Adaptive Energy Storage System Control for Microgrid

Stability Enhancement
by

Tan Zhang

A Thesis

Submitted to the Faculty

of the

WORCESTER POLYTECHNIC INSTITUTE

In partial fulfillment of the requirements for the

Degree of Doctor of Philosophy

in

Electrical and Computer Engineering

April 2018
© Copyright by Zhang, Tan (2018)

APPROVED:

Dr. Alexander E. Emanuel, Dissertation Advisor, ECE Department, WPI

Dr. John A. Orr, Advisor, ECE Department, WPI

Dr. Aleksandar Stanković, ECE Department, Tufts University

 I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis,
including any required final revision, as accepted by my examiners.

I understand that my thesis may be made electronically available to the public.

For discussion with the author, please send email to: [email protected].

2
 Abstract

Microgrids are local power systems of different sizes located inside the distribution sys-
tems. Each microgrid contains a group of interconnected loads and distributed energy
resources that acts as a single controllable entity with respect to the grid. Their island-
ing operation capabilities during emergencies improve the resiliency and reliability of
the electric energy supply. Due to its low kinetic energy storage capacity, maintaining
microgrid stability is challenging under system contingencies and unpredictable power
generation from renewable resources.
This dissertation highlights the potential benefits of flexibly utilizing the battery en-
ergy storage systems to enhance the stability of microgrids. The main contribution of this
research consists in the development of a storage converter controller with an additional
stability margin that enables it to improve microgrid frequency and voltage regulation
as well as its induction motor post-fault speed recovery. This new autonomous control
technique is implemented by adaptively setting the converter controller parameters based
on its estimated phase-locked loop frequency deviation and terminal voltage magnitude
measurement. This work also assists in the microgrid design process by determining the
normalized minimum storage converter sizing under a wide range of microgrid motor in-
ertia, loading and fault clearing time with both symmetrical and asymmetrical fault types.
This study evaluates the expandability of the proposed control methodologies under
an unbalanced meshed microgrid with fault-induced feeder switching and multiple con-
tingencies in addition to random power output from renewable generators. The favorable
results demonstrate the robust storage converter controller performance under a dynamic
changing microgrid environment.
 Acknowledgements

If there is a God, I want to thank you for guiding me through this journey. For the past
seven years, my life has been wonderful here at WPI. My family is the most important
part of my life. And I miss you here during my study. I thank you for the continuous love,
supports and sacrifices. I love you.
Words cannot express my gratefulness to my beloved mentors: Prof. Alexander Eige-
les Emanuel and John Andrew Orr. For me, they are not just my advisors, but my Amer-
ican grandfather and father. They nurture my growth, both as a researcher and as a man.
Thank you for guiding me with the patience, love and faith. Now it is the time for me to
carry on your torch, and pass on to the next generation.
I consider myself very lucky to have Dr. Aleksandar Stanković as my dissertation
examiner. It is inspiring to listen his insights on power system. I also appreciate Dr.
Marija D. Ilić on willingness to provide the written feedback on my research.
Finally, I thank you all my friends who share my journey here at WPI. It is like a long
marathon, and you help me go through those ups and downs. The memories we created
along the way is my life treasure.

i
 I Ching

D EDICATED TO

The people in my WPI journey

”I don’t think about the miles that are coming down the road, I don’t think about the
mile I am on right now, I don’t think about the miles I’ve already covered. I think about
what I’m doing right now, just being lost in the moment.”
– Ryan Hall, U.S. Olympic marathoner.

”When you’re in the day-to-day grind, it just seems like it’s another step along the
way. But I find joy in the actual process, the journey, the work. It’s not the end. It’s not
the end event.”
– Cal Ripken, Jr., retired MLB shortstop and 19-time All-Star.

”Doing my PhD feels like running a marathon. The most beautiful part of it is not
crossing the finish line, but inside my each step of this journey itself.”
– Tan Zhang, Electrical engineer and marathoner.

ii
Contents

1 Introduction 1
1.1 Thesis Statement, Objectives and Contributions . . . . . . . . . . . . . . 1
1.2 Historical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Grid Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.2 Distributed Generation (DG) and Storage . . . . . . . . . . . . . 5
1.2.3 Microgrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Thesis Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Literature Review 9
2.1 Energy Storage System (ESS) . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Voltage-Sourced Converter (VSC) Control . . . . . . . . . . . . . . . . . 11
2.3 Microgrid Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Microgrid Stability and Control . . . . . . . . . . . . . . . . . . . . . . . 13
2.5 Chapter2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Fundamentals of Microgrid Stability, Modeling and Control 16

3.1 Microgrid Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Microgrid Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.1 Electric networks . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.2 Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

iii
 3.3 Synchronous Generator Modeling and Control . . . . . . . . . . . . . . . 20
3.3.1 Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.2 Real power and frequency control . . . . . . . . . . . . . . . . . 22
3.3.3 Reactive power and voltage control . . . . . . . . . . . . . . . . 25
3.4 Battery Energy Storage System (BESS) Modeling and Control . . . . . . 26
3.4.1 Converter modeling and current-mode control . . . . . . . . . . . 26
3.4.2 Real/Reactive power/current capacity limitations . . . . . . . . . 27
3.5 Chapter3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4 Adaptive Frequency Regulation 29

4.1 Microgrid System for Chapter 4 . . . . . . . . . . . . . . . . . . . . . . 29
4.2 Adaptive Primary Nonlinear Droop Control . . . . . . . . . . . . . . . . 30
4.3 Simultaneous Fast Secondary Frequency Control . . . . . . . . . . . . . 34
4.3.1 Approach#1: Fixed Gain . . . . . . . . . . . . . . . . . . . . . . 34
4.3.2 Approach#2: Adaptive Gain . . . . . . . . . . . . . . . . . . . . 35
4.4 Cases Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.4.1 Case#A: Unstable BESS operation with linear droop . . . . . . . 37
4.4.2 Case#B: Stable BESS operation with nonlinear droop . . . . . . . 40
4.5 Chapter4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5 Adaptive Voltage Regulation 46

5.1 Microgrid System for Chapter 5 . . . . . . . . . . . . . . . . . . . . . . 47
5.2 Induction Motor (IM) loads Impacts on Microgrid Dynamics and Stability 48
5.3 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.4 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.4.1 Algorithm for adaptive voltage regulation set point: Vref (t) . . . . 53
5.4.2 Alternative: With Regulation Delays and Power Priority Switch . 56

iv
 5.5 Simulation Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.5.1 Transient Study#A: Impacts of BESS size on motor speed recovery 59
5.5.2 Transient Study#B: Advantage of adaptive voltage regulation . . . 66
5.5.3 Sensitivity Study#A: Minimum BESS size under symmetrical fault 70
5.5.4 Sensitivity Study#B: Minimum BESS size under asymmetrical fault 73
5.5.5 Transient Study#C: Multiple IMs recovery . . . . . . . . . . . . 74
5.6 Chapter5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

6 Adaptive Control Evaluations on CIGRE Benchmark System 79

6.1 Microgrid System for Chapter 6 . . . . . . . . . . . . . . . . . . . . . . 80
6.2 Study#1: Normal Voltage Regulation . . . . . . . . . . . . . . . . . . . . 82
6.2.1 Case#A: Impact of BESSs secondary control approaches . . . . . 83
6.2.2 Case#B: Microgrid with and without BESSs - A comparison study 89
6.2.3 Case#C: Impact of asymmetrical fault . . . . . . . . . . . . . . . 93
6.2.4 Case#D: Multiple faults and renewable generation . . . . . . . . 98
6.3 Study#2: Adaptive Voltage Regulation . . . . . . . . . . . . . . . . . . . 109
6.3.1 Case#A: Without voltage regulation control delay . . . . . . . . . 109
6.3.2 Case#B: With voltage regulation control delay . . . . . . . . . . 121
6.4 Chapter6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

7 Conclusions 128
7.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

A System Parameters 131

A.1 Radial Feeder for Chapter 4 and 5 . . . . . . . . . . . . . . . . . . . . . 131
A.2 Microgrid data for Chapter 6 CIGRE Benchmark System . . . . . . . . . 131

v
 A.2.1 Feeder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
A.2.2 Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
A.2.3 Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
A.3 Synchronous generators . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
A.4 Prime movers and governing systems . . . . . . . . . . . . . . . . . . . . 134
A.5 Excitation systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
A.6 Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
A.7 Induction Motors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

vi
List of Abbreviations
AGC: Automatic Generation Control

BESS: Battery Energy Storage System

DER: Distributed Energy Resources

DFIG: Doubly-Fed Induction Generator

DG: Distributed Generation

EMS: Energy Management System

ESS: Energy Storage System

FIDVR: Fault Induced Delayed Voltage Recovery

HV: High Voltage

IM: Induction Motor

LFC: Load-Frequency Control

MV: Medium Voltage

PCC: Point of Common Coupling

PSS: Power System Stabilizer

PLL: Phase-Locked Loop

RES: Renewable Energy Source

VSC: Voltage-Sourced Converter

vii
List of Symbols
Chapter 3

D: Load damping constant

H: Inertia constant (s) of the synchronous machine

δ: rotor angle, rad.

idref : Converter reference real current

iqref : Converter reference reactive current

id : Converter real current

iq : Converter reactive current

î: Converter current magnitude

Pa : Per unit accelerating power applied to the rotor.

∆Pm : Generator input mechanical power change (pu)

∆Pe : Generator output electrical power change including the losses (pu)

∆PL : Non-frequency-sensitive load change (pu)

∆PD : Frequency-sensitive load change (pu)

ωs : Synchronous angular velocity (radians/s)

ωs : Per unit synchronous angular velocity

∆ω: Per unit angular speed deviation of the rotor.

Rs : Generator governor speed regulation

viii
 Ta : Per unit accelerating torque applied to the rotor.

Chapter 4

fdb : frequency regulator dead-band

fr : Generator rotor speed

fs : Estimated PLL frequency

Idmax : maximum d axis current.

KA : Gain parameter for adaptive nonlinear droop control

KB : Gain parameter for adaptive secondary frequency control

KI : Secondary frequency control gain

KImax : maximum allowable value of KI

KP : Primary droop gain

PB : Battery real power

0
Pref 1 : Converter nonlinear droop reference power under base power of converter
rated power

Pref 1 :: Converter nonlinear droop reference power under system base power

Pc : BESS rated power in pu under the system base.

Pmax : Maximum allowable converter output power

Rc : Converter capacity saturation droop

V̂ P CC : PCC voltage magnitude

ix
Chapter 5
0
∆f : frequency deviation with respect to the nominal value

ωIM : IM speed (rev/min)

fsw : frequency deviation threshold for power priority switch

HIME : equivalent motor inertia constant (s)

Iqmax : maximum q axis current.

Imax : maximum allowable converter current.

Isat : converter saturation current

KR : adaptive voltage reference value ramping rate

JIM : IM inertia (Kg.m2 ).

KIM : IM load linear interpolation factor.

KVI : voltage regulator integral gain

QB : reactive power response of the BESS

Qmax : maximum allowable reactive power

SB : BESS converter power rating in MVA

S B : normalized minimum BESS inverter size under microgrid loading base

Sm : total microgrid demand in M W .

trs : reset time (s)

TM : load torque

x
 Tr : rated torque of the IM

Te : IM electromechanical torque

Tp : IM breakdown torque

Vdb : voltage regulator dead-band

Vs : Source voltage

VT : IM terminal voltage

Vref (t): voltage reference point for voltage regulator

VrefN : normal voltage reference point

VrefE : emergency voltage reference point

VthH : higher threshold voltage to deactivate the adaptive voltage regulation

VthL : lower threshold voltage to activate the adaptive voltage regulation

Zs : short circuit source impedance

Chapter 6

PRE : renewable power output.

xi
List of Figures

3.1 Microgrid stability problems with usual reasons and improvement strategies. 18
3.2 Block diagram for transfer function relating (a). generator speed and
power; (b). generator speed and power including the effects of the load
damping. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Speed-droop governor schematic . . . . . . . . . . . . . . . . . . . . . . 23
3.4 Traditional Droop Control Concept . . . . . . . . . . . . . . . . . . . . . 24
3.5 Two generators parallel operation with both primary and secondary fre-
quency control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Generic excitation control system block diagram (a). DC type; (b). AC
type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7 Overall converter modeling and control block diagrams . . . . . . . . . . 27

4.1 Single-Line diagram of the chapter 4’s studied microgrid system . . . . . 30

4.2 Adaptive Nonlinear Droop: (a) Block diagram, (b) Resultant Droop Curve 31
4.3 Converter Frequency Regulator Block Diagram . . . . . . . . . . . . . . 34
4.4 Adaptive Secondary Frequency Controller Block Diagram . . . . . . . . 35
4.5 Estimated PLL frequency fs for linear droop with Rc = 1% and sec-
ondary controller gain of 1000: (a). Overall unstable response, (b). Mag-
nified oscillogram for 0.67 < t < 0.71s, (c). Magnified oscillogram for
1.48 < t < 1.52s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

xii
4.6 Magnified PCC voltage oscillogram vs. time: for 0.53 < t < 0.73s, inad-
missible waveform quality for linear droop with Rc = 1% and secondary
controller gain of 1000. . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.7 Unstable control performance: The BESS power vs. time for linear droop
with Rc = 1% and secondary controller gain of 1000. . . . . . . . . . . . 39
4.8 Estimated PLL frequency fs for adaptive nonlinear droop with Rc = 1%
and secondary controller gain of 1000. . . . . . . . . . . . . . . . . . . . 41
4.9 Stable control performance: The BESS power vs. time for adaptive non-
linear droop with Rc = 1% and secondary control gain of 1000. . . . . . 42
4.10 PCC voltage dq components vs. time with and without BESS adaptive
control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.11 Magnified PCC voltage oscillogram vs. time: (a). With BESS adaptive
control, for 0.53 < t < 0.73s, (b). With BESS adaptive control, for
load#2 switched in at t = 1.5s, (c). Without BESS, for 0.53 < t < 0.73s,
(d). Without BESS, for load#2 switched in at t = 1.5s. . . . . . . . . . . 44
4.12 Synchronous generator rotor frequency fr with and without BESS adap-
tive control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5.1 Single-Line diagram of the chapter 5’s studied system . . . . . . . . . . . 47

5.2 Induction motor free accelerating experiment single-line diagram: (a).
Direct start, (b). With step-down transformer. . . . . . . . . . . . . . . . 49
5.3 Induction motor free accelerating torque-speed curves . . . . . . . . . . . 51
5.4 Torque-speed characteristics for an induction motor under different ter-
minal voltage VT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.5 BESS PCC voltage regulator . . . . . . . . . . . . . . . . . . . . . . . . 52
5.6 Vref (t) operating state decision tree . . . . . . . . . . . . . . . . . . . . 54
5.7 The alternative Vref (t) decision tree . . . . . . . . . . . . . . . . . . . . 57

xiii
5.8 Induction motor speed ωIM1 under different BESS sizes . . . . . . . . . . 60
5.9 V̂ P CC with and without 2.5 MVA BESS . . . . . . . . . . . . . . . . . . 61
5.10 Induction motor torque (Te1 ) versus time with and without 2.5 MVA BESS 61
5.11 Induction motor torque (Te1 ) versus speed (ωIM1 ) with and without 2.5
MVA BESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.12 BESS converter current d/q frame components versus time . . . . . . . . 63
5.13 BESS power versus time: (a). Real power, (b). Reactive power. . . . . . . 65
5.14 Generator frequency fr versus time with and without 2.5 MVA BESS . . 66
5.15 Induction motor speed ωIM1 with and without adaptive voltage regulation 68
5.16 V̂ P CC with and without adaptive voltage regulation . . . . . . . . . . . . 68
5.17 PCC voltage waveform oscillograms under adaptive voltage regulation,
1.40 < t < 1.55 s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.18 Torque of induction motor (Te1 ) versus time with and without adaptive
voltage regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.19 BESS converter current d/q frame components versus time, with and
without adaptive voltage regulation . . . . . . . . . . . . . . . . . . . . . 70
5.20 Normalized minimum BESS inverter size S B as a function of equivalent
motor inertia constant HIME with fault clearing time as the parameter:
(a). 34 % IM loading, (b). 47 % IM loading, (c). 55 % IM loading, (d).
61 % IM loading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.21 Minimum normalized BESS Size S B as a function of fault clearing time
under single-phase fault . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.22 Speed of the two induction motors versus time with 3.1 MVA BESS . . . 75
5.23 V̂ P CC with and without 3.1 MVA BESS . . . . . . . . . . . . . . . . . . 76
5.24 BESS power versus time under two induction motors loading: (a). Real
power, (b). Reactive power. . . . . . . . . . . . . . . . . . . . . . . . . . 77

xiv
5.25 Generator frequency fr versus time with two induction motors loading
and a 3.1 MVA BESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

6.1 GIGRE Test Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.2 Estimated PLL Frequency fs ; (left column): unstable under secondary
controller with KI = 3000, (right column): stable under secondary con-
troller with adaptive gain using KB = 3 × 105 : (a) fs1 , unstable; (b) fs2 ,
unstable; (c) fs3 , unstable; (d) fs1 , stable; (e) fs2 , stable; (f) fs3 , stable . . 85
6.3 The BESSs power vs. time; (left column): unstable under secondary con-
troller with KI = 3000, (right column): stable under secondary controller
with adaptive gain using KB = 3 × 105 : (a) PB1 , unstable; (b) PB2 , un-
stable; (c) PB3 , unstable; (d)PB1 , stable; (e) PB2 , stable; (f) PB3 , stable . . 86
6.4 The PCCs voltage magnitude V̂ P CC vs. time; (left column): unstable
under secondary controller with KI = 3000, (right column): stable under
secondary controller with adaptive gain using KB = 3 × 105 : (a) V̂ P CC1 ,
unstable; (b) V̂ P CC2 , unstable; (c) V̂ P CC3 , unstable; (d) V̂ P CC1 , stable;
(e) V̂ P CC2 , stable; (f) V̂ P CC3 , stable . . . . . . . . . . . . . . . . . . . . 87
6.5 The synchronous generators rotor frequency fr (Hz) vs. time; (left col-
umn): unstable under secondary controller with KI = 3000, (right col-
umn): stable under secondary controller with adaptive gain using KB =
3 × 105 : (a) fr1 : 5 M V A hydroelectric generator, unstable; (b) fr2 :
1.8 M V A diesel generator, unstable; (c) fr3 : 3.125 M V A diesel gen-
erator, unstable; (d) fr1 : 5 M V A hydroelectric generator, stable; (e) fr2 :
1.8 M V A diesel generator, stable; (f) fr3 : 3.125 M V A diesel generator,
stable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

xv
6.6 The synchronous generators rotor frequency fr (Hz) vs. time; with and
without BESSs: (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A
diesel generator; (c) fr3 : 3.125 M V A diesel generator . . . . . . . . . . . 90
6.7 Induction motors speed ωIM vs. time, with and without BESSs: (a) ωIM1 :
2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP . . . . 91
6.8 The PCCs voltage magnitude V̂ P CC vs. time; with and without BESSs:
(a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3 . . . . . . . . . . . . . . . . . . . . . 92
6.9 Estimated PLL Frequency fs under 12 cycles asymmetrical fault: (a)
fs1 ; (b) fs2 ; (c) fs3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.10 The BESSs power vs. time under 12 cycles asymmetrical fault: (a) PB1 ;
(b) PB2 ; (c) PB3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.11 The synchronous generators rotor frequency fr (Hz) vs. time under 12
cycles asymmetrical fault: (a) fr1 : 5 M V A hydroelectric generator; (b)
fr2 : 1.8 M V A diesel generator; (c) fr3 : 3.125 M V A diesel generator . . 96
6.12 The PCCs voltage magnitude V̂ P CC vs. time under 12 cycles asymmetrical
fault: (a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3 . . . . . . . . . . . . . . . . . . 97
6.13 Renewable Energy Power Output PRE under multiple faults (4 and 3 cy-
cles fault clearing time respectively): (a) PRE1 ; (b) PRE2 . . . . . . . . . 100
6.14 Estimated PLL Frequency fs under multiple faults and renewable gener-
ation ramping events with normal voltage regulation: (a) fs1 ; (b) fs2 ; (c)
fs3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.15 BESSs current magnitudes transients under multiple faults and renewable
generation ramping events with normal voltage regulation: (a) BESS1 ;
(b) BESS2 ; (c) BESS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

xvi
6.16 BESSs real power vs. time under multiple faults and renewable genera-
tion ramping events with normal voltage regulation: (a) PB1 ; (b) PB2 ; (c)
PB3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.17 BESSs reactive power vs. time under multiple faults and renewable gen-
eration ramping events with normal voltage regulation: (a) QB1 ; (b) QB2 ;
(c) QB3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.18 Induction motors speed ωIM vs. time, under multiple faults and renew-
able generation ramping events with normal voltage regulation: (a) ωIM1 :
2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP . . . . 105
6.19 The synchronous generators rotor frequency fr (Hz) vs. time under mul-
tiple faults and renewable generation ramping events with normal voltage
regulation: (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A
diesel generator; (c) fr3 : 3.125 M V A diesel generator . . . . . . . . . . . 106
6.20 The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and
renewable generation ramping events with normal voltage regulation: (a)
V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3 . . . . . . . . . . . . . . . . . . . . . . . 107
6.21 Magnified PCC voltage oscillogram vs. time under multiple faults and
renewable generation ramping events with normal voltage regulation: (a)
VP CC1 , first fault-induced islanding transient, 0.05 < t < 0.45 s; (b)
VP CC2 , first fault-induced islanding transient, 0.05 < t < 0.45 s; (c)
VP CC3 , first fault-induced islanding transient, 0.05 < t < 0.45 s; (d)
VP CC1 , second (islanding) fault-induced feeder switching transient, 3.5 <
t < 3.9 s; (e) VP CC2 , second (islanding) fault-induced feeder switching
transient, 3.5 < t < 3.9 s; (f) VP CC3 , second (islanding) fault-induced
feeder switching transient, 3.5 < t < 3.9 s . . . . . . . . . . . . . . . . . 108

xvii
6.22 Renewable Energy Power Output PRE under multiple (two) faults (7 and
4 cycles fault clearing time respectively): (a) PRE1 ; (b) PRE2 . . . . . . . 112
6.23 The synchronous generators rotor frequency fr (Hz) vs. time under multi-
ple faults and renewable generation ramping events with adaptive voltage
regulation (no regulation delay): (a) fr1 : 5 M V A hydroelectric generator;
(b) fr2 : 1.8 M V A diesel generator; (c) fr3 : 3.125 M V A diesel generator 113
6.24 The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and
renewable generation ramping events with adaptive voltage regulation (no
regulation delay): (a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3 . . . . . . . . . . . . 114
6.25 Magnified PCC voltage oscillogram vs. time under multiple faults and
renewable generation ramping events with adaptive voltage regulation
(no regulation delay): (a) VP CC1 , first fault-induced islanding transient,
0.05 < t < 0.45 s; (b) VP CC2 , first fault-induced islanding transient,
0.05 < t < 0.45 s; (c) VP CC3 , first fault-induced islanding transient,
0.05 < t < 0.45 s; (d) VP CC1 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s; (e) VP CC2 , second (islanding) fault-
induced feeder switching transient, 3.5 < t < 3.9 s; (f) VP CC3 , second
(islanding) fault-induced feeder switching transient, 3.5 < t < 3.9 s . . . 115
6.26 Induction motors speed ωIM vs. time, under multiple faults and renewable
generation ramping events with adaptive voltage regulation (no regulation
delay): (a) ωIM1 : 2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d)
ωIM4 : 500 HP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.27 Estimated PLL Frequency fs under multiple faults and renewable gen-
eration ramping events with adaptive voltage regulation (no regulation
delay): (a) fs1 ; (b) fs2 ; (c) fs3 . . . . . . . . . . . . . . . . . . . . . . . . 117

xviii
6.28 BESSs current magnitudes transients under multiple faults and renewable
generation ramping events with adaptive voltage regulation (no regulation
delay): (a) BESS1 ; (b) BESS2 ; (c) BESS3 . . . . . . . . . . . . . . . . . 118
6.29 BESSs real power vs. time under multiple faults and renewable genera-
tion ramping events with adaptive voltage regulation (no regulation de-
lay): (a) PB1 ; (b) PB2 ; (c) PB3 . . . . . . . . . . . . . . . . . . . . . . . 119
6.30 BESSs reactive power vs. time under multiple faults and renewable gen-
eration ramping events with adaptive voltage regulation (no regulation
delay): (a) QB1 ; (b) QB2 ; (c) QB3 . . . . . . . . . . . . . . . . . . . . . . 120
6.31 The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (in-
cluding regulation delay): (a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3 . . . . . . . 122
6.32 Induction motors speed ωIM vs. time, under multiple faults and renew-
able generation ramping events with adaptive voltage regulation (includ-
ing regulation delay): (a) ωIM1 : 2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 :
2250 HP; (d) ωIM4 : 500 HP . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.33 The synchronous generators rotor frequency fr (Hz) vs. time under multi-
ple faults and renewable generation ramping events with adaptive voltage
regulation (including regulation delay): (a) fr1 : 5 M V A hydroelectric
generator; (b) fr2 : 1.8 M V A diesel generator; (c) fr3 : 3.125 M V A diesel
generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

xix
6.34 Magnified PCC voltage oscillogram vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (in-
cluding regulation delay): (a) VP CC1 , first fault-induced islanding tran-
sient, 0.05 < t < 0.45 s; (b) VP CC2 , first fault-induced islanding tran-
sient, 0.05 < t < 0.45 s; (c) VP CC3 , first fault-induced islanding tran-
sient, 0.05 < t < 0.45 s; (d) VP CC1 , second (islanding) fault-induced
feeder switching transient, 3.5 < t < 3.9 s; (e) VP CC2 , second (island-
ing) fault-induced feeder switching transient, 3.5 < t < 3.9 s; (f) VP CC3 ,
second (islanding) fault-induced feeder switching transient, 3.5 < t < 3.9 s126

A.1 CIGRE system loading . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

A.2 Diesel governor block diagram . . . . . . . . . . . . . . . . . . . . . . . 134
A.3 MATLAB hydraulic turbine and governing system block diagram: (a).
Overall model; (b). hydraulic turbine; (c). gate servomotor . . . . . . . . 135
A.4 PLL schematic and associated control block diagram . . . . . . . . . . . 137

xx
Chapter 1

Introduction

1.1 Thesis Statement, Objectives and Contributions

Deregulation of the electric utility industry as well as the need for a more sustainable and
efficient future power system drive a broad range of technological developments including
the proliferation of distributed generation (DG) units in the electrical distribution grid [1].
The increasing concerns regarding grid reliability and resiliency under both physical dis-
turbances and cyber attacks support development of the capacity of islanding portions of
distribution systems with groups of interconnected distributed energy resources (DERs)
and loads as ”Micro-Grids” during emergencies [2–4]. Consensus has been reached that
microgrids will play a significant role in enhancing the performance of modern distribu-
tion systems by improving the quality and reliability of the electric energy supply [5].
The realization of microgrid benefits requires each microgrid to maintain stable fre-
quency and voltage operation by implementing robust control of its DERs [6], especially
during the unplanned islanding events such as fault-triggered islanding incidents [7]. This
requires power engineers to develop innovative control strategies that can maintain elec-
tric power system stability under these challenging contingencies [1, 6].

1
 The following conditions highlight the microgrid stability and control challenges:

1. The randomly varying power generation from distributed Renewable Energy Sources
(RESs), unforeseeable load switchings, faults and their induced electric network re-
configurations, all create large disturbances on both microgrid frequency and volt-
age.

2. Low kinetic energy storage capacity and reduction of rotational inertia as syn-
chronous generators are replaced by DERs with power converter interfaces within
a microgrid further reduce ride-through capability.

3. The high degree of interaction among different DG units with various interfaces and
differing controller characteristics can negatively impact overall microgrid stability.

Energy storage systems (ESSs) equipped with converter interfaces are capable of pro-
viding rapid real and reactive current/power support to a microgrid control system. They
can be seen as either loads or generators depending on their modes of operation. Under
proper control methodologies, those capabilities can be indispensable for facilitating a
reliable and stable microgrid operation. Although previous work has proposed several
new strategies using ESSs for microgrid frequency and voltage control under disturbance
conditions [7, 8], their control methods are usually based on fixed microgrid topologies.
Thus a microgrid with frequent system modifications or expansions may require multi-
ple re-tunings of the controllers. This can make numbers of studies that address adaptive
control approaches less practical [9–13]. It is of fundamental importance for an energy
storage system to have simple yet effective adaptive control methodologies in order to
make its microgrid truly robust under an unpredictable dynamic changing environment.
Unfortunately, this has not been available to date and is the subject of this dissertation.
The goal of this dissertation is to address the aforementioned shortcomings by propos-
ing a set of adaptive ESS control methodologies for both microgrid frequency and voltage

2
stability enhancements. This research has the following objectives:

1. To propose an autonomous adaptive frequency regulation strategy for fast microgrid

frequency regulation via control of a battery energy storage system (BESS).

2. To propose an autonomous adaptive voltage regulation strategy for BESS converters

to enhance both voltage recovery and induction motor (IM) speed recovery within
a fault-triggered islanded microgrid, as the continuation of the work in Objective 1.

3. To evaluate the proposed adaptive ESS frequency and voltage regulation control
methodologies with a realistic network topology.

This study has made the following contributions in the context of microgrid stability
enhancements via adaptive ESS control analysis, design and implementation:

1. Introduced an improved fast frequency regulation strategy that combines a phase-

locked loop (PLL) frequency measurement based adaptive nonlinear droop control
with a simultaneous fast secondary control using BESS. It demonstrates the ad-
ditional stability margin of the controller over the traditional frequency regulation
method without performance sacrifice.

2. Developed an autonomous adaptive voltage regulation strategy for BESS converters

that implements a simple yet effective way of achieving both faster voltage recovery
and longer maximum permissible fault clearing/islanding time for a microgrid with
one or more IMs.

3. Facilitated the microgrid design process by presenting the normalized minimum

BESS inverter sizing under a wide range of microgrid motor inertia, loading and
fault clearing times with both symmetrical and asymmetrical fault types.

3
 4. Investigated the microgrid dynamics under CIGRE medium voltage (MV) distribu-
tion network benchmark system [14] with unbalanced loads, multiple IMs, renew-
able generators with converter interfaces and distributed synchronous generators
with different prime resources (diesel fuel or hydropower). It also studied the im-
pacts of voltage regulation delays inside the adaptive voltage regulation algorithm
on microgrid dynamics.

5. Evaluated and demonstrated the robustness as well as expandability of the pro-

posed adaptive ESS frequency and voltage regulation control methodologies under
a meshed microgrid with fault-induced feeder switching and multiple contingencies
in addition to random power output from RESs.

1.2 Historical Background

1.2.1 Grid Evolution

There is little doubt regarding the assertion that an interconnected electric grid may be the
largest and most complex machine ever devised by man [15]. Since the birth of Thomas
Edison’s historic Pearl Street DC generation station in New York City on September 4,
1882, the electric power utility industry has grown to be one of the largest enterprises in
the world over the past 136 years [16]. The transformer development by William Stanley
Jr. as well as Nikola Tesla’s inventions of polyphase AC machines and related systems
broke the boundary of short distance DC power delivery to create an era of large-scale AC
electric power distribution, signified by the commissioning of the Niagara Power Project’s
three-phase AC long-distance power transmission system on August 26, 1895. The large-
scale electric grid was formed by gradually interconnecting isolated utility systems for
achieving better reliability and economies of scale.

4
 Several regulatory and market actions either shaped or redirected the grid evolution.
Those movements started with the regulated monopolies concept introduced by Samuel
Insull as the mark of the beginning of the regulation era following the Public Utility
Holding Company Act (PUHCA) of 1935. The deregulation took place after the 1973
oil crisis through the Public Utility Regulatory Policies Act (PURPA) of 1978 with the
Energy Policy Act of 1992 afterward. These enactments motivated engineers to develop
distributed generation and renewable energy techniques. Although the incentives for bet-
ter efficiency yielded to promotion of ”alternative” fuel sources, it inflated the cost of
building new infrastructures and hence discouraged the new investments on transmission
lines [17]. Consequently, the electric grid is evolving toward a more de-centralized na-
ture. Without proper actions to account these grid changes, the interconnected power grid
continuously operates near its capacity limits under this market-based system. The 2000
California energy crisis began to show somewhat problematic nature of the new compet-
itive energy market. A lacking grid infrastructures investments further contributed to the
Great Northeast power blackout of 2003. Those challenges all motivated the Energy Pol-
icy Act of 2005. Today, the need for a more reliable, high-quality electric power to digital
societies in an environmentally friendly and sustainable manner leads to the birth of the
idea of a ”Smart Grid” around the globe [18].

1.2.2 Distributed Generation (DG) and Storage

As previously stated, increasing amounts of electric power are generated from various
DERs since 1970s (e.g., In 2016, DERs accounted for about 2 % of the installed gen-
eration capacity in the United States 1 ). The major incentives include better use of the
diverse energy resources of different types for greater energy savings and delivering ef-
ficiency. The environmental concerns of electric power generation lead to the continued
1
source: https://www.ferc.gov/CalendarFiles/20180215112833-der-report.pdf

5
integration of renewable resources, especially wind and solar. DG can be categorized into
dispatchable and non-dispatchable generation. The diesel-fueled generators, for example,
are dispatchable. Most of the renewable generation, however, is non-dispatchable, with
the exception of hydroelectric power generation. From the technical perspective, different
DGs need different interfaces in order to be connected to their host utility grid. These can
be divided into three groups: (i) synchronous generators, (ii) induction machines or (iii)
power converters. Due to the increasing grid integration of non-synchronous generator
based DERs, the overall power grid inertia is decreasing. Both the randomness of non-
dispatchable resources and the reduction in power system inertia weaken the electrical
grid and create operational difficulties in maintaining a reliable electric power supply.
The traditional power grid has very limited energy storage capability (mostly in the
form of pumped hydro). However, the necessity of developing and deploying ESSs has
been widely accepted in the face of the grid operational challenges under high penetration
of DERs. It will be possible to make non-dispatchable resources dispatchable with the
help of storage. As one will see in this thesis, the ESS can be a critical component for
stable and reliable operation of a microgrid.
Both technical and economic aspects must be addressed in demonstrating the feasibil-
ity of both DGs and storage systems [19]. On the technical side, the goal is to effectively
manage the DGs and storage in order to maximize their operational benefits in the electric
grid. The microgrid concept presented in the following subsection serves as one of the
most viable solutions to this task.

1.2.3 Microgrids

The U.S. Department of Energy defines a microgrid as ”a group of interconnected loads

and distributed energy resources within clearly defined electrical boundaries that acts as a
single controllable entity with respect to the grid. A microgrid can connect and disconnect

6
from the grid to enable it to operate in both grid-connected or islanded mode [20].”
Microgrid systems have been identified as key components of the future ”Smart Grid”.
They provide multiple benefits that include [20]:

1. Capability for systematical integrating various types DERs and storage devices into
the grid.

2. Ensuring a greater grid resiliency under physical and cyber attacks.

3. Ability to be self-healing from power system disturbances.

4. Potential to serve as ”black start” entities for a faster restoration process during a
blackout.

5. Meeting the electric energy delivery quality needs for critical and sensitive loads.

6. Supporting bulk power system by providing ancillary services such as demand shift-
ing, deferral infrastructures investments, and frequency regulations.

1.3 Thesis Layout

The remaining six chapters of this dissertation are structured as follows:

Chapter 2 presents the literature summary related to the studied topics on energy stor-
age, converter control, microgrid operation and stability.
Chapter 3 reviews the fundamentals of the microgrid stability definition, modeling
concepts and control principles of its DERs as essential tools and background knowledge
for this research.
Chapter 4 proposes the adaptive frequency regulation strategy in order to help the mi-
crogrid frequency quickly return to its nominal value with a zero steady state error after

7
disturbances. This improved control methodology combines a PLL frequency measure-
ment based adaptive nonlinear droop control with a simultaneous fast secondary control
using BESS. Two case studies show the additional stability margin gained by adaptively
changing the controller gains based on the estimated PLL frequency deviation. Simula-
tion results reveal the robustness of the control under both fault induced islanding and
load switching during microgrid autonomous operation.
Chapter 5 introduces the new adaptive voltage regulation strategy for BESS converters
as the continuation of the work in Chapter 4. This is implemented by adaptively setting
the voltage regulator reference point after sensing the low voltage condition. This new au-
tonomous control strategy outperforms conventional voltage regulation by enabling IMs
to achieve a higher electromechanical torque for successful fault ride-through. The re-
sults also assist in the microgrid design process by determining the normalized minimum
converter size as a function of the IM inertia, loading, fault types and clearing time.
Chapter 6 evaluates the effectiveness and robustness of the overall proposed adap-
tive ESS control methods, combining both adaptive frequency regulation from Chapter 4
and adaptive voltage regulation from Chapter 5. The studied meshed microgrid is based
on the CIGRE MV distribution network with both synchronous generator and inverter
based resources. The microgrid loadings are mixed with RLC impedance and IM loads.
The results demonstrate significant enhancements of the microgrid stability and resiliency
under multiple disturbances with the help of the proposed BESS’s autonomous control.
The investigated transient events include fault-induced islanding (both symmetrical and
asymmetrical faults), isolated microgrid internal fault with topology change induced by
feeder switching, as well as, random power generation from RESs.
Chapter 7 summarizes the study and suggests future work. An appendix details the
system parameters including the simulated microgrid network, generators and load data.

8
Chapter 2

Literature Review

This presentation aims to give a comprehensive view of the current art and identify the
gaps regarding this dissertation’s research topics.

2.1 Energy Storage System (ESS)

The potential opportunities and operating benefits of grid-tied ESSs have long been well-
recognized [21]. The DOE/EPRI Electricity Storage Handbook in 2005 [8] provides an
extensive information resource guide on the stationary ESSs regarding their use cases,
benefits, technologies, and development stages. The costs and evaluation tools for both
the economic and technical analyses are also detailed. A recent technology development
review is given in [22] for high power storage devices such as supercapacitors and fly-
wheels. A good technical analysis regarding the flywheels design, modeling and associ-
ated passivity-based power electronic control for power system stability enhancement un-
der wind power plants is presented by Bachovchin from Carnegie Mellon University [23].
The study in [24] also uses a passivity-based control design method and shows the po-
tential to combine the application of STATCOM and battery energy storage to enhance

9
the transient stability of large-scale multimachine power systems with synchronous and
doubly-fed induction generators (DFIGs). The work in [25] shows a method for dimen-
sioning a storage device to provide a primary frequency reserve. Another piece of work
related to this high power ESS application on electric grid frequency regulation is given
by Leitermann from MIT [26], it highlights the characteristics of the frequency regula-
tion signal that are important for dividing the regulation loads between an energy storage
unit and a traditional generator. It also points out the potential for a near-instantaneous
frequency response from an ESS. This dissertation provides a novel adaptive ESS control
approach inside microgrids for a fast frequency regulation service.
The importance of ESS on microgrid operation and control is reflected in several
studies [27–31]. The ability of a storage device to provide inertial response by adding ad-
ditional converter control loop to account for frequency rate of change is shown in [27].
It mentions adaptive control for frequency controller parameter tuning as a future work.
The same group of researchers experimentally demonstrate the microgrid performance
enhancements and possible resynchronization accomplished by additional monitoring and
hysteresis triggering circuits [28]. A multiple ESS control study is conducted in [29]. It
uses a deadband droop in combination with SoC gain scheduling to set power references.
The Consortium for Electric Reliability Technology Solutions (CERTS) report [30] de-
tails the battery storage systems field testing results including parallel operation with syn-
chronous generators. The capacity limits of the storage converter must be respected for a
reliable operation. Finally, the authors in [31] investigate the short-term ESS frequency
control support in a French island with high penetration of wind or solar generation. It
demonstrates the benefits for high power storage devices on mitigating the impacts of
non-inertia type generation on the dynamic performance of island systems under major
generation outage.

10
2.2 Voltage-Sourced Converter (VSC) Control

The effective utilizations of power-electronic converters hold a key role on the integration
of DERs for a more flexible and reliable future power systems [32–36]. The authors in
[37] present an excellent early analysis on vector control for converters. Good overview of
control and synchronization methods can be found in [38, 39]. Several papers provide the
converter PLL design and analyses tools [40–43]. A novel grid synchronization method
by means of virtual oscillator control is presented in [44]. Some literature advocates
for the potential of a converter to mimic synchronous machine dynamics and control as a
”virtual” synchronous machine [45,46]. The work in [47] expands on this thesis converter
modeling and study by detailing the design and control guidelines for an LCL-filter-based
three-phase active rectifier. The converter scaling rule proposed in [48] further assists this
dissertation research’s sensitivity study on BESS sizing.
In the view of VSC in power systems, the book [49] by Yazdani gives a comprehen-
sive presentation for VSC modeling, control and applications. Load sharing techniques
for converters in both AC and DC grids have been investigated for decades under the um-
brella of droop control [50–58]. These works introduce several modified droop control
methodologies such as real/reactive current droop [52], signal injection and measure-
ment based droop control in [53], droop control with inverter output impedance turning
in AC grids [55, 56], as well as probable adaptive droop control in Multi-Terminal DC
(MTDC) grids [57, 58]. The studies in [59, 60] tackle the flexible converter real & reac-
tive power/current control methodologies during grid faults for a better fault ride-through.
The DERs’ VSC control strategies have profound impacts on microgrid dynamics and en-
ergy supply quality [61–68]. The report [61] gives a detailed summary as well as future
trends on the power converter microgrid operation modes and control structures with three
configurations: grid-forming, grid-feeding and grid-supporting.

11
2.3 Microgrid Operation

Extensive amount of literature focuses on the microgrid operation and energy manage-
ment system (EMS). The reports of [69–71] show some of the early work on CERTS
microgrid and its autonomous control strategy. Recent publications rectify this ongoing
effort by providing experimental results on model and control validation of the natural gas
generator set [72], updating the refinement of the EMS design and implementation with
parallel BESS, RES, as well as synchronous generators operating via traditional droop
control to achieve P lug − and − play functionality [73]. It indicates the importance for a
fast frequency regulation capability via batteries sources to improve the grid stability [73].
Another series of work from Katiraei [1, 74–76] provide good inspirations for his pioneer
study on real/reactive power management of DERs on microgrid dynamics. Its real power
control is based on the standard frequency droop with restoration integral control under a
large time constant. Moreover, three reactive power management strategies based on: (1).
voltage-droop, (2). voltage regulation and (3). power factor correction are explored.
A classic paper dealing with microgrid islanded operation with both primary and sec-
ondary control is [6]. In [77], the authors propose a voltage-power droop/frequency-
reactive power boost (VPD/FQB) control scheme to droop voltage reference for real
power sharing and frequency reference for reactive power sharing inside a converter fed
microgrid. The work in [78] introduces a combined angle-frequency droop controller for
both converters and synchronous generators using nonlinear control law for better power
sharing and more flexible control. The voltage and frequency dependent load impacts
on microgrid operation can be found in [79], where the dependency between microgrid
voltage and frequency as well as possible instability are shown to be caused by the load
dynamics. The propositions of utilizing a distributed secondary control system with lo-
cal communication are shown in [80, 81]. The study in [82] demonstrated the feasibility

12
for stand-alone DFIGs with autonomous frequency control during islanded operation via
PLL. Several papers [83–85] conducted good literature surveys on the present practice
and future trends of distributed and hierarchical control for microgrids.
The aforementioned works indicate the value of flexible converter-based dispatchable
DER control strategies on enhancing microgrid operations. This dissertation research on
the efficacy of BESS application for autonomous fast frequency regulation builds on these
past works and combines both distributed primary and secondary control using PLL.

2.4 Microgrid Stability and Control

The microgrid and bulk power system stability shares the same fundamental concepts
[86, 87]. Due to the high level penetration of converter interfaced DERs, many of the mi-
crogrid stability problems are either originated or heavily influenced by their converters’
dynamics and control [88–90]. Due to this unique aspect, many researchers have been
dedicating on enhancing the microgrid stability performance via modifying the convert-
ers control loops [36, 91–94]. Some new techniques and methods include introducing a
combination of repetitive and deadbeat control with feedforward compensation for distur-
bance rejection [92], secondary control based on potential-function [93], real-time small
signal stability analysis for droop gain schedule [91] and droop gain variation for increas-
ing virtual inertia [94].
There are various methods to analyze and predict the stability of microgrids [95–101].
All of them serve important roles on providing analytical tools for microgrid controllers
design. The small signal models of the converter based microgrid is well-established
in [96,97]. The active loads impacts on system small-signal stability are analyzed in [98],
it shows the active loads with a large dc voltage controller gain may lead to instability.
Similarly, the system is more prone to become unstable when the microgrid system unbal-

13
ance level increases [99]. A combination of small-signal and time-domain simulation are
usually used for a comprehensive stability analysis [7, 95, 99, 102, 103]. Another popular
analytic tool for grid-connected inverters is the impedance-based stability criterion [101],
which is based on an old concept from [104].
Numerous control methodologies have been proposed to enhance the microgrid stabil-
ity [9–13,105–107]. Two studies in [105,106] present novel frequency regulation methods
of microgrid by either utilizing the kinetic energy from the IM loads [105] or load voltage
sensitivity via voltage regulation [106]. Both methods can reduce the amount of needed
energy storage in microgrid. There are extensive references that fall within the category of
adaptively controlling converter in microgrid power system [9–13, 107]. An early adap-
tive droop adjustment based on the microgrid small signal model is introduced in [9].
The authors in [10] use the on-line grid parameter estimation combined with small signal
stability analysis to detect the islanding and adjust the inverters droop control parameters.
The droop scheduling scheme based on bifurcation theory is presented in [11] to find the
worst primary reserve share that is closest to instability. A mode-adaptive droop control
method was proposed in [12]. Two nonlinear droop controls are proposed in [13, 107]
to enhance microgrid frequency regulation. Nonetheless, it may either require multiple
small signal stability analyses to determine the acceptable droop gains [13] or lead to in-
stability due to a high droop gain when the frequency is close to the nominal value [107].
Both aspects are problematic for the practical implementations. On the contrary, the adap-
tive frequency regulation strategy developed in this dissertation aims to provide a simple,
efficient and robust control action under dynamic changing microgrid environments.

14
2.5 Chapter2 Summary

The prior work not only indicates a great potential for ESSs to enhance the microgrids
stability via fast converter control but also points out the need for a more flexible, simpler
yet more effective way of utilizing the limited converter capacity. Building on top of the
prior work, this dissertation research developed a set of adaptive ESS control strategies for
both microgrid frequency and voltage stability enhancements. This work focuses on the
real time control process under the inspirations from [108, 109]. In a microgrid that may
encounter frequent topology changes and load switchings, the control laws must also be
adaptable to those uncertainties. This is achieved by adjusting the BESS converters con-
trollers parameters adaptively via utilizing their PLL and terminal voltage measurements.

15
Chapter 3

Fundamentals of Microgrid Stability,

Modeling and Control

The purpose of this chapter is to provide essential background review for this dissertation
research. The presentation aims to:

1. Provide precise definition of microgrid stability problems investigated in this study.

2. Clarify microgrid modeling techniques as well as assumptions in this work regard-

ing the electric networks, loads, synchronous generators and BESS.

3. Review the synchronous generator’s governing control system and its primary droop
and secondary controls during power system frequency regulation process.

4. Present the synchronous generator’s voltage regulation methodology via excitation

control systems and associated power system stabilizer (PSS).

5. Discuss the overall power converter current − mode control and its real/reactive
power/current constrains during operation.

16
3.1 Microgrid Stability

This research considers the general microgrid stability as same as the formal power sys-
tem stability definition in [87], which is:”the ability of an electric power system, for a
given initial operating condition, to regain a state of operating equilibrium after being
subjected to a physical disturbance, with most system variables bounded so that practi-
cally the entire system remains intact.”
While analyzing the microgrid stability problems, one must take advantage of pre-
vious large-scale power system stability research efforts while identify their critical dis-
tinctions. In order to develop appropriate stability enhancement strategies, it is of funda-
mental importance to understand the physical nature of the instability phenomena that are
unique for microgrids, which are reflected in the thesis statements (Ch. 1.1). This dis-
sertation categorizes the microgrid stability problems based on previous work in [15, 87],
under following considerations:

• The physical nature and cause of the microgrid instability;

• The size of the disturbance relative to the studied microgrid systems;

• The devices, processes, and the time span that are critical for determining microgrid
stability; and

• The appropriate method to assess and predict the microgrid stability.

Figure 3.1 presents the overall picture of the microgrid system stability problem. The
unique origins of the microgrid instability and probable enhancement methods are em-
phasized in bold at the necessary places.

17
Figure 3.1: Microgrid stability problems with usual reasons and improvement strategies.

This study focuses on microgrids frequency and voltage stability. The frequency sta-
bility is ”the ability of a power system to maintain steady frequency following a severe
system upset resulting in a significant imbalance between generation and load. [87]” And
the voltage stability refers to ”the ability of a power system to maintain steady voltage at
all buses in the system after being subjected to a disturbance from a given initial operat-
ing condition. [87]” Similar to the bulk power system, the microgrid frequency stability is
largely depends on the real power demand-supply balance and the reactive power balance
is critical for microgrid voltage stability. Moreover, this thesis analyses make sure to keep
the overall microgrid stability in mind since different categories of microgrid instability
phenomena can have high degree of overlapping. For examples, during the low voltage
condition caused by the faults, the motors may lose stability by means of running toward
the stall and this indeed can further worse both microgrid voltage and frequency [110].
Moreover, the instability of the power converter controllers can cause microgrid syn-
chronous generators loss synchronism and lead to inadmissible overall system voltage

18
profiles at the same time.
In the case of high penetration of DERs with converter interfaces and stochastic power
generation from RESs, maintaining a stable microgrid frequency even during normal op-
erations can be difficult due to the significant reduction of microgrid rotational inertia as
well as slow real power control from synchronous generators. Similarly, the insufficient
microgrid capabilities to fast regulate system voltage after the contingencies can result a
sustained low voltage condition or even voltage collapse in the case with substantial IM
loads. This research considers the short-term microgrid frequency and voltage stabilities
in a time frame from a fraction of a second to a few seconds. The goal is to develop a
set of novel adaptive ESS control strategies for a fast and accurate real/reactive power
support to ensure a stable and reliable microgrid operation under severe disturbances.

3.2 Microgrid Modeling

3.2.1 Electric networks

The microgrid electric network is modeled under the general settings of the MV electric
power distribution [111]. The distributed parameter line models are used to model the
feeder [112]. To evaluate the expandability of the proposed control, this dissertation case
studies consider all three types of feeder, which include: (i). radial, (ii). loop and (iii).
meshed network [113].
Chapter 4 and 5 use a radial feeder topology, adapted from [1, 114], as the first plat-
form to carry on the proposed ESS adaptive control strategies studies under a microgrid
of full feeder scale. Chapter 6 expands the work into a full substation microgrid with
loop and/or meshed distribution systems based on the CIGRE MV distribution network
benchmark for the further control verifications [14].

19
3.2.2 Loads

The studied microgrid systems include both static and dynamic load models [15].

Static load model: RLC impedance

Instead of widely used ZIP (constant impedance, constant current, constant power)
model [115], this thesis chooses the linear RLC impedance as the static load model
to avoid the convergence issues during the low voltage conditions caused by the studied
contingencies (e.g., faults).

Dynamic load model: induction motor

The induction motor is the workforce of the industry and typically consume 60% to 70%
of the total energy supplied by a power system [15]. Therefore, this research consid-
ers three-phase single squirrel-cage induction machine as the dynamic load model. The
nonlinear differential equations based on reference-frame theory that described the induc-
tion machine are simulated [116]. The parameters of the machine are taken from [117],
summarized in Appendix A.7.

3.3 Synchronous Generator Modeling and Control

3.3.1 Modeling

The modeling and analysis of a synchronous machine has always been a central part in
understanding the power system operation and study the power system stability [118].
This work uses a 6th order state-space model as the electric part of the machine, repre-
sented in the rotor reference frame (dq frame) [119]. It takes the dynamics of the stator,
field and damper windings into considerations. The mechanical system is represented by

20
a single equivalent rigid mass as [120]:

2H d2 δ
= Ta = Pa (3.1)
ωs dt2

where:
H = synchronous machine inertia constant in (M J/M V A) or simply (s), as equal
to the machine kinetic energy in joules at the rated speed divided by the rated apparent
power of the machine.
δ = rotor angle, the electrical equivalent angular displacement of the rotor (radians)
with respect to the synchronously rotating reference axis.
Ta =per unit accelerating torque applied to the rotor.
Pa =per unit accelerating power applied to the rotor
ωs =synchronous angular velocity (radians/s)
and with:

d2 δ d∆ω d∆ω
= = ωs (3.2)
dt 2 dt dt

where:
∆ω = per unit angular speed deviation of the rotor.
One also deducts the following equation governing the alternator rotor dynamic [15]:

d∆ω 1
= Pa (3.3)
dt 2H

21
3.3.2 Real power and frequency control

Speed-droop governor [15]

Following the (3.3), the equivalent generators and load composite response to a small
rotor frequency deviation (denoted by ∆ω) can be mathematically expressed as:

Pa = ∆Pm − ∆Pe
(3.4)
∆Pe = ∆PL + D∆ω = ∆PL + ∆PD

where
∆Pm = Generator input mechanical power change (pu)
∆Pe = Generator output electrical power change including the losses (pu)
∆PL = Non-frequency-sensitive load change (pu)
∆PD = Frequency-sensitive load change (pu)
D = Load damping constant
Figure 3.2 shows the generator and its load response block diagrams.

Figure 3.2: Block diagram for transfer function relating (a). generator speed and power;
(b). generator speed and power including the effects of the load damping.

A speed-droop governor control system senses this generator per unit rotor speed de-
viation (∆ω) with respect to its per unit synchronous angular velocity (ωs ), inputs to a ser-
vomotor (modeled by an integrator with amplification gain K) and provide the valve/gate
position actuating signal ∆Y to adjust the generator input mechanical power Pm . Fig. 3.3
shows such speed-droop governor schematic.

22
 Figure 3.3: Speed-droop governor schematic

The steady-state feedback loop with droop gain Rs results a speed-droop governor
characteristic. The parameter Rs also refers as speed regulation expressed in percent %.
For example, Rs = 5% means the generator governor control system will change 100%
of its power output when experiences a 5% rotor frequency deviation. Appendix A.4
summarizes the hydraulic and diesel engine governor systems used in this study.

Primary droop control and power sharing [15]

With speed-droop governors, multiple synchronous generators can operate in parallel and
autonomously share the load change when they all experience a sudden rotor frequency
deviation. The concept of two units’ power sharing via primary droop control is explained
in Fig. 3.4.

23
 Figure 3.4: Traditional Droop Control Concept

It shows that each generator picks up load based on their own speed regulation Rs :

∆Pm1 Rs
= 2 (3.5)
∆Pm2 Rs1

Hence a generator governor with a smaller percent Rs will pick up more load change
during a frequency disturbance.

Secondary frequency restoration control [15]

A secondary control, also referred as automatic generation control (AGC) or load −

f requency control (LFC), adjusts the selective generators’ load reference setpoints to
restore the power system frequency after the primary droop control. The idea is illustrated
in Fig. 3.5 two units system. It shows the supplementary integral control action on units #
1. Meanwhile, it also represents the two units primary control loop and overall generator-
load response.

24
Figure 3.5: Two generators parallel operation with both primary and secondary frequency
control.

The design of secondary frequency restoration (i.e., AGC) control system via syn-
chronous generators for a interconnected power grid is a nontrivial task and must consider
both technical and economical aspects [121]. Comparing to the primary droop control,
the AGC regulation takes much longer time (30 seconds to minutes). On the contrary, it
is probable to utilize both primary and secondary frequency control actions via fast BESS
converter control to accomplish a rapid microgrid frequency regulation under dynamic
changing operational conditions, proposed in Chapter 4.

3.3.3 Reactive power and voltage control

An excitation system controls the synchronous machine terminal voltage and reactive
power. This work uses a DC excitation system (Type DC1C) for hydroelectric generator
and AC excitation system (Type AC1C) for diesel generator [122]. A multi-band power
system stabilizer (PSS) (Type P SS4C) are also equipped at hydroelectric generator’s
DC excitation system to enhance the damping of power system oscillation [119, 122].

25
The input signal to PSS is the generator shaft speed.
The generic DC and AC excitation control systems are shown in Fig. 3.6 [123].
Both of them are composed of an amplifier connected in series with a main excitor [1].
The associated excitation systems parameters (gains, time-constants, limiters, non-linear
function relations and constant coefficients) are summarized in Appendix A.5.

Figure 3.6: Generic excitation control system block diagram (a). DC type; (b). AC type

3.4 Battery Energy Storage System (BESS) Modeling and

Control

3.4.1 Converter modeling and current-mode control

This study is based on a simplified simulation model of the BESS as average model of
its converter. The overall converter modeling and control block diagram is summarized
in Fig.3.7 [49]. The BESS is represented as three linear, dependent voltage sources,
connected to PCC through series RL branches. This approach allows one to focus on the
converter control strategies’ impact on system performance without worrying about the
dynamics of the battery sources at the DC side and converter switching harmonics [49].
The figure also shows the associated d − q frame control block diagrams and signal flow.
The current − mode control is used to regulate the converter injected real and reactive

26
currents with respect to the PCC voltage. The reference real current (idref ) and reactive
current (iqref ) are obtained from converter frequency and voltage regulator respectively.
The settings of the PLL and current controller are adopted from [49]. Their parameters
can be found in Appendix A.6.

Figure 3.7: Overall converter modeling and control block diagrams

3.4.2 Real/Reactive power/current capacity limitations

The converter output current magnitude î can be expressed as:

q
î = (id )2 + (iq )2 (3.6)

where
id = Converter real current (pu)
iq = Converter reactive current (pu)
The main converter capacity constrain is its current magnitude î should not exceed its
maximum allowable value Imax . Under a fast current − mode control, the id ≈ idref ,
q
iq ≈ iqref and hence î ≈ îref = (idref )2 + (iqref )2 . The level of priority between idref
and iqref can significantly affect the BESS regulations on microgrid frequency and voltage

27
since the dynamic allocations of idref and iqref within the current limit will determine the
converter output real and reactive power in real-time.
One must realize that a stable real power transfer requires a well-established voltage.
Therefore, this study use the reactive power/current priority mode as the default condition.
However, it may be beneficial to switch to real power/current priority when the microgrid
experiences prolonged over/under-frequency. This aspect particularly reflects in Chapter
5 where a power/current priority switch condition is proposed.

3.5 Chapter3 Summary

This chapter provided necessary background knowledge to understand and conduct this
dissertation research by reviewing the fundamentals of microgrid stability, modeling and
control. First, it established the context of the research via defining the studied micro-
grid frequency and voltage stability problems. Second, it explained the modeling con-
cepts for microgrid network, loads, synchronous generators and BESS power convert-
ers. Third, it presented the synchronous generator governor/frequency control and ex-
citor/voltage control methodologies. Finally, it discussed the overall power converter
current − mode control and clarified the BESS converter’s controller priority mode set-
tings regarding its real/reactive power/current capacity limitations during operation.

28
Chapter 4

Adaptive Frequency Regulation

Governor droop control is widely used in the power system for primary frequency con-
trol while achieving the load sharing among synchronous generators [15] followed by
a slow acting secondary frequency control provided by automatic generation control
(AGC) [121, 124]. This approach can create frequency regulation issues in fast chang-
ing microgrid systems with small rotational inertia [1].
To tackle such challenges, this chapter introduces an improved fast frequency reg-
ulation strategy that combines a PLL frequency measurement based adaptive nonlinear
droop control with a fast simultaneous secondary control using BESS. In particular, this
study sheds light on stability enhancement of microgrids with synchronous generator.

4.1 Microgrid System for Chapter 4

Figure 4.1 represents the studied system single-line diagram. It shows a 15 kV class radial
feeder connected to its upper 69 kV utility grid (X/R = 22.2 ; 1000-MVA short-circuit
capacity) through substation step-up transformer. These distribution system parameters
were modified from [1, 114]. The line is divided by sections and protected by their dedi-

29
cated circuit breakers as numbered.

Figure 4.1: Single-Line diagram of the chapter 4’s studied microgrid system

A 3.125 MVA diesel generator equipped with excitation and governor control sys-
tem [125] is in operation at the feeder end. The governor has a droop setting: Rs = 1%.
Feeder loads are distributed along the main line with load #1 = 1.56 M W and 1.17 M var
(50% diesel generator nameplate capacity, power factor 0.8 inductive) and load #2 =
0.94 M W and 0.70 M var (30% diesel generator nameplate capacity, power factor 0.8
inductive). A 3.125 MVA BESS is installed at the point of common coupling (PCC), ca-
pable of controlling its real and reactive power exchange through the hosted network by
utilizing the PCC voltage phase angle information obtained from its PLL.

4.2 Adaptive Primary Nonlinear Droop Control

This and the following session describe the proposed adaptive fast frequency regulation
strategy used in the BESS converter’s power controller. All quantities are expressed in
pu, with base values of 13.8 kV , 10 M V A and 60Hz.
Figure 4.2 shows the new proposed adaptive nonlinear droop control strategy. The
block diagram is shown in Fig. 4.2(a). The converter nonlinear droop reference power
0
Pref 1 :

30
 0
Pref 1 = KA |f0 − fs |(f0 − fs )

= KA |∆f |∆f (4.1)

= KP ∆f

Figure 4.2: Adaptive Nonlinear Droop: (a) Block diagram, (b) Resultant Droop Curve

31
 0
The Pref 1 is converted to Pref 1 by:

0
Pref 1 = Pref 1 Pc (4.2)

where:
Pc is the BESS rated power in pu under the system base.
There is also a real time power limit which is denoted by Pmax :

Pmax = V̂ P CC Idmax (4.3)

Where V̂ P CC is the PCC voltage magnitude and Idmax is the maximum d axis current.
The droop gain KP is expressed as:

KP = KA |∆f | (4.4)

Equation (4.4) means that the droop gain is linearly proportional to the measured
absolute value of frequency deviation |∆f | from the converter PLL (after considering the
dead-band frequency fdb ). As a result, the converter output power (4.1) is proportional
to the measured frequency deviation squared. Unlike the previous work [107], when the
frequency is close to the nominal value, the droop gain of the converter’s power control
loop will be small, resulting in a smaller response compared to the traditional linear droop.
Meanwhile, as the measured frequency deviation ∆f is increasing, the converter power
controller gain is ramping up rapidly, thus leading to a more significant power output.
Such response helps to bring the microgrid frequency back to normal and the sensitivity
of the frequency control is not sacrificed.
The variable droop control gain leads to a nonlinear droop curve, Fig. 4.2(b). The
curve is plotted in generator reference, assuming a zero dead-band frequency (i.e., fdb =
0), with converter power on the X − axis and PLL frequency estimation on the Y − axis.

32
This configuration results in a negative droop curve slope. When the power is positive,
the converter is generating power output to its connected system as a generator. On the
other side, when the power is negative, the converter acts like a load which absorbs power.
At each point along the droop line, there is its equivalent droop R = |∆f /∆P |, defined
in the same way as the traditional droop control. The equivalent droop is getting smaller
as the absolute frequency deviation becoming larger, which is an indicator of a more
sensitive converter frequency regulating response during high frequency deviation. Near
the nominal frequency, the slope of the curve becomes infinite, corresponding to a zero
droop gain KP = 1/R.
The equivalent droop at the beginning of converter capacity saturation points (i.e.,
0 0
fs = fmax , Pref 1 = −1, Pref 1 = −Pc or fs = fmin , Pref 1 = 1, Pref 1 = Pc ) is defined as
capacity saturation droop, Rc , expressed in precent:.

r
1
P erccent Rc = × 100 = ∆fM × 100 (4.5)
KA

where:
∆fM : Maximum frequency deviation at the converter capacity saturation points, ∆fM =
fmax − f0 = f0 − fmin .
It is straightforward to determine KA by given Rc . The most important advantages of
this nonlinear droop method is that it provides additional stability margin around the equi-
librium point by dynamically changing the droop controller gain. It further avoids back
and forth oscillations that yield possible instability that have been observed in systems
using linear controllers when the frequency is close to the nominal value [88]. Further-
more, one may be able to achieve a smaller Rc without causing instability compared to
the linear droop case to achieve better frequency regulation.

33
4.3 Simultaneous Fast Secondary Frequency Control

This section introduces a fast secondary integral controller with a very small time constant
(sub-cycle). This is achieved by a fast converter control coupled with storage as the
primary fuel source. In addition, the secondary controller is acting at the same time as
the converter adaptive nonlinear droop control. Two probable approaches are presented
in the following two subsections. The first method uses a fixed integral controller gain
while the second strategy proposes an integral controller with adaptive gain.

4.3.1 Approach#1: Fixed Gain

If the fixed integral controller gain KI is used, the complete power controller block dia-
gram is shown in Fig. 4.3.

Figure 4.3: Converter Frequency Regulator Block Diagram

The final converter power reference Pref is the sum of both power references from
primary adaptive nonlinear droop Pref 1 and secondary controller Pref 2 :

Z t
Pref = Pref 1 + Pref 2 = KA |∆f |∆f Pc + KI ∆f dt (4.6)
0

34
This method brings the benefit of both sensitive primary control and fast secondary inte-
gral control together. For instance, with a sudden frequency deviation, the adaptive droop
control will first change converter power based on the nonlinear droop curve. Meanwhile,
the secondary controller quickly integrates the frequency errors and provide an additional
control signal to move the power set point at the nonlinear droop curve for a new system
equilibrium with zero frequency error just like the traditional AGC [121], however much
faster. In the end, one has the ability to use converter based generators for regulating
the isolated microgrid frequency around its nominal value under the dynamic changing
system conditions.

4.3.2 Approach#2: Adaptive Gain

The secondary gain can also be adaptive using the same idea from the previous presented
adaptive nonlinear droop control with the following modified block diagram, Fig. 4.4:

Figure 4.4: Adaptive Secondary Frequency Controller Block Diagram

The expression for the integral gain is:

35
 KI = KB |∆f | (4.7)

An excessive secondary control may also lead to undesirable frequency oscillation

in the power system [126]. This adaptive secondary control addresses such concern by
dynamically changing its integral controller gain KI . Equation (4.7) means that the sec-
ondary integral gain KI is linearly proportional to the measured absolute value of fre-
quency deviation |∆f | from the converter PLL. As the results,the converter will have a
large KI while the |∆f | is big. This contributes a quick integration of the frequency error
for the needed secondary power reference Pref 2 and hence also a more aggressive fre-
quency regulation action. Meanwhile, the KI is small when the system frequency is close
to its nominal value. This leads to a much milder response and a better stability margin
comparing to an integral controller with a fixed gain.
The value for KB can be determined by the desired KI value at a specific |∆f |. It
is often desirable to set a limit KImax as the maximum allowable value of KI to further
avoid the excessive control action. The case study on impacts of adaptive KI gain on the
stability of the converter controller is given in Ch. 6.2.1.

4.4 Cases Studies

To examine the effectiveness of the adaptive storage control on microgrid operation, a

fault induced islanding event is conducted in MATLAB® Simulink® model for the stud-
ied microgrid system shown in Fig. 4.1. It should be noted that since this chapter is
dedicated to frequency regulation strategy, the PCC voltage regulation scheme is adopted
from [49]. Prior to the fault, the system works in the steady state with local load #1 in
service and load #2 disconnected. The BESS is in the standby mode, connected to the
PCC and synchronized to the host utility by its PLL. The diesel synchronous generator

36
generates 2.5 MW (80% of its nameplate capacity) to the grid. At t = 0.55 s, a permanent
three phase-fault strikes on the feeder between circuit breaker #3 and #4. At t = 0.6 s, 3
cycles after the fault, abnormal conditions are detected and the BESS starts to activate its
converter for voltage and frequency regulation. The fault is cleared by opening breakers
#3 and #4 at t = 0.63 s, 5 cycles after the fault. As a result, part of the distribution
system functions as a microgrid at the downstream of the PCC. The microgrid is in au-
tonomous operation after the islanding. Finally, at t = 1.5 s, the load #2 is energized.

4.4.1 Case#A: Unstable BESS operation with linear droop

Figures 4.5 and 4.7 illustrate the unstable BESS performance for a linear droop control
with fdb = 6 × 10−4 pu (36 mHz), Rc = 1% and secondary controller with a fix gain
of 1000. The estimated PLL frequency fs in Hz is plotted in Fig. 4.5. The overall fs
unstable response versus time is shown in Fig. 4.5(a). Meanwhile, Fig. 4.5(b) and 4.5(c)
present the magnified fs oscillogram for 0.67 < t < 0.71 s and 1.48 < t < 1.52 s.
Fig. 4.5(a) reveals a significant frequency excursion during the fault (0.55 < t < 0.63 s)
and subsequent islanding transients. The fs returns to near the nominal value around
t = 0.67 s as shown in Fig. 4.5(b). It is also illustrated that the Rc = 1% of linear
droop leads to an unstable oscillation of fs centering around base frequency (60Hz), with
oscillating frequency higher than 400 Hz, thus resulting in poor voltage supply quality,
Fig. 4.6. The load #2 switching in at t = 1.5 s leads fs unstable oscillation to swing in
Fig. 4.5(c).

Figure 4.7 presents the BESS power versus time. From here it can be learned that
the unstable controller leads to the inadmissable power fluctuation. Before the load #2
is switched on at t = 1.5 s, the mean value of BESS power is negative, indicating its

37
Figure 4.5: Estimated PLL frequency fs for linear droop with Rc = 1% and secondary
controller gain of 1000: (a). Overall unstable response, (b). Magnified oscillogram for
0.67 < t < 0.71s, (c). Magnified oscillogram for 1.48 < t < 1.52s.

38
Figure 4.6: Magnified PCC voltage oscillogram vs. time: for 0.53 < t < 0.73s, inadmis-
sible waveform quality for linear droop with Rc = 1% and secondary controller gain of
1000.

Figure 4.7: Unstable control performance: The BESS power vs. time for linear droop
with Rc = 1% and secondary controller gain of 1000.

39
charging mode. Afterward, the mean power is positive, showing the BESS discharging as
a generator. Nevertheless, the Rc = 1% linear droop cause the BESS power controller to
be unstable.

4.4.2 Case#B: Stable BESS operation with nonlinear droop

In this second case study, the linear droop is replaced with an adaptive droop control
under the same fdb = 6 × 10−4 pu (36 mHz), Rc = 1% and secondary controller with a
fix gain of 1000. The responses of the BESS and the system performance, both with and
without BESS, are summarized in Figs. 4.8 to 4.12.
Figure 4.8 shows the converter PLL estimated frequency fs under adaptive nonlinear
droop control. Compared with previous case shown in Fig. 4.5(a), it is learned that the
PLL remains stable and fs quickly returns to 60 Hz with the help of the BESS. Despite
the load #2 switching on at t = 1.5 s, the PLL returns to normal operation after a short
disturbance.

The BESS power oscillogram is presented in Fig. 4.9. One will observe a stable
control performance of the BESS power controller. The power ramps up from negative
as a load to positive as a generator at t = 1.5 s to support the new load #2 in service.
It indicates that the proposed controller has the ability to quickly follow the load change.
Looking at both Fig. 4.7 and 4.9, it also reveals that the BESS power under adaptive
nonlinear droop control is a replica of the mean value of BESS power under linear droop
control. This adaptive control preserves the sensitivity of the regulation without causing
controller instability.

40
Figure 4.8: Estimated PLL frequency fs for adaptive nonlinear droop with Rc = 1% and
secondary controller gain of 1000.

41
Figure 4.9: Stable control performance: The BESS power vs. time for adaptive nonlinear
droop with Rc = 1% and secondary control gain of 1000.

The PCC voltage d − q components are compared in Fig. 4.10 for cases with adaptive
BESS control and without BESS. Excursions are observed at the switching instant, t =
1.5 s, for both d and q axis components. In the case with BESS, the voltage d−axis
component recovers to 5% of its nominal value after the fault induced islanding event
almost 100 times quicker than the case without the BESS. The new load #2 energized
at t = 1.5 s causes prolonged low voltage sags for the case without the BESS. With the
BESS contribution, after a brief disturbance at t = 1.5 s, the PCC voltage immediately
returns to normal.

42
Figure 4.10: PCC voltage dq components vs. time with and without BESS adaptive
control

Figure 4.11 depicts the magnified PCC voltage oscillograms versus the time with
and without BESS. Fig. 4.11(a) and 4.11(c) compare the system performance with and
without BESS before, during and after the fault for 0.53 < t < 0.73 s. It is shown that
in Fig. 4.11(a), the voltage returns within 5% of its nominal value within 1 cycle after
the islanding. Conversely, the voltage without the BESS stays below 0.8 pu during first
5 cycles after the islanding. The impacts of load #2 switching in at t = 1.5 s is shown
in Fig. 4.11(b) and 4.11(d). In Fig. 4.11(b), thanks to BESS performance, the voltage
returns to nominal value within 1 cycle. Without BESS, the additional loading causes the
obvious voltage sags, Fig. 4.11(d).

43
Figure 4.11: Magnified PCC voltage oscillogram vs. time: (a). With BESS adaptive
control, for 0.53 < t < 0.73s, (b). With BESS adaptive control, for load#2 switched in
at t = 1.5s, (c). Without BESS, for 0.53 < t < 0.73s, (d). Without BESS, for load#2
switched in at t = 1.5s.

Finally in Fig. 4.12, the distributed synchronous generator rotor speed is compared
with and without BESS. The three-phase fault causes a severe generator speed excursion
that measured over 1 Hz before clearing. With the BESS adaptive control, the rotor
frequency quickly returns to 60 Hz within 0.5 s after the islanding. Even with a large
load #2 (30% of the generator nameplate) when switching on at t = 1.5s, the generator
speed varies less than 0.1 Hz when the BESS is used. Without the BESS, the alternator
takes considerably longer to regulate its speed despite a fast acting diesel unit in service
with a Rs = 1% droop. Moreover, the additional load #2 leads to a speed change of

44
Figure 4.12: Synchronous generator rotor frequency fr with and without BESS adaptive
control

more than 0.75 Hz. All of the above results demonstrate the superior performance of the
proposed adaptive energy storage control method.

4.5 Chapter4 Summary

It has been demonstrated that with the droop gain depending on the PLL frequency es-
timation, the adaptive nonlinear droop control gains additional stability margin over the
traditional linear droop control by its flexibility. Moreover, this chapter highlights the op-
erational benefits of adaptably controlling the BESS on the frequency stability enhance-
ment of the microgrid.

45
Chapter 5

Adaptive Voltage Regulation

This report extends the previous chapter 4’s solution [127,128] by describing a new adap-
tive voltage regulation strategy that goes hand in hand with IM speed recovery after the
islanding. The specific contributions of this advocated adaptive voltage regulation ap-
proach are:

1. An increase in the microgrid motor’s probability of riding-through the fault-induced

islanding even under high load torque with prolonged fault clearing time, thereby
increasing the maximum permissible fault cleaning time while assuring that the
islanded microgrid motors’ speeds recover.

2. Provision for an autonomous control strategy, with no additional sensors and mon-
itoring systems needed to achieve the benefits for the IMs.

3. Capability for modifications with a broad adaptability into various power system
configurations based on the required standards [5].

46
 Figure 5.1: Single-Line diagram of the chapter 5’s studied system

5.1 Microgrid System for Chapter 5

Figure 5.1 represents the studied system single-line diagram. It shows a 15 kV class
radial distribution feeder connected to 69 kV utility source (X/R = 22.2; 1000-MVA
short-circuit capacity) via a substation step-up transformer. These distribution system
parameters (see Appendix I) were adapted from [1, 114]. The loadings of the feeder are
mixed with both linear RLC impedance and IM loads. There are two 2.3 kV IMs [117],
connected to the 13.8 kV feeder through their transformers; a large unit with 2250 HP
rated power and a smaller motor with 500 HP rated power.
A 3.125 M V A diesel generator equipped with excitation and governor control system
[125] is in operation at the feeder end. The governor has a droop setting of 1%. A BESS is
installed at the PCC for both frequency and voltage regulation. The feeder is sectionized
by its dedicated circuit breakers as numbered. It is assumed that a fault happened between
the circuit breaker #3 and #4 leading to fault-induced islanding. These system settings
capture the essential features of an autonomous microgrid with multiple IM sizes. This
topology enables one to test the performance, verify the concept and conduct sensitivity
studies for the proposed adaptive voltage regulation strategy without loss of generality.

47
5.2 Induction Motor (IM) loads Impacts on Microgrid

Dynamics and Stability

It has been learned that load characteristics have significant impacts on the dynamics and
the stability of a microgrid [129]. The extensive use of IMs, and their nonlinear transient
behavior have triggered special interest [7]. During the low voltage condition caused by
the fault, the microgrid IMs may begin to stall and consume excessive reactive power be-
fore tripping off, which leads to fault induced delayed voltage recovery (FIDVR) or even
voltage collapse [130]. It is essential for a truly reliable microgrid to be able to achieve a
successful motor fault ride-through under a broad range of conditions. Previously, several
approaches have been proposed in order to avoid motor stalling after the fault-induced
islanding [7, 131]. For example, [131] utilizes dynamic inverter frequency settings inside
an inverter based microgrid that are meant to increase the motor torque at a low speed.
The study in this chapter also focuses on the microgrid voltage stability issue caused by
the IM stalling.

5.3 Concept

Figure 5.2 shows the one-line diagrams of a 2250 HP IM connected to a stiff voltage
source Vs either directly or through a step-down transformer (8% short circuit reactance).
The machine parameters are taken from [117].

48
Figure 5.2: Induction motor free accelerating experiment single-line diagram: (a). Direct
start, (b). With step-down transformer.

The idea behind utilizing adaptive voltage regulation to enhance the IM fault ride-
through can be illustrated by simulating the IM free acceleration experiments under the
two configurations shown in Fig. 5.2. Fig. 5.3 plots the free acceleration characteristics
of the motor under 0.8 < Vs < 1.2 pu with either direct connection, Fig. 5.2(a), or use
of a transformer, Fig. 5.2(b). It illustrates the impacts of supply voltage Vs magnitude on
the IM transient torque-speed curves. It is evident that when the IM is equipped with a
transformer, the motor operates with a lower terminal input voltage VT and the developed
electromechanical torque is less. For stable motor operation with a constant torque load
TM equal to its rated torque Tr , it is necessary for the breakdown torque Tp > TM = Tr
to result in two equilibrium points. The left equilibrium point with a positive torque-
speed curve slope is unstable while the right point is stable with a negative slope. Since
the torque of the machine is proportional to the magnitude squared of its stator terminal
voltage VT [116], both Tp and the speed range in between the two equilibrium points are
reduced significantly with a reduced Vs magnitude. The above observations indicate that
the motor is prone to be disconnected under a high torque load and low supply voltage,

49
especially with a transformer connection. With a supply voltage Vs > 1.0 pu, the mo-
tor will have higher electromechanical torque Te at a lower speed, resulting in a higher
stability margin, Fig. 5.3.
When there is a small dip in the microgrid motor terminal voltage VT , the motor may
be able to slow slightly until reaching a new stable equilibrium point at a lower voltage
and speed without stalling. However, under the power system fault, the terminal voltage
of the IM can experience a severe sag. This condition may lead to Tp < TM , and hence
a continuing decline of the motor speed. The speed reduction rate is dependent on the
unbalance level between Te and TM , which is affected by the fault types, clearing speed,
as well as, the motor inertia and loading. If Vs cannot recover sufficiently in order to
develop a Te > TM , the motor will be tripped [110].
In reality, the motor supply voltage Vs has a short circuit source impedance Zs . Sim-
ilar to the transformer impedance effects in Fig. 5.2(b), the excessive reactive power
consumed by the IM at a low speed will result in a large current flow and inadmissible
voltage drop across Zs . Nevertheless, if one can regulate the post-islanding Vs > 1.0 pu
for a time even as brief as a few cycles, it will boost the voltage across the motor termi-
nals. This higher VT will contribute to a higher electromechanical torque Te > TM during
the low speed condition that may permit a successful disturbance ride-through. An illus-
trative case is given in Fig. 5.4. It plots four torque-speed characteristics of an IM under
terminal voltage VT of different magnitudes.
Assuming that the motor carries a full load torque TM = Tr , it will operate at the
intersection of curve 1 and load torque within the stable region during normal operation.
If a fault happens near the motor terminal, its VT will experience a severe sag and the
motor will operate under the curve 2. With Tp < TM , the motor speed will decline. Under
the post-fault recovery voltage #1, curve 3, the IM can recover before its speed reaching
Su1 . Curve 4 is under the post-fault recovery voltage #2 with a higher magnitude and the

50
 Figure 5.3: Induction motor free accelerating torque-speed curves

Figure 5.4: Torque-speed characteristics for an induction motor under different terminal
voltage VT

51
 Figure 5.5: BESS PCC voltage regulator

developed electromechanical torque. Thanks to this higher voltage, the motor can recover
before reaching to the Su2 with a longer fault clearing time. This case shows that it is
feasible to gain additional stability margin by adaptively regulating a higher post-fault VT
which is the incentive of proposed adaptive voltage regulation methodology.
This concept requires additional resources that can provide a large and rapid dynamic
reactive current support. The traditional synchronous generator excitation control for
voltage regulation is slow and thus is not suitable for this task [122]. Luckily, with a fast
current − mode control ability [49], the BESS converter is an ideal candidate for this
type of voltage support. By helping both voltage recovery and IM speed recovery, this
concept improves the stability as well as the resiliency of the overall microgrid.

5.4 Implementation

Figure 5.5 shows the block diagram of the BESS PCC voltage controller. All quantities
are expressed in pu, with base values of 13.8 kV , 10 M V A and 60Hz.
The proposed adaptable voltage regulation strategy is implemented by setting the
BESS voltage reference point Vref (t) > 1.0 pu following the fault induced islanding. The
converter will act like an over-excited synchronous machine, with much quicker reaction
time, supplying dynamic reactive current/power to its PCC within its current rating.

52
5.4.1 Algorithm for adaptive voltage regulation set point: Vref (t)

The BESS is operated under reactive power priority mode with two possible operat-
ing states: normal or emergency. The algorithm for determining the Vref (t), based on
the previously described adaptive voltage regulation concept can be summarized in Fig.
5.6. During the normal operation when the PCC voltage magnitude V̂ P CC > VthL , the
PCC voltage regulator reference Vref (t) stays at VrefN (usually VrefN = 1.0 pu) as the
normal operation mode. If the V̂ P CC < VthL , the regulator will switch to emergency
mode and the Vref (t) will ramp up at a rate of KR to VrefE , a value higher than VrefN .
Then the BESS converter will inject reactive current/power within its current limit (i.e.,
iqref ≤ Iqmax = Imax ; Qmax = V̂ P CC Iqmax ) to regulate PCC voltage toward VrefE . To
avoid the excessive over-voltage condition caused by this operation, the algorithm will
ramp down the Vref (t) as soon as the PCC voltage magnitude hits an upper threshold,
V̂ P CC ≥ VthH with VrefN < VthH ≤ VrefE . It should also be noted that the algorithm
must carefully take the system standards [5,132] and motor protection schemes [133,134]
into consideration so that the actual over voltage duration is permissible (i.e., not trigger-
ing unwanted protections or sacrificing motor performance and life [135]). Moreover, the
Vref (t) will ramp back down to VrefN at the rate of KR if at least one of the following
three conditions is true so that the BESS inverter capacity can be utilized for frequency
regulation:

1. Insufficient capacity for over-excitation: V̂ P CC < (VrefN − Vdb ) for a time duration
trs1 , where Vdb is the voltage regulator deadband.

2. Capacity saturation around normal voltage reference: îref ≥ Isat and |V̂ P CC −
q
VrefN | ≤ Vdb for a time duration trs2 , where îref = (idref )2 + (iqref )2 and Isat ≤
Imax .

53
Figure 5.6: Vref (t) operating state decision tree
54
 3. Insufficient capacity for a successful PCC voltage regulation to VthH : Vref (t) =
VrefE but V̂ P CC < VthH for a time duration trs3 .

Considering the excessive amount of reactive power required by the IMs at a reduced
speed, the current capacity of the BESS converter may not even be large enough for a
pumped PCC voltage regulation to a value greater than VrefN . The condition 1 accounts
for such situation and adaptably ramps down the Vref (t). Meanwhile, it does not conflict
with the contributions that a BESS converter can bring for the voltage recovery and IM
speed recovery since the converter has already utilized all its capacity for the microgrid
stability enhancement.
The condition 2 is mainly caused by probable different voltage regulation references
among the generator units in the system and thus the incentive for ramping down the
Vref (t) under such case is for a better voltage control co-ordination among the power
system apparatuses inside the microgrid.
Despite the fact that the capacity of the converter may be large enough to successfully
regulate V̂ P CC > VrefN following the disturbance, it may still not be capable to contribute
to V̂ P CC > VrefH in a timely manner, therefore it is appropriate to also return the Vref (t)
back to VrefN for a better microgrid voltage profile as indicated in condition 3.
It is worth emphasizing again that this proposed adaptive autonomous voltage regula-
tion strategy is for a rapid microgrid stability enhancement accounting the IMs speed re-
covery. It does not fully address the optimization issues related to reactive power reserve
which traditionally disposed by many centralized voltage control methods that require
communications and hence also a longer regulation time [136].
Finally, even through it is feasible to also change the voltage regulator gain KVI (t)
during the control for a more robust action against possible controller instability, this
work uses a fixed gain: KVI = 420, determined by the design method promoted in [49].

55
5.4.2 Alternative: With Regulation Delays and Power Priority Switch

The previous algorithm for Vref (t) determines that the Vref (t) will ramp up from VrefN
to VrefE whenever the V̂ P CC < VthL . Similarly, Vref (t) will ramp down from VrefE
to VrefN whenever the V̂ P CC ≥ VthH . Moreover, the converter is assumed to operate
under reactive power priority mode. This section introduces an alternative approach that
accounts the voltage regulation delays and power priority switch that could have profound
impacts on the operations of the BESS and hence the real-time microgrid dynamics.
Fig. 5.7 shows this alternative Vref (t) algorithm. The following new conditions are
implemented:

1. Voltage regulation delay #1, trd1 : The voltage regulator will switch to emergency
mode and ramp up Vref (t) at a rate of KR from VrefN to VrefE only after V̂ P CC <
VthL for an aggregate of more than trd1 s.

2. Voltage regulation delay #2, trd2 : The voltage regulator will return to normal mode
and ramp down Vref (t) at a rate of KR from VrefE to VrefN only after V̂ P CC ≥ VthH
for an aggregate of more than trd2 s.

3. Power Priority Switch: The converter will switch from reactive power priority mode
to real power priority mode and subsequently set voltage regulator to normal oper-
0
ation mode when îref ≥ Isat and |∆f | ≥ fsw for an aggregate of more than tsw1 s,
0
where ∆f = f0 − fs and fsw is the frequency deviation threshold value based on
the standards.

The incentive for the new voltage regulation delays, trd1 and trd2 , is to avoid unnec-
essary mode switching for the voltage regulator. The voltage reference doesn’t need to
ramp up high when the voltage sag duration is short and hence the normal voltage regula-
tion could have already contributed to a successful IM recovery. It could also be helpful

56
Figure 5.7: The alternative Vref (t) decision tree

57
for Vref (t) to stay at VrefE for a short period of time to ensure a better voltage regulation
coordinations with multiple distributed DERs at different locations inside the microgrid.
Furthermore, it is necessary to dedicate BESS mainly for frequency regulation when the
microgrid experiences inadmissible frequency fluctuations, which is addressed by the new
condition 3. To illustrate the effects of those new sets of conditions on the behaviors of
BESS converter and microgrid operation, two case studies are presented in Ch. 6.3.

5.5 Simulation Studies

This section summarizes case studies for Fig. 5.1 microgrid performance. Three cases
demonstrate the microgrid transients with one or multiple IMs, showing the effectiveness
and advantages of the BESS adaptive voltage regulation on motor post-islanding recov-
ery. Two additional sensitivity studies help to further estimate the normalized minimum
required BESS inverter size for successful microgrid motors fault ride-through under both
symmetrical and asymmetrical faults.
Prior to the fault, the system works in steady state with diesel synchronous generator,
generating 2.5 MW (80% of its nameplate capacity) to the grid. The BESS is in the
standby mode, connected to the PCC and synchronized to the host utility by its PLL. At
t = 0.55 s, a permanent fault strikes the middle of the feeder between circuit breaker
#3 and #4. At t = 0.6 s, 3 cycles after the fault, abnormal conditions are detected
and the BESS starts to activate its converter for voltage and frequency regulation. The
fault is cleared by opening breakers #3 and #4. As a result, part of the distribution
system functions as a microgrid downstream of the PCC. The studies are based on Vref (t)
algorithm in Fig. 5.6 with the following voltage regulator parameters: VrefN = 1.00 pu,
VrefE = 1.35 pu, Vdb = 0.05 pu, VthL = 0.50 pu, VthH = VrefE − Vdb = 1.30 pu,
KR = 7 V /s, trs1 = 0.100 s (6 cycles), trs2 = 0.067 s (4 cycles), trs3 = 1.000 s and

58
Isat = 0.90 Imax .

5.5.1 Transient Study#A: Impacts of BESS size on motor speed re-

covery

In this first case, only the 2250 HP IM1 is in operation with load torque equals to 90% of
its rated torque, TM1 = 0.9 Tr1 . The microgrid is islanded 3 cycles after the fault. Figs.
5.8 to 5.14 illustrate the responses of the BESS and the system performance, both without
and with different sizes of BESS.
Figure 5.8 presents the IM1 rotor speed, ωIM1 , without and with 3 different sizes
of BESS inverter in the range of (0.5 < SB < 2.5 M V A) in steps of 1 M V A. One
can find the 2.5 M V A BESS inverter interface impedance to PCC and the associated
current − control scheme parameters in the Appendix A.6. These values are scaled
according to [48] in order to account for the inverter power rating changes. The speed
ωIM1 is unable to recover if SB < 2.5 M V A. It is also learned that ωIM1 can reach a
higher value after islanding with a larger inverter’s support, indicating that the motor is
more likely to ride through the disturbance with a large BESS.
In order to understand the key implementation stages of the adaptive voltage regu-
lation, it is instructive to observe the PCC voltage profiles, summarized in Fig. 5.9 for
scenarios without and with a 2.5 MVA BESS. The fault causes a severe voltage sag.
Without the BESS, the voltage oscillates around 0.5 pu even after the islanding due to the
excessive reactive power consumed by the motor. This is resolved by tripping the motor
slightly after t = 1.3 s. Even by doing so, the PCC voltage magnitude still does not
return to its nominal value in a timely manner. Instead, it experiences a large swing due
to the slow action of the diesel generator excitation system [125]. On the other hand, the
V̂ P CC quickly recovers with the help of the BESS after the islanding. One can see that the

59
 Figure 5.8: Induction motor speed ωIM1 under different BESS sizes

Vref (t) ramps up from VrefN to VrefE after sensing the low voltage condition at t = 0.6 s,
V̂ P CC < VthL = 0.5 pu. In this specific case, the 2.5 MVA BESS is not large enough to
regulate V̂ P CC > (VrefN − Vdb ) = 0.95 pu within the time duration of trs1 = 0.1 s and
thus Vref (t) ramps down at t = 0.7 s. The voltage regulator then helps the V̂ P CC profile
back to VrefN as the IM1 recovers its speed.
The oscillograms of IM1 electromechanical torque Te1 with and without a 2.5 MVA
BESS are presented in Fig. 5.10. Thanks to the direct contribution of the BESS to the
voltage recovery, it clearly shows the immediate boost of the torque after the fault clear-
ing. This contributes to a successful motor fault ride-through. On the other hand, without
the BESS, the prolonged voltage sag leads to an unrecoverable motor torque Te1 .
The IM1 electromagnetic torques Te1 (%) versus speed ωIM1 (%) in Fig. 5.11 gives
an alternative view of BESS influences on the induction motor speed recovery. A new
stable equilibrium point is achieved with the help of the BESS after the islanding. It

60
 Figure 5.9: V̂ P CC with and without 2.5 MVA BESS

Figure 5.10: Induction motor torque (Te1 ) versus time with and without 2.5 MVA BESS

61
shows that the torque trajectory increases to the level greater than the load torque which
helps the motor return to its original speed. However, without the BESS, the delivered
motor torque remains low. As a consequence, its speed is decreasing toward the tripping
condition.
The BESS responses during the above event are captured in Figs. 5.12 and 5.13. The
converter d/q current components in time are plotted in Fig. 5.12. After sensing the
fault, the iq ramps to the negative converter current capacity limit to generate the needed
maximum reactive power for the PCC voltage support. This condition is reflected in
Fig. 5.13(b) reactive power response. The reactive power magnitude is increasing as the
PCC voltage is recovering and eventually reaching the converter apparent power rating
(0.25 pu).
The BESS frequency regulator [127] is activated immediately after the V̂ P CC recovers
to VrefN , Fig. 5.9. This condition results in the dynamic allocation between id and iq , Fig.
5.12, meant to achieve both microgrid frequency and voltage control. Fig. 5.13(a) and
(b) show the associated converter output real and reactive power in time. Their responses
closely follow the reference signals.

62
Figure 5.11: Induction motor torque (Te1 ) versus speed (ωIM1 ) with and without 2.5 MVA
BESS

Figure 5.12: BESS converter current d/q frame components versus time

63
 To complete the picture, the diesel generator frequency fr oscillograms with and with-
out the BESS are shown in Fig. 5.14. Without the BESS, the generator frequency experi-
ences an excursion that is close to 2 Hz. With 2.5 M V A BESS, the generator frequency
stays within 0.5 Hz deviation from the nominal 60 Hz. This clearly shows the beneficial
contribution of the BESS to the microgrid frequency stability [127].

64
Figure 5.13: BESS power versus time: (a). Real power, (b). Reactive power.

65
 Figure 5.14: Generator frequency fr versus time with and without 2.5 MVA BESS

5.5.2 Transient Study#B: Advantage of adaptive voltage regulation

The second Transient Analysis demonstrates the advantages of adaptive voltage regula-
tion over the normal voltage regulation on IM recovery enhancement. The microgrid is
defined by the same parameters as in Transient Analysis 1 except that the three-phase
fault clearing takes 7 cycles (4 cycles longer). Consequently, it requires a large 7 M V A
BESS to recover the motor, determined through the same time-domain simulation ap-
proach used in the Transient Analysis 1. Considering that the total amount of load inside
the islanded microgrid is only 2.47 M W , the BESS size may seem to be too large to be
economical. However, if one considers the fact that for most of the time the microgrid
is operated under the grid-connected mode where the BESS can participate in multiple
grid side services such as load shifting, outage avoidance or feeder deferral [19], then it
is possible that a microgrid with a large BESS may be economically feasible. The BESS
responses and the microgrid dynamics during the islanding event are presented in Figs.

66
5.15 to 5.19 under either adaptive or normal voltage regulation.
Figure 5.15 presents the IM1 speed during transients. It shows that the IM1 speed can
recover under the adaptive voltage regulation but not with normal voltage regulation. One
can interpret the concept by studying both V̂ P CC profiles in Fig. 5.16 and motor torque
responses in Fig. 5.18. The prolonged fault clearing leads to a rapid decline of the motor
speed. Thus, even through with the BESS’s help, the V̂ P CC can recover to nominal value
immediately after the islanding under normal voltage regulation, Fig. 5.16, the nonlinear
transient torque Te1 is still not high enough to overcome the motor load torque TM1 , Fig.
5.18. As a consequence, the motor speed will decrease to the point of tripping. On
the contrary, adaptive voltage regulation increases the Vref (t) during the post-islanding
recovery phase. It helps the V̂ P CC ramp up to a higher value as shown in Fig. 5.16.
This contributes to Te1 > TM1 in Fig. 5.18; hence the motor can accelerate even from
a relatively low speed with a large load torque. The Vref (t) immediately ramps down as
soon as the V̂ P CC > VthH = 1.30 pu. Moreover, the detailed oscillograms for the VP CC
voltage waveform during this recovery phase (1.40 < t < 1.55 s) is shown in Fig. 5.17.
It illustrates the adaptive voltage regulation on the control of the PCC voltage magnitude
without sacrificing the waveform quality during the transient. Additional simulations
indicate that the critical islanding time for this case under normal voltage regulation is
6 cycles. It is 10 cycles with the adaptive voltage regulation. With an improvement of
66 %, this justifies the benefits of utilizing adaptive voltage regulation on enhancing IM
speed recovery.
Looking at the BESS converter current d/q frame components in time, Fig. 5.19, one
can see that part of the difference between the adaptive and normal voltage regulation
strategies lies in the different real time allocations between the real and reactive current
components within the converter current capacity. With the adaptive voltage regulation,
the converter utilizes all its capacity to regulate PCC voltage toward to a higher level. The

67
Figure 5.15: Induction motor speed ωIM1 with and without adaptive voltage regulation

Figure 5.16: V̂ P CC with and without adaptive voltage regulation

68
Figure 5.17: PCC voltage waveform oscillograms under adaptive voltage regulation,
1.40 < t < 1.55 s.

Figure 5.18: Torque of induction motor (Te1 ) versus time with and without adaptive volt-
age regulation

69
Figure 5.19: BESS converter current d/q frame components versus time, with and without
adaptive voltage regulation

maximum reactive power (maximum iq ) support duration is significantly longer compared

to the normal voltage regulation. This contributes to a successful motor fault ride-through.
Overall, this case study helps one to understand the concept as well as the benefits of
adaptive voltage regulation on enhancing the microgrid’s IM speed recovery for achieving
an uninterrupted high-quality power supply.

5.5.3 Sensitivity Study#A: Minimum BESS size under symmetrical

fault

This case study assumes that IM1 is carrying a constant torque load PIM1 and it considers
four different PIM1 under the range of 30%Pr1 ≤ PIM1 ≤ 90%Pr1 in steps of 20%Pr1 ,
where Pr1 is the IM1 rated power. This will result in the motor load making up from 34%
to 61% of the total microgrid real power demand.

70
 As in the previous Transient Analyses #A and #B, multiple simulations were con-
ducted to find the minimum BESS inverter size for a successful motor recovery. The
results under different motor loadings are then normalized with respect to their total mi-
crogrid real power demand respectively and summarized in Fig. 5.20’s four sub-graphs.
Each graph plots the normalized minimum BESS converter size S B versus equivalent
motor inertia constant HIME [15], defined as:

stored kinetic energy in MJ at operated speed

HIME = (5.1)
motor loading in MW

or
1
PN 2
2 n=1 JIMn ωomn
HIME = PN (s) (5.2)
n=1 PIMn

where
0
JIMn = KJn JIM n
(5.3)

is the total polar moment of inertia for the nth microgrid motor, including its rotor plus
mechanical load and ωomn is the motor operated speed at the load of PIMn .
For this sensitivity study with only IM1 :

1
J ω2
2 IM1 om1
HIME = (s) (5.4)
PIM1

The effect of motor inertia is determined by simulating cases where the total inertia
0
of IM1 ’s rotor and load combined is equal to KJ1 times the base JIM1
(Appendix A.7) as
in (5.3). KJ1 ranges from 1-19 in steps of 3. Family of curves under fault clearing time
from 3-9 cycles in steps of 2 cycles are plotted in each sub-figure, Fig. 5.20.

71
Figure 5.20: Normalized minimum BESS inverter size S B as a function of equivalent
motor inertia constant HIME with fault clearing time as the parameter: (a). 34 % IM
loading, (b). 47 % IM loading, (c). 55 % IM loading, (d). 61 % IM loading.

The combined impacts of motor inertia, loading and fault clearing time can be sub-
0
stantial. For example, if the IM1 with base inertia JIM1 = JIM1
(HIME = 0.75 s) is 90%
loaded (which accounts for approximately 60% of the total microgrid loading), the re-
quired minimum BESS inverter can be as high as 300% of the original microgrid demand
under 9 cycles of fault clearing.
The results can be summarized in the following three rules:

1. A high inertia motor will require a smaller BESS.

2. As the fault clearing time increases, the motor will need a larger BESS for a suc-
cessful ride-through.

72
 3. The microgrid loading has direct impacts on the required BESS inverter size. A
microgrid with larger percentage of IM loading demands a bigger size inverter to
ride through the fault-triggered islanding incident.

An important advantage in generating the normalized minimum BESS converter size

results, like one in Fig. 5.20, is that this kind of comprehensive presentation provides
useful engineering guidelines for approximate sizing of the BESS for a given microgrid.
Later on, these results are tested by sizing the BESS based on an additional transient case
study with a loading of two IMs.

5.5.4 Sensitivity Study#B: Minimum BESS size under asymmetrical

fault

If the fault type is asymmetrical, for example a single-phase to ground fault instead of
three-phase, the minimum required BESS size versus fault clearing time is plotted in Fig.
5.21. The system is operated with the IM1 under 90% loading and HIME = 0.75 s.
Comparing Figs. 5.20 and 5.21, one will observe the significant reduction of the size
of the needed for BESS under this less severe fault, even though the fault clearing time
can be much longer in the single-phase fault case than three-phase. In fact, no BESS is
required for IM1 to ride through disturbance as long as the single-phase fault is cleared
within 9 cycles. Even with HIME = 0.75 s and a prolonged fault clearing time of 30
cycles, the microgrid only needs S B = 60% to recover.
Similarly, though not reported in this chapter, comparing with a three-phase fault, it
will require less BESS capacity to recover motors under line-to-line and double line-to-
ground faults with an assumption of the same fault clearing/islanding time.

73
Figure 5.21: Minimum normalized BESS Size S B as a function of fault clearing time
under single-phase fault

5.5.5 Transient Study#C: Multiple IMs recovery

The final Transient Analysis is conducted under assumptions that both IM1 and IM2 in
Fig. 5.1 are operated with 90% of their rated power under a three-phase fault and a
clearing time of 3 cycles. The purpose of this case study is two-fold: first, it tests the sizing
guidelines demonstrated by Sensitivity Study 1. Secondly, it verifies the effectiveness
of the BESS adaptive voltage regulation strategy on enhancing multiple microgrid IMs’
speed recovery during fault-induced islanding incidents.
0 0
The HIME is calculated as 0.72 s based on (5.2) with JIM1 = JIM 1
, JIM2 = JIM 2
.
The IM loading accounts for 64% of the total 2.88 M W microgrid demand. From Fig.
103%
5.20(d), S B = 100%
under 3 cycles fault clearing with HIME = 0.72 s. The BESS

74
 Figure 5.22: Speed of the two induction motors versus time with 3.1 MVA BESS

converter size is calculated by:

SB = S B Sm KIM
103% 64%
= · 2.88 · (5.5)
100% 61%
≈ 3.1 M V A

where
Sm = 2.88 is the total microgrid demand in M W .
and
64%
KIM = 61%
is the IM load linear interpolation factor.

Figs. 5.22 to 5.25 summarize the system performance and responses of the chosen
3.1 M V A BESS. With the adaptive voltage regulation, both IM1 and IM2 ’s speed are

75
 Figure 5.23: V̂ P CC with and without 3.1 MVA BESS

recovered as shown in Fig. 5.22. The PCC voltage profile is presented in Fig. 5.23.
Similar to the Transient Analysis 1, the Vref (t) ramps up from VrefN to VrefE after sensing
the low voltage condition, V̂ P CC < VthL = 0.5 pu. Due to the capacity limitation of the
BESS converter, the V̂ P CC cannot immediately recover beyond (VrefN − Vdb ) = 0.95 pu
and thus the Vref (t) ramps down after waiting for a duration of trs1 = 0.1 s. Nevertheless,
comparing the V̂ P CC with and without the BESS, one can see that the BESS still plays
a critical role on both the voltage recovery and the IM speed recovery. And the V̂ P CC
smoothly increases to VrefN as the IMs speed is recovering. The BESS initially utilizes
all its inverter capacity for voltage support by injecting maximum reactive power, Fig
5.24(b). After the recovery of both voltage and IM speed, the frequency regulator controls
the real power, Fig 5.24(a), which helps the diesel generator frequency return to normal
as shown in Fig. 5.25.

76
Figure 5.24: BESS power versus time under two induction motors loading: (a). Real
power, (b). Reactive power.

77
Figure 5.25: Generator frequency fr versus time with two induction motors loading and
a 3.1 MVA BESS

5.6 Chapter5 Summary

Following the previous chapter 4’s work [127], this report highlights the additional bene-
fits of adaptively controlling the BESS converter to enhance microgrid’s induction motor
speed recovery during fault-triggered islanding incidents by means of adaptive voltage
regulation. It proposes a simple yet effective way of achieving both faster voltage recov-
ery and longer maximum permissible fault clearing/islanding time for a microgrid with
one or more IMs. The study also facilitates the microgrid design process by presenting
the normalized minimum BESS inverter sizing under a wide range of microgrid motor
inertia, loading and fault clearing times.
This work is mainly focused on technical considerations of the adaptive voltage reg-
ulation strategy. To determine the economic benefit, it is necessary to combine the other
possible grid-support services that can be provided by the BESS [19].

78
Chapter 6

Adaptive Control Evaluations on

CIGRE Benchmark System

This chapter presents two microgrid case studies under CIGRE MV distribution network
benchmark system [14]. The pu quantities are with the base values of 12.47 kV, 10 MVA
and 60 Hz for all cases. The objective is to evaluate the proposed adaptive energy storage
system frequency and voltage regulation control methodologies introduced in chapters 4
and 5 with a realistic network topology.
The first study covers four cases with multiple BESSs under normal voltage regula-
tion and adaptive frequency regulation strategies. It shows the effectiveness of BESSs and
advantages of adaptive secondary control on the stability enhancement of a meshed micro-
grid with fault-induced feeder switching. This microgrid has unbalanced loads, multiple
IMs, renewable generators with converter interfaces and distributed synchronous genera-
tors with different prime resources (diesel fuel or hydropower). Both symmetrical and
asymmetrical faults are considered. It demonstrates the robust performance of the pro-
posed converter controller under multiple contingencies in addition to the random power
output from renewable energy sources. The second case study further illustrates the ben-

79
efits of adaptive voltage regulation as well as the impacts of its voltage regulation delays
on the microgrid dynamics.

6.1 Microgrid System for Chapter 6

This section describes two modified CIGRE MV distribution network benchmark systems
as the final microgrid systems under investigation. It is assumed that a fault happened at
the high voltage (HV) subtransmission line leads to the entire MV distribution network
as a fault-induced islanded microgrid. The first network contains three alternators and
BESSs. The second network uses the same network topology however contains only
alternators. Both microgrid systems have the similar amount of local generating capacities
(in a range between 30 to 40 MVA, including synchronous generators, BESSs and
RESs), enough to meet the total local demands (total of about 20 MVA). Appendix A.2
summarizes the detailed network and loading data.
Fig. 6.1 shows the topology of the studied microgrid network. It shows two transform-
ers that feed two 12.47 kV feeders via a 115 kV subtransmission system (X/R = 10;
5000-MVA short-circuit capacity). It is assumed that both feeders are fed by the same
substation with all five switches (S1 , S2 , S3 , S4 and S5 ) closed. Effectively, two feeders
will create a meshed microgrid network if it is being islanded from the substation bus
#0. It should be emphasized that there are more feeders served by the two transformers
that are not modeled in details in this topology [14]. This explains why there are much
larger sizes of DERs and load values at the buses #1 and #12 at the beginning of the two
feeders than those given for the other nodes.
The loadings of the system are unbalanced and distributed under a mixture of both
linear RLC impedance and IM loads. Four 2.3 kV IMs of various sizes (two 500 HP, two
2250 HP) are connected to their 12.47 kV hosted feeders at different locations through

80
Figure 6.1: GIGRE Test Network

81
associated transformers. Each IM operates at 90% of its rated power. The original RLC
impedance loading data [14] is modified at nodes with IMs by subtracting associated IM
loading demand in order to maintain the similar amount of total local load demands.
Both synchronous generators and power inverter based DERs are added at the various
nodes as shown in the schematic. One 5 M V A hydroelectric generator (Rs1 = 5% droop)
coupling with hydroelectric turbine is installed at the beginning of the feeder #1, bus
#1. There are also two diesel generators; a smaller generator with 1.8 M V A capacity
(Rs2 = 3% droop) is at bus #7 and a larger one with size of 3.125 M V A (Rs3 = 1%
droop) is at bus #12. Three BESSs are installed separately at bus #1, #12 and #14. Two
renewable generators, with a random power output via their power converters under the
capacity of 3.5 M V A and 0.5 M V A, are located at bus #12 and #14 respectively.
The second network benchmark has the same topology as in the previous Fig. 6.1,
however all three BESSs are being replaced by the diesel generators with an almost alike
total output power capacity. Consequently, it is being used in the comparative case study
(study #1, case #B) to show the effects of BESSs on the microgrid dynamics.

6.2 Study#1: Normal Voltage Regulation

Four cases are presented in this study. The BESSs are operated under normal voltage
regulation (i.e., Vref = 1.0 pu) with power priority switch control, described in Ch. 5.4.2,
Fig. 5.7, implemented. Case #A demonstrates the additional stability margin gained by
the adaptive secondary control. Case #B compares two microgrids dynamics with and
without BESSs. Case #C illustrates a Single Line-to-Ground fault in order to show the
performance of microgrid and BESSs under asymmetrical fault. Finally, case #D further
tests the effectiveness of the proposed adaptive control strategies by applying a second
fault inside an islanded microgrid along with random power supplied by the renewable

82
generators.

6.2.1 Case#A: Impact of BESSs secondary control approaches

This case illustrates the effects of BESS converter secondary frequency control approaches
on the microgrid stability. It uses the topology shown in Fig. 6.1 as the studied micro-
grid system. Three BESSs (each 6 M V A) are all equipped with adaptive nonlinear droop
control with Rc = 3%. A secondary controller with KI = 3000 is assumed in the first
scenario for a fixed gain control approach. For the second approach with adaptive sec-
ondary control gain, it sets the KB = 3 × 105 which corresponds to KI (|∆f |) = 3000 at
|∆f | = 1% (0.6 Hz). In order to fully appreciate the advantages of adaptive secondary
control, it further assumes that KImax = ∞. The only constrain for KI (|∆f |) during
the transients in this case is a hard limitation (15 Hz) for |∆f | via converter PLL con-
troller settings, Appendix A.6. The BESS converter frequency regulator has a frequency
dead-band of fdb = 6 × 10−4 pu (36 mHz).
Prior to the fault, the system works in steady state with all the loads in service. All
three BESSs are in the standby mode, connected to their PCCs and synchronized to the
utility grid through associated PLL. The hydroelectric generator (DG#1) generates 3 MW
(60% of its nameplate capacity) to the grid. Similarly, two diesel generators (DG#2:
1.8 M V A and #3: 3.125 M V A) are delivering 1.26 MW (70% capacity) and 2.5 MW
(80% capacity) respectively. The output of the renewable generators is assumed to be
zero during the 2 seconds’ simulation.
At t = 0.067 s (4 cycles after the simulation starts), a permanent three-phase fault
strikes on the upper level subtransmission line close to the substation (bus #0). This fault
is cleared in 4 cycles. As the result, a meshed microrgrid is formed at t = 0.133 s. Three
BESSs detect the fault disturbance at different time and start to participate the frequency
and voltage regulations at t = 0.100; 0.117; 0.133 s respectively, corresponding to the

83
fault detection time of 2, 3 and 4 cycles. Figures. 6.2 to 6.5 compare three BESSs
responses and the microgrid dynamics during this fault-induced islanding event under a
secondary controller with either fixed or adaptive gain.
Figure. 6.2 presents three BESSs’ estimated PLL frequencies in Hz. It is evident
from the left column that the fault disturbance leads to a significant fluctuation for all
three PLL’s frequency estimations. They further evolve into unstable oscillations around
base frequency (60Hz) under the secondary controller with a fixed gain. On the contrary,
the right column shows oscillograms that, with an adaptive secondary controller, all three
frequency estimations have acceptable transients and quickly return to nominal frequency
by the help of adaptive frequency regulation.
Figure. 6.3 summarizes three BESSs’ real power responses. The BESSs controllers
with a fixed secondary control gain leads to inadmissible power fluctuations. With an
adaptive secondary controller, all three BESSs’ power responses are stable. Moreover,
it shows that each BESS quickly shares power under discharge mode and this action
contributes to a fast microgrid frequency regulation. All three BESSs’ power responses
reach to a new equilibrium in less than one second after the start of the regulation.
Figure. 6.4 shows three BESSs’ terminal (PCC) voltage magnitudes, V̂ P CC in time.
The fault causes severe sags of voltage at all nodes inside the microgrid. The unstable
controllers result in poor voltage supply quality as shown on the left. The right column
indicates a rapid voltage recovery with stable controllers under the adaptive secondary
control strategy.
Finally, three alternators’ rotor speeds are summarized in Fig. 6.5. The unstable sec-
ondary frequency control causes unavoidable frequency deviations for all three generators
as shown on the left. With the adaptive secondary controllers, the BESSs quickly regulate
the microgrid frequency and all three generators’ rotor speeds are back to 60 Hz in less
than a second (right column).

84
Figure 6.2: Estimated PLL Frequency fs ; (left column): unstable under secondary con-
troller with KI = 3000, (right column): stable under secondary controller with adaptive
gain using KB = 3 × 105 : (a) fs1 , unstable; (b) fs2 , unstable; (c) fs3 , unstable; (d) fs1 ,
stable; (e) fs2 , stable; (f) fs3 , stable

85
Figure 6.3: The BESSs power vs. time; (left column): unstable under secondary con-
troller with KI = 3000, (right column): stable under secondary controller with adaptive
gain using KB = 3 × 105 : (a) PB1 , unstable; (b) PB2 , unstable; (c) PB3 , unstable; (d)PB1 ,
stable; (e) PB2 , stable; (f) PB3 , stable

86
Figure 6.4: The PCCs voltage magnitude V̂ P CC vs. time; (left column): unstable under
secondary controller with KI = 3000, (right column): stable under secondary controller
with adaptive gain using KB = 3 × 105 : (a) V̂ P CC1 , unstable; (b) V̂ P CC2 , unstable; (c)
V̂ P CC3 , unstable; (d) V̂ P CC1 , stable; (e) V̂ P CC2 , stable; (f) V̂ P CC3 , stable

87
Figure 6.5: The synchronous generators rotor frequency fr (Hz) vs. time; (left col-
umn): unstable under secondary controller with KI = 3000, (right column): stable
under secondary controller with adaptive gain using KB = 3 × 105 : (a) fr1 : 5 M V A
hydroelectric generator, unstable; (b) fr2 : 1.8 M V A diesel generator, unstable; (c) fr3 :
3.125 M V A diesel generator, unstable; (d) fr1 : 5 M V A hydroelectric generator, stable;
(e) fr2 : 1.8 M V A diesel generator, stable; (f) fr3 : 3.125 M V A diesel generator, stable

88
6.2.2 Case#B: Microgrid with and without BESSs - A comparison

study

Case #B performs two simulations based on the case #A topology, Fig.6.1. The renewable
generators are assumed to have a zero output power during both simulations and hence
their impacts can be omitted in this case study. The first simulated microgrid has both syn-
chronous generators and BESSs. The second simulation is done with only synchronous
generators. It can be thought as if each 6 M V A BESS in Fig.6.1 is being replaced by two
diesel generators (each generator size is 3.125 M V A) with droop setting Rs = 1% to en-
sure a similar generating capacity in both microgrids. Meanwhile, they are in the standby
mode during the beginning of the simulation (zero output power) similar to three BESSs
in case #A. The microgrid is formed under the same contingency as in the case#A. Since
the fundamental difference between two simulation scenarios is their microgrid genera-
tor types, comparing two simulation results will provide good insights for the effects of
BESSs on the microgrid performance.
Fig. 6.6 summarizes the rotor speeds for Fig. 6.1’s three synchronous generators,
under the situation either with or without BESSs. With the BESSs help, all three gener-
ators’ rotors are returned to their nominal value in one second. Without the BESSs, all
three generators experience a frequency deviation that is close to 2 Hz, despite the higher
system inertia contributed by the additional diesel generators.
Looking at the IMs speed, Fig. 6.7, one will notice that all IMs can recover their
speed in both simulations. Nonetheless, the scenario with BESSs has a faster IMs speed
recovery thanks to a faster PCCs voltage regulation provided by the BESS’s converters.
One interesting observation is that the IMs’ speed dips are less in the case with only
synchronous generators likely because of the higher system inertia. This indicates the
possible trade-off between the system rotational inertia and BESS’s fast power response.

89
Figure 6.6: The synchronous generators rotor frequency fr (Hz) vs. time; with and with-
out BESSs: (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A diesel generator;
(c) fr3 : 3.125 M V A diesel generator

90
Figure 6.7: Induction motors speed ωIM vs. time, with and without BESSs: (a) ωIM1 :
2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP

Figure 6.8 shows that the PCCs voltage will have a higher value during the fault in
the case without BESSs due to the higher short-circuit MVA contributed by the additional
synchronous generators. However, the voltage takes significant longer time to recover
after the fault clear because the synchronous generators’ excitation system response speed
is much slower than the BESS’s voltage regulating action [122]. Comparing with the rapid
voltage recovery in the case with BESSs, it clearly demonstrates the effectiveness of the
BESSs control on the microgrid voltage quality and stability enhancement.

91
Figure 6.8: The PCCs voltage magnitude V̂ P CC vs. time; with and without BESSs: (a)
V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3

92
6.2.3 Case#C: Impact of asymmetrical fault

The proposed control strategies are further tested under an asymmetrical fault of Single
Line-to-Ground type. All three BESSs are operated under adaptive frequency control
with nonlinear droop Rc = 3% and secondary adaptive gain with KB = 1 × 105 which
corresponds to KI (|∆f |) = 1000 at |∆f | = 1% (0.6 Hz). The fault location and starting
moment are the same as in the previous case #A and #B, except the clearing time in this
case is 12 cycles (8 cycles longer). Fig. 6.9 to 6.12 summarize the BESSs and microgrid
dynamics.
Three BESSs’ estimated PLL frequencies are plotted in Fig. 6.9. It shows that all
three frequency estimations are stable and quickly return to nominal value with the help
of BESSs after the fault disturbance. Three BESSs’ power fluctuates around zero during
the fault, Fig. 6.10. Nevertheless, all three BESSs rapidly share the power mismatching
in the islanded microgrid and this quickly help all generators’ rotor speeds recover to their
nominal value as shown in Fig. 6.11. Similarly, Fig. 6.12 illustrates that the microgrid
voltages at different nodes can immediately recover within Vdb = 5% of its nominal value
after the islanding event under the proposed adaptive control strategies.

93
Figure 6.9: Estimated PLL Frequency fs under 12 cycles asymmetrical fault: (a) fs1 ;
(b) fs2 ; (c) fs3

94
Figure 6.10: The BESSs power vs. time under 12 cycles asymmetrical fault: (a) PB1 ;
(b) PB2 ; (c) PB3

95
Figure 6.11: The synchronous generators rotor frequency fr (Hz) vs. time under 12 cycles
asymmetrical fault: (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A diesel
generator; (c) fr3 : 3.125 M V A diesel generator

96
Figure 6.12: The PCCs voltage magnitude V̂ P CC vs. time under 12 cycles asymmetrical
fault: (a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3

97
6.2.4 Case#D: Multiple faults and renewable generation

The final case performs a 10 s simulation with the system configuration shown in Fig. 6.1.
The BESSs controllers have the same settings as in the case #C. The system experiences
two faults. The first contingency is the same as in case #A and #B, which happens 4
cycles after the simulation begins. This leads to the fault-induced islanding 4 cycles later,
at t = 0.133 s. The second three-phase fault happens at t = 3.567 s inside the islanded
microgrid close to bus #6. It is cleared by opening switches S3 and S4 at t = 3.617 s, 3
cycles after the fault. As a result of this feeder switching induced by the second fault, the
microgrid’s topology changes. In addition, both renewable generators have their random
power output with fast ramp rates as shown in Fig. 6.13. It is evident that the second
fault inside the microgrid causes a glitch at the first renewable generator’s power output.
The responses of the BESSs and the system performance are summarized in Figs. 6.14 to
6.21.
Figure 6.14 plots three BESSs’ estimated PLL frequencies. It shows multiple fre-
quency transients caused by several disturbances. The initial fault leads to the largest
frequency excursion in all three PLL estimations. They all quickly return to normal with
the help of three BESSs after the islanding. The second fault inside the microgrid causes
feeders to switch and results a significant frequency oscillation. Fortunately, the oscilla-
tion damps out with the help of both BESSs and power system stabilizer (PSS) equipped
at the hydroelectric generator’s excitation system. With the BESSs adaptive frequency
regulation, the frequency fluctuations are negligible under unpredictable power genera-
tion from renewable resources. All of these observations clearly show the robustness of
the proposed control strategies inside a microgrid with dynamic changing environments.
The BESSs output current in time during the above events are captured in Fig. 6.15,
which includes d/q components and the resulted total current magnitude. It shows the
autonomous frequency and voltage regulation actions of the three BESSs within their

98
current capacities. The associated real and reactive power responses are plotted in Figs.
6.16 and 6.17. All three figures clearly show the impacts of multiple disturbances on the
responses of all BESSs. Similar to the previous cases, all three BESSs share the real power
for a collaborative frequency regulation action. However, due to the different BESSs
locations, the reactive power is not shared autonomously. Nevertheless, the microgrid
stability is enhanced significantly with all four IM’s speed successfully recovered under
BESSs’ contributions, Fig. 6.18.
Three generators’ rotor frequencies are plotted in Fig. 6.19. Two faults cause sig-
nificant rotor speed disturbances. The second fault has a great impact on the islanded
microgrid’s frequency. The hydroelectric generator rotor oscillation is damped with the
help of the PSS. Under the adaptive frequency regulation of the BESSs, all of generators’
rotors are returning to the nominal 60 Hz shortly after the faults are cleared. Moreover,
all generators’ speed varies less than 0.2 Hz even under the rapid random power output
from the renewable resources (e.g., PRE1 has a step change of 3.5 M W at t = 5 s, it also
ramps up at a rate of 4 M W/s, 7.0 < t < 7.5 s).
Three PCCs voltage magnitudes are presented in Fig. 6.20. Two faults cause signifi-
cant voltage sags. Nevertheless, all of them quickly recover to the nominal value with the
help from BESSs’ voltage regulator. One can also clearly see the voltage magnitudes os-
cillations resulted from the generators’ excitation systems responses and rotor dynamics,
3.65 < t < 5.00 s.
Finally, Fig. 6.21 depicts the magnified PCCs voltage oscillograms before, during and
after two faults, 0.05 < t < 0.45 s and 3.5 < t < 3.9 s. Thanks to the BESSs voltage
regulation action, it demonstrates the rapid voltage recovery with a high waveform supply
quality under both faults at all three PCCs locations inside the microgrid. It illustrates
that the microgrid resiliency is significantly enhanced with the help of proposed adaptive
energy storage control methods even under multiple large disturbances.

99
Figure 6.13: Renewable Energy Power Output PRE under multiple faults (4 and 3 cycles
fault clearing time respectively): (a) PRE1 ; (b) PRE2

100
Figure 6.14: Estimated PLL Frequency fs under multiple faults and renewable generation
ramping events with normal voltage regulation: (a) fs1 ; (b) fs2 ; (c) fs3

101
Figure 6.15: BESSs current magnitudes transients under multiple faults and renewable
generation ramping events with normal voltage regulation: (a) BESS1 ; (b) BESS2 ; (c)
BESS3

102
Figure 6.16: BESSs real power vs. time under multiple faults and renewable generation
ramping events with normal voltage regulation: (a) PB1 ; (b) PB2 ; (c) PB3

103
Figure 6.17: BESSs reactive power vs. time under multiple faults and renewable genera-
tion ramping events with normal voltage regulation: (a) QB1 ; (b) QB2 ; (c) QB3

104
Figure 6.18: Induction motors speed ωIM vs. time, under multiple faults and renewable
generation ramping events with normal voltage regulation: (a) ωIM1 : 2250 HP; (b) ωIM2 :
500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP

105
Figure 6.19: The synchronous generators rotor frequency fr (Hz) vs. time under multiple
faults and renewable generation ramping events with normal voltage regulation: (a) fr1 :
5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A diesel generator; (c) fr3 : 3.125 M V A
diesel generator
106
Figure 6.20: The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and
renewable generation ramping events with normal voltage regulation: (a) V̂ P CC1 ; (b)
V̂ P CC2 ; (c) V̂ P CC3

107
Figure 6.21: Magnified PCC voltage oscillogram vs. time under multiple faults and
renewable generation ramping events with normal voltage regulation: (a) VP CC1 , first
fault-induced islanding transient, 0.05 < t < 0.45 s; (b) VP CC2 , first fault-induced is-
landing transient, 0.05 < t < 0.45 s; (c) VP CC3 , first fault-induced islanding transient,
0.05 < t < 0.45 s; (d) VP CC1 , second (islanding) fault-induced feeder switching transient,
3.5 < t < 3.9 s; (e) VP CC2 , second (islanding) fault-induced feeder switching transient,
3.5 < t < 3.9 s; (f) VP CC3 , second (islanding) fault-induced feeder switching transient,
3.5 < t < 3.9 s

108
6.3 Study#2: Adaptive Voltage Regulation

Both adaptive frequency and voltage regulations are implemented in this study. Case #A
shows the advantages of the adaptive voltage regulation on the microgrid recovery when
a prolonged fault clearing is applied. Case #B indicates that it is possible to achieve
better microgrid dynamics by including appropriate voltage regulation delays inside the
adaptive voltage regulation algorithm.
Fig. 6.1 shows the studied system topology. It also experiences similar transients as
in study #1, case #D with the assumption that both faults happen at the same moment
(t = 0.067 s and t = 3.567 s). However, the clearing time of fault#1 takes 7 cycles (3
cycles longer, at t = 0.183 s) and fault#2’s clearing time is 4 cycles (1 cycle longer, at
t = 3.633 s). Consequently, the BESS #1 and #3 sizes increase to 10 MVA to ensure the
stable microgrid operation.

6.3.1 Case#A: Without voltage regulation control delay

In this first case, three BESSs have the same adaptive frequency control settings as in
study #1, case #C and #D. The voltage controller Vref (t) is implemented based on Fig.
5.6. The power priority switch control introduced in Fig. 5.7, Ch. 5.4.2, is also in
action. The adaptive voltage controller has the following settings: VrefN = 1.00 pu,
VrefE = 1.30 pu, Vdb = 0.05 pu, VthL = 0.50 pu, VthH = VrefE − Vdb = 1.25 pu,
KR = 7 V /s, trs1 = 0.100 s (6 cycles), trs2 = 0.067 s (4 cycles), trs3 = 1.000 s,
fsw = 0.01 pu (0.6 Hz), tsw1 = 0.083 s (5 cycles) and Isat = 0.90 Imax . Figs. 6.23 to
6.26 picture the BESSs adaptive control contributions on the overall microgrid stability
enhancement.
Two renewable generators’ random power outputs are plotted in Fig. 6.22 similar to
the previous study case #D. The second fault causes a short disturbance of power gener-

109
ation at 3.567 < t < 3.633 s. Both the multiple faults and the randomness of renewable
power test the BESS’s controller performance. Three alternators’ rotor speeds are evalu-
ated in Fig. 6.23. Similar to the previous study #1, case #D, all synchronous generators’
rotor speeds experience large excursions during the faults. One can see that the longer
fault durations lead to a greater deviation of rotor speeds. Nevertheless, all three syn-
chronous generators’ frequency return to 0.5 Hz within their nominal value in less than
one second after the fault clearing under the BESSs’ contributions. Even though the sys-
tem experiences the renewable output power with a 2 M W ramping up in half second
(2.0 < t < 2.5 s), their impacts on the microgrid system frequency is inconsequential,
Fig. 6.23.
The BESSs’ adaptive voltage regulations contribute to a fast PCCs voltage recovery as
shown in Fig. 6.24. The fault-induced low voltage conditions activate the adaptive voltage
control algorithm. Therefore all three voltage reference signals are ramping up during
both faults. This results a higher PCC recovery voltage magnitudes after faults clearing.
One will notice that the voltage references ramp down even before the voltage magnitudes
reach VthH = 1.25 pu during the fault#1 transient (this is not the case in the fault#2).
The reasons can be explained as following: the prolonged low voltage condition seen
by BESS#1 results an early ramping down of Vref1 (t) (condition 1: insufficient capacity
for over-excitation). However, this is not the case for both BESS#2 and #3 and hence
their voltage references are still staying at VrefE . The different voltage references among
BESSs eventually lead both the Vref2 (t) and Vref3 (t) back to VrefN . This action results in
a better voltage control coordination (condition 2).
Figure 6.25 depicts the detailed oscillograms for PCCs voltage waveforms during two
faults transient, 0.05 < t < 0.45 s and 3.5 < t < 3.9 s. It clearly indicates higher
recovery voltage magnitudes after the fault clearing showing the effectiveness of adaptive
voltage regulation. The waveforms remain sinusoidal despite a short disturbance.

110
 Thanks to the adaptive voltage regulation, all four IMs speeds can ride-through the
prolonged faults disturbances during islanding event, Fig. 6.26. Meanwhile, the effects
of renewable output power on motor speed variations are inconsequential. Conversely,
two 2250 HP motors (IM #1 and #3) under normal voltage regulation cannot recover their
speed.
The BESSs responses are summarized in Figs. 6.27 to 6.30. All three BESSs’ con-
verters PLL remain stable under several consecutive large disturbances, Fig. 6.27. These
frequency transient deviations activate the fast adaptive frequency control, which quickly
brings the system frequency back to normal. Since all BESSs are equipped with both
adaptive nonlinear droop and secondary control, the frequency regulation actions are very
small near the nominal value but significant during the large disturbances. This is favor-
able for a stable operation of microgrid under a dynamic changing environment. Conse-
quently, the renewable generation causes very little frequency deviations, well within the
acceptable range.
One can see the three BESSs rapidly change their output currents during multiple
transients as shown in Fig. 6.28. All of them are reaching steady state within a second
after each disturbance. During the two faults-on periods, all three BESSs’ currents are
saturated, indicating that they all contribute their best efforts to enhance microgrid stabil-
ity. During normal operations, each BESS has an adequate capacity for both voltage and
frequency regulations. The associate real and reactive powers are presented in Fig. 6.29
and 6.30. Both figures demonstrate the fast BESSs’ regulation actions. In addition, their
responses closely follow the corresponding reference signals.

111
Figure 6.22: Renewable Energy Power Output PRE under multiple (two) faults (7 and 4
cycles fault clearing time respectively): (a) PRE1 ; (b) PRE2

112
Figure 6.23: The synchronous generators rotor frequency fr (Hz) vs. time under multi-
ple faults and renewable generation ramping events with adaptive voltage regulation (no
regulation delay): (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A diesel
generator; (c) fr3 : 3.125 M V A diesel generator
113
Figure 6.24: The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (no regulation delay):
(a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3

114
Figure 6.25: Magnified PCC voltage oscillogram vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (no regulation de-
lay): (a) VP CC1 , first fault-induced islanding transient, 0.05 < t < 0.45 s; (b) VP CC2 ,
first fault-induced islanding transient, 0.05 < t < 0.45 s; (c) VP CC3 , first fault-induced
islanding transient, 0.05 < t < 0.45 s; (d) VP CC1 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s; (e) VP CC2 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s; (f) VP CC3 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s

115
Figure 6.26: Induction motors speed ωIM vs. time, under multiple faults and renewable
generation ramping events with adaptive voltage regulation (no regulation delay): (a)
ωIM1 : 2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP

116
Figure 6.27: Estimated PLL Frequency fs under multiple faults and renewable generation
ramping events with adaptive voltage regulation (no regulation delay): (a) fs1 ; (b) fs2 ; (c)
f s3

117
Figure 6.28: BESSs current magnitudes transients under multiple faults and renewable
generation ramping events with adaptive voltage regulation (no regulation delay): (a)
BESS1 ; (b) BESS2 ; (c) BESS3

118
Figure 6.29: BESSs real power vs. time under multiple faults and renewable generation
ramping events with adaptive voltage regulation (no regulation delay): (a) PB1 ; (b) PB2 ;
(c) PB3

119
Figure 6.30: BESSs reactive power vs. time under multiple faults and renewable genera-
tion ramping events with adaptive voltage regulation (no regulation delay): (a) QB1 ; (b)
QB2 ; (c) QB3

120
6.3.2 Case#B: With voltage regulation control delay

The purpose of this case study is to show that it is beneficial to introduce appropriate
control delays in the voltage regulation process. Such approach helps to avoid unnec-
essary voltage fluctuations and to achieve better microgrid frequency regulation. For
this purpose, the BESSs’ voltage regulators use the alternative decision tree as shown
in Ch. 5.4.2, Fig. 5.7. In order to determine the Vref (t), this alternative algorithm
introduces two voltage regulation delays. The adaptive voltage controller has the fol-
lowing settings: VrefN = 1.00 pu, VrefE = 1.25 pu, Vdb = 0.05 pu, VthL = 0.50 pu,
VthH = VrefE − Vdb = 1.20 pu, KR = 7 V /s, trs1 = 0.100 s (6 cycles), trs2 = 0.067 s (4
cycles), trs3 = 1.000 s, fsw = 0.01 pu (0.6 Hz), tsw1 = 0.083 s (5 cycles), trd1 = 0.100 s
(6 cycles), trd2 = 0.017 s (1 cycles) and Isat = 0.90 Imax . Figs. 6.31 to 6.34 highlight
the results for this case study.
The PCC voltage profiles under this alternative adaptive voltage regulation strategy
are shown in Fig. 6.31. The Vref (t) only ramps up after waiting for trd1 (6 cycles) under
the low voltage condition. Since the first fault lasts 7 cycles, it triggers the adaptive
voltage regulation actions. As the result, all three PCC voltages magnitudes have a higher
post-fault/islanding value as indicated. This is beneficial for all IMs speed recover as
shown in Fig. 6.32. Moreover, the Vref (t) does not immediately ramp down when V̂ P CC
hits VthH = 1.20 pu. Instead, it ramps down to VthN after waiting for trd2 (1 cycles). By
doing so all BESSs can have time to regulate their V̂ P CC to the VthE . Meanwhile, it can
achieve both higher probability for successful IMs speed recovery and a more uniformed
voltage profiles across the microgrid.

121
Figure 6.31: The PCCs voltage magnitude V̂ P CC vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (including regulation
delay): (a) V̂ P CC1 ; (b) V̂ P CC2 ; (c) V̂ P CC3

122
Figure 6.32: Induction motors speed ωIM vs. time, under multiple faults and renewable
generation ramping events with adaptive voltage regulation (including regulation delay):
(a) ωIM1 : 2250 HP; (b) ωIM2 : 500 HP; (c) ωIM3 : 2250 HP; (d) ωIM4 : 500 HP

123
 Since the true benefits of adaptive voltage regulation can only be shown under the
cases with a prolonged fault clearing, it is unnecessary to ramp the Vref (t) up during
a short fault disturbance. Therefore, the Vref (t) stays at VthN during the second fault
event because it only takes 4 cycles to clear the fault. Nevertheless, all PCC’s voltages
quickly recover to their nominal values which contributes to a successful IMs fault ride-
through. This setting also gives more converter capacity for a better microgrid frequency
regulation, which is evident in Fig. 6.33. Comparing with generator rotor speeds in
previous case #A, Fig. 6.23, 3 < t < 4 s, one can see that this new alternative achieves
1 Hz less frequency excursions during the second fault for all three alternators. This
additional accomplishment demonstrates the benefits of having appropriate control delays
inside the adaptive voltage regulation algorithm. Finally, Fig. 6.34 details the magnified
PCC voltage oscillograms during both faults. The adaptive voltage regulation contributes
to a higher voltage magnitude shortly after clearing the fault #1. This higher voltage
magnitude only lasts about 6 cycles and then quickly goes back to the nominal value. On
the contrary, the voltage magnitude directly recovers to nominal value under fault #2. The
quality of the voltage waveforms are excellent during both disturbances recovery.

124
Figure 6.33: The synchronous generators rotor frequency fr (Hz) vs. time under multiple
faults and renewable generation ramping events with adaptive voltage regulation (includ-
ing regulation delay): (a) fr1 : 5 M V A hydroelectric generator; (b) fr2 : 1.8 M V A diesel
generator; (c) fr3 : 3.125 M V A diesel generator
125
Figure 6.34: Magnified PCC voltage oscillogram vs. time under multiple faults and re-
newable generation ramping events with adaptive voltage regulation (including regulation
delay): (a) VP CC1 , first fault-induced islanding transient, 0.05 < t < 0.45 s; (b) VP CC2 ,
first fault-induced islanding transient, 0.05 < t < 0.45 s; (c) VP CC3 , first fault-induced
islanding transient, 0.05 < t < 0.45 s; (d) VP CC1 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s; (e) VP CC2 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s; (f) VP CC3 , second (islanding) fault-induced feeder
switching transient, 3.5 < t < 3.9 s

126
6.4 Chapter6 Summary

This chapter presents extensive microgrid case studies under CIGRE MV distribution
network with both synchronous generator and inverter based resources. They help to
evaluate the effectiveness and robustness of proposed adaptive energy storage control
methods under a meshed microgrid with topology change induced by feeder switching.
The microgrid loadings are mixed with RLC impedance and IM loads. It has been shown
that the microgrid stability and resiliency can be greatly enhanced under multiple distur-
bances with the help of the proposed BESSs’ autonomous control actions. The transient
events under investigation include fault-induced islanding (with both symmetrical and
asymmetrical fault types), isolated microgrid internal fault, as well as, random power
generating from renewable resources. The obtained results also demonstrate the advan-
tages of both adaptive secondary frequency control and appropriate voltage regulation
delays inside the adaptive voltage regulation algorithm.

127
Chapter 7

Conclusions

This dissertation research highlights the efficacy of flexibly utilizing the battery energy
storage systems (BESSs) to enhance the stability of microgrids. This study introduces a
set of autonomous adaptive control strategies for BESS converters to achieve fast micro-
grid frequency and voltage regulations. The following two sections summarize the main
discoveries, contributions and suggest the future studies.

7.1 Summary

This work discovers the following aspects of adaptive energy storage system control:

• By adaptively changing the frequency controller gains based on the estimated Phase-
Locked Loop (PLL) frequency deviation, the resulted nonlinear droop and sec-
ondary control can work together to achieve an enhanced microgrid frequency reg-
ulation with additional stability margins.

• By adaptively setting the voltage regulator reference point after sensing the low
voltage condition, it enables the inductor motors to achieve a higher electrome-
chanical torque for a successful fault-ride through.

128
• Although it is viable to help a heavily loaded induction motor ride-through a pro-
longed fault clearing time under the proposed adaptive voltage regulation with a
large storage capacity, its economical feasibility needs further justifications.

• The real/reactive power/current allocations within the converter capacity limitation

have significant impacts on the microgrid dynamic performances. It may be neces-
sary to switch the converter power priority settings during real-time operations to
meet the microgrid systems requirements, e.g., switching the priority from reactive
power to real power when the microgrid experiences inadmissible frequency.

• The adaptive voltage regulation may not be necessary when the detected voltage sag
durations are short. Therefore, some appropriate voltage regulation delays inside
the adaptive voltage regulation settings can indeed be favorable.

• The proposed adaptive energy storage system control strategies are expandable un-
der different microgrid configurations. They are also very robust under an unpre-
dictable dynamic changing microgrid environment.

The dissertation has accomplished the following major contributions:

1. The study proposes a set of improved robust autonomous adaptive energy storage
control strategies for both microgrid frequency and voltage regulation.

2. The study thoroughly investigates the proposed adaptive energy storage control
strategies effects on microgrid stability and dynamic performances via extensive
simulation case studies. The results verify the control efficacy, facilitate the micro-
grid design process and inspire the further methodology improvements.

129
7.2 Future Work

The following areas, in continuations of this study, can be worth exploring:

• This research has demonstrated the additional stability margins gained by adaptive
nonlinear droop and secondary frequency control. The following technical con-
cerns regarding the adaptive control demand additional painstaking investigations:
(i): real-time adjustments for Rc based on locally available information for a better
regulation performance, (ii): possibilities of other nonlinear control laws based on
the PLL frequency measurement, e.g., the controller gains are resulted from non-
linear function of measured frequency deviation, and (iii): development of stability
criterion for the proposed nonlinear frequency control law.

• This research has verified the effectiveness of proposed adaptive control for a short-
term microgrid stability enhancement, it can expand to account the effects of a
microrgrid controller with communication capabilities in order to study the coordi-
nated control strategies among distributed generators and associated energy/power
managements/optimizations over longer period of time.

• A future study may plan to tackle the optimization of the adaptive voltage regula-
tion settings under different system standards, protection schemes and motor shaft
load/torque-speed characteristics.

• The additional studies may take other un-modeled dynamics into consideration,
such as the dynamics of converters DC sources or other types of nonlinear loads.

• The possible reconfigurations between multiple microgrids via multi-stages of is-

landing events or resynchronizations can be used to further examine the proposed
control methodologies and study the aggregated impacts of microgrids on higher
level grid dynamics.

130
Appendix A

System Parameters

A.1 Radial Feeder for Chapter 4 and 5

The studied microgrid in Ch. 4 and 5 are modeled using the following feeder data [114].
The Figs. 4.1 and 5.1 show the lengths of the feeder.

Positive-sequence impedance and susceptance:

Z1 = 0.1906 + j0.3896 Ω/km, B1 = 4.301 µS/km

Zero-sequence impedance and susceptance:

Z0 = 0.4806 + j1.2058 Ω/km, B0 = 2.003 µS/km

A.2 Microgrid data for Chapter 6 CIGRE Benchmark

System

A.2.1 Feeder

The feeder line parameter for CIGRE benchmark system is shown below:

131
Positive-sequence impedance and susceptance:

Z1 = 0.282 + j0.703 Ω/km, B1 = 3.193 µS/km

Zero-sequence impedance and susceptance:

Z0 = 0.466 + j1.243 Ω/km, B0 = 1.826 µS/km

A.2.2 Transformer

The information below details the two transformers’ parameters at CIGRE system HV-
MV substation.

Transformer connects node 0 and 1: Z = 0.104 + j1.24 Ω at 12.47 kV side;

Srated = 15 M V A.

Transformer connects node 0 and 12: Z = 0.13 + j1.55 Ω at 12.47 kV side;

Srated = 12 M V A.

A.2.3 Loads

The Fig. A.1 below provides a table that summarizes the CIGRE system loading condi-
tions implemented in Chapter 6 case studies.
The single-phase subnetwork has the loading of: 172.25 kVA with power factor of
0.94 inductive. In addition to the above table, there is a parasitic load (10% of their
individual rated power) at each synchronous generator and IM terminal, Fig.6.1 (Bus
#15: 0.5 MW; #16: 0.18 MW; #17: 0.3125 MW; #18: 0.168 MW; #19: 0.037 MW; #20:
0.168 MW; #21: 0.037 MW). The terminal capacitor bank values for IMs are: Bus #18:
0.6 Mvar; #19: 0.15 Mvar; #20: 0.6 Mvar; #21: 0.15 Mvar.

132
 Figure A.1: CIGRE system loading

A.3 Synchronous generators

The synchronous machines’ parameters can be found in the following table [1, 125]:

Generator number 1 2 3
Power rating (MVA) 5 1.8 3.125
Voltage rating(KV) 13.8 2.4 2.4
Xd (pu) 2.86 2.38 1.56
Xq (pu) 2.0 1.1 1.06
0
Xd (pu) 0.7 0.264 0.296
00
Xd (pu) 0.22 0.201 0.177
00
Xq (pu) 0.21 0.376 0.177
Xl (pu) 0.2 0.2 0.088
0
Td0 (pu) 3.4 2.47 3.7
00
Td0 (pu) 0.01 0.018 0.05
00
Tq0 (pu) 0.05 0.009 0.05
H (s) 2.9 1.035 1.0716

133
 Figure A.2: Diesel governor block diagram

A.4 Prime movers and governing systems

This appendix describes the two types of synchronous generator’s prime movers used in
this research work: diesel engine and hydraulic turbine with their governing system.
A Woodward diesel engine control system shown in Fig. A.2 is used in the study
[125, 137] and its block diagram is shown below:
The information below summarized the parameters used in this diesel governor model:
K = 40; T1 = 0.01 s; T2 = 0.02 s; T3 = 0.2 s; T4 = 0.25 s; T5 = 0.009 s;
T6 = 0.0384 s; TD = 0.024 s; TM IN = 0; TM AX = 1.1.
Similarly, the hydraulic turbine and governing system MATLAB block diagram is
shown in the following Fig. A.3 [119].
The following parameters are used in this study:

10
Servo-motor: Ka = 3
; Ta = 0.07 s, Position ranging (pu): [0.01, 0.97518], speed limits
range (pu/s): [-0.1, 0.1].

Droop setting and main P ID regulator: 5 % droop, P ID settings: KP = 1.163;

KI = 0.105; KD = 0; Time constant for derivative (s): 0.01.

Hydraulic turbine: beta=0; turbine gain=1.0361; tw = 2.67 s.

134
Figure A.3: MATLAB hydraulic turbine and governing system block diagram: (a). Over-
all model; (b). hydraulic turbine; (c). gate servomotor

135
A.5 Excitation systems

The excitation systems parameters used in this study according to Fig. 3.6 are summarized
as below:

DC excitation system for hydro generator:

KA = 300; τA = 0.001 s; KE = 1; τE = 0 s; VRmax = 11.5;

AC excitation system for diesel generator:

KA = 400; τA = 0.02 s; KE = 1; τE = 0.8 s; KD = 0.38; VRmax = 14.5; Exciter

saturation function pairs: [VE1 = 4.18, SE (VE1 ) = 0.1], [VE2 = 3.14, SE (VE2 ) = 0.03].

A.6 Inverter

The data for a 2.5 M V A converter used in Ch. 5 is presented below. The values can be
scaled according to [48] for different sizes of inverter.

Interface impedance to PCC (refer to the 13.8 kV side) [127]:

Z = 0.3809 + j3.8090 Ω

Current controller [49]:

Kp = 0.5305, Ki = 20

Phase-Locked Loop (PLL)

Figure A.4 shows the schematic diagram for the PLL and associated control block dia-
gram. It utilizes the feedback control to realize voltage phasor abc− to dq− transforma-
tion phase shift ρ such that ωt − ρ = 0 (Vq = 0). The integrator in the Fig. A.4(a) refers

136
 Figure A.4: PLL schematic and associated control block diagram

as voltage-controlled oscillator (VCO). It resets the ρ to zero whenever it reaches 2π. The
control block diagram is based on the following PLL control law:

dρ
= ω 0 (t) = H(p)V̂ (ωt − ρ)
dt (A.1)
0 0
ω (0) = ωs and ωmin ≤ ω (t) ≤ ωmax

The p = d(•)/dt is the differentiation operator, ωs is the nominal power system an-
gular frequency. The initial condition and range constrain of ω 0 (t) ensure a satisfactory
behavior of the PLL [49]. In this study, the ωmin = 345.6 rad/s and ωmax = 471.2 rad/s.
With per-unitization, the following values are chosen for the PLL compensator H(s)

137
based on the method proposed in [49]:

2.68 × 105 (s2 + 568, 516)(s2 + 166s + 6889)

H(s) = [(rad/s)/V ] (A.2)
s2 (s2 + 1508s + 568, 516)(s2 + 964s + 232, 324)

A.7 Induction Motors

The following table summarizes the induction motor loads information used in this thesis
[117].

Motor number 1 2
Power rating (HP) 2250 500
Voltage rating(KV) 2.3 2.3
Stator RS (Ω) 0.029 0.262
Stator/Rotor inductance LS = Lr 0 (µH) 600 3200
Rotor Rr 0 (Ω) 0.022 0.187
Mutual inductance Lm (mH) 34.6 143.3
0
Base inertia JIM (Kg.m2 ) 63.87 11.06
Damping Coefficient D (N.m.s) 1.607 0.402

138
Bibliography

[1] F. Katiraei, Dynamic analysis and control of distributed energy resources in a

micro-grid. Ph.D. dissertation, U. Toronto, Canada, 2005.

[2] H. You, V. Vittal, and X. Wang, “Slow coherency-based islanding,” IEEE Transac-
tions on Power Systems, vol. 19, pp. 483–491, Feb 2004.

[3] G. Andersson, P. Donalek, R. Farmer, N. Hatziargyriou, I. Kamwa, P. Kundur,

N. Martins, J. Paserba, P. Pourbeik, J. Sanchez-Gasca, R. Schulz, A. Stankovic,
C. Taylor, and V. Vittal, “Causes of the 2003 major grid blackouts in north america
and europe, and recommended means to improve system dynamic performance,”
IEEE Transactions on Power Systems, vol. 20, pp. 1922–1928, Nov 2005.

[4] R. H. Lasseter, “Smart distribution: Coupled microgrids,” Proceedings of the IEEE,

vol. 99, pp. 1074–1082, June 2011.

[5] “IEEE draft standard for interconnection and interoperability of distributed energy
resources with associated electric power systems interfaces,” IEEE P1547/D7.0,
September 2017, pp. 1–147, Jan 2017.

[6] J. A. P. Lopes, C. L. Moreira, and A. G. Madureira, “Defining control strategies

for microgrids islanded operation,” IEEE Transactions on Power Systems, vol. 21,
pp. 916–924, May 2006.

[7] A. H. K. Alaboudy, H. H. Zeineldin, and J. Kirtley, “Microgrid stability charac-

terization subsequent to fault-triggered islanding incidents,” IEEE Transactions on
Power Delivery, vol. 27, pp. 658–669, April 2012.

[8] A. A. Akhil etc, DOE/EPRI Electricity Storage Handbook in Collaboration with

NRECA. SAND2015-1002, Sandia National Laboratories.

[9] Y. A. R. I. Mohamed and E. F. El-Saadany, “Adaptive decentralized droop con-

troller to preserve power sharing stability of paralleled inverters in distributed gen-
eration microgrids,” IEEE Transactions on Power Electronics, vol. 23, pp. 2806–
2816, Nov 2008.

139
[10] J. C. Vasquez, J. M. Guerrero, A. Luna, P. Rodriguez, and R. Teodorescu,
“Adaptive droop control applied to voltage-source inverters operating in grid-
connected and islanded modes,” IEEE Transactions on Industrial Electronics,
vol. 56, pp. 4088–4096, Oct 2009.

[11] G. Diaz, C. Gonzalez-Moran, J. Gomez-Aleixandre, and A. Diez, “Scheduling of

droop coefficients for frequency and voltage regulation in isolated microgrids,”
IEEE Transactions on Power Systems, vol. 25, pp. 489–496, Feb 2010.

[12] J. Kim, J. M. Guerrero, P. Rodriguez, R. Teodorescu, and K. Nam, “Mode adap-

tive droop control with virtual output impedances for an inverter-based flexible ac
microgrid,” IEEE Transactions on Power Electronics, vol. 26, pp. 689–701, March
2011.

[13] Y. Sun, W. Huang, G. Wang, W. Wang, D. Wang, and L. Zhang, “Study of control
strategy of dg based on nonlinear droop characteristic,” in 2012 China Interna-
tional Conference on Electricity Distribution, pp. 1–4, Sept 2012.

[14] “Benchmark Systems for Network Integration of Renewable and Distributed En-
ergy Resources, CIGRE WG (TF): WG C6.04,” April 2014.

[15] P. Kundur, N. Balu, and M. Lauby, Power System Stability and Control. EPRI
power system engineering series, McGraw-Hill, 1994.

[16] G. Masters, Renewable and Efficient Electric Power Systems. Wiley - IEEE, Wiley,
2013.

[17] S. Blumsack, “How the free market rocked the grid,” IEEE Spectrum, vol. 47,
pp. 44–59, December 2010.

[18] E. Santacana, G. Rackliffe, L. Tang, and X. Feng, “Getting smart,” IEEE Power
and Energy Magazine, vol. 8, pp. 41–48, March 2010.

[19] T. Zhang, The Economic Benefits of Battery Energy Storage System in Electric
Distribution System. MS Thesis, Worcester Polytechnic Institute (WPI), USA,
2013.

[20] M. Smith and D. Ton, “Key connections: The u.s. department of energy?s micro-
grid initiative,” IEEE Power and Energy Magazine, vol. 11, pp. 22–27, July 2013.

[21] P. F. Ribeiro, B. K. Johnson, M. L. Crow, A. Arsoy, and Y. Liu, “Energy stor-

age systems for advanced power applications,” Proceedings of the IEEE, vol. 89,
pp. 1744–1756, Dec 2001.

[22] M. Farhadi and O. Mohammed, “Energy storage technologies for high-power ap-
plications,” IEEE Transactions on Industry Applications, vol. 52, pp. 1953–1961,
May 2016.

140
[23] K. D. Bachovchin, Design, Modeling, and Power Electronic Control for Transient
Stabilization of Power Grids Using Flywheel Energy Storage Systems. Ph.D. dis-
sertation, Carnegie Mellon University (CMU), PA, 2015.

[24] A. Kanchanaharuthai, V. Chankong, and K. A. Loparo, “Transient stability and

voltage regulation in multimachine power systems vis statcom and battery energy
storage,” IEEE Transactions on Power Systems, vol. 30, pp. 2404–2416, Sept 2015.

[25] A. Oudalov, D. Chartouni, and C. Ohler, “Optimizing a battery energy storage sys-
tem for primary frequency control,” IEEE Transactions on Power Systems, vol. 22,
pp. 1259–1266, Aug 2007.

[26] O. Leitermann, Energy storage for frequency regulation on the electric grid. Ph.D.
dissertation, Massachusetts Institute of Technology (MIT), MA, 2015.

[27] I. Serban, R. Teodorescu, and C. Marinescu, “Energy storage systems impact on the
short-term frequency stability of distributed autonomous microgrids, an analysis
using aggregate models,” IET Renewable Power Generation, vol. 7, pp. 531–539,
Sept 2013.

[28] I. Serban and C. Marinescu, “Control strategy of three-phase battery energy storage
systems for frequency support in microgrids and with uninterrupted supply of local
loads,” IEEE Transactions on Power Electronics, vol. 29, pp. 5010–5020, Sept
2014.

[29] E. Liegmann and R. Majumder, “An efficient method of multiple storage control in
microgrids,” IEEE Transactions on Power Systems, vol. 30, pp. 3437–3444, Nov
2015.

[30] A. S. Khalsa and S. Baktiono, “Certs microgrid test bed battery energy storage
system report: Phase 1,” tech. rep., 10/2016 2016.

[31] G. Delille, B. Francois, and G. Malarange, “Dynamic frequency control support by

energy storage to reduce the impact of wind and solar generation on isolated power
system’s inertia,” IEEE Transactions on Sustainable Energy, vol. 3, pp. 931–939,
Oct 2012.

[32] F. Blaabjerg, Z. Chen, and S. B. Kjaer, “Power electronics as efficient interface

in dispersed power generation systems,” IEEE Transactions on Power Electronics,
vol. 19, pp. 1184–1194, Sept 2004.

[33] J. M. Carrasco, L. G. Franquelo, J. T. Bialasiewicz, E. Galvan, R. C. PortilloGu-

isado, M. A. M. Prats, J. I. Leon, and N. Moreno-Alfonso, “Power-electronic sys-
tems for the grid integration of renewable energy sources: A survey,” IEEE Trans-
actions on Industrial Electronics, vol. 53, pp. 1002–1016, June 2006.

141
[34] D. Boroyevich, I. Cvetkovi, D. Dong, R. Burgos, F. Wang, and F. Lee, “Future elec-
tronic power distribution systems a contemplative view,” in 2010 12th International
Conference on Optimization of Electrical and Electronic Equipment, pp. 1369–
1380, May 2010.
[35] J. D. van Wyk and F. C. Lee, “On a future for power electronics,” IEEE Journal of
Emerging and Selected Topics in Power Electronics, vol. 1, pp. 59–72, June 2013.
[36] A. Tabesh and R. Iravani, “Multivariable dynamic model and robust control of a
voltage-source converter for power system applications,” IEEE Transactions on
Power Delivery, vol. 24, pp. 462–471, Jan 2009.
[37] C. Schauder and H. Mehta, “Vector analysis and control of advanced static var
compensators,” IEE Proceedings C - Generation, Transmission and Distribution,
vol. 140, pp. 299–306, July 1993.
[38] A. Timbus, M. Liserre, R. Teodorescu, and F. Blaabjerg, “Synchronization methods
for three phase distributed power generation systems - an overview and evaluation,”
in 2005 IEEE 36th Power Electronics Specialists Conference, pp. 2474–2481, June
2005.
[39] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. V. Timbus, “Overview of control
and grid synchronization for distributed power generation systems,” IEEE Trans-
actions on Industrial Electronics, vol. 53, pp. 1398–1409, Oct 2006.
[40] S.-K. Chung, “A phase tracking system for three phase utility interface inverters,”
IEEE Transactions on Power Electronics, vol. 15, pp. 431–438, May 2000.
[41] D. Jovcic, “Phase locked loop system for facts,” IEEE Transactions on Power Sys-
tems, vol. 18, pp. 1116–1124, Aug 2003.
[42] P. Rodrguez, R. Teodorescu, I. Candela, A. V. Timbus, M. Liserre, and F. Blaab-
jerg, “New positive-sequence voltage detector for grid synchronization of power
converters under faulty grid conditions,” in 2006 37th IEEE Power Electronics
Specialists Conference, pp. 1–7, June 2006.
[43] P. Rodriguez, A. Luna, I. Candela, R. Mujal, R. Teodorescu, and F. Blaabjerg,
“Multiresonant frequency-locked loop for grid synchronization of power convert-
ers under distorted grid conditions,” IEEE Transactions on Industrial Electronics,
vol. 58, pp. 127–138, Jan 2011.
[44] B. B. Johnson, S. V. Dhople, A. O. Hamadeh, and P. T. Krein, “Synchronization of
parallel single-phase inverters with virtual oscillator control,” IEEE Transactions
on Power Electronics, vol. 29, pp. 6124–6138, Nov 2014.
[45] H. P. Beck and R. Hesse, “Virtual synchronous machine,” in 2007 9th International
Conference on Electrical Power Quality and Utilisation, pp. 1–6, Oct 2007.

142
[46] Q. C. Zhong, “Virtual synchronous machines: A unified interface for grid integra-
tion,” IEEE Power Electronics Magazine, vol. 3, pp. 18–27, Dec 2016.

[47] M. Liserre, F. Blaabjerg, and S. Hansen, “Design and control of an lcl-filter-based

three-phase active rectifier,” IEEE Transactions on Industry Applications, vol. 41,
pp. 1281–1291, Sept 2005.

[48] V. Purba, S. V. Dhople, S. Jafarpour, F. Bullo, and B. B. Johnson, “Reduced-order

structure-preserving model for parallel-connected three-phase grid-tied inverters,”
in 2017 IEEE 18th Workshop on Control and Modeling for Power Electronics
(COMPEL), pp. 1–7, July 2017.

[49] A. Yazdani and R. Iravani, Voltage-Sourced Converters in Power Systems: Model-

ing, Control, and Applications. Wiley, 2010.

[50] J. M. Guerrero, L. Hang, and J. Uceda, “Control of distributed uninterruptible

power supply systems,” IEEE Transactions on Industrial Electronics, vol. 55,
pp. 2845–2859, Aug 2008.

[51] M. C. Chandorkar, D. M. Divan, and R. Adapa, “Control of parallel connected

inverters in standalone ac supply systems,” IEEE Transactions on Industry Appli-
cations, vol. 29, pp. 136–143, Jan 1993.

[52] K. D. Brabandere, B. Bolsens, J. V. den Keybus, A. Woyte, J. Driesen, and R. Bel-

mans, “A voltage and frequency droop control method for parallel inverters,” IEEE
Transactions on Power Electronics, vol. 22, pp. 1107–1115, July 2007.

[53] A. Tuladhar, H. Jin, T. Unger, and K. Mauch, “Control of parallel inverters in

distributed ac power systems with consideration of line impedance effect,” IEEE
Transactions on Industry Applications, vol. 36, pp. 131–138, Jan 2000.

[54] J. W. Simpson-Porco, F. Drfler, and F. Bullo, “Synchronization and power sharing

for droop-controlled inverters in islanded microgrids,” Automatica, vol. 49, no. 9,
pp. 2603 – 2611, 2013.

[55] J. M. Guerrero, L. G. de Vicuna, J. Matas, M. Castilla, and J. Miret, “Output

impedance design of parallel-connected ups inverters with wireless load-sharing
control,” IEEE Transactions on Industrial Electronics, vol. 52, pp. 1126–1135,
Aug 2005.

[56] J. M. Guerrero, J. Matas, L. G. de Vicuna, M. Castilla, and J. Miret, “Decentral-

ized control for parallel operation of distributed generation inverters using resistive
output impedance,” IEEE Transactions on Industrial Electronics, vol. 54, pp. 994–
1004, April 2007.

143
[57] N. R. Chaudhuri and B. Chaudhuri, “Adaptive droop control for effective power
sharing in multi-terminal dc (mtdc) grids,” IEEE Transactions on Power Systems,
vol. 28, pp. 21–29, Feb 2013.

[58] J. Yu, Hybrid AC-High Voltage DC Grid Stability and Controls. Ph.D. dissertation,
Arizona State University (ASU), AZ, Nov. 2017.

[59] P. Rodriguez, A. V. Timbus, R. Teodorescu, M. Liserre, and F. Blaabjerg, “Flexible

active power control of distributed power generation systems during grid faults,”
IEEE Transactions on Industrial Electronics, vol. 54, pp. 2583–2592, Oct 2007.

[60] A. Camacho, M. Castilla, J. Miret, R. Guzman, and A. Borrell, “Reactive power

control for distributed generation power plants to comply with voltage limits during
grid faults,” IEEE Transactions on Power Electronics, vol. 29, pp. 6224–6234, Nov
2014.

[61] J. Rocabert, A. Luna, F. Blaabjerg, and P. Rodrguez, “Control of power converters

in ac microgrids,” IEEE Transactions on Power Electronics, vol. 27, pp. 4734–
4749, Nov 2012.

[62] S. Barsali, M. Ceraolo, P. Pelacchi, and D. Poli, “Control techniques of dispersed

generators to improve the continuity of electricity supply,” in 2002 IEEE Power En-
gineering Society Winter Meeting. Conference Proceedings (Cat. No.02CH37309),
vol. 2, pp. 789–794 vol.2, 2002.

[63] A. Yazdani and P. P. Dash, “A control methodology and characterization of dynam-

ics for a photovoltaic (pv) system interfaced with a distribution network,” IEEE
Transactions on Power Delivery, vol. 24, pp. 1538–1551, July 2009.

[64] Y. Li, D. M. Vilathgamuwa, and P. C. Loh, “Microgrid power quality enhancement

using a three-phase four-wire grid-interfacing compensator,” IEEE Transactions
on Industry Applications, vol. 41, pp. 1707–1719, Nov 2005.

[65] M. Prodanovic and T. C. Green, “High-quality power generation through dis-

tributed control of a power park microgrid,” IEEE Transactions on Industrial Elec-
tronics, vol. 53, pp. 1471–1482, Oct 2006.

[66] M. B. Delghavi and A. Yazdani, “A control strategy for islanded operation of a dis-
tributed resource (dr) unit,” in 2009 IEEE Power Energy Society General Meeting,
pp. 1–8, July 2009.

[67] M. B. Delghavi and A. Yazdani, “A unified control strategy for electronically inter-
faced distributed energy resources,” IEEE Transactions on Power Delivery, vol. 27,
pp. 803–812, April 2012.

144
[68] J. He, Y. W. Li, D. Bosnjak, and B. Harris, “Investigation and active damping of
multiple resonances in a parallel-inverter-based microgrid,” IEEE Transactions on
Power Electronics, vol. 28, pp. 234–246, Jan 2013.
[69] R. Lasseter and P. Piagi, “Providing premium power through distributed resources,”
in Proceedings of the 33rd Annual Hawaii International Conference on System
Sciences, pp. 9 pp.–, Jan 2000.
[70] R. H. Lasseter and P. Paigi, “Microgrid: a conceptual solution,” in 2004 IEEE 35th
Annual Power Electronics Specialists Conference (IEEE Cat. No.04CH37551),
vol. 6, pp. 4285–4290 Vol.6, June 2004.
[71] P. Piagi and R. H. Lasseter, “Autonomous control of microgrids,” in 2006 IEEE
Power Engineering Society General Meeting, pp. 8 pp.–, 2006.
[72] A. A. Renjit, M. S. Illindala, R. H. Lasseter, M. J. Erickson, and D. Klapp, “Model-
ing and control of a natural gas generator set in the certs microgrid,” in 2013 IEEE
Energy Conversion Congress and Exposition, pp. 1640–1646, Sept 2013.
[73] E. Alegria, T. Brown, E. Minear, and R. H. Lasseter, “Certs microgrid demonstra-
tion with large-scale energy storage and renewable generation,” IEEE Transactions
on Smart Grid, vol. 5, pp. 937–943, March 2014.
[74] F. Katiraei, M. R. Iravani, and P. W. Lehn, “Micro-grid autonomous operation dur-
ing and subsequent to islanding process,” IEEE Transactions on Power Delivery,
vol. 20, pp. 248–257, Jan 2005.
[75] F. Katiraei and M. R. Iravani, “Power management strategies for a microgrid
with multiple distributed generation units,” IEEE Transactions on Power Systems,
vol. 21, pp. 1821–1831, Nov 2006.
[76] F. Katiraei, R. Iravani, N. Hatziargyriou, and A. Dimeas, “Microgrids manage-
ment,” IEEE Power and Energy Magazine, vol. 6, pp. 54–65, May 2008.
[77] C. K. Sao and P. W. Lehn, “Control and power management of converter fed micro-
grids,” IEEE Transactions on Power Systems, vol. 23, pp. 1088–1098, Aug 2008.
[78] S. M. Ashabani and Y. A. R. I. Mohamed, “General interface for power manage-
ment of micro-grids using nonlinear cooperative droop control,” IEEE Transac-
tions on Power Systems, vol. 28, pp. 2929–2941, Aug 2013.
[79] H. H. Zeineldin and J. L. Kirtley, “Micro-grid operation of inverter based dis-
tributed generation with voltage and frequency dependent loads,” in 2009 IEEE
Power Energy Society General Meeting, pp. 1–6, July 2009.
[80] Q. Shafiee, J. M. Guerrero, and J. C. Vasquez, “Distributed secondary control for
islanded microgrids; a novel approach,” IEEE Transactions on Power Electronics,
vol. 29, pp. 1018–1031, Feb 2014.

145
[81] F. Guo, C. Wen, J. Mao, and Y. D. Song, “Distributed secondary voltage and fre-
quency restoration control of droop-controlled inverter-based microgrids,” IEEE
Transactions on Industrial Electronics, vol. 62, pp. 4355–4364, July 2015.

[82] Y. Zhang and B. T. Ooi, “Stand-alone doubly-fed induction generators (dfigs) with
autonomous frequency control,” IEEE Transactions on Power Delivery, vol. 28,
pp. 752–760, April 2013.

[83] J. M. Guerrero, J. C. Vasquez, J. Matas, L. G. de Vicuna, and M. Castilla, “Hierar-

chical control of droop-controlled ac and dc microgrids; a general approach toward
standardization,” IEEE Transactions on Industrial Electronics, vol. 58, pp. 158–
172, Jan 2011.

[84] A. Bidram and A. Davoudi, “Hierarchical structure of microgrids control system,”

IEEE Transactions on Smart Grid, vol. 3, pp. 1963–1976, Dec 2012.

[85] D. E. Olivares, A. Mehrizi-Sani, A. H. Etemadi, C. A. Caizares, R. Iravani, M. Kaz-

erani, A. H. Hajimiragha, O. Gomis-Bellmunt, M. Saeedifard, R. Palma-Behnke,
G. A. Jimnez-Estvez, and N. D. Hatziargyriou, “Trends in microgrid control,” IEEE
Transactions on Smart Grid, vol. 5, pp. 1905–1919, July 2014.

[86] R. Majumder, “Some aspects of stability in microgrids,” IEEE Transactions on

Power Systems, vol. 28, pp. 3243–3252, Aug 2013.

[87] P. Kundur, J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares,

N. Hatziargyriou, D. Hill, A. Stankovic, C. Taylor, T. V. Cutsem, and V. Vittal,
“Definition and classification of power system stability IEEE/CIGRE joint task
force on stability terms and definitions,” IEEE Transactions on Power Systems,
vol. 19, pp. 1387–1401, Aug 2004.

[88] R. Majumder, Modeling, stability analysis and control of microgrid. Dissertation,

Queensland University of Technology, Australia, 2010.

[89] M.-S. Ali, Control Strategies for the Next Generation Microgrids. Dissertation,
University of Toronto, Toronto, ON. Canada, 2012.

[90] J. Elizondo and J. L. Kirtley, “Effect of inverter-based dg penetration and control in

hybrid microgrid dynamics and stability,” in 2014 Power and Energy Conference
at Illinois (PECI), pp. 1–6, Feb 2014.

[91] E. Barklund, N. Pogaku, M. Prodanovic, C. Hernandez-Aramburo, and T. C.

Green, “Energy management in autonomous microgrid using stability-constrained
droop control of inverters,” IEEE Transactions on Power Electronics, vol. 23,
pp. 2346–2352, Sept 2008.

146
 [92] M. B. Delghavi and A. Yazdani, “Islanded-mode control of electronically coupled
distributed-resource units under unbalanced and nonlinear load conditions,” IEEE
Transactions on Power Delivery, vol. 26, pp. 661–673, April 2011.
[93] A. Mehrizi-Sani and R. Iravani, “Potential-function based control of a microgrid
in islanded and grid-connected modes,” IEEE Transactions on Power Systems,
vol. 25, pp. 1883–1891, Nov 2010.
[94] N. Soni, S. Doolla, and M. C. Chandorkar, “Improvement of transient response in
microgrids using virtual inertia,” IEEE Transactions on Power Delivery, vol. 28,
pp. 1830–1838, July 2013.
[95] N. Jayawarna, X. Wu, V. Zhang, N. Jenkins, and M. Barnes, “Stability of a micro-
grid,” in The 3rd IET International Conference on Power Electronics, Machines
and Drives, 2006. PEMD 2006, pp. 316–320, April 2006.
[96] F. Katiraei, M. R. Iravani, and P. W. Lehn, “Small-signal dynamic model of
a micro-grid including conventional and electronically interfaced distributed re-
sources,” IET Generation, Transmission Distribution, vol. 1, pp. 369–378, May
2007.
[97] N. Pogaku, M. Prodanovic, and T. C. Green, “Modeling, analysis and testing of au-
tonomous operation of an inverter-based microgrid,” IEEE Transactions on Power
Electronics, vol. 22, pp. 613–625, March 2007.
[98] N. Bottrell, M. Prodanovic, and T. C. Green, “Dynamic stability of a microgrid
with an active load,” IEEE Transactions on Power Electronics, vol. 28, pp. 5107–
5119, Nov 2013.
[99] E. Nasr-Azadani, C. A. Caizares, D. E. Olivares, and K. Bhattacharya, “Stability
analysis of unbalanced distribution systems with synchronous machine and dfig
based distributed generators,” IEEE Transactions on Smart Grid, vol. 5, pp. 2326–
2338, Sept 2014.
[100] J. Sun, “Small-signal methods for ac distributed power systems;a review,” IEEE
Transactions on Power Electronics, vol. 24, pp. 2545–2554, Nov 2009.
[101] J. Sun, “Impedance-based stability criterion for grid-connected inverters,” IEEE
Transactions on Power Electronics, vol. 26, pp. 3075–3078, Nov 2011.
[102] A. C. R. Kamel and K. Nagasaka, “Detailed analysis of micro-grid stability during
islanding mode under different load conditions,” Engineering, vol. Vol. 3, no. No.
5, pp. pp. 508–516, 2011.
[103] E. N. Azadani, C. Canizares, and K. Bhattacharya, “Modeling and stability anal-
ysis of distributed generation,” in 2012 IEEE Power and Energy Society General
Meeting, pp. 1–8, July 2012.

147
[104] R. D. MIDDLEBROOK, “Input filter considerations in design and application of
switching regulators,” IAS’76, 1976.

[105] J. Elizondo, R. Y. Zhang, P. H. Huang, J. K. White, and J. L. Kirtley, “Inertial

and frequency response of microgrids with induction motors,” in 2016 IEEE 17th
Workshop on Control and Modeling for Power Electronics (COMPEL), pp. 1–6,
June 2016.

[106] M. Farrokhabadi, C. A. Caizares, and K. Bhattacharya, “Frequency control in iso-

lated/islanded microgrids through voltage regulation,” IEEE Transactions on Smart
Grid, vol. 8, pp. 1185–1194, May 2017.

[107] Y.Z. Zhang, “Capacity-based adaptive droop control for battery energy storage op-
eration,” in 2017 IEEE Power and Energy Society General Meeting, pp. 1–5, July
2017.

[108] J. Han, “Control theory, is it a model analysis approach or a direct control ap-
proach?,” Syst. Sci. Math., vol. 9, pp. 328–335, May 1989 (in Mandarin Chinese).

[109] J. Han, “From PID to Active Disturbance Rejection Control,” IEEE Transactions
on Industrial Electronics, vol. 56, pp. 900–906, March 2009.

[110] J. C. Das, “Effects of momentary voltage dips on the operation of induction

and synchronous motors,” IEEE Transactions on Industry Applications, vol. 26,
pp. 711–718, Jul 1990.

[111] T. Short, Electric Power Distribution Handbook, Second Edition. Taylor & Francis,
2014.

[112] H. W. Dommel, “Digital computer solution of electromagnetic transients in single-

and multiphase networks,” IEEE Transactions on Power Apparatus and Systems,
vol. PAS-88, pp. 388–399, April 1969.

[113] H. Willis, Power Distribution Planning Reference Book, Second Edition. Power
Engineering (Willis), CRC Press, 2004.

[114] W. H. Kersting, “Radial distribution test feeders,” in 2001 IEEE Power Engineering
Society Winter Meeting. Conference Proceedings (Cat. No.01CH37194), vol. 2,
pp. 908–912 vol.2, 2001.

[115] “Standard load models for power flow and dynamic performance simulation,” IEEE
Transactions on Power Systems, vol. 10, pp. 1302–1313, Aug 1995.

[116] P. Krause, O. Wasynczuk, S. Sudhoff, and S. Pekarek, Analysis of Electric Machin-

ery and Drive Systems. IEEE Press Series on Power Engineering, Wiley, 2013.

148
[117] J. J. Cathey, R. K. Cavin, and A. K. Ayoub, “Transient load model of an induc-
tion motor,” IEEE Transactions on Power Apparatus and Systems, vol. PAS-92,
pp. 1399–1406, July 1973.

[118] E. Kimbark, Power System Stability, Volumes I, II, III, 3 Volume Set. IEEE Press
Series on Power Engineering, Wiley, 1995.

[119] “Matlab simscape power systems blocks reference,” tech. rep., MathWorks, 1 Ap-
ple Hill Drive, Natick, MA 01760-2098, January 2018.

[120] J. Grainger and W. Stevenson, Power system analysis. McGraw-Hill series in elec-
trical and computer engineering: Power and energy, McGraw-Hill, 1994.

[121] N. Jaleeli, L. S. VanSlyck, D. N. Ewart, L. H. Fink, and A. G. Hoffmann, “Un-

derstanding automatic generation control,” IEEE Transactions on Power Systems,
vol. 7, pp. 1106–1122, Aug 1992.

[122] “IEEE Recommended Practice for Excitation System Models for Power System
Stability Studies,” IEEE Std 421.5-2016 (Revision of IEEE Std 421.5-2005), pp. 1–
207, Aug 2016.

[123] I. C. Report, “Excitation system models for power system stability studies,” IEEE
Transactions on Power Apparatus and Systems, vol. PAS-100, pp. 494–509, Feb
1981.

[124] Y. G. Rebours, D. S. Kirschen, M. Trotignon, and S. Rossignol, “A survey of fre-

quency and voltage control ancillary services mdash;part i: Technical features,”
IEEE Transactions on Power Systems, vol. 22, pp. 350–357, Feb 2007.

[125] K. E. Yeager and J. R. Willis, “Modeling of emergency diesel generators in an 800

megawatt nuclear power plant,” IEEE Transactions on Energy Conversion, vol. 8,
pp. 433–441, Sep 1993.

[126] Ibraheem, P. Kumar, and D. P. Kothari, “Recent philosophies of automatic gener-

ation control strategies in power systems,” IEEE Transactions on Power Systems,
vol. 20, pp. 346–357, Feb 2005.

[127] T. Zhang, J. A. Orr, and A. E. Emanuel, “Adaptable energy storage system control
for microgrid stability enhancement,” in 2018 IEEE Power and Energy Society
General Meeting, pp. 1–5, August 2018[Accepted].

[128] Tan Zhang, Adaptive Converter Control for Fast Frequency Regulation. Provisional
Patent Number 62/645,393, filed March 20, US, 2018.

[129] J. Elizondo and J. L. Kirtley, “Effect of DG and induction motor load power rating
on microgrid transient behavior,” in ISGT 2014, pp. 1–5, Feb 2014.

149
[130] G. K. Stefopoulos and A. P. Meliopoulos, “Induction motor load dynamics: Im-
pact on voltage recovery phenomena,” in 2005/2006 IEEE/PES Transmission and
Distribution Conference and Exhibition, pp. 752–759, May 2006.

[131] J. Elizondo, P.-H. Huang, J. L. Kirtley, and M. S. Elmoursi, “Enhanced critical

clearing time estimation and fault recovery strategy for an inverter-based micro-
grid with IM load,” in 2016 IEEE Power and Energy Society General Meeting
(PESGM), pp. 1–5, July 2016.

[132] “NEMA MG 1-2016,” Mar 2017.

[133] D. Kosterev, A. Meklin, J. Undrill, B. Lesieutre, W. Price, D. Chassin, R. Bravo,

and S. Yang, “Load modeling in power system studies: WECC progress update,”
in 2008 IEEE Power and Energy Society General Meeting (PESGM), pp. 1–8, July
2008.

[134] D. James and J. D. Kueck, “Commercial Building Motor Protection Response Re-
port, No. PNNL-24468,” June 2015.

[135] A. H. Bonnett, “The impact that voltage and frequency variations have on ac induc-
tion motor performance and life in accordance with nema mg-1 standards,” in Pulp
and Paper, 1999. Industry Technical Conference Record of 1999 Annual, pp. 16–
26, June 1999.

[136] H. Vu, P. Pruvot, C. Launay, and Y. Harmand, “An improved voltage control
on large-scale power system,” IEEE Transactions on Power Systems, vol. 11,
pp. 1295–1303, Aug 1996.

[137] “Powerworld simulator 18 block diagrams,” tech. rep., PowerWorld Corporation,

2001 S First Street Champaign, IL 61820, May 2014.

150