optimal schedules for the SHAs, whereas, for a DRSREOD-based HEMS, it computes the optimal schedules combined with the optimal dispatch for the RESs, the ESS, and the grid to achieve the HEMS objectives. The functions of HEMC can be given as [14]: • Receive data from HAs and sensors installed in a home. It also receives DR signaling information from the utility via a SM. • Provide rich interface to its users for controlling appliances usage, monitoring or setting user comfort options [15]. • Perform optimal HEMS operation based on the algorithms installed. It also transmits the control signals information to the required HAs to fully automate DR programs. • Tackle and control scalability challenges pertaining to the appliances with various parameter settings. • HEMC must ensure cooperative and reliable energy consumption of HAs based on the agreement between the consumer and utility (if any).

1.2.5

Home area network

The HAN enables local communication between the smart HDs and the HEMC [16]. The major protocols used for enabling HAN are Zigbee, Bluetooth, Z-wave, WiFi, etc. [17]-[19]. In [17], Zigbee is proposed for communication because of

5


Chapter 1

1.3. MODELS FOR HEMO

features like low initial investment, low power, and low range wireless connectivity. The authors used Bluetooth technology for communication HAs and HEMC to minimized energy usage [18]. Fletcher et al. used WiFi for monitoring and recording of home energy. A server is used to save data for smart grid applications [19].

1.2.6

Smart meter

The SM is used for two way communication between the consumer and the utility in order to exchange the pricing and consumption information [20]. The SMs can be considered as the foundation of HEMS. The major functions related to SM are listed as: • Perform two-way communication between utility and consumers and share real-time (RT) information accordingly. • The primary function of SM involves measuring the amount of energy used by the consumers during various time slots of the day. • Sharing of demand and response information regarding load consumption with both consumers and utility. • Any other data collection requirement for additional value-added services.

1.3

Models for HEMO

A generalized model for HEMO is based on the elements including response classes for HAs, RESs and ESSs; dynamic tariffs; problem formulation using control and decision parameters; nature of control parameters; objective functions (OBJFs) and constraints; coordination requirements among consumers; algorithms/ optimization techniques to carry out HEMO; and information, automation and communication systems to implement the model. Recent literature on HEMO have focused on any one or more of the aforementioned modeling elements. Further,

6


Chapter 1


based on the nature of the modeling elements, HEMO models can also be reviewed for dichotomous approaches designated as DR versus DRSREOD-based; RT- versus day-ahead (DA)-based; deterministic versus stochastic; single objective optimization (SOO) versus multi objective optimization (MOO); Individual HEMO (I-HEMO) versus co-ordinated HEMO (CoHEMO); and conventional techniques versus advanced heuristics-based techniques. In this approach of reviewing of research, the models are analyzed based on the performance parameters for the efficacy of the approach. Furthermore, tradeoffs (TOs) among the dichotomous approaches can also be investigated based on the achieved performance parameters for the respective approach. Research on HEMO based on the aforementioned approaches for reviewing of research is summarized below.

1.3.1

Demand response versus DRSREOD-based HEMO models

Based on the integration of RESs and ESSs, the models for HEMO are classified as DR versus DRSREOD. The methodology for DR-based HEMO is primarily based on time-shifted/ curtailed operations of HAs making use of dynamic tariffs. There is a multitude of HAs in each household that can be included in modeling for HEMO. In [21], the authors presents the ownership rates of various responsive devices in Canada. A dataset including more than 300 commercially available smart HAs is presented in [22]. HEMS components including SMs, software/ hardware based HEMCs, HANs, etc., are presented in [23]. DR-based HEMS comprising the aforementioned components enables reducing CE and peak demand making use of HEMO. The methodology for DRSREOD-based HEMO is based on the time shifted and curtailed operations of SHAs combined with the dispatch schemes of RESs, ESSs and the grid. RESs are the most preferred choice for a sustainable supply in future [8]. Various types of RESs and the methodologies for their integration 7


Chapter 1


in HEMS are reviewed in [24]. However, the power supplied by RESs is intermittent in nature and consequently, such sources are complex to be modeled for optimal dispatch [24]. RES-based HEMS are integrated with ESSs to introduce dispatch-ability. The storage battery (SB) is one of the most popular types of ESS that can be integrated with RES-based HEMS to ensure reliable operation of HDs. The technology for lead-acid batteries and the methods for better adaptation of such batteries with RESs are discussed in [25]. Furthermore, ESSs that can be integrated with PV units including SBs, super capacitors, fuel cells (FCs), pumped hydro, compressed air, flywheels and thermal energy storage technologies are reviewed in [26]. In [27], the authors reviewed the operation of community batteries that enables time-shifts of the PV energy usage and the demand load simultaneously; based on dynamic tariffs. The authors conclude that the optimal performance for HEMS operation is based on the full discharge of batteries during peak hours. Various schemes for optimal operation of DRSREOD-based HEMS with hydrogen backup technology are reviewed in [28]. The surplus RE has used to generate hydrogen through electrolyzers. The stored hydrogen was used to generate electrical energy by FCs when the energy from RESs has not sufficient to supply the load. In [29], the authors reviewed approaches for integrating hydrogen energy technology into HESs. Algorithms for optimal EM are also analyzed. Prosumers based EM and sharing (PEMS) are reviewed for optimization techniques and communication technologies in [30]. In [31], the authors analyzed the efficacy of approaches for DR, integration of ESSs and synergization of electric vehicles (EVs) in HEMSs in order to achieve the objective for peak load shaving. DRSREOD-based HEMS provides added benefits by reducing peak demands; enabling the sale of surplus RE; reducing the net CE (CEnet) and greenhouse gas (GHG) emissions; and improving the parameters for the reliability and power quality of the energy systems.

8


Chapter 1

1.3.2


DA versus RT HEMO models

Based on timescale, HEMO models are classified as the DA versus RT [21]. The DA-based approach can be applied to large-scale HEMO problems due to mild constraints on solving time; however, RT-based approach bears scalability issues because of solving time constraints. The later approach offers the best performance when used to complement DA-approach. HEMO models can also be classified for the tariff: price versus incentive-based. In [32], the research is reviewed for price-based DR including RT pricing (RTP), DA pricing (DAP) and critical peak pricing (CPP) and incentive-based DR including direct load control, curtailing schemes and demand bidding. Further, models are also reviewed for progressive tariffs which penalize high consumption of electricity and electricity saving feed-in tariffs, also provide incentives to reduce consumption of electricity in [33].

1.3.3

Deterministic versus stochastic HEMO models

For uncertainty of data, the models for HEMO are classified as deterministic versus stochastic. A number of parameters included in the formulation of HEMO problem are stochastic in nature and carry data uncertainty. In deterministic approach, forecasted values of the said parameters are taken as ex-ante. While in stochastic approach, the uncertainty of parameter is tackled using stochastic methods. The authors in [21, 34, 35] have reviewed the research on HEMO analyzing the TOs among methods used to handle data uncertainties. In addition, the predicted data for demands and the energy harnessed from RESs are integral parts of HEMO models. In [36], the authors reviewed methods for data predictions classifying them as data-driven and large scale building energy prediction. Learning algorithms used for the prediction tools are classified as regression; ANN and support vector machine. SMs record information about electricity consumption in near RT called SM big data. In [37], the authors presented a technique to compress the said data for demand predictions. 9


Chapter 1

1.3.4


Single versus multi objective HEMO models

For the number of OBJFs, HEMO models are classified as SOO versus MOO. In real life, most HEMO models are MOO in nature. Classification based on the number of objectives enables comparison of pros and cons of the two approaches. Further, the efficacy of an approach for the HEMO model can also be analyzed. The authors in [35] and [38] have reviewed the research on HEMO modeling and optimal sizing of HES, respectively; analyzing the TOs for the number of OBJFs.

1.3.5

Individual versus co-ordinated HEMO models

Based on coordination among consumers, the models for HEMO are classified as I-HEMO versus CoHEMO. I-HEMO is characterized using the advantages like local independence, fast convergence, easy implementation and its drawbacks like peak rebounds. CoHEMO is characterized using the advantages like avoidance of peak rebounds, more optimal operation of HDs. The approach is associated with disadvantages of scheduling conflicts and scalability issues that can be resolved by incorporating game-theory (GT) based approach like using the heuristics. Further, large communication infrastructure requirement for coordinated HEMS results in increased system implementation cost. The authors in [21, 34, 39, 40] have reviewed the research on HEMO analyzing the TOs for coordination among consumers. Moreover, the authors in [41]-[44] have analyzed distributed and centralized approaches for EVs charging, cellular base stations and MGs respectively.

1.3.6

Conventional versus advanced heuristics-based techniques

For discreteness of decision variables, the HEMO problems are classified as a combinatorial problem using advanced meta-heuristics versus continuous variable problem using conventional techniques. HEMO problems especially the ones based 10


Chapter 1

1.4. ENERGY CHALLENGES AND OPPORTUNITIES

on CoHEMO approach are very large in scale. Conventional optimization (COP) methods like linear programming (LP) are applicable to small-scale problems only. While combinatorial based advanced meta-heuristics give near optimal solution for large-scale NP-hard problems in matchless time. The proposed classification helps in analyzing TOs among the approaches based on performance parameters achieved by the respective models for HEMO. The authors in [32], [43], and [45] have reviewed optimization techniques for microgrids (MGs) optimal operations, hybrid energy systems (HESs) sizing and HEMO using meta-heuristics.

1.4 1.4.1

Energy challenges and opportunities Power shortfalls and load shedding

Over the past few decades, demand for energy has increased at a drastic pace, while energy generation capabilities have not been upgraded at a sufficient rate to catch up with the rising demand. This imbalance between demand and generation has resulted in power shortfalls. This can place networks in undesirable situations and can lead to system instability and load shedding (LSD) in developing countries [46]. A HEMS is used to implement DSMin a home. This provides an opportunity to shift the peak load to the off-peak hours and also helps in minimizing the LSD problem. Thus, a HEMS achieves its function through price-based DR and the optimized dispatch of distributed energy sources, especially RESs [47]. Price-based DR consists of the scheduling of consumer loads, based on load shifting from peak to off-peak periods, to achieve a smoother utility demand profile. More than 24% of the loads from systems installed in homes in developing countries are elastic in nature. Load elasticity carries a large hidden potential to smooth the utility demand profile through loads shifting via price-based DR [48]. This benefits the utility by reducing the generation cost through the exclusion of costly peaking 11


Chapter 1

1.4. ENERGY CHALLENGES AND OPPORTUNITIES

plants from dispatch and by encouraging the sale of unutilized energy available during off-peak periods. Utilities encourage consumers to engage in such desired load shifting by offering lower energy prices during off-peak hours. Consumers benefit through the reduction in CE obtained by shifting their loads toward offpeak times.

1.4.2

Green house gas emissions

A rampant rise in green house gas (GHG) emissions, the consequent climate changes and the related environmental issues have raised serious concerns over the quality of the life on the earth. In order to mitigate the serious environmental issues, various proposals have been discussed at the highest international forums to confine GHG emissions. The Kyoto protocol of United Nations Framework Convention on climate change has signed by 192 countries all over the world which proposes a reduction in GHG emissions through selling of emission commodities [49]. Such a trading sets penalties and quantitative limitations on emissions by polluters that may include utilities, independent MG operators, and the prosumers having fossil fuel based generation deployed with DRSREOD integrated with load shedding-compensating dispatchable generators (LDGs) (DRSREODLDG)-based HEMSs. This has incentivized utilities to reduce not only the cost of generation of energy; but also the supply-side emissions making use of RESs installed for DRSREOD-based HEMS. The research on HEMS now seems to focus on reducing the GHG emissions along with the other well-known objectives for (CE), time based discomfort (T BD), etc. In [50], a scheme for DR-based HEMS is presented. Non-critical house loads are shifted towards off-peak hours to minimize the daily cost of generation and the supply-side emissions. It is validated that implementation of the DR program effectively reduces the cost of generation on the supply-side; however, the emission on this side is reduced only when peak demand is met by high emission fuels based peaking plants. The DRSREOD-based HEMS 12


Chapter 1

1.5. DISSERTATION PROBLEM STATEMENT, FOCUS, AND CONTRIBUTIONS

on the other hand, through an optimal operation of HDs, can easily be used for reducing the supply-side emissions along with the reductions in the CEnet for the consumer and cost of generation for the utility. In [51], authors present a scheme for optimal scheduling of SHAs integrated with the optimal dispatch of an RES and an SB. The objectives include reductions in CEnet, temperature based discomfort, peak load, and the GHG emissions. The supply-side emissions are computed using GHG emission factors (EFTs) for the energy mix adopted at different times of the day. The supply-side emissions are reduced through an optimal operation of local RESs and storage batteries (SBs) during high emission times.

1.5

Dissertation problem statement, focus, and contributions

Most of the research conducted on HEMSs over the last decade has been based on DR only [14], [54]; research on DR-synergized RESs has been actively pursued only for the past few years [55]-[57]. Researchers are also currently pursuing the integration of ESSs into their models to introduce dispatchability for RESs [58][64].

1.5.1

Thesis problem statement

The present research on optimizing the size of hybrid energy systems for HEMSs is primarily focusing on the new infrastructure to be installed [65]-[67]. The issue of DG sizing to ensure power availability under LS conditions for an existing DRSREOD-based HEMS while considering CE and T BD TOs is seldomly addressed. Due to the rapid increase in demand, consumers in developing countries are facing widespread LS due to energy-deficient systems. This study proposes a novel method for the optimal sizing of a DG that can supply MS loads during LS hours in parallel with SB and PV power supplies in an existing DRSREOD-based 13


Chapter 1


HEMS. The DG sizing problem considers the HEMS TOs between CE and T BD as well as the maximum scheduled LS hours. To the best of our knowledge, it is rare to find research on the optimal DG sizing to cope with LS for an existing DRSREOD-based HEMS infrastructure while considering the TOs between DG size required to cope with DG size required to cope with LS (P gsize), CE and T BD. Further, most of the DR models considered in present research are based on DS [53, 57]. These limit scheduling flexibility and opportunities for achieving the maximum reduction in CE by making use of more off-peak hours for load shifting, direct use of RESs and more optimal use of the ESS during peak hours. The handling of multi-objective (MO) is a major issue in the present research. Most of the researchers have used WSM as priori to handle the issue. The method solves the problem combining all of the objectives into a single compound function (SCF) while assigning fractional weights as priorities to the constituent objective functions (OFs). Such methods do not provide any feedback to the consumers that may enable to improve their selection of choices. The MO issue can be resolved by achieving a diversified set of TO solutions. A few authors achieved such TO solutions by using PO method combined with the evolutionary algorithms like genetic algorithm (GA) for non-dominated solutions [7]. The techniques like WSM combined with advanced meta-heuristics, however, have rarely been used to achieve such TO solutions as posteriori. Furthermore, most of the algorithms proposed in the recent research have not been tested for their performance for a diversified set of test problems (TPs).

1.5.2

Thesis focus

This thesis initially presents an algorithm for a DRSREOD-based HEMS that considers the aforementioned improvements/conditions. The proposed heuristic algorithm combines DR with optimal dispatch based on the excess available PV energy, maximum charge/discharge rates and SOC. PV technology is regarded as 14


Chapter 1


the preferred source for supplying the load. Excess PV energy is stored in the SB in accordance with the limiting SB parameters (SOC max, maximum charge rate), and the rest is sold to the grid. The SB supplies the load during peak hours in parallel with the grid in accordance with the limiting SB parameters (SOC min, maximum discharge rate). MS is proposed for the SHAs, with some of them classified as AS (w.r.t. the preferred ending time for completing their operation) and the others classified as DS (w.r.t. the preferred start time for starting their operation), to improve the HEMS performance. A time of use (ToU) tariff with an inclined block rate (IBR) scheme is incorporated to limit the peak load. The proposed DRSREOD-based algorithm minimizes CE and T BD while considering the underlying TOs. Most researchers have previously incorporated TOs between HEMS performance parameters by means of the weighted sum method (WSM) [14, 53, 54]. Fossil-based LDGs are integrated into DRSREOD-based HEMSs to supply the load during LSD hours. Such a LDG adds a vital benefit of uninterrupted supply of power to DRSREOD-based HEMS. An algorithm for optimal sizing of LDG for DRSREODLDG-based HEMS was proposed in our recent research [7]. The proposed sizing was based on the TO analysis for the parameters including CE, T BD and size of LDG. An uninterrupted supply of power through the integration of LDG was ensured; however, the operational schemes for HEMS were remained to be analyzed for the emissions released during the LDG operations. To implement an eco-efficient operation of DRSREODLDG-based HEMS, optimal TOs between CEnet, T BD and minimal GHG emissions (T EM iss) need to be computed. This dissertation introduces a method to harness a diversified set of solutions to decision vector T st and the related TOs for CEnet, T BD and minimal T EM iss for an eco-efficient HEMS operation. The proposed method for an eco-efficient operation of DRSREODLDG-based HEMS is based on a three-step approach. In step-1, a set of primary solutions in terms of T st and the related TOs for CEnet, T BD and minimal T EM iss are 15


Chapter 1


generated using algorithm 1. The algorithm is based on a heuristic derived from our previous studies on HEMS in [7]. The proposed heuristic takes into account PV availability, the state of charge, the related rates for the storage system and the parallel operation of the sources. To achieve maximum reduction in CEnet, SHAs are modeled for mixed scheduled (MS) as already validated in [7]. This research formulates the TOs parameters for: CEnet to include the CE purchased from the grid, CE sold to the grid and the CE supplied by the LDG; T EM iss to include the energy supplied by the LDG during LSD hours, EF T based on the calorific value of the fuel, the consumption efficiency of the LDG and the related EFTs for GHGs; and T BD to include the delay in the starting times of delay scheduling (DS) type and advanced completion of the job of advanced scheduling (AS) type for HAs. The TOs solutions obtained in step-1 are analyzed for T EM iss as related to the TOs between CEnet and T BD as shown in Fig. 1.2. 1.80

1.60

1.40

TEMiss (Lbs.), TBD

1.20

1.00 R² = 0.0111 0.80

0.60

0.40

R² = 0.8501 0.20

0.00 20.00

25.00

30.00

35.00

40.00

45.00

50.00

55.00

CEnet (Cents) TBD

TEMiss

Expon. (TBD)

Linear (TEMiss)

Figure 1.2: Un-even trends for T EM iss as related to CEnet and T BD

The plot in Fig. 1.2 reveals a highly un-even relation between T EM iss and the related parameters for CEnet and T BD. This un-even trend for T EM iss is exploited to screen out/ exclude a set of TOs with larger values of T EM iss using a constraint filtration mechanism as presented in algorithm 2 (step-2 and step-3). In 16


Chapter 1


step-2, an average value based constraint filter (AVCF) for T EM iss is developed and applied to filter out the TOs with extremely high values of T EM iss. In step3, average surface fits for T EM iss are developed in terms of CEnet and T BD using polynomial based regression. The most suitable polynomial is selected after cross-validation of 25 number of polynomial model that fits on their capabilities in order to reduce T EM iss and T BD, and to maximize the number of diverse TOs for CEnet and T BD. The average surface fit based constraint filter (ASCF) with the selected polynomial formulation is applied to screen out the TOs with even marginally higher values of T EM iss. The solutions for an eco-efficient HEMS operation are thus achieved including diversified TOs for CEnet, T BD and minimal T EM iss.

This research proposes a decomposed-weighted-sum PSO (DWS-PSO) algorithm for MOO for price-driven demand response (PDDR) and PDDR synergized with RESs and ESS optimal dispatch (PDDR-RED) based HEMSs. The obtained TO solution helps the consumer in making decisions as per his needs. A method to carry out the diversified performance analysis (DPA) for a HEMS algorithm based on the construction of the TPs and the formulation of the performance metrics is also proposed.

1.5.3

Thesis contributions

In this thesis, firstly the current state of the art in HEMO focusing on DRSREODbased HEMS is reviewed. DRSREOD and DRSREODLDG algorithms are then proposed to harness diversified TOs. Following are the major contributions of our work: 1. An overview of HEMO for DRSREOD-based HEMS is presented. The pursuit may be a useful reference for new researchers to familiarize them with the field. 17


Chapter 1


2. Models for DRSREOD-based HEMO in the most recent literature are reviewed for dichotomous approaches for integration of renewable-based distributed energy sources (RDESs), dynamic tariffs, mutual coordination, uncertainty of data, multi-objectivity and optimization techniques. The performance achieved and the main features of the proposed models are furnished. The performance of the models is analyzed based on the efficacy of the modeling approaches. Further, the TOs among the approaches are also investigated. The accomplishment is a ready reference for new researchers to access the recent literary work on HEMO. 3. Challenges related to multi-objectivity, data uncertainty, optimization techniques, problem complexity and coordination among HEMS are analyzed. Furthermore, salient issues regarding standardization requirements for HAs and formalization needed to handle diversification in HEMO are investigated. Solutions are proposed to resolve the outstanding challenges. The work may help researchers in understanding the most recent approaches visualizing the future trends and exploring advanced methods for HEMO. 4. A DRSREOD algorithm based on MS is proposed that outperforms a DSbased algorithm for CE and T BD TOs for DR as well as for a DRSREODbased HEMS. 5. This thesis proposes a novel method for the optimal sizing of a DG that can supply MS loads during large-scale load shedding (LS) hours in parallel with SB and PV power supplies in an existing DRSREOD-based HEMS. The DG sizing problem considers the HEMS TOs between CE and T BD(M ) as well as the maximum scheduled LS hours. To the best of our knowledge, it is rare to find research on the optimal DG sizing to cope with LS for an existing DRSREOD-based HEMS infrastructure while considering the TOs between P gsize, CE and T BD(M ). 6. An innovative method is proposed to harness diversified TOs between CEnet,

18


Chapter 1

1.6. THESIS ORGANIZATION

T BD and minimal T EM iss for an eco-efficient operation of DRSREODLDGbased HEMS. 7. The TOs for such HEMS have rarely been computed by combining a multiobjective GA or pareto optimization (MOGA/PO) based heuristic and regression based constraint filtration. The polynomial model fit for regression is based on its capabilities to reduce the TOs parameters for eco-efficient HEMS operation. 8. Most of the authors use the WSM to handle multi-objectivity for similar problems. This thesis presents a diverse set of TOs that are critically analyzed to enable the consumer choosing the best option. 9. Trends exhibited by the TO parameters are analyzed based on vital factors affecting these parameters, e.g. loss of unused energy from the PV unit. 10. The proposed method validated to minimize the emissions from a local LDG for a DRSREODLDG-based HEMS; however, it is easily extendable to reduce the supply-side emissions as well. 11. A DWS-PSO algorithm for TO solutions to HEMS operations is proposed. 12. A diversified set of TPs based on standardized DPSs and scheduling models for DS and MS are introduced. 13. Performance metrics for DPA for HEMS algorithms are also proposed. The metrics include the reduction in CEnet and the gradient of the TO line for the reduction in CEnet and the TBD. 14. A method for DPA for PDDR-, RED-, and PDDR-RED- based HEMS algorithms is proposed. The method makes use of the proposed set of TPs and the defined metrics.

1.6

Thesis organization

The thesis is organized into 4 major parts stated as: • The preliminary related work is given in Chapter 2. 19


Chapter 1

1.6. THESIS ORGANIZATION

• A survey conducted on HEMS is presented in Chapter 3. • The DRSREOD and DRSREODLDG algorithms, model, problem formulation, and proposed solution are summarized in Chapter 4. • The proposed algorithms results for CE, T BD, and TOs relationship are presented and discussed in Chapter 5. • Finally, conclusion and future work are given in Chapter 6. The related work (Chapter 2) and survey (Chapter 3) chapters give an in-depth overview of state of the art literature and domain knowledge pertaining to HEMS equipped with RESs and ESSs. Furthermore, the survey part also gives a comparative analysis of various HEMS. In Chapter 4, DRSREOD and DRSREODLDG algorithms are discussed along with proposed system models, and problem formulation. The DRSREOD algorithm and its results as given in Chapter 5 are condensed from our publication [7]. Finally, conclusion and future work are stated in Chapter 6.

20


Chapter 2 Related work

21

Chapter 2

2.1

2.1. RELATED WORK

Related Work

The related work is discussed as HEMO components review, categories of HEMS, and emissions reduction sections.

2.1.1

Components for HEMO

A generalized model for HEMO encompasses the following elements [68]-[73]: (a) Response models of HAs, RESs and ESSs; (b) Dynamic tariff; (c) Problem formulation using control and decision parameters; (d) Nature of control parameters; (e) OBJFs and constraints; (f) Coordinative requirements; (g) Algorithms/ optimization techniques to solve HEMO problem; and (h) A platform to validate the results through simulations. In recent literature reviews on HEMO, the authors have focused on any one or more of the aforementioned elements especially response models of HDs, tariff schemes, OBJFs and optimization methods. However, the nature of modeling elements enables classifications of HEMO models for dichotomous approaches for integration of RDESs; timescale for scheduling; type of tariffs; mutual coordination among consumers; uncertainty of data; multi-objectivity; and optimization techniques. The said classifications introduce a unique method to review the literature on HEMO by analyzing the performance of HEMO models for the suitability of an approach. Furthermore, pros and cons of approaches can also be investigated taking into account the performance of related models. In [21], the authors present a survey of research on HEMO. Various types of HDs including non-controllable appliances (NCAs), SHAs, elastic HAs, RESs, ESS, and dynamic tariffs comprising ToU, RTP and CPP are introduced. The reviewed models take into account OBJFs for minimizing CE, T BDs, peak load and maximizing the usage of RESs. The models consider constraints related to HAs (for T BD or energy usage); RESs; ESS (for SOC, charge/ discharge rates, damages, SOC left for the next day, cost, etc.); energy balance; and tariffs. In 22


Chapter 2

2.1. RELATED WORK

addition, the HEMO models are classified as I-HEMO versus CoHEMO; deterministic versus stochastic; and DA versus RT. I-HEMO models are primarily based on DR and are formulated as SOO. LP, dynamic programming (DP), heuristics and meta-heuristics like GA are used to compute optimal solutions. The deterministic approach for HEMO is considered viable despite of carrying scheduling errors due to uncertainty. The stochastic approaches take into consideration the uncertainties of parameters for RESs, device stop times, outdoor temperature, tariffs, etc. Stochastic optimization (SOP) and robust optimization (RO) methods are used to tackle uncertainty. RO methods handles uncertainty with lesser computational complexity using the variance of price. In addition, the methods based on the stochastic gradient of the OBJF are also discussed. I-HEMO-based systems are prone to peak rebounds as the approach does not consider the joint effect of scheduling on the grid. CoHEMO-based systems enable avoiding peak rebounds using joint resource management (RM). However, this approach is associated with disadvantages of large data transmission, scalability and privacy issues. Further, RM is achieved by minimizing a shared utility function that does not resolve conflicts among cooperative consumers. A game theoretic approach is proposed to avoid such conflicts. For deterministic CoHEMO models, shared utility functions are achieved using CP, heuristics, EAs and Pareto multi-objectivity. CE has formulated as a convex function of the centralized generation and solved using CP. For stochastic CoHEMO, bounded uncertainty, Gaussian and unknown distribution models have been used to handle uncertainty. The authors have classified the models for time scale as well. The DA approach can be applied to large-scale HEMO problems due to mild constraints on solving time; however, this approach is sensitive to errors in data forecasting. The RT approach bears scalability issues and gives less accurate solutions because of solving time constraints; however, offers the best performance when it is used to complement DA-approach. In [32] the authors classify the research on HEMO for dynamic tariffs as price- versus incentive-based DR. Models for price-based tariffs are classified for RTP, DAP 23


Chapter 2

2.1. RELATED WORK

and CPP approaches. Incentive-based tariff are classified for direct load control, curtailing schemes and demand bidding approaches. Further, in [33], the authors review the models for progressive tariffs which penalize high consumption of electricity and saving feed-in tariffs which provide incentives to reduce consumption of electricity. The performance of the two approaches was compared based on price elasticity and an incentive elasticity. The former approach shows a better performance for mobilizing the consumers for electricity saving. In [23], authors presented a survey of research on smart home activities including optimization, automation, control and communication. A flowchart showing research areas for the activities and the information of reputed manufacturers/ products for HDs, controllers, SMs, communication systems, etc., is furnished. The reviewed models are primarily based on deterministic approach for I-HEMO and are formulated for SOO. The models take into account HDs including NCAs, SHAs, heating, ventilation and air-conditioning (HVAC) system, RESs and SBs and are included with fixed, ToU, RTP, and regulated approaches for tariffs. The problem formulations are based on OBJFs to minimize CE, discomfort, peak load, GHG emissions, to maximize local energy performance and use of local generation. Problems for HEMO have been solved using LP techniques and meta-heuristics including artificial neural network (ANN), PSO, GA and tabu search (TS). The GT with Nash-equilibrium is analyzed for CoHEMO. To handle uncertainty, methods comprising DA scheduling with RT adjustment, SOP and RO have been analyzed. In [38], authors present a review of HEMO modeling focusing the objectives, uncertainty of data and infrastructures requirements. Challenges such as device heterogeneity, multi-objectivity, forecast uncertainty, infrastructure implications and computational limitations are discussed. Heterogeneity has been related to the diversity and energy consumption pattern of devices that are based on dwelling characteristics, lifestyle and the home occupancy. The objective for well-being is generally related to the inconvenience and the discomfort due to time shifting or shift from desired energy states of SHAs. Penalty factors for undesired shifts 24


Chapter 2

2.1. RELATED WORK

based on bounded value, linear deviation, quadratic deviation and binary/ linear methods are also discussed. The reviewed models take into account the objectives for optimizing costs, well-being, load profiles and GHG emissions. Multiobjectivity has been tackled using methods comprising WSM, bounded objectives and physical programming. Optimal solutions are computed using mathematical optimization, heuristics and meta-heuristics. Uncertainties in forecasting of REs, home occupancy, energy consumption and weather are also discussed. Methods like RT scheduling, SOP, RO, GT, chance-constrained optimization (CCO), model predictive control (MPC) and stochastic decision process (SDP) are used to tackle uncertainty. Coordination in HEM results in maximal reductions in CE and system peaks with minimal impact on the user’s comfort. Approaches for co-ordinated HEM (CoHEM) comprise optimization, GT and multi-agent systems (MASs). In [39], modeling for CoHEM is reviewed focusing on the aforementioned approaches. Dynamic tariffs with IBR are used to participate in DSM through burst, regular loads and RDESs. Communication for DSM is enabled within a home using HAN and within a community using neighborhood area network (NAN). Data for electricity price (EP ), local generation and consumption is collected at a gateway that communicates with the consumer and the aggregator for the relevant data. Based on the centralization, coordination among models is classified as centralized and decentralized coordinations (CC and DC). In CC, a central operator (CO) manages electricity usage of all of the coordinative homes. Information regarding homes is accessible to the CO who computes schedules for HAs to implement a CoHEM strategy. Whereas, in DC, consumers schedule their HDs directly after communicating with each other to gather profile information of the others. The CC results in an efficient RM of consumers; however, suffers from the drawback of scalability. On the other hand, DC provides more independence of choices to the consumers; however, results in a higher aggregated system cost as compared to that for CC due to larger requirements of communication networks. The TOs 25


Chapter 2

2.1. RELATED WORK

among the coordinative approaches have analyzed as follows: (a) MAS has an advantage of a decentralized, intelligent and single template agent. Its drawbacks include separate modeling of each agent and the cost resulting from a larger number of communicating agents in RT. (b) GT being a strategic decision-making process is suitable for DC. Further, new players can easily enter into the game. Major drawbacks of using GT include the frequent message exchanges, large communication infrastructure requirements and difficulty in reaching an equilibrium. (c) Optimization methods compute optimal schedules for HDs for an individual as well as for coordinated homes. Coordinated optimization avoids peak rebounds. Scalability, privacy and conflicts among the consumers are the major drawbacks that can be avoided using enhanced meta-heuristics, coding methods and incorporating the GT approach. In [40], authors review various configuration and the control strategies to operate and implement PV, WT and SB based HESs. Operational schemes based on centralized control, distributed control, hybrid control and multiple control are analyzed. A hybrid scheme with a combination of the centralized and distributed control is proposed. The scheme resulted in global optimization. The disadvantage of multiple communications related to DC was proposed to overcome by use of MAS based methodologies for AI techniques. The hybrid control technique is less prone to failures; however, accompanies highly complex information conversion system. The multiple control method is same as hybrid control technique with the exception of better control on the system in addition to working on the basis of current information. In [74], the authors propose an advanced concept of zero energy building (ZEB) to achieve self-sustainability. A novel HEMS is proposed to minimize the grid-dependency of buildings. The optimal EM of ZEBs is a multi-objective problem which involves decisions regarding power dispatch, power flow and load scheduling. The controller is designed for a MAS, comprising three intelligent agents designed for the generation, load and the battery. A suitablydesigned control algorithm is used for controlling these agents carries out the 26


Chapter 2

2.1. RELATED WORK

instantaneous power management decisions. In [41], authors review various energy optimization approaches for charging of EVs. Various OBJFs including reduced power losses support for RESs for charging of EVs, minimized peak load, minimized total CE to the consumer, maximized aggregator profit and frequency/ voltage regulations are formulated. The optimized charging of EVs enhance the performance/ life of EVs batteries; conserve the energy by minimizing the load and power losses; and contribute to the voltage and frequency regulations. The centralized, distributed and hybrid charging approaches for charging of EVs suited to different OBJFs are studied. The hybrid approach combines the features of centralized and distributed approaches for charging of EVs battery. The approach addressed the limitations and demerits of the constituent approaches. In [42], the authors overviewed optimization strategies for the source and consumption management in cellular base stations powered by RESs. The OBJFs to reduce the cost and GHG emissions are taken into consideration. The possibilities of cooperation among stations for optimal EM are also analyzed. In [34], DSM models are reviewed for optimization and the GT approaches. The models are classified as individual versus coordinated and deterministic versus stochastic. The reviewed models are included with OBJFs to optimize the operational cost of the system, consumer discomforts, use of RESs, GHG emissions and the peak demand. Most of the reviewed models are based on DR; however, a few models based on DRSREOD are also analyzed. The computations for I-HEMO are carried out through LP, mixed integer LP (MILP), DP, heuristics including greedy iterative algorithm and meta-heuristics including evolutionary algorithm (EA), binary PSO (BPSO), GA and simulated annealing (SA). The uncertainty for EP , outdoor temperature, hot water usage, fixed load and solar irradiation for PV units are tackled using cost value at risk and fuzzy controllers. DSM optimization models incorporate the following features: inclusion of scheduled power outages, effect of information exchange between consumer and service provider, 27


Chapter 2

2.1. RELATED WORK

self-learning agent based DSM approach through heterogeneous clusters of devices, priorities for HAs, integrated planning based on DSM, etc. GT is used to introduce coordinative features in DSM. In centralized design, DSM problem is solved in order to reduce peak-to-average ratio P AR using LP, PSO, etc. In decentralized approach, non-cooperative consumers competitively minimize CE with minimum discomfort by controlling their energy profiles. GT models incorporate features, e.g., non-cooperative Nash game; GT extension with MPC using real data and stochastic forecasting; customers selection in a round robin fashion; repeated energy scheduling games as non-stationary algorithm to reduce total cost; dynamic game at two levels, at lower level between HAs for energy consumptions and at upper level between consumers based on market price; etc. A formulation to maximize the load factor for the utility through load shifting, peak clipping, valley filling and flexible loading for DSM problem is also presented. In [24], authors present state of the art work on application of optimization methods to renewable-based HESs focusing on system sizing, operational dispatch and optimal DSM. Heuristics and meta-heuristics are used to overcome scalability issues faced by conventional methods. Heuristics are classified as trajectory versus population-based and nature-inspired versus non-nature-inspired. Trajectory-based heuristics include hill climbing, SA, TS, greedy search, iterative local search, etc. Population-based heuristics include GA, EA, scattered search, memetic algorithms, ant colony optimization (ACO), PSO, differential evolution (DE), bee colony optimization (BCO), etc. Further, MO trajectory methods including Pareto based ES, MO-SA, etc., and MO population-based methods including MO-TS, non-dominated GA, Pareto SA, single front GA, strength Pareto EA, Pareto envelope-based selection are discussed. Furthermore, hybrid approaches such as genetic-TS and MO-based genetic local search, memetic-Pareto based ES, SA-TS, etc., are also discussed in detail. Models for optimal sizing of HES based on solar, wind, FC, DGs, SB, pumped hydro, etc., are solved using aforementioned SO based meta-heuristics. The objectives and constraints include minimizing the 28


Chapter 2

2.1. RELATED WORK

system cost and GHG emissions and maximizing the reliability and the energy supply security. An optimization method based on a hybrid of PSO-simplex outperform population-based methods like PSO, BCO, ANN, ACO and GA. On the other hand, fuzzy-adaptive PSO outperforms the heuristics including GA, PSO, DE, ACO and TS. MO-based HES sizing models are solved using Pareto based GA and fuzzy-SP. Optimal dispatch of HES is computed in order to meet the objectives including minimizing system operating cost, combined intermittence, peak load, GHG emissions and loss of power supply probability. HEMO models are solved for single as well as for compounded objectives using heuristics and meta-heuristics like PSO, EA and neuro-fuzzy. In [43], the authors analyze MG operations based on the power trading with the main grid. In case of main grid failure, MG restores its operation for islanded mode by providing a continuous supply to critical loads by efficient integrated operation of RESs, DR and LSD. MG operation is classified for centralized and decentralized approaches. The authors present a critical analysis of decision-making strategies and solution methods for optimal MGs operations. The methods are based on LP; non LP (NLP); meta-heuristic including GA, PSO, etc; and artificial intelligence (AI) including fuzzy logic, neural network and MAS. Uncertainty in data related to the renewables and load demand are handled through SOP, RO and MPC. A comparative analysis of communication technologies for MGs is also discussed. In [32], research related to the HEMS is reviewed for load scheduling controllers. Controllers for optimal scheduling of HDs are classified as rule-based; AI-based using ANN, fuzzy logic, adaptive neural fuzzy inference system, etc.; HEMObased using LP, mixed integer non-linear programming (MINLP) and GT and GA and PSO based meta-heuristics. In [35], authors investigate optimization techniques to solve the sizing problem for HES focusing on efficacy of approaches for MOO. Optimization problems are

29


Chapter 2

2.1. RELATED WORK

classified as limited versus unlimited based on parameters’ limits, discrete (combinatorial) versus continuous (heuristics) based on discreteness of decision variables, deterministic versus stochastic based on data uncertainty and SOO versus MOO-based on the number of objectives. The formulation for objectives functions of MOO is based on two approaches: (a) Integrated OBJF, where MO problem is transformed into a SO problem using WSM or changing all OBJFs into the constraints except the one and (b) Pareto-based MOO using the Paretodominance concept. This method solves HEMO problems as MOO and delivers non-dominated solutions as Pareto optimal set (POS) and the corresponding values of objectives as Pareto fronts (PFs). The tradeoff solutions enable the consumer to select the best HES components meeting his needs from a diversified set of options. OBJFs for HES sizing include optimizing the system cost, GHG emissions and the system reliability. While parameters for the system sizing and the operational strategy are taken as decision variables. The SOO problems for HEMO can be solved using COP methods as well as using meta-heuristics like GA and PSO. The MOO problems for HEMO are most widely solved using PO based heuristics like MO-PSO, MO-EA, MO-GA and non-dominated sorting GA. The COP methods including the least squared method, iterative method, MINL method and NL constrained method can also be applied to solve MOO problems, although used rarely. The DSM has emerged as an efficient method for energy management focusing the consumer side. A HEMS is used to implement DSM in a home. The DSM utilizes DR programs which are classified as price-driven and incentive-driven. Price-driven DR (PDDR) is the most important DR category that makes use of a control signal for energy management. Utilities offer dynamic pricing signal (DPS) in order to encourage consumers to shift their load towards off-peak hours. The DPSs offered by the renowned utilities in the past few years include TOUP, RTP, CPP, and IBR [75]-[76]. Shifted operations of HAs for dynamic pricing result in a reduced cost of generation 30


Chapter 2

2.1. RELATED WORK

for the utility whereas the consumer is benefited with the reduced CE. Numerous algorithms have been developed to solve the problem for PDDR-based HEMS while making use of the aforementioned DPSs [77]. Hussain et al. proposed a PSO algorithm for DS of SHAs in order to minimize the CE. The performance of the algorithm was validated for DA-RTP, two step (2S) 2S-ToUP, and 3S-ToUP combined with IBR [78]. Research on hybrid meta-heuristics has recently been expedited to solve PDDR- based HEMS [79]-[82]. The author in [81] modeled the problem for HEMS as MOO and solved it for the CE and a PAR using WSM as priori. All of the DPSs except IBR have been used for a diversified performance analysis (DPA). Javaid et al. proposed EDE, teacher learning-based optimization (TLBO) and a hybrid of the two techniques. The objective functions (OFs) were defined to reduce CE, PAR, and TBD. The algorithms for TLBO, EDE, and the hybrid were evaluated for DA-RTP and CPP [82]. Rasheed et al. presented a GA- based algorithm for an optimal operation of RES- based HEMS for DA-RTP. The HAs are modeled for shifted operations and for thermostatic control whereas non-SHAs (NSHAs) are dispatched to a dispatchable generator (DG) [67]. RESs are required to be integrated with ESSs in order to introduce dispatch-ability to PDDR-RED- based HEMS. Moghaddam et al. proposed an optimization model for energy scheduling taking RESs, ESSs, domestic thermo-electrical systems, and HAs into account. The HEMS problem was formulated to minimize the CE and maximize the consumer’s comfort. The OFs are combined through WSM as priori. The algorithm was validated for different scenarios for weather and for DPS including RTP, ToUP and flat rates [83]. Shakeri et al. developed an algorithm for HEMS operation to minimize overall CE for 3S-ToUP. Operation of HAs was shifted towards the SB in order to reduce the discomfort. The SBs were charged from the solar energy during daytime and from the grid during off-peak times [84]. Wang et al. proposed a pareto tribe evolution (PTE) with Nash equilibrium-based decision for MOO for multiple HEMS. Three OFs including consumer satisfaction,

31


Chapter 2

2.2. CATEGORIES OF HEMSS

CE, and PAR were simultaneously optimized for 4S-ToUP [85]. The optimal performance of an algorithm, however, does vary if evaluated for a diversified set of TPs [86]. Some of the algorithms, however, might be competitive for a diversified set of TPs [87].

2.2

Categories of HEMSs

Research on HEMSs over the last decade has mainly focused on DR only. Various schemes for tariffs and SHAs have been considered to implement DR. Various objectives, including reducing CE, T BD and P AR/peak load, have been formulated, and various optimization techniques, such as LP, MILP, and advanced heuristic methods, have been adopted to obtain the optimal solutions. Following the widespread installation of RESs and ESSs as part of optimal smart grid operations, researchers have begun to focus on optimal DR synergized with renewables (DRSR) and DRSREOD-based HEMS operations. Various algorithms have been presented for the integration of RESs and ESSs into HEMSs. Furthermore, work is also being done on the optimal sizing of HESs (PV/SB/DG) for HEMSs. Related work on HEMSs is discussed in this section under the categories of DR, DRSR, and DRSREOD with optimal DG sizing to cope with LS.

2.2.1

DR-based HEMSs

In most DR-based HEMSs, SHA scheduling is based on DS with the starting slot of the SHA operating time range as the preferred point or delayed/AS with an intermediate point in the SHA operating range as the preferred start time. T BD is computed from the distance between the actual start time and the preferred point. In [14], the authors present a method for the DS of SHAs based on a GA in which peak re-emergence is handled through an IBR scheme. Objectives are combined via the WSM to achieve CE and T BD TOs. In [7], a DS algorithm based on PSO is presented. The CE performance parameters are compared for RTP and 32


Chapter 2


ToU tariffs using IBR. A multi-stage ToU tariff is found to yield better results in reducing CE. In [52], a two-level HEMS framework is presented. SHAs are scheduled for delayed operation to reduce CE via communication with the operator. The operator solves a multi-objective problem to minimize the total load deviation to modify the desired demand from customers. The rewards offered to customers for each kW of deviation from the desired load are also minimized. In [53], the authors segregate the scheduling horizon for the HEMS into 4 windows. Appliances are classified based on their presence in the home, activity orientation and delay tolerance. The SHAs are operated in designated windows to increase user comfort by means of the combined CE and SHA delay results obtained through the WSM. In [63], an algorithm is presented for the optimal scheduling of SHAs, the charging/discharging of EVs and the preferred periods for NSHAs considering customer preferences. Objectives regarding CE and the interruption cost of the SHAs are combined using the WSM. Outages are managed using EVs to increase satisfaction. This research proposed the idea of MS for SHAs, classifying them as AS and DS to enable more SHAs to operate during off-peak/PV availability hours to reduce CE and T BD. MS not only enhances the satisfaction of HEMS objectives but also provides more diverse options to the consumer when shifting the operation of an HA. Our proposed algorithms demonstrate the advantages of MS over DS for DR-based HEMSs.

2.2.2

DRSR-based HEMSs

In [55], an algorithm for the optimal operation of SHAs and EVs considering predicted RES capacities and power-purchase agreements with retailers is presented. A diary of consumption showing interest in the usage of a SHA/EV in certain desired time slots for a DA is prepared by the consumer for all possible load profiles. To address the large number of possible combinations, a GA is used for optimal scheduling to maximize the difference between the amount the consumer could 33


Chapter 2


pay and the cost of obtaining that energy. In [56], an algorithm is proposed for a prosumer-based HEMS simultaneously participating in generation and efficient consumption. The past history of SHA operation is used as the predicted demand. Related appliances are clustered together for operation in three time windows. The frustration from delayed/AS of the SHAs and CE are combined. The profit from selling PV energy and the penalty cost to the consumer for not providing the promised PV energy are modeled. In [57], the performances of heuristic HEMCs based on GA, particle swarm and ACOtechniques are evaluated. Appliances are modeled as fixed, shiftable and elastic. The HEMS problem is formulated as a multiple knapsack problem. For tradeoff analysis, the WSM is used to combine the OBJFs for CE and T BD. Because the power supplied by RESs is intermittent in nature, the related dispatch problem is quite complex if they are used without an ESS.

2.2.3

DRSREOD-based HEMSs

An ESS is integrated into an RES-based HEMS to introduce dispatchability. In [58], an algorithm is presented for the priority-based scheduling of PV/SB/grid sources to maximize PV usage. In the absence of PV power, the SB is utilized with RT prioritization of appliances for operation. HAs are classified as controllable, with user-defined operating time intervals; semi-controllable, with flexible power usage; or uncontrollable. In [59], a mechanism for dynamic HEMS operation is presented. The SB is charged from the RESs, where SOC indicates the contribution from the RESs. Grid availability, SOC and the sign of the change in the SOC value are considered for optimal operation. During discharge, the operation of lower-priority appliances is shifted toward off-peak periods, and the air conditioner is operated at a high setting. In [60], a HEMS algorithm for prioritybased SHA scheduling based on RTP and RES management is presented. During peak hours, energy from the RESs and the SB is used to supply appliances, and 34


Chapter 2


during off-peak hours, the RESs are used to charge the SB while loads are supplied from the grid. In [61], a collaborative strategy for DR with offline scheduling for EV charging and bi-directional power utilization of the SB/EVs and PV sources is presented. The DR strategy is based on the preferred EV charging time. The net CE in terms of the difference between the costs of the energy bought from and sold to grid is minimized. The ESS is charged using PV energy or energy from the grid during off-peak periods and is discharged during peak periods. A penalty function is included to adjust the priority for selling energy in the following order: PV energy, SB energy and EV energy. In [63], an idea is presented for the participation of aggregated homes in locational-marginal-price-based DR through the shifting of dryer operation and the control of the target temperature for air conditioners. All percentiles of the annual marginal price data are chosen for appliance shifting/control. To further reduce electricity bills, PV and SB power supplies are considered to supply the shifted/controlled loads based on the PV availability, SOC, maximum charge/discharge rates and marginal prices. In [64], a method is proposed for the optimal operation of HAs in neighborhood homes with a two-step optimization strategy subject to power limits to ensure the fair usage of transformer capacity. All possible bi-directional power flows from the PV sources, the ESS, EVs and the grid are considered for each house and between neighborhood houses. In the first step, a uniformly distributed capacity is allocated to each home, and in the second step, CE is increased only for homes that require excess capacity allocation. Our proposed algorithm for a DRSREOD-based HEMS for a prosumer is based on the MS of SHAs synergized with the shared parallel operation of PV sources, the SB and the grid to maximize the HEMS objectives, including CE and T BD. The algorithm demonstrates the advantages of MS over DS for DRSREOD-based HEMSs. Direct/clear relations between the objectives are computed using a MOGA/PO approach for tradeoff analysis, enabling the selection of the most feasible tradeoff solution for the consumer. The proposed heuristic-based algorithm considers the PV availability, the SOC, the charge/discharge rates and a ToU tariff 35


Chapter 2


with an IBR strategy to generate optimal schedules/operating schemes for the SHAs, the SB and the grid.

2.2.4

DRSREOD-based HEMSs with optimally sized DGs to cope with LS

Present research is mainly focused on the design of a new hybrid energy system infrastructure for HEMSs. In [66], the optimal sizing of PV and SB units for a HEMS is investigated, incorporating the effect of DR-based load shifting. PV energy is used to supply the load or is sold to the grid. Batteries are charged from the grid during off-peak periods and feed energy to the grid during peak periods. A flat feed-in tariff equal to 67% of the peak-hour tariff is considered. Energy is sold in the following order of priority: PV energy, SB energy and EV battery energy. This DR-based PV/SB sizing results in a more economical design compared with the sizing with unscheduled loads. In [65], a harmony search algorithm is used for the optimal sizing of a PV/DG system. The surface area of the PV system and the nominal power of the DG are treated as the decision variables. The probability of power supply loss is used as the reliability index for the design of the hybrid energy system. Three new pitch adjustment mechanisms are introduced to enhance diversification and intensification in the algorithm. The results are compared with those obtained using the original harmony search algorithm, PSO and a GA. In [67], a scheduling algorithm is presented to minimize CE for energy from the grid while maintaining user comfort and P AR. Energy constraints are used to formulate knapsack capacity limits for each time slot to reduce P AR. Air conditioners are modeled for thermostatic control, SHAs are shifted using a GA, and fixed appliances are dispatched to a local DG with the minimum generation cost. A preferred intermediate position with adequate slots on both sides is proposed for the operation of each SHA to evaluate T BD in the advanced and delayed modes. In a number of developing countries with energy-deficient power 36


Chapter 2

2.3. EMISSIONS REDUCTION

supply networks, utilities are subjecting consumers to LSD to balance demand and generation. This research presents an innovative method for the optimal sizing of a DG to cope with LSD for a HEMS, considering the existing PV/SB infrastructure as well as CE and T BD TOs. The recent work related to our research on improved algorithms for DR-based HEMSs with MS, for DRSREOD-based HEMSs integrating the MS of shiftable appliances with optimal dispatch, and for the optimal sizing of a DG to cope with LS is summarized in Table 2.1 and Table 2.2, respectively.

2.3

Emissions reduction

With the installation of smart grid technologies enabling DSM, a widespread deployment of DR- and DRSREOD-based HEMSs has been carried out throughout the world in the past few years [88], [89]. In recent years, authors have presented various models and methods for the optimal operation of such systems as [53] and [63]. The objectives for optimal HEMS operation include minimizing CE, T BD, P AR, peak/ permanent demands and daily cost of generation. Further, utilities owning energy deficient power networks in developing countries are subjecting their consumers to LSD to balance demand and generation. In such power networks, consumers deploy a LSD-compensating DG in DRSREOD-based HEMS to ensure an uninterrupted supply of power [7]. The aforementioned objectives for optimal HEMS operation have been achieved using optimization techniques like LP, MILP, advanced heuristics, etc. Additionally, the issue regarding serious environmental concerns over the use of fossil fuels has been raised at international forums consistently in the past few decades. Recently, worldwide consensus has been reached to reduce the GHG emissions by selling them as commodities [49]. Such trading sets quantitative limitations on the emissions made by polluters that may include utilities, independent MG operators and the prosumers having local fossil fuel based generations. The 37


Chapter 2


Table 2.1: Related work relevant to the proposed HEMS algorithms for DR, DRSREOD and generator sizing Tariff+HEMS Components DAP/IBR + SHAs [14]

Objectives CE, P AR, T BD(D)

ToU/DAP/IBR + SHAs [78]

CE

ToU + SHAs [52]

CE

ToU + SHAs [53]

CE, T BD(D)

ToU/IBR + SHAs + EV [82]

CE, peak load, satisfaction

RTP + SHAs + EV + PV + WT [55]

CE

RTP + SHAs + PV [56]

CE, Frustration due to time shifting

ToU/IBR + SHAs + Elastic + PV [57]

CE, T BD(D), PAR

ToU + SHAs + Curtailable + Fixed + PV + SB [58]

CE, Peak demand

Salient features of the HEMS DS of SHAs; IBR to avoid peak re-emergence; TOs between CE and T BD(D) managed using the WSM DS of SHAs; CE compared for RTP and ToU tariffs with IBR

Achievements

Limitations

CE and P AR reduced by 15.3% and 25%

RESs and ESS not included

CE reduced by 39% for 3-stage ToU

Two-level framework; DS demand conveyed to operator, who minimizes load deviation arising from consumer demand and rewards for deviation Horizon divided into 4 windows; HAs classified in terms of occupancy, activity and delay tolerance are operated in designated windows; CE and T BD objectives are combined through the WSM for user comfort Optimal scheduling for SHAs and charging/discharging of EVs; Preferred periods for NSHAs; CE and the SHAs’ interruption costs are combined using the WSM Optimal scheduling of HAs and EVs; Usage diary showing interest in SHA and/or EV usage; Difference between budget and CE maximized Prosumer-based HEMS; Predicted demand; delayed/AS w.r.t. intermediate position; HAs clustered for operation in 3 time windows; Frustration and CE combined; Penalty cost for not providing PV energy Evaluation of HEMS algorithms based on GA, BPSO and ACO; Fixed HAs, SHAs and elastic HAs; Knapsack-based formulation; CE and T BD(D) combined using the WSM Priority-based resource scheduling; Maximized PV usage; SB used after PV; SHA operation based on RT priority adjustment; HAs classified as controllable, semi-controllable (flexible power) and fixed

CE reduced by 11.94%

T BD, RESs and ESS not included; PAR not reduced due to limited off-peak hours T BD, RESs and ESS not included

Gain of 0.185 for user comfort, compared with 0.149 for unscheduled loads

Fixed windows limit consumer convenience by requiring SHAs to operate in one window; RESs and SB not included

BPSO

CE reduced by 22% through optimal SHA scheduling

T BD and RESs not included

CPLEX solver

Reduction of 22% in CE obtained for a home in Spain

User T BD and ESS not included; EVs used as a load

GA

CE reduced by 11% for DR and further reduced through the sale of PV energy

Starting/ending operating time limits not modeled, affecting user convenience; SB not included

LP

GA-based algorithm outperforms BPSO and ACO for CE, T BD and P AR Savings in CE and sold units/day of 15.96% and 90, respectively

Only DS is modeled; SB is formulated but not simulated

GA, BPSO, ACO

T BD, P chg max and P dis max for SB and parallel PV/SB/grid operation not considered; P grid used only during off-peak periods, although it may be needed below SOC min

Heuristic based on resource priorities

38

Optimization method GA

PSO

MILP/ CPLEX solver


Chapter 2


Table 2.2: Related work continued. Tariff+HEMS Components ToU + SHAs + PV + WT + SB [59]

Objectives CE; Maximum RES usage

RTP + SHAs + PV + WT + SB + EV [60]

CE

DAP + PV + SB + EV [61]

CE

RTP/LMP + Curtailable + SHAs + PV + SB [63]

CE

DAP/ToU + PV + EV + SB [64]

CE

ToU + SHAs + PV + EV + SB [66]

NP V C

Fixed load + PV + DG [65]

N P V C, emissions

ToU + Fixed + SHAs + Thermostatic + PV + DG [67]

CE, P AR, T BD

Salient features of the HEMS SOC as indicator of RES contribution; Grid availability, SOC, and change in SOC used for optimal operation; Lower-priority HA operation shifted toward off-peak periods during SB discharge Priority-based SHA scheduling; RESs and SB supply HAs during peak periods; Load supplied from grid and RESs charge SB during off-peak periods

Achievements

Limitations

Cost savings for HEMSs with and without RESs are 25% and 28%

DR for EV scheduling integrated with SB/EV/PV power utilization; Difference between CE and CEsold minimized; SB charged from PV/grid during off-peak periods and discharged during peak periods; Penalty function adjusts priority of PV, SB and EVs for energy selling DR for aggregated homes; LMP-based HA shifting and AC temperature control; PV/SB integrated to supply loads based on PV, SOC, Pdis max and Pchg max HA scheduling integrated with PV/EV/SB in a neighborhood; 2-step optimization strategy for transformer capacity usage; Initial uniformly distributed capacity allocation; CE increased for homes needing excess capacity PV/SB sizing incorporating the effects of DR; PV energy used to supply loads or sold to the grid; SB/EV charged from the grid during off-peak periods and fed to the grid during peak periods; Energy sold in order of priority: PV, EV and SB Optimized sizing of PV/DG system; Probability of loss of power supply used as reliability index; Enhanced diversification and intensification applied in HSA Knapsack formulation with applied capacity limits to reduce P AR; Fixed HAs dispatched to DG; Preferred intermediate position for each SHA to evaluate T BD

Net cost reduction of 65% achieved with DR while shifting EVs during off-peak periods and selling PV, ESS and EV energy CE reduced by 9.5% for DR and by 28.6% for DRSREOD for 1000 homes

T BD, grid power usage during off-peak periods, and emergency supply during grid unavailability are not incorporated T BD and P dis max and P chg max rates for SB not included; Grid/PV/SB sharing not formulated SHA scheduling and T BD not included; Algorithm for parallel grid/ESS/EV operation for load sharing not included

39

Cost is reduced by 33%

Proposed scheme avoids transformer overloading while reducing peaks and CE

Sizing incorporating DR results in more economical design

Optimization method Heuristic based on SOC

BPSO

MILP/ CPLEX solver

T BD, parallel load sharing between grid and SB, and PV sharing of load and sold energy are not included T BD not considered; Algorithm for parallel grid/ESS/EV operation for load sharing not presented T BD not analyzed for sizing; DG for LS not included

LP

GAMS/ CPLEX solver

MILP/ CPLEX solver

HSA with more intensification yields optimal sizing

Size of DG in existing PV system not considered; DR TOs and SB not included in sizing

HSA

Maximum savings in CE, user comfort and reduced P AR

Consumer cannot opt for MS, thereby limiting consumer convenience; Tradeoff of T BD and CE for DG not analyzed

Knapsack + GA


Chapter 2


present scenario based on polluter pays principle has incentivized utilities to reduce not only the generation cost; however, the supply-side emissions as well while making use of the RESs installed for DRSREOD-based HEMSs [51],[90], [91]. Further, MG operators having RESs, ESSs and DGs also include T EM iss as an objective in the optimal dispatch scheme for their systems [62], [92], [93]. Furthermore, in energy-deficient power networks, DRSREODLDG-based HEMSs having LSDcompensating DGs are used to ensure an uninterrupted supply of power during LSD hours [7]. The operation of LDG in such HEMSs; however, does accompany the release of emissions, that needs to be minimized.

The section of related work includes the recent research on models and methods to achieve important objectives for DR and DRSREOD-based HEMSs including reductions in T EM iss (supply-side), CEnet, and T BD; for MGs including reductions in T EM iss and CEnet ; and for DRSREODLDG-based HEMS including reductions in T EM iss (local), CEnet and T BD .

2.3.1

Emissions reduction using DR-based HEMSs

Most of the research on DR-based HEMS has focused on objectives like CE, P AR, peak load and discomfort [53], [54]. Such systems have limited capabilities to play a role in the reduction of GHG emissions. In [50], a scheme for DR based HEMS is presented. Non-critical house loads are shifted towards off-peak hours to minimize the daily cost of generation and emissions for the supply-side. It is validated that implementation of DR program effectively reduces the cost of generation on the supply-side; however, the emissions on this side are reduced only when peak demand is met by peaking plants based on high emission fuels like coal, diesel, etc.

40


Chapter 2

2.3.2


Emissions reduction using DRSREOD-based HEMSs

Most of the models for DRSREOD-based HEMS presented in the recent past are based on optimal scheduling of SHAs integrated with the optimal dispatch of RESs and ESSs. HEMS problems for these models have been solved to reduce CE and discomfort for the consumer, and to minimize peak load/ PAR and cost of generation for the utility [56]-[63]. Recently, in the context of worldwide concerns over GHG emissions, authors have focused on the reduction in emission as an objective for DRSREOD-based HEMS. In [51], authors present a scheme for optimal scheduling of SHAs integrated with the optimal dispatch of RES, SB, and the power grid. The objectives include reductions in CEnet, temperature based discomfort, peak load, and the supply-side emissions. Such emissions are computed using emission coefficients for the energy mix adopted by the utility during various times of the day. An optimal dispatch of local RESs and SBs results in the reduction of net supply-side emissions by supplying the load during high emission hours. MILP has been used to solve the model. In [90], an operating mechanism of major HAs including heating and cooling appliances integrated with the optimal dispatch of PV and SB is presented. The algorithm for RT HEMS operation is based on user preferences, home occupancy, DA emissions and climate forecasts. The objectives for reduction in CE, electric consumption, T EM iss, and the peak demand are formulated. The net cost of emission includes carbon footprint of the customer from the grid electricity usage minus carbon reduction from injecting emission-free electricity from RES. In [91], a prosumer based algorithm is presented to maximize the sum of benefits to the consumer and the utility. The emission trading has been considered as a mean of mitigating this commodity. The utility is profited by reducing his carbon footprints while purchasing energy from locally installed RESs and ESSs during high emission times. The fitness function maximizes the welfare including consumption-based satisfaction and monetary benefits from RESs and ESSs to consumers and benefits of the reduced peak load, generating cost and 41


Chapter 2


emissions to the utility. A dynamic selling with dynamic buy-back pricing scheme is also proposed to implement the model. For scalability and user privacy, the problem is solved using Lagrange multipliers. The objectives in all of the above research are combined using the WSMD.

2.3.3

Emissions reduction in MGs

In MGs, RESs and ESSs are integrated with DGs to enhance the quality and the reliability of the power supply. In [62], a solution for DRSREOD-based HEMS operations for a stand-alone home including WT, DG, and SB is computed using PSO. The local fossil fueled DG is operated at rated power for an improved efficiency and reduced emissions. A separate objective function for emissions; however, is not included. An optimal dispatch for an MG is computed in [92] using GA. Additional constraints for ESS charge/discharge rates, DG start/stop and supply capacity are considered. Total emission is computed using EFTs for the grid, power supplied from the local DG and the ESS. The model does not include load shifting while computing the dispatch for power sources. A method to compute an optimal dispatch of RESs and DGs for a MG is presented in [93]. The dispatch is based on costs of energy from WT, PV and DG, EM iss and CE from/to main grid for a fixed load profile. The WT and the PV are the preferred sources. The SB is discharged based on its SoC if local RESs are not able to meet the demand; else, the load is supplied through the economic dispatch of the DG, FC, SB and the grid. Non-critical loads are disconnected when local sources are insufficient. The DG is operated at rated power to minimize EM iss. DR based load shifting is not included.

2.3.4

Emissions reduction in DRSREODLDG-based HEMS

In developing countries with energy-deficient power supply networks, utilities are subjecting consumers to LSD in order to maintain the balance between demand 42


Chapter 2

2.4. CONCLUSION OF THE CHAPTER

and generation of energy [7], [46]. While a number of consumers in developing countries are participating in DSM making use of DRSREOD-based HEMSs; LSD-compensating DGs are deployed in such HEMSs to ensure an uninterrupted supply of power. An algorithm for optimal sizing of an LDG for DRSREODbased HEMS was presented in our recent research [7]; however, such a DG does introduce emissions when operated during LSD hours. Based on the recent scenario for quantitative restrictions on carbon emissions, research on the optimized operation of DRSREODLDG-based HEMS focusing reduction in T EM iss looks pertinent. A simulation-based posteriori method for an eco-efficient operation of DRSREODLDG-based HEMS takes into account the TOs between CEnet, T BD, and minimal T EM iss is proposed. A three-step approach is followed. At step-1, primary tradeoff solutions for CEnet, T BD, and T EM iss are generated using a heuristic proposed for an optimal operation of DRSREODLDG-based HEMS. The heuristic, that uses MOGA/ PO to search optimal TOs, is detailed in algorithm 1. At step-2, an AVCF is used to filter out the TOs with extremely high values of T EM iss. Whereas, an ASCF is used to screen out the TOs with marginally high values of T EM iss at step-3. The ASCF was developed using advanced regression techniques.

2.4

Conclusion of the Chapter

Related work is summarized as: (a) Optimization is an efficient approach to implement I-HEMO- as well as CoHEMO-based models. I-HEMO-based models are locally defined, independent, converge rapidly and easy to implement; however, can result in peak rebounds. CoHEMO-based models are able to avoid peak rebounds; however, carry drawbacks like scalability, scheduling conflicts and compromised privacy. Such drawbacks may be avoided making use of enhanced meta-heuristics, parallel processing, coding methods, MASs and GT approach. (b) Most of the research reviews on HEMO are focused on DR; very few have reviewed research 43


Chapter 2

2.4. CONCLUSION OF THE CHAPTER

on HEMO for DRSREOD-based approach. (c) Recent research works on optimal HES sizing and optimal HES dispatch have been reviewed for the computing methods classification as conventional versus advanced heuristics; however, a few reviewers analyze DRSREOD-based HEMO focusing the same classification. (d) Very few reviewers have analyze DRSREOD-based HEMO for SOO versus MOO based classification, especially focusing on GA/EA based Pareto tradeoff analysis. (e) Related work on HEMSs is discussed under the categories of DR, DRSR, and DRSREOD with optimal DG sizing to cope with LS. (f) Finally, the state of the art pertaining to emission reduction is focused and elaborated.

44


Chapter 3 Survey

45

Chapter 3

3.1. OVERVIEW OF HEMO MODELING

Chapter summary This Chapter reviews the state of the art work in HEMO focusing on eminent models for DRSREOD-based HEMO in recent literature. The related schemes for HDs, dynamic tariffs, objectives/constraints, optimization algorithms/ techniques and validating platforms are presented. The main features for modeling and the relevant performance parameters are furnished. The HEMO models are analyzed for classifications from integration of RDESs (DR- versus DRSREOD-based), mutual coordination (I-HEMO versus COHEMO), uncertainty of data (stochastic versus deterministic), multi-objectivity (SOO versus MOO) and optimization techniques (conventional versus meta-heuristics-based). Further, the TOs among the approaches are also investigated based on the performance parameters achieved in the reviewed models. Challenges and issues in the field of HEMO are analyzed and solutions are proposed to resolve the issues.

3.1

Overview of HEMO modeling

A generalized model to solve the problem for DRSREOD-based HEMO constitutes the elements including HAs; RESs; ESSs; fossil fuel based DGs; dynamic tariff; control parameters; problem formulation based on control and decision parameters; OBJFs and constraints; coordinative requirements among consumers and algorithms/ optimization techniques to solve the problem [94]-[98]. The main objective of HEMO for DRSREOD-based HEMS is to compute the optimal schedules for SHAs and synergize them with the optimal dispatch of the RESs, the ESSs and the grid in order to reduce peak/overall demand for the utility and reduce CE for the consumer while keeping T BD within acceptable limits. A time horizons of adequate length is also specified in order to execute HEMO. The specified horizon is divided into N number of slots. The problem for HEMO in order to compute optimal schedules for SHAs (in terms of decision vector (T st)) is formulated as follows: 46


Chapter 3


P schd sh =

N X k X

X(a, n)

(3.1)

n=1 a=1

where the power term X(a, n) (kWh/ slot) for the ath SHA is computed based on the conditions,    P app(a) : f or T st(a) + LoT (a) > n ≥ T st(a) X(a, n) =   0 : f or T st(a) > n ≥ T st(a) + LoT (a) where P app is the vector of per slot power of SHAs. P schd sh is the scheduled power vector for SHAs. The load vector for NSHAs is added to the vector P schd sh in order to compute the net scheduled load vector, P schd. The OBJFs and constraints for HEMO problem are formulated as equalities or inequalities in terms of control parameters. These parameters are related to HAs, RESs, ESSs, DGs and dynamic tariffs. Control algorithms/ methods based on optimization techniques are devised to synergize the scheduling of HAs with the optimal dispatch of RESs units, ESSs and the power grid in order to optimize the objectives for HEMO while meeting the constraints. The problem for HEMO is solved in order to achieve a decision vector T st against the optimal tradeoff values for the objectives. Salient elements of HEMO model and the techniques to solve HEMO problem are presented in subsequent sub-sections.

3.1.1

HAs

HAs are designated for DR classifications as follows: (a) NCAs, e.g., lights and fans are operated as and when needed. Such appliances can not be included in scheduling. (b) SHAs can be shifted towards off-peak hours for optimal scheduling. Each SHA is to be operated for length of operation time (LoT ) between two limits. The consumer specifies these limits based on his own convenience. SHAs are classified into interruptible and non-interruptible. Appliances of the former type, 47


Chapter 3


e.g., pool pumps, air conditioners and electric geysers can be interrupted once started and hence may be operated in two or more separated sets of time slots. Whereas, the latter type of SHAs, e.g., washing machines and dryers are needed to be operated until completion of the job without interruption. (c) Curtail-able HAs are the ones that can be curtailed in power; however, cannot be shifted in time. Their usage can be reduced when EP is high. (d) Elastic HAs are the ones that are fully flexible both in terms of temporal shifting and power (or energy) consumption regulation; however, must consume a fixed quantity of energy within a given time frame to complete their jobs. (e) Model-based appliances (MBAs) control physical parameters to a specified level, e.g., thermostatically controlled air conditioner or heater. Such appliances control building temperature during each time slot based on temperature in the previous time slot and the required temperature in the next one [21, 99, 100].

3.1.2

RESs

Solar PV plants and the wind power units are the most widely used types of RESs in homes [59]. Their integration into a HEMS results in reductions in CE, the demand supplied from the grid, the peak load and GHG emissions. However, the power supplied by them is intermittent in nature and consequently, they are complex to be modeled for scheduling problems [51, 101]. Various methods, e.g., integration of ESSs are adapted to introduce dispatchability to RESs.

3.1.3

ESSs

These systems are integrated into RESs-based HEMSs in order to introduce flexibility in RESs dispatch. Normally, surplus energy available from the RESs is stored in ESSs or sold to the grid. The stored energy is later used to supply the load during peak hours to minimize the CE [102]. Integration of ESSs into

48


Chapter 3


RESs-based HEMSs results in improved parameters for economy, reliability and the power quality.

3.1.4

Fossil-based DGs

The supply of energy from a RES is intermittent in nature and lacks dispatchability. ESSs are used to store the excess energy from RESs for use during peak hours. Therefore, the amount of energy stored in an ESS greatly depends on RESs. DGs are integrated into HEMS to supply the emergency loads and to ensure the reliability of the power supply [7].

3.1.5

Electric Tariffs

Dynamic tariffs are the key to compute optimal scheme for HEMS operations. The major types of them include RTP, DAP and ToU tariffs. RTP is typically communicated by the utility to the consumer on an hourly basis, whereas DAP is communicated on a DA basis. ToU tariffs comprise two or more rates for the electricity during peak, off-peak and mid-peak hours of the day for a specified period (typically 3-6 months) [103, 104]. In combination with tariffs, utilities charge higher rates called IBR, at higher power levels. Such schemes are introduced in order to discourage consumers from concentrating their loads at off-peak hours. The schemes enable avoiding the re-emergence of peaks during scheduled operations. Multi-stage ToU tariff allows more optimized scheduling solutions as compared to two-stage ToU tariff [78]. ToU- and DAP-based algorithms provide solutions in very little computation time are viable for real-time household applications [105]. On the other hand, RTP-based algorithms offer more optimal solutions; however, need large computational time due to the uncertainty of data and problem complexity [106]. Dynamic tariffs [107]-[115] offered by various utilities are given in Table 3.1.

49


Chapter 3

3.1. OVERVIEW OF HEMO MODELING Table 3.1: Dynamic Tariffs

Tariff RTP RTP/IBR DAP

DAP/IBR ToU ToU/IBR

3.1.6

Utility/ Operator New York independent system operator (NYISO); ISO New England Illinois Power Company (IPC) Commonwealth Edison, Illinois based utility; NYC Company; Ontario EP; EU DAP data www.nordpoolspot.com Ameren Illinois power company Multi-rate tariffs in Turkey, 3 Stage ToU by Baltimore NYISO Pacific gas and electric/ San Diego gas and electric/ Southern California Edison

References [107, 108, 109] [110] [105, 111, 112, 90] [14, 113, 114] [59, 104, 115] [54]

Main Objectives for HEMO Problem

The main objectives for HEMO problem generally include minimizing the CE for energy from the grid, minimizing the T BD for the consumer, reducing the peak load P AR and reducing GHG emissions. To achieve the aforementioned these objectives, the HEMO problem is formulated for SHAs scheduling while simultaneously computing P schd and synergizing the scheduling with RES, ESS and power grid dispatch for N time slots over a specified scheduling horizon.

3.1.6.1

Minimization of CE

For DRSREOD-based HEMO problem, the OBJF to minimize the net CE to be paid by the consumer is formulated as follows: N X M inimize (P E × EP − SE × SP )

(3.2)

n=1

where P E, EP , SE, and SP are the energy purchased from the utility, EP per KWh, energy sold to the utility and feed-in electricity price, respectively [21].

50


Chapter 3 3.1.6.2

3.1. OVERVIEW OF HEMO MODELING Minimization of Discomfort Level

Discomfort is a measure of the loss of quality of service based on the delivery of energy. It may be based on time shift of an HA from its preferred operating time or deviation from an ideal state like the preferred temperature of a room, etc. Various measures: called penalty functions are used to estimate the discomfort level. Penalty functions are classified based on the time shifts [14, 105, 110, 111] , state deviations [108, 115, 116], binary decisions and linear parameters for undelivered services [117, 118]. The OBJF to minimize the discomfort due to time shifting of SHAs is expressed as follows:

M inimize

K X

((T st − Alpha)/(Beta − LOT − Alpha + 1))/K

(3.3)

a=1

where T st, Alpha, Beta, LoT and K are the vectors for scheduled starting times, proposed limits for the start and stop times of SHAs, length of operation of shiftable home appliances and the total number of SHAs. The parameters are computed by taking the average of the normalized delays for all of SHAs.

3.1.6.3

Maximization of the Usage of RESs

The OBJF to maximize the fraction of demand covered by local renewable generations is formulated as follows:

M inimize

N X

(Eren − Enet)/(Hs × P tot)

(3.4)

n=1

where Eren, Enet, P tot, and Hs are per slot energy generated by RESs, net generation of the energy in the system, total power demand and hours/slot, respectively [21].

51


Chapter 3 3.1.6.4

3.1. OVERVIEW OF HEMO MODELING Minimization of P AR

The OBJF to minimize the P AR of the demand to be supplied by the grid is expressed as follows:

M inimize(P eak(P L)/Average(P L))

(3.5)

where P L is the scheduled load to be supplied by the grid.

3.1.6.5

Minimization of GHG emission

The OBJF for minimizing the emissions from the fossil fuel based DG is formulated as below: M inimize

N X (F cons × EF )

(3.6)

n=1

where EF (Kg/Liter) is the emission factor for the consumed fuel F con for nonrenewable generations [51].

3.1.7

Constraints for HEMO Problem

Constraints for HEMO problems are based on the characteristics of HDs, tariffs, consumer requirement, etc., and are to be taken into account. Some major constraints for HEMO problem are presented in subsequent sub-sections.

3.1.7.1

HA Constraints

Scheduling constraints are imposed on the HAs to satisfy the user’s preferences; these constraints include defined time deadlines for the completion of operation [14] and non-interruptibility constraints [111, 119], etc. These constraints are implemented by introducing upper and lower bounds on the T st vector.

52


Chapter 3 3.1.7.2

3.1. OVERVIEW OF HEMO MODELING Tariff Constraints

The tariffs issued by some utilities impose maximum power consumption limits and offer lower rates for respecting those limits [14].

3.1.7.3

ESS Constraints

The SB is subject to constraints on its SOC which must be within certain minimum and maximum allowable levels [120]. Another constraint can be imposed for reserving a certain amount of energy in SB at the start of next day or at any instant to satisfy critical loads [99].

3.1.7.4

RDESs Constraints

Constraints are laid to ensure that energy injected into the grid is not greater than excess renewable generation exceeding the load [21] and a diesel generator once started or stopped to be remained in that state for a certain minimum time [120].

3.1.7.5

Constraint for the Balance of Energy

This constraint ensures that in each time slot, the total energy generated is equal to the total energy consumed by the load, or the sum of the energy inputs is equal to the sum of the energy outputs for the system.

3.1.7.6

Other Constraints

In addition, the constraints can be imposed on the time/ temperature based comfort level [114], budget for the daily energy consumption [121], life cycle cost of ESS [21], etc.

3.1.8

Scheduling Resolution

Scheduling efficiency depends on homogeneity of resolution of data and resolution for scheduling. Existing methods optimize the HEMS operation for a specified 53


Chapter 3


scheduling horizon with a fixed time resolution, typically one hour. Parameters like loads and solar power are assumed constant throughout each step. Low resolution is adequate for hourly prices; however, inadequate to capture variation in solar power which needs higher resolution. The scheduling algorithm must account for higher frequency fluctuation of data with high resolution for scheduling; however, it increases computational burden. Conversely, low-resolution results in suboptimal schedules [122]. Increased resolution up to 10 min is a tradeoff that addresses above mentioned issues and results in optimal computing cost using heuristics [62].

3.1.9

Modeling Uncertainty of Data

Various parameters included in HEMO modeling carry uncertainty. Based on uncertainty, the models are classified as deterministic versus stochastic. The former assigns forecasted values of uncertain parameters as ex-ante; e.g. forecasts for RESs, HAs preferences, environmental parameters, NCAs operation, EV availability, etc. Forecasts are never fully accurate; introduce errors and result in compromised efficiency of the scheduling model. Various methods are used to address uncertainty in scheduling process to improve scheduling inefficiencies. A detailed comparison of the methods is presented in Section-5.5.

3.1.10

Coordination in HEMO

Based on coordination, HEMO models are classified as I-HEMO and CoHEMO. IHEMO-based HEMSs are locally defined, independent and are fast in convergence. The approach may cause peak rebounds as joint effect of scheduling on the grid is not considered. CoHEMO-based systems inherently avoid peak rebounds due to joint RM. CoHEMO can reduce consumers’ bills and the system peak demand with a limited impact on consumer’s discomfort more efficiently as compared to I-HEMO [21]. Coordination in HEMO based on the configuration are classified 54


Chapter 3


as follows: (a) CC, where a CO manages electricity usage of all homes. CO computes optimal schedules for HDs to implement CoHEMO strategy based on the information provided by individual homes. (b) DC, where the end users schedule their HDs directly after gathering profile information of the others [39].

3.1.11

Techniques to Solve Problem for HEMO

Most of the problems for HEMO are non-linear (NL), non-convex constrained and highly multi-dimensional in nature. Such problems have a large number of solutions that grow exponentially with the problem size [55] when solved using conventional techniques. In addition, these problems can also be solved as a combinatorial optimization. Techniques use to solve such problems are classified as conventional techniques versus advanced heuristics-based as discussed in the subsequent subsections.

3.1.11.1

Conventional Techniques

Conventional techniques include the following: • LP techniques are used to solve linear problems that consist of linear OBJFs and constraints. The design vector Xn is computed to minimize the OBJF subject to the laid constraints as follows:

M inimize

N X

(Cn × Xn)

(3.7)

n=1

where Cn is a constant. LP techniques give solution in polynomial-time (PT) for small-scale problems. The most popular LP techniques include simplex and interior point methods. Linear problems may not be able to represent HEMO accurately due to non-linearity of OBJFs including parameters for loads, energy prices, etc.

55


Chapter 3


• NLP techniques are used for problems having nonlinear OBJFs or constraints. The NL problems, if convex in nature, can be solved using LP techniques. • Quadratic programming (QP) techniques are used for quadratic problems that have quadratic OBJFs and linear constraints. Such problems are formulated to find Xn while minimizing the OBJF. Problems with positive definite objectives as per Karush-Kuhn-Tucker conditions are said to be convex. Such problems can be transformed into linear function and can be solved using LP techniques, e.g., Lagrange function. For indefinite objectives, problem becomes NP-hard. • CP: A function f (X) is said to be convex, if for any pair of points X1 and X2 and for all h, 0 ≤ h ≤ 1, the following condition is satisfied:

f (h × X2 + (1 − h) × X1 ) ≤ h × f (X2 ) + (1 − h) × f (X1 )

(3.8)

For convexity, the line segment joining any of the two points lies completely above or on graph of the function as per equation 4.3. A convex problem has a convex OBJF, linear equality constraints, and concave inequality constraints. In CP technique, the convex problem is transformed into a linear problem to obtain a solution. In addition to conventional LP methods, gradient, grid search and secant methods are also used to solve convex problems. If a solution exists, the method is sure to be converged. • MILP: LPs; including integers as well as continuous variables are called mixed integer linear (MIL) that are NP-complete in complexity. Such problems require enumeration based algorithms for MILP, e.g., branch-and-bound (BB) for solution. • MINLP: Problems for MINLP can be very difficult to solve and may not guarantee that a solution is found even if it exists [24].

56


Chapter 3


• DP for a problem is used to find Xi =[X1 , X2 , ..., Xk ] that minimizes the function f (Xi ) as follows:

M inimize

n X

Ri =

i=1

n X

ri (si+1 , xi )

(3.9)

i=1

DP is based on the concept of sub-optimization and principle of optimality. The original problem is broken down into a number of sub-problems. Optimal solutions to smaller sub-problems are recursively calculated and used to obtain the feasible solution to the original problem. Ri ,xi ,si are return function, design variables and state parameters, respectively. DP solves problems in O(n2 ) or O(n3 ) times for which a simple approach takes exponential time. The COP techniques have been used for linear and convex problems successfully for small-scale HEMO problems. For large scale, non-linear and non-convex problems conventional methods become computationally impracticable and problems appear as NP-hard [122]. Advanced heuristic-based techniques have emerged as optimal choices for nearest optimum solutions to such problems.

3.1.11.2

Advanced Meta-heuristic Techniques

Over the last decade, heuristic tools have been used very successfully for the robust solutions of HEMO problems. Meta-heuristics are the general form of heuristics applied to a large number of problems with minor modifications for specific cases. The tools have solved optimization problems that were believed to be impossible in the past, such as non-convex and NP-hard problems in very short computational times [123]. Meta-heuristics used in recent research to solve HEMO problems include the followings: • GA search the solutions on the basis of natural selection and genetics. It searches multiple paths for maxima (or minima)and escapes from local solutions by means of niching method [124]. It uses parameter coding instead of actual parameters, thereby enabling it to develop the next state from 57


Chapter 3


the current state with minimal computation. A GA evaluates the fitness of each string to guide its search after evaluating the performance of one or more fitness functions. To handle constraints, a GA uses chromosome rejection, repairing and other genetic operators. GA performance decreases for multi-constrained problems where solutions are on constraints boundary. • ES differs from GA in that ES uses floating point parameters while GA operates on binary strings. GA relies on crossover while ES uses mutation to move in search space. EA has the natural capability for constraints handling and is suitable for optimization of real-valued function where fitness landscape is rugged and numerous local solutions exist [123]. • DE uses differences of pairs of objective vectors for mutation process. Provision of topographical information for OBJFs in DE results in improved global optimization capability. Crossover in DE is non-uniform and based on child vector parameters. It is a robust, accurate and fast method to solve nonlinear constrained optimization problems. • PSO is based on swarm behavior of birds and fish schooling called swarm intelligence. It generates populations of random solutions. Each solution called a particle, is allocated a randomized velocity to move the particle in solution hyperspace. PSO is very simple to implement and solves even the most difficult MINLP problems very quickly. Constraints are taken using strategies for adjustment of individual position within the boundary and dynamic space reduction [123, 125, 126]. • The ACO is based on ant’s method to search food, i.e., a pheromone on the shortest path will accumulate faster than on any other. The algorithm uses a probabilistic model to generate solutions. Parameter values for the probabilistic model are updated at run-time. High-quality solutions are generated over time [127]. • TS is based on the gradient-descent search. The method uses memory that

58


Chapter 3

3.2. STATE OF THE ART IN HEMO MODELING

stores previously visited states in a Tabu list. Aspiration helps in the inclusion of all neighboring states of a present state in the list. Overridden Tabu obstacle enables state with better fitness function to move out of local minima. Diversification introduces randomness in this deterministic search. Special constraints are easily handled by TS in coding even they are not easily described using the algebraic form. • SA is an excellent method to solve large sized and complicated problems without differentiability, continuity and convexity requirements. It is based on annealing process to form high-quality crystals. SA functionality is based on perturbation mechanism, cost function, solution space and cooling schedules. Local minima is escaped using cost function with smoothing approach [123].

3.1.11.3

Expert Systems

Expert Systems (ES) are the methods for optimal decision making based on defined rules. ES does the job of human beings with higher availability, reliability, lower cost and response time. Commonly used ESs are ANN and fuzzy logic. ANN function is based on the genetics of brain that makes use of back and forth propagation. ANN has the capability to recognize patterns that appear too complex to other computational techniques. Fuzzy logic function is based on the concept of partial truth. Data uncertainty is introduced through membership functions in crisp form. Fuzzy set allots degree of membership (0,1) to each element. The fuzzy system applies sets based on fuzzy rules to nonlinear complex problems [128]. In HEMO modeling, ESs have been used to forecast uncertain parameters [103, 129].

3.2

State of the Art in HEMO modeling

This Section presents eminent models on HEMO taking into account the most recent literature. Major elements of HEMO modeling including HAs; RESs; ESSs; 59


Chapter 3


fossil-based DGs; dynamic tariffs; objectives functions/ constraints; stochastic approaches; coordinative approaches; optimization method/ techniques and simulation platforms are analyzed. Salient features for HEMO modeling and the performance achieved for each model are furnished. The models are classified for dichotomous approaches as DR versus DRSREOD; deterministic versus stochastic; I-HEMO versus CoHEMO; SOO versus MOO; and conventional techniques versus advanced heuristic-based optimization. The performance parameters exhibit the efficacy of the approaches for the proposed models. The main classification [130][172] (Tables 3.2-3.5) for DR versus DRSREOD-based are reflected, respectively. Table 3.2: Models for DR-based HEMO (Deterministic) Ref. Tariff+HAs+ Objectives Salient features of HEMO

Performances/Optimization/

Type of co-

Achieve-

Simula-

ordination

ments

tion method

[104] 3

level

Cost

of

Control of EV charging lev-

50%

more

LP/

ToU+

charging

els based on tariff, Controlled

EV

sup-

IEEE 37-

SHAs

EV

and

charging (C/C) is investigated

plied

with

node test

(EVs)+

peak load

on test feeder selecting various

C/C

as

number of houses and %age

compared

penetrations of EV

to 30% w/o

I-HEMO

feeder

C/C [78] DAP/IBR+ SHAs+ HEMO

I-

CE

SHAs scheduling; CE com-

39% reduc-

LP/PSO/

pared for RTP and multi-stage

tion in CE

MAT-

ToU tariffs using IBR; 3 stage

LAB

ToU tariff enabled a maximum reduction in CE

60


Chapter 3

3.2. STATE OF THE ART IN HEMO MODELING Table 3.2: Models for DR-based HEMO (Deterministic)

Ref. Tariff+HAs+ Objectives Salient features of HEMO


Type of co-

Achieve-

Simula-

ordination

ments

tion method

Scheduling considering prefer-

47% reduc-

MILP/CPLEX

SHAs+

ences for SHAs timings; SHAs

tions in CE

(BB, cut-

I-HEMO

operation distributed in dis-

ting

tinct phases; An energy phase

Plane)/

is not started unless finishing

MAT-

of preceding one

LAB

[105] DAP+

CE

[14] DAP/IBR+ SHAs+

I-

HEMO

NLP/GA/MATLAB

SHAs scheduling;

P AR and

avoid

T BD

peak; T BD for delays from

ductions in

preferred starting time; WSM

CE/P AR

re-emergence

IBR to

15.3%/

CE,

of

the

19.7%

re-

used for TOs between MOs CE,

SHAs scheduling; For the de-

1.3%/ 25%

MILP/

SHAs+

P AR,

sired reduction in CE, a pa-

reduc-

Interior

I-HEMO

and T BD

rameter introduced in T BD

tions

function for delayed opera-

CE/P AR

[110] RTP/IBR+

in

Point/CPLEX

tions, Discrete values for strict/medium/ no cost reduction may be assigned; WSM used for TOs

61


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

and

Scheduling based on daily en-

7%

peak load

ergy requirement and con-

tion

in

sumer preferences; Loads clas-

CE

with

sified as (a) time changeable

maximum

that operates as per power in

reduction in

specified sequence (b) power

peak load

[115] ToU+SHAs+ CE I-HEMO

reduc-

MILP/goal programming

changeable that operates between max/min power and (c) non-changeable that operates as per preferences

I-HEMO

re-

MILP/

and

SHAs scheduling based on

43.7%

Peak load

power threshold and consump-

ductions in

CPLEX

tion forecast; Optimization is

CE

(Branch

[130] ToU+SHAs+ CE

performed when an event trig-

and cut)/

gers system and initiated con-

MAT-

troller actions handle real-life

LAB

dynamics of the household [131] RTP+

CE

and

Scheduling algorithm to find

-

Binary

Ther-

cooling

starting times of SHAs and

LP/TS/

mostat+

/time

set points of permanent ser-

Nu-

I-HEMO

based dis-

vices through controlled phys-

merical

comfort

ical variable for the NP-hard

Simula-

problem;

tion

Energy consump-

tions satisfied maximal power constraint

62


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

CE, peak

Optimal scheduling for SHAs

22% reduc-

CPLEX

SHAs+EV+ load and

and charge/discharge of EV;

tion in CE

Solver

I-HEMO

satisfac-

Preferred periods for must run

tion

HAs; CE and SHAs inter-

Scheduling based on division

User

LP/BPSO

of horizon in 4 windows; HAs

fort

classified for occupancy, ac-

creased

tivity and delay tolerance are

from 0.149

operated in designated win-

to 0.185

[54] ToU/IBR+

ruption cost combined using WSM [53] ToU+SHAs+ CE I-HEMO

and

T BD

comin-

dows; CE and T BD combined through WSM for user comfort [132] DAP+SHAs+ Exchanged Hybrid thermal+

energy

elastic+

with

I-HEMO

grid

based the

energy HEMO

appliances for

optimal

Reduction

QP/greedy

in

algo-

energy

energy usage ensuring energy

exchange

rithm,

supply for electricity and heat

by

GA/

demand;

greedy

CHP and heaters

GA/ is

MAT-

run in 100%, 50%, 20% and

58.98/47.86%; LAB

idling modes; Greedy algo-

Greedy

rithm

far faster

proposed

computational

to

reduce

is

complexity;

Results compared with other meta-heuristics

63


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

[107] RTP+SHAs+ CE

for

A CC based HEMO frame-

13%/

12%

GA, Stack-

CoHEMO

retailer

work for the retailer and

reduc-

(CC)

and

consumers as a 1-leader, N-

tions

follower game based on GA/

CE/P AR

game/C-

heuristic; Bill reduced through

for

PLEX

SHAs scheduling;

sumer

the

user

Retailer’s

in

con-

elberg

profit increases through reducing P AR when RTP used as compared to flat tariff; A two way communication considered

CoHEMO (CC)

LP/BPSO/

and

A CC based SHAs schedul-

Maximum

peak load

ing participated by neighbor-

22.2%

ing consumers and retailer;

19.74%

Three schemes are discussed

ductions in

comprising: (a) unscheduled

CE/P AR

(b) SO viz CE for the con-

for (c)

[133] ToU+SHAs+ CE

/ re-

MATLAB

sumer, reduced P AR for the retailer (c) double objectives viz tariff for the user and minimum power for the retailer

64


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

[134] DAP+SHAs+ CE CoHEMO

and

T BD

(CC)

A queuing based framework

An average

NLP/

for

reduction

RELOAD

HEMO; Optimal scheduling

of 39% in

(DP)

of

CE

centrally

essential

coordinated

loads

classi-

using

fied as delay-sensitive with

proposed

higher

scheme

and

delay-tolerant

with lower priorities; Delays for

delay-tolerant

can

be

reduced upgrading it for a high-priority queue with given probability [135] RTP+SHAs+ CE

and

A CC based HEMO frame-

Well spread

LP/MO-

CoHEMO

energy

work; Consumers provide a

population

EA using

(CC)

based

discomfort level to the retailer

reduces

PO/

comfort

to participate in DR; Retailer

crowding

MAT-

proposes scheduling and cor-

and

LAB

responding cost for a set of

puting cost;

homes; Efficient MO-EA de-

CE reduced

com-

veloped based on dominance and crowding distances used for offspring selection; TOs for cost/discomfort drawn as PFs.

65


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

[136] CoHEMO

CE

(CC, DC)

DP

and

SHAs

optimal

scheduling/-

4.71%

re-

source sharing for neighbor-

ductions in

hoods;

CE

Optimization for a

LP/GA

using

group of SHs is further coor-

proposed

dinated at neighborhood level

scheme

through a scheme used by the network operator; Objective load suiting each group’s demand provided; Difference b/w scheduled and objective load minimized Scheduling based on the two-

11.94%

MILP

CoHEMO

level framework;

reduction in

(CPLEX

(CC, DC)

demand conveyed to the oper-

CE

Solver)/

ator who minimizes load de-

proposed

viation from scheduled de-

scheme

[137] ToU+SHAs+ CE

Scheduled

using

CPLEX

mand and rewards for deviation; WSM used to combine objectives

66


Chapter 3




Type of co-

Achieve-

Simula-

ordination

ments

tion method

[138] RTP+SHAs+ T BD and

Peak load for aggregated de-

Peak

peak load

mand of several houses have

flattens

stat+

minimized; Scheduling flexi-

shifting

CoHEMO

bility with delayed, advance

SHAs up to

(DC,

CC

and mixed directions w.r.t pre-

a threshold

and Mixed)

ferred times; Mean squared de-

only;

viation of schedules from pre-

approach to

ferred timings used for penalty

reduce CE

Thermo-

load

LP

An

factor for convenience; CC, DC and mixed approach for CoHEMO

Table 3.3: Models for DR-based HEMO (Stochastic) Ref. Tariff+HAs+ Objectives Salient features of HEMO


Type of co-

Achieve-

Simula-

ordination

ments

tion method

[111] DAP+SHAs+ CE I-HEMO

A

scheduling

algorithm

34-40% re-

for interruptible and non-

ductions in

interruptible loads based on

CE

SDP/MDP

SDP formulated as MDP; Price uncertainty considered

67


Chapter 3


[108] RTP+SHAs+ CE

mechanism

24-27% re-

MILP

dealing uncertainty in RTP;

ductions in

(CPLEX

Stochastic and robust op-

CE

solver)

An algorithm to obtain op-

20% reduc-

Immune

tions in CE

colonial

A

I-HEMO

scheduling

timization approaches compared; Robust scheme shows better computational performance while stochastic ones result in better cost reduction;

Coordination of two

approaches WSM

used

recommended; to

combine

CE/risk to increase CE due to uncertainty [116] DAP+

CE

and

thermo-

cooling

timal temperature scheduling

static+

based dis-

for air conditioner (AC) ac-

I-HEMO

comfort

cording to DAP and out-

GA

door temperature forecasts; To manage uncertainty in EP and temperature for 24h advanced prediction, parameters are modeled using fuzzy set

68


Chapter 3


An MPC based HEMO model

20-30% re-

QP

cooling

with multi-stage optimization

ductions in

slow,

based dis-

to incorporate uncertainty of

CE as com-

Integer

comfort

HVAC state; Kalman state

pared with

SP

prediction used for forecast;

MPC

fast time

[109] RTP+SHAs+ CE I-HEMO

and

at

at

Problem bifurcated into slow

scale/

and fast scale optimizations

Numeri-

for information received at

cal

different times, e.g., hourly energy for devices and one minute resolution for thermal data A scheduling algorithm based

44.6-56.9%

NLP

cooling

on machine learning and data

reductions

HVAC,

based dis-

structure;

in CE

IP for de-

comfort

nism for HVAC developed and

ferrable/

integrated with DR-based op-

MAT-

timization to regulate HVAC

LAB,

loads with temperatures in

eQUEST

[143] DAP+SHAs+ CE I-HEMO

and

Learning mecha-

comfort range; Data structure stores and captures current HAs behavior

69


for

Chapter 3


An algorithm to shift non-

3.7%/ 20%

Heuristic

GHG

critical house loads to min-

reduc-

and SP/

emissions

imize daily generation cost

tions

and GHG emissions for supply

CE/ GHG

side; DR effectively reduces

emission

[50] DAP/ToU+ CE SHAs+

I-

HEMO

and

in

Software based

supply side generation cost; however, emission is reduced only if coal used as peaking plants [139] RTP/IBR+ SHAs+

CE P AR

I-HEMO

and

A multi-stage load manage-

15.7/ 25.5%

ment algorithm considering

reduc-

load uncertainty, Statistical

tions

estimates of future load de-

CE/P AR

MILP

in

mand included; On revealing information of appliances over time, operation schedules are updated [140] RTP+SHAs+ CE I-HEMO

in-

QP/Lagrange

An optimal RTP scheme is de-

LU

rived for service providers un-

creases

multipli-

der three types of load un-

optimal CE

ers (LMs)

certainty (LU) models; Effect of LU on power consumption and generation capacity shown; Unknown load distribution model incurs an optimal price higher than the Gaussian model

70


Chapter 3


[141] DAP/RTP+ CE

and

A two-step linear-sequential

20% reduc-

NLP/

tions in CE

Software

thermo-

heating

fast, robust and flexible algo-

stat+

based dis-

rithm for optimal scheduling

I-HEMO

comfort

of water heaters; First step

based

gives optimal schedules; Second provides adjustments to optimal schedule to account for uncertainties for forecast errors in EP / hot water usage A stochastic model of HAs

Survival

Correlation

HAs+

load; A comparison of strate-

multi-state

using

I-HEMO

gies for discrete-time MDP

approach

Pearson’s

and survival analysis; Predic-

outper-

coeffi-

tion based on time-varying

formed

cient

fractional

MDP

[142] Dynamic+

-

power

demands;

Rigorous procedures used to assess how effectively results are generalized to different data sets;

For data with

un-identified trends, the algorithm works as an outlier filter

71


Chapter 3


[114] DAP+

CE

to

A framework for neighbor-

28.6%

Elastic+

retailer/

ing homes’ collaboration to

duction

CoHEMO

consumer

balance

CE

(DC)

ply;

demand

and

sup-

rein

problem

Stochastic/ MDP,

Multi-stage stochastic

optimization

MS-

MC,

for

demand uncertainty; Monte

LMs/

Carlo (MC) used to deal with

MAT-

uncertainty; Scheduling based

LAB

on message exchange between neighbors

Table 3.4: Models for DRSREOD-based HEMO (Deterministic) Ref. Tariff+HDs+ Objectives Salient features of HEMO

Performances/Optimization

Type of co-

Achieve-

/Simu-

ordination

ments

lation methods

CE,

A versatile HAs scheduling al-

Performance MINLP,

WT+SB+

T BD and

gorithm based on CP, Conver-

better than

NP-hard/

I-HEMO

maximum

sion of MILP to CP based on

CPLEX for

Interior

usage

the characteristics of HAs re-

large-scale

point

sulted in efficient scheduling

problems

method/

[100] RTP+PV+

RESs

of

CP Solver

72


Chapter 3


28%/

PV+WT+SB+maxi-

ations based on grid availabil-

reduction

SOC

I-HEMO

ity, SOC and change in SOC

in CE with

based

in a slot; SB supplies loads

/without

heuristic/

for SOC greater than maxi-

RESs

Exper-

and

mized usage

of

RES

25%

NLP/

A HEMO model for HAs oper-

[59] ToU+SHAs+ CE

mum SOC; Loads are shifted

iment

towards off-peaks for SOC

homes

less than minimum SOC; For intermediate SOC/ SB discharged, SHAs of lower priority are shifted towards offpeaks and AC setting is increased [102] ToU+SHAs+ CE

and

An

optimum

loads/sources

Reduction

LP, NLP

Data

in

/

PSO

maximum

management;

I-HEMO

usage

tainty for PV, load and EP ;

(LP/AD-

with

LP, action dependent heuristic

HDP/

learning

DP (ADHDP) and PSO are

PSO):

compared ;

LP algorithm

terministic:

anism,

offered the best solution with

7/6.6/6.1%,

ADHDP

low complexity; For extended

Stochastic:

models for NLP like ADHDP

5.3/4.9/4.3%

RESs

of

and

offline

uncer-

CE

PV+SB+

learning

De-

mech-

based

PSO validated; UF minimizes difference

between

demand

and the supply of energy from RESs, SB and the grid.

73


Chapter 3


[62] RTP+SHA+ Generation An algorithm for HEMO so-

PSO

con-

LP/

WT+MT+SB+ Cost, MT

lutions for stand-alone HES

verges

90

PSO,

I-HEMO

efficiency

including WT, micro turbine

times faster

SQP

and GHG

(MT) and SB; PSO converged

than SQP

MAT-

emissions

90 times faster than SQP

LAB

which makes it suitable for HEMO for multisource systems An algorithm for optimal op-

22% reduc-

NLP

EV+PV+WT+

eration of SHAs, EV with pre-

tion in CE

/GA

I-HEMO

dicted RTP, RESs and agreed

/Exper-

power-purchase with retailer;

imental

Matrix showing all possible

home

[55] RTP+SHAs+ CE

load profiles for SHAs and EV prepared; Difference between consumer budget for specified energy and CE to be achieved is maximized [129] RTP+Fixed+ CE SB+ HEMO

I-

A strategy with SB and grid

30.5%

configuration to reduce CE

duction

for a specified load; Optimal

CE

rein

NLP/ ADHDP

charge during low price and discharge during high price hours to reduce CE; ANN used for self-learning and function approximation to implement ADHDP

74


/

Chapter 3


An algorithm for prosumer

4.2-11%

frustra-

based HEMO to participate in

reduction in

tion

generation and efficient con-

CE

[56] RTP+SHAs+ CE PV+

I-

HEMO

and

for

T BD

LP

sumption; Past history of HAs schedules used for predicted demand pattern; Scenario for HEMO with/without PV analyzed

[144] RTP+SHAs+ Cost and

An algorithm with separate

Different

MILP,

thermostat

comfort

subroutines for control of de-

weights for

NP-hard/

+PV+SB+

(climatic/

mand, SB and thermal load

objectives

heuristic/

I-HEMO

shifting)

based on predicted EP and

in WSM are

CPLEX

PV; Comfort for thermal and

discussed

delay from preferred appliances schedules are considered;

Objectives

combined

through WSM CE, peak

An algorithm for the op-

22%/ 10.5%

MILP/

SHAs+EV+ load and

timal scheduling of SHAs,

reduction in

CPLEX

I-HEMO

satisfac-

charge/discharge of EV and

CE/ P AR

solver/

tion

preferred periods for fixed

[54] ToU/IBR+

HAs;

CPLEX

Consumer preferences

for SHAs considered; Load interruption cost/outages reduced using EV to increase satisfaction; Objectives combined using WSM

75


Chapter 3


T BD and

based algorithms based on

performs

BPSO,

Elastic+

P AR

GA, BPSO and ACO; Fixed,

BPSO/

ACO

SHAs and elastic HAs con-

ACO

sidered;

HEMO

Knapsacks

formulation;

CE/

HEMO

LP/ GA,

SHAs+

I-

of

out-

An

PV+

evaluation

GA

CE,

[57] ToU/IBR+

based T BD

combined using WSM [60] RTP+SHAs+ CE

An

algorithm

for

priority

33%/

21%

PV+WT+

based SHAs scheduling; RESs

reduction in

SB+EV+

and SB supply HAs in peaks

CE

I-HEMO

hours;

without

Load supplied from

grid and RESs charge SB in

LP/BPSO

with/

RES

off-peaks; Sharing of grid, PV and SB not included [93] ToU+Fixed+ System operation PV+WT+FC+ MT+SB+ I-HEMO

cost

An optimal dispatch of RDESs

Optimal

HIGA,

based on costs of fuel for MT,

operating

DP,

O&M, GHG emissions and

cost

EA/Ac-

CE from/to main grid for a

RDESs for

smart MG; WT/PV preferred

DP>GA>EA>experi-

sources; SB discharges based

HIGA

for

tual

ment

on SOC, if local sources are not able to meet demand; For insufficient local sources, noncritical loads are disconnected; Habitat isolation GA (HIGA), DP, GA and EA are used

76


Chapter 3


A DRSREOD based HEMO

65% reduc-

MILP/

PV+ESS+EV+

algorithm; Impact of trading

tion in CE

GAMS,

I-HEMO

of energy from RDESs on the

CPLEX

reduction of CE examined;

solver

[61] ToU+SHAs+ CE

The total daily electricity consumption cost viz the difference between energy bought and energy sold to grid minimized; V2H, V2G, G2V options incorporated to minimize peak loads/ further reduce CE [66] ToU+SHAs+ Net PV+EV+SB+present I-HEMO

value

of

HEMO

MILP/GAMS,

ried out after DR-based load

DR-based

CPLEX

leveling; Sizing including DR

HES sizing

solver

and

An algorithm for scheduling

28% reduc-

NP-hard/

peak load

largest load first in each slot;

tion in CE

Minmax

For same loads,

with

algo-

[145] RTP+SHAs+ CE I-

Economical

resulted in economical design

cost

SB+

Optimal sizing for a HES car-

the task

with longest run time is as-

and SB

DR

rithm/

signed first; Generally cumula-

MAT-

tive consumption in a time slot

LAB

taken as critical while considering total cost per timeslot; SB is used to smooth the demand

77


Chapter 3


Operating

An algorithm for the economic

Optimal

NLP/

cost

operation of HES considering

schedules

CCA

SHAs+FC+ HES

predicted electrical/ thermal

of

Software

SB+

loads; Optimal load schedul-

s/sources

simula-

ing decreased system opera-

generated

tion

-

GA/

[146] 3

stage

ToU+

for

I-

HEMO

load-

/

tion cost; Effect of SB efficiency and electricity/gas tariffs on HES operation cost analyzed; Colonial competitive algorithm (CCA) is used to solve the optimization problem [92] ToU+Fixed+ Operating

An algorithm for the opti-

PV+WT+

cost,

mum dispatch of RESs, ESS

Case

FC+ESS+

GHG

and the main grid; Constraints

studies

I-HEMO

emissions

for ESS charge/discharge, gen-

for

win-

and RES

erators start/stop, emissions

ter

and

usage

and supply capacity consid-

Summer

ered; Validated for the varied load for different seasons

78


Chapter 3


[91] DAP+PV+ SB+

I-

HEMO

Welfare

A prosumer based algorithm

7%

to

to maximize sum of benefits to

tion

in

sumer/

users and the utility; For scal-

CE

for

util-

ability and user privacy, prob-

consumer

lem solved using LMs; Fit-

and 136.6%

ness function maximizes wel-

increment

fare including consumption-

in profit for

based satisfaction, monetary

utility

ity

con-

and

privacy

reduc-

CP/Interior point method, LMs

benefits from RES/SB to users and benefits for the reduced peak load, generating cost and GHG emissions to utility [147] RTP+SHAs+ CE

and

An algorithm for the load

4.06% duction

residences

gradient

with

method,

user sat-

management

able+ SB+

isfaction

and the utility; LMs as EP s

CE

and hourly consumption data

3.48%

exchanged

satisfaction

between

utility

QP/ Sub-

in

curtail-

I-HEMO

for

re-

dis-

LMs

and consumer to converge for

optimal

operations

of

shiftable/curtailable HAs and generation; due

to

Dis-satisfaction shifting/curtailed

operation of HAs

79


Chapter 3


LP/

[67] ToU+SHA+ CE,P AR

A Knapsack formulation for

22.63/

and T BD

HEMO model; Demand capac-

22.77%

stat+

ity limits applied during shift-

ductions in

sack,

PV+DG+

ing of HAs to reduce P AR;

CE/P AR

GA

I-HEMO

Fixed HAs dispatched to DG;

at

Preferred intermediate posi-

discomfort

thermo-

re-

Knap-

50%

tion proposed for each SHAs to evaluate T BD [58] ToU+SHA+ CE

and

A

priority

based

resource

15.96% sav-

LP/resource

curtailable

Peak

scheduling; Maximized PV us-

ings in CE;

priorities

load+

demand

age; SB used after PV; SHA

90 units sol-

based

Fixed

operation based on RT pri-

d/day

heuristic

HAs+PV+

ority adjustment; HAs classi-

SB+

fied as controllable and semi-

A PV/DG system optimal siz-

Optimal

NLP/HSA

and GHG

ing;

Loss of power supply

sizing with

emissions

probability used as reliabil-

more inten-

ity index; Enhanced diversifi-

sification

I-

controllable (flexible power)

HEMO [65] Fixed+PV+ NPVC DG+ HEMO

I-

cation and intensification applied in harmony search algorithm (HSA).

80


Chapter 3


CE (from

A realistic HEMO framework

CE reduced

LP/

grid

to reduce overall CE consider-

by

simplex,

ing thermal dynamics of hous-

V2G

(SB)+PV+

es/EVs and life degradations

economical

CPLEX/

I-HEMO

cost of EV’s SB; PV genera-

for old SB;

MAT-

tion, smart charging/discharg-

Avoiding

LAB

ing of EV and space heat-

high

ing control included; Expen-

raises

sive fuel usage for hybrid EV

life

[112] DAP+ thermostat+

EV

and

PV/SB)

8-33%; more

barrier,

SOC SB

reduced [148] ToU+Fixed+ CE SB+

I-

and

privacy

HEMO

A HEMO model studied from

Heuristic

Convex/

privacy-cost tradeoff perspec-

reduced

DP,

tive using SB; Privacy mea-

CE/

vari-

ter filling

sured as the variation of power

ance

by

algorithm

from grid for a given pro-

13.5/52.4%

wa-

file; TOs formulated as WSM; Impact of SB size on tradeoff studied; Larger SB capacity improved privacy; Off-line heuristic shows lesser computing as compared to online DP [149] ToU+EV(SB)+ Maximized A HEMO algorithm is used Fixed+

RES’s us-

Energy

is

to maximize renewable en-

shared from

SHAs+PV+ age

ergy usage; Optimal shifting

RES

I-HEMO

of EVs charging and HAs to-

creased

wards slots with maximum

from

RESs penetration

68.2∼100%

81

MILP/GAMS

in-


Chapter 3


59.2%

ing to benefit the utility

duction

and households; At start of

in

CE

algo-

CoHEMO

the day,

with

DR

rithm/

(CC)

trol agents iteratively compute

and

high

Actual

EP and schedules/CE, re-

capacity SB

neighbors

50% reduc-

NLP/

tion in CE

BB /Ex-

and

SHAs+MBA+utility EV+SB+

welfare

utility and con-

re-

CP/

An algorithm for schedul-

[99] RTP+Fixed+ CE

Gradient

spectively; Forecasted load is sent to the utility which recomputes EP and communicates it to the consumer; Equilibrium is based on the gradient algorithm; Selfish welfare by a home maximizes welfare for others and the utility; Lower price of SB enhances DR benefits [103] ToU+SHAs+ Maximized A HEMS with neural fuzzy inCE sub-

ference based predictor to fore-

HEMO

ject

cast consumer demand based

periment

(CC)

genera-

on the lifestyle, environment

home

tion cost

and social factors; ToU tar-

less than

iff is based on predicted load;

CE

HEMO problem formulated

PV+

Co-

to

for the aggregator to maximize EP as function of CE, elasticity coefficient and consumed power; Utility and consumer are benefited

82


Chapter 3


A mechanism based on two

35-65%

NP-hard

SHAs+PV+ RESs

algorithms for a cooperative

reduction in

LP/

usage

neighborhood; First algorithm

CE

duino

[106] ToU+NCAs+ CE

WT+SB+

and

CoHEMO

schedules high power loads to-

2560

(CC)

wards off-peaks taking user’s

board

Ar-

expected load profile using LP; Second reallocates starting time of loads when surplus RES power is detected based on random broadcast consensus negotiation [120] SHAs+PV+ Cost WT+SB+

dispatch

Diesel+FC+ and CoHEMO

of

SB

efficiency

ad-

DP,

An algorithm for a forward

Full

base camp for optimal EM in

vantage

simplex

a MG with distributed energy

of

method/

resources (DERs); The total

need larger

MAT-

daily cost of the system oper-

reserve

LAB

ation minimized and total SB

capacity

RESs

SB

efficiency during charge/discharge maximized using priority list method; Coordinated smart MG implemented

83


Chapter 3


[64] DAP/ToU+ CE

An

algorithm

for

optimal

Reduced

MILP/

SHAs+PV+

scheduling of HAs integrated

peaks, CE,

GAMS,

EV+SB+

with PV, EV and SB in a

transformer

CPLEX

CoHEMO

neighborhood; Two steps op-

overloading

Solver

timization strategy for trans-

avoided

former capacity usage; Uni-

using

formly distributed capacity al-

HEMO

located;

Co-

CE increased for

homes needing excess capacity [63] RTP+SHAs+ CE Thermostat

usage

+PV+SB+

RES

and

A control algorithm for aggre-

CE

re-

LP/1000

of

gated homes to participate in

duced

by

experi-

DR program for an optimal

9.5/28.6%

mental

scheduling of SHAs and AC

for

homes

temperature control; PV and

DR/DRSREOD-

SB are integrated to supply

based

loads based on the availabil-

HEMO

CoHEMO

Co-

ity of PV, SOC, charge/ discharge rates, EP , etc. [150] ToU+SHAs+ Economy

A scheduling algorithm for a

CE reduced

GA/

49.5%

Case

com-

study

PV+SB+

and com-

households’ MG including PV

by

CoHEMO

fort

and SB that effectively reduces

with

shifted

CE, P AR while maintaining

fort of only

power

the comfort; Economy is for-

0.86

for

mulated by summing reduction in CE in each slot while comfort is based on the sum of power shifts in each slot

84


Chapter 3


An algorithm to find energy

Half hourly

LP/ Lin-

mix for optimal sources dis-

optimal

prog/

patch and load profile for a vir-

sched-

MAT-

hydro+

tual power plant (VPP) in a

ules

Flywheel+

MG; Direct load control exer-

sources/DLC;

CoHEMO

cised where required; UF in-

VPP

cludes direct costs for sources;

for

The levelized CE used for

nation

[151] Dynamic+ DLC+PV+ WT+

Cost

of

dispatch

P-

for

LAB

used coordi-

comparing the energy mix

Table 3.5: Models for DRSREOD-based HEMO (Stochastic) Ref. Tariff+HDs+ Objectives Salient features of HEMO

Performances/Optimization

Type of co-

Achieve-

/Simu-

ordination

ments

lation methods

[51] ToU+SHAs+ CE,

A MPC based scheme for

28% reduc-

MILP/

PV+EV+

thermal

optimal scheduling of HAs

tion in CE

GAMS,

ESS+

discom-

with RES; Stochastic model

at 41.7% in-

ILOG,

I-HEMO

fort,

uses MC simulation to repre-

creased dis-

CPLEX

total/-

sent uncertainty in PV, tem-

comfort

solver

peak load

perature, N-controllable loads;

and GHG

WSM used to combine objec-

emission

tives

85


Chapter 3

[101] ToU/Dyn.+


Minimized An energy consumption/pro-

MILP/

Reduction

SHAs+PV+ daily bill

duction control mechanism for

in

CE

is

ESS+

optimal usage of PV and

10,22,20

I-HEMO

HAs; Prediction for PV based

and

30

on linear regression and load

(%)

with

stochastic models considering

SB,

PV,

periodicity in people habit in

DR

and

using HAs; Difference between

DRSREOD

actual/predicted values used

usage

Case Study

for training [117] ToU/CPP+

CE

and

An optimization of energy

Scheme

Co-

SHAs+

comfort

services of end-users by the

hedge con-

evolutionary

Elastic+

for

ther-

scheduling of available DERs;

sumer from

PSO

thermo-

mal/ time

forecasted solar irradiation,

over

stat+

delay

energy demands, EV avail-

due to in-

PV+EV+

ability, events occurrences; A

accurate

I-HEMO

probabilistic method to gen-

prediction

cost

erate robust schedules when accurate forecasts cannot be made; WSM to combine objectives

86


Chapter 3


[121] RTP+SHAs+ CE

and

An algorithm for curtailmen-

30% reduc-

MILP/

thermo-

HVAC

t/scheduling of HVAC and EV

tion in CE

Heuristic

stat+ PV+

based dis-

minimizing temperature devi-

without in-

optimiza-

I-HEMO

comfort

ation from the desired set

creased dis-

tion/

points; Heuristic model based

comfort

Real

on multi-scale MPC; In fast

measure-

temperatures and thermal dy-

ment

namic parameters are computed (loads are allocated) [152] RTP+CHP+ CE SB+

A robust programming based

I-

-

LP

mechanism to manage uncer-

HEMO

tainty in load scheduling; Battery and CHP used to absorb uncertain parameters A two-stage HEMO mecha-

CF

DAP+

Flow

nism to achieve objectives w/o

creased

MILP,

Aggre-

(CF)

changing user habits; First

2.43%;

DP/

gate load+

for

stage: Forecasted demand/RE

4.75%

SB+PV+

sumer/

generated using ANNs and

certainty

I-HEMO

demand

planned demand curve sent to

for

fore-

uncer-

utility on daily basis; Second

cast

error

tainty for

stage (pre-planning): new de-

10.35%

utility

mand curve minimizing dif-

con-

in-

NLP/

Cash

[153] Fix/ToU/

un-

MATLAB

ference from first stage, sent to utility on hourly bases to reduce demand uncertainty; Shifting of algorithm start time improved CF and uncertainty

87


Chapter 3


An algorithm schedules HAs

31.2-40.9%

LP/

PV+SB+

by considering uncertainties in

reduction in

C++/

I-HEMO

certain operating time and use

CE

Software

[113] DAP+SHAs+ CE

Based

RE; Adaptation variable used to handle stochasticity in energy consumption and HAs operation; Offline schedule of RE are adapted to runtime dynamic scheduling while adjusting RE variation [157] DAP+

CE

An

optimized

operational

17.3%

rein

MILP/ DP/

forecast+

management for the load, SB

duction

PV+SB+

and PV through 2-stage week-

CE

I-HEMO

long stochastic optimization;

2 stage ap-

Simula-

In the first stage, problem

proach

tion

solved for longer horizon using

PV-storage

stochastic MILP for end of

scheduling

using

for

Software

day SOC with minimum computation; Second stage based on detailed shorter horizon/daily efficient solution using DP; Uncertainty in PV and demand are managed through MDP and occupancy

88


Chapter 3


An algorithm for HEMO based

6.93% more

MILP/

thermo-

on SOP considering uncer-

reduction in

2PEM-

stat+ SB+

tainty in fixed load and EP ;

CE

GPSO

I-HEMO

Model solved using two point

[130] RTP+SHAs+ CE

/CPLEX

estimation method (2PEM), gradient PSO (GPSO) based hybrid algorithm Cost

-

NLP/

of

A mechanism for HEMO for

cast+ ther-

PV, WT

optimal operation of HVAC

GA/

mostat+

and SB

with renewable-based HES;

GAMS

PV+WT+SB+

Historical wind speed, solar ir-

and

I-HEMO

radiance and load data used

MAT-

for stochastic modeling; Fuzzy

LAB

[154] RTP+ fore-

C-Means clustering used to group data for seasonal variations; GA with two-point estimation used to minimize system cost.

89


Chapter 3


[155] RTP/FIT+

CE

An

MPC

based

optimal

12%

re-

SHAs+PV+

scheduling method for PV

duction

SB+

based multi sources and load;

CE

MPC

disturbing

I-

HEMO

manages

scheduling

errors due to uncertainties

comfort

in predicting EP s, weather

level

in w/o

MILP

/

CPLEX solver, GAMS

condition and activity level; Parameters

updated

in

a

receding horizon at each iteration to give more accurate decision to schedule SHAs, SB, PV and energy trade A real-time mechanism to pur-

CE reduced

NLP/

chase energy during off-peak

by

20%

SOC

hours, SB/PV utilization and

using

load

based

PV+SB+

temperature management dur-

manage-

heuristic/

I-HEMO

ing peak hours; In peaks HAs

ment for 26

MAT-

are supplied from SB; For

HAs

LAB

[156] RTP+SHAs+ CE (from thermo-

grid

and

stat+

PV/SB)

lesser SOC, temperature setting increased; If HAs power is still more than set value, then check tariff for average value to connect HAs; Otherwise, warn user from purchase of energy/stop HAs

90


Chapter 3


[107] ToU/CPP+

CE

and

An

optimization

method

Robustness

SHAs+PV+ comfort

called moving window error

against

EV+

for shift-

correcting algorithm (MWA)

forecast

ing/cur-

to improve energy consump-

errors

tailing of

tion/production schedules for

energy

homes with periodic forecasts

I-

HEMO

MILP

updates through MPC [158] RTP+SHAs+ CE SB+

I-

and

privacy

HEMO

A rolling online SOP learn-

Better pri-

MILP/

ing from the past and antici-

vacy

stochastic

pates for future at each slot for

well as re-

opti-

computing robust and optimal

duced bills

mization,

scheduling; RT EP uncertain-

achieved

CPLEX

ties dealt using MC method;

with larger

solver

Use of SBs enabled in dis-

SB capacity

as

guising actual appliance power profile during scheduling to increase privacy; TOs for privacy, CE and SB capacity are analyzed [159] (-

Maximized An algorithm for dynamic

Power con-

LP/ Real home

)+forecast+ usage effi-

updated HA priority based

sumption

PV+SB+

ciency of

scheduling for SB using cloud

reduced by

I-HEMO

RESs

computing; HAs are priori-

7.3%

tized, e.g., for AC, priority is updated according to time, environment and SOC before scheduling; Efficient use of PV based on optimal charge/discharge of SB

91


Chapter 3


Load management using max-

Faster

Event

imum loading of PV/wind en-

turn

based+

ergy to reduce CE; Resource

investment

BPSO

PV+WT+

surplus calculated; DP tech-

for RESs

Virtual

I-HEMO

nique failed to solve knapsack

Smart

problem because of complexity

Home

[160] RTP+

CE

reof

2 dimensional

for multi-dimensionality; Cognition and adaptiveness factors for PSO improved its performance CE,

Households are considered as

20%/

50%

MILP/

SHAs+

energy

energy hubs each with energy

reduction in

GLPK

thermo-

usage,

production, storage and con-

CE/peak

Solver

stat+

GHG

sumption; Novel models for

load

PV+SB+

emissions

major HAs, PV and SB pre-

I-HEMO

and peak

sented; An algorithm for sys-

demand

tem operation based on users

[90] ToU/RTP+

preferences,

occupancy and

day ahead emissions and climate forecasts presented

92


/

Chapter 3


CE (cash

An optimal scheduling of PV-

PV-based

MILP/

flow)

based multi sources, load and

ancillary

BB,

curtailed+

power exchanges with the grid

services

Simplex

PV+SB+

to maximize profit using BB;

increase

(CPLEX

I-HEMO

To deal with uncertainty for

consumer

MI

variations in consumption due

profit by 10

Solver)

to events occurring and PV,

times

[161] ToU/FIT+ SHAs+

re-scheduling is done using the simplex method; Penalty factor deals uncertainty in transmitted power Energy

An MPC based method for ef-

1.35% / 2%

thermo-

saving

ficient operation of HAs for

reduction in

stat+ PV+

and

room luminance and temper-

consump-

I-HEMO

comfort

ature management using pre-

tion

dicted solar radiations and ex-

respective

ternal temperature; Controller

cases

[162] ToU+

QP

for

takes care about prediction errors due to actual input parameter that might result in reduced room lux and heat pump usage;

Reduction in

consumption for comfort/ saving TOs proposed

93


Chapter 3

[163] RTP+


An

forecast+

GHG

method; ANN forecast model

duction

PV+SB+

emission

is used for proximal load pre-

CE

diction;

emission

I-HEMO

MPC

based

HEMO

35%

Cost and

Weather condition,

re-

NLP/ DP

in

/ Simula-

and

tions

solar flux, EP and emission data obtained through internet; Proposed MPC attains 98% optimality [164] RTP+ pre-

Profit

An MPC based control for the

20.3% more

MILP

dicted+

with

state of EV battery to op-

profit while

/PSO

PV+EV+

selling

timize CE for the user and

adopting

I-HEMO

energy

profit for the utility; MPC

DRSREOD

used for EV battery; Load estimated using the regressive model on 19 days data of a home An MPC based model for

11.5%

Cost

MG having ESS with high

3.1% reduc-

horizon

curtailed+

uncertainty of load, PV and

tion in cost

method/

PV+ESS+

EP ;

in

Simula-

I-HEMO

charge/discharge

Elastic+

Economic

dispatch, of

ESS,

DAP

/

LP/Rolling

Operating

[165] RTP+

MPC/

tions

external grid and load curtailment

considered;

with EP

MPC

and DAP based

optimization compared

94


Chapter 3


An MPC based HEMO model

32-42%

MILP/

EV+SB+CHP+

considering uncertainty in the

reduction in

CPLEX

I-HEMO

predicted wind, PV, load and

CE

solver/

[166] RTP+SHAs+ CE

EP ; Three case studies inves-

MAT-

tigated: smart loads and tra-

LAB,

ditional DGs; EV with V2H

YALMIP

capability; and availability of ESS; MPC for EV resulted in > 32% and for ESS resulted in additional > 10% cost reduction as compared with base scheme [167] RTP+SHA+ CE

for

An MPC based approach ap-

58.8%

thermostat

elec-

plied to homes for optimal

duction

+PV+SB+

trical/

HAs scheduling, heating de-

CE

re-

MILP/

in

MATLAB,

HTP+CHP+ thermal

vices and local generations

CPLEX

demand

taking care of user preferences,

as

and

RESs generation, demand and

based

comfort

EP ; At each time step, pre-

solver

I-HEMO

vious scheduling adjusted to follow fresh local generation, heat demand and EP

95


BB

Chapter 3


[168] RTP+SHA+ Maximized A mechanism for HEMO for PV+WT+

usage

of

standalone

-

NLP/ SMCM

renewable-based

Diesel+SB+ RESs

HES; WT and PV used as

I-HEMO

main sources supported with DG and SB; To optimize PV and extract maximum available power, fuzzy logic is used for temperature, irradiation and the generated voltage at MPP; HAs shifting is based on sliding mode control method (SMCM)

[169] RTP+SHAs+ CE

A

mechanism

for

optimal

GPSO-

MILP/GPSO/

thermostat

operation of HAs consider-

2PEM

MAT-

+PV+SB+

ing uncertainties due to fore-

outperform

LAB,

I-HEMO

casted errors of EP and fixed

GPSO-

CPLEX

load;

CCO based formula-

LHM in 8

tion to accommodate uncer-

times lesser

tainties; GPSO used for opti-

comput-

mum solution; two-point esti-

ing

mate method (2PEM)/ CCO

and in CE

using output probability den-

with

sity function (OPDF) is used

32.8/35.2%

time

ratio

to handle uncertainty viz compared with Latin hypercube method (LHM)

96


Chapter 3


A method for load scheduling

Cost

SHAs+PV+

based on cost efficiency (ra-

ficiency

Newton

SB+

tio of consumption benefit to

increased

method/

electricity cost) concept intro-

by 11.60%,

MAT-

duced to improve economic ef-

further im-

LAB

ficiency of consumption; Im-

proved

pact of RDESs included; To

RDES

I-

HEMO

ef-

CP/

[170] RTP/DAP+ CE

by

cater for consumption uncertainty, penalty factor included in formulation A mechanism to minimize

13-20%

SP/

Elastic+

CE for load serving entity

reduction in

LMs/ Ex-

PV+WT+

(LSE) that supplies power

CE

periment

SB+

to neighborhood having re-

Total CE

[48] ToU+

Co-

HEMO

sources and load; LSE coordi-

(DC)

nates for optimal use of energy at homes; RESs, load and CE are stochastic processes based on Lyapunov based cost minimization and used to jointly consider energy and demand management decisions; TOs between CE and SB computed

97


Chapter 3

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO

A MPC based HEMO mech-

25.4%

HEMO+

anism with two-tiered ap-

duction

CoHEMO

proach; First, implement a

peak load

[171] RTP+

I-

Peak load

re-

MILP/

in

linprog for MPC,

centralized MPC to reduce

GAMS

peak of net demand based on

/CPLEX

thermostat set points; Second, optimally schedule HAs with maximum power to further reduce peak The HEMO schemes for the

Framework

MILP/

PV+WT+

individual home,

for

DP / 39

CHP+

VPP; Three-step methodol-

HEMO and

houses

I-HEMO+

ogy is proposed based on

CoHEMO

neighbor-

CoHEMO

global planning, local predic-

concepts

hood

tion and control with DERs;

for

Micro-CHP scheduling based

presented

[172] RTP+SHA+ CE

MG and

I-

MG

on predicted heat demand; Prediction errors managed using MPC by reducing sum of squared mismatch between planned and produced power.

3.3

Salient Issues and Challenges in HEMO

The issues related to standardization requirement for HAs: formalization to handle diversification; and coordination complexities need attention for their resolution.

98


Chapter 3


Challenges regarding handling of multi-objectivity; management of data uncertainty; development of optimal computational methods; and handling of problem complexities [122] still need to be addressed. The issues and challenges in the field of HEMO are analyzed in sub-sections 3.3.1 to 3.3.7. The aforementioned analysis is mainly based on the pros and cons of the modeling approaches, reflected through the performance parameters of the related models.

3.3.1

Standardization Requirements for Response Classes

Various DR responsive classes with their names used in the present literature are presented in Table 3.6. It is found that the same appliance has been modeled under more than one response class that indicates flexible nature of HAs for HEMO modeling, e.g., refrigerators have been considered as non-controllable in [138], as un-interruptible in [56], as interruptible in [144] and as regulating loads in [173]. On the other hand, conflicts in definitions of various response classes are revealed. For example, name deferrable exists in both shiftable as well as in elastic response classes. Non-deferrable HAs are modeled like non-controllable and deferrable HAs like the non-interruptible ones in [59]. Deferrable HA is functioning like an interruptible class in [152]. Further, response class named non-controllable exists with more than 11 names in the recent research. The issue seeks the attention of relevant international bodies to include standardized nomenclature and definition for each of the response class in the relevant standard.

3.3.2

Formalizing Requirements to Handle Diversification of Control Parameters and Optimization Approaches

The recent literature on HEMO is highly diversified with respect to frameworks for modeling, control parameters and optimization techniques. Each model has strength in one and weakness in some other perspective. Hence the performance

99


Chapter 3

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO Table 3.6: Response Classifications of HDs

Response class Non- controllable appliance

Examples

References

Frequency

Real, must run, nondeferrable, fixed, sensitive, non-responsive, delay sensitive, essential, critical, non-changeable, nonshiftable Curtailable Controllable, nonappliances sensitive, adjustable Shift able, interruptSHAs ible, controllable, (Intermulti-phase, deruptible) ferrable (interruptible)

TV, fans, lights, computer, refrigerator, etc.

47

SHAs (Noninterruptible)

Non-interruptible, deferrable, shift-able, plan-able, deferrable (non-interruptible), burst load

Washing machine, dish washer, dryer, refrigerator, etc.

Elastic appliances

Deferrable, multi-duty cycle with time shifting, power changeable Thermostatic control, thermal dynamic model, regulatable

HVAC, pool pump, EV, etc.

[14, 110, 59, 114, 136, 107, 151, 64, 159, 106, 157, 99, 100, 103, 54, 146, 63, 51, 150, 101, 179, 92, 143, 134, 50, 108, 117, 118, 144, 139, 140, 56, 147, 115, 165, 166, 137, 67, 58, 65, 142, 156, 132, 149] [107, 151, 93, 160, 150, 179, 147, 58] [57, 14, 111, 110, 59, 136, 107, 133, 61, 66, 55, 99, 105, 100, 104, 54, 51, 150, 121, 130, 64, 50, 108, 109, 117, 118, 144, 139, 56, 147, 161, 166, 137, 53, 67, 58, 156, 169, 170, 132, 149] [14, 111, 110, 59, 113, 136, 107, 133, 106, 105, 100, 103, 54, 63, 51, 101, 150, 121, 179, 143, 130, 50, 108, 109, 144, 139, 56, 138, 115, 166, 137, 53, 67, 58, 156, 169, 170, 132] [57, 48, 114, 60, 99, 54, 64, 117, 115, 164, 171, 112, 132, 149] [159, 145, 99, 100, 63, 51, 121, 143, 134, 109, 117, 144, 138, 141, 171, 67, 156, 112, 169, 132] [57, 48, 59, 113, 120, 60, 62, 102, 151, 93, 154, 106, 157, 61, 66, 55, 99, 100, 103, 54, 146, 63, 51, 150, 101, 153, 156, 121, 92, 130, 91, 64, 56, 144, 117, 118, 147, 155, 67, 58, 65, 112, 159, 160, 90, 161, 162, 163, 164, 165, 166, 167, 168, 172, 152, 169, 170, 149] [48, 59, 113, 120, 168, 136, 60, 62, 102, 151, 93, 159, 154, 106, 157, 61, 66, 145, 99, 100, 146, 63, 51, 150, 101, 92, 130, 64, 90, 144, 147, 129, 161, 158, 152, 155, 163, 165, 166, 167, 58, 156, 153, 112, 148, 169, 170, 149] [120, 168, 93, 146, 51, 166, 67, 65]

20

MBA

Name used in research

Renewable, Noncontrollable energy sources

Energy storage

HVAC, pool pump, etc. HVAC, pool pump,EV, rice cooker, ironing, refrigerator

HVAC

alternate

PV, WT, FC, etc.

Electrical storage systems

SB, fuel cell, electrolyzer, pump hydro, fly wheel, EVs, etc.

Dispatchable Controllable, sources renewable

Non-

Diesel generator

100


8 41

38

14

58

48

8

Chapter 3


of various HEMO models cannot be compared easily on the same bases. For example, reduction in CE is the main performance parameter in majority of HEMO models. Different studies on HEMO claim different values for the reduction in CE based on the number of SHAs and scheduling resolution. Such reductions have a strong tradeoff with computational cost due to changing number of variables [105]. Reduction in CE also depends on the type of tariff, proposed starting/ending times of appliances, length of HAs operation, user priorities [78], starting time of scheduling horizon [153], the inclusion of NCAs, baseline cost, etc. Table 3.7: Reduction in CE and P AR for DR-based HEMO

References [59] [101] [14] [110] [115] [114] [157] [109] [48] [107] [133] [54] [137] [50] [139] [159] [171] Average(%)

Tariff

Reduction in CE (%) ToU 25 ToU 30 RTP/IBR 15.3 RTP+IBR 1.3 ToU 7 DAP 28.6 DAP 17.3 RTP 20 ToU 20 RTP 13 ToU 22.2 ToU 22 ToU 11.94 DAP 3.7 RTP+IBR 15.7 7.3 RTP 16.27

Reduction P AR (%) 19.7 25 12 19.74 25.5 25.8 21.29

in

As an example, HEMSs incorporating NCAs and the similar non-incorporating NCAs are compared for an achieved reduction in CE. Refer to an average reduction in CE given in Table 3.7 and Table 3.8, it is observed that huge amount of reduction in CE has been achieved by HEMO models; not including NCAs as compared to the HEMO models incorporating NCAs. These values for respective configurations of HEMS are found to be 36.53% and 16.27%. As HEMO models 101


Chapter 3 Table 3.8:

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO Reduction in CE and P AR for DR-based HEMO (Notincorporating NCAs)

ReferencesTariff [78] DAP/ToU [105] DAP [111] DAP [108] RTP [179] ToU [133] ToU [143] DAP [141] RTP [60] RTP [171] RTP Average(%)

Reduction in CE (%) 39 47 40 27 43.7 22.2 56.9 20 33 36.53

Reduction in P AR (%) 19.74 25.8 22.7

in Table 3.7 include control parameter for NCAs load while the models in Table 3.8 do not include the same parameter, performance of these DR-based HEMO models cannot be compared for an overall reduction in CE. It demonstrates that performance parameters like CE are highly sensitive to the varying values of the aforementioned control parameters, e.g., NCAs load. On the other hand, focusing on optimization techniques, LP models provide global minimized solution resulting in more reduction in CE as compared to heuristics methods that provide approximate optimal solutions. Computational cost; however, is a tradeoff that makes LP techniques even impracticable due to high computational cost for large-scale problems [102]. Based on the least computing cost, heuristics can easily outperform conventional methods. Whereas, the authors in a number of studies have not provided the computational cost that restricts a fair evaluation of the models. A vast diversity in HEMO calls for serious efforts to formalize the frameworks for HEMO modeling. Standard testbeds based on such frameworks may be developed to analyze and to compare the performance of various HEMO models on the same bases. Such formalization may facilitate in improving the quality of research on HEMO. Various models for DR-based HEMO including HAs, dynamic tariffs, objectives, 102


Chapter 3


features/ approaches for HEMO and optimization methods are summarized in Table 3.7 and Table 3.9. For a comparison between DR and DRSREOD-based approaches for HEMO through the reductions in CE and (P AR); achieved by eminent models on HEMO are presented in Table 3.7 and Table 3.9. DRSREODbased HEMS outperforms DR-based HEMS for a 1.85 and 1.4 times larger values of average reductions in CE and P AR.

3.3.3

Coordinated Approaches and Configurations for HEMO

I-HEMO-based systems are locally defined, independent and are fast in convergence; however, those systems can create peak rebounds. CoHEMO-based systems achieve HEMO objectives considering joint resource management (RM) and can avoid peak rebounds using coordination. Coordination in HEMO reduces CE and system peaks with limited impact on the consumer comfort. In addition, it ensures efficient adaptation of overall demand to RESs and optimizes the joint performance of coordinated consumers by exploiting their inherent difference of demands. Table 3.10 furnishes the reductions in CE achieved by various researchers using I-HEMO and CoHEMO based approaches. Although researchers have used diverse methods to solve their problems like an average reduction in CE can be used to evaluate its performance. The 2.2 times more reduction in CE is achieved through CoHEMO as compared to that achieved through I-HEMO which validates the superiority of CoHEMO approach over I-HEMO. Coordination in CoHEMO is classified into CC where a CO and and DC. In CC, a CO manages electricity usage of all homes; whereas in DC, the end users schedule their HDs directly without any omniscient central entity after communicating with each other or contacting a central entity to gather others’ profile information [39]. CC based approach results in a more efficient RM for the consumers. Disadvantages of this approach include scalability and privacy issues, and scheduling conflicts raised among consumers due to RM by minimizing a shared 103


Chapter 3


Table 3.9: Reduction in CE and P AR for DRSREOD-based HEMO

References

Tariff

[99] [59] [51] [103] [106] [121] [55] [129] [56] [152] [113] [155] [156] [54] [60] [61] [145] [63] [150] [67] [58] [112] [148] [90] [163] [164] [165] [171] [166] [167] [170] Average(%)

RTP ToU ToU ToU ToU RTP RTP RTP RTP DAP RTP DAP ToU+IBR RTP ToU RTP RTP ToU ToU ToU DAP ToU ToU RTP RTP RTP RTP RTP RTP RTP

Reduction in CE (%) 59.2 28 28 50 50 30 22 30.5 11 15.86 42.45 12 20 22 21 65 28 28.6 50 22.63 15.96 33 19.6 20 35 20.3 8.6 32 58.8 11.6 30.18

Reduction P AR (%) 10.5 33 50 25.8 29.83

in

utility function. The drawbacks can be avoided using enhanced meta-heuristics, parallel processing, coding methods and incorporating GT approach [21]. DC based approach provides more independence of choices to the consumers; however, the aggregated cost is usually higher. Furthermore, the bandwidth required for communication and large convergence time are big drawbacks of this approach. 104


Chapter 3

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO Table 3.10: Reduction in CE for I-HEMO versus COHEMO

Type of coordination for HEMO I-HEMO

Range of reduction in CE (%) 0 - 10

References

Frequency

[102, 56, 110, 115, 130, 50, 91, 147, 159, 162] I-HEMO 10 - 20 [14, 157, 155, 139, 58, 148, 165] I-HEMO 20 - 30 [59, 51, 116, 156, 55, 109, 54, 145, 67, 112, 90] I-HEMO 30 - 40 [129, 101, 121, 113, 163, 166] I-HEMO 40 - 50 [130] I-HEMO 50 - 60 [132] CoHEMO 0 - 10 [136, 63] CoHEMO 10 - 20 [107, 137] CoHEMO 20 - 30 [114, 48, 133, 63] CoHEMO 30 - 40 [134] CoHEMO 40 - 50 [103, 106, 150] CoHEMO 50 - 60 [99] Average reduction in CE (I-HEMO) = 19.38% Average reduction in CE (CoHEMO) = 42.88%

10 7 11 6 1 1 2 2 4 1 3 1

Research on MASs and GT for their application to DC is still in progress. MAS has advantages of decentralization, intelligence and single template agents. While drawbacks of such systems comprise separate modeling of each agent and cost resulting from a larger number of communicating agents in real-time. GT, being a strategic decision-making process is suitable for DC and can easily include new players. Developments of frameworks for more efficient coordination in HEMO is hot area in recent research.

3.3.4

Handling Multi-objectivity

Based on the number of objectives, HEMO problems are classified as SOO versus MOO. Most of the researchers have dealt HEMO problem as SOO for a unique solution, while in real life such problems are MOO with TOs among objectives. CE is the main objective that has been included in HEMO model by most of the researchers. Objectives incorporated by various researchers in modeling HEMO problem are summarized in Table 3.11.

105


Chapter 3

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO Table 3.11: Objectives Focused in HEMO

Objectives Cost

Comfort Demand profiling (DPr) (Peak, P AR, consumption) Optimum resource utilizations (ORU) (for RDES, SB) Cost and comfort (convenience, satisfaction social welfare) Cost and DPr Cost and ORU Comfort and DPr Cost and privacy Cost, comfort and DPr Cost, privacy and GHG emission Cost and GHG emission Cost, GHG emission and ORU Cost, DPr and emission Cost, comfort, DPr, GHG emission

References (DRbased) [111, 114, 107, 160, 54]

References (DRSREOD-based)

Frequency 35

[171, 132]

[48, 59, 113, 136, 78, 60, 102, 93, 154, 157, 61, 66, 105, 103, 146, 101, 130, 64, 129, 141, 172, 161, 152, 155, 164, 165, 166, 156, 169] [121] [151, 162]

-

[168, 159, 149]

3

[131, 134, 108, 109, 139, 140, 135, 141, 116] [133, 104, 179, 115] [111] [14, 110, 143] -

[55, 99, 100, 150, 117, 118, 144, 56, 147, 167, 170]

20

[145, 63, 153]

7

[120, 160, 106] [158, 148] [91]

3 1 2 3 1

[50] -

[163] [62, 92]

2 2

-

[90] [51]

1 1

The CE exhibits TOs with other objectives, the most important one is with the discomfort that has been worked on by the maximum number of researchers. Suitable handling of multi-objectivity has been a challenge in solving HEMO problems. In recent research, the following methods have been used to handle multiobjectivity for HEMO problems:

3.3.4.1

WSM

In this method, weighted sum of all the objectives is taken to define a new OBJF, by providing that the constituent functions are linearly scalable. The method has 106


1 4

Chapter 3


been used in [14] to get weighing TOs between the cost and the comfort. It has been the simplest method to incorporate TOs in MOO problems and been used in most of the research on HEMO.

3.3.4.2

Bounded Objective Method (BOM)

In this method, all of the objectives are changed into constraints while imposing acceptable lower and upper limits for them except one most important objective, i.e., a MOO problem is transformed into an SOO problem. For example, the objective for minimizing CE may be transformed into a constraint by including the range of the consumer budget for a specified period. Similarly, the objective for thermal comfort can also be transformed into a constraint by introducing deviation limits from an acceptable value of temperature [121].

3.3.4.3

Physical Programming (PP)

The method needs an in-depth knowledge of the system. Suitable functions are used to model TOs among various objectives, e.g., an occupant may know the value of discomfort as a function of temperature. It is similar to weighted sum if constituent functions are linearly related to the others.

3.3.4.4

Pareto Optimization (PO)

MOO problems have conventionally been solved by transforming them into an SOO problem which is of least use for HEMC designer. They need all possible optimal solutions for all conflicting objectives simultaneously, known as tradeoff analysis. PO-based methods provide relationships among solutions based on Pareto-dominance concept. The method computes sets of non-dominated solutions called POS. The corresponding values of objectives are termed as PFs that give TOs among objectives F1 (X), F2 (X),..., Fk (X). Pareto based TOs are used for an optimal HEMC design having diversification in decision making. Such a

107


Chapter 3


design enables consumer choosing the best available option in accordance with his needs. [123].

3.3.4.5

Utility Function (UF)

UFs Ui (f (i)) are defined for a set of OBJFs based on the importance of individual OBJFs; where i is the number of OBJFs. An overall UF is obtained by summing up all of the individual UFs. In [102], an overall UF is designed to minimize the difference in power supplied from RESs and the load. Table 3.12: Methods to Handle Multi-objectivity in HEMO Techniques WSM BO PP PO UF GP

References [14, 110, 160, 122, 154, 61, 55, 51, 150, 101, 134, 50, 117, 118, 90, 144, 139, 163, 137, 148, 116] [105, 121, 92, 143, 64, 108, 109, 141, 167, 112] [120, 62, 133, 106, 99, 100, 91, 140, 56, 138, 147] [160, 131, 129, 135] [102, 151, 132] [115]

Frequency 21 10 11 4 3 1

Methods used to handle multi-objectivity in last six years’ research are summarized in Table 3.12. Most of researchers have opted for WSM that solved HEMO problem as SOO without providing clear TOs among objectives. From recent few years, a trend towards PO-based methods has been observed that provides clear and optimal TOs among the objectives to offer diverse choices to the consumer.

3.3.5

Modeling Uncertainty of Data

Based on the uncertainty of data, HEMO problem has been classified as deterministic versus stochastic. Deterministic models assume all uncertainty resolved ex-ante; where all parameters are assigned forecasted values prior to the model run. RESs, environmental parameters, must run appliances, EV availability, etc., all include uncertainties and result in forecasting errors. For example, solar energy from a PV panel varies with changing angle of the sun, cloud’s shadow, etc. 108


Chapter 3


While deterministic model generally subtracts forecasted RES energy from load to calculate the demand to be supplied by the utility. In case of any reduction in RE due to forecast error, demand from utility will exceed from the designated power that results in inefficient scheduling or even system tripping for exceeding power capacity limits. Various methods have been used to address uncertainty in scheduling process in order to improve scheduling inefficiencies. Some methods like ”heuristic-based real-time scheduling” manage uncertainty without considering forecast, e.g., in [59], the old and present data have been used for the real-time heuristic-based solution. While others use day ahead scheduling with real-time adjustment based on point-forecast to address uncertainty, e.g., authors in [113] used this scheme in order to manage uncertainty in consumption patterns of HAs. Major stochastic HEMO models used to manage uncertainty in parameters are as follows:

3.3.5.1

SOP

In this method, schedules are corrected after receiving fresh information through recourse variables for each recourse stage. The resulting multi-stage SOP suffers from a curse of dimensionality due to exponentially increasing number of variables. This limits SOP application up to two stages only [24]. In [157], a two-stage week-long stochastic optimization method for a PV-ESS system is presented. At first stage, problem is simplified and solved for one week longer horizon using stochastic MILP to find an end of the day SOC constraint. At second stage, a detailed shorter horizon daily scheduling solution based on DP is introduced for improved efficiency and minimum computation time. However, because of larger computational burden, the method seems not preferable for HEMO problems.

3.3.5.2

RO

In this method, a set of bounds and a distribution for the uncertain parameter are included in order to enable the mechanism minimizing the impact of the worst-case 109


Chapter 3


scenario. For a minimization problem, it minimizes upper-bound of the objective. The model may not provide the most optimal scheduling; however, it is appropriate to challenge adverse consumers. In [108], an RO based algorithm computes optimal schedules considering RTP uncertainties [109].

3.3.5.3

CCO

The method is used to solve the problems that contains uncertainties and shows good characteristics to satisfy constraints with feasible probability level. In contrary to RO, the method is unbounded. In [130], a CCO based method for HEMO problem considering uncertainties of NCA type loads and CE is presented.

3.3.5.4

SDP

This method is based on optimization through a finite state-space model. To model future uncertainty, present state and an action leads to a distribution over possible states in future have been taken. The solution is obtained from the final state and from finding the optimal track back to initial. In [111], an algorithm is used to find a series of price threshold that reduces CE through optimal scheduling. Computing cost is low for a moderate horizon and 1-hour resolution. Method becomes burdensome at 15 minutes resolution [111].

3.3.5.5

Stochastic Fuzzy Optimization

The method is based on fuzzy logic. Truth values in fuzzy logic are taken between 0∼1. Non-crisp values are applied to forecasts the future parameters. Optimal schedules of home devices is computed in smaller time using confidence level in fuzzy parameters. In [154], historical data based on fuzzy C-means has been used to stochastically model RESs generation and the load.

110


Chapter 3 3.3.5.6

3.3. SALIENT ISSUES AND CHALLENGES IN HEMO MPC

This method is based on a control algorithm that repeatedly revises scheduling process over remaining part of the time horizon based on new forecasts. MPC has the capability to receive proximal data. As errors for proximal forecasts are smaller as compared to distal forecasts, they schedules for first few steps in scheduling horizon. In [51], an MPC based algorithm is developed to represent uncertainties in RES generation, outdoor temperature, water usage and NCAs operation. In [155], activity level, weather condition and the RTP are main updated inputs for the MPC based HEMO model. For rescheduling, algorithm runs for a horizon of 96 intervals. First interval decisions are applied while rest of the output decisions are rejected. Time horizon is then shifted for one interval forward after next iteration. The optimal solution at each iteration is based on the current and forecasted information of uncertain parameters. MPC uses local optima at each stage that may lead to myopic decisions. Scheduling horizon is to be sufficiently long to avoid myopic optimization. Table 3.13: Methods to Handle Stochasticity in Solving HEMO Problem Stochastic method References Real-time scheduling (without considering [59, 173, 156] forecasts) Models incorporating uncertainty using point forecast DA scheduling with real-time adjustment [113, 141, 153, 170] Models considering uncertainty within models SO [114, 157, 108, 158, 174] SDP [111, 157] RO [168, 108, 117, 139, 152, 175] Fuzzy programming [154, 116] CCO [130, 140, 176, 169] MPC [110, 48, 113, 51, 121, 109, 118, 90, 172, 161, 158, 174, 152, 155, 162, 163, 164, 165, 171, 166, 167]

Frequency 3

4 5 2 6 2 4 21

Various methods found in present research to incorporate uncertainty in HEMO models are summarized in Table 3.13. Most of the stochastic optimization methods 111


Chapter 3


address uncertainty issues within the model formulation that introduces curse of dimensionality. The number of decision variables grows exponentially due to the number of uncertain parameters and number of time stages that increases the problem complexity. Further, these methods need large historical data to solve the problem which is hardly available in most of the cases. MPC has been emerged as an attractive option to incorporate uncertainty in HEMO modeling for high scheduling efficiency. Such models are characterized with low computational burden without a need of large historical data.

3.3.6

Optimal Computational Techniques for HEMO Models

For small-scale linear HEMO problems, conventional methods for LP like simplex [120] and interior point methods [110] have been used. For nonlinear, quadratic and convex problems; including quadratic programming (QP) [62], convex programming (CP) and LMs [48]; iterative searches including secant search [177] and grid search [178]; and gradient methods [99] have been used for global optimal solutions. MILP methods based on integer-continuous variables, e.g., BB and cutting plane (CTP) have also been used to solve small-scale linear HEMO problems [179]. For large scale problems, conventional methods including LP and CP become computationally impracticable and problems appear as NP-hard [122]. Further, very little work has been done on integer programming (IP) techniques (like BB and CTP algorithms) for their application to problems based on MINLP. DP is suitable for multistage decision problems which may solve a number of complex problems; however, suffers from the drawback of the curse of dimensionality. For large scale complex HEMO problems, advanced meta-heuristics have emerged as a unanimous choice to obtain the nearest optimum solution in the minimum possible

112


Chapter 3


time. Various computational techniques used to solve HEMO problems in recent literature are summarized in Table 3.14. Table 3.14: Computational Techniques to Solve HEMO Problems Conventional techniques LP Interior point method Simplex method Linprog (MATLAB) QP CP GP NLP Gradient method Secant method Grid search method MILP BB and (CPLEX) DP

Genetic (GA) PSO

CTP

algorithm

Ant colony optimization TS ES CCA Harmony search algorithm

References (DR-based)

References (DRSREOD-based)

Frequency

[104, 138, 141] [110] [62, 109, 140] [115] [77] [177] [178] [139]

[113, 106, 63, 56, 165, 152] [100, 91] [120, 161, 112] [151] [162, 132] [48, 99, 91, 147, 140, 148, 170] [59, 168, 103, 163] [99, 147] [101, 121, 118, 172, 157, 121, 137, 153, 169, 149] [61, 54, 51, 64, 90, 161, 158, 155, 171, 166, 167] [157, 172, 163, 153, 148]

9 3 3 1 5 7 1 5 2 1 1 11

[105, 179, 108]

14

[111, 108, 134, 93, 159, 129] Meta-heuristics techniques [14, 107, 116] [57, 136, 93, 55, 150, 92, 154, 67, 169, 132] [133] [57, 60, 62, 102, 130, 117, 164, 160, 53] [57]

11

[131] [135, 129, 93] [146] -

1 3 1 1

[65]

Heuristics techniques Action dependent [50, 129] heuristic Greedy Search [132] [145] MDP [157] Other heuristics [58, 156, 148] Total papers focusing on conventional techniques = 73 Total papers focusing on advanced heuristics = 38 Nos.

Recently advanced heuristic techniques have been combined within themselves and with other type of techniques to form diversified and better performing tools. 113


13 10 1

2 2 1 3

Chapter 3


Such techniques which combine advantages of compounded techniques are called hybrid computational techniques. Further, heuristic techniques are also combined with expert systems to form hybrid computational intelligence tools such as neurofuzzy, neuro-swarm, fuzzy-GA, fuzzy-PSO, etc. Such tools are suitable for handling dynamics, predictivity and stochasticity. Research on their applications in HEMO, techno-economic analysis, economic dispatch, energy planning and parallel computing is getting pace from past few years. Meta-heuristic-based hybrid computational techniques [180]-[193] found in the recent research are summarized in Table 3.15. Table 3.15: Hybrid Computational Techniques to Solve HEMO Problems Fuzzy Fuzzy ANN [128] ES [128] GA [116] PSO [128] ACO [128] DE TS DP [128]

ANN [128] [128] [128] [180] [181] [182] [153]

ES [128] [128] [128] [128] [183] [128]

GA [116] [128] [128] [184] [185] [186] [187] [128]

PSO [128] [180] [184] [188] [189] [190]

ACO [128] [181] [183] [185] [191] [128]

DE [182] [186] [188] [191] [192] -

TS [187] [189] [192] [193]

DP [128] [153] [128] [128] [190] [128] [193] -

Use of an optimal computational technique is of vital importance in solving HEMO problem. An active research on mixed integer programming, meta-heuristics, hybrid tools and their application in solving non-linear HEMO problems is still continued. Hybrid techniques shall be reviewed in detail in our future work for their application to HEMO problems.

3.3.7

Computational Burden and Problem Complexity

A higher resolution is more suitable to address the problem of insufficiency of data resolution and to achieve optimal reduction in CE, e.g. a slot length of 10 minutes results in 10% more saving in CE as compared to the saving with a slot length of 60 minutes. Yet, high resolution increases the number of time 114


Chapter 3


steps that results in larger number of variables and implicitly it gets an increased computational burden. An increase in the slot length from 10 to 20 minutes results in a loss of just 2∼3% of the saving in CE but corresponds to a decrease in problem complexity by 70%. So generally slot length of less than 10 minutes is avoided for lesser problem complexity. Further, a larger number of schedulable appliances are required for more reduction in the CE, however, again it results in an increased number of variables and thus the computational burden as well. Furthermore, an effort to minimize the number of variables selecting smaller scheduling horizon results in a sub-optimal solutions [106]. Conventional methods show an exponential rise in the computing time with the problem size that depends on the number of variables and constraints. Number of variables (problem size) on the other hand increases exponentially with the product of the number of time slots in a scheduling horizon (N ) and the number of schedulable devices (N dev). Computing times claimed in various papers with the related variables are summarized in Table 3.16. Empirical relations between T ∼N dev and T ∼N , giving a measure of the problem complexity in terms of problem size may be correlated. Such a relation based on the data given in Table 3.16 for [105] is formulated as equations 4.5 and 4.6.

T = 4 × e0.955×N dev with R2 > 0.95

(3.10)

T = 0.47 × e0.034×N with R2 > 0.95

(3.11)

Heuristic-based methods solve HEMO problem as a combinatorial optimization problem. Typically the number of possible combinations (N c) in an optimization problem are related to the number of schedulable devices (n). For 24 number of slots with HAs’ operating time which is randomly selected between 1-24 hours. It is an exponential graph between N c and the computing time which is presented

115


Chapter 3


Table 3.16: Computing Time with Related Parameters for Deterministic HEMO Models Reference Optimization method

No HDs

[100]

CP: Interior point method MILP: BB and CTP

[14] [110] [62]

[105]

[138] [144] [113] [135]

[132] [141]

[93]

[61] [148]

16

Resolution (minutes/slot) 60

Horizon (hours)/ of slots 24

1, 2 and 3

10,20,30 and 60

GA MILP

16 10-25

12 60

24/24-144, repeats for HDs 24/120 24/24

LP: QP and PSO LP

120

3

24/ 480

NA

60

24/24

MILP:CPLEX and Heuristic LP EA and Approx. EA

NA

15

24/96

5-30 5-25 Consumers 15 1

60 -

24/24 -

60

-/30 24/24

3 Feeders

60

24/ 24

7 -

60 1

24/24 24/1440

QP: Heuristic LP:DAP with real adjustment Modified-GA compared with DP,GA and ES MILP:CPLEX CP: DP and heuristic

of

No

Computing (sec)

time

Few seconds on CP solver 28,4,0.5,0.3; 44,6,1,0.4; and 68,8,2,0.5 msecs Very Low Very high for Ndev>24 PSO=0.012 and QP=0.957 Exponential rise with Ndev Heuristics=46 and CPLEX=10 0.8-2.9 25-10033 minutes and 15-4978 minutes 0.5 11

DP=9.13,GA=6.7, ES=5.28 and Modified-GA=3 0.11 DP impracticable; heuristic very fast/practicable

in [55] and formulated using Equation 4.7.

T = 0.557 × e2.48×n

(3.12)

It is also observed that very few publications included computing times for their pertinent algorithms. Although, the remaining time is acceptable for most of the deterministic models; however, in RT stochastic based applications, the parameter gets significance. Tables 3.16 and 3.17 present the value of computational time 116


Chapter 3


Table 3.17: Computing Time with Related Parameters for Stochastic HEMO Models Reference Optimization method

No. HDs

[111]

SDP

[108]

-

Resolution (minutes/slot) 15

Horizon (hours)/ No. of slots 24/96

SOP and RO

6

5

24/288

[114]

SOP: MDP, MC and Lagrange

24/96

[153]

MILP and SDP RO 2-stage with SDP and MDP

4 per 15 home (50400 homes) Aggregated60 5-30 60 Occupancy 1 hour/ based 15min

24/24 7Days/1Day for respective stages 24/24 and 24/96 24/-

[113] [157]

[139]

RO

[160]

SDP and 2 Dim-BPSO Linprog and MPC CCO with Gradient-PSO

[171] [169]

[170]

CP, used factor for uncertainty

20 35 4

of

and

60 and 15 -

24/24

40 houses 5

60 60

12/12 MPC 24/24

10

60

24/24

for

Computing (sec)

time

Reasonably low upto 96 slots Robust=60/slot and Stochastic=140/slot 3.3-23.4 hours

MILP 5.8 times faster than SDP 2.8-10 Low with 2 stages HEMO 0.028 and 0.098 SDP=very large, PSO=very low 0.2 GPSO-CCO=0.1 and GPSO-LHS= 110 2.36

claimed by various researchers for their proposed models based on the deterministic and stochastic approaches, respectively. Important HEMO parameters including N dev, N and number of recourse stages affecting the computing time are also included in these tables. Computational times for optimization models presented in the aforementioned tables are analysed as follows: • In deterministic modeling, small problems take low computation time when solved using LP techniques. The time rises exponentially with the problem size based on N dev and N . For larger size problems, meta-heuristics based methods give near-optimal solutions in very small computation time. Such 117


Chapter 3


methods may be used even for very large and complex problems. • Stochastic modeling carries curse of dimensionality due to recourse variables. The resulting high problem complexity causes an exponentially increased computation time. SOP has been used only up to two stages even for small-scale problems. The problem dimensionality increases infinitely with increasing number of stages. MPC with LP or heuristics and RO methods are used to manage computation time within acceptable range while handling uncertainty in HEMO problems.

118


Chapter 4 Models, problem formulation and proposed solutions

119

Chapter 4

4.1

4.1. PROPOSED SYSTEM MODEL

Proposed system model

The proposed HEMS architecture is already shown in Fig. 1.1. The DR-based HEMS is based on load shifting toward off-peak hours using dynamic tariffs. In the DRSREOD-based HEMS model, the load shifting of appliances is synergized with the optimal dispatch of the PV units, the storage battery and the grid by means of dynamic tariffs. The HEMS controller is a vital component that carries all of the computational intelligence, based on a control algorithm, necessary for optimizing the HEMS operations. The main objective of the DRSREOD-based HEMS is to compute the optimal schedules for shiftable appliances based on DS/MS and synergize with the optimal dispatch of the PV units, the storage battery and the grid to reduce peak/overall demand for the utility and reduce CE for the consumer while keeping T BD within acceptable limits. For the execution of HEMS operations, a time horizon of 24 hours is adopted. Each shiftable appliance is to be operated once within a proposed interval for a specified length of time in the scheduling horizon. For scheduling, the time horizon is divided into N slots. The value of CE for energy from the grid decreases with increasing N ; however, the computational burden simultaneously increases. A tradeoff analysis conducted by previous researchers suggests an optimal slot length of 10 minutes, corresponding to 144 slots in a 24-hour scheduling horizon, and the same time division has been adopted in this optimization model. The operating scheme focuses on the shared parallel operation of the PV units, the storage battery and the grid based on P pv, SOC and the maximum charge/discharge rates. The PV units are the preferred source from which to supply the scheduled loads. Any excess PV energy in a time slot is stored in the storage battery to be used during peak hours to reduce CE. Furthermore, the optimal size for a local DG for a DRSREOD-based HEMS is computed, considering the CE and T BD TOs. The proposed DG supplies the scheduled load during load shedding hours in parallel with the PV units and the

120


Chapter 4


storage battery to avoid power interruptions. The grid is not included in the dispatch scheme during load shedding hours.

4.1.1

HAs

HAs are modeled based on their DR classes. NSHAs/fixed loads, e.g., lights and fans, are operated as and when needed and cannot be scheduled. SHAs can be shifted toward off-peak hours for optimized scheduling. Each SHA is to be operated for LOT time slots between two limits given by Alpha and Beta, as illustrated in Figs. 2-4. The consumer specifies these parameters based on his own convenience. SHAs are classified into interruptible and non-interruptible appliances. Appliances of the former type, e.g., pool pumps, air conditioners and electric geysers, can be interrupted once started and hence may be operated in two or more separated sets of time slots. Those of the latter type, e.g., washing machines and dryers, need to be operated until completion without interruption. SHAs, being the most flexibly available for scheduling, and NSHAs are both included in our model. We further classify SHAs into groups for AS and DS. In AS, the operation of an SHA is shifted such that the job will be completed before the preferred ending time specified by the consumer (Beta). T BD is computed by measuring the shift of the actual ending time of SHA operation in advance of the preferred ending time. In DS, the operation of an SHA is shifted such that the start of the job is delayed from the preferred start time (Alpha). T BD is computed by measuring the actual time delay of the start of SHA operation relative to the preferred start time. An MS model is proposed in which each SHA is explicitly selected for either DS or AS. MS not only results in greater reductions in CE and T BD while making more optimal use of the available energy sources (PV units, the SB and the grid) but also provides the consumer with more diverse options for scheduling his SHAs on a time line. As an example, a consumer may select a rice cooker/oven for AS operation during the evening. In this case, T BD is calculated based on the advance 121


Chapter 4


completion of the job relative to the specified ending time limit. Reducing the T BD for the consumer means ensuring the availability of freshly cooked rice/food for the consumer in the evening as close as possible to his dining time. The SHAs, with their preferred operating time intervals, and the NSHAs considered in our simulations are listed in Tables 4.1 and 4.2. Table 4.1: SHAs and scheduling specifications for the DS and MS scenarios

SHA

Power (kWh)

LOT (slots)

Air Conditioner 1 (Reversible) Air Conditioner 2 (Reversible) Air Conditioner 3 (Reversible) Air Conditioner 4 (Reversible) Dishwasher 1 Dishwasher 2 Electric Geyser 1 Rice Cooker/Oven 1 (Manual) Computer/Laptop (Manual) Washing Machine Water Pump Electric Geyser 2 Rice Cooker/Oven 2 (Manual) Iron (Manual)

1

4.1.2

18

MS (Start/End Limits) 01-36 (DS)

DS (Start/End Limits) 01-36 (DS)

1

9

37-54 (DS)

37-54 (DS)

1

9

103-120 (DS)

103-120 (DS)

1

12

121-144 (DS)

121-144 (DS)

0.6 0.6 0.8 0.4

3 3 6 3

49-102 (DS) 127-144 (DS) 01-36 (DS) 73-81 (DS)

49-102 (DS) 127-144 (DS) 01-36 (DS) 73-81 (DS)

0.1

6

114-144 (DS)

114-144 (DS)

0.7 0.7 0.8 0.4

9 3 6 3

93-123 (AS) 37-117 (AS) 55-121 (AS) 100-117 (AS)

114-144 114-144 115-126 114-120

0.6

3

55-117 (AS)

114-144 (DS)

(DS) (DS) (DS) (DS)

Electricity tariffs

Dynamic tariffs are the key to implementing DR/DRSREOD-based HEMSs. The major types of dynamic tariffs include RTP, DAP and ToU tariffs. RTP is typically communicated by the utility to the consumer on an hourly basis, whereas 122


Chapter 4

4.1. PROPOSED SYSTEM MODEL Table 4.2: NSHAs considered for scheduling

Fixed Load 01 Light + 02 Fans + 01 Refrigerator 02 Lights + 02 Fans + 01 Refrigerator 01 Lights + 02 Fans + 01 Refrigerator 02 Lights + 02 Fans + 01 Refrigerator 03 Lights + 03 Fans + 01 Refrigerator + 01 TV 04 Lights + 03 Fans + 01 Refrigerator + 01 TV

Power (kWh) 0.2

Start/End (slots) 01-36

0.25

37-54

0.2

55-78

0.25

79-108

0.3

109-114

0.35

115-144

DAP is communicated on a day-ahead basis. ToU tariffs comprise two or more rates for electricity during peak, off-peak and mid-peak hours of the day for a specified period (typically 3-6 months). In combination with tariffs, utilities charge higher rates at higher power levels, in a scheme called IBR, to discourage users from concentrating loads at specific times, which may lead to the re-emergence of peaks. The ToU tariff scheme combined with IBR is formulated as follows:

EP = [(EP 1, P T 1, IBR1), (EP 2, P T 2, IBR2), (EP 3, P T 3, IBR3)]

where EP 1, EP 2 and EP 3 are the normal tariffs at peak, off-peak and mid-peak times, respectively, and a tariff of EP × IBR applies above power threshold values of P T 1, P T 2 and P T 3 at the corresponding times. ToU- and DAP-based algorithms provide solutions that require very little computation time and are viable for real-time household applications [194]. A 2-stage ToU tariff scheme with an 123


Chapter 4


IBR value of 1.4 was used in our simulations, consistent with the two EP levels used by British Columbia Hydro [195]. For the application of the IBR factor, a threshold power demand of 2.4 kW was considered in all scenarios [196]. Although this specific 2-stage ToU tariff scheme was adopted for the simulations, the proposed algorithm is generic in nature and works equally well for DAP schemes.

4.1.3

RESs

Solar PV units and wind power units are the most widely used types of RESs in homes [59, 197]. The integration of RESs into a HEMS results in overall reductions in CE, the demand supplied from the grid and the peak load. However, the power they supply is intermittent in nature, and consequently, they give rise to complex scheduling models [101]. Our model is based on the forecasted irradiation levels for PV operation. Most researchers have not included the cost of generations from local RESs in their models [63, 102], and they have treated RESs as part of the existing infrastructure. By virtue of rebate-based incentives and continued research, the cost of RESs has been greatly reduced, and as EPs have increased, they have now become a popular means of harvesting energy. The PV units are treated as part of the existing infrastructure in our model. The power obtained from the PV units is formulated as follows:

Ppv = Parea × Iirrad × ηpv × ηconv

(4.1)

where Ppv = PV power in kWh Parea = PV plate area in meter2 Iirrad = PV irradiation in kWh/meter2 ηpv = PV electrical efficiency ηconv = Converter efficiency

124


Chapter 4


In the simulations of our model, we included a PV system with the specifications given in Table 4.3. Solar irradiation data as measured by the Pakistan Engineering Council in Islamabad were used in our simulations [198]. Table 4.3: PV system specifications

Parameter Total capacity Rating of each panel Number of panels and panel area Efficiency of PV panels

4.1.4

Value 5 kW 250 W 20, 32 m2 15%

ESS

The SB is one of the most popular types of ESS that is integrated with a PV system to introduce flexibility in PV dispatch. Normally, surplus energy available from the PV system is stored in the SB (or sold to the grid). The stored energy is later consumed to supply the load when required or during peak hours to increase economy [102]. When the PV system is unable to fully support the necessary load, the difference is supplied from the SB and/or the main grid. With decreasing RES/ESS prices, more consumers are deploying PV/SB systems for parallel operation with the grid to improve the economy, reliability and quality of their power supply. The inverter and the SB are key components in the proposed DRSREOD-based HEMS. The specifications of these components that were considered in the simulations are given in Table 4.4. The efficiencies of the inverter and the SB are included in the formulation. The net loss for the SB is assumed to be 20% and is considered during charging.

4.1.5

LDG

The PV energy supply is intermittent in nature and lacks dispatchability. The SB is used to store the excess PV energy from the RESs for use during peak hours. 125


Chapter 4

4.1. PROPOSED SYSTEM MODEL Table 4.4: SB and inverter specifications

Parameter Inverter rating Inverter efficiency SB Ah SB voltage SB capacity SB charge rate SB discharge rate Minimum SOC Maximum SOC SB efficiency

Value 5 kW 70% 600 Ah 48 V 4.8 kWh/slot 0.48 kWh/slot 0.32 kWh/slot 30% 95% 80%

Therefore, the amount of energy stored in the SB greatly depends on the RESs. For energy-deficient/unreliable grids subjected to LS, an optimally sized LDG is proposed for use by a consumer with an existing DRSREOD-based infrastructure. The LDG participates in dispatch along with the PV system and the SB during LS hours as described by Eq. 4.22, and its size is computed as per Eq. 4.14. The LDG sizing algorithm also considers the TOs between P gsize, CE and T BD. The specifications of the LDG to cope with the LSD are used in the simulation of proposed model as given in Table 4.5. The emission factor is computed as per Eq. 4.6 using the pertinent data given in [202], [203]. The cost of energy for the LDG is according to the levelized cost of energy for such units given in [204]. Table 4.5: LDG specifications

Parameter LDG rating Power factor Emission factor Levelized cost of generation for the LDG

126

Value 1 kVA 0.8 1.6 Lbs./ kWh 17 Cents/ kWh


Chapter 4

4.2

4.2. FORMULATING THE HEMS OPTIMIZATION PROBLEM

Formulating the HEMS optimization problem

The HEMS optimization problem is formulated for one or more objective functions under few constraints. The problem is solved to obtain a set of inputs/decision variables that optimize the output values of the performance parameters while satisfying the constraints. The HEMS problem is formulated based on the control parameters given below:

A = [a1 , a2 , .., ak ] = SHAs available for scheduling T = [1, 2, 3, .., N ] = Slot numbers in the scheduling horizon P app = [P1 , P2 , .., Pk ] = Per-slot power ratings of the SHAs LOT = [LOT1 , LOT2 , .., LOTk ] = Lengths of operation of the SHAs Alpha = [Alpha1 , Alpha2 , .., Alphak ] = Starting slots for the operating time intervals of the SHAs Beta = [Beta1 , Beta2 , .., Betak ] = Ending slots for the operating time intervals of the SHAs EP = [EP1 , EP2 , .., EPN ] = ToU EPs in cents/kWh IBR = [IBR1 , IBR2 , .., IBRN ] = Factor by which to multiply EP for loads greater than P T T s = [T s1 , T s2 , .., T sk ] = Decision vector consisting of the start time for each SHA

In our proposed algorithm, the vector T s is generated via a GA. Based on T s, a decision vector Xa (dim: 1 × N ) is derived to specify the scheduled power of the ath SHA over the complete scheduling horizon, as follows:    P app(a) : f or T s(a) + LOT (a) > i ≥ T s(a) X a (i) =   0 : f or T s(a) > i ≥ T s(a) + LOT (a) 127


Chapter 4


Vectors X1 , .., Xk are developed in a similar way for each SHA, considering the corresponding T s(a) for each SHA, numbered as A = 1, 2, .., k. The decision vectors X1 , .., Xk are combined into a matrix, denoted by P ower matrix, as follows:

P ower matrix = [X1 , X2 , ..., Xk ]t

(4.2)

P ower matrix is summed in a columnwise manner to obtain a scheduling vector P schd sh that specifies the power requirement in each time slot in the scheduling horizon. Accordingly, a power scheduling vector, based on MILP, is formulated as follows: P schd sh =

N X k X

X(a, n)

(4.3)

n=1 a=1

where X(a, n) is a generalized element of the derived P ower matrix. The load vector for fixed appliances, P load f ix, is added to P schd sh to compute the final scheduled load vector, P schd, as follows:

P schd = P schd sh + P load f ix

4.2.1

(4.4)

Objectives for the HEMS problem

The main objectives for a HEMS generally include minimizing the CE for energy from the grid, minimizing the T BD for the consumer, reducing the peak load/P AR, and reducing emissions.

To achieve the aforementioned these objectives, the

HEMS problem is formulated for SHA scheduling while simultaneously computing P schd and synergizing the scheduling with RES, ESS and power grid dispatch for N time slots over a specified scheduling horizon.

128


Chapter 4 4.2.1.1

4.2. FORMULATING THE HEMS OPTIMIZATION PROBLEM Minimization of CE/ CEnet

The objective function corresponding to minimizing the cost of the energy purchased from the utility (CE) for a HEMS without a PV/SB system is formulated as follows: N X

M inimize

(P schd × EP )

(4.5)

n=1

where P schd is the scheduling vector as given in Eq. 4.4 and the EP is the tariff vector as given in Eq. 4.1.2. Because all of the power is supplied by the grid in this case, P schd is equal to P grid. For a DRSREOD-based HEMS with a PV system and an SB, the vector for P grid is formulated as follows:

P grid = P schd − P pv + P sold + P chg − P dis

(4.6)

In this case, the objective function for minimizing the cost of the energy purchased from the grid (CE) is formulated as follows:

M inimize

N X

(P grid × EP )

(4.7)

n=1

where P grid and EP are vectors specifying the energy purchased from the utility and its price per kWh, respectively, for each time slot. The objective function for maximizing the cost of the energy sold to the grid (CEsold) is formulated as follows:

M aximize

N X

(P sold × EP f )

(4.8)

n=1

where P sold is the amount of energy in kWh sold to the grid by the consumer and the EP f is the feed-in tariff in cents/kWh. In our model, we take the value of EP f to be 0.7 × EP . The net bill to be paid to the utility (in cents) is formulated

129


Chapter 4


as follows:

N T bill = CE − CEsold

(4.9)

For a DRSREODLDG-based HEMS with PV, SB, the grid and the LDG, the objective function to minimize the net cost of energy to be paid by the consumer, CEnet, is formulated as follows:

M inimize

N X

(P gd × P E + P gn × P Eg − P sold × P Ef )

(4.10)

n=1

where P gd and P E are the energy purchased from the utility and its price in Cents/kWh; P sold and P Ef are the energy sold to the utility by the consumer and its feed-in price in Cents/kWh; and P gn and P Eg are the energy supplied from the LDG and its levelized price in Cents/kWh. The objective function for CEnet has been derived from the work in [7] by including the cost of energy for the LDG operation and the cost of the excess PV energy sold to the utility. A factor P sold × CEM iss can be excluded from CEnet as a reward for reducing the supply-side emissions through the PV energy sold to the utility.

4.2.1.2

Minimization of T BD for the consumer

The average T BD in DS due to delay in the start times of the SHAs, denoted by T BD(D), is obtained by taking the average of the normalized delays for all Pk1 appliances. This quantity is formulated as follows: T BD(D) = a=1 ((T s − Alpha)/ (Beta − LOT − Alpha + 1))/k1 where Alpha and Beta represent the flexible time ranges specified by the consumer, indicating the start and end limits for SHA operation. LOT is a vector consisting of the lengths of operation time required for each SHA to complete its job. T s is the decision vector consisting of the start times for all of the SHAs, which are the values to be varied to find the optimal solution. The numerator and denominator

130


Chapter 4


in Eq. 4.2.1.2 represent the actual and maximum delays, respectively, in starting the operation of an appliance. Meanwhile, k1 is the number of SHAs designated for DS. Conceptual diagrams illustrating T BD(D) are shown in Figs. 4.1 and 4.2. Alpha(a)

Beta(a)

t(a)=Beta (a)-LOT(a) Range for starng me of appliance LOT(a)

LOT(a)

Preered starng me

Figure 4.1: Time range available for appliance operation in DS Alpha(a)

Beta(a)

t(a) = Beta(a)-LOT(a) Maximum delay in start time=Beta(a) -LOT(a)-Alpha(a)

Actual delay in start time=Ts(a)- Alpha(a)

Tsa

LOT(a)

Figure 4.2: Maximum and actual values of the time delay in DS

The T BD(D) will take its minimum value of 0 when the corresponding SHA starts operation at Alpha, i.e., the start of the operating time range specified by the consumer. It will attain its maximum value of 1 when the T s(a) is equal to Beta(a) − LOT (a)+1, i.e., the latest start time for the SHA that results in completion of the job at the latest allowed time Beta(a) specified by the consumer. These limits/bounds must be respected when selecting T s and are formulated as follows:

Lb = Alpha and U b = Beta − LOT + 1 131

(4.11)


Chapter 4


The average T BD in AS due to the advance completion of the jobs of the SHAs, denoted by T BD(A), is obtained by averaging the normalized advance-completion times for all of the appliances. This quantity is formulated as follows: T BD(A) = Pk2 a=1 ((Beta − T s − LOT + 1)/(Beta − LOT − Alpha + 1))/k2 The numerator and denominator in Eq. 4.2.1.2 represent the actual and maximum forward shifts, respectively, in the job ending time for an appliance, and k2 is the number of SHAs designated for AS. The conceptual diagrams illustrating the T BD(A) are shown in Figs. 4.3 and 4.4. The T BD(A) will take its minimum value of 0 when the corresponding SHA completes its job at Beta(a), i.e., when T s(a) + LOT (a)-1 is equal to Beta(a). It will attain its maximum value of 1 when T s(a) is equal to Alpha(a), i.e., when the ending time for the operation of the appliance is Alpha(a) + LOT (a)-1. In this scheme, the user is concerned with the ending times of the SHA operation (Beta), and T BD(A) is computed by considering the distances between the actual and desired ending times of the SHAs, i.e., the distance of T s + LOT -1 from Beta. By contrast, the DS scheme focuses on the start times of the SHA operation, and the T BD(D) is calculated based on the distances between the actual and desired start times of the SHAs, i.e., the distance of T s from Alpha. In the MS, some appliances are selected for AS, whereas the others are designated for DS. The average T BD for a total of k appliances in the MS mode, denoted by T BD(M ), is formulated as follows:

T BD(M ) = T BD(D) + T BD(A)

(4.12)

For the simulation of the MS scheme and its comparison with the DS scheme, some of the SHAs subjected to DS in the simple DS scenario were changed to AS 132


Chapter 4

4.2. FORMULATING THE HEMS OPTIMIZATION PROBLEM Alpha(a)

Beta(a)

t(a)=Alpha (a)+LOT(a) Range for ending time of appliance

Preffered ending time

LOT(a)

LOT(a)

Figure 4.3: Time range available for appliance operation in AS Alpha(a)

Beta(a)

Ts(a) Alpha(a)+LOT(a)

Maximum advanced shift of ending time=beta(a)-LOT(a)-alpha(a)

LOT(a)

Actual advanced shi of ending me =Beta(a)-Ts(a)LOT(a)

LOT(a)

Figure 4.4: Maximum and actual values of the advance-completion time in AS

mode in the MS scenario. To model the AS of these appliances, the preferred Beta values were calculated from the Alpha values for the same appliances in the DS scenario. Based on these preferred Beta values, new Alpha values were assigned to these SHAs corresponding to earlier times, as shown in Table IV.

4.2.1.3

Minimization of the peak load

The objective function for minimizing the peak load supplied from the grid is formulated as follows: M inimize P eak(P grid)

(4.13)

For a DR-based HEMS, P grid is equal to P schd, whereas for a DRSREOD-based HEMS, P grid includes the effects of the DR as well as the overall demand reduction 133


Chapter 4


due to the optimal dispatch of the RESs and ESS as described by Eq. 4.6.

4.2.1.4

Minimal size of the generator required to cope with LS

The objective function for minimizing the size of the dispatchable generator while maintaining the ability to supply the required load during LS is formulated as follows: M inimize P eak(P gen)

(4.14)

where P gen is computed by implementing an energy balance constraint as expressed in Eq. 4.24.

4.2.1.5

Minimization of emissions

The objective function for minimizing the emissions from non-RES generation units is formulated as follows:

M inimize

N X (F cons × EF )

(4.15)

n=1

where EF (kg/liter) is the emission factor and F cons (liters) is the amount of the consumed fuel. EF depends on the type of fuel and the engine characteristics [31].

4.2.2

Constraints for the HEMS Problem

Constraints for the HEMS problem are introduced based on various components of the HEMS, as described below.

4.2.2.1

HA constraints

Scheduling constraints are imposed on the HAs to satisfy the user’s preferences; these constraints include defined time deadlines for the completion of operation (Alpha/Beta) [14] and non-interruptibility constraints [119]. These constraints 134


Chapter 4


are implemented by introducing the upper and lower bounds on the T s vector as per Eq. 4.11.

4.2.2.2

Tariff constraints

The tariffs issued by some utilities impose maximum power consumption limits and offer lower rates for respecting those limits [121]. Our model is based on a ToU tariff scheme combined with IBR. For a power demand of > 0.4 kW/slot, a penalty factor of 1.4 times the regular tariff is applied as the IBR, as per the practice of British Columbia Hydro [14].

4.2.2.3

SB constraints

The SB is subject to constraints on its SOC, which must be within certain minimum and maximum allowable levels [120]. These constraints ensure a satisfactory service life of the SB and are formulated as follows:

SOC min < SOC < SOC max

(4.16)

A second set of constraints on the SB pertains to its maximum charge and discharge rates (P chg and P dis), which may not exceed the maximum permissible rates P chg max and P dis max, respectively. These constraints are formulated as follows: P chg ≤ P chg max and P dis ≤ P dis max

(4.17)

Furthermore, SOC(i + 1), i.e., the state of charge of the SB in the next time slot, depends on its SOC(i) in the present slot and on whether the SB is charging or discharging in the present slot. Accordingly, the following constraint is implemented for the SB:

135


Chapter 4


SOC(i + 1) =

   SOC(i) + 0.8 × P chg : when charging   SOC(i) − P dis :

(4.18)

when discharging

A net energy loss of 20% is assumed for the SB during charging.

In addition, the SB should be either charging or discharging in each time slot. To implement this constraint, a binary variable BS is formulated as follows:

BS =

   1 : when charging

(4.19)

  0 : when discharging The following formulation enforces the necessary constraints for the charging/discharging status of the SB in each time slot:

P chg = P chg × BS and P dis = P dis × (1 − BS)

4.2.2.4

(4.20)

Energy balance constraint

This constraint ensures that in each time slot, the total energy generated is equal to the total energy consumed by the load, or the sum of the energy inputs is equal to the sum of the energy outputs for the system [93]. This constraint for a DRSREOD-based HEMS is formulated as follows:

P grid + P pv + P dis = P schd + P chg + P sold

(4.21)

This constraint is implemented in the proposed algorithm 2 for a DRSREODbased HEMS to compute the optimal schedules and dispatch plans for the loads and energy sources, respectively. Meanwhile, for the optimal sizing of a DG for a DRSREOD-based HEMS, two additional parameters, P gen and P dl, are introduced to reformulate the constraint 136


Chapter 4


as follows:

P grid + P gen + P pv + P dis = P schd + P chg + P sold + P dl

(4.22)

This constraint is implemented to compute the optimal power to be supplied by the DG during LS time slots together with the optimal tradeoff schedules and dispatch plans for the loads and energy sources, respectively.

4.2.3

Optimization techniques for solving the HEMS problem

HEMS optimization is a combinatorial optimization problem. Most HEMS problems are non-linear, non-convex constrained, and multi-dimensional in nature and have a large number of solutions that grows exponentially for large-scale problems [55]. Optimization techniques for solving such problems include both conventional and advanced heuristic methods.

4.2.3.1

Conventional techniques

Conventional techniques include LP, convex programming (CP) and MILP. LP is used for linear models. This method yields a solution in polynomial time (PT) for small-scale problems only. Furthermore, linear models are not able to represent most HEMS problems accurately. The CP is certain to converge if a solution exists. This method is applied in cases in which the non-linear models can be transformed into linear ones. MILP is used for models that include integers as well as continuous variables, called MILP models, which are NP-complete in complexity. LP methods have been successfully used only for small-scale HEMS problems. For large-scale linear and non-linear problems, conventional methods become computationally impracticable, and the problems become NP-hard [122].

137


Chapter 4 4.2.3.2

4.2. FORMULATING THE HEMS OPTIMIZATION PROBLEM Advanced heuristic techniques

Over the last decade, advanced heuristic techniques have emerged as the single best choice for obtaining near-optimal robust solutions to complex HEMS problems. Metaheuristics are a general class of heuristics that can be applied to a large number of problems with only minor modifications for specific cases. These advanced tools have been used to solve optimization problems that were believed to be impossible in the past, such as non-convex and NP-hard problems, in very short computation times [123]. The renowned metaheuristic tools used to solve HEMS optimization problems include GAs, particle swarm optimization, ant colony optimization, and evolutionary programming. GAs, as an efficient and robust metaheuristic approach, have been used very successfully to solve combinatorial optimization problems. A GA relies on natural selection and genetics, searches multiple paths, explores multiple maxima/minima in parallel and can escape from local minima by means of niching methods [124]. It uses parameter coding instead of actual parameters, thereby enabling it to develop the next state from the current state with minimal computation. A GA evaluates the fitness of each string to guide its search after evaluating the performance of one or more fitness functions. To handle constraints, a GA uses chromosome rejection, repair and other genetic operators [123]. However, such metaheuristic techniques have primarily been used to solve single-objective optimization problems only. To solve multi-objective optimization problems with metaheuristics, they have generally been transformed into single-objective optimization problems.

4.2.4

Techniques for handling multi-objectivity in the HEMS problem

Most of the HEMS problems encountered in real life are MOO problems with mutually conflicting objectives. Minimizing the CE is the main objective in the majority of the published research [14, 64]. Minimizing the T BD is the second 138


Chapter 4


most important objective [14, 53, 56, 57, 67] from the consumer perspective. A tradeoff exists between the CE and the T BD. The solution to the HEMS problem actually consists of an optimal set of solutions, each offering some optimized tradeoff between the objectives. Various methods have been used to consider important TOs between objectives. The most widely used approach is the WSM [14, 53, 54, 57]. This is an apriori method that transforms a multi-objective HEMS problem into a single-objective optimization problem to obtain a single optimal solution. Such methods do not provide a clear understanding of the relation between the objectives to allow a consumer to choose among specific preferences and do not even enable him to improve the solution. Potentially better solutions that are feasible for a specific consumer may be missed because this method does not allow for any feedback regarding the given preferences. A tradeoff analysis between the objectives in HEMS problems is very important because it enables consumers to make decisions after evaluating a diverse set of available optimal choices. Pareto-based MOO, a posteriori method, provides a diverse set of optimal solutions for multiple objectives based on the concept of Pareto dominance. The MOO problem for a HEMS with decision vector T s and m objectives for Pareto-based optimization is formulated as follows: minimize the objective vector F (T s) = [F1 (T s), F2 (T s), .., Fm (T s)] subject to the given constraints. A solution T s1 is said to dominate another T s2 when T s1 is better than T s2 in at least one objective and is no worse in any other. The set of non-dominated solutions composes the Pareto-optimal set, or Pareto front. The recently introduced MOGA includes features for implementing Pareto optimization. POSs/PFs providing optimal TOs between the CE and the T BD for a HEMS have been computed in this study by using MOGA with the Pareto optimization features to provide HEMS consumers with diverse options. A consumer can then choose the best available option in accordance with his needs. Another objective, namely, the identification of the minimal DG size necessary to cope with load shedding, has also been modeled in this study. The minimal P gsize is selected based on a POS computed using 139


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING MOGA for the TOs between P gsize, CE and T BD. This method ensures that no potentially superior solution is missed. The only disadvantage of this method is its longer computing time due to additional computations. However, with deterministic models, ToU/DAP tariffs and the use of advanced metaheuristics such as MOGA, the technique can be used very successfully for identifying tradeoff-based solutions to the HEMS problem that are both optimal and feasible, as validated in this research.

4.3

Algorithms for a DR-based HEMS, a DRSREODbased HEMS and optimal DG sizing

Three algorithms are proposed in this study: - Algorithm 1 for a DR-based HEMS with either DS or MS of the SHAs - Algorithm 2 for a DRSREOD-based HEMS with either DS or MS of the SHAs - Algorithm 3 for optimal DG sizing to cope with LS in a DRSREOD-based HEMS with MS The algorithms are presented in the following subsections.

4.3.1

Algorithm 1 for a DR-based HEMS with DS or MS

This algorithm computes a set of solutions that provide optimal TOs between the CE and the T BD and a solution with the minimal CE based on SHA scheduling. For optimal scheduling, T s is generated heuristically, within the specified bounds, using MOGA. T end (the vector of times at which the SHAs complete their operation) is computed by adding the LOT values specified by the consumer. A power matrix (dim: k × N ) is generated with parameter values equal to the power values of the corresponding SHAs for the time slots from T s to T end. The vector P schd is obtained by summing up the power matrix and adding P load f ix, as shown in lines 9-18. IBR is applied for slots with P schd values greater than T P . The CE 140


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING Algorithm 1 Algorithm for a DR-based HEMS with either DS or MS for the SHAs Input: EP , IBR, P T , P app, Stype, Alpha, Beta,

—–Computing tariffs with IBR—–

LOT , P load f ix

19:

Output: Optimal TOs for CE and T BD in PF/POS

20:

form with the corresponding set of T s, minimized CE

21:

and related T BD

22:

1: Initialize input parameters

23:

2: Call MOGA

—–Computing fitness function for CE—–

3: Initialize Ts within bounds using Eq. 4.11

24:

4: for iter = 1:Ng max

—–Computing fitness function for TBD—–

5:

25:

6:

if iter > 1 Generate new Ts populations within bounds

for j = 1: N if Pschd(j) > PT EP(j) = IBR × EP(j) end end

Compute CE = sum(EP × Pschd)

for a = 1:k

26:

if Stype = DS

using GA operations

27:

7:

LOT(a)-Alpha(a)+1)

end

T BD(D)(a) = (Ts(a)-Alpha(a))/(Beta(a)-

—–Computing Pschd vector for DR-based load

28:

scheduling—–

29:

8:

Tend = Ts+LOT-1

LOT(a)+1)/(Beta(a)-LOT(a)-Alpha(a)+1)

9:

for i = 1:k

30:

10:

for j = 1:N if (j ≥ T s(i) &&j ≤ T end(i))

11:

Power matrix(i,j) = Papp(i)

12: 13:

else

14:

Power matrix(i,j) = 0

15: 16:

end end

else T BD(A)(a)

=

(Beta(a)-Ts(a)-

end

31:

end

32:

Compute TBD(M or D) = (sum(T BD(D))+sum(T BD(A)))/k

33: end 34: Exit MOGA; return results as CE/TBD TOs and corresponding Ts 35: Selection of a feasible tradeoff solution by the consumer

17:

end

18:

Pschd = sum(Power matrix)+ Pload fix

and the T BD (for MS or DS) are computed using the equations specified in lines 24, 27, 29 and 32. Tradeoff solutions for the CE and the T BD are obtained using MOGA in POS/PF form, and the solution with the minimal CE is selected as the most feasible solution as a reference for the consumer.

4.3.2

Algorithm 2 for a DRSREOD-based HEMS with DS or MS

This algorithm computes a set of solutions that provide optimal TOs between the CE and the T BD and a solution with the minimal CE based on DR-based SHA 141


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING

Algorithm 2 Algorithm for a DRSREOD-based HEMS with either DS or MS for the SHAs Input: EP , IBR, T P , P app, Stype, Alpha, Beta,

——-Dispatch when PV energy ≤ Pschd——-

LOT , P load f ix, SOC(0),

23:

SOC max, SOC min, P chg max, P dis max, P pv

24:

Output: Optimal TOs for CE and T BD in PF/POS

SOC min) && (EP(j) ≤ price set))

form with T s and minimized CE and T BD

25:


26:

2: Call MOGA

27:


price set))

4: for iter = 1: Ng max

28:

5:

SOC min)

if iter > 1

6:

Generate new Ts populations within bounds

if (SOC(j) ≤ SOC min)

30:

7:

31:

|

((SOC(j) >

Pgrid(j) = -Pres(j) SOC(j+1) = SOC(j) elseif ((SOC(j)> SOC min) && (EP(j)>

Pdis(j) = min(Pdis max,-Pres(j),SOC(j)-

29:

using GA operations end

case (Ppv(j) ≤ Pschd(j)) do

if Pdis(j) == Pdis max Pgrid(j) = Pschd(j)-Ppv(j)- pdis max elseif Pdis(j) == SOC(j)-SOC min

—-Computing Pschd vector for DR-based load

32:

scheduling—–

SOC min)

8: Compute Pschd using the method given in algorithm

33:

end

1, lines 8-18

34:

SOC(j+1) = SOC(j)-Pdis(j)

—–Computing dispatch for the PV system, SB

35:

end

and grid—–

36:

endcase

9:

——-Computing tariffs with IBR———

for j = 1:N

10:

Pgrid(j) = Pschd(j)-Ppv(j)-(SOC(j)-

37:

Pres(j) = Ppv(j)-Pschd(j)

——-Dispatch when PV energy > Pschd——-

38:

11:

case (Ppv(j) > Pschd(j)) do

39: 40:

if Pgrid(j)> TP EP(j) = IBR × EP(j) end

12:

if SOC(j) ≥ SOC max

13:

Psold(j) = Pres(j)

—–Computing fitness function for CE—–

14:

SOC(j+1) = SOC(j)

41:

15: 16:

else

Compute CE = sum(EP × Pgrid)

—–Computing fitness function for TBD—–

Pch(j) = min(Pch max,Pres(j),SOC max-

SOC(j)) 17:

end

42:

Compute TBD(M or D) using the method given

in algorithm 1, lines 25-32 if Pch(j) 6= Pres(j)

18:

Psold(j) = Pres(j)-Pch(j)

43: end 44: Exit MOGA; return results as CE/TBD TOs and

19:

end

corresponding Ts

20:

SOC(j+1) = SOC(j)+0.8* Pch(j)

45: Selection of a feasible tradeoff solution by the con-

21:

end

22:

endcase

sumer

142


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING scheduling synergized with the optimal dispatch of the PV system, the SB and the grid. First, the vector P schd is computed as in algorithm 1. The PV system is used as the preferred energy source to directly supply P schd. The dispatch planning is mainly based on the excess PV energy in each slot, denoted by P res, which is the arithmetic difference between P pv and P schd. For each slot in the scheduling horizon, two main cases arise with regard to the relative values of these two quantities. In each case, SOC and the maximum charge/discharge rates play major roles in the dispatch. In the first case, in which P pv is greater than P schd, as shown in line 11, the excess energy is stored in the SB if the SOC is less than its maximum value; otherwise, it is sold to the grid. The SB charging state depends on the condition given in line 16. If a value other than P res is computed, as shown in line 17, it indicates that either the maximum charge rate or the residual capacity of the SB before reaching the maximum SOC is restricting the complete storage of the excess PV energy in the SB. Hence, any excess energy left after charging the SB is sold to the grid. Notably, 20% of the energy is lost due to the net SB loss, and hence, the SOC is increased by only 80% of P ch in line 20. In the second case, in which P pv is less than or equal to P schd, as shown in line 23, the PV energy is insufficient to completely supply the load. The residual energy in this case will be supplied from the grid if the SOC is less than or equal to its minimum limit or from the SB if the SOC is greater than its minimum limit. Moreover, the SB still will not be discharged if cheap energy is available from the grid, as shown in line 24. The discharging state of the SB depends on the condition given in line 28. If the minimum computed value is equal to the maximum discharge rate or to the residual capacity of the SB before discharging to the minimum SOC, then one of these constraints restricts the ability to supply the full load from the SB, and the remaining load must be supplied from the grid, as shown in lines 30 and 32. For each slot in the scheduling horizon, one of the above 143


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING two cases will hold, and P grid will be computed accordingly. IBR is applied, and the CE and T BD(D or M ) are computed for each iteration of MOGA. The tradeoff solutions between these two objectives are obtained in POS/PF form, along with the solution with the minimal CE as the most feasible case.

4.3.3

Algorithm 3 for optimal DG sizing to cope with LS in a DRSREOD-based HEMS with MS

This HEMS optimization problem includes the DR-based MS of the SHAs synergized with the optimal dispatch of the PV system, the SB and the grid under normal grid conditions and the integration of an optimally sized DG for use during the LS hours. To identify the optimal size for a DG to cope with the maximum LS, the algorithm solves an optimization problem to compute the TOs between the CE, T BD(M ) and P gsize. First, the vector P schd is generated. The PV system is regarded as the preferred source to directly supply P schd. The dispatch planning is mainly based on the PV excess energy in each slot, denoted by P res, which is the arithmetic difference between P pv and P schd. Two main cases arise with regard to the relative values of these two quantities, and in each case, the SOC, the maximum charge/discharge rates, the grid status and the power from the DG play major roles in dispatch. In the first case, in which excess PV energy is available, as shown in line 11, the energy is stored in the SB if the SOC is less than its maximum value; otherwise, it is sold to the grid. However, during LS hours, the excess energy that would be sold to the grid is instead supplied to a dummy load, as shown in line 14. The SB charging state depends on the condition given in line 20. If a value other than P res is computed, it indicates that either the maximum charge rate or the limiting value of the SOC is restricting the complete storage of the excess PV energy in the SB. Hence, any excess energy left after charging the SB is sold to the grid, as

144


4.3. ALGORITHMS FOR A DR-BASED HEMS, A DRSREOD-BASED HEMS Chapter 4 AND OPTIMAL DG SIZING Algorithm 3 Algorithm for optimal DG sizing to cope with LS in a DRSREODbased HEMS with MS Input: EP , IBR, T P , P app, Stype, Alpha, Beta,

——-Dispatch when PV energy ≤ Pschd——-

LOT , P load f ix, P gds,

31:

SOC(0), SOC max, SOC min, P chg max, P dis max,

32:

P pv

SOC min) && (EP(j)≤ price set))

Output: Optimal TOs for P gsize, CE and T BD with

33:

T s and minimized P gsize

34:


35:

2: Call MOGA

36:


37:

4: for iter = 1: Ng max

38:

5:

39:

if iter > 1

6:

Generate new Ts populations within bounds

case (Ppv(j)≤ Pschd(j)) do if (SOC(j)≤ SOC min)

Pgrid(j) = -Pres(j) else Pgen(j) = -Pres(j) end SOC(j+1) = SOC(j) elseif ((SOC(j)> SOC min) && (EP(j)>

price set)) 40:

7:

SOC min)

Pdis(j) = min(Pdis max,-Pres(j),SOC(j)-

if Pdis(j) == Pdis max

—-Computing Pschd vector for DR-based load

41:

scheduling—–

42:

8: Compute Pschd using the method given in algorithm

pdis max

1, lines 8-18

43:

—–Computing dispatch for the PV system, SB,

44:

Pload d(j)

Pload d(j)

(SOC(j)-SOC min)

9:

45:

end

46:

if Pgds(j) == 0

10:

Pres(j) = Ppv(j)-Pschd(j)

=

Pschd(j)-Ppv(j)-

elseif Pdis(j) == (SOC(j)-SOC min)

grid and DG—– for j = 1:N

=

——-Dispatch when PV energy > Pschd——-

47:

Pgen = Pload d(j)

11:

case (Ppv(j)> Pschd(j)) do

48:

Pload d(j) = 0

12:

if SOC(j)≥ SOC max

49:

Pschd(j)-Ppv(j)-

end

13:

if Pgds(j) == 0

50:

14:

Pdl = Pres(j)

51:

end

52:

endcase

53:

Pgrid(j) = Pgrid(j)+Pload d(j)

15:

else

16: 17: 18: 19: 20:

Psold(j) = Pres(j)

SOC(j+1) = SOC(j)-Pdis(j)

end

54:

SOC(j+1) = SOC(j)

given in algorithm 2, lines 37-39 55:

else Pch(j) = min(Pch max,Pres(j),SOC max-

SOC(j))

((SOC(j)>

if Pgds(j) == 1

using GA operations end

|

Compute IBR-based tariffs using the method

end

—–Computing fitness functions for CE, TBD and Pgsize—–

21:

if Pch(j)6= Pres(j)

56:

Compute CE = sum(EP × Pgrid)

22:

if Pgds(j) == 0

57:

Compute TBD(M) using the method given in al-

23:

Pdl = Pres(j)-Pchg(j)

24:

else

25:

Psold(j) = Pres(j)-Pchg(j)

26:

end

gorithm 1, lines 25-32 58:

Compute Pgsize = Peak(Pgen)

59: end 60: Exit MOGA; return results as Pgsize/CE/TBD

27:

end

TOs and corresponding Ts

28:

SOC(j+1) = SOC(j)+0.8* Pch(j)

61: Selection of a feasible tradeoff solution for Pgsize

29:

end

30:

endcase

as per Table 5.2 and Fig. 5.20

145


Chapter 4

4.4. DRSREODLDG-BASED HEMS

shown in line 25. However, during LS hours, the excess energy that would be sold to the grid is instead supplied to a dummy load. In the second case, in which P pv is less than or equal to P schd, as shown in line 31, the PV energy is insufficient to completely supply the load. The residual energy in this case will be supplied from the grid if the SOC is less than or equal to its minimum limit or from the SB otherwise. Moreover, the SB still will not be discharged if cheap energy is available from the grid, as shown in line 32. However, during LS hours, the DG will supply the load in place of the grid, as shown in line 36. The discharging state of the SB depends on the condition given in line 40. If the minimum computed value is equal to the maximum discharge rate or to the residual capacity of the SB before discharging to the minimum SOC, then one of these constraints is restricting the ability to supply the full load from the SB, and the remaining load must be supplied from the grid, as shown in lines 42 and 44. However, during the LS hour, the DG will supply the load in place of the grid, as shown in line 47. For each slot in the scheduling horizon, one of the above two cases will hold, and the vectors P grid and P gen will be computed for dispatch accordingly. The CE is computed by applying IBR, as shown in line 56. The T BD(M ) values for the MS are computed using algorithm 1, lines 25-32, for each MOGA iteration. Tradeoff solutions between P gsize, CE and T BD are obtained in the form of a Pareto-optimal set using Table 5.2 and Fig. 5.20.

4.4

DRSREODLDG-based HEMS

The architecture for DRSREODLDG-based HEMS is similar as shown in Fig. 1.1. A three-step simulation based posteriori method is proposed to provide tradeoff solutions for an eco-efficient operation of DRSREODLDG-based HEMS. The method makes use of algorithm 4 and algorithm 5 to harness eco-efficient schemes for HEMS operation in terms of T st and the related TOs for CEnet, T BD, and minimal T EM iss. At step-1, primary TOs solutions for CEnet, T BD, and T EM iss 146


Chapter 4


are generated making use of algorithm 4. algorithm 4 is based on a MOGA/ PO based heuristic proposed in this work. At steps-2, the primary tradeoff solutions are passed through an AVCF to filter out the TOs with extremely high and above average values of T EM iss. The filtrate is then passed through an ASCF to screen out the TOs with even the marginally higher values of T EM iss at step-3. The proposed filtration mechanism comprising AVCF and ASCF is detailed in algorithm 5. The formulation for ASCF is elaborated in subsections given below.

4.4.1

Constrained filtration of tradeoffs to HEMS optimization

The tradeoff solutions achieved for a multi-objective optimization problem can be passed through an adequately designed filtration mechanism in order to apply a constraint on any one or more of the specified tradeoffs. Such filtration mechanism enables harnessing the tradeoffs with enhanced efficiency. In this research, a filtration mechanism has been proposed to screen out the tradeoffs with larger values of T EM iss as related to the tradeoffs for CEnet and T BD. This mechanism comprises of an AVCF and an ASCF. AVCF makes use of an average value of the tradeoff parameter T EM iss to filter out the tradeoffs with above average and extremely high values of T EM iss. While, ASCF takes into account an average surface fit for T EM iss in terms of CEnet and T BD to screen out the tradeoffs with higher values of T EM iss. ASCF has been developed using polynomial based regression technique elaborated in sub-sections 4.4.2.

4.4.2

Regression based surface fitting techniques to develop ASCF

Regression models are used to establish a relation between the dependent and the independent variables in a set of data (zi , xi , yi ). In order to develop a surface for the variable z in terms of variables x and y, a regression model is fit to the set 147


Chapter 4


of input data. Major models for surface fitting include interpolant, lowess, and polynomial. A polynomial surface fit model has been used in this research for its flexibility in application to the input data. For polynomial surfaces, general model is designated as Poly(kl), where k is the degree in x and l is the degree in y. The degree of x in each term will be less than or equal to k, and the degree of y in each term will be less than or equal to l. The maximum for the sum of k and l is m. The degree of the polynomial is the maximum of the values of k and l. A linear regression model (LRM) for i number of observations for the independent variables (xi , yi ) defines a curved surface for the dependent variable zi in a 3D-space. The LRM for the surface in terms of an order-m polynomial may be represented as:

zi =

k X l X

A(k, l) × xk × y l + ui

(4.23)

f =1 g=1

where 0 ≤ k + 1 ≤ m. zi is called dependent variable or regressand, and xi and yi are called independent variable or regressors. The first term in Eq. 4.23 is deterministic and represents the conditional mean of zi based on the given values of xi and yi . The second term ui , called the error term is random in nature. The term is added or subtracted from the first term to realize the actual data. A(k, l) are called regression coefficients (RCs). In LRMs, they are assumed to be fixed numbers. The term linear in LRM refers to the linearity of the RCs. The fitting of a model is based on the estimates for the RCs. The estimation is carried out based on the minimization of the least squares of the error term called least square method. Based on Eq. 4.23, ui is the difference between the actual value of zi and the one obtained through the regression. For an optimal linear coefficient (LC) for surface fitting, the sum of the squared error term (SSE) to be minimized is given as follows:

SSE =

i X e=1

2

ui = (zi −

k X l X

2 k

l

A(k, l) × x × y )

(4.24)

f =1 g=1

148


Chapter 4


As ui is a function of RCs, the minimum value of the SSE may be computed by taking partial derivatives of the same with respect to each of the RCs and equating the expression to zero. Based on the estimated a(k,l), a sample model zis is formulated and the error term is also known as residuals which is computed as ei = zi − zis . The regression coefficients a(k, l) are the estimators of A(k, l) and ei is the estimator of the error term ui . The numerical values taken by an estimator are called estimate. The least-squares solution to the problem is a vector a(k, l), which estimates the unknown vector of coefficients A(k, l). In present research, SSE has been used to estimate the model fit for an average surface for T EM iss in terms of CEnet and T BD. It is assumed that the observed data is of equal quality and thus has constant variance; however, the fit might be unduly influenced by the data of poor quality. Methods like weighted-least-squares regression are applied to reduce the influence of the low quality data on estimating the model fits [205], [206]. In present research, use of AVCF to screen out the tradeoffs with extremely high values of T EM iss, inherently improves the data quality for the model fit for ASCF. The model fit in this research is based on minimization of SSE that may be improved using methods like minimization of root mean square error and root mean square error of approximation. However, the improved well-fitting is of minimal value, if it is not based on the ideas from a theory validation point of view and in such cases an extensive cross-validation is required [201]. Accordingly, various polynomial models fit, 25 in number have been examined for their capabilities to reduce the average value of tradeoffs parameters for T EM iss and the number of diverse tradeoffs available for CEnet and T BD after the application of filtration mechanism.

149


Chapter 4

4.5

4.5. ALGORITHMS FOR ECO-EFFICIENT TOS FOR DRSREODLDG-BASED HEMS

Algorithms for eco-efficient TOs for DRSREODLDGbased HEMS

A three-steps approach has been used to achieve eco-efficient tradeoff solutions for DRSREODLDG-based HEMS. In step-1, schemes for optimal HEMS operation and the related TOs for T EM iss, CEnet, and T BD are computed using algorithm 4. The tradeoff solutions thus obtained are passed through a filtration mechanism to harness the ones with bare minimum T EM iss in terms of CEnet and T BD using algorithm 5. The filtration mechanism is completed in two stages designated as step-2 and step-3. In step-2, an AVCF for T EM iss is developed and applied to the primary TOs to filter out the ones with extremely high and above average values of T EM iss. The remaining TOs are then passed to step-3 to screen out the TOs with even the marginally higher values of T EM iss. In step-3, an ASCF is used to filter out the TOs with T EM iss parameters residing above the average surface fit for T EM iss. Eco-efficient solutions including bare minimum T EM iss and diversified TOs for CEnet and T BD are thus achieved for DRSREODLDGbased HEMS operation. The followings algorithms have been proposed to harness the eco-efficient tradeoff solutions for DRSREODLDG-based HEMS: • algorithm 4 to generate primary TOs for DRSREODLDG-based HEMS (Step-1). • algorithm 5 for filtration mechanism to harness eco-efficient TOs for DRSREODLDGbased HEMS (Step-2 and step-3). The algorithms are presented in the following subsections.

150


Chapter 4


Algorithm 4 Algorithm to generate operating schemes and the primary TOs for DRSREODLDG-based HEMS (Step-1) Input: P E, P Ef , P Eg, IBR, T P , EF T , P a, Styp, ST slot, EN slot, LoT , P load nsh, P gds, SoC(init), SoC mx, SoC mn, P ch mx, P ds mx, P pv Output: Optimal TOs for T EM iss, CEnet and T BD with T st for SHAs 1: Initialize input parameters 2: Call MOGA 3: Initialize Tst within bounds STslot and ENslotLoT+1 4: for Iterat = 1: Ng mx 5: if Iterat > 1 6: Generate new Tst populations within bounds using GA operations 7: end —-Computing P schd vector for DR-based scheduling—– 8: Tend = Tst+LoT-1 9: for i = 1:k 10: for j = 1:N 11: if (j ≥ T st(i) &&j ≤ T end(i)) 12: Power matrix(i,j) = Pa(i) 13: else 14: Power matrix(i,j) = 0 15: end 16: end 17: end 18: Pschd = sum(Power matrix)+ Pload nsh —–Computing dispatch for the PV system, SB, grid and the LDG—– 19: for j = 1:N 20: Pres(j) = Ppv(j)-Pschd(j) ——-Dispatch when PV energy > Pschd——21: case (Ppv(j)> Pschd(j)) do 22: if SoC(j)≥ SoC mx 23: if Pgds(j) == 0 24: Pdl = Pres(j) 25: else 26: Psold(j) = Pres(j) 27: end 28: SoC(j+1) = SoC(j) 29: else 30: Pch(j) = min(Pch mx,Pres(j),SoC mxSoC(j)) 31: if Pch(j)6= Pres(j) 32: if Pgds(j) == 0 33: Pdl = Pres(j)-Pch(j) 34: else 35: Psold(j) = Pres(j)-Pch(j) 36: end 37: end 38: SoC(j+1) = SoC(j)+0.8* Pch(j) 39: end 40: endcase ——-Dispatch when PV energy ≤ Pschd——41: case (Ppv(j)≤ Pschd(j)) do 42: if (SoC(j)≤ SoC mn) | ((SoC(j)> SoC mn)

&& (PE(j)≤ price set)) 43: if Pgds(j) == 1 44: Pgd(j) = -Pres(j) 45: else 46: Pgn(j) = -Pres(j) 47: end 48: SoC(j+1) = SoC(j) 49: elseif ((SoC(j)> SoC mn) && (PE(j)> price set)) 50: Pds(j) = min(Pds mx,-Pres(j),SoC(j)SoC mn) 51: if Pds(j) == Pds mx 52: Pload d(j) = Pschd(j)-Ppv(j)pds mx 53: elseif Pds(j) == (SoC(j)-SoC mn) 54: Pload d(j) = Pschd(j)-Ppv(j)(SoC(j)-SoC mn) 55: end 56: if Pgds(j) == 0 57: Pgn = Pload d(j) 58: Pload d(j) = 0 59: end 60: SoC(j+1) = SoC(j)-Pds(j) 61: end 62: endcase 63: Pgd(j) = Pgd(j)+Pload d(j) ——-Computing tariffs with IBR——— 64: if Pgd(j)> TP 65: PE(j) = IBR × PE(j) 66: end 67: end —–Computing fitness 68: T EM iss = EFT × —–Computing fitness 69: CEnet = sum(PE sum(PEf × Psold)

function forT EM iss—– sum(Pgn) functions for CEnet—– × Pgd)+sum(PEg × Pgn)-

—–Computing fitness function for T BD—– 70: for a = 1:k 71: if Styp = DS 72: T BD(D)(a) = (Tst(a)-STslot(a))/(ENslot(a)LoT(a)-STslot(a)+1) 73: else 74: T BD(A)(a) = (ENslot(a)-Tst(a)LoT(a)+1)/(ENslot(a)-LoT(a)-STslot(a)+1) 75: end 76: end 77: Compute T BD = (sum(TBD(D))+sum(TBD(A)))/k 78: end 79: Exit MOGA; return results as T EM iss, CEnet and T BD TOs and corresponding T st 80: Goto algorithm 5 to harness eco-efficient tradeoff solutions

151


Chapter 4


Algorithm 5 Algorithm for filtration mechanism to harness eco-efficient TOs for DRSREODLDG-based HEMS (Step-2 and step-3) Input: TOs from algorithm 4 for CEnet, T BD and T EM iss Output: Eco-efficient tradeoff solutions for CEnet, T BD and minimal T EM iss —–Step-2: Filtration of tradeoff solutions using AVCF for T EM iss—– 1: Do —–Computing average value of T EM iss—– 2: TEMiss avg = Average (TEMiss) —–Computing residuals for T EM iss w.r.t T EM iss avg —– 3: TEMiss Resid avg = Average (TEMiss)- TEMiss —–Filtration based on average value of T EM iss—– 4: Filter out/ Exclude solutions with negative TEMiss Resid avg 5: Collect the remaining solutions for refined filtration in step-3 6: End do

4.5.1

—–Step-3: Refined filtration of TOs based on average surface of T EM iss—– 7: 8:

Do Tabulate CEnet, TBD and TEMiss

—–Computing average surface for T EM iss—– 9: Generate an average surface using polynomial option (Ploy41) —–Computing residuals for T EM iss w.r.t average polynomial surface—– 10: TEMiss Resid avgs = TEMiss on surface Actual value of TEMiss —–Filtration based T EM iss —– 11: Filter out Miss Resid avgs 12:

on TOs

average with

surface negative

of TE-

Collect the remaining TOs as eco-efficient

tradeoff solutions 13:

Enddo

Algorithm 4 to generate operating schemes and the primary TOs for DRSREODLDG-based HEMS (Step1)

This algorithm computes a set of primary tradeoff solutions for optimized HEMS operation based on MS of SHAs synergized with the optimal dispatch of the PV system, the SB, the grid, and an LDG. The LDG supplies the load only during LSD hours in coordination with the PV unit and the SB. TOs for CEnet, T BD, and T EM iss are based on the underlying scheme for HEMS operation. At the start, vector T st for SHAs is generated that is followed by the production of P schd vector. The PV system is regarded as the preferred source to directly supply P schd. The dispatch planning is mainly based on the excess PV energy in each slot denoted by P res which is the arithmetic difference between P pv and P schd. Two main cases arise with regard to the relative values of these two quantities and in each case, SoC, the maximum charge/discharge rates, the grid status and 152


Chapter 4


the power from the LDG play major roles in the dispatch. In the first case, where excess PV energy is available, as shown on line 21, the energy is stored in the SB if SoC is less than its maximum value; otherwise, it is sold to the grid. However, during LSD hours, the excess energy that would be sold to the grid is instead supplied to a dummy load as shown on line 24. The SB charging state depends on the condition given on line 30. If a value other than P res is computed, it indicates that either the maximum charge rate or the limiting value of SoC is restricting the complete storage of the excess PV energy in the SB. Hence, any excess energy left after charging the SB is sold to the grid, as shown on line 35. However, during LSD hours, the excess energy that would have been sold to the grid is instead supplied to a dummy load, as shown on line-33. In the second case, in which P pv is less than or equal to P schd, as shown on line 41, the PV energy is insufficient to completely supply the load. The residual energy, in this case, will be supplied from the grid if SoC is less than or equal to its minimum limit or from the SB otherwise. Moreover, the SB will still also not be discharged if cheap energy is available from the grid as given on line 42. However, during LSD hours, the LDG will supply the load in place of the grid, as shown on line 46. SB shall supply the load only during peak hours when the cost of energy is greater than a maximum price limit. The discharging state of the SB depends on the condition given on line 50. If the minimum computed value is equal to the maximum discharge rate or to the residual capacity of the SB before discharging to the minimum SoC, then one of these constraints is restricting the ability to supply the full load from the SB, and the remaining load must be supplied from the grid, as shown on lines 52 and 54. However, during LSD hour, the LDG will supply the remaining load in place of the grid as shown on line 57. For each slot in the scheduling horizon, one of the above two cases will hold, the vectors P pv, P gd, P ds, and P gn will be computed for dispatch accordingly. Similarly, the loads for each slot is computed for P schd, P ch, P dl, and P sold. T EM iss is computed (applying EF T ) for the net generation from LDG as shown on line 68. CEnet is 153


Chapter 4


computed by arithmetically adding CE (applying P E/IBR), cost of generation from LDG (applying P Eg) and cost of energy sold to the grid (applying P Ef ) as shown on line 69. The T BD values are computed on line 77 after adding T BD(A) and T BD(D) on lines 72 and 74. The values for the mentioned objective functions are computed for each MOGA iteration. The algorithm provides Pareto optimal sets comprising one hundred operating schemes for HEMS in terms of T st and the related TOs for CEnet, T BD and T EM iss.

4.5.2

Algorithm 5 for filtration mechanism to harness ecoefficient TOs for DRSREODLDG-based HEMS (Step2 and Step-3)

The algorithm completes the filtration process in two steps as stated below: Step-2: An AVCF based on the average value of T EM iss is developed taking into account all of the primary TOs generated through algorithm 4 as shown on line-2. The residuals for T EM iss (T EM iss Resid avg) for each solution are then computed as given on line-3. A tradeoff solution with the value of T EM iss Resid avg less than 0 indicates an above average value for T EM iss, and all such TOs are filtered out at the step shown on line-4. The tradeoff solutions with average (or less than average) T EM iss values are collected and forwarded to step-3 for further processing as shown on line-5. Step-3: An ASCF based on the average surface fit (using polynomial-based regression) is developed making use of the tradeoff solutions forwarded from step-2 as shown on line-9. The residuals for T EM iss (T EM iss Resid avgs) for each solution are then computed by taking the difference between the T EM iss and the average surface fit of T EM iss computed in terms of CEnet and T BD as shown on line-10. A tradeoff solution with the value of T EM iss Resid avgs less than 0 indicates the T EM iss value greater than the respective value on the average surface fit, and all such TOs are filtered out at step shown on line-11. The remaining 154


Chapter 4

4.6. DWS-PSO- BASED SYSTEM MODEL

tradeoff solutions with the T EM iss values equal to (or less than) the respective values on the average surface fit are selected and declared final eco-efficient TOs for DRSREODLDG-based HEMS operation as shown on line-12.

4.6

DWS-PSO- based system model

The components of HEMS are similar as already stated in Fig. 1.1, including HAs, RESs, an ESS, a HEMS controller, a local communication network, a smart meter for the exchange of information between the consumer and the utility for energy pricing, and the consumer’s electricity profile. In PDDR- based HEMS, the operations of HAs are shifted in time for a specified DPS in order to minimize the peak demand for the utility and the cost of energy for the consumer. Whereas, in PDDR-RED- based HEMS, the shifted operations of SHAs are integrated with the dispatch scheme for renewable energy sources, ESSs, and the grid in order to reduce peak as well as overall demand and CEnet for the utility and the consumer, respectively. While minimizing the CE/ CEnet for the consumers, TBD should remain within acceptable limits. HEMS operation has been modeled for a time horizon of 24 hours length. HAs are supposed to be operated for specified lengths of operational time (LOOT) vector within the proposed time intervals (as per the vectors STslt and ENslt) as given in section 4.7. The dispatch model for the power sources is based on the operation of the PV unit, the SB system, and the power grid in parallel based on the availability of the PV energy, and vector of states of charge (SOCG) and the limiting values of charge/ discharge rates for the SB. The PV units are modeled as the preferred source of energy to supply the scheduled loads. The un-used PV energy is preferably stored in the SB that is used to supply the load during high pricing time to reduce CEnet. The excess energy, in case of a fully charged storage unit, is transmitted for monetary benefits. The strategy for scheduling of HAs and the dispatch protocols for the local resources is taken from [7] where a MOGA- based algorithm was used to implement the strategy. This 155


Chapter 4


research makes use of DWS-PSO- based algorithms to implement the strategies for optimal operations of PDDR- and PDDR-RED- based HEMS. The proposed algorithms, as detailed in section V, are designed to handle MO and are based on the TO solutions. A procedure for DPA based on the construction of a diversified set of TPs and the performance metrics has also been proposed and the developed algorithms have been validated for DPA accordingly.

4.6.1

HAs

The HAs are classified into NSHAs and SHAs [7],[80]. The NSHAs comprising of electric lamps, fans, etc., are assumed to work as and when required and they cannot opt for scheduling. The forecasted load for NSHAs used in the simulation section is presented in Fig. 4.5. The SHAs are supposed to be scheduled towards the off-peak hours and the PV energy harnessing hours for optimized HEMS operation. In order to maximize the reduction in the cost of energy, SHAs are modeled as AS and DS. MS (a combination of AS and DS)- based model for SHAs enables more reduction in the cost of energy making use of an enhanced flexibility in the price-driven shifting of HAs and an increased direct usage of the PV energy from the PV unit. Technical specifications along with the consumer’s defined settings for the accomplishment of the required operations of AS and DS type SHAs used in the simulation section are described in Table 4.6 [7].

4.6.2

DPSs

Dynamic prices are the key to implement PDDR as well as PDDR-RED- based HEMSs. They are introduced to motivate consumers to modify their energy consumption profiles. Such modifications in the consumption profiles enable reducing the peak demand, the overall demand as well as the greenhouse gas (GHG)- emissions for the utility. The types of DPSs include DA-RTP, ToUP, and CPP that are discussed next. 156


Chapter 4


0.5 0.45

Fixed load (kWh/ slot)

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 4.5: Fixed load profile for a smart home Table 4.6: SHAs and scheduling specifications for the DS and MS scenarios

SHA

Power (kWh)

LOOT (slots)

AC 1 (Reversible) AC 2 (Reversible) AC 3 (Reversible) AC 4 (Reversible) Dishwasher (DW) 1 Dishwasher (DW) 2 Electric Geyser (EG)1 Rice Cooker (RC) 1 (Manual) Computer (CP) (Manual) Washing Machine (WM) Water Pump (WP) Electric Geyser (EG) 2 Rice Cooker (RC) 2 (Manual) Iron (IR) (Manual)

1 1 1 1 0.6 0.6 0.8

4.6.2.1

18 9 9 12 3 3 6

MS (Start/End Limits) 01-36 (DS) 37-54 (DS) 103-120 (DS) 121-144 (DS) 49-102 (DS) 127-144 (DS) 01-36 (DS)

DS (Start/End Limits) 01-36 (DS) 37-54 (DS) 103-120 (DS) 121-144 (DS) 49-102 (DS) 127-144 (DS) 01-36 (DS)

0.4

3

73-81 (DS)

73-81 (DS)

0.1

6

114-144 (DS)

114-144 (DS)

0.7

9

93-123 (AS)

114-144 (DS)

0.7 0.8

3 6

37-117 (AS) 55-121 (AS)

114-144 (DS) 115-126 (DS)

0.4

3

100-117 (AS)

114-120 (DS)

0.6

3

55-117 (AS)

114-144 (DS)

ToUP

The TOUP schemes are based on predefined price values. The price pattern is maintained typically for a period of 3 157 to 6 months. PhD Newthesis pricesby: areBilal proposed on Hussain

Chapter 4


the yearly based on the operational cost and long-term investments of the utilities.

Tariff (Cents/kWh)

15

10

5

0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 4.6: 2S-ToUP scheme

Tariff (Cents/kWh)

15

10

5

0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 4.7: 3S-ToUP scheme

ToUP schemes are offered for a specified period of time and are based on different electricity rates for peak, mid-peak and off-peak times. The 2S-ToUP and 3SToUP schemes are adopted by the national transmission and distribution company in Pakistan and Baltimore gas and electric company, USA has been shown in Fig. 4.6 and Fig. 4.7 [78]. The algorithms for HEMS are designed for shifting of SHAs from the on-peak hours to the off-peak hours based the price signal received from the utility.

158


Chapter 4 4.6.2.2

4.6. DWS-PSO- BASED SYSTEM MODEL DA-RTP

In DA-RTP scheme, the electricity price varies on an hourly basis. In such schemes, the price signal is typically communicated on 24 hours ahead basis. This helps and motivates the consumer to participate in the PDDR programme. RTP enables utility companies to better distribute the price of electricity reflecting the demandsupply elastics. The nature of the scheme characterized by a diversified price elasticity is ought to motivate the consumer to adjust their demand more precisely [75]. The DA-RTP scheme adopted from Ameron Illinois Power Company (AIPC), USA, is shown in Fig. 4.8 [78]. 10 9

Tariff (Cents/kWh)

8 7 6 5 4 3 2 1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 4.8: DA-RTP scheme

4.6.2.3

CPP

The CPP is an event-based scheme. To manage the energy in an event, a control signal for extra high electricity prices is communicated to affect electricity demand during critical peak hours. The scheme can also be imposed if the system is expected to be severely constrained due to the extremely cold/ warm period for a limited number of hours. The consumers can participate in CPP based PDDR programme for very high incentives for reducing the CEnet by either reducing

159


Chapter 4


their peak demands or shifting energy consumption towards off-peak times. CPPbased PDDR program is event based and not a daily based DR program. Further, CPP values are always higher than the corresponding TOUP values. A CPP scheme based on the proposed scheme for 2S-ToUP with the critical peak price as double of the daily peak time price has been shown in Fig. 4.9 [75]. The aforementioned scheme for CPP, based on the criteria adopted by San Diego gas and electric company which has been used in the simulation section. 30

Tariff (Cents/kWh)

25

20

15

10

5

0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 4.9: CPP scheme (2-stage)

4.6.2.4

IBR

At high levels of demands, the aforementioned standardized DPSs are introduced with higher rates of electricity, called IBR. Such schemes are introduced in order to discourage the consumers from over-shifting of the loads towards the off-peak hours. The scheme enables avoiding the re-emergence of peaks that may appear as a result of PDDR- based scheduling. A DPS with the related IBR vector is computed as follows:

DEP = [(DEP 1, P T 1, IBR1), ..., (DEP n, P T n, IBRn)]

160

(4.25)


Chapter 4


where DEP 1, DEP 2 to DEP n are the normal energy prices during the nth time slots. Further DEP n×IBRn pricing schemes are applied with the power threshold levels of P T 1, P T 2 to P T n in the respective time slots. An IBR value of 1.4 as opted by British Columbia Hydro has been adopted in the proposed models for DPSs. IBR factor beyond a power threshold of 2.4 kW (0.4 kWh/ slot) has been applied while simulating the model for various DPSs [7].

4.6.3

RESs

Solar PV units are one of the most widely used types of RESs at homes. The proposed model for HEMS operation is included with the PV unit based on the forecasted value of solar irradiations. The data for solar irradiations measured by the Pakistan Engineering Council in Islamabad is applied for the simulated operations of PDDR-RED- based HEMS. The profile of electrical energy harnessed from the PV unit is displayed in Fig. 4.10. The cost of electricity generation from the local PV unit has not been included in the model. 1 0.9 0.8

Ppv (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time Slots

Figure 4.10: PV energy profile

161


Chapter 4

4.6.4


SBs

SBs are introduced into the PV- based HEMSs in order to introduce flexibility in the dispatch of the PV energy. The surplus PV energy is stored in the SB that can be used to supply the load during peak hours to minimize the CE. Technical specifications of the PV unit, the SB and the inverter that were used in the simulations are given in Table 4.7. Table 4.7: PV unit, SB and inverter specifications [7]

Parameter Total capacity of the PV unit Rating of each panel Number of panels and panel area Efficiency of PV panels Inverter rating Inverter efficiency SB Ah SB voltage SB capacity SB charge rate SB discharge rate Minimum SOCG Maximum SOCG SB efficiency

4.7

Value 5 kW 250 W 20, 32 m2 15% 5 kW 70% 600 Ah 48 V 4.8 kWh/slot 0.48 kWh/slot 0.32 kWh/slot 30% 95% 80%

Formulating the HEMS optimization problem

The energy management problem is formulated for specified objective under a set of constraints. The problem is solved for a set of input variables that can provide 162


Chapter 4


the optimal values of the desired objectives. The formulation is founded on the scheduled load profile based on the following control parameters:

B = [b1 , b2 , .., bk ]

T = [1, 2, 3, .., N ]

P app = [P1 , P2 , .., Pk ]

LOOT = [LOOT1 , LOOT2 , .., LOOTk ]

ST slt = [ST slt1 , ST slt2 , .., ST sltk ]

EN slt = [EN slt1 , EN slt2 , .., EN sltk ]

DEP = [DEP1 , DEP2 , .., DEPN ]

IBR = [IBR1 , IBR2 , .., IBRN ]

T st = [T st1 , T st2 , .., T stk ]

The formulation for the scheduled load is based on the input vector for T st. In the proposed algorithms, the specified vector is generated through PSO. A decision vector Pa (dim: 1 × N ), based on the generated T st, is derived for the scheduled power profile of the bth shiftable home appliance, as follows:

P b (i) =

   P app(b) : f or T st(b) + LOOT (b) > i ≥ T st(b)   0

: f or T st(b) > i ≥ T st(b) + LOOT (b)

163


Chapter 4


Similarly, vectors P1 , .., Pk are developed for each of the SHA, based on the respective value of T st(b) for the SHA, numbered as B = 1, 2, .., k.

The decision vectors P1 , .., Pk are joined into a matrix called as P ower matrix as given below:

P ower matrix = [P1 , P2 , ..., Pk ]t

(4.26)

P ower matrix is summed up column-wise to develop a scheduling vector P sch sh. The developed vector specifies the power requirement for each of the slot in the scheduling horizon. The power scheduling vector for HEMS problem is thus formulated as follows: P sch sh =

N X k X

P (b, n)

(4.27)

n=1 b=1

where P (b, n) is a generalized element of the derived P ower matrix. The final scheduled load vector, P sch is then developed by adding the load vector for NSHAs, P load nsh, to the vector P sch sh as follows:

P sch = P sch sh + P load nsh

4.7.1

(4.28)

Objectives for optimal HEMS operation

The major objectives for HEMS include minimizing the CE to be supplied from the grid, minimizing the discomfort borne by the consumer, and reducing the peak load. These objectives are discussed next.

4.7.1.1

Minimization of CE

The OF for minimizing the CE for a PDDR- based HEMS is formulated as follows:

M inimize

N X

(P sch × DEP )

(4.29)

n=1

164


Chapter 4


where DEP and P sch are the pricing vector and the load scheduling computed through Eq. 4.25 and Eq. 4.28, respectively. For DR-RED- based HEMSs Pgd is computed using algorithm 2 as follows:

P gd = P sch − P pv + P sold + P ch − P ds

(4.30)

In this case, the OF for minimizing the CEnet is formulated as follows:

M inimize

N X

(P gd × DEP − P sold × DEP f )

(4.31)

n=1

where P gd and P sold are the vectors denoting the purchased/ sold energies from/ to the utility, and DEP and EP f are the respective vectors for the pricing. We have assumed a value of EP f equal to 0.7 × DEP in the simulations section.

4.7.1.2

Minimization of T BD

For DS-type SHAs, the average T BD due to the delayed starting of SHAs, denoted by T BD(D), is formulated as follows:

T BD(D) =

k1 X

((T st − ST slt)/ (EN slt − LOOT − ST slt + 1))/k1

(4.32)

b=1

where k1 is the number of SHAs chosen for DS. Further, T BD(D) assumed a minimum value of 0 when the SHAs started their operations at ST slt. The parameter achieved a maximum value of 1 when the SHAs started their operations at EN slt(b) − LOOT (b)+1. These limits must be followed while computing the vector T st and are formulated below:

Lb = ST slt and U b = EN slt − LOOT + 1

165

(4.33)


Chapter 4


For AS-type SHAs, the average T BD due to the advanced completion of the jobs, designated as T BD(A), is computed as follows:

T BD(A) =

k2 X

((EN slt − T st − LOOT + 1)/(EN slt − LOOT − ST slt + 1))/k2

b=1

(4.34) where k2 is the number of SHAs chosen for AS. The T BD(A) assumes a minimum value of 0 when all of the SHAs complete their jobs at their proposed job ending time in EN slt(b), and will achieve a maximum value of 1 when T st(b) equals ST slt(b). In MS, some of the SHAs are selected for AS, whereas the others for DS. In this mode, the average value of T BD for a total number of k SHAs, designated as T BD(M ), is expressed as follows [7]:

T BD(M ) = T BD(D) + T BD(A)

4.7.1.3

(4.35)

Minimization of Ppeak

The objective to minimize the peak load fed from the grid is computed as follows:

M inimize P eak(P gd)

4.7.2

(4.36)

Constraints

The constraints were applied for HAs, pricing schemes, SB, and the energy balance as per our previous research designed for MOGA [7].

4.7.3

Meta-heuristic techniques to solve energy management problems

Over the past few years, meta-heuristics tools have been used very successfully for obtaining robust solutions to complex HEMS optimization problems. The renowned meta-heuristics include GA, PSO, ACO, and evolutionary programming 166


Chapter 4


[7]. In this study, we have introduced PSO to reach the optimal solutions to PDDRand PDDR-RED- based HEMSs. The technique uses swarm behavior of the birds known as swarm intelligence. PSO is simple, faster in convergence and has an ability to quickly search in extra large search spaces. The technique is capable of solving a diversified set of complex optimization problems very quickly. At the start, the algorithm generates a population of solutions called as particles that move towards the best position in the search space with random velocities. Each particle remembers his own best and global (swarm’s) best positions and moves with j th particle velocity (in ith iteration) formulated as follows:

Vj (i) = Vj (i − 1) + clr × ran1 × [P best − Xj (i − 1)] + slr × ran2 × [Gbest − Xj (i − 1)] (4.37) where j = 1, 2, . . . ,N clr= cognitive learning rate slr= social learning rate ran1 and ran2 are uniformly distributed random numbers

New position of j th particle is found as:

Xj (i) = Xj (i − 1) + Vj (i)

(4.38)

where j = 1, 2, . . . ,N

Each of the particle corresponds to the values of decision variables that are related to the values of FFs as F1 (X1 (i)), F2 (X2 (i)), ...FN (XN (i)). The method is said to be converged if positions of all particles converge to the same set of values, 167


Chapter 4


otherwise more iterations are carried out [123].

4.7.4

Handling of multi-objectivity in energy managment problems

Most of the problems for HEMS in real life are MOO with mutually conflicting objectives. The main objective in the recent research is minimizing the CE [77][82], [67],[86], whereas, minimizing T BD is the second most important objective [7], [53], [57], [67], [83], [84], and [86] from the consumer’s perspective. A TO exists between the reduction in CE and the T BD that makes home energy management problems interesting to the issue of MO. The following methods, in recent research, have been used to handle MO: PO [7], [85]-[86] , E-constraint [15], [55], and WSM [14], [53], [81], [57], [83] and [137]. Algorithms for MOO based on PO, like NSGA, have been used to compute a Pareto-optimal set (POS) containing non-dominated TO solutions for the specified objectives. The method provides a diverse set of TOs between the objectives for CE and TBD that helps consumer making decisions as per their needs. In recent research, PO has been the one and only method used for TO solutions as posteriori. E-constraint method was introduced by Haimes et. al. in 1971. In this method, just one OF is kept while the rest of the OFs are transformed into constraints within the user-specified values. The vector for the constrained values has to be chosen very carefully so that it remains within minimum/ maximum limits of the constrained OFs. The method was used by Khan et al. to compute the value of the discomfort using the value of the consumer budget as constraint [15]. Algorithms based on WSM solve MOO problems while transforming all of the OFs into a SCF using respective weights for individual OFs. The said SCF for WSM is formulated as follows [200]:

M inimizeF (X) =

M X

W mF m(X)

(4.39)

m=1

168


4.8. ALGORITHMS FOR A PDDR- AND PDDR-RED- BASED HEMSS Chapter 4 USING DWS-PSO where

Pk

m=1

W m = 1.

The method is characterized by assigning of fractional weights to the constituent OFs in order to reflect their priorities while computing the SCF. The method has mostly been used as priori. Use of WSM with a posteriori approach, as an alternate to PO, has been focused in this research for its application to HEMS. Moreover, WSM can be combined with population-based meta-heuristic like PSO for posteriori in order to achieve the TO solutions. In decomposition approach for WS-PSO, a set of weighting vectors are deployed, and a MOO is decomposed into a number of SO sub-problems that are optimized simultaneously in a single run. The solution obtained through decomposition WS-PSO based algorithm provides a set of TOs that helps consumers to make decisions after evaluating a diverse set of choices.

4.8

Algorithms for a PDDR- and PDDR-REDbased HEMSs using DWS-PSO

The following algorithms have been proposed: - Algorithm 1 for a PDDR-based HEMS using DWS-PSO - Algorithm 2 for a PDDR-RED- based HEMS using DWS-PSO

The algorithms are presented in the following subsections.

4.8.1

Algorithm 6 for a PDDR-based HEMS using DWSPSO

The algorithm computes a set of solution that provides optimal TOs for CE and T BD for PDDR- based scheduling of SHAs using PSO. For optimal scheduling, T st is heuristically generated based on the specified bounds using PSO. The computations for Psch, CE, and, TBD mentioned at line numbers 10-34, are taken 169


4.8. ALGORITHMS FOR A PDDR- AND PDDR-RED- BASED HEMSS Chapter 4 USING DWS-PSO Algorithm 6 Algorithm for a PDDR-based HEMS using DWS-PSO Input: DEP , IBR, P T , P app, SHtype, ST slt, EN slt,

21:

LOOT , P load nsh, W T V

22:

Output: Optimal TOs for CE and T BD and the re-

23:

lated set of T st

24:


25:

2: M = Count (WTV); counting the required number

—–Computation of the fitness function for CE—

of TO solutions

–

3: For p= 1:M

26:

4:

—–Computation for fitness function for TBD—

5:

Do Initialize Tst within the defined bounds as per

Eq. 4.33 6: 7: 8:

if Psch(j) > PT DEP(j) = IBR × DEP(j) end end

Compute CE = sum(DEP × Psch)

– 27:

for itter = 1:Ng mx

for j = 1: N

for b = 1:k

28:

if itter > 1

if SHtype = DS

29:

Generate new populations for Tst within

the defined bounds using PSO operations

30:

9:

31:

end

T BD(D)(b)=(Tst(b)-STslt(b))/(ENslt(b)-

LOOT(b)-STslt(b)+1) else T BD(A)(b)=(ENslt(b)-Tst(b)-

—–Computation for Psch vector for PDDR-

LOOT(b)+1)/(ENslt(b)-LOOT(b)-STslt(b)+1)

based HEMS—–

32:

10:

33:

end

34:

Compute T BD = (sum(T BD(D))+sum(T BD(A)))/k

11:

Tend = Tst+LOOT-1 for i = 1:k

12:

for j = 1:N

Power matrix(i,j) = 0

17:

20:

35:

end end

= sum (W T V (1) ×

36:

end

37:

end DO

38: end; return of results from DWS-PSO for CE,

end

T BD and T st for M number of TOs

Psch = sum(Power matrix)+ Pload nsh

—–Computation

Compute SCF

CE)+W T V (2) × T BD )

else

16:

19:

(SCF ) to apply WSM———–

Power matrix(i,j) = Papp(i)

14: 15:

18:

—–Computation of single compound function

if (j ≥ T st(i) &&j ≤ T end(i))

13:

end

for

tariffs

combined

39: Selection of a feasible TO solution by the consumer with

IBR—–

from [7]. The value of SCF is then computed for the FFs for CE and T BD and the relative weights to be allocated to each of the FF. This important computation is presented at 35 line number. The TO solutions for CEnet and T BD achieved using DWS-PSO are mentioned at 38 line number.

170


4.8. ALGORITHMS FOR A PDDR- AND PDDR-RED- BASED HEMSS Chapter 4 USING DWS-PSO

4.8.2

Algorithm 7 for a PDDR-RED-based HEMS using DWS-PSO

This algorithm provides a set of solutions for optimal TOs between CEnet and T BD for PDDR-RED- based HEMS operations using DWS-PSO. The algorithm combines the scheduling of SHAs with the dispatch scheme of the PV system, the storage unit, and the power grid. For optimal scheduling, T st is generated using PSO based on the specified bounds. The computations for Psch; dispatch scheme for PV units, SB, and the power grid; DEP combined with IBR; CE, and TBD as given at line numbers 10-44 are taken from [7]. The SCF is computed while making use of the values of FFs for CE, T BD, and the relative weights allocated to each of the FF. This important computation is presented at line number 45. The TO solutions for CEnet and T BD achieved using DWS-PSO are mentioned at line number 48.

171


4.8. ALGORITHMS FOR A PDDR- AND PDDR-RED- BASED HEMSS Chapter 4 USING DWS-PSO

Algorithm 7 Algorithm for a PDDR-RED-based HEMS using DWS-PSO Input: DEP , IBR, P T , P app, SHtype, ST slt, EN slt,

25:

LOOT , P load nsh,

26:

SOCG mx,

SOCG(init),

SOCG mn,

P ch mx,

case (Ppv(j) ≤ Psch(j)) do if

(SOCG(j)

≤

SOCG mn)

((SOCG(j) > SOCG mn) && (DEP(j) ≤ price set))

P ds mx, P pv, W T V

27:

Pgd(j) = -Pres(j)

Output: Optimal TOs for CE and T BD and the re-

28:

SOCG(j+1) = SOCG(j)

lated set of T st

29:


(DEP(j)> price set))

2: M = Count (WTV); counting the required number

30:

of TO solutions

Pres(j),SOCG(j)-SOCG mn)

3: for p= 1:M

31:

4:

32:

do

5:

Initialize Tst within the defined bounds as per

6: 7:

elseif ((SOCG(j)> SOCG mn) &&

Pds(j)

for itter = 1: Ng mx

min(Pds mx,-

Pgd(j)

=

Psch(j)-Ppv(j)-

Pds mx elseif

Pds(j)

==

SOCG(j)-

SOCG mn

if itter > 1

8:

=

if Pds(j) == Pds mx

33:

Eq. 4.33

34:

Generate new populations for Tst within

35:

9:

36:

end

Pgd(j)

=

Psch(j)-Ppv(j)-

(SOCG(j)-SOCG mn)

the defined bounds using PSO operations

end SOCG(j+1) = SOCG(j)-Pds(j)

—-Computation of Psch vector for PDDR-based

37:

load scheduling—–

38:

10:

——-Computation of tariffs with IBR———

Compute Psch using the method given in

algorithm 1, lines 10-20

39:

—–Computation of the dispatch for the PV sys-

40:

tem, SB and grid—–

41:

11:

42:

12:

for j = 1:N

–

13:

case (Ppv(j) > Psch(j)) do

43:

14:

if SOCG(j) ≥ SOCG mx

16: 17: 18:

end end

Compute CEnet = sum (DEP × Pgd -

SOCG(j+1) = SOCG(j)

44:

Compute T BD using the method given in

algorithm 1, lines 27-34

Pch(j) = min(Pch mx,Pres(j),SOCG mx—–Computation of single compound function (SCF ) to apply WSM———– if Pch(j) 6= Pres(j)

45:

Compute SCF

= sum (W T V (1) ×

CEnet)+W T V (2) × T BD )

Psold(j) = Pres(j)-Pch(j) end SOCG(j+1) = SOCG(j)+0.8*

Pch(j)

24:

DEP(j) = IBR × DEP(j)

DEPf× Psold )

else

22:

23:

if Pgd(j)> PT

—–Computation of fitness function for T BD—–

20: 21:

endcase

Psold(j) = Pres(j)

SOCG(j)) 19:

end

—–Computation of fitness function for CEnet—

Pres(j) = Ppv(j)-Psch(j)

——Dispatch when PV energy > Psch——-

15:

|

46:

end

47:

end DO

48: end; return of results from DWS-PSO for CEnet, end endcase

T BD and T st for M number of TOs 49: Selection of a feasible TO solution by the consumer

——-Dispatch when PV energy ≤ Psch——-

172


Chapter 5 Simulation results and discussion

173

5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS

5.1

Simulations for the optimal operation of DRand DRSREOD-based HEMSs

Simulations were conducted using MATLAB 2015. The simulations reported in subsections 5.1.1 and 5.1.2 are based on algorithm 1. They demonstrate the validity of this algorithm for DR-based HEMSs and enable a comparison of the performance parameters between a HEMS with MS and a HEMS with DS. The simulations reported in subsections 5.1.3 and 5.1.4 are based on algorithm 2. They demonstrate the validity of this algorithm for DRSREOD-based HEMSs and similarly enable a comparison of the performance parameters between the MS and the DS. Four scenarios, as listed below, are addressed and critically analyzed:

-A DR-based HEMS with DS (based on algorithm 1) -A DR-based HEMS with MS (based on algorithm 1) -A DRSREOD-based HEMS with DS (based on algorithm 2) -A DRSREOD-based HEMS with MS (based on algorithm 2)

For the simulations, a 2-stage ToU tariff scheme was considered. It consists of a rate of 15 cents/kWh during the peak hours from 19:00 to 23:00 (slot numbers 115-138) and a rate of 9 cents/kWh during the rest of the day, as shown in Fig. 5.24. The detailed specifications and other information for the NSHAs, SHAs, PV system, SB and inverter used to implement the scheduling simulations are given in Tables 4.1 to 4.4.

The hardware and software used for the simulations include the following: Machine: Core i7-4790 CPU @3.6 GHz with 16 GB of RAM Platform: MATLAB 2015a Optimization tool: MOGA/PO with the following parameters: 174


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS Population size: 100 Population type: Double vector Generation size: 1400 Crossover fraction: 0.8 Elite count: 0.05 x population size Pareto fraction: 0.35 Pareto plot function: gaplotpareto

20 18 16

Tariff (cents/kWh)

14 12 10 8 6 4 2 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time Slots

Figure 5.1: Two-stage ToU tariff scheme for August-September 2016

5.1.1

Simulation of a DR-based HEMS with DS

From Fig. 5.25, which shows the optimal tradeoff solutions for the CE and the T BD(D) obtained using DR, it is evident that as the CE decreases, the T BD(D) increases. The user may select the maximum feasible reduction in the CE based on the acceptable T BD. Fig. 5.26 shows the unscheduled and scheduled load curves. It is evident that the DR-based DS shifts most of the load in the forward direction toward the off-peak hour that starts at 11:00 pm (slots 139-144). However, loads ranging from 0.18 to 0.26 kW/slot are scheduled in the peak time slots 117-128, and loads of 0.34 kW/slot are scheduled in the peak time slots 133-138 due to the limited number

175


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS 220

CE (cents)

215

210

205

200

195 0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

TBD (D)

Figure 5.2: Tradeoff between CE and T BD(D) for a DR-based HEMS with DS 1 without scheduling with DR based DS

0.9 0.8

Pgrid (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.3: Scheduled loads for a DR-based HEMS with DS

of off-peak slots available in the delayed direction, as is required for DS, which restricts any further decrease in the CE.

5.1.2

Simulation of a DR-based HEMS with MS

Fig. 5.27 shows the optimal tradeoff solutions for the CE and the T BD(M ) and reveals that as the CE decreases, the T BD(M ) again increases due to the MS of the SHAs. A user may select a maximum feasible reduction in the CE based on his/her maximum bearable discomfort level. It is observed that a greater reduction in the CE is achieved in the MS case compared with the DS case for the same T BD. This is because some of the SHAs that are initially expected to operate during peak hours continue to operate during peak hours even after DS due to 176


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS the limited number of off-peak hours available in the delayed direction, whereas in the MS case, some of these SHAs that would otherwise be operating during peak hours can be designated for AS, thereby enabling the scheduler to shift them toward off-peak hours in the earlier direction after other SHAs have been shifted toward off-peak hours in the delayed direction. Because the scheduler is able to shift more SHA loads to off-peak hours in MS, a greater reduction in the CE is achieved in the MS compared with the DS case for the same T BD.

From Fig.

196

194

CE (cents)

192

190

188

186

184 0.12

0.14

0.16

0.18

0.2

0.22

0.24

0.26

0.28

TBD (M)

Figure 5.4: Tradeoff between CE and T BD(M ) for a DR-based HEMS with MS

1 without scheduling with DR based MS

0.9 0.8

Pgrid (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.5: Scheduled loads for a DR-based HEMS with MS

5.28, it can be seen that the loads selected for the DS are shifted toward off-peak hours in the forward direction (slots 139-144), whereas the loads designated for AS are shifted toward off-peaks hours in the backward direction (slots 99-114). Hence, the loads that continued to operate during the peak time slots 117-128 177


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS in the DS case are instead shifted toward off-peak hours in the earlier direction. Thus, the 0.34 kW/slot loads operating during the peak time slots 133-138 in the DS scenario are reduced to 0.23 kW/slot in the same set of slots. This shifting of more of the load toward off-peak slots (some in the delayed direction and some in the earlier direction) in the MS results in a greater reduction in the CE. Therefore, the MS-based DR shows better performance compared with the DS-based model in terms of both the CE and the T BD.

5.1.3

Simulation of a DRSREOD-based HEMS with DS

A simulation was performed to validate the performance of a HEMS based on DR synergized with the optimal dispatch of the PV system, the SB and the grid to achieve the maximal reduction in the CE through simultaneous reduction of the overall demand and minimization of the grid load during peak hours. P pv is utilized directly by the loads, and any excess power from the PV system is stored in the SB to be utilized during peak hours (or during off-peak hours to avoid an IBR-based penalty) to reduce the CE. Fig. 5.29 shows an approximately linear 126 124 122

CE (cents)

120 118 116 114 112 110 108 106 0

0.05

0.1

0.15

0.2

0.25

0.3

TBD (D)

Figure 5.6: Tradeoff between CE and T BD(D) for a DRSREOD-based HEMS with DS

relation between the CE and the T BD(D), with a slope of 70.8 cents per unit of T BD(D). From Fig. 5.30, it is evident that the PV-SB combination manages to supply almost all of the load during peak hours by virtue of the SB. However, 178


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS 0.8 without scheduling/PV/SB with DS,PV and SB

0.7

Pgrid (kW/slot)

0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.7: Scheduled loads for a DRSREOD-based HEMS with DS

a small part of the load must still be supplied by the grid during peak hours in slots 116-120. The detailed simulation results presented in Fig. 5.31 show that the 1 Pschd (kW/ slot) Ppv Pgrid Pdis Pchg Psold SOC of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.8: Load, PV, SB and energy parameters for a DRSREOD-based HEMS with DS

grid supplies all of the load in the morning during off-peak hours. At 5:20 (slot 133), the power from the PV system begins to gradually rise. Some of the load is supplied by the PV system, while the grid also supplies some power in parallel when the available PV energy is less than the load demand. If the available PV energy is greater than the load demand, the SB is charged. Once the SB is fully charged, any excess PV energy is sold to the grid. Beginning in the 103rd slot, the load demand becomes higher than the PV output, and because this is an off-peak period, the grid supplies power in parallel with the PV system to satisfy the load demand. In the 115th slot, the peak period starts, and the SB is discharged during 179


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS slots 115-120 (based on the maximum discharge rate) to supply the load demand to save money while operating in parallel with the grid and the PV system. In slots 121-132, the load demand is small and can be almost fully supplied from the SB until the SB has discharged to its lower limit, at the end of slot 132. A small amount of power is then supplied by the grid to support the load during slots 133-138 (peak hours). Then, the off-peak period starts in the 139th slot, and the grid again supplies power in the range of 0.225-0.34 kW until the 144th slot. Thus, the grid supplies a small amount of power in the peak time slots 115-120 and 133138. Furthermore, the grid supplies a very low power level of 0.04-0.06 kW/slot in parallel with the SB during the peak time slots 130-132.

The bar charts shown

1 Pschd Pchg Psold

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.9: Power constraints for P schd, P chg and P sold for a DRSREODbased HEMS with DS

1 Pgrid Ppv Pdis

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.10: Power constraints for P grid, P pv and P dis for a DRSREODbased HEMS with DS

in Figs. 5.32 and 5.33 graphically illustrate the two sides of the energy balance 180


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS constraint, i.e., the balance of the energy generated and the energy consumed in each slot, as achieved through heuristic algorithm 2 for DS as per Eq. 4.24.

5.1.4

Simulation of a DRSREOD-based HEMS with MS

Fig. 5.34 shows the tradeoff relation between the CE and the T BD(M ). When the CE decreases, the T BD(M ) increases in an approximately linear fashion, with a slope of 80 cents per unit of T BD(M ), somewhat greater than the corresponding value of 70.8 in DS. This finding indicates that the consumer can achieve a faster/greater reduction in CE in the MS case than in the DS case for a given increase in T BD. 100

95

CE (cents)

90

85

80

75

70 0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

TBD (M)

Figure 5.11: Tradeoff between CE and T BD(M ) for a DRSREOD-based HEMS with MS

0.8 without scheduling/PV/SB with MS,PV and SB

0.7

Pgrid (kW/slot)

0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.12: Scheduled loads for a DRSREOD-based HEMS with MS

181


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS 1 Pschd (kW/slot) Ppv Pgrid Pdis Pchg Psold SOC of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.13: Load, PV, SB and energy parameters for a DRSREOD-based HEMS with MS

From Figs. 5.35 and 5.36, it can be seen that some of the loads that were operated during peak hours in slots 115-120 in the DS scenario are shifted toward off-peak hours in the earlier direction (slots 104-115) to achieve a greater reduction in CE. Furthermore, the rest of the aforementioned load and the peak-hour load in slots 133-138 can be completely supplied by the SB in this scenario, unlike in the DSbased scenario, in which some of this load is supplied from the grid. Thus, the SB can fully supply the load for almost the entire time during peak hours, unlike in the corresponding DS scenario. Furthermore, the system can supply some of the peak load that is shifted in the earlier direction directly through the PV power (instead of selling that excess energy from the PV system to the grid at a cheap rate), as is evident from Fig. 5.36. 1 Pschd Pchg Psold

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.14: Power constraints for P schd, P chg and P sold for a DRSREODbased HEMS with MS

182


5.1. SIMULATIONS FOR THE OPTIMAL OPERATION OF DR- AND Chapter 5 DRSREOD-BASED HEMSS 1 Pgrid Ppv Pdis

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.15: Power constraints for P grid, P pv and P dis for a DRSREODbased HEMS with MS

Table 5.1: Comparison of the maximum reductions in CE, the net bill, and the peak load for different HEMS categories HEMS category

CE (cents)

Reduction in CE (%)

CEsold (cents)

Net bill (cents)

Average T BD

Peak load (kW)

Without HEMS (unscheduled/base case) DR-based HEMS (DS) DR-based HEMS (MS) DRSREOD-based HEMS (DS) DRSREOD-based HEMS (MS)

218.99

-

-

-

0

0.61

Peak load reduction (%) -

198.55

9.33

-

-

0.40

0.34

43.83

185.04

15.50

-

-

0.26

0.39

35.61

107.02

50.68

70.54

36.49

0 .27

0.34

43.83

74.64

65.92

55.80

18.84 (48.37% less than DS/DRSREOD)

0.45

0.33

45.75

Figs. 5.14 and 5.15 graphically illustrate the two sides of the energy balance constraint, i.e., the balance of the energy generated and the energy consumed in each slot, as achieved through the heuristic algorithm 2 for MS as per Eq. 4.24.

5.1.5

Critical analysis of HEMS scheduling (A-D)

The simulation results for the various scenarios are compared in Table VIII in terms of the maximum reduction in the CE, the corresponding T BD, the net bill and the peak load reduction. The reduction in the CE with the application of the proposed HEMS algorithms is computed as follows, taking a base CE value 183


Chapter 5

5.2. SIMULATIONS FOR DG SIZING TO COPE WITH LS IN A DRSREOD-BASED HEMS WITH MS

of 218.99 cents/day for the unscheduled load scenario:

HEM S %RCE = (218.99 − CE)/(218.99)

(5.1)

Eqs. 4.9 and 4.13 are used to compute the net bill paid by the consumer and the peak load that must be supplied by the grid. It is concluded that the MS results in a much larger reduction in the CE compared with the DS in a DR-based HEMS. Similarly, the MS also results in a larger reduction in the CE and in the net bill compared with the DS for a DRSREODbased HEMS. In a DR-based HEMS, the T BD is much lower in the MS scenario than in the DS scenario, even at the minimal CE. In a DRSREOD-based HEMS, the T BD incurred in the MS case for the maximal reduction in the CE is slightly higher; however, for the same T BD of 0.27, the CE value for MS is only 81 cents/day, compared with 107 cents/day for the DS. This result reveals that for the same level of T BD, a much greater reduction in the CE can be achieved in the MS scenario than in the DS scenario. An excellent reduction in the peak load is also achieved in both the MS- and DS-based HEMSs. Thus, the MS approach is recommended to achieve the maximal reduction in CE, a lower T BD, a greater peak load reduction and enhanced user convenience by means of diverse scheduling options.

5.2

Simulations for DG sizing to cope with LS in a DRSREOD-based HEMS with MS

Simulations based on algorithm 3, presented in section V, were run to investigate the optimal sizing of a DG for a consumer already participating in an energydeficient power supply network in a developing country via a DRSREOD-based HEMS. The algorithm computes the optimal size of a DG to cope with a scheduled LS enforced by the utility while synergizing price-based DR with the optimal 184


Chapter 5


dispatch of the PV system, the SB, the grid and the DG. The scheduled LS on an hourly basis for a maximum of 4 hours applied at 10:00, 16:00, 20:00 and 23:00 was considered in the simulations. MOGA/POS was used to obtain optimal tradeoff solutions for P gsize, the percentage reduction in the CE and the value of T BD(M ) under optimal HEMS operation. The loads supplied from the grid with and without a DRSREOD-based HEMS with a DG for the minimal CE are shown in Fig. 5.16. Fig. 5.17 presents the corresponding detailed loads and generation parameters under LS with the application of a DRSREOD-based HEMS with a DG for the minimal CE. Figs. 5.18 and 5.19 graphically illustrate the two sides of the energy balance constraint, i.e., the balance of the energy generated and the energy consumed in each slot, as achieved through heuristic algorithm 3 by incorporating a DG into the dispatch planning during LS hours as per Eq. 4.22. A very small fraction of power is dissipated in the dummy load only when surplus PV energy is available after all HAs have been supplied and the SB has been charged to its maximum SOC during LS hours. 0.8 without scheduling/PV/SB/DG/LS with MS, PV, SB, DG, LS

0.7

Pgrid (kW/slot)

0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.16: Power supplied from the grid with and without a DRSREODbased HEMS with an additional DG to cope with LS

Tradeoff solutions for P gsize, CE and T BD(M ) for a DRSREOD-based HEMS with a DG are graphically presented in Fig. 5.20. The data are classified based on the CE/T BD TOs to assist the consumer in selecting an optimal and feasible

185


Chapter 5

5.2. SIMULATIONS FOR DG SIZING TO COPE WITH LS IN A DRSREOD-BASED HEMS WITH MS 1 Pschd (kW/slot) Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstat Pgen

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.17: Load, P pv, P chg, P dis, P gen and energy parameters for a DRSREOD-based HEMS with a DG to cope with LS 1 Pschd Pchg Psold Pdl

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.18: Power constraints for P schd, P chg, P sold and P dl for a DRSREOD-based HEMS with an LS-compensating generator

solution for the size of the DG. Table 5.2 further elaborates on the data classification for the optimal selection of a DG to cope with the LS in a DRSREOD-based HEMS.

5.2.1

Critical analysis of DG sizing to cope with LS in a DRSREOD-based HEMS with MS

First, five alternative scenarios (numbered I-V and described below) are discussed to investigate the appropriate sizes for DGs to cope with LS in a home with no load

186


Chapter 5

5.2. SIMULATIONS FOR DG SIZING TO COPE WITH LS IN A DRSREOD-BASED HEMS WITH MS 1 Pgrid Ppv Pdis Pgen

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.19: Power constraints for P grid, P pv, P dis and P gen for a DRSREOD-based HEMS with an LS-compensating generator

0.60

2.50

0.50 2.00

TBD (M)

1.50 0.30 1.00

Pgsize (kW)

0.40

0.20

0.50 0.10

0.00

49.63 51.54 53.67 55.04 57.32 58.07 59.25 60.49 62.86 63.50 64.08 64.70 65.82 67.06 68.05 69.99 70.18

0.00

% Reduction in CE Class I

Class II

Class III

TBD (M)

Class IV

Pgsize (kW)

Figure 5.20: Generator size classification based on CE/T BD(M ) TOs

scheduling as well as for smart homes with load scheduling via DR-/DRSREODbased HEMSs as presented in section VI (A-D), without considering the TOs between P gsize, CE and T BD. The DG sizes computed in these scenarios are then used as references/base cases to validate the benefits of the proposed algorithm for optimal DG sizing in a DRSREOD-based HEMS considering the aforementioned 187


Chapter 5

5.2. SIMULATIONS FOR DG SIZING TO COPE WITH LS IN A DRSREOD-BASED HEMS WITH MS Table 5.2: Generator sizing based on CE/T BD(M ) TOs

Sr. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

Class

I

II

III

IV

CE (cents) 63.83 65.30 65.45 65.73 68.60 69.97 71.10 72.12 73.34 74.84 75.23 77.29 78.04 78.65 78.85 79.93 81.04 81.33 83.18 86.51 88.00 89.23 90.68 91.82 93.07 93.46 96.31 98.45 99.55 101.47 106.11 106.11 110.30 110.30

Reduction in CE (%) 49.63 49.63 51.54 51.54 53.67 54.54 55.04 56.02 57.32 57.50 58.07 58.59 59.25 59.81 60.49 62.02 62.86 62.99 63.50 63.99 64.08 64.36 64.70 65.65 65.82 66.51 67.06 67.53 68.05 68.67 69.99 70.11 70.18 70.85

Range

49.6357.32

57.5062.02

62.8666.51

67.0670.85

TBD(M) 0.17 0.17 0.19 0.19 0.18 0.18 0.19 0.21 0.22 0.23 0.23 0.26 0.27 0.27 0.27 0.34 0.31 0.30 0.38 0.37 0.44 0.37 0.43 0.44 0.46 0.48 0.33 0.35 0.34 0.37 0.40 0.48 0.50 0.50

Range

0.170.22

0.230.34

0.300.48

0.330.50

Pgsize (kW) 0.35 0.35 0.35 0.35 0.41 0.41 0.41 0.41 0.41 0.92 0.72 0.35 0.63 1.11 1.01 0.72 1.32 1.52 0.35 0.53 0.62 0.72 0.95 1.01 0.95 0.95 1.62 1.62 1.71 1.35 1.35 1.95 1.95 1.95

Range

0.350.41

0.3 51.11

0.351.52

1.3 51.95

TOs.

Scenario I: In this scenario, unscheduled house loads are considered, and the consumer selects a DG to supply the necessary load during LS hours when he/she is neither participating in DR nor using a PV/SB system. When hourly scheduled LS at 10:00, 16:00, 20:00 and 23:00 is assumed, a peak load of 2.04 kW occurs during LS at 20:00 (121st slot), and hence, the DG should be sized for 2.04 kW. When the LS schedule is shifted toward peak hours of 7-11 p.m., the peak load reaches 3.66 kW in the 115th slot. Thus, a DG that can supply a maximum of 3.66 kV can be safely chosen in this scenario. Scenario II: The consumer is participating in DR based on DS. Under scheduled LS, a peak load of 2.04 kW occurs during LS hours at 23:00 (139th slot), and the

188


Chapter 5


DG should accordingly be sized for 2.04 kW. Scenario III: The consumer is participating in DR based on MS. A peak load of 2.34 kW occurs during LS hours at 16:00 (100th slot), and the DG should accordingly be sized for 2.34 kW. Scenario IV: The consumer is participating in the energy network via a DRSREODbased HEMS with DS to maximize the reductions in the CE and the T BD (without considering the DG requirements and LS effects). A peak load of 2.04 kW occurs during LS hours at 23:00, and the DG should accordingly be sized for 2.04 kW. Scenario V: The consumer is participating in the energy network via a DRSREODbased HEMS with MS to maximize the reductions in the CE and the T BD (without considering the DG requirements and LS effects). A peak load of 1.35 kW occurs during LS hours at 23:00, and the DG in this scenario should accordingly be sized for 1.35 kW. Scenario VI: This is the actual scenario for computing the appropriate size for a DG to cope with LS in a DRSREOD based HEMS. The algorithm for this scenario was developed based on scenario V. All of the computations for scenario V are performed. Additionally, 4 imposed LS hours and the dispatch of the DG during these LS hours are included in the algorithm. The DG size is included as a third fitness function, along with the CE and the T BD, in determining the POS. This scenario for DG selection using the proposed tradeoff-based classification given in Fig. 5.20 is of immense interest for comparison with scenarios I-V. The maximum supply capacity of the DG required in scenario VI ranges from 0.41 to 1.95 kW for the various classes, far less than the required DG capacities of 3.65 kW, 2.04 kW, 2.34 kW and 2.04 kW in reference scenarios I-V, respectively. When comparing scenario VI with reference scenario V, the consumer finds that scenario VI offers a great flexibility (multiple choices providing diverse options) in selecting the DG size, with capacities ranging from 0.41 to 1.95 kW based on the CE/T BD TOs, compared with the fixed DG capacity requirement of 1.35 kW in scenario V. Based 189


Chapter 5


on Fig. 5.20 and Table 5.2, the salient features of the proposed method and the underlying classification for the selection of a DG to cope with LS in a DRSREODbased HEMS based on scenario VI can be summarized as follows: Class I: In this class, the percentage reduction in the CE ranges from 49.63 to 57.32, with corresponding T BD levels from 0.17 to 0.22 and a DG sized for 0.41 kW. Consumers opting for this class may enjoy a cost reduction of up to 57.32% with the lowest T BD levels of up to 0.22. The DG size necessary to manage the power supply interruptions is also the lowest, i.e., with a power supply capacity of 0.41 kW. Therefore, comfort-conscious consumers should choose this class, with its minimal T BD, reasonable reduction in the CE and a minimum capital cost for a DG to ensure an uninterrupted supply of power. Class II: In this class, the percentage reduction in the CE ranges from 57.50 to 62, with corresponding T BD levels from 0.23 to 0.34 and a DG sized for 1.11 kW. Consumers opting for this class can achieve a cost reduction of up to 62% (greater than in class I) with an accompanying increase in T BD up to 0.34 (also greater than in class I). A DG power supply capacity of almost 2.71 times that in class I is required to ensure an uninterrupted supply of power. This class seems less attractive for consumers due to the larger required DG size compared with class I. Class III: In this class, the percentage reduction in the CE ranges from 62.86 to 66.51, with corresponding discomfort levels from 0.30 to 0.48 and a DG sized for 1.52 kW. Consumers opting for this class can achieve a maximum cost reduction of 66.51% (greater than in class II) with a T BD level of 0.48 (also greater than in class II). Consumers may opt for this class to achieve a greater reduction in the CE with a mildly increased T BD. However, this class is expected to be more attractive if the DG sizes available in the market for this class overlap with those sized for class II. Class IV: In this class, the percentage reduction in CE ranges from 67 to 70.85, with corresponding T BD levels from 0.33 to 0.50 and a DG sized for 1.95 kW. 190


Chapter 5


Consumers opting for this class will achieve a maximum cost reduction of 70.85% (the largest in all classes) accompanied by a maximum T BD level of 0.50 (also the largest in all classes). Consumers who are not concerned about the T BD should choose this class to achieve the maximum possible reduction in CE. This class may be more attractive to consumers if the DG sizes available in the market for this class overlap with those sized for class III.

80

2.5

70 2.0

50

1.5

40 1.0

30

Pgsize (kW)

% Reduction in CE

60

20 0.5 10

0

0.17 0.18 0.19 0.19 0.22 0.23 0.27 0.27 0.31 0.34 0.35 0.37 0.38 0.43 0.44 0.48 0.50

0.0

TBD (M) Class-I

Class-II

% Reduction in CE

Class-III

Pgsize (kW)

Figure 5.21: Generator size classification based on T BD(M )/CE TOs

Alternatively, when the T BD is placed on the x axis, as in Fig. 5.21, a different set of classes for selecting an optimal DG size is obtained that is predominantly based on the T BD. It is observed that in class I, a maximum T BD of 0.23 is incurred to achieve a cost reduction of up to 57% with a DG sized for 0.9 kW. Similar trends are observed for class II and class III. This T BD-focused classification will be of greater interest to comfort-conscious consumers.

191


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM

5.3

DRSREODLDG based HEMS operation and the filtration mechanism

The simulations reported in subsection 5.3.1 are based on algorithm 1. They demonstrate the validity of MOGA/ PO based heuristic for DRSREODLDG-based HEMS to compute operational schemes for SHAs in terms of vector T st and the primary TOs for CEnet, T BD and T EM iss. The results of simulations enable analyzing the trends exhibited by the tradeoff parameters taking into consideration vital factors affecting these parameters. The critical analysis of the primary TOs enabled designing a filtration mechanism to extract desired set of eco-efficient tradeoff solutions with minimal T EM iss. The simulations reported in subsection 5.3.2 are based on algorithm 2. They demonstrated the validity of the filtration mechanism to harness eco-efficient TOs. Regression based polynomial formulations and the procedure to finalize the model fits for the proposed mechanism are also elaborated in subsection 5.3.2. Simulations have been conducted for the following: -DRSREODLDG-based HEMS operation to compute primary TOs for HEMS ( based on algorithm 1/ step-1). -Application of filtration mechanism to harness eco-efficient TOs for HEMS (based on algorithm 2/ step-2 and step-3).

5.3.1

Simulations for DRSREODLDG-based HEMS operation to compute primary TOs using algorithm 1

Simulations were performed to validate DRSREODLDG-based HEMS operation using algorithm 1. Operating schemes for SHAs in terms of T st and the primary TOs were computed. The trends exhibited by the tradeoff parameters were analyzed. Critical analysis for validating the relation between the tradeoff parameter: T EM iss and the TOs for CEnet, T BD, enabled designing a filtration mechanism required to harness the desired eco-efficient tradeoff solutions with minimal 192


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM T EM iss from a large set of primary TOs.

For the simulations, a 2-stage ToU tariff scheme with an IBR value of 1.4 was considered. It consists of a rate of 15 Cents/kWh during the peak hours from 19:00 to 23:00 (slot numbers 115-138) hours and a rate of 9 Cents/kWh during the rest of the day, as shown in Figure 5.22. For the application of the IBR factor, a threshold power demand of 2.4 kW was considered. A feed-in tariff, P Ef , valued at 0.7 times of P E was considered for the PV energy sold to the grid. 20 18 16

Tariff (cents/kWh)

14 12 10 8 6 4 2 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time Slots

Figure 5.22: Two-stage ToU tariff scheme

The hardware and the software used for simulations included the followings: Machine: Core i7-4790 CPU @3.6 GHz with 16 GB of RAM Platform: MATLAB 2015a Optimization tool: MOGA/PO with the following parameters: Population size: 100 Population type: Double vector Generation size: 1400 Crossover fraction: 0.8 Elite count: 0.05 x population size Pareto fraction: 1

193



1.5

Step-1 Solutions Generation: Primary solutions generated through algorithm-1

1.6

1.4 1.3

1.4

TEMiss (Lbs.)

1.2

1.2 1.1

1 1

0.8 0.9

0.6

0.8

0.4

0.7

0.5

0.4

0.3

0.2

30

TBD

35

40

50

45

0.6

CEnet (Cents)

Figure 5.23: Primary tradeoff solutions with un-even surface for T EM iss generated through Algorithm 1.

Table 5.3: PRIMARY TOs FOR DRSREODLDG-BASED HEMS USING ALGORITHM 1 (STEP-1) Sr.No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

CEnet (Cents)

T BD

T EM iss (Lbs.)

52.87 52.87 51.74 51.74 51.02 51.02 50.39 50.3 49.5 48.78 48.78 48 47.51 47.51 46.77 45.81 45.72 45.18 45.01 44.62 44.36 44.08 43.96 43.96 43.8 43.74 43.57 43.45 43.27 43.07 42.73 41.9 41.33 41.15 40.92 40.87 40.69 40.43 40.31 40.06 40.06 39.22 38.94 38.69 38.47 38.11 37.88 37.56 36.89 36.66

0.17 0.17 0.17 0.18 0.17 0.17 0.17 0.18 0.17 0.18 0.18 0.18 0.19 0.19 0.18 0.19 0.19 0.18 0.19 0.2 0.2 0.19 0.22 0.22 0.2 0.2 0.2 0.19 0.24 0.22 0.2 0.2 0.21 0.22 0.23 0.22 0.22 0.27 0.21 0.26 0.26 0.23 0.25 0.24 0.23 0.24 0.25 0.26 0.32 0.27

0.67 0.67 0.67 0.67 0.75 0.75 0.81 0.67 0.84 0.8 0.8 0.75 0.81 0.81 0.85 0.57 0.99 0.6 0.56 0.6 1.36 0.88 1.55 1.55 0.74 1.15 0.56 1.18 0.56 0.79 0.99 0.68 0.57 0.8 0.56 0.66 0.74 0.64 0.66 1.41 1.41 1.03 0.68 0.74 0.85 1.03 0.6 0.56 0.56 0.6

T EM iss Resid avg 0.11 0.11 0.11 0.11 0.03 0.03 -0.03 0.11 -0.06 -0.02 -0.02 0.03 -0.03 -0.03 -0.07 0.21 -0.21 0.18 0.22 0.18 -0.58 -0.1 -0.77 -0.77 0.04 -0.37 0.22 -0.39 0.22 -0.01 -0.21 0.1 0.21 -0.02 0.22 0.13 0.04 0.14 0.13 -0.63 -0.63 -0.25 0.1 0.04 -0.07 -0.25 0.18 0.22 0.22 0.18

P dl (kWh) 1.87 1.87 1.87 1.87 1.79 1.79 1.69 1.71 1.58 1.55 1.55 1.63 1.38 1.38 1.45 1.61 1.24 1.66 1.9 1.5 1.24 1.31 1.09 1.09 1.49 1.24 1.74 1.24 1.34 1.26 1.09 1.43 1.46 1.24 1.58 1.4 1.32 1.01 1.25 0.84 0.84 0.96 1.28 1.18 0.88 0.84 1.21 1.11 1.28 1.16

Sr.No. 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

194

CEnet (Cents)

T BD

T EM iss (Lbs.)

36.65 36.34 36.18 36.18 35.82 35.56 35.13 35.03 33.9 33.86 33.85 33.68 33.67 33.02 32.98 32.96 32.67 32.57 32.37 32.24 32.01 32.01 31.99 31.92 31.88 31.76 31.76 31.45 31.27 31.15 30.82 30.49 30.34 30.3 30.15 30.02 29.65 29.65 29.03 28.77 28.75 27.98 27.87 27.87 27.36 26.98 26.98 26.8 26.66 26.22

0.28 0.28 0.31 0.31 0.28 0.29 0.3 0.3 0.32 0.33 0.31 0.34 0.29 0.59 0.33 0.36 0.34 0.47 0.33 0.33 0.35 0.35 0.35 0.51 0.43 0.57 0.57 0.35 0.52 0.35 0.32 0.35 0.4 0.36 0.43 0.53 0.37 0.37 0.36 0.37 0.37 0.38 0.44 0.44 0.46 0.41 0.49 0.45 0.51 0.48

0.56 0.7 1.27 1.27 0.73 0.57 0.6 0.57 0.57 0.64 0.92 0.56 0.65 0.57 1.11 0.56 0.56 0.56 0.65 0.7 1.23 1.23 1.2 0.56 0.65 0.62 0.62 0.7 0.56 0.9 0.73 0.7 0.7 0.73 1.09 0.56 0.7 0.7 0.73 0.73 0.85 0.73 0.85 0.85 0.65 0.73 0.73 0.65 0.73 0.65

T Emiss Resid avg 0.22 0.08 -0.49 -0.49 0.05 0.22 0.18 0.22 0.22 0.14 -0.14 0.22 0.13 0.21 -0.33 0.22 0.22 0.22 0.13 0.08 -0.45 -0.45 -0.42 0.22 0.13 0.17 0.17 0.08 0.22 -0.12 0.05 0.08 0.08 0.05 -0.31 0.22 0.08 0.08 0.05 0.05 -0.07 0.05 -0.07 -0.07 0.13 0.05 0.05 0.13 0.05 0.13

P dl (kWh) 1.06 0.91 0.59 0.59 0.99 0.94 0.86 0.86 0.84 0.74 0.57 0.81 0.63 0.27 0.33 0.62 0.7 0.73 0.51 0.54 0.17 0.17 0.17 0.63 0.59 0.27 0.27 0.42 0.53 0.32 0.34 0.32 0.25 0.29 0.03 0.33 0.14 0.14 0.21 0.22 0.09 0.04 0.01 0.01 0.11 0.04 0.04 0.11 0.04 0.11


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM The primary TOs for CEnet, T BD and T EM iss, generated through simulation for an optimal DRSREODLDG-based HEMS operation are presented in Table 5.3. Due to space limitation, the related T st vector is not shown in this table (however, it is presented with the final eco-efficient TOs in Table 5.6). The primary TOs are graphically shown in Figure 5.23. The trends exhibited by the tradeoff parameters and the relationship between them has been analyzed to approach a filtration mechanism that enables harnessing TOs with diversified options for CEnet, T BD and minimal value of T EM iss. Refer to Table 5.3, each tradeoff solution is related to a unique T st. The decision vector T st is generated through the MOGA based on the vectors ST slot and EN slot. The vector T st for each of the solution corresponds to a unique demand profile, P schd. To supply this demand, a dispatch scheme for energy sources and ESS based on the parameters P pv, P gd, P gn, P ch, and P ds is computed through the heuristic proposed in algorithm 1. Preferably, the load is supplied from the PV unit. The extra energy from the PV unit is stored in the SB after supplying the load. The SB is discharged to supply the load during the peak hours for making use of the stored energy. The LDG supplies the load in coordination with the SB during LSD hours only. The excess PV energy is sold to the grid viz designated as P sold after supplying the load and charging the SB. However, during the LSD hours, the excess energy from the PV is ought to be dissipated into the dummy load viz designated as P dl. The PV, SB, and the charger system are considered a part of the existing infrastructure and their cost is not included in computation. The tradeoff parameter CEnet is based on the dispatch from various sources to supply the scheduled load and the energy sold to the utility according to Eq. 4.17. The rates for energies including P E, P Ef and P Eg in different slots play vital role in the computation of CEnet. The loss of the harnessed PV energy due to the unavailability of the grid, given by P dl, is another important factor affecting the value of CEnet. The parameter T EM iss primarily depends on the energy supplied by the LDG, P gn, during LSD hours. The EF T for the LDG is also 195


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM important while evaluating T EM iss. The T BD is based on the time shift of SHAs from their preferred times of operation and is computed using Eq. 4.5. The relationships between the tradeoff parameters for the primary tradeoff solutions are graphically presented in Figure 1.2 and Figure 5.24. 1.60

TEmiss (Lbs.), TBD

1.40

1.20

1.00

0.80

0.60

0.40

0.20

52.87 51.02 48.78 45.81 44.36 43.74 42.73 40.87 40.06 38.11 36.65 35.56 33.85 32.96 32.01 31.76 30.82 30.02 28.75 26.98

0.00

CEnet (Cents)

TBD

TEmiss

Figure 5.24: Relations among primary TOs for CEnet, TBD and TEMiss using algorithm 1

The trends exhibited by the tradeoff parameters comprising CEnet, T BD and T EM iss are analyzed in subsections below. The primary TOs with extreme values of the parameters have especially been investigated. 5.3.1.1

Trends for CEnet

The objective to minimize the CEnet is mainly based on the following factors:

1. Maximized usage of the PV energy to supply the load directly: This avoids the loss of energy in the SB due to storage/re-use of the PV energy while supplying the load (a net loss of 20% has been assumed for the SB). The energy thus saved enables to reduce the demand from the grid and the LDG which ultimately results in a reduced value of CEnet. 2. Maximized usage of the stored PV energy to supply the load during the peak hours: This reduces the energy to be supplied from the grid during the peak hours as well as from the LDG during the peak LSD hours that results in a reduced value of CEnet. 196


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM 3. Selling of the extra PV energy to the utility: A direct usage of the energy from the PV unit is better than selling it to the utility as P Ef is generally lesser than the P E (P Ef is assumed as 70% of the P E). However, it is beneficial to sell the PV energy to the utility, if surplus of it is available after supplying P schd and the charging load. The above-mentioned factors enable reducing the CEnet parameter through an optimal use of the PV energy based on the P E, P Ef , P Eg and the SB efficiency. Other factors to reduce CEnet parameter include the followings: 4. Load shifting towards the off-peak hours: The load left after being supplied from the PV and the SB unit should have been shifted towards off-peak hours. This shifting minimizes the CEnet based on the tariff P E. 5. Load to be supplied by the LDG during LSD hours: The algorithm enables supply of the energy from the LDG during LSD hours. If more load is shifted towards LSD hours, LDG is required to supply that load in coordination with the PV/SB at a higher cost of energy (P Eg) that results in an increased value of CEnet. 6. Loss of the harnessed PV energy: The dummy load P dl has been identified as a factor of vital importance for reducing CEnet. Figure 5.25 reveals a direct relationship between the CEnet and the P dl. The P dl needs to be minimized to achieve an optimal value of CEnet. A larger P dl indicates a loss of the PV energy due to lesser shifting of the load (including charging of the SB) towards the LSD hours having the harnessed PV that results in a larger CEnet.

To investigate the variations in CEnet parameter based on the above mentioned 6 number of factors, solution-1 and solution-100 with the maximum and the minimum values of CEnet are analyzed as case studies. The analysis is based on the related HEMS operation including the power profiles for the loads and the dispatch scheme for the power sources and the SB. 197


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM 2.50

2.00

Pdl (kWh)

1.50 R² = 0.8904

1.00

0.50

0.00 0.00

10.00

20.00

30.00

40.00

50.00

60.00

CEnet (Cents) Pdl

Linear (Pdl)

Figure 5.25: Relation between CEnet and P dl for DRSREODLDG-based HEMS

Solution-1 shows a CEnet value of 52.87 Cents, the largest of all solutions. This largest value of CEnet may be analyzed based on the above-mentioned factors by focusing on the power profiles for this solution shown in Figure 5.26.

1 Pschd Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstatus EMiss

Power (kW/ slot), EMiss (Lbs.)

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.26: Power and emission profiles for DRSREODLDG-based HEMS operation for solution-1

First, a very small portion of the load (P schd) has been supplied directly from the PV energy that is available from time slot no. 37. Some of the available PV energy has been used to charge the SB while most of the PV energy is sold to the utility at cheap rates (P Ef equals 70% of P E). A part of the load, instead of being supplied directly from the PV unit, is shifted towards the off-peak slots and 198


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM supplied from the grid at the off-peak time rate. This load thus has been supplied at a net 30% increased cost of the energy as compared to the cost of energy sold to the grid. Second, a load larger than the capacity of the SB is shifted towards the peak-time slots. An average load of 0.21 kWh is thus supplied from the grid during peak time slot nos. 132-134. The CEnet could be reduced if the load exceeding the capacity of the SB was shifted towards off-peak time. Third, a net load of 0.348 kWh has been supplied from the LDG during LSD based slot nos. 139-144 at a rate of P Eg (viz higher than P E). This load is based on NSHAs only and it can not be shifted. However, the LDG also supplies a load of 0.068 kWh during slot no. 102 that may be shifted towards the grid/ PV to reduce the CEnet. Fourth, the least of the load has been shifted within the PV harnessed LSD hours starting from slot nos. 61 and 97. Under this scenario, 1.87 kWh of the PV energy has been lost/ dumped during slot nos. 63-66 and slot nos. 97-101. More load could be shifted towards the mentioned slots to minimize the loss of the harnessed energy from the PV and thus to reduce the CEnet. In brief, a load shifting resulted in a non-optimal use of the PV energy, a very large value of the P dl and other aforementioned factors resulted in the largest value of CEnet for this solution. Solution-100, on the other hand exhibits the lowest CEnet value of 26.22 Cents that is again based on the aforementioned factors. The lowest value of CEnet may again be analyzed by focusing on the corresponding power profiles for the solution as shown in Figure 5.27.

First, a larger portion of the load (P schd), as compared to solution-1, has been supplied directly from the PV that is available from time slot no. 37. The harnessed PV energy has been used to charge the SB as well as to supply the maximum of the load, while a smaller value of the PV energy is sold to the utility at cheap rates. Second, the remaining load viz smaller as compared to solution-1 has been shifted towards the peak time slots so that the SB is able to supply most of the 199


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM 1 Pschd Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstatus EMiss


0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots


said load. Accordingly, an average load of 0.189 kWh is left to be supplied by the grid during the peak time slot nos. 135-137 that is smaller as compared to the same load in solution-1. Third, the LDG supplies a total energy of 0.054 kWh during slot nos. 100-101, that is smaller as compared to the same parameter in solution-1. Fourth, most of the load has been shifted towards the PV harnessed LSD hours and hence P dl exhibits a minimal value 0.11 kWh. In brief, a load shifting enabling an optimal use of the PV energy, minimized value of P dl and other aforementioned factors resulted in the lowest CEnet for this solution. Similarly, the solutions with intermediate value of CEnet may also be validated by focusing the same above mentioned factors affecting CEnet.

5.3.1.2

Trends for T BD

The value of T BD is based on the total time shifts of the SHAs from the preferred times (ST slot or EN slot based on type of scheduling) provided by the consumers. It depends on the decision vector T st and computed using Eq. 4.5 through algorithm 1. The simulations reveal an exponential relation between the CEnet and T BD as shown in Figure 5.28. The T BD increases exponentially while reducing the CEnet. The relationship between the CEnet and T BD is very important in the context of the consumer’s welfare. The optimal solutions provide diverse choices to the consumer for TOs between CEnet and T BD. However, it has been 200


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM observed that CEnet cannot be reduced beyond a specific value after the T BD reaches a knee-point value. A knee-point value of 0.48 for T BD may be realized from Figure 5.28. 0.70

0.60

0.50

TBD

0.40

0.30

R² = 0.8501

0.20

0.10

0.00 20.00

25.00

30.00

35.00

40.00

45.00

50.00

55.00

CEnet (Cents) TBD

Expon. (TBD)

Figure 5.28: Relation between CEnet and T BD for a DRSREODLDG-based HEMS

On the other hand, the relation between the T BD and T EM iss for DRSREODLDGbased HEMS is highly un-even as shown in Figure 5.29. Such relations are not possible to be defined using standard techniques.

1.80 1.60 1.40

TEMiss (Lbs.)

1.20 1.00

R² = 0.0351

0.80 0.60 0.40 0.20 0.00 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

TBD

TEMiss

Linear (TEMiss)

Figure 5.29: Relation between T BD and T EM iss for a DRSREODLDGbased HEMS

201


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM 5.3.1.3

Trends for T EM iss

The variation in T EM iss is analyzed based on the primary TOs presented in the Figure 5.24/ Table 5.3. Figure 5.24 exhibits an extremely uneven variations in T EM iss as related to CEnet (and T BD), especially around the center of the data. The solution-23 with the largest, solution-27 with the smallest and solution73 with moderate values of T EM iss are analyzed as case studies. 1 Pschd Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstatus EMiss


0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots


Solution-23 exhibits a T EM iss value of 1.55 Lbs., the largest of all solutions. The value of T EM iss parameter depends on profile for P gn parameter. The profile for this solution is analyzed by focusing on the power/ emission profiles shown in Figure 5.30. The value of T EM iss mainly depends on the operation of the LDG during four number of LSD hours discussed as follows. The loads shifted in the first LSD hour (starting at slot no. 61) and in the third LSD hours (the peak time hour starting at slot no. 121) are completely supplied by the PV and the SB respectively. So, in actual, the LDG has to operate only during the second LSD hour (starting at slot no. 97) and during the fourth LSD hour (starting at slot no. 139) to supply the shifted load as neither the grid nor the SB is available to supply within these hours. During the fourth LSD hour, a fixed load made up of NSHAs is supplied by the LDG completely. As no other source is available to supply during this hour, the fixed load has been supplied by the LDG in all scenarios. Focusing 202


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM the second LSD hour, PV is available to supply the shifted load; however, the demand exceeding the energy harnessed from the PV (named excess demand) is only supplied through the LDG. This excess demand to be supplied by the LDG during the second LSD hour combined with the fixed demand in the fourth LSD hour, in fact, determines the net value of T EM iss. A maximum shifting of the excess demand out of the second LSD hour results in the minimization of the T EM iss. For solution-23, a maximum excess demand supplied through the LDG during the second LSD hour resulted in a maximum T EM iss value of 1.55Lb. for this solution. The CEnet parameter in this scenario assumes a near average value of 43.96 Cents that is based on the combined effect of the related parameters’ values including: a PV energy loss of 1.09 kW; a supply of an average load of 0.2 kWh through the grid during peak time slot nos. 132-134; and a maximum supply of 0.98 kWh of energy from the LDG at a higher cost of value (P Eg ). 1 Pschd Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstatus EMiss


0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots


Solution-27 exhibits a T EM iss value of 0.56 Lbs, the lowest in all solutions and the power profiles shown in Figure 5.31. The minimum value of T EM iss in this scenario is because of zero loading of LDG during the second LSD hour. On the other hand, the CEnet parameter shows a near average value of 43.57 Cents that is nearly similar to the CEnet value in solution-23. The value is again based on the combined effect of the related parameters’ values including: a PV energy loss of 1.75 kW; supply of an average load of 0.23 kW by the grid during the peak time 203


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM slot nos. 132-134; and a minimum supply of 0.35 kWh of energy from the LDG at a higher cost, P Eg. Solution-73 shows a moderate T EM iss value of 1.20 Lbs. corresponding to the power profiles shown in Figure 5.32. The excess load during the second LSD hour has not been completely shifted out of this hour and so the same has been supplied through the LDG. The T EM iss for this solution, therefore, is higher as compared to its value for solution-27. A much lower CEnet of value 31.99 Cents as compared to the value of CEnet in solution-27 is based on a more efficient shifting of the load and a smaller value of P dl in solution-73. 1 Pschd Ppv Pgrid Pdis Pchg Psold Pdl SOC of SB(%) Pgdstatus EMiss


0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots


5.3.1.4

Critical analysis for T EM iss and TOs for CEnet and T EM iss

The relation between T EM iss parameter and the TOs for CEnet and T BD is analyzed based on the primary TOs (sorted on CEnet), presented in Table 5.3. The TOs are graphically shown in Figure 5.33. Based on the variations in T EM iss, the data may be divided into three classes. Class-1, including solution nos. 01-20 at the beginning of the data, class-2, including solution nos. 21-73 around the center of the data, and class-3, including solution nos. 74-100 at the end of the data. Class-1 is characterized by the TOs with minimal values of T EM iss combined with maximal values of CEnet; and class-3 by the TOs with minimal values of 204


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM both of the T EM iss and CEnet parameters. Whereas, class-2 around the middle of the data, including more than 50% of the TOs, exhibits highly un-even/ irregular trend for T EM iss as related to the TOs for CEnet and T BD. It includes an uneven distribution of the data with the minimal, average as well as extremely high values of the T EM iss. Such trends indicate the presence of numerous solutions with comparable values of the TOs for CEnet and T BD; however, with large variations in the related values for T EM iss. Solutions-23 and 27, graphically shown as points A and B respectively in Figure 5.33 are an example of such large variation in the T EM iss parameter. For comparable values of (43.96, 0.22) and (43.57, 0.2) for CEnet and T DB; the solutions exhibit extremely varied values of 1.55 Lbs. (maximum of all solutions) and 0.56 Lbs. (minimum of all solutions) for T EM iss. 1.80

A

1.60

1.40 D

TEMiss (Lbs.), TBD

1.20

1.00 R² = 0.0111 0.80

C

0.60

B

0.40

R² = 0.8501

0.20

0.00 20.00

25.00

30.00

35.00

40.00

45.00

50.00

55.00

CEnet (Cents) TBD

TEMiss

Expon. (TBD)

Linear (TEMiss)

Figure 5.33: Variations in T EM iss with the related TOs for CEnet and T BD

Solution-69 and solution-72 shown as point C and D are another example of similar large variations in T EM iss. For comparable values of (32.37, 0.33) and (32.01, 0.35) for CEnet and T DB; the solutions show largely varied respective values of 0.65 Lbs. and 1.23 Lbs. for T EM iss.

205


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM Figure 5.33 reveals a large number of data points especially in class-2 exhibiting large variations in T EM iss with very small corresponding variation in the respective tradeoff values of CEnet and T BD. The finding regarding the existence of a large number of multiple comparable TOs for CEnet and T BD with extremely varied values of T EM iss in the primary TOs was exploited to design a mechanism to harness eco-efficient TOs for DRSREODLDG-based HEMS operation. A filtration mechanism was proposed to screen out the TOs with larger values of T EM iss in order to harness eco-efficient TOs with minimal T EM iss and a set of diverse TOs for CEnet and T BD. The proposed mechanism, based on an average value constraint filter and an average surface based constraint filter, is presented in algorithm 2. Table 5.4: TRADEOFFS ACHIEVED AFTER APPLYING AVCF BASED ON ALGORITHM 2 (STEP-2) Sr. No.

CEnet (Cents)

T BD

T EM iss (Lbs.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

52.87 52.87 51.74 51.74 51.02 51.02 50.3 48 45.81 45.18 45.01 44.62 43.8 43.57 43.27 41.9 41.33 40.92 40.87 40.69 40.43 40.31 38.94 38.69 37.88 37.56 36.89 36.66 36.65 36.34 35.82 35.56 35.13

0.17 0.17 0.17 0.18 0.17 0.17 0.18 0.18 0.19 0.18 0.19 0.2 0.2 0.2 0.24 0.2 0.21 0.23 0.22 0.22 0.27 0.21 0.25 0.24 0.25 0.26 0.32 0.27 0.28 0.28 0.28 0.29 0.3

0.67 0.67 0.67 0.67 0.75 0.75 0.67 0.75 0.57 0.6 0.56 0.6 0.74 0.56 0.56 0.68 0.57 0.56 0.66 0.74 0.64 0.66 0.68 0.74 0.6 0.56 0.56 0.6 0.56 0.7 0.73 0.57 0.6

T EM iss Resid avgs 0.02 0.02 0.02 0.00 -0.06 -0.06 0.00 -0.08 0.08 0.07 0.09 0.04 -0.10 0.08 0.03 -0.04 0.06 0.05 -0.04 -0.12 -0.06 -0.03 -0.07 -0.12 0.02 0.06 0.03 0.02 0.06 -0.08 -0.10 0.05 0.03

P dl (kWh)

Sr. No.

CEnet (Cents)

T BD

T EM iss (Lbs.)

1.87 1.87 1.87 1.87 1.79 1.79 1.71 1.63 1.61 1.66 1.9 1.5 1.49 1.74 1.34 1.43 1.46 1.58 1.4 1.32 1.01 1.25 1.28 1.18 1.21 1.11 1.28 1.16 1.06 0.91 0.99 0.94 0.86

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

35.03 33.9 33.86 33.68 33.67 33.02 32.96 32.67 32.57 32.37 32.24 31.92 31.88 31.76 31.76 31.45 31.27 30.82 30.49 30.34 30.3 30.02 29.65 29.65 29.03 28.77 27.98 27.36 26.98 26.98 26.8 26.66 26.22

0.3 0.32 0.33 0.34 0.29 0.59 0.36 0.34 0.47 0.33 0.33 0.51 0.43 0.57 0.57 0.35 0.52 0.32 0.35 0.4 0.36 0.53 0.37 0.37 0.36 0.37 0.38 0.46 0.41 0.49 0.45 0.51 0.48

0.57 0.57 0.64 0.56 0.65 0.57 0.56 0.56 0.56 0.65 0.7 0.56 0.65 0.62 0.62 0.7 0.56 0.73 0.7 0.7 0.73 0.56 0.7 0.7 0.73 0.73 0.73 0.65 0.73 0.73 0.65 0.73 0.65

206

T Emiss Resid avgs 0.06 0.07 -0.01 0.07 0.00 -0.04 0.07 0.09 0.04 0.01 -0.04 0.03 -0.02 -0.05 -0.05 -0.03 0.04 -0.04 -0.02 -0.03 -0.05 0.06 -0.01 -0.01 -0.02 -0.02 -0.01 0.04 0.00 -0.05 0.06 -0.06 0.05

P dl (kWh)


0.86 0.84 0.74 0.81 0.63 0.27 0.62 0.7 0.73 0.51 0.54 0.63 0.59 0.27 0.27 0.42 0.53 0.34 0.32 0.25 0.29 0.33 0.14 0.14 0.21 0.22 0.04 0.11 0.04 0.04 0.11 0.04 0.11


5.3.2

Simulations for filtration mechanism to harness ecoefficient TOs using algorithm 2

The simulation for filtration mechanism is based on algorithm 2. The mechanism completes its task in two steps as follows:

• Application of an AVCF to the primary TOs to filter out the the TOs with extremely high and above average values of T EM iss (step-2)

• Application of an ASCF to the filtrate of step-2 to filter out the TOs with marginally higher values of T EM iss (step-3)

5.3.2.1

Simulation for filtration using AVCF (step-2)

This step includes the formulation and application of a constraint filter based on the average value of T EM iss for the primary tradeoff solutions presented in Table 5.3. Following are the software and hardware tools used to demonstrate the solution space, to formulate and apply the filter to validate the AVCF based filtration:

Machine: Core i7-4790 CPU @3.6 GHz with 16 GB of RAM Platform: MATLAB 2015a Regression model = Linear interpolation Interpolation surface model = linearinterp Method = Linear least square Normalize = off Robust = off AVCF formulation and application: T EM iss Resid avg = average(T EM iss) − T EM iss

207


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM Exclude = T EM iss Resid avg < 0

Where T EM iss Resid avg is the decision element for the filter. The exclude option provided with the surface fitting function can be used to screen out the TOs based on the formulation of the decision element. As per the formulation for T EM iss Resid avg, a tradeoff solution with a negative value of the decision element T EM iss Resid avg indicates the above average value for T EM iss. The application of AVCF thus screens out the TOs with extremely high as well as above the average values of T EM iss. The function of the AVCF to screen out the undesired TOs with larger values of T EM iss are graphically shown in Figure 5.34. The selected solutions after the application of the AVCF are presented in Table 5.4.

Step-2 Avg. Filteration: Selected solutions (with below avg. TEMiss). Filtered out/ excluded solutions (with above avg. TEMiss)

1.6

1.4 1.3

1.4

TEMiss (Lbs.)

1.5

1.2

1.2

1.1

1

1

0.8 0.9

0.6 0.8

0.4 0.7

0.5

0.4

TBD

0.3

0.2

30

35

40

45

50 0.6

CEnet (Cents)

Figure 5.34: Application of AVCF to screen out the TOs with larger T EM iss values

5.3.2.2

Simulation for filtration using ASCF (step-3)

This step includes the formulation and application of a constraint filter based on the average surface fit for T EM iss. The average surface fit for T EM iss in terms of CEnet and T BD is generated using polynomial based regression for the TOs achieved after the application of AVCF presented in Table 5.4. Following are the software and hardware tools used to develop the surface fit and to formulate and apply the filter to validate the AVCF based filtration: 208


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM Machine: Core i7-4790 CPU @3.6 GHz with 16 GB of RAM Platform: MATLAB 2015a Regression model = Polynomial Polynomial surface model = Poly41 Method = Linear least square Normalize = off Robust = off ASCF formulation and application: average surf ace f it = sfit( CEnet , T BD) T EM iss Resid avgs = average surf ace f it − T EM iss Exclude = T EM iss Resid avgs < 0

Where average surf ace f it is the value of emission obtained through the average surface fit based polynomial for the respective CEnet and T BD tradeoff. And T EM iss Resid avgs is the decision element for the filter. The exclude option provided with the surface fit function has been used to screen out the TOs based on the formulation of the decision element. As per the formulation for T EM iss Resid avgs in this research, a tradeoff solution with a negative value of the decision element T EM iss Resid avg indicates the average surface fit for T EM iss. The application of ASCF thus screened out the TOs with higher values of T EM iss lying above the average surface fit for T EM iss. Various polynomial model fit options were coupled with the ASCF. The best model fit for the polynomials was achieved after comparison of the actual TOs for DRSREODLDG-based HEMS problem exhibited by various polynomial models ranging from Poly11 to Poly55. The tradeoff solutions harnessed through each polynomial based ASCF were analyzed for the average value of T EM iss and the number of diverse TOs harnessed for CEnet and T BD. Poly11 based ASCF achieved the minimum average T EM iss value of 0.58 Lbs.; however, the filter harnessed the least number of tradeoff solutions that did not include the desired 209


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM solutions like ones with CEnet value below 30 Cents. Poly12 based ASCF, on the other hand, included the TOs with minimal CEnet value less than 30 Cents; however, on the other hand, it lacked the diversification due to lesser number of tradeoff solutions. The options with the average T EM iss value equal or less than 0.59 were focused and poly41 was selected based on the lesser average values for T EM iss and T BD (0.59 Lbs. and 0.3) and more number of diverse solutions for TOs between CEnet/T BD. In this way, the model fit is based on an optimal set of the performance TOs for DRSREODLDG-based HEMS problem [56]. A summary comparing the performance of polynomial based ASCFs is given in Table 5.5 below. Table 5.5: A COMPARISON OF PERFORMANCE PARAMETERS FOR POLYNOMIAL BASED ASCF Regression Model Poly11 Poly12 Poly13 Poly14 Poly15 Poly21 Poly22 Poly23 Poly24 Poly25 Poly31 Poly32 Poly33 Poly34 Poly35 Poly41 Poly42 Poly43 Poly44 Poly45 Poly51 Poly52 Poly53 Poly54 Poly55

No. of Tradeoffs harnessed 29 29 32 31 33 34 35 35 35 34 32 36 37 35 35 33 33 32 33 33 33 35 37 34 35

Average CEnet (Cents) 37.01 36.98 37.48 37.66 38.49 38.41 37.46 37.46 37.01 36.95 37.62 37.82 37.52 37.67 37.51 38.01 38.47 37.56 37.57 36.59 37.67 35.78 35.6 36.32 36.16

Average T BD 0.31 0.31 0.31 0.31 0.3 0.3 0.31 0.31 0.31 0.31 0.31 0.31 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.31 0.31 0.32 0.33 0.31 0.31

Average T EM iss (Lbs.) 0.58 0.58 0.6 0.59 0.6 0.6 0.6 0.6 0.6 0.61 0.59 0.6 0.61 0.61 0.61 0.59 0.6 0.6 0.6 0.61 0.6 0.61 0.61 0.61 0.62

SSE

R2

0.28 0.21 0.18 0.17 0.17 0.19 0.17 0.17 0.17 0.16 0.19 0.17 0.17 0.16 0.13 0.19 0.17 0.16 0.16 0.13 0.18 0.15 0.14 0.13 0.13

0.05 0.31 0.41 0.42 0.44 0.35 0.42 0.42 0.42 0.48 0.37 0.43 0.43 0.47 0.55 0.37 0.44 0.45 0.47 0.55 0.4 0.51 0.53 0.56 0.56

The proposed polynomial model, poly41, for ASCF is based on the following formulation:

z(x, y) = p00 +p10 ×x+p01 ×y+p20 ×x2 +p11 ×x×y+p30 ×x3 +p21 ×x2 ×y+p40 ×x4 +p31 ×x3 ×y (5.2)

210


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM The proposed polynomial model is based on the coefficients (with 95% confidence bounds) as follows: p00 = 5.48 (-41.39, 52.35) p10 = -0.3234 (-4.894, 4.248) p01 = -9.079 (-77.88, 59.73) p20 = 0.00699 (-0.1564, 0.1704) p11 = 0.6176 (-5.337, 6.572) p30 = -4.498e-05 (-0.002571, 0.002482) p21 = -0.013 (-0.1842, 0.1582) p40 = -1.359e-08 (-1.426e-05, 1.423e-05) p31 = 6.749e-05 (-0.001563, 0.001698)

The eco-efficient solutions harnessed after the application of Poly41 surface filter are graphically shown in Figure 5.35. The final set of tradeoff solutions for ecoefficient operation of DRSREODLDG-based HEMS are tabulated as Table 5.6.

Table 5.6: ECO-EFFICIENT SOLUTIONS FOR DRSREODLDG-BASED HEMS USING ALGORITHM 2 (STEP-3) CEnet (Cents) 52.87 52.87 51.74 50.3 45.81 45.18 45.01 44.62 43.57 43.27 41.33 40.92 37.88 37.56 36.89 36.66 36.65 35.56 35.13 35.03 33.9 33.68 33.67 32.96 32.67 32.57 32.37 31.92 31.27 30.02 27.36 26.8 26.22

T BD 0.17 0.17 0.17 0.18 0.19 0.18 0.19 0.2 0.2 0.24 0.21 0.23 0.25 0.26 0.32 0.27 0.28 0.29 0.3 0.3 0.32 0.34 0.29 0.36 0.34 0.47 0.33 0.51 0.52 0.53 0.46 0.45 0.48

T EM iss (Lbs.) 0.67 0.67 0.67 0.67 0.57 0.6 0.56 0.6 0.56 0.56 0.57 0.56 0.6 0.56 0.56 0.6 0.56 0.57 0.6 0.57 0.57 0.56 0.65 0.56 0.56 0.56 0.65 0.56 0.56 0.56 0.65 0.65 0.65

P dl (kWh) 1.87 1.87 1.87 1.71 1.61 1.66 1.9 1.5 1.74 1.34 1.46 1.58 1.21 1.11 1.28 1.16 1.06 0.94 0.86 0.86 0.84 0.81 0.63 0.62 0.7 0.73 0.51 0.63 0.53 0.33 0.11 0.11 0.11

T s1

T s2

T s3

T s4

T s5

T s6

T s7

T s8

T s9

T s10

T s11

T s12

T s13

T s14

6 6 6 6 5 5 6 5 5 5 6 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 5 6 6 7 3 5 6

39 39 39 40 40 39 39 40 40 40 41 42 42 40 42 42 41 42 42 40 42 42 42 41 43 44 42 44 44 44 41 44 44

104 104 104 104 104 104 104 104 104 105 104 104 105 104 104 104 104 105 105 105 105 105 104 104 105 105 105 106 105 104 105 105 105

123 123 123 123 123 123 123 123 123 123 123 124 123 123 123 123 123 123 123 124 123 123 123 124 123 124 124 124 124 124 125 125 124

60 61 60 61 60 60 60 60 61 60 60 60 60 60 62 62 60 60 62 60 62 60 60 60 62 75 60 78 95 81 58 58 59

128 128 128 128 128 128 128 128 128 128 129 129 129 128 132 130 129 129 130 129 130 130 130 132 130 132 131 133 135 135 133 134 135

4 5 5 5 4 4 5 5 5 5 5 6 6 5 5 6 5 7 7 7 7 7 5 7 8 8 5 11 11 12 8 10 7

73 73 73 73 73 73 73 73 73 73 73 73 74 74 75 74 74 74 74 74 74 74 74 75 74 76 74 76 75 77 74 74 74

119 119 119 119 119 119 119 119 120 119 119 119 119 120 122 120 120 118 120 117 120 119 120 121 121 120 119 121 120 130 117 119 117

107 107 107 107 107 107 107 105 107 107 107 106 106 104 105 106 103 105 104 104 103 103 104 104 103 103 103 103 103 103 93 93 93

108 108 107 107 107 106 105 107 104 107 104 104 103 105 91 98 103 97 98 97 97 95 99 64 96 84 99 78 80 63 61 61 85

102 102 102 102 93 94 92 94 92 63 93 90 94 62 78 91 62 86 64 62 64 62 62 62 63 62 61 61 61 61 60 61 61

114 114 114 114 114 114 114 114 114 113 114 114 113 112 108 110 112 114 114 111 110 110 113 110 109 106 112 105 104 104 104 107 103

95 95 95 95 94 94 94 94 94 94 94 93 90 95 97 95 93 64 94 89 94 76 90 95 93 63 82 62 62 98 61 94 61

211


5.3. DRSREODLDG BASED HEMS OPERATION AND THE FILTRATION Chapter 5 MECHANISM Step-3 Avg. Surface Filteration: Selected eco-efficient solutions (under avg. surface for TEMiss). Filtered out/ excluded solutions (above avg. surface for TEMiss).

1

0.8

0.6

TEMiss (Lbs.)

0.4

0.5 0.2

0

0 -0.2

-0.4

0.4

-0.5 30

35

40

45

CEnet (Cents)

0.2 50

TBD

Figure 5.35: Eco-efficient solutions selected using average surface filtration based on algorithm 2 (Step-3)

5.3.3

Critical tradeoff analysis of solutions for eco-efficient DRSREODLDG-based HEMS operation

The final tradeoff solutions for eco-efficient HEMS operation harnessed through algorithm 1 and algorithm 2, are analyzed in this section for percentage reduction in CEnet, T BD, and T EM iss. The values of CEnet, T BD and T EM iss obtained without using the proposed method are 68.32 Cents, zero and 1.354 Lbs., respectively and the same have been used as base values in this analysis. For critical tradeoff analysis (CTA), the finalized TOs are classified for percentage reduction in CEnet, T BD and T EM iss as presented in Table 5.7. Following are the main features of the proposed classification: Class-I: In this class, the percentage reduction in CEnet ranges from 22.61 to 36.23, with the corresponding discomfort levels from 17% to 20%. The comfortconscious consumers are likely to opt this class due to minimal T BD and a reasonable reduction in CEnet. Maximal reduction in T EM iss ranging from 50.53% to 58.58% ensures eco-efficiency. Class-II: In this class, the percentage reduction in CEnet ranges from 36.67 to 52.18 with the corresponding discomfort levels from 21% to 36%. The reduction in T EM iss ranges from 51.72% to 58.58%. With the double-tailed polynomial 212


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO trend for T EM iss as shown in Figure 5.36, the class lies in the minimal range for emission. The class exhibits the best tradeoff solutions taking into account CEnet, discomfort and T EM iss. Most of the consumers are likely to choose this class for a fairly high welfare in terms of CEnet and the discomfort for the consumer with bottom minimal T EM iss. The class is regarded as the best for eco-efficiency. Class-III: In this class, the percentage reduction in CEnet ranges from 52.33 to 61.63 with the corresponding discomfort levels from 33% to 53%. Consumers opting for this class will achieve a maximum cost reduction of 61.63% (the largest in all classes) accompanied by a maximum discomfort level of 53% (also the largest in all classes). Consumers who are not conscious about the discomfort should choose this class to achieve the maximum possible reduction in CEnet. Maximal reduction in T EM iss ranging from 50.53% to 58.58% for this class ensures eco-efficiency. Further, the last three solutions in this class offer the maximum reduction in CEnet reaching upto 61.63% with a relatively low level of discomfort of value down to 45% as compared to the other members of this class. The consumers satisfied with these typical operating schemes may avail the maximum welfare through the largest reduction in CEnet at a relatively low level of discomfort. CTA given in Table 5.7 along with the respective scheduled times T st in Table 5.6 enables consumer to choose the best eco-efficient option in accordance with his needs after consulting a diverse set of available optimal choices for CEnet, T BD and minimal T EM iss.

5.4

Simulations for PDDR- and PDDR-RED- based HEMS using DWS-PSO

The simulations reported in this section were carried out to validate the performance of DWS-PSO based algorithm for optimal operations of PDDR- and PDDR-RED- based HEMSs. In order to achieve the TO solutions, WSM was 213


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO

Table 5.7: CRITICAL TRADEOFF ANALYSIS FOR ECO-EFFICIENT DRSREODLDG-BASED HEMS OPERATION Classes

I

II

III

CEnet (Cents) 52.87 52.87 51.74 50.30 45.81 45.18 45.01 44.62 43.57 43.27 41.33 40.92 37.88 37.56 36.89 36.66 36.65 35.56 35.13 35.03 33.90 33.68 33.67 32.96 32.67 32.57 32.37 31.92 31.27 30.02 27.36 26.80 26.22

Reduction in CEnet (%) 22.61 22.61 24.26 26.38 32.95 33.87 34.12 34.70 36.23 36.67 39.50 40.11 44.56 45.02 46.00 46.33 46.36 47.95 48.58 48.73 50.38 50.70 50.71 51.75 52.18 52.33 52.62 53.28 54.22 56.05 59.96 60.78 61.63

Range (%)

T BD 0.17 0.17 0.17 0.18 0.19 0.18 0.19 0.20 0.20 0.24 0.21 0.23 0.25 0.26 0.32 0.27 0.28 0.29 0.30 0.30 0.32 0.34 0.29 0.36 0.34 0.47 0.33 0.51 0.52 0.53 0.46 0.45 0.48

22.61 36.23

36.67 52.18

52.33 61.63

TBD

TEMiss

Range (%)

T EM iss (Lbs.) 0.67 0.67 0.67 0.67 0.57 0.60 0.56 0.60 0.56 0.56 0.57 0.56 0.60 0.56 0.56 0.60 0.56 0.57 0.60 0.57 0.57 0.56 0.65 0.56 0.56 0.56 0.65 0.56 0.56 0.56 0.65 0.65 0.65

17 - 20

21 - 36

33 53

Poly. (TBD)

Reduction in T EM iss (%) 50.53 50.53 50.53 50.53 58.11 55.74 58.58 55.74 58.58 58.58 58.11 58.58 55.74 58.58 58.58 55.90 58.58 58.19 55.90 58.19 58.19 58.58 51.72 58.58 58.58 58.58 51.72 58.58 58.58 58.27 51.72 51.72 51.72

Range (%)

50.53 58.58

51.72 58.58

51.72 58.58

Poly. (TEMiss)

0.80

0.70

R² = 0.6214

TEMiss (Lbs.), TBD

0.60

0.50

R² = 0.8931 0.40

0.30

0.20

0.10

0.00 0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

% Reduction in CEnet

Figure 5.36: Relation between % Reduction in CEnet, TBD and TEMiss for eco-efficient TOs

214


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO combined with the PSO. The SCF for DWS-PSO- based algorithm has been computed using Eq. 4.39. The SCF combines the OFs for the CE (CEnet) and the T BD through relative weights that impart priorities to the respective objectives. The simulation was carried out for a set of a critical pair of weights for CE and the T BD as follows:

W T V = [(1, 0), (0, 1)]

(5.3)

where W T V is a vector comprising pairs of weights to compute the required SCF to implement DWS-PSO. The first pair of weights provides the solution with the minimal value of CEnet based on optimal scheduling of SHAs, whereas the second pair provides the solution with the minimal value of T BD. To evaluate the algorithm’s performance, a set comprising eight number of TPs is designated. Four TPs are based on the diversified DPS implemented in various parts of the world. DPS including 2S-ToUP, 3S-ToUP, DA-RTP, and CPP are selected for implementation. In order to avoid the re-emergence of the peak load, each of the DPS has been combined with IBR. A factor of 1.4 has been applied as IBR for an energy consumption of above 0.4 kW/ slot. The remaining four TPs in the set are based on the modeling of SHAs for DS or MS. Problem based on any one of these approaches differ from its counterpart for the bounds laid down for the operation of SHAs. Taking the aforementioned bases into the account, the following set of TPs was proposed for the DPA of DWS-PSO- based algorithms for PDDR- as well as PDDR-RED- based HEMS: -A 2S-ToUP scheme (with DS) -A 2S-ToUP scheme (with MS) -A 3S-ToUP scheme (with DS) -A 3S-ToUP scheme (with MS) -A DA-RTP scheme (with DS) -A DA-RTP scheme (with MS) 215


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO -A CPP scheme (with DS) -A CPP scheme (with MS)

Detailed specifications for the SHAs, PV system, SB and inverter used to implement the algorithms for simulated operations of PDDR- and PDDR-RED- based HEMSs are given in Table 4.6 and Table 4.7. The hardware and software used for the simulations include the following: Machine: Core i7-4790 CPU @3.6 GHz with 16 GB of RAM Platform: MATLAB 2015a Optimization tool: Particle swarm optimization Swarm size: 140 Maximum iterations: 2800 Inertia range: 0.1-1.1

5.4.1

Simulations for PDDR- based HEMS using DWSPSO

The performance of the algorithm is analyzed for the TO solutions for CE and the T BD based on DWS-PSO. The SCF for WSM is achieved by combining the OFs of CE and T BD through the relative weights. In order to minimize the parameter for the CE a pair of weights of value (1, 0) was adopted reflecting a maximum priority to the reduction in CE. Whereas, in order to minimize the value of T BD a pair of weights of value (0, 1) was adopted that reflects a maximum priority to the reduction in T BD. Simulations were carried out for DPA of DWS-PSO algorithm for PDDR- based HEMS. The algorithm was tested for a set of TPs for HEMS based on 2S-ToUP (DS/ MS), CPP (DS/ MS), 3S-ToUP (DS/ MS) and DA-RTP (DS/ MS) for the TOs for CE and T BD. 216


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO The percentage reduction in CE based on the optimal HEMS operations is computed as follows: HEM S %RCE = (CEbase − CE)/(CEbase )

(5.4)

where CEbase is the base value of the CE for the unscheduled load scenario. The said values for each of the DPSs are presented in Table 5.8. The maximum reductions in CE along with the related T BD for PDDR- based HEMS using DWS-PSO with values of weights as (1, 0) are shown in Fig. 5.27. The simulation results of the proposed algorithm for the designated set of TPs are summarized in Table 5.8. The table furnishes the achieved performance parameters for the maximal reductions in the CE, the related values of T BD, and the peak load for the complete set of TPs. 70

60

50

40

30

20

10

0 2S-TOU (DS)

2S-TOU (MS)

3S-TOU (DS)

3S-TOU (MS)

DA-RTP (DS)

DA-RTP (MS)

CPP (DS)

CPP (MS)

-10

-20 Reduction in CE (%)

TBD (%)

Reduction in Peak load (%)

Figure 5.37: DWS-PSO for PDDR- based HEMS: A comparative performance for maximal reduction in CE using weights of (1,0)

The minimal values of T BD were achieved while selecting the value of weight (0, 1) for DWS-PSO. The algorithm minimized the value of T BD to zero and accordingly all of the loads were operated as per the preferred starting times (STslt) and

217


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Table 5.8: DWS-PSO FOR PDDR- BASED HEMS: REDUCTIONS IN CE, PEAK LOAD AND TBD FOR A DIVERSIFIED SET OF TPs Tariff Scheme

2S-TOU 3S-TOU RTP CPP

PDDR Base values of CE HEMS (Cents), Ppeak (kW) Approach and TBD DS MS DS MS DS MS DS MS

218.99; 0.608 and 1 206.85; 0.608 and 1 68.48; 0.608 and 1 327.44; 0.608 and 1

After PDDR- based Performance Achieved through HEMS Scheduling PDDR- based HEMS CE Ppeak Reduction in CE (%) TBD Reduction (Cents) (kW) (Performance Metric 1) (%) in Ppeak (%) 197.69 0.36 9.7 46.6 41.12 184.46 0.34 15.8 62.18 43.80 173.88 0.36 15.9 58.16 41.07 167 0.34 19.3 62.8 43.80 54.58 0.39 20.3 59.37 35.58 54.45 0.34 20.5 64.07 43.80 255.97 0.68 21.8 61 -11.02 220.46 0.34 32.7 65 43.80

ending times (ENslt) of HAs for DS and AS based models, respectively. Under this scenario percentage reductions in CE, peak load, and T BD remained zero.

5.4.2

Simulation results discussion for PDDR- based HEMS

The performance of the algorithm for reductions in CE, the related values of T BD, and Ppeak achieved for different TPS were analyzed while comparing the profile of the load before scheduling with the optimally scheduled profile achieved after the application of the algorithm. The performance of the algorithm for achieving the maximum reductions in CE through optimal scheduling was based on shifting of the load from the peak time (with higher electricity price (EP) towards the off-peak (with lower EP). The problem based on a 2-stage ToUP scheme was considered as a benchmark for analyzing the performance of the other TPs. The scheme consists of a price of 15 cents/ kWh during the peak time from 19:00 to 23:00 hours (slot numbered 115-138) and 9 cents/ kWh during the rest of the day as shown in Fig. 4.6. While moving from peak towards off-peak times the price coefficient reduces by 0.62 times. The algorithms use this price elasticity as a pressure for shifting of the peak load towards off-peak times. While pressurizing, the algorithm has to take into account the limiting bounds (ST slt and EN slt) and the LOOT of each of the SHAs. For DS scenario, the algorithm can shift the load from peak hours slots numbered 115-138 towards the off-peak hour’s slots numbered 139-144 in the 218


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO forward direction. Refer to Table 4.6, such shifting is applicable to HAs like AC-4, DW-2, CP, WM, WP, and IR. Further, some of the HAs are bounded to operate during the peak hours like EG-2 and RC-2, due to the specified values ST slt and EN slt, and cannot be shifted on a timeline; called as non-pressurized (NP) HAs. Similarly, the AS-type HAs can be pressurized for advanced shifting. The algorithm generates combinations of starting times of SHAs as vector Tst based on the constraint for each of the aforementioned types of HAs. The load profile corresponding to Tst vector is computed as per Eq. 4.28, and the same profile has been used to discuss the algorithm performance for a specified set of TPs.

The scenario for the 2S-ToUP scheme (with DS) showed that the algorithm attempted the DS- type HAs for their shifted operations in the forward direction in order to achieve an optimal load profile that might result in a minimum CE. The simulated load profile for this scenario is shown in Fig. 5.38. The algorithm curtailed the peak time load ranging from 0.35 to 0.6 kWh to a value of 0.15-0.35 kWh. A load of 0.31 kWh was shifted from the peak time towards the off-peak time slots numbered 139-144. That shifting of the load resulted in a 9.7% reduction in the CE under this scenario. 1 without scheduling DR (2S-TOU)- based DS, Wt

0.9

1,0

0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.38: Load profile with maximal reduction in CE for 2S-ToUP (DS)

219


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Refer to the scenario for the 2S-ToUP scheme (with MS), some of SHAs was modeled as AS while the others as DS. The algorithm attempted the DS- type HAs like AC 3, AC 4, DW 2, and CMP for their shifted operations towards the off-peak slots in the forward direction. Whereas, the AS- type HAs like WM, WP, EG 2, RC-2, and IR were pressurized for their shifted operations towards the off-peak slots in the advanced direction. The simulated load profile after MS of HAs is shown in Fig. 5.39. The algorithm curtailed the peak time load ranging from 0.35 to 0.6 kWh, to a value of 0.25 kWh supplied during slots numbered 133-138. The peak hour load was reduced because loads of 0.31 and 0.15 kWh were shifted towards off-peak time slots numbered 139144 (forward direction) and numbered 92-102 (advanced direction), respectively. This bi-directional shifting of the load out of the peak hours resulted in 15.8% reduction in the CE for this scenario. 1 without scheduling DR (2S-TOU)- based MS, Wt

0.9

1,0

0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.39: Load profile with maximal reduction in CE for 2S-ToUP (MS)

The scenario for the 3S-ToUP scheme (with DS) revealed that the algorithm attempted for the shifted operation of the SHAs from peak time slots numbered 103-102 towards the off-peak time slots numbered 133-144 in the forward direction. The simulated scheduled load profile for this scenario is shown in Fig. 5.40. The algorithm curtailed the peak time load ranging from 0.35 to 0.6 kWh to a 220


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO lower range of 0.15-0.3 kWh. A load of 0.33 kWh was shifted from the peak time towards the off-peak time. The said shifting under this scenario resulted in 15.9% reduction in CE. 1 without scheduling DR (3S-TOU)- based DS, Wt 1,0

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.40: Load profile with maximal reduction in CE for 3S-ToUP (DS)

Refer to the scenario for the 3S-ToUP scheme (with MS), the increased number of the pricing stages and multiple options for shifting of the load in the forward and in the advanced directions enabled the algorithm for a larger load shifting towards the slots with cheaper EP. Accordingly, the algorithm successfully shifted the peak time load, initially supplied during slots numbered 103-132, towards the off-peak time slots numbered 133-144 in the forward and the mid-peak/ off-peak time slots numbered 1-102 in the advanced direction. Further, the algorithm could move the mid-peak time load towards the off-peak time as well. The simulated load profile after MS of HAs for 3S-ToUP is shown in Fig. 5.41. The algorithm curtailed the peak time load ranging from 0.35-0.6 kWh to a lower range of 0.05-0.2 kWh; only the fixed and NP loads were fed during the peak hours. Loads of 0.31 and 0.2-0.3kWh supplied during peak hours were shifted towards offpeak slots numbered 133-144 in the forward and towards the mid-peak/ off-peak slots numbered 1-102 in the advanced directions respectively. This bi-directional 221


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO shifting of the load out of the relatively higher energy pricing slots resulted in 19.3% reduction in CE for the scenario of 3S-ToUP (MS). 1 without scheduling DR (3S-TOU)- based MS, Wt 1,0

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.41: Load profile with maximal reduction in CE for 3S-ToUP (MS)

The scenario for DA-RTP scheme (with DS) depicted that a larger number of pricing stages are provided to implement PDDR. The peak times EP ranges from 4.5-5.3 Cents/ kWh during slots numbered 85-132. To achieve an optimal value of CE, the algorithm shifted the peak time load towards the slots with relatively lower EP. As the model is based on DS, the peak load could only be shifted in the forward direction to reduce the CE. The simulated load profile for this scenario is shown in Fig. 5.42. The algorithm curtailed the peak hour load ranging from 0.35 to 0.6 kWh to a lower range of 0.1-0.2 kWh. The load pertaining to NP type HAs like AC-3, WM and RC remained for supplying during the peak hours. A load of 0.3-0.35 kWh was shifted from the peak towards the off-peak hours in the forward direction. The DS-based shifting under RTP resulted in 20.3% reduction in CE. In scenario DA-RTP scheme (with MS), the EP changes on an hourly basis. Such a model for DEP provides a larger number of pricing stages to implement PDDR. The multi-stage scheme combined with MS for HAs introduces diversified and larger number of options for shifting of the load in the forward as well as in the 222


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 without scheduling DR (RTP)- based DS, Wts 1,0

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.42: Load profile with maximal reduction in CE for DA-RTP (DS)

advanced directions. The simulated load profile after MS of HAs for DA-RTP is shown in Fig. 5.43. The algorithm curtailed the peak hours load ranging from 0.35-0.6 kWh to a lower range of 0.2-0.3 kWh; only the fixed and NP-type loads like AC-3, WM and RC were fed during the peak hours. The algorithm shifted most of the load from the peak time slots numbered 85-132 towards the slots with lesser EP in the forward direction (slots numbered 133-144) and in the advanced direction (slots numbered 11-60). The said shifting towards the time slots with diversified and more reduced EPs resulted in 20.5% reduction in CE under this scenario. The CPP is an event-based scheme that is used to manage extra critical power demands. The scheme for CPP is based on extra high electricity prices during the highly critical peak hours. Such prices motivate the consumers to shift their load from the extra critical peak times towards the off-peak times. The scenario for CPP scheme (with DS) showed that the algorithm attempted the HAs supplied during the critical peak times for their shifted operations towards the off-peak time in the forward direction. The simulated load profile for this scenario is shown in Fig. 5.44. The algorithm curtailed the peak time load ranging from 0.35 to 0.6 kWh to a lower range of 0.05-0.18 kWh. The curtailed peak time load was shifted 223


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 without scheduling DR (RTP)- based MS, Wt 1,0

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.43: Load profile with maximal reduction in CE for DA-RTP (MS)

towards the off-peak slots numbered 139-144 in the forward direction. A load shift in the forward direction combined with a very large reduction in the pricing co-efficient (0.32 times) while moving from critical-peak towards off-peak time resulted in 21.8% reduction in CE. However, the peak load was increased due to larger shifting of the demand towards the limited one hour off-peak period in the forward direction. The proposed value of 1.4 times of the normal EP for IBR was not sufficient in that case to avoid the re-emergence of the peak load. 1 without scheduling DR (CPP)- based DS, Wt1,0

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.44: Load profile with maximal reduction in CE for CPP (DS)

224


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Refer to the scenario for CPP (with MS), some of the SHAs are modeled as AS while the others as DS. The algorithm can shift the operations of peak time loads like AC 3, AC 4, DW 2, and CMP towards the off-peak slots numbered 139-144 in the forward direction. Whereas, the operations of loads like WM, WP, EG 2, RC-2, and IR can be shifted towards the off-peak slots numbered 1-114 in the advanced direction. The simulated load profile after MS of HAs is shown in Fig. 5.45. The algorithm curtailed the peak hours load range of 0.35-0.6 kWh to a lower value of 0.22 kWh supplied during slots numbered 133-138. The load curtailed out of the peak time load was shifted towards off-peak time slots numbered 139-144 in the forward and slots numbered 92-102 in the advanced directions. This bi-directional shifting of the load out of the peak time and the related very large reduction in the pricing co-efficient resulted in 32.7% reduction in CE for this scenario. Further, MS-based load shifting successfully avoided the emergence of the peak load as well. 1 without scheduling DR (CPP)- based MS, Wt

0.9

1,0

0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.45: Load profile with maximal reduction in CE for CPP (MS)

225



5.4.3

A DPA of DWS-PSO algorithm for PDDR- based HEMS

Two metrics were established for DPA of DWS-PSO for PDDR- based HEMS. The first metric pertains to the maximum reduction in the CE. The said metric for the diversified set of TPs have been furnished in Table 5.8. The second metric, the gradient of the TO line between percentage reduction in CE and T BD, is defined for the responsiveness of the algorithm for the reduction in the CE while increasing the value of T BD. The second metric was computed while drawing the TO solutions for the two parameters as follows: (i) The solution for the maximal reduction in the CE was computed by selecting weights of (1, 0) for the CE and T BD respectively while minimizing the SCF given in Eq. 4.39. (ii) The solution for a minimal T BD was achieved while selecting values of weights as (0, 1) for the respective objectives of CE and T BD. These values of weights allocated a maximum priority to the minimization of the value of T BD while computing the SCF. Under this scenario, T BD is minimized to zero and accordingly, all of the loads are operated as per the preferred starting times (ST slt) and ending times (EN slt) of HAs for DS and AS, respectively as shown in Fig. 5.46. The percentage reductions in CE, peak load, and T BD remained zero. The aforementioned two scenarios for maximal and minimal reductions in CE and the corresponding TBDs are reflected in Fig. 5.47. The information revealed by Fig. 5.47 are summarized in Table 5.9. The tabulated data was used for DPA of the DWS-PSO algorithm for PDDR- based HEMS. In this regards, the following conclusions were made: (i) When tested for the diversification in the modeling of SHAs, the algorithm showed better performance for MS- based HEMS as compared to the ones based on DS. The MS- based HEMSs outperformed for metric 1 and 2. 226



1 without scheduling DR (2S-TOU)-, (3S-TOU)-, RTP-, and CPP- based, DS/ MS, Wts 0,1

0.9 0.8

Pgd (kWh)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.46: Load profile with minimal value of TBD with weights = (0, 1)

35 y = 0.5026x

30

25 y = 0.3578x y = 0.3419x y = 0.32x

20

% Reduction in CE

y = 0.3067x y = 0.2741x y = 0.2536x

15

y = 0.2087x

10

5

0 0

10

20

30

40

50

60

70

-5 TBD ( %) 2S-TOU (DS)

2S-TOU (MS)

3S-TOU (DS)

3S-TOU (MS)

DA-RTP (DS)

DA-RTP (MS)

CPP (DS)

CPP (MS)

Figure 5.47: Tradeoffs between CE and TBD for PDDR-based HEMS

227


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Table 5.9: PERFORMANCE METRICS OF DWS-PSO- BASED ALGORITHM FOR PDDR-BASED HEMS Tariff Scheme


HEMS Approach

DS MS DS MS DS MS DS MS

Reduction in CE (%) 0 ∼09.73 0 ∼15.77 0 ∼15.94 0 ∼19.26 0 ∼20.30 0 ∼20.50 0 ∼21.83 0 ∼32.67

Performance metrics for PDDR-based Average reduction in CE per unit of TBD (Performance metric 2) 0 ∼46.60 0.2087 0 ∼62.18 0.2536 0 ∼58.16 0.2741 0 ∼62.80 0.3067 0 ∼59.37 0.3419 0 ∼64.07 0.32 0 ∼61.00 0.3578 0 ∼65.00 0.5026 Range of TBD (%)

HEMS Relative value of performance metric 2 1 1.22 1.31 1.47 1.64 1.53 1.71 2.41

(ii) When tested for diversification for DPSs, the algorithm exhibited the best performance for TP based on CPP. The metric 1 stood at a value of 32.67% for a reduction in the CE. Whereas, the metric 2 showed a value of 0.5026 kWh that was 2.41 times greater than the corresponding benchmark value of 0.2087 kWh for the TP for 2S-ToUP (DS). This highest value of metric 2 for CPP based TPs seems sufficient to motivate the consumers to participate in PDDR for critical peak reduction. The scheme, however, has event-based application and is not available on daily basis for PDDR. (iii) The algorithm could not avoid the peak rebounds due to the lesser value of IBR ratio for CPP (DS)- based TPs. The algorithm is proposed to be used with CPP (MS)- based scheme for achieving the highest performance for metric 1 and 2 and avoidance of the peak rebounds simultaneously. (iv) When tested for the diversification for DPSs used on daily basis, the performance of the algorithm was ranked 1-6 for DA-RTP (MS), DA-RTP (DS), 3S-ToUP (MS), 3S-ToUP (DS), 2S-ToUP (MS), 2S-ToUP (DS) based TPs respectively. This indexing of the algorithm for DPA is based on the proposed metrics 1 and 2. (v) The algorithm showed the best performance for CPP (MS) and DA-RTP (MS) for the event-based and daily based pricing schemes respectively.

228



5.4.4

Simulations for PDDR-RED based HEMS using DWSPSO

Simulations were also carried out for DPA of DWS-PSO algorithm for PDDRRED- based HEMS. The algorithm was tested on a set of TPs for HEMS for 2S-ToUP (DS/ MS), CPP (DS/ MS), 3S-ToUP (DS/ MS), and DA-RTP (DS/ MS).

The algorithm was run for a TO solution for maximum reduction in CEnet using respective weights of values (1, 0) for the objectives of CEnet and T BD while minimizing the SCF given in Eq. 4.39. Further, the TO solution for minimal T BD was also achieved while selecting weights of values (0, 1) for the respective objectives of CEnet and T BD. Both of these TO solutions were simulated for DPA of the DWS-PSO algorithm for PDDR-RED- based HEMS. The capability of DWS-PSO for maximum reductions in CEnet along with the related T BD for PDDR-RED- based HEMS, while using weights of values (1, 0), is reflected in Fig. 5.48. In order to demonstrate the performance of the proposed algorithm for the designated set of TPs the results of the simulation for a maximal reduction in the CEnet using the weights of values (1, 0) are summarized in Table 5.10. The table furnishes the achieved performance parameters for the maximal reductions in the CEnet, and the corresponding values of T BD and peak load for the complete set of TPs. The capability of DWS-PSO for achieving a minimal value of T BD for PDDRRED- based HEMS was also analyzed. Under this scenario, the weights of values (0, 1) were assigned to CEnet and T BD while computing SCF through Eq. 4.39. The algorithm minimized the value of T BD to zero and accordingly all of the loads were operated as per the preferred starting times (ST slt) and ending times (EN slt) of HAs for DS and AS, respectively. However, contrarily to the same 229


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 140

120

100

80

60

40

20

0 2S-TOU (DS)

2S-TOU (MS)

3S-TOU (DS)

3S-TOU (MS)

DA-RTP (DS)

DA-RTP (MS)

CPP (DS)

CPP (MS)

-20 Reduction in CE (%)

TBD (%)


Figure 5.48: DWS-PSO for PDDR-RED- based HEMS: A comparative performance for maximal reduction in CEnet using weights of (1,0) Table 5.10: DWS-PSO FOR PDDR-RED- BASED HEMS: MAXIMUM REDUCTIONS IN CE, PEAK LOAD AND TBD FOR DIVERSIFIED TPs USING WEIGHTS (CE, TBD) = 1,0 Dynamic HEMS Pricing Approach Scheme



Without PDDR-REDbased Operation

With PDDR-RED- based HEMS Operation

CEnet (Cents), CEnet CEgrid CEsold Ppeak (kW) and TBD (Cents) 218.99; 0.608 and 0 206.85; 0.608 and 0 68.48; 0.608 and 0 327.44; 0.608 and 0

103.86 72.49 83.38 58.83 23.00 18.58 115.15 72.49

70.54 54.30 67.95 52.93 29.94 27.07 70.54 54.30

33.33 18.19 15.43 5.89 -6.94 -8.49 44.61 18.19

Ppeak (kW) 0.39 0.33 0.36 0.33 0.39 0.34 0.68 0.33

Performance achieved with PDDR-RED- based HEMS operation Reduction in TBD Reduction in CEnet (%) (%) Ppeak (%) (Per. metric 1) 84.8 58.37 35.53 91.7 62.04 45.23 92.5 60.25 41.12 97.2 58.17 45.23 110.1 73.1 35.58 112.4 61.39 43.80 86.4 59.1 -11.02 94.4 63.7 45.23

scenario for PDDR- based HEMS, a reduction of CEnet was achieved as shown in Fig. 5.49. This reduction was based on the optimal dispatch of the existing power sources at home, including PV and SB system for supplying a non-scheduled (fixed) load corresponding to a zero value of T BD. The strategy for optimal dispatch had already been embedded in the algorithm-2. The RED- based strategy, as a part of algorithm-2, was primarily based on (a) preferred direct usage of the energy from the PV system (b) storage of the excess energy from the PV into the SB during off-peak hours (c) discharge of the SB during peak hours to supply the load (d) 230


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO parallel operation of the SB with the power grid under the limiting conditions for SOCG and the discharge rates (e) selling of the excess energy to the grid. The proposed scheme was designed for the optimal dispatch of the PV, SB, and the power grid for supplying of the consumer load and selling back the extra PV energy to the utility in order to minimize the CEnet. 120

100

80

60

40

20

0 2S-TOU (DS)

2S-TOU (MS)

3S-TOU (DS)

3S-TOU (MS)

Reduction in CE (%)

TBD (%)

DA-RTP (DS)

DA-RTP (MS)

CPP (DS)

CPP (MS)


Figure 5.49: DWS-PSO for RED- based HEMS: A comparative performance at minimal TBD using weights of (0,1)

The simulation results for RED- based HEMS using weight values (0, 1) for a minimum value of T BD are summarized in Table 5.11. The table includes the related parameters for the reduction in CEnet, the T BD, and the peak load achieved under the specified scenario. The achieved parameters demonstrated the performance of the algorithm for the proposed dispatch strategy of the sources including the PV system, the SB, and the power grid while retaining all of the loads at their un-scheduled positions. The algorithm achieved the maximum reductions of 96.6% in the CEnet for the TP based on DA-RTP when applied merely for the optimal dispatch of the sources without opting any PDDR based shifting of HAs. A detailed discussion on the

231


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Table 5.11: DWS-PSO FOR RED- BASED HEMS: REDUCTIONS IN CE AND PEAK LOAD FOR DIVERSIFIED TPs USING WEIGHTS (CE, TBD) = (0, 1) FOR MINIMAL TBD Without PDDR-REDWith PDDR-RED- based Performance achieved with Tariff HEMS based HEMS Operation HEMS Operation PDDR-RED- based HEMS Operation Scheme Approach Reduction CEnet (Cents), CEgrid CEsold CEnet Ppeak Reduction in CEnet (%) TBD Ppeak (kW) and TBD (Cents) (Cents) (Cents) (kW) in Ppeak (%) (Per. metric 1) DS 124.78 77.53 47.25 0.33 78.4 0 45.23 2S-TOU 218.99; 0.608 and 0 MS 124.78 77.53 47.25 0.33 78.4 0 45.23 DS 115.25 75.44 39.82 0.34 80.7 0 43.80 3S-TOU 206.85; 0.608 and 0 MS 115.25 75.44 39.82 0.34 80.7 0 43.80 DS 34.36 32.06 2.30 0.34 96.6 0 43.80 RTP 68.48; 0.608 and 0 MS 34.36 32.06 2.30 0.34 96.6 0 43.80 DS 174.71 77.53 97.18 0.33 70.3 0 45.23 CPP 327.44; 0.608 and 0 MS 174.71 77.53 97.18 0.33 70.3 0 45.23

application of the algorithm for the complete set of TPs for RED and PDDR-REDbased HEMSs is carried out in the next section.

5.4.5

Simulation results discussion for PDDR-RED based HEMS

In this section, the performance of the algorithm is discussed for solving PDDRRED- based HEMS problems for the proposed set of TPs. Further, the algorithm is used to solve HEMS problems with respective weights of (1, 0) and (0, 1) for the TO solutions with a maximal reduction in the CEnet and a minimal (zero) value of T BD, respectively. The latter scenario led to a RED- based operation of HEMS based on the implementation of dispatch strategy of the power sources while keeping the loads at their un-scheduled positions. The TP for 2S-ToUP (non-shifted) was solved with weights of (0, 1) for a minimal T BD. The resulting energy profiles are shown in Fig. 5.50. Under this scenario, the SB was charged through the PV energy available during off-peak slots numbered 45-63. The SB transported that energy towards the peak hours and supplied some of the peak time load during slots numbered 115-126. However, due to the limited discharge rates and the capacity of the SB, the load during peak time slots numbered 113-118 and 126-138 had to be supplied through the grid at higher prices. The RED- based operation resulted in 78.42% reductions in the 232


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO CEnet based on the 218 Cents per day for the 2S-ToUP scheme in PDDR- based HEMS. The load shifting was not taken into accounts under this scenario and the reduction in the CEnet is based on the optimal dispatch strategy implemented for the PV/ SB system (installed at a home) and the power grid. 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.50: Load profile for minimal TBD using 2S-ToUP (Non-shifted) for PDDR-RED- based HEMS

Refer to TP for 2S-ToUP (DS), which was solved with weights of (1, 0) for a maximal reduction in CEnet. The resulting energy profiles are depicted in Fig. 5.51. The SB was charged with the PV energy during off-peak slots numbered 45-63. Based on DS, the peak time load is shifted forward towards slots numbered 138-144. The remaining un-shifted load lying within the peak-hours was supplied through the SB. However, due to the limited capacity of the SB, 0.37kWh of the load was supplied through the grid during peak time slots numbered 136-138. The algorithm, while combining the effect of PDDR- based DS with the optimal dispatch of the PV system, the SB, and the grid has achieved a reduction of 84.78% in CEnet for 2S-ToUP with DS. The results for TP for 2S-ToUP (MS) with weights of (1, 0) is shown in Fig. 5.52 for a maximal reduction in CEnet. The SB was charged with the PV energy available during off-peak slots numbered 45-70. Based on MS, some of the peak 233


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.51: Load profile for maximal reduction in CE using 2S-ToUP (DS) for PDDR-RED- based HEMS

time load was shifted forward towards the off-peak time slots numbered 138-144, whereas the rest of the load was shifted in the advanced direction towards the offpeak slots numbered 45-115. The AS load was supplied through the PV system directly in order to achieve a higher reduction in the CEnet. The remaining unshifted peak hours load was completely supplied through the SB that was charged through the PV units ahead in time. The algorithm achieved a 91.69% reduction in the CEnet for TP based on 2S-ToUP (MS). The TP for CPP solved with weights of (0, 1) for minimal T BD are shown in Fig. 5.53. As already discussed that the CPP scheme used in the present simulations is based on 2S-ToUP, however, the peak price for this scheme is double from that in 2S-ToUP. As compared to the peak/ off-peak prices of 15/ 9.25 Cents/kWh in 2S-ToUP scheme, a set of prices of 30/ 9.25 Cents/kWh is adopted for the CPP scheme as per the criteria mentioned by San-Diego electric supply company. Under the proposed CPP scenario, the SB was charged through the extra PV energy during slots numbered 45-63. The energy stored in the SB was used to supply the peak time load during slots numbered 115-126. However, due to the limited discharge rate and the capacity of the SB, the battery could not supply 234


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 Psch (kW/slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.52: Load profile for maximal reduction in CE using 2S-ToUP (MS) for PDDR-RED- based HEMS

all of the peak time load completely. Consequently, a large amount of the load was supplied through the grid at higher prices during peak times slots numbered 113-118 and 126-138. The RED- based operation resulted in 70.32% reductions in the CEnet. A bit lesser reduction in CEnet in this scenario as compared to the 2S-ToUP is due to the supply of a large amount of the load at double the price during peak hours in this scenario. The reduction in CEnet in this scenario is due to the optimal dispatch strategy implemented for the PV, SB and the grid without combining the PDDR- based shifting of the load. The energy profiles for CPP (DS)- based HEMS are shown in Fig. 5.54 for a maximal reduction in CEnet having weights of value (1, 0). The SB was charged with the PV energy during off-peak slots numbered 38-64. Based on DS, the peak time load was shifted forward towards slots numbered 138-144. The peak time load that remained un-shifted was supplied through the SB. However, due to the limited capacity of the SB, a load of 0.37 kWh was supplied through the grid during peak time slots numbered 136-138. The rest of the peak time load was shifted towards the off-peak slots numbered 139-144 in the forward direction. Due to the limited number of off-peak slots in the forward direction and for the large 235


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.53: Load profile for minimal TBD using CPP (Non-shifted) for PDDR-RED- based HEMS

peak to off-peak prices ratio, the load shifted in the forward direction exceeded the limiting value of 0.4 kWh in spite of the implementation of IBR ratio. So a large peak load was observed during off-peak slots numbered 139-144 in this scenario. That high peak load can be managed by opting MS based scheduling or by increasing the values of IBR for these specific slots. The algorithm, while combining the effect of PDDR- based DS with the optimal dispatch of the PV system, the SB, and the grid, achieved a reduction of 86.38% in CEnet for CPP (DS)- based TPs. Refer to TP for CPP (MS) solved with weights of (1, 0) for a maximal reduction in CEnet, the resulting energy profiles are shown in Fig. 5.55. The SB was charged through the extra PV energy during slot numbered 45-70. Based on MS, some of the peak time load was shifted forward towards the off-peak slots numbered 138144, whereas the rest of the load was shifted in the advanced direction towards the off-peak slots numbered 45-115. The advanced shifted load was supplied through the PV system directly to reduce the CEnet. The remaining un-shifted peak time load was completely supplied through the SB. The algorithm achieved a 94.44% reduction in the CEnet for TP based on CPP (MS). 236



1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.54: Load profile for maximal reduction in CE using CPP(DS) for PDDR-RED- based HEMS

1 Psch (kW/slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.55: Load profile for maximal reduction in CE using CPP(MS) for PDDR-RED- based HEMS

237


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO The TP for 3S-ToUP (non-shifted) solved with weights of (0, 1) for minimal T BD showed that the SB was charged with the PV energy available during off-peak time slots numbered 45-63, as shown in Fig. 5.56. The energy stored in the SB during off-peak hours was used to supply the peak time loads during slots numbered 103-117. However, due to the limited discharge rate and the capacity of the SB, the load during peak time slots numbered 117-132 was supplied through the grid at higher prices. The RED- based operations resulted in 80.75% reduction in the CEnet based on the optimal dispatch strategy implemented for the PV/ SB system and the power grid. The scenario did not take the shifting of the load into consideration. 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.56: Load profile for minimal TBD using 3S-ToUP (Non-shifted) for PDDR-RED- based HEMS

Refer to TP for 3S-ToUP (DS) solved with weights of (1, 0) for a maximal reduction in CEnet, the resulting energy profiles are shown in Fig. 5.57. The SB was charged with the extra PV energy available during off-peak slot numbers 42-66. Based on DS, the peak time load is shifted forward towards slots numbered 138-144. The remaining un-shifted load lying within the peak-hours was supplied through the SB. However, due to the limited capacity of the SB, loads of values 0.18 and 0.37 kWh were supplied through the grid during peak time slots numbered 126-132 238


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO and 133-138 respectively. The algorithm, while combining the effect of DS with the optimal dispatch of the PV/SB system and the grid, achieved a reduction of 92.54%.in CEnet. 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.57: Load profile for maximal reduction in CE using 3S-ToUP (DS) for PDDR-RED- based HEMS

The TP for 3S-ToUP (MS) is solved with weights of (1, 0) for a maximal reduction in CEnet is revealed in Fig. 5.58. The SB was charged with the PV energy available during off-peak slots numbered 45-70. Based on MS, some of the peak time load was shifted forward towards the off-peak slots numbered 138-144. The rest of the load was shifted in the advanced direction towards the mid peak slots numbered 45-102 where it was supplied through the PV system directly to reduce the CEnet. The remaining un-shifted peak time load was completely supplied through the already charged SB. The algorithm achieved a 97.15% reduction in the CEnet for the designated TP. The energy profiles are shown in Fig. 5.59 for RTP (non-shifted) solved with weights of (0, 1). Under this scenario, the SB was charged with the PV energy during slot numbers 45-63. The stored energy of the SB was used to supply the load when the price of energy was above its average value (more than 2.5 cents per kWh) during slots numbered 102-120. However, due to the limited discharge rate 239


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 1 Psch (kW/slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.58: Load profile for maximal reduction in CE using 3S-ToUP (MS) for PDDR-RED- based HEMS

and the capacity of the SB, the load during the slots numbered 121-132 was to be supplied through the grid at peak times pricing of 4.5 Cents/ kWh. The REDbased operation resulted in 96.64% reduction in the CEnet based on the optimal dispatch strategy implemented for the PV, SB and the grid for a fixed load. 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.59: Load profile for minimal TBD using DA-RTP (Non-shifted) for PDDR-RED- based HEMS

Refer to TP for DA-RTP (DS) solved with weights of (1, 0) for a maximal reduction 240


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO in CEnet, the resulting energy profiles are shown in Fig. 5.60. The SB was charged through the extra PV energy available during off-peak slot numbers 4266. Based on DS, the peak time load was shifted forward towards lower pricing slots numbered 138-144. The un-shifted load that remained within the peak-hours was supplied through the SB. However, due to the limited capacity of the SB, loads of values 0.18 and 0.37 kWh were supplied through the grid during peak time slots numbered 126-132 and 133-138 respectively. The algorithm, while combining the effect of DS with the optimal dispatch of the power sources, achieved a reduction of 92.54%.in CEnet. 1 Psch (kW/ slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/ slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.60: Load profile for maximal reduction in CE using DA-RTP (DS) for PDDR-RED- based HEMS

The TP for DA-RTP (MS) solved with weights of (1, 0) for a maximal reduction in CEnet are shown in Fig. 5.61. The SB was charged during slots numbered 43-64. Based on MS, some of the peak time load was shifted forward towards the slots numbered 138-144 with below averaged electricity prices, whereas most of the load was shifted in the advanced direction towards the lower pricing slots numbered 1-96. The load shifted in the advanced direction was supplied through the PV system directly. The NP- type load that could not be shifted out of the peak time was completely supplied by the SB during slots numbered 99-118. The 241


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO algorithm achieved a maximal value of 112.40% for the reduction in CEnet. The algorithm showed the best performance for DA-RTP (MS)- based HEMS for a maximal reduction in the CEnet. 1 Psch (kW/slot) Ppv Pgd Pds Pch Psold SOCG of SB(%)

0.9 0.8

Power (kW/slot)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

9

18 27 36 45 54 63 72 81 90 99 108 117 126 135 144

Time slots

Figure 5.61: Load profile for maximal reduction in CE using DA-RTP (MS) for PDDR-RED- based HEMS

5.4.6

A DPA of DWS-PSO algorithm for PDDR-REDbased HEMS

Two metrics were established for DPA of DWS-PSO for PDDR-RED- based HEMS. The metric 1 is composed of the maximum reduction in the CEnet achieved by the algorithm. Whereas, the metric 2 comprised the gradient of the TO line for the percentage reduction in CEnet and T BD. The later one signifies the responsiveness of the model for the reduction in CE while increasing the value of T BD. The maximal reduction in the CEnet was achieved using the values of the weights as (1, 0) for CEnet and T BD respectively while solving the TP. These maximal reductions in the CEnet, representing metric 1, achieved for the diversified of TPs are presented in Table 5.10. Further, the minimal reduction in CEnet was also 242


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO achieved using the values of weights as (0, 1). These weights assign maximum priority to the minimization of the T BD. Under this scenario, the algorithm minimized the T BD to zero and the corresponding (minimal) reduction in CEnet was achieved. This reduction of CEnet, as furnished in Table 5.11, is based on the RES, the ESS and the grid optimal dispatch (RED) while supplying a nonshifted (fixed) load. The maximum and minimum values of the CEnet and the related values of T BD achieved for the diversified set of TPs are furnished in Table 5.10 and Table 5.11 and drawn for averaged TO lines as shown in Fig. 5.62. The gradient of the TO line for the percentage reduction in CEnet and the corresponding T BD represents metric 2. This metric has special significance in respect to the responsiveness of the model for PDDR. 115 y = 0.2566x + 96.643

110 y = 0.1846x + 96.643

105

100 y = 0.282x + 80.749

% Reduction in CEnet

y = 0.3787x + 70.321

95 y = 0.1957x + 80.754 y = 0.2139x + 78.425

90 y = 0.2717x + 70.321

85

y = 0.1089x + 78.424

80

75

70

65 0

2S-TOU (DS)

10

2S-TO (MS)

20

3S-TOU (DS)

30

40 TBD(%)

3S-TOU (MS)

50

RTP (DS)

60

RTP (MS)

70

CPP (DS)

80

CPP (MS)

Figure 5.62: Tradeoffs between CEnet and TBD for PDDR-RED- based HEMS

The information gathered from Fig. 5.62 are summarized in Table 5.12. Table 5.12 was used for DPA of the DWS-PSO algorithm for PDDR-RED- based HEMS. The results furnished in the table are analyzed as follows: (i) RED- based HEMS operation using (0, 1) weights: In this approach, the consumer can avail a substantial reduction in CEnet even without shifting 243


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO Table 5.12: PERFORMANCE METRICS OF DWS-PSO FOR PDDR-REDBASED HEMS Tariff Scheme

HEMS Approach

Reduction in CEnet with RED only (%)


78.42 78.42 80.75 80.75 96.64 96.64 70.32 70.32


Performance metrics for PDDR-RED- based HEMS Average reduction in Relative value Range CEnet per unit of TBD of performance of TBD (%) (Performance metric 2) metric 2 78.42 ∼84.78 0 ∼58.37 0.109 1.00 78.42 ∼91.69 0 ∼62.04 0.214 1.96 80.75 ∼92.54 0 ∼60.25 0.196 1.80 80.75 ∼97.15 0 ∼58.17 0.282 2.59 96.64 ∼110.13 0 ∼73.10 0.185 1.70 96.64 ∼112.40 0 ∼61.39 0.257 2.36 70.32 ∼86.38 0 ∼59.10 0.272 2.50 70.32 ∼94.44 0 ∼63.70 0.379 3.48 Reduction in CEnet (%)

his load as per the results furnished in Table 5.11. The reduction in CEnet is achieved by simply implementing the optimal strategy for the dispatch of the photovoltaic system, the storage unit, and the power grid for a fixed load. For RED- based HEMS operation, the algorithm showed the best result for DA-RTP based TP while achieving a value of 96.64% for the reduction in the CEnet. This higher value indicates that a larger amount of the unscheduled load lying within the peak/ relatively higher pricing slots was supplied by the PV/ SB system. Refer to Fig. 5.61, the y-intercept in the graph indicates the initial contribution for the reduction in CEnet from the RED that can be achieved without facing any T BD. (ii) PDDR-RED- based HEMS operation using (1, 0) weights: Refer to Table 5.12, column 4, the algorithm successfully reduced the CEnet for all of the TPs. (iii) The algorithm showed better performance for MS- based as compared to the DS- based HEMSs for minimizing the CEnet (metric 1). Further, refer to column 7, MS-based outperformed their counterparts for DS- based HEMSs for the increase in the reduction of CEnet per unit of the T BD (metric 2). (iv) Performance of the algorithm for a diversified set of TPs: The algorithm exhibited the best performance for the HEMS problem based on DA-RTP while achieving maximum reductions in the CEnet of value 110.13% for DSand 112.40% for MS- based models. Although the second metric for the reductions in CEnet per unit of T BD is a bit lower in this approach, a very high value of the initial contribution in the reduction of CEnet (96.64%) 244


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO from the RED makes this scheme the best for its application for HEMS. The algorithm while solving HEMS problems for RTP simultaneously reduced the values of Ppeak as well by 35.58 and 43.80% for DS and MS-based models respectively. (v) The DWS-PSO- based algorithm was ranked 1-8 for its performance for metric 1 for the proposed TPs as: DA-RTP (MS), DA-RTP (DS), 3S-ToUP (MS), CPP (MS), 3S-ToUP (DS), 2S-ToUP (MS), CPP (DS), and 2S-ToUP (DS). The proposed indexing is based on metric 1 whereas the respective values of metric 2 for the individual TPs signifies the responsiveness of the model for PDDR. (vi) The algorithm achieved a much higher value of metric 1 for CPP- based HEMS as compared to its fundamental model for 2S-ToUP. A reduction in CEnet of value 94.44% was achieved for CPP (MS)- based as compared to the respective value of 91.69% for the 2S-ToUP (MS)- based HEMS. Further, CPP- based HEMS, with the maximum relative value of 3.48 for the metric 2, validated their maximal responsiveness for a reduction in CEnet with the increase in T BD. That demonstrate the ability of the CPP (MS)- based models for motivating the consumers to participate in PDDR for reducing the event based critical peaks.

5.4.7

Summary of DWS-PSO results

Following is the summary of the results of DPA for DWS-PSO- based algorithm when tested for a designated set of TPs for PDDR- based HEMS: (i) The respective performance metrics for the algorithm in percent when tested for the HEMS problems for 2S-TOU (DS); 2S-TOU (MS); 3S-TOU (DS); 3STOU (MS); DA-RTP (DS); DA-RTP (MS);CPP (DS); and CPP (MS) are 9.7; 15.8; 15.9; 19.3; 20.3; 20.5; 21.8; and 32.7 for metric 1 and 0.2087;

245


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO 0.2536;0.2741;0.3067;0.3419;0.32;0.3578; and 0.5026 for metric 2. The relative values of metric 2, taking 2S-TOU (DS) based HEMS as bench-mark problem with value equal to one, are : 1; 1.22; 1.31; 1.47; 1.64; 1.53; 1.71; and 2.41. (ii) When tested for diversified problems for HEMS based on modeling of SHAs, the algorithm for MS- outperformed DS- based HEMS when evaluated for metric 1 and 2. (iii) The algorithm for CPP (MS)- based HEMS outperformed all the models for other DPSs when evaluated for metric 1 and 2. Especially, the highest value of metric 2 for this model, indicating the fastest reduction in CE while increasing the value of T BD, shows the capability of this model to motivate the consumers for an active participation for reducing event- based critical peaks. However, the CPP- based scheme is not meant for application on daily basis. (iv) When tested for the diversified problems for HEMS for DPSs used on daily basis, DA-RTP- based HEMS outperformed the models for other DPSs with respective values of 20.5% and 0.32 for metrics 1 and 2. The operation of the algorithm for PDDR-RED- based HEMS using weights of (1, 0) resulted in a maximal reduction in CEnet; that reflected the value of the metric 1. While using weights of values (0, 1) for the minimal value of T BD, the algorithm achieved a substantial reduction in CEnet for the RED- based operations for the optimal dispatch of the photovoltaic system, the storage unit, and the power grid without shifting of the loads. The algorithm operations using weights of (1, 0) and (0, 1) for CEnet and T BD enabled the computations of the proposed metrics 1 and 2 for the algorithm. Following is the summary of the results of DPA for the proposed algorithm when tested for a designated set of TPs for PDDR-REDbased HEMS: (i) The respective performance metrics in percent for the TP of 2S-TOU (DS); 2S-TOU (MS); 3S-TOU (DS); 3S-TOU (MS); DA-RTP (DS); DA-RTP (MS);CPP 246


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO (DS); and CPP (MS) are 84.78; 91.69; 92.54; 97.15; 110.13; 112.40; 86.38; and 94.44 for metric 1 and 0.109; 0.214; 0.196; 0.282; 0.185; 0.257; 0.272; and 0.379 for metric 2. The relative values of metric 2, taking 2S-TOU (DS) based HEMS as bench-mark problem with value equal to one, are: 1.00; 1.96; 1.80; 2.59; 1.70; 2.36; 2.50; and 3.48 (ii) When tested for diversified problems for HEMS based on modeling of SHAs, the algorithm for the MS- outperformed DS- based HEMS for metric 1 and 2. (iii) The algorithm was ranked 1-8 for the designated TPs for DA-RTP (MS), DA-RTP (DS), 3S-ToUP (MS), CPP (MS), 3S-ToUP (DS), 2S-ToUP (MS), CPP (DS), and 2S-ToUP (DS) based on metric 1. (iv) When tested for the diversified problems for HEMS for DPSs used on daily basis, the algorithm exhibited the best performance for DA-RTP (MS)- based HEMS while achieving a maximum value of 112.40% for metric 1 for such HEMSs. Interestingly, a very high value of the reduction in CEnet for the RED- based operation for such HEMSs (96.64%) indicates large benefits to the consumer due to the high responsiveness of the scheme for optimal dispatch of the local sources. However, a smaller value of metric 2 indicates lesser responsiveness of such HEMSs for reductions in CEnet while increasing T BD. (v) The algorithm performed the best for the DA-RTP (MS)- and 3S-ToUP (MS)- based HEMSs for the TO for metric 1 and 2. These schemes achieved respective values of 112.40% and 97.15% for metric 1 and 2.36 and 2.59 for metric 2. (vi) The performance of the algorithm for CPP- based schemes was acknowledged through the highest value of 3.48 for metric 2 for CPP (MS)- based HEMSs. This highest value of metric 2 reflecting the maximum reduction in CEnet with increasing value of the T BD shows the potential of CPP- based models for motivating the consumers participating in this event- based PDDR to 247


5.4. SIMULATIONS FOR PDDR- AND PDDR-RED- BASED HEMS USING Chapter 5 DWS-PSO reduce the critical peaks. (vii) The algorithm can also be used for RED- based optimal HEMS operations without shifting of the load. Under this scenario, DA-RTP- based HEMS showed the best results while achieving a value of 96.64% for the reduction in CEnet. That highest reduction in CEnet for RED- based HEMSs was due to the supplying of the peak/ relatively higher priced load through the PV/ SB system. (viii) An outstanding performance for a set of diversified TPs demonstrates the robustness of the proposed DWS-PSO algorithm for HEMS applications.

248


Chapter 6 Conclusion and Future Work

249

Chapter 6

6.1

6.1. CONCLUSIONS, FUTURE WORK

Conclusions, Future Work

6.1.1

Conclusions

The related work (Chapter 2) and survey (Chapter 3) results showed that there are 147 research papers (inclusive 18 review/survey papers) have been analyzed by focusing on the last six years’ publications on DRSREOD-based HEMO. Followings are the conclusions based on the detailed analysis made in this thesis while taking into account the eminent approaches, facts and figures and the performance parameters found in the reviewed literature on HEMO: 1. A review of HEMO modeling based on the dichotomous approaches for integration of RDESs, mutual coordination among consumers, uncertainty of data, multi-objectivity and optimization techniques is conducted. The performance of proposed models are analyzed based on the suitability of the underlying approaches. Further, TOs between dichotomous approaches are investigated based on the performance of the related models. 2. Refer to Tables 3.7 and 3.9, DRSREOD-based HEMO enhances HEMS potential for reduction in CE, peak demand and GHG emissions as compared to DR-based HEMO. An average reduction in CE and P AR are obtained using DRSEROD based HEMO and figured as 30.18% and 29.83%, respectively. While the same performance parameters for DR-based HEMO are averaged as 16.27% and 21.26%. 3. Refer to Table 3.6, response class for SHA has been the one most widely and diversely modeled for its effectiveness in HEMO. SHAs are modeled based on different operating schemes in 41 papers. Elastic and MBA classes are the future trends based on the advancements in power electronic sensors and controls. They are modeled in 14 and 20 number of papers, respectively. Conflicts in nomenclatures and functional definitions for HDs exist in the present literature that calls for standardization requirement for responsive 250


Chapter 6


classifications of HDs. 4. Refer to Tables 3.4 and 3.5 for DRSREOD-based HEMO, SBs have been integrated with RESs to introduce dispatchability to RESs in HEMO. RESs combined with ESSs are modeled in 48 number of papers. The dispatch using HEMO is based on availability of PV, SOC, charge/ discharge rates, EP , etc. Furthermore, increasing number of EVs in the perspective of toxic emissions are expected to largely affect the electric demand profiles in future. Work on HEMO with practical objectives/ constraints like optimal life of ESSs, EVs usage for V2G, V2H and G2V operations and CHP integrations for optimal resources utilization are the future trends in HEMO. 5. Refer to Section 3.3.2, recent research on HEMO is highly diverse in nature. Parameters for modeling are varied largely making it complex to compare the performance of various models. In addition to other parameters, the inclusion of NCA in modeling also affects the performance parameter like CE largely. Refer to Table 3.7 and Table 3.8, quite large reduction in CE has been achieved using HEMO models; even not including the NCAs as compared to the ones incorporating NCAs. The respective average figures are 36.53% and 16.27%. The performance of these two types of models cannot be compared on the same ground. Serious efforts are needed to formalize methods that may enable evaluation of various HEMO models on adequately selected bases. Further, standardized simulation tools and test beds need to be developed for performance evaluations of HEMO models and facilitation for quality research in future. 6. CoHEMO results in more optimal system operation as compared to I-HEMO. CoHEMO-based systems outperform I-HEMO-based systems for peak rebounds management and more reduction in CE. Such system optimizes the joint performance of coordinated consumers exploiting their inherent difference of demands and efficient adaptation of overall demand to RESs and ESSs. Refer to Table 3.10, an average reduction in CE achieved through 251


Chapter 6


CoHEMO is 1.64 times greater than the one achieved through I-HEMO. The respective figures of 31.975% and 19.52% confirm the superiority of CoHEMO approach over I-HEMO. Future trend includes a mixed coordinated scheme with I-HEMO at the local level for freedom of choices and easy implementation along with a centralized controller coordinating for scheduling some of the flexible/direct loads and sources at the consumer end. More productive research on multi agent systems and game theoretic approaches for their application to CoHEMO in near future is predicted. 7. Refer to Table 3.11 which includes the objectives for HEMO problems, most of the researchers have focused on SO based HEMO. In SOO problems, CE and DPr are focused in 35 and 5 research papers, respectively. For MOO problems, the vital TOs between CE and discomfort have been considered in 20 papers. Objectives including GHG emissions and ORU are predicted as future trends. 8. Refer to Table 3.12 including methods to tackle multi-objectivity, WSM or PP have been used in majority of the works. Such methods combine all objectives into a single one that limits provision of clearly diverse solutions for the consumers. PO-based heuristics for MOO providing multiple tradeoff solutions for diverse decision making is predicted as the future trend. Recently, tools to simulate PO-based MOO problems have been made available on the platform like MATLAB. 9. Refer to Tables 3.14, 3.15, 3.16 and 3.17, including computational techniques and LP based computation due to scalability issues which have been used for small-scale problems only. While advanced meta-heuristics have been used very successfully even for the very large-scale complex HEMO problems. Problem complexity increases exponentially with the number of schedulable HDs, the number of time slots, the inclusion of stochastic based variables in modeling, etc. Future trends include improved algorithms based on metaheuristics, meta-heuristics-LP hybrids and NLIP based techniques like BB 252


Chapter 6


to manage problem complexities. 10. Most of the HEMO models in recent literature are deterministic in nature. Refer to Tables 3.2, 3.4 and 3.16, the approach is simple, fast in computation and easily implementable. A compromised scheduling efficiency because of the uncertainty of data is the only drawback in this approach. Small-scale deterministic problems are solved using LP. Typical computing time like 2.9 seconds is claimed for 30 HAs and 24 number of slots. For large scale complex problems, meta-heuristics are used to search near optimal solutions in the matchless minimum time. Refer to Tables 3.13 and 3.17 including methods to address uncertainty, SOP suffers from curse of dimensionality. RO method is used for risk adverse solutions while tackling problem complexity because of uncertainty. In addition, these methods, especially SOP requires accurate historical data for implementation as well. Recently, MPC based on the usage of proximal forecast data has been emerged as a method with lower problem complexity and minimized forecast errors. A typical computing time like 0.2 seconds using MPC is claimed for scheduling of HAs in 40 homes with 480 number of scheduling slots. Research to handle uncertainty for DRSREOD-based HEMO models still remains open with advancements in MPC method predicted as the future trend. Real-time and historical data management needs to be expedited to facilitate research on alternate methods to handle uncertainty. 11. Standardization/ formalization; Multi-objectivity uncertainty management; meta-heuristics, hybrid techniques, NLIP, DP based approaches and parallel processing; HEMO-based residential energy planning; coordinated HEMO, multi-agent systems, game theory, zero energy buildings with negative GHG emission, energy hubs with multi-energy systems; and MGs with multilevel decision making are promising research areas in the field of HEMO. Next, a heuristic algorithm for a DRSREOD-based HEMS for the optimal sharingbased parallel operation of the PV system, the SB and the grid is presented and 253


Chapter 6


validated. A scheme for the MS of SHAs is proposed and incorporated into the algorithm. A higher reduction in the CE with a lower T BD is achieved with MS as compared with the DS scenario for both DR- and DRSREOD-based HEMSs. The peak load for the HEMS is reduced by using an IBR scheme with ToU tariffs. Using MS and MOGA/PO, an aposteriori method of handling multi-objectivity, provides more scheduling flexibility and diversity in decision-making for the consumer. For a DR-based HEMS, a 15.5% reduction in the CE is achieved with the MS, compared with 9.3% in the DS case. This maximum reduction in the CE that is achieved with the MS is also accompanied by a lesser T BD than in the DS scenario, with values of 0.26 and 0.40 for MS and DS, respectively. For a DRSREOD-based HEMS, a 65.92% reduction in the CE is achieved with MS, compared with 50.68% in the DS case. At this minimum CE, the MS results in a slightly higher value of the T BD than that in the DS scenario, with values of 0.45 and 0.27 for MS and DS, respectively. However, for the same T BD level of 0.27, the MS outperforms the DS in terms of the CE, with costs of 88 and 107 cents/day for the MS and the DS, respectively. Considering the energy sold to the utility in a DRSREODbased HEMS, the net bill paid to the utility is 48.37% less in the MS scenario than in the DS scenario. Although the reduction in the CE varies widely when different architectures/parameters are used for modeling, the 65.9% reduction in the CE achieved in this research is even greater than the maximum reduction of 65% achieved in [106] and [61] for DRSREOD-based HEMSs using LP and MILP, respectively. An algorithm for selecting the optimal size of a DG to cope with the LS in a DRSREOD-based HEMS in a developing country, considering the TOs between the CE, T BD and P gsize, is also developed and validated. The required DG power supply capacities as identified directly, without any tradeoff analysis, for HEMS classes designated as unscheduled, DR (with DS), DR (with MS), DRSREOD (with DS) and DRSREOD (with MS) are 3.66, 2.04, 2.34, 2.04 and 1.35 kW, respectively. The proposed algorithm for DG sizing provides the consumer with 254


Chapter 6


multiple choices for DG selection from among a set of tradeoff-based classes with DG supply capacities ranging from 0.41 to 1.95 kW. The problem of optimal DG sizing to cope with the LS in a DRSREOD-based HEMS considering the CE and T BD TOs has vital applications for consumers participating in energy-deficient power supply networks in developing countries. Finally, a simulation-based posteriori method for eco-efficient operations of a DRSREODLDG-based HEMS is proposed. The method computes an optimal set of diversified TOs for CEnet and T BD against minimal T EM iss. Based on simulations, the method completed its function in three-steps. First, a MOGA/ PO based heuristic was used to generate a set of 100 TOs for HEMS operation. Second, an average value constraint filter for T EM iss was applied to filter out the solutions with extremely high values of T EM iss, leaving a set of 66 TOs with reduced T EM iss. Third, an average surface fit for T EM iss was formulated in terms of CEnet and T BD, using an optimal polynomial model for regression. A constraint filter based on the proposed surface fit was applied to screen out the TOs with marginally higher values of T EM iss. The method delivered an eco-efficient set of 33 TOs between CEnet and T BD against a minimal T EM iss. The TOs are classified to enable the consumer choosing the best eco-efficient option. Class-I offered a maximum reduction of 36.23% for CEnet against 20% value of T BD while reduction in T EM iss maintained above 50.53%. Class-II offered a maximum reduction of 52.18% for CEnet against 36% value of T BD while reduction in T EM iss maintained above 51.72%. Class-III offered a maximum reduction of 61.63% for CEnet against 53% value of T BD while reduction in T EM iss maintained above 51.72%. The best eco-efficient solution for a consumer comprised maximized reduction of 60.78% in CEnet against a 45% value of T BD and a 51.72% reduction in T EM iss. An overall reductions achieved for CEnet ranges from 22.61% to 61.63% against the T BD of 17 to 53% while reductions in T EM iss kept within 50.53 to 58.58%. Relationship between the tradeoff parameters and various factors affecting their 255


Chapter 6


trends are analyzed as follows: CEnet reduces exponentially with an increasing T BD; CEnet increases linearly with an increasing loss of the PV energy (P dl); the relation between T EM iss and the related TOs for CEnet and T BD remained undefined when analyzed for the primary TOs data (Fig. 1.1); however, T EM iss exhibited a double-tailed polynomial relation with CEnet when the parameters were analyzed for the final eco-efficient TOs (Fig. 5.36); an uneven/ irregular trend for T EM iss as related to the TOs between CEN et and T BD was exploited to design the proposed filtration mechanism for T EM iss. DWS-PSO based algorithms are presented for PDDR-and PDDR-RED- based HEMSs. The performance of the algorithms was analyzed for the TO solutions using respective weights of (1, 0) and (0, 1) for CEnet and T BD. The first set of weights allocated the maximal priority to the reductions in CE while the second set of weights was used for minimizing the value of T BD. The proposed algorithms were tested for a diversified set of TPs based on DPSs comprising 2S-ToUP, 3SToUP, DA-RTP, DA-RTP and IBR combined with HEMS models for DS and MS. Performance metrics were proposed to evaluate the performance of the algorithms. Metric-1 dealt with the capability to maximize the reduction in the value of CE and metric-2 with the maximization in the average reduction in CE per unit of T BD. Metric-2 was based on the gradient of the TO line for the percentage reduction in the CE and T BD and that represented the responsiveness of the model for reduction in CEnet with an increasing value of T BD. The detailed numerical results are summarized as in subsection 5.4.7.

6.1.2

Future Work

Future work will address improvements in the performance parameters, including CE, T BD and P gsize, and the computation time for a DRSREOD-based HEMS through the following means:

256


Chapter 6


• Varying the slot length from 10 minutes up to 60 minutes. • Using a hybrid function with MOGA while reducing the population/generation sizes. • Varying the crossover fraction from 0.2 to 1 instead of using the default value of 0.8. • Varying the crossover function type from the default “crossover scattered” type to other available options. • Using parallel processing and vectorized options in MOGA. • Introducing emissions as a fitness function for DG selection. • Comparing the performance of MOGA with the performance of other metaheuristic and hybrid methods for HEMS analysis. Further, the work can be extended in terms of improved performance and diversification of the tradeoff parameters including CEnet, T BD and T EM iss for DRSREODLDG-based HEMS through the following means: • Use of MOGA with varied value of crossover fraction and type of crossover function from the opted default values. • Use of other meta-heuristic and hybrid methods to generate the primary TOs and comparing their performance with MOGA. • Activate normalize and robust options available for surface fitting with the polynomial model for regression. These options are not activated in this research. • Use of other type of surface fits and the related options to achieve more efficient and diversified solutions. • Additions of constraints regarding the life of the storage devices, starting of the LDG, and the operation of the LDG near rated power. • The development of a scheme for the integrated reduction of the carbon

257


Chapter 6


commodities for the consumers and the utility through DRSREODLDGbased HEMS. • Minimization of the sum of the PV energy losses in HEMS.

258


Chapter 7 References

259

Bibliography [1] Nadeem, Z., Javaid, N., Malik, A. W., & Iqbal, S. (2018). Scheduling appliances with GA, TLBO, FA, OSR and their hybrids using chance constrained optimization for smart homes. Energies, 11(4), 888. [2] Hafeez, G., Javaid, N., Iqbal, S., & Khan, F. A. (2018). Optimal residential load scheduling under utility and rooftop photovoltaic units. Energies, 11(3), 611. [3] Hussain, H. M., Javaid, N., Iqbal, S., Hasan, Q. U., Aurangzeb, K., & Alhussein, M. (2018). An Efficient Demand Side Management System with a New Optimized Home Energy Management Controller in Smart Grid. Energies, 11(1), 190. [4] Venizelou, V., Philippou, N., Hadjipanayi, M., Makrides, G., Efthymiou, V., & Georghiou, G. E. (2018). Development of a novel time-of-use tariff algorithm for residential prosumer price-based demand side management. Energy, 142, 633-646. [5] Lokeshgupta, B., & Sivasubramani, S. (2018). Multi-objective dynamic economic and emission dispatch with demand side management. International Journal of Electrical Power & Energy Systems, 97, 334-343. [6] Javaid, N., Ullah, I., Akbar, M., Iqbal, Z., Khan, F. A., Alrajeh, N., & Alabed, M. S. (2017). An intelligent load management system with renewable energy integration for smart homes. IEEE Access, 5, 13587-13600. 260

References

BIBLIOGRAPHY

[7] Hussain, B., Hasan, Q. U., Javaid, N., Guizani, M., Almogren, A., & Alamri, A. (2018). An Innovative Heuristic Algorithm for IoT-Enabled Smart Homes for Developing Countries. IEEE Access, 6, 15550-15575. [8] UN Climate Change Agreement, Paris, April 2016. Available online: https://ec.europa.eu/clima/policies/strategies en, (accessed on: 8 June 2018). [9] Global Status report by Renewable Energy Policy Network for 21st Century, Renewable, 2016. Available online:

http://www.ren21.net/status-of-

renewables/global-status-report/, (accessed on: 8 June 2018). [10] Ahmad, A., Khan, A., Javaid, N., Hussain, H. M., Abdul, W., Almogren, A.,. & Azim Niaz, I. (2017). An optimized home energy management system with integrated renewable energy and storage resources. Energies, 10(4), 549. [11] Aslam, S., Iqbal, Z., Javaid, N., Khan, Z. A., Aurangzeb, K., & Haider, S. I. (2017). Towards efficient energy management of smart buildings exploiting heuristic optimization with real time and critical peak pricing schemes. Energies, 10(12), 2065. [12] Rahim, M., Khalid, A., Javaid, N., Ashraf, M., Aurangzeb, K., & Altamrah, A. (2018). Exploiting game theoretic based coordination among appliances in smart homes for efficient energy utilization. Energies, 11(6), 1426. [13] Maleki, A., & Askarzadeh, A. (2014). Artificial bee swarm optimization for optimum sizing of a stand-alone PV/WT/FC hybrid system considering LPSP concept. Solar Energy, 107, 227-235. [14] Zhao, Z., Lee, W. C., Shin, Y., & Song, K. B. (2013). An optimal power scheduling method for demand response in home energy management system. IEEE Transactions on Smart Grid, 4(3), 1391-1400.

261


References

BIBLIOGRAPHY

[15] Khan, A., Javaid, N., & Khan, M. I. (2018). Time and device based priority induced comfort management in smart home within the consumer budget limitation. Sustainable Cities and Society, 41, 538-555. [16] Khan, A., Javaid, N., Ahmad, A., Akbar, M., Khan, Z. A., & Ilahi, M. (2018). A priority-induced demand side management system to mitigate rebound peaks using multiple knapsack. Journal of Ambient Intelligence and Humanized Computing, 1-24. [17] Saha, A., Kuzlu, M., Pipattanasomporn, M., & Rahman, S. (2016). Enabling residential demand response applications with a ZigBee-based load controller system. Intelligent Industrial Systems, 2(4), 303-318. [18] Choi, M., Han, J., & Lee, I. (2016, January). An efficient energy monitoring method based on Bluetooth Low Energy. In Consumer Electronics (ICCE), 2016 IEEE International Conference on (pp. 359-360). IEEE. [19] Fletcher, J., & Malalasekera, W. (2016). Development of a user-friendly, lowcost home energy monitoring and recording system. Energy, 111, 32-46. [20] Nagesh, D. R., Krishna, J. V., & Tulasiram, S. S. (2010, January). A realtime architecture for smart energy management. In Innovative Smart Grid Technologies (ISGT), 2010 (pp. 1-4). IEEE. [21] Barbato, A., & Capone, A. (2014). Optimization models and methods for demand-side management of residential users: A survey. Energies, 7(9), 57875824. [22] Pritoni, M., Ford, R., Karlin, B., & Sanguinetti, A. (2018). Home energy management (HEM) database: A list with coded attributes of 308 devices commercially available in the US., 16, 71-74..

262


References

BIBLIOGRAPHY

[23] Saad al-sumaiti, A., Ahmed, M. H., & Salama, M. M. (2014). Smart home activities: A literature review. Electric Power Components and Systems, 42(34), 294-305. [24] Banos, R., Manzano-Agugliaro, F., Montoya, F. G., Gil, C., Alcayde, A., & Gómez, J. (2011). Optimization methods applied to renewable and sustainable energy: A review. Renewable and Sustainable Energy Reviews, 15(4), 17531766. [25] May, G. J., Davidson, A., & Monahov, B. (2018). Lead batteries for utility energy storage: A review. Journal of Energy Storage, 15, 145-157. [26] Akbari, H., Browne, M. C., Ortega, A., Huang, M. J., Hewitt, N. J., Norton, B.,& McCormack, S. J. (2018). Efficient energy storage technologies for photovoltaic systems. Solar Energy, ISSN 0038-092X, https://doi.org/10.1016/j.solener.2018.03.052. [27] Parra, D., Norman, S. A., Walker, G. S., & Gillott, M. (2017). Optimum community energy storage for renewable energy and demand load management. Applied energy, 200, 358-369. [28] Vivas, F. J., De las Heras, A., Segura, F., & And´ ujar, J. M. (2018). A review of energy management strategies for renewable hybrid energy systems with hydrogen backup. Renewable and Sustainable Energy Reviews, 82, 126-155. [29] Eriksson, E. L. V., & Gray, E. M. (2017). Optimization and integration of hybrid renewable energy hydrogen fuel cell energy systems–A critical review. Applied Energy, 202, 348-364. [30] Zafar, R., Mahmood, A., Razzaq, S., Ali, W., Naeem, U., & Shehzad, K. (2018). Prosumer based energy management and sharing in smart grid. Renewable and Sustainable Energy Reviews, 82(1), 1675-1684.

263


References

BIBLIOGRAPHY

[31] Uddin, M., Romlie, M. F., Abdullah, M. F., Halim, S. A., & Kwang, T. C. (2018). A review on peak load shaving strategies. Renewable and Sustainable Energy Reviews, 82(3), 3323-3332. [32] Shareef, H., Ahmed, M. S., Mohamed, A., & Al Hassan, E. (2018). Review on Home Energy Management System Considering Demand Responses, Smart Technologies, and Intelligent Controllers. IEEE Access, 6, 24498-24509. [33] Prasanna, A., Patel, M. K., Mahmoodi, J., & Brosch, T. (2018). Recent experiences with tariffs for saving electricity in households. Energy Policy, 115, 514-522. [34] Esther, B. P., & Kumar, K. S. (2016). A survey on residential demand side management architecture, approaches, optimization models and methods. Renewable and Sustainable Energy Reviews, 59, 342-351. [35] Tezer, T., Yaman, R., & Yaman, G. (2017). Evaluation of approaches used for optimization of stand-alone hybrid renewable energy systems. Renewable and Sustainable Energy Reviews, 73, 840-853. [36] Ahmad, T., Chen, H., Guo, Y., & Wang, J. (2018). A comprehensive overview on the data driven and large scale based approaches for forecasting of building energy demand: A review. Energy and Buildings, 165, 301-320. [37] Wen, L., Zhou, K., Yang, S., & Li, L. (2018). Compression of smart meter big data: A survey. Renewable and Sustainable Energy Reviews, 91, 59-69. [38] Beaudin, M., & Zareipour, H. (2015). Home energy management systems: A review of modelling and complexity. Renewable and Sustainable Energy Reviews, 45, 318-335. [39] Celik, B., Roche, R., Suryanarayanan, S., Bouquain, D., & Miraoui, A. (2017). Electric energy management in residential areas through coordination of multiple smart homes. Renewable and Sustainable Energy Reviews, 80, 260-275. 264


References

BIBLIOGRAPHY

[40] Sawle, Y., Gupta, S. C., & Bohre, A. K. (2018). Review of hybrid renewable energy systems with comparative analysis of off-grid hybrid system. Renewable and Sustainable Energy Reviews, 81(2), 2217-2235. [41] Amjad, M., Ahmad, A., Rehmani, M. H., & Umer, T. (2018). A review of EVs charging: From the perspective of energy optimization, optimization approaches, and charging techniques. Transportation Research Part D: Transport and Environment, 62, 386-417. [42] Ahmed, F., Naeem, M., Ejaz, W., Iqbal, M., & Anpalagan, A. (2018). Resource management in cellular base stations powered by renewable energy sources. Journal of Network and Computer Applications, 112, 1-17. [43] Zia, M. F., Elbouchikhi, E., & Benbouzid, M. (2018). Microgrids energy management systems: A critical review on methods, solutions, and prospects. Applied Energy, 222(2018), 1033-1055. [44] Liu, Y., Yu, S., Zhu, Y., Wang, D., & Liu, J. (2018). Modeling, planning, application and management of energy systems for isolated areas: A review. Renewable and Sustainable Energy Reviews, 82, 460-470. [45] Wagh, M. M., & Kulkarni, V. V. (2018). Modeling and Optimization of Integration of Renewable Energy Resources (RER) for Minimum Energy Cost, Minimum CO2 Emissions and Sustainable Development, in Recent Years: A Review. Materials Today: Proceedings, 5(1), 11-21. [46] Latest/Current Islamabad

Load

Electric

Management

Supply

Company,

schedule, Pakistan,

official

website

available

at

http://www.iesco.com.pk/index.php/customer-services/load-sheddingschedule [Accessed on 22-3-17] [47] Gellings, C. W. (2009). The smart grid: enabling energy efficiency and demand response. Taylor Francis Inc, United States. 265


References

BIBLIOGRAPHY

[48] Guo, Y., Pan, M., Fang, Y., & Khargonekar, P. P. (2013). Decentralized Coordination of Energy Utilization for Residential Households in the Smart Grid. IEEE Trans. Smart Grid, 4(3), 1341-1350. [49] Breidenich, C., Magraw, D., Rowley, A., & Rubin, J. W. (1998). The Kyoto protocol to the United Nations framework convention on climate change. American Journal of International Law, 92(2), 315-331. [50] Tsagarakis, G., Thomson, R. C., Collin, A. J., Harrison, G. P., Kiprakis, A. E., & McLaughlin, S. (2016). Assessment of the cost and environmental impact of residential demand-side management. IEEE Transactions on Industry Applications, 52(3), 2486-2495. [51] Wu, H., Pratt, A., & Chakraborty, S. (2015, July). Stochastic optimal scheduling of residential appliances with renewable energy sources. In Power & Energy Society General Meeting, 2015 IEEE (pp. 1-5). IEEE. [52] Ahmed, M. S., Mohamed, A., Khatib, T., Shareef, H., Homod, R. Z., & Ali, J. A. (2017). Real time optimal schedule controller for home energy management system using new binary backtracking search algorithm. Energy and Buildings, 138, 215-227. [53] Mahmood, D., Javaid, N., Alrajeh, N., Khan, Z. A., Qasim, U., Ahmed, I., & Ilahi, M. (2016). Realistic scheduling mechanism for smart homes. Energies, 9(3), 202. [54] Rastegar, M., & Fotuhi-Firuzabad, M. (2015). Outage management in residential demand response programs. IEEE Transactions on Smart Grid, 6(3), 1453-1462. [55] Lujano-Rojas, J. M., Monteiro, C., Dufo-Lopez, R., & Bernal-Agustin, J. L. (2012). Optimum residential load management strategy for real time pricing (RTP) demand response programs. Energy policy, 45, 671-679. 266


References

BIBLIOGRAPHY

[56] Adika, C. O., & Wang, L. (2014). Autonomous appliance scheduling for household energy management. IEEE transactions on smart grid, 5(2), 673-682. [57] Rahim, S., Javaid, N., Ahmad, A., Khan, S. A., Khan, Z. A., Alrajeh, N., & Qasim, U. (2016). Exploiting heuristic algorithms to efficiently utilize energy management controllers with renewable energy sources. Energy and Buildings, 129, 452-470. [58] Rajalingam, S., & Malathi, V. (2016). HEM algorithm based smart controller for home power management system. Energy and Buildings, 131, 184-192. [59] Boynuegri, A. R., Yagcitekin, B., Baysal, M., Karakas, A., & Uzunoglu, M. (2013, May). Energy management algorithm for smart home with renewable energy sources. In Power Engineering, Energy and Electrical Drives (POWERENG), 2013 Fourth International Conference on (pp. 1753-1758), Istanbul, Turkey, May 13-17, IEEE. [60] Gudi, N., Wang, L., Devabhaktuni, V., & Depuru, S. S. S. R. (2010, September). Demand response simulation implementing heuristic optimization for home energy management. In North American Power Symposium (NAPS), 2010 (pp. 1-6), Arlington, TX, IEEE. [61] Erdinc, O., Paterakis, N. G., Mendes, T. D., Bakirtzis, A. G., & Catalao, J. P. (2015). Smart household operation considering bi-directional EV and ESS utilization by real-time pricing-based DR. IEEE Transactions on Smart Grid, 6(3), 1281-1291. [62] Kani, S. A. P., Nehrir, H., Colson, C.,& Wang, C. (2011, July). Real-time energy management of a stand-alone hybrid wind-microturbine energy system using particle swarm optimization. In Power and Energy Society General Meeting, 2011 IEEE (pp. 1-1), IEEE.

267


References

BIBLIOGRAPHY

[63] Liu, M., Quilumba, F. L., & Lee, W. J. (2015). A collaborative design of aggregated residential appliances and renewable energy for demand response participation. IEEE Transactions on Industry Applications, 51(5), 3561-3569. [64] Paterakis, N. G., Erdin¸c, O., Pappi, I. N., Bakirtzis, A. G., & Catalão, J. P. (2016). Coordinated operation of a neighborhood of smart households comprising electric vehicles, energy storage and distributed generation. IEEE Transactions on smart grid, 7(6), 2736-2747. [65] Askarzadeh, A. (2017). Distribution generation by photovoltaic and diesel generator systems: Energy management and size optimization by a new approach for a stand-alone application. Energy, 122, 542-551. [66] Erdinc, O., Paterakis, N. G., Catalão, J. P., Pappi, I. N., & Bakirtzis, A. G. (2015, January). Smart households and home energy management systems with innovative sizing of distributed generation and storage for customers. In System Sciences (HICSS), 2015 48th Hawaii International Conference on (pp. 1462-1471), Kauai, IEEE. [67] Rasheed, M. B., Javaid, N., Awais, M., Khan, Z. A., Qasim, U., Alrajeh, N., & Javaid, Q. (2016). Real time information based energy management using customer preferences and dynamic pricing in smart homes. Energies, 9(7), 130. [68] Javaid, N., Hafeez, G., Iqbal, S., Alrajeh, N., Alabed, M. S., & Guizani, M. (2018). Energy Efficient Integration of Renewable Energy Sources in the Smart Grid for Demand Side Management. IEEE Access. [69] Samuel, O., Javaid, S., Javaid, N., Ahmed, S., Afzal, M., & Ishmanov, F. (2018). An Efficient Power Scheduling in Smart Homes Using Jaya Based Optimization with Time-of-Use and Critical Peak Pricing Schemes. Energies, 11(11), 3155.

268


References

BIBLIOGRAPHY

[70] Samuel, O., Javaid, N., Ashraf, M., Ishmanov, F., Afzal, M., & Khan, Z. (2018). Jaya based Optimization Method with High Dispatchable Distributed Generation for Residential Microgrid. Energies, 11(6), 1513. [71] Iqbal, Z., Javaid, N., Mohsin, S., Akber, S., Afzal, M., & Ishmanov, F. (2018). Performance Analysis of Hybridization of Heuristic Techniques for Residential Load Scheduling. Energies, 11(10), 2861. [72] Aslam, S., Javaid, N., Khan, F. A., Alamri, A., Almogren, A., & Abdul, W. (2018). Towards Efficient Energy Management and Power Trading in a Residential Area via Integrating a Grid-Connected Microgrid. Sustainability, 10(4), 1245. [73] Iqbal, Z., Javaid, N., Iqbal, S., Aslam, S., Khan, Z. A., Abdul, W., ... & Alamri, A. (2018). A Domestic Microgrid with Optimized Home Energy Management System. Energies, 11(4), 1002. [74] Sharma, S., Singh, M., & Prakash, S. (2017, July). A novel energy management system for modified zero energy buildings using multi-agent systems. In Smart Grid and Smart Cities (ICSGSC), 2017 IEEE International Conference on (pp. 267-271). Singapore, IEEE. [75] Yan, X., Ozturk, Y., Hu, Z. and Song, Y., 2018. A review on price-driven residential demand response. Renewable and Sustainable Energy Reviews, 96, pp.411-419. [76] Yang, C., Meng, C. and Zhou, K., 2017. Residential electricity pricing in China: The context of price-based demand response. Renewable and Sustainable Energy Reviews. [77] Anees, A., Hussain, I., AlKhaldi, A.H. and Aslam, M., 2018. Linear triangular optimization technique and pricing scheme in residential energy management systems. Results in Physics, 9, pp.858-865. 269


References

BIBLIOGRAPHY

[78] Hussain, B., Hasan, Q.U., 2016, October. Demand side management for smart homes in Pakistan. In Emerging Technologies (ICET), 2016 International Conference on (pp. 1-6). IEEE. [79] Latif, U., Javaid, N., Zarin, S.S., Naz, M., Jamal, A. and Mateen, A., 2018 Cost Optimization in Home Energy Management System using Genetic Algorithm, Bat Algorithm and Hybrid Bat Genetic Algorithm. [80] Naz, M., Iqbal, Z., Javaid, N., Khan, Z.A., Abdul, W., Almogren, A. and Alamri, A., 2018. Efficient Power Scheduling in Smart Homes Using Hybrid Grey Wolf Differential Evolution Optimization Technique with Real Time and Critical Peak Pricing Schemes. Energies, 11(2), p.384. [81] Khalid, A., Javaid, N., Guizani, M., Alhussein, M., Aurangzeb, K. and Ilahi, M., 2018. Towards dynamic coordination among home appliances using multiobjective energy optimization for demand side management in smart buildings. IEEE Access, 6, pp.19509-19529. [82] Javaid, N., Ahmed, A., Iqbal, S. and Ashraf, M., 2018. Day Ahead Real Time Pricing and Critical Peak Pricing Based Power Scheduling for Smart Homes with Different Duty Cycles. Energies, 11(6), p.1464. [83] Anvari-Moghaddam, A., Monsef, H. and Rahimi-Kian, A., 2015. Costeffective and comfort-aware residential energy management under different pricing schemes and weather conditions. Energy and Buildings, 86, pp.782793. [84] Shakeri, M., Shayestegan, M., Reza, S.S., Yahya, I., Bais, B., Akhtaruzzaman, M., Sopian, K. and Amin, N., 2018. Implementation of a novel home energy management system (HEMS) architecture with solar photovoltaic system as supplementary source. Renewable Energy, 125, pp.108-120.

270


References

BIBLIOGRAPHY

[85] Wang, D., Zhang, X., Qu, K., Yu, T., Pan, Z. and Liu, Q., 2018. Pareto tribe evolution with equilibrium-based decision for multi-objective optimization of multiple home energy management systems. Energy and Buildings, 159, pp.1123. [86] Meng, X.B., Gao, X.Z., Lu, L., Liu, Y. and Zhang, H., 2016. A new bioinspired optimisation algorithm: Bird Swarm Algorithm. Journal of Experimental and Theoretical Artificial Intelligence, 28(4), pp.673-687. [87] Li, K., Wang, R., Zhang, T. and Ishibuchi, H., 2018. Evolutionary manyobjective optimization: a comparative study of the state-of-the-art. IEEE Access. [88] Bodansky, D. (2016). The Paris climate change agreement: a new hope. American Journal of International Law, 110(2), 288-319. [89] Sawin, J. L., Martinot, E., Sonntag-O’Brien, V., McCrone, A., Roussell, J., Barnes, D., ... & Ellenbeck, S. (2013). Renewables 2010Global status report. Available Online:

http : //www.ren21.net/wp −

content/uploads/2018/06/17 − 8652G SR2018F ullReportw eb− 1.pdf [90] Bozchalui, M. C., Hashmi, S. A., Hassen, H., Ca˜ nizares, C. A., & Bhattacharya, K. (2012). Optimal operation of residential energy hubs in smart grids. IEEE Transactions on Smart Grid, 3(4), 1755-1766. [91] Chiu, T. C., Pai, C. W., Shih, Y. Y., & Pang, A. C. (2014),”Optimized DayAhead Pricing With Renewable Energy Demand-Side Management for Smart Grids,” in IEEE Internet of Things Journal, 4(2), 374-383. [92] Elsied, M., Oukaour, A., Gualous, H., Hassan, R., & Amin, A. (2014, June). An advanced energy management of microgrid system based on genetic algorithm. In Industrial Electronics (ISIE), 2014 IEEE 23rd International Symposium on (pp. 2541-2547), Istanbul, Turkey, IEEE. 271


References

BIBLIOGRAPHY

[93] Liao, G. C. (2013, February). The optimal economic dispatch of smart Microgrid including Distributed Generation. In Next-Generation Electronics (ISNE), 2013 IEEE International Symposium on (pp. 473-477), Kaohsiung, IEEE. [94] Javaid, N., Hussain, S. M., Ullah, I., Noor, M. A., Abdul, W., Almogren, A., & Alamri, A. (2017). Demand side management in nearly zero energy buildings using heuristic optimizations. Energies, 10(8), 1131. [95] Manzoor, A., Javaid, N., Ullah, I., Abdul, W., Almogren, A., & Alamri, A. (2017). An intelligent hybrid heuristic scheme for smart metering based demand side management in smart homes. Energies, 10(9), 1258. [96] Javaid, N., Naseem, M., Rasheed, M. B., Mahmood, D., Khan, S. A., Alrajeh, N., & Iqbal, Z. (2017). A new heuristically optimized Home Energy Management controller for smart grid. Sustainable Cities and Society, 34, 211-227. [97] Kazmi, S., Javaid, N., Mughal, M. J., Akbar, M., Ahmed, S. H., & Alrajeh, N. (2017). Towards optimization of metaheuristic algorithms for IoT enabled smart homes targeting balanced demand and supply of energy. IEEE Access. [98] Javaid, N., Ahmed, F., Ullah, I., Abid, S., Abdul, W., Alamri, A., & Almogren, A. S. (2017). Towards cost and comfort based hybrid optimization for residential load scheduling in a smart grid. Energies, 10(10), 1546. [99] Li, N., Chen, L., & Low, S. H. (2011, July). Optimal demand response based on utility maximization in power networks. In Power and Energy Society General Meeting, 2011 IEEE (pp. 1-8). IEEE. [100] Tsui, K. M., & Chan, S. C. (2012). Demand response optimization for smart home scheduling under real-time pricing. IEEE Transactions on Smart Grid, 3(4), 1812-1821.

272


References

BIBLIOGRAPHY

[101] Barbato, A., Capone, A., Carello, G., Delfanti, M., Falabretti, D., & Merlo, M. (2014). A framework for home energy management and its experimental validation. Energy Efficiency, 7(6), 1013-1052. [102] Squartini, S., Boaro, M., De Angelis, F., Fuselli, D., & Piazza, F. (2013, June). Optimization algorithms for home energy resource scheduling in presence of data uncertainty. In Intelligent Control and Information Processing (ICICIP), 2013 Fourth International Conference on (pp. 323-328), Beijing, China, June 9-11, IEEE. [103] Ozturk, Y., Senthilkumar, D., Kumar, S., & Lee, G. K. (2013). An Intelligent Home Energy Management System to Improve Demand Response. IEEE Trans. Smart Grid, 4(2), 694-701. [104] Shafiee, S., Fotuhi-Firuzabad, M., & Rastegar, M. (2012). Impacts of controlled and uncontrolled PHEV charging on distribution systems. 9th IET Int. Conference on Advances in Power System Control, Operation and Management (pp. 1-6), Hong Kong, IEEE. [105] Sou, K. C., Weimer, J., Sandberg, H., & Karl Henrik, J. (2011). Scheduling smart home appliances using mixed integer linear programming. In 50th IEEE Conference on Decsion and Control and European Control Conference (pp. 5144-5149), Orlando, FL, USA, December 12-15, IEEE. [106] Cabras, M., Pilloni, V., & Atzori, L. (2015, June). A novel smart home energy management system: Cooperative neighbourhood and adaptive renewable energy usage. In Communications (ICC), 2015 IEEE International Conference on (pp. 716-721), London, UK, June 8 - 12, IEEE. [107] Meng, F. L., & Zeng, X. J. (2014, July). An optimal real-time pricing for demand-side management: A Stackelberg game and genetic algorithm approach. In Neural Networks (IJCNN), 2014 International Joint Conference

273


References

BIBLIOGRAPHY

on (pp. 1703-1710), Beijing, China, IEEE. [108] Chen, Z., Wu, L., & Fu, Y. (2012). Real-time price-based demand response management for residential appliances via stochastic optimization and robust optimization. IEEE Transactions on Smart Grid, 3(4), 1822-1831. [109] Jia, L., Yu, Z., Murphy-Hoye, M. C., Pratt, A., Piccioli, E. G., & Tong, L. (2011, December). Multi-scale stochastic optimization for home energy management. In Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), 2011 4th IEEE International Workshop on (pp. 113-116), IEEE. [110] Mohsenian-Rad, A. H., & Leon-Garcia, A. (2010). Optimal residential load control with price prediction in real-time electricity pricing environments. IEEE Trans. Smart Grid, 1(2), 120-133. [111] Kim, T. T., & Poor, H. V. (2011). Scheduling power consumption with price uncertainty. IEEE Transactions on Smart Grid, 2(3), 519-527. [112] Salpakari, J., Rasku, T., Lindgren, J., & Lund, P. D. (2017). Flexibility of electric vehicles and space heating in net zero energy houses: an optimal control model with thermal dynamics and battery degradation. Applied energy, 190, 800-812. [113] Chen, X., Wei, T., & Hu, S. (2013). Uncertainty-aware household appliance scheduling considering dynamic electricity pricing in smart home. IEEE Transactions on Smart Grid, 4(2), 932-941. [114] Chang, T. H., Alizadeh, M., & Scaglione, A. (2013). Real-time power balancing via decentralized coordinated home energy scheduling. IEEE Transactions on Smart Grid, 4(3), 1490-1504. [115] Dehnad, A., & Shakouri, H. S. (2013). A novel model of intelligent electrical load management by goal programming for smart houses, respecting consumer preferences. Energy and Power Engineering, 5(10), 622. 274


References

BIBLIOGRAPHY

[116] Hong, Y. Y., Lin, J. K., Wu, C. P., & Chuang, C. C. (2012). Multi-Objective Air-Conditioning Control Considering Fuzzy Parameters Using Immune Clonal Selection Programming. IEEE Trans. Smart Grid, 3(4), 1603-1610. [117] Pedrasa, M. A. A., Spooner, T. D., & MacGill, I. F. (2010, June). The value of accurate forecasts and a probabilistic method for robust scheduling of residential distributed energy resources. In Probabilistic Methods Applied to Power Systems (PMAPS), 2010 IEEE 11th International Conference on (pp. 587-592), Singapore, IEEE. [118] Beaudin, M., Zareipour, H., & Schellenberg, A. (2012, October). Residential energy management using a moving window algorithm. In Innovative Smart Grid Technologies (ISGT Europe), 2012 3rd IEEE PES International Conference and Exhibition on (pp. 1-8), Berlin, IEEE. [119] Chavali, P., Yang, P., & Nehorai, A. (2014). A distributed algorithm of appliance scheduling for home energy management system. IEEE Transactions on Smart Grid, 5(1), 282-290. [120] Nguyen, T. A., & Crow, M. L. (2012, May). Optimization in energy and power management for renewable-diesel microgrids using dynamic programming algorithm. In Cyber Technology in Automation, Control, and Intelligent Systems (CYBER), 2012 IEEE International Conference on (pp. 11-16), Bangkok, IEEE. [121] Yu, Z., Jia, L., Murphy-Hoye, M. C., Pratt, A., & Tong, L. (2013). Modeling and stochastic control for home energy management. IEEE Transactions on Smart Grid, 4(4), 2244-2255. [122] Beaudin, M., Zareipour, H., Bejestani, A. K., & Schellenberg, A. (2014). Residential energy management using a two-horizon algorithm. IEEE Transactions on Smart Grid, 5(4), 1712-1723.

275


References

BIBLIOGRAPHY

[123] Lee, K. Y., & El-Sharkawi, M. A. (Eds.). (2008). Modern heuristic optimization techniques: theory and applications to power systems (Vol. 39), United Kingdom, John Wiley & Sons. [124] Sareni, B., & Krahenbuhl, L. (1998). Fitness sharing and niching methods revisited. IEEE transactions on Evolutionary Computation, 2(3), 97-106. [125] Bonabeau, E., Marco, D. D. R. D. F., Dorigo, M., Théraulaz, G., & Theraulaz, G. (1999). Swarm intelligence: from natural to artificial systems (No. 1). Oxford university press. [126] Eberhart, R. C., Shi, Y., & Kennedy, J. (2001). Swarm intelligence. Elsevier. [127] Blum, C., & Dorigo, M. (2004). The hyper-cube framework for ant colony optimization. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 34(2), 1161-1172. [128] Momoh, J. (2012). Smart grid: fundamentals of design and analysis (Vol. 63). John Wiley & Sons. [129] Huang, T., & Liu, D. (2013). A self-learning scheme for residential energy system control and management. Neural Computing and Applications, 22(2), 259-269. [130] Paterakis, N. G., Erdinc, O., Bakirtzis, A. G., & Catalao, J. P. (2015). Optimal household appliances scheduling under day-ahead pricing and loadshaping demand response strategies. IEEE Transactions on Industrial Informatics, 11(6), 1509-1519. [131] Ha, L. D., Ploix, S., Zamai, E., & Jacomino, M. (2006, September). Tabu search for the optimization of household energy consumption. In Information Reuse and Integration, 2006 IEEE International Conference on (pp. 86-92), Waikoloa Village, HI, IEEE.

276


References

BIBLIOGRAPHY

[132] Shi, H., Blaauwbroek, N., Nguyen, P. H., & Kamphuis, R. I. (2016). Energy management in Multi-Commodity Smart Energy Systems with a greedy approach. Applied energy, 167, 385-396. [133] Zhou, Y., Chen, Y., Xu, G., Zhang, Q., & Krundel, L. (2014, June). Home energy management with PSO in smart grid. In Industrial Electronics (ISIE), 2014 IEEE 23rd International Symposium on (pp. 1666-1670), Istanbul, Turkey, IEEE. [134] Liu, Y., Yuen, C., Yu, R., Zhang, Y., & Xie, S. (2016). Queuing-based energy consumption management for heterogeneous residential demands in smart grid. IEEE Transactions on Smart Grid, 7(3), 1650-1659. [135] Salinas, S., Li, M., & Li, P. (2013). Multi-objective optimal energy consumption scheduling in smart grids. IEEE Transactions on Smart Grid, 4(1), 341-348. [136] Imamura, A., Yamamoto, S., Tazoe, T., Onda, H., Takeshita, H., Okamoto, S., & Yamanaka, N. (2013, August). Distributed demand scheduling method to reduce energy cost in smart grid. In Humanitarian Technology Conference (R10-HTC), 2013 IEEE Region 10 (pp. 148-153), Sendai, IEEE. [137] Rastegar, M., Fotuhi-Firuzabad, M., & Moeini-Aghtaie, M. (2018). Developing a two-level framework for residential energy management. IEEE Transactions on Smart Grid, 9(3), 1707-1717. [138] Hassan, N. U., Pasha, M. A., Yuen, C., Huang, S., & Wang, X. (2013). Impact of scheduling flexibility on demand profile flatness and user inconvenience in residential smart grid system. Energies, 6(12), 6608-6635. [139] Samadi, P., Mohsenian-Rad, H., Wong, V. W., & Schober, R. (2013). Tackling the load uncertainty challenges for energy consumption scheduling in smart grid. IEEE Transactions on Smart Grid, 4(2), 1007-1016. 277


References

BIBLIOGRAPHY

[140] Tarasak, P. (2011, October). Optimal real-time pricing under load uncertainty based on utility maximization for smart grid. In Smart Grid Communications (SmartGridComm), 2011 IEEE International Conference on (pp. 321326), Brussels, IEEE. [141] Du, P., & Lu, N. (2011). Appliance commitment for household load scheduling. IEEE transactions on Smart Grid, 2(2), 411-419. [142] Sancho-Tomás, A., Sumner, M., & Robinson, D. (2017). A generalised model of electrical energy demand from small household appliances. Energy and Buildings, 135, 350-366. [143] Zhang, D., Li, S., Sun, M., & O’Neill, Z. (2016). An optimal and learningbased demand response and home energy management system. IEEE Transactions on Smart Grid, 7(4), 1790-1801. [144] Agnetis, A., de Pascale, G., Detti, P., & Vicino, A. (2013). Load scheduling for household energy consumption optimization. IEEE Transactions on Smart Grid, 4(4), 2364-2373. [145] Xiao, J., Chung, J. Y., Li, J., Boutaba, R., & Hong, J. W. K. (2010, October). Near optimal demand-side energy management under real-time demandresponse pricing. In Network and Service Management (CNSM), 2010 International Conference on (pp. 527-532), Niagara Falls, IEEE. [146] Karami, H., Sanjari, M. J., Hosseinian, S. H., & Gharehpetian, G. B. (2014). An optimal dispatch algorithm for managing residential distributed energy resources. IEEE Transactions on Smart Grid, 5(5), 2360-2367. [147] Gatsis, N., & Giannakis, G. B. (2012). Residential Load Control: Distributed Scheduling and Convergence With Lost AMI Messages. IEEE Trans. Smart Grid, 3(2), 770-786.

278


References

BIBLIOGRAPHY

[148] Tan, O., Gomez-Vilardebo, J., & G¨ und¨ uz, D. (2017). Privacy-cost trade-offs in demand-side management with storage. IEEE Transactions on Information Forensics and Security, 12(6), 1458-1469. [149] Mesaric, P., & Krajcar, S. (2015). Home demand side management integrated with electric vehicles and renewable energy sources. Energy and Buildings, 108, 1-9. [150] Gao, F., Tang, W., Yan, T., Ma, J., Yan, W., Zhang, X., & Yang, D. (2015, November). Day-ahead energy optimal scheduling of household microgrid considering the user satisfaction. In Power and Energy Engineering Conference (APPEEC), 2015 IEEE PES Asia-Pacific (pp. 1-5), Brisbane, Australia, IEEE. [151] Aloini, D., Crisostomi, E., Raugi, M., & Rizzo, R. (2011, December). Optimal power scheduling in a virtual power plant. In Innovative Smart Grid Technologies (ISGT Europe), 2011 2nd IEEE PES International Conference and Exhibition on (pp. 1-7), Manchester, IEEE. [152] Conejo, A. J., Morales, J. M., & Baringo, L. (2010). Real-time demand response model. IEEE Transactions on Smart Grid, 1(3), 236-242. [153] Di Piazza, M. C., La Tona, G., Luna, M., & Di Piazza, A. (2017). A twostage Energy Management System for smart buildings reducing the impact of demand uncertainty. Energy and Buildings, 139, 1-9. [154] Arabali, A., Ghofrani, M., Etezadi-Amoli, M., Fadali, M. S., & Baghzouz, Y. (2013). Genetic-algorithm-based optimization approach for energy management. IEEE Transactions on Power Delivery, 28(1), 162-170. [155] Alrumayh, O., & Bhattacharya, K. (2015, October). Model predictive control based home energy management system in smart grid. In Electrical Power and Energy Conference (EPEC), 2015 IEEE (pp. 152-157), London, IEEE. 279


References

BIBLIOGRAPHY

[156] Shakeri, M., Shayestegan, M., Abunima, H., Reza, S. S., Akhtaruzzaman, M., Alamoud, A. R. M., ... & Amin, N. (2017). An intelligent system architecture in home energy management systems (HEMS) for efficient demand response in smart grid. Energy and Buildings, 138, 154-164. [157] Keerthisinghe, C., Verbiˇc, G., & Chapman, A. C. (2014, September). Evaluation of a multi-stage stochastic optimization framework for energy management of residential PV-storage systems. In Power Engineering Conference (AUPEC), 2014 Australasian Universities (pp. 1-6), Perth, Australia, IEEE. [158] Chen, Z., & Wu, L. (2013). Residential appliance DR energy management with electric privacy protection by online stochastic optimization. IEEE Transactions on Smart Grid, 4(4), 1861-1869. [159] Byun, J., Hong, I., & Park, S. (2012). Intelligent cloud home energy management system using household appliance priority based scheduling based on prediction of renewable energy capability. IEEE Transactions on Consumer Electronics, 58(4), 1194-1201. [160] Lugo-Cordero, H. M., Fuentes-Rivera, A., Guha, R. K., & Ortiz-Rivera, E. I. (2011, June). Particle swarm optimization for load balancing in green smart homes. In Evolutionary Computation (CEC), 2011 IEEE Congress on (pp. 715-720), New Orleans, Los Angeles, IEEE. [161] Clastres, C., Pham, T. H., Wurtz, F., & Bacha, S. (2010). Ancillary services and optimal household energy management with photovoltaic production. Energy, 35(1), 55-64. [162] Zanoli, S. M., Pepe, C., Orlietti, L., & Barchiesi, D. (2015, October). A Model Predictive Control strategy for energy saving and user comfort features in building automation. In System Theory, Control and Computing (ICSTCC),

280


References

BIBLIOGRAPHY

2015 19th International Conference on (pp. 472-477), Cheile Gradistei, Romania, IEEE. [163] Sun, C., Sun, F., & Moura, S. J. (2015, July). Data enabled predictive energy management of a PV-battery smart home nanogrid. In American Control Conference (ACC), 2015 (pp. 1023-1028), Chicago, IL, IEEE. [164] Yoshimura, Y., Kondo, T., Kawanishi, M., Narikiyo, T., & Sato, A. (2015, November). Model predictive control of ev storage battery with hems based on particle swarm optimization. In Innovative Smart Grid Technologies-Asia (ISGT ASIA), 2015 IEEE (pp. 1-5), Bangkok, Thailand, IEEE. [165] Chen, Z., Zhang, Y., & Zhang, T. (2015, May). An intelligent control approach to home energy management under forecast uncertainties. In Power Engineering, Energy and Electrical Drives (POWERENG), 2015 IEEE 5th International Conference on (pp. 657-662), Riga, Latvia, IEEE. [166] Zhang, Y., Wang, R., Zhang, T., Liu, Y., & Guo, B. (2016). Model predictive control-based operation management for a residential microgrid with considering forecast uncertainties and demand response strategies. IET Generation, Transmission & Distribution, 10(10), 2367-2378. [167] Parisio, A., Wiezorek, C., Kyntäjä, T., Elo, J., & Johansson, K. H. (2015, August). An MPC-based energy management system for multiple residential microgrids. In Automation Science and Engineering (CASE), 2015 IEEE International Conference on (pp. 7-14), Gothenburg, Sweden, IEEE. [168] Dahmane, M., Bosche, J., El-Hajjaji, A., & Dafarivar, M. (2013, October). Renewable energy management algorithm for stand-alone system. In Renewable Energy Research and Applications (ICRERA), 2013 International Conference on (pp. 621-626), Madrid, Spain, IEEE.

281


References

BIBLIOGRAPHY

[169] Huang, Y., Wang, L., Guo, W., Kang, Q., & Wu, Q. (2018). Chance constrained optimization in a home energy management system. IEEE Transactions on Smart Grid, 9(1), 252-260. [170] Ma, J., Chen, H. H., Song, L., & Li, Y. (2016). Residential load scheduling in smart grid: A cost efficiency perspective. IEEE Transactions on Smart Grid, 7(2), 771-784. [171] Perez, K. X., Baldea, M., & Edgar, T. F. (2016, July). Integrated smart appliance scheduling and HVAC control for peak residential load management. In American Control Conference (ACC), 2016 (pp. 1458-1463), Boston, MA, USA, IEEE. [172] Hurink, J. L., & Smit, G. J. (2010). Management and control of domestic smart grid technology. IEEE Transactions on Smart Grid, 1, 109-119. [173] Koch, S., Zima, M., & Andersson, G. (2009). Potentials and applications of coordinated groups of thermal household appliances for power system control purposes. Sustainable Alternative Energy (SAE), 1-8. [174] Scott, P., Thiébaux, S., Van Den Briel, M., & Van Hentenryck, P. (2013, September). Residential demand response under uncertainty. In International Conference on Principles and Practice of Constraint Programming (pp. 645660), Springer, Berlin, Heidelberg. [175] Liu, P., & Fu, Y. (2013, July). Optimal operation of energy-efficiency building: A robust optimization approach. In Power and Energy Society General Meeting (PES), 2013 IEEE (pp. 1-5), Vancouver, BC, Canada, IEEE. [176] Mathieu, J. L., Vayá, M. G., & Andersson, G. (2013, November). Uncertainty in the flexibility of aggregations of demand response resources. In Industrial Electronics Society, IECON 2013-39th Annual Conference of the IEEE (pp. 8052-8057), Vienna, Austria, IEEE. 282


References

BIBLIOGRAPHY

[177] Alkhalil, F., Degobert, P., Colas, F., & Robyns, B. (2009, April). Fuel consumption optimization of a multi-machines microgrid by secant method combined with IPPD table. In International Conference on Renewable Energies and Power Quality ICREPQ, Valencia, Spain (Vol. 15, No. 17.04). [178] Jokar, P., Arianpoo, N., & Leung, V. C. (2013, June). Intrusion detection in advanced metering infrastructure based on consumption pattern. In Communications (ICC), 2013 IEEE International Conference on (pp. 4472-4476), Budapest, IEEE. [179] Di Giorgio, A., & Pimpinella, L. (2012). An event driven smart home controller enabling consumer economic saving and automated demand side management. Applied Energy, 96, 92-103. [180] Xu, L., Qian, F., Li, Y., Li, Q., Yang, Y. W., & Xu, J. (2016). Resource allocation based on quantum particle swarm optimization and RBF neural network for overlay cognitive OFDM System. Neurocomputing, 173, 1250-1256. [181] Niu, D., Wang, H., & Cai, C. (2010, March). Application of Neural Networkbased Combining Forecasting Model Optimized by Ant Colony In Power Load Forecasting. In Power and Energy Engineering Conference (APPEEC), 2010 Asia-Pacific (pp. 1-4), Chengdu, China, IEEE. [182] Liao, G. C. (2014). Hybrid improved differential evolution and wavelet neural network with load forecasting problem of air conditioning. International Journal of Electrical Power & Energy Systems, 61, 673-682. [183] Tsai, M. S., & Chu, C. C. (2011, September). Applications of hybrid EPACO for power distribution system loss minimization under load variations. In Intelligent System Application to Power Systems (ISAP), 2011 16th International Conference on (pp. 1-7), Hersonissos, Greece, IEEE.

283


References

BIBLIOGRAPHY

[184] Fang, N., Zhou, J., Zhang, R., Liu, Y., & Zhang, Y. (2014). A hybrid of real coded genetic algorithm and artificial fish swarm algorithm for short-term optimal hydrothermal scheduling. International Journal of Electrical Power & Energy Systems, 62, 617-629. [185] Ghanbari, A., Kazemi, S. M., Mehmanpazir, F., & Nakhostin, M. M. (2013). A cooperative ant colony optimization-genetic algorithm approach for construction of energy demand forecasting knowledge-based expert systems. Knowledge-Based Systems, 39, 194-206. [186] Trivedi, A., Srinivasan, D., Biswas, S., & Reindl, T. (2016). A genetic algorithm–differential evolution based hybrid framework: case study on unit commitment scheduling problem. Information Sciences, 354, 275-300. [187] Mohammadi, M., & Nafar, M. (2013). Optimal placement of multitypes DG as independent private sector under pool/hybrid power market using GAbased Tabu Search method. International Journal of Electrical Power & Energy Systems, 51, 43-53. [188] Saber, A. Y., & Rahman, D. M. F. (2011, July). Economic load dispatch using particle swarm differential evolution optimization. In Power and Energy Society General Meeting, 2011 IEEE (pp. 1-8), Detroit, MI, USA, USA, IEEE. [189] Talbi, N., & Belarbi, K. (2011, December). Optimization of fuzzy controller using hybrid Tabu search and particle swarm optimization. In Hybrid Intelligent Systems (HIS), 2011 11th International Conference on (pp. 561-565). Melacca, Malaysia, IEEE. [190] N¨ uesch, T., Ott, T., Ebbesen, S., & Guzzella, L. (2012, June). Cost and fuel-optimal selection of HEV topologies using particle swarm optimization and dynamic programming. In American Control Conference (ACC), 2012 (pp. 1302-1307), Montreal, QC, Canada, IEEE.

284


References

BIBLIOGRAPHY

[191] Rahmat, N. A., Musirin, I., Abidin, A. F., & Ahmad, M. R. (2014, September). Economic load dispatch with valve-point loading effect by using differential evolution immunized ant colony optimization technique. In Proc. Austral. Univ. Power Eng. Conf.(AUPEC), 5(3), 91-99. [192] Li, C., Zhao, H., & Chen, T. (2010, August). The hybrid differential evolution algorithm for optimal power flow based on simulated annealing and tabu search. In Management and Service Science (MASS), 2010 International Conference on (pp. 1-7), Wuhan, China, IEEE. [193] Jirutitijaroen, P., & Singh, C. (2006, October). A hybrid method for multiArea generation expansion using tabu-search and dynamic programming. In Power System Technology, 2006. PowerCon 2006. International Conference on (pp. 1-5), Chongqing, China, IEEE. [194] Kin Cheong Sou, James Weimer, Henrik Sandberg, and Karl Henrik Johansson,”Scheduling SH Appliances Using Mixed Integer Linear Programming”, 2011 50th IEEE Conference on Decision and Control and European Control Conference (CDC-ECC) Orlando, FL, USA, December 12-15, 2011 [195] avilable:http://www.bchydro.com/youraccount/content/residential r

−

ates.jsp]. [196] Inclining Block Rate in British Columbia Hydro Co Aug. 2012 [Online] [197] Pedrasa, M. A. A., Spooner, T. D., & MacGill, I. F. (2010). Coordinated scheduling of residential distributed energy resources to optimize smart home energy services. IEEE Transactions on Smart Grid, 1(2), 134-143. [198] Pakistan

Engineering

Council,

PEC,

[Online].

Available:

http://www.pec.org.pk/. [Accessed 28 September 2016].

285


References

BIBLIOGRAPHY

[199] Sharafi, M., & ElMekkawy, T. Y. (2014). A dynamic MOPSO algorithm for multiobjective optimal design of hybrid renewable energy systems. International Journal of Energy Research, 38(15), 1949-1963. [200] Jakob, W., and Blume, C.,2014. Pareto Optimization or Cascaded Weighted Sum: A Comparison of Concepts.Algorithms [201] Denis, D. J. (2015). Applied univariate, bivariate, and multivariate statistics. John Wiley & Sons., United States of America. [202] Diesel ABLE

Generator Sales,

Fuel

Consumption

Melbourne,

Chart

Australia,

in

Litres,

Available

Online:

https://www.ablesales.com.au/source/Diesel Generator Fuel Consumption Chart in Litres.pdf (accessed on February 19, 2018) [203] Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M., & Troxler, T. G. (2014). 2013 supplement to the 2006 IPCC guidelines for national greenhouse gas inventories: Wetlands. IPCC, Switzerland. [204] An Overview of Electricity Sector in Pakistan, Islamabad Chamber of Commerce and Industry, Pakistan, 2012, [Available Online] http : //icci.com.pk/data/downloads/63/12936190481 .pdf (accessed on February 19, 2018) [205] Curve

and

surface

fittings

in

Matlab-2015a,

https://www.mathworks.com/help/curvefit/fit.html

Available

(accessed

on

Online: Feb.

19, 2018) [206] Gujarati, Damodar N. ”Econometrics by example.” (2011), Palgrave Macmillan, United States.

286
