UNIVERSITI TEKNOLOGI MARA UNIT COMMITMENT IN POWER SYSTEM USING MULTI-AGENT EVOLUTIONARY PROGRAMMING INCORPORATING PRIORITY LISTING OPTIMISATION TECHNIQUE

UNIVERSITI TEKNOLOGI MARA

UNIT COMMITMENT IN POWER SYSTEM USING MULTI-AGENT EVOLUTIONARY PROGRAMMING

INCORPORATING PRIORITY LISTING OPTIMISATION TECHNIQUE

MUHAMMAD NAZREE BIN CHE OTHMAN

T hesis subm itted in fulfilm ent o f the requirem ents for the degree o f

Master of Science

AUTHOR’S DECLARATION

I declare that the w ork in this thesis was carried out in accordance w ith the regulations o f U niversiti Teknologi M ARA. It is original and is the result o f my own work, unless otherw ise indicated or acknow ledge as referenced work. This thesis has not been subm itted to any other academ ic institution or non-academ ic institution for any degree or qualification.

I, hereby, acknow ledge that I have been supplied w ith the Academ ic Rules and Regulations for Post Graduate, Universiti Teknologi M ARA, regulating the conduct o f my study and research.

Nam e o f Student Student I.D. No.

Program m e Faculty Title

M uham m ad Nazree Bin Che Othm an 2009840114

M aster o f Science

Faculty o f Electrical Engineering

U nit C om m itm ent in Power System U sing M ulti-agent Evolutionary Program m ing Incorporating Priority Listing O ptim isation Technique.

Signature o f Student :

Date

ii

ABSTRACT

Escalating cost o f electrical energy generation becom es one o f the m ajor issues faced by alm ost all electrical pow er producers around the globe. High price o f fuel, generating unit unavailability, fix loaded requirem ent, the m ust run generators, m axim um or m inim um pow er capacity or even generating units derated capacity are factors that influenced this cost increase. M ost o f generating units are restricted to its scheduled interchange w ith sufficient operating reserve and regulating m argin in order to prepare the pow er system operation for em ergency response. Furtherm ore, the tim e variance o f the dem and for electricity has increased the com plexity in ensuring the request is satisfied while the costs o f generation is minimised. As dem ands for electricity grows from time to time, pow er system s becom e larger and m ore com plex, which, lead to the issues o f high operation cost and resulting long com putational time for the execution process. In pow er system operation and planning study, the m ain platform in discussing this issue is referred as unit com m itm ent (UC) problem . This research presents an approach to solve the UC problem using a newly developed M ulti-agent Evolutionary Program m ing incorporating Priority Listing optim isation technique (M AEP-PL). The objective o f this study is to search for generation scheduling such that the total operating cost can be m inim ised w hen subjected to a variety o f constraints, while at the same time reducing its com putational time. The proposed technique assim ilates the concepts o f Priority Listing (PL), m ulti-agent system (M AS) and Evolutionary Program m ing (EP) as its basis. In the proposed technique, determ inistic PL technique was first applied to produce a population o f initial solutions. The search process then being refined using heuristic EP-based algorithm with m ulti-agent approach to produce the final solution. The developed technique was tested on the 10 generating units test system then up to 20 generating units in a 24-hours scheduling period and the results

TABLE OF CONTENTS

Page

AUTHOR’S DECLARATION

ⁱⁱ

ABSTRACT

ⁱⁱⁱ

ACKNOWLEDGEMENT

^iv

TABLE OF CONTENTS

^V

LIST OF TABLES

^viii

LIST OF FIGURES

^ix

LIST OF ABBREVIATIONS

^X

CHAPTER ONE: INTRODUCTION

1.1 Research Background 1

1.2 Problem Statem ent 2

1.3 Research O bjectives 4

1.4 Scope o f Research 4

1.5 Significant o f the Research 6

1.6 O rganization o f Thesis 7

CHAPTER TWO: LITERATURE REVIEW

2.1 Introduction 9

2.2 Unit C om m itm ent Problem 9

2.2.1 O n -o ff State and Unit Com m itm ent Econom ic Dispatch 11

2.3 D eterm inistic O ptim ization Techniques 11

2.3.1 D ynam ic Program m ing 12

2.3.2 Lagrange Relaxation 12

2.3.3 Branch and Bound 13

2.3.4 M ixed Integer Program m ing 14

2.3.5 Priority Listing 15

v

Page

2.4 Evolution A lgorithm O ptim ization Techniques 16

2.4.1 Evolutionary Program m ing 17

2.4.2 A pplications o f Evolutionary Program m ing 17

2.5 M ulti-agent System 18

2.6 C hapter C onclusion 19

CHAPTER THREE: METHODOLOGY

3.1 Introduction 21

3.2 Research M ethodology 21

3.2.1 K now ledge A cquisition and Background Studies 21

3.2.2 Design of Algorithm 22

3.2.3 C onstruction and Execution 24

3.2.4 Experim entation and O bservation 24

3.2.5 Evaluation and Analysis 25

3.3 Modelling the Unit Commitment Problem 25

3.3.1 O n -o ff State o f G enerator 25

3.3.2 U nit C om m itm ent Econom ic D ispatch (U C-ED) 26 3.3.3 U nit C om m itm ent O bjective Function 26 3.3.4 System Requirem ent and G enerating Unit Constraints 28

i Generating Unit Capability Limit Constraint 28

ii Pow er Balance C onstraint 29

iii Spinning Reserve Constraint 29

iv G enerating Unit M inim um On-tim e Constraint 29 v G enerating Unit M inim um Off-time Constraint 30

vi Fuel Constraint 30

vii Crew Constraint 30

viii M ust-run Unit/s C onstraint 30

3.3.5 D ata Specification 31