CONTROL OF DISTRIBUTED GENERATION FOR GRID-CONNECTED AND

INTENTIONAL ISLANDING OPERATIONS

By

Irvin Joel Balaguer Álvarez

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Electrical Engineering

2011
 ABSTRACT

CONTROL OF DISTRIBUTED GENERATION FOR GRID-CONNECTED AND

INTENTIONAL ISLANDING OPERATIONS

By

Irvin Joel Balaguer Álvarez

The current model for electricity generation and distribution in the United States is

dominated by centralized power plants. The power at these plants is typically combustion (coal,

oil, and natural) or nuclear generated. Centralized power models, like this, require distribution

from the center to outlying consumers. This system of centralized power plants has many

disadvantages. Electric utilities are becoming more and more stressed since existing transmission

and distribution systems are facing their operating constraints with growing load. Greenhouse

gas emissions have resulted in a call for cleaner renewable power sources. Under such

circumstances, distributed generation (DG) with alternative sources; such as fuel-cell, wind-

turbine, bio-mass, micro-turbine and solar-cell systems; has been considered as a promising

solution to the above problems.

DG is defined as small, modular electricity generators located close to the end customer's

load connection point. DGs can enable utilities to decrease investment costs in transmission and

distribution system upgrades while still meeting increasing power demands. Also, DGs provide

customers with improved quality and reliability of energy supplies without imposing undesirable

effects on environment. In general, DG can be intended as small sized power plants that are

designed to be installed and operated within a local load center.

 This research presents the development and test of a control strategy for DG capable of

working in both intentional islanding (stand-alone) and grid-connected modes. In the grid-

connected mode of operation, the DG is connected to the utility. The utility, which is assumed to

be stiff, sets the voltage at the terminal of the DG inverter. The inverter controls the power being

injected into the grid by controlling the injected current. Thus, in this mode, the inverter operates

in the current control mode. In the stand-alone mode, the inverter supplies power to the load. It

has to maintain the voltage at the terminals of the load, irrespective of any changes in the load.

Thus, in this mode, the inverter operates in the voltage controlled mode.

The stand-alone control features an output voltage controller capable of handling deficit of

generated power (load shedding) and synchronization for grid reconnection with a seamless

transition from stand-alone to grid-connected operation modes. The grid-connected mode with

current control is also enabled for the case of power grid connection. This grid-connected control

features an output current controller capable of loss of main detection, synchronization with the

grid, and seamless transition from grid-connected to stand-alone operation modes with minimum

interruption to the load.

The operational principle and control method of the proposed system are explained in detail.

A 10kW DG inverter has been designed, built and set up for testing. Simulation and

experimental results are provided in order to verify the validity of the developed DG system.
To the only wise God our Savior, be glory and majesty, dominion and power, both now and ever.

Amen. (Jude 1:25 King James Version)

To my wife Raquel, my son Kevin, and my daughters Kelly and Kiara, for their patience,

unconditional love, and support.

iv
 ACKNOWLEDGMENTS

First and foremost I would like to thank God for giving me the drive, ability and patience to

strive for the completion of this dissertation.

I am forever grateful to Dr. Fang Z. Peng for providing me the opportunity to work with him

and for his kind words of advice, guidance and encouragement throughout my stay at Michigan

State University.

I would also like to thank Dr. Robert Schlueter, Dr. Elias Strangas, and Dr. Robert Ofoli for

agreeing to be members of my doctoral guidance committee, for their invaluable comments and

time.

I extend a special thank you to Dr. Percy Pierre and Dr. Barbara O‟Kelly of the Sloan

Engineering Program in MSU's College of Engineering and to Professor Aníbal Romney,

Chairperson of the Department of Electronics Engineering Technology of the University of

Puerto Rico-Aguadilla Campus, for the generous financial support without which this work

would not have been possible.

I deeply appreciate the continued support, cooperation, and valuable discussions that I

engaged in with the Power Electronics and Motor Drives Laboratory colleagues who I was

involved with: Dr. Eduardo Ortiz, Dr. Lihua Chen, Dr. Miaosen Shen, Dr. Haiping Xu, Dr.

Heung-Geun Kim, Dr. Nam-Sup Choi, Dr. Yuan Li, Dr. Honnyong Cha, Dr. Yi Huang, Miss Qin

Lei, Mr. Uthane Supatti, Mr. Shuitao Yang, Mr. Craig Rogers, Mr. Matthew Gebben, Mr. Jorge

Cintrόn, Mr. Dong Cao, Miss Xi Lu, Miss Wei Qian, Mr. Varun Chengalvala, Mr. Richard

Badin, Mr. Alan Joseph, Mr. Joel Anderson, and Mr. Carlos Niño. This long list is still not

complete and I apologize to those not listed as you all contributed to my experience in the lab.

v
 A special thanks to Miss Qin Lei, Mr. Shuitao Yang, Mr. Jorge Cintrόn, and Mr. Uthane

Supatti for all their help with the hardware and experimental setups. To my proofreaders Mr.

Craig Rogers, Mr. Matthew Gebben, Mrs. Deana Hutchins, Mr. Erick Rivas, Mrs. Aurelia Rivas,

Mr. Jason Lounds, Mr. Alex Fitzpatrick, and Mrs. Deb Fitzpatrick, thanks all you very much.

I would also like to express my gratitude to Mrs. Sheryl Hulet, former graduate secretary and

Mrs. Meagan Kroll, the current graduate secretary at the Department of Electrical and Computer

Engineering at Michigan State University for their tremendous help with all the procedural paper

work throughout my time in the Ph.D. program.

From the bottom of my heart, I‟m most grateful to my parents, Mr. Ramón Balaguer and

Mrs. Carmen Rita Álvarez, and my brothers, Victor and José, for their unwavering support,

constant encouragement, and extraordinary understanding throughout my graduate studies. I‟m

equally grateful to my parents-in-law, Mr. Adalberto Lebrón, Mrs. Esther Salas, and my

brothers-in-law, Adalberto Jr. and Marcos, for their good wishes, and specifically, for their faith

in my abilities. I truly believe that this work could only have been brought to completion via the

blessings and good wishes from all of them.

A special thank you goes out to all of our extended family from the Lansing Seventh-day

Adventist Church, Greater Lansing Adventist School, Lansing Spanish Seventh-day Adventist

Church, and the University Seventh-day Adventist Church in East Lansing.

Finally, words cannot articulate the companionship and support provided to me by my lovely

wife, Raquel, my son Kevin and my two daughters, Kelly and Kiara. I would have left this work

long ago without their presence in my life. Obtaining a Ph.D. is a complex undertaking. At times

it is uncomfortable and challenging, and at others utterly rewarding. There have been many

peaks and troughs along the way. Raquel has been my pillar through them all. She has always

vi
been willing to make sacrifices in her life to accommodate my work schedule. I could not have

asked for more, and I genuinely thank her for the role she has played in making this endeavor a

success.

vii
 TABLE OF CONTENTS

LIST OF FIGURES .......................................................................................................................x

CHAPTER 1. INTRODUCTION .................................................................................................1

1.1. The Distributed Generation Concept ....................................................................................1
1.2. Benefits associated with DG .................................................................................................3
1.3. Issues associated with DG ....................................................................................................5
1.4. Scope of the Dissertation ....................................................................................................11
1.5. Outline of the Dissertation ..................................................................................................12

CHAPTER 2. LITERATURE REVIEW ...................................................................................13

CHAPTER 3. SYSTEM DESCRIPTION ..................................................................................37

3.1. Circuit Topology .................................................................................................................37
3.2. Voltage Source Inverter ......................................................................................................38
3.3. LCL filter ............................................................................................................................41
3.3.1. Specifications for the design of LCL filter ..................................................................42
3.3.2. Design of L ................................................................................................................43
I
3.3.3. Design of C f ..............................................................................................................44
3.3.4. Design of L ...............................................................................................................44
g
3.4. Simulation and Experimental Set-ups .................................................................................45
3.4.1. Simulation Set-up.........................................................................................................45
3.4.2. Experimental Set-up.....................................................................................................46

CHAPTER 4. PROPOSED CONTROL FOR GRID-CONNECTED OPERATION OF DG49

4.1. Introduction .........................................................................................................................49
4.2. Circuit Topology .................................................................................................................49
4.3. Controller ............................................................................................................................50
4.3.1. Synchronization Controller for Grid Reconnection: Proposed Algorithm ..................52
4.4. Transfer Functions ..............................................................................................................54
4.5. Simulation Results ..............................................................................................................56
4.6. Experimental Results ..........................................................................................................60

CHAPTER 5. PROPOSED CONTROL FOR INTENTIONAL ISLANDING OPERATION

OF DG WITH SEAMLESS TRANSITION FROM GRID CONNECTED OPERATION .63
5.1. Introduction .........................................................................................................................63
5.2. Grid Condition Detection ....................................................................................................64
5.3. Intelligent Load Shedding ...................................................................................................68
5.4. Transition from Grid-connected to Stand-alone: Proposed Controller ...............................73
5.5. Transfer Functions ..............................................................................................................75
5.6. Simulation Results ..............................................................................................................77

viii
 5.7. Experimental Results ..........................................................................................................82

CHAPTER 6. CONCLUSIONS AND FUTURE WORKS ......................................................86

6.1 Conclusions ..........................................................................................................................86
6.2 Future Works .......................................................................................................................87

APPENDICES ..............................................................................................................................89
Appendix 1. Transfer Functions of the current controlled inverter ...........................................90
Appendix 2. Derivation of the Load Shedding Equations .......................................................100
Appendix 3. Transfer Functions of the voltage controlled inverter .........................................104
Appendix 4. Block Diagram of the 2407A DSP Controller ....................................................109

REFERENCES ...........................................................................................................................110

ix
 LIST OF FIGURES

Fig. 2. 1 Real power compensator with synchronization function ............................................... 15

Fig. 2. 2 Reactive power compensator with synchronization function ......................................... 15

Fig. 2. 3 Overview of proposed unified controller........................................................................ 15

Fig. 2. 4 System configuration for simulated and experimental results........................................ 16

Fig. 2. 5 Islanding Detection Flowchart........................................................................................ 17

Fig. 2. 6 Re-closure Flowchart ...................................................................................................... 19

Fig. 2. 7 Proposed islanding detection algorithm ......................................................................... 21

Fig. 2. 8 Flowchart of the control strategy implemented to produce intentional islanding .......... 22

Fig. 2. 9 Block diagram of voltage control ................................................................................... 24

Fig. 2. 10 Block diagram of voltage control with positive feedback ............................................ 24

Fig. 2. 11 Diagram of magnitude of voltage feedback in dq frame .............................................. 29

Fig. 2. 12 Diagram of frequency feedback in dq frame ................................................................ 29

Fig. 2. 13 Islanding-detection scheme .......................................................................................... 30

Fig. 2. 14 Re-closure scheme ........................................................................................................ 31

Fig. 2. 15 Flowchart diagram of the modes of the VSI................................................................. 32

Fig. 2. 16 Block diagram of the whole proposed controller using the synchronization control
loops .............................................................................................................................................. 32

Fig. 3. 1 Schematic diagram of the grid-connected inverter system ............................................. 37

Fig. 3. 2 Typical diagram of micro-source generation system ..................................................... 38

Fig. 3. 3 DC to Three-Phase AC Inverter Diagram ...................................................................... 39

Fig. 3. 4 PWM Operation of the Bridge........................................................................................ 41

x
Fig. 3. 5 LCL Filter configuration circuit ..................................................................................... 42

Fig. 3. 6 Simulated system ............................................................................................................ 46

Fig. 3. 7 Experimental set-up ........................................................................................................ 48

Fig. 4. 1 Schematic diagram of the grid-connected inverter system ............................................. 50

Fig. 4. 2 Block diagram of the current controller for grid-connected ........................................... 51

Fig. 4. 3 Synchronization controller ............................................................................................. 54

Fig. 4. 4 Block diagram of the current controlled inverter ........................................................... 54

Fig. 4. 5 LCL Filter and parallel RLC load................................................................................... 55

Fig. 4. 6 Current controller for grid-connected operation ............................................................. 57

Fig. 4. 7 Phase voltages and currents during grid-connected operation ....................................... 58

Fig. 4. 8 Synchronization for grid re-connection .......................................................................... 59

Fig. 4. 9 Phase voltage without (top) and with (bottom) synchronization .................................... 60

Fig. 4. 10 Line to line voltages and phase current during grid-connected operation .................... 61

Fig. 4. 11 Transition from stand-alone to grid connected operation ............................................. 62

Fig. 5. 1 Simulation waveforms of load voltages and grid currents during the grid disconnection
....................................................................................................................................................... 63

Fig. 5. 2 PLL structure .................................................................................................................. 65

Fig. 5. 3 Loss of Main Detection .................................................................................................. 66

Fig. 5. 4 Voltage or frequency change at grid disconnection ....................................................... 67

Fig. 5. 5 Islanding Detection Algorithm ....................................................................................... 67

Fig. 5. 6 Voltage Transients under various Active Power Differences ........................................ 69

Fig. 5. 7 Frequency Transients under various Reactive Power..................................................... 70

Fig. 5. 8 System to implement load shedding ............................................................................... 71

Fig. 5. 9 System in per unit to implement load shedding ............................................................. 71

xi
Fig. 5. 10 Voltage controlled inverter ........................................................................................... 74

Fig. 5. 11 Block diagram of the voltage controlled inverter ......................................................... 75

Fig. 5. 12 Simulated systems ........................................................................................................ 78

Fig. 5. 13 From grid-connected to stand-alone operation with severe transients ......................... 79

Fig. 5. 14 From grid-connected to stand-alone operation without severe transients .................... 79

Fig. 5. 15 From grid-connected to stand-alone operation without severe transients .................... 80

Fig. 5. 16 Phase voltage Va without load shedding algorithm...................................................... 81

Fig. 5. 17 Phase voltage Va with load shedding algorithm ........................................................... 81

Fig. 5. 18 Transition from grid-connected to stand-alone operation with severe transients ......... 82

Fig. 5. 19 Transition from grid-connected to intentional islanding operation without severe

transients ....................................................................................................................................... 83

Fig. 5. 20 Line to line voltages and phase currents during intentional islanding operation ......... 84

Fig. 5. 21 DG line to line voltages and phase currents during intentional islanding operation .... 84

Fig. 5. 22 Implementation of the load shedding algorithm ........................................................... 85

Fig. A1. 1 Block diagram of the current controlled inverter ........................................................ 90

Fig. A1. 2 LCL Filter and parallel RLC load ................................................................................ 91

Fig. A2. 1 System to implement load shedding .......................................................................... 100

Fig. A3. 1 Block diagram of the voltage controlled inverter ...................................................... 104

Fig. A3. 2 LCL Filter and parallel RLC load .............................................................................. 105

Fig. A4. 1 Block diagram of the 2407A DSP controller ............................................................. 109

xii
CHAPTER 1. INTRODUCTION

The current model for electricity generation and distribution in the United States is

dominated by centralized power plants. The power at these plants is typically combustion (coal,

oil, and natural) or nuclear generated. Centralized power models like this require distribution

from the center to outlying consumers.

This system of centralized power plants has many disadvantages. Electric utilities are

becoming more and more stressed since existing transmission and distribution systems are facing

their operating constraints with growing load. Greenhouse gas emissions have resulted in a call

for cleaner renewable power sources. Under such circumstances, distributed generation (DG)

with alternative sources such as fuel-cell, wind-turbine, bio-mass, micro-turbine and solar-cell

systems, has been considered as a promising solution to the above problems.

1.1. The Distributed Generation Concept

DG is defined as small, modular electricity generators located close to the end customer's

load connection point. They can enable utilities to decrease investment costs in transmission and

distribution system upgrades while still meeting increasing power demands and provide

customers with improved quality and reliability of energy supplies without imposing undesirable

effects on environment [1-2]. In general, DG can be intended as small sized power plants that are

designed to be installed and operated within a local load center.

To effectively and efficiently connect any of the DG sources to the existing power systems

power electronics-based conversion systems need to be developed for the proper control and

conditioning of the energy to be delivered [3]. Specifically, with most of the DG sources such as

1
variable frequency AC microturbines, DC fuel cells, DC photovoltaic cells and low power

variable frequency AC wind turbines, electrical power is generated as DC voltage or converted

to DC voltage, then converted to AC using a voltage sourced inverter. This voltage sourced

inverter performs the interface function between the DC bus and the AC world. Through the

proper control and conditioning of the DG, benefits such as voltage support and improved power

quality, diversification of power sources, reduction in transmission and distribution losses,

transmission and distribution capacity release and improved reliability, among others, can

enhance the utility grid without having to add or replace the existing transmission/distribution

system.

Some issues and concerns must be addressed when dealing with DG systems. Particularly,

attention must be taken into account when a design is being done for the support of power

delivery to the utility and when incorporating the concepts of both grid-connected and stand-

alone. During the grid-connected operation, each DG system is usually operated to provide or

inject pre-set power to the grid, which is the current control mode in stiff synchronization with

the grid [4-6]. When the main grid is cut off from the DG system, stand-alone operation or

intentional islanding, the DG system has to detect this islanding situation and must be switched

to a voltage control mode to provide constant voltage to the local sensitive loads [7-9]. The trend

is that they should be able to work in stand-alone mode but also connected to the power grid

(grid-connected) [10]. Thus some new challenges on the control side of the DG occurred. Among

these challenges are: reasonable voltage regulation in stand-alone mode, grid-connection mode

enabled, and automatic detection of grid disconnection. Also grid disconnection detection and an

automatic mode switching are required. In order to manage these challenges, relative complex

control strategies need to be developed.

2
1.2. Benefits associated with DG

In distribution systems, DG can provide benefits for the consumers as well as for the utilities,

especially in sites where there are deficiencies in the transmission system. Some of the expected

benefits of DG are [11-19]:

 Green house emissions reductions: By the increase of the use of renewable energy units,

as well as high efficiency generation units, operated in an optimum manner, the green

house emissions will be decreased with respect to the conventional generation.

 Energy efficiency: By an adequate planning and operation of the generation and storage

units of the DG, the electric and heat generation or combined heat and power (CHP) can

be mixed, increasing the energy efficiency of the installation. It can also be made in a

profitable way.

 Reduced transmission and distribution investments: DG helps bypass „„congestion‟‟ in

existing transmission grids. DG could serve as a substitute for investments in

transmission and distribution capacity. DG can postpone the need for new infrastructure.

Because of opportunities for integration in buildings, DG development often occurs in the

same location as demand. In such cases, if production output is concurrent with demand –

such as demand for air-conditioning in hot regions – network reinforcement may be

unnecessary while generation remains in the same order of magnitude as demand.

Moreover, normal development of the grid in response to growing demand may also be

postponed or even avoided as DG has the net effect of decreasing demand in that area.

 Minimization of the electric losses: On-site production reduces the amount of power that

must be transmitted from a centralized plant, and avoids resulting transmission losses and

3
 distribution losses, as well as the transmission and distribution costs, a significant part of

the total electricity cost, due to the fact that generation buses and consumption are closer.

 Network (voltage) support: The connection of distributed generators to networks

generally leads to a rise in voltage in the network.

 Quality of supply improvement: In areas where voltage support is difficult, installation of

a distributed generator may improve quality of supply. As the demand for more and

better quality electric power increases, DG can provide alternatives for reliable, cost-

effective, premium power for homes and businesses.

 New market opportunities and enhanced industrial competitiveness: DG can also

stimulate competition in supply; adjusting price via market forces. In a free market

environment, DG operator can buy or sell power to the electricity grid, exporting only at

peak demand and purchasing power at off-peak prices.

 Reduction of the energy costs: Thanks to the intelligence and control capabilities of the

DG, its operator will schedule the operation of its generation and storage sources,

depending on the electricity and gas actual prices, climatic conditions and their forecast.

 Locality, i.e. improved utilization of local resources: Distributed energy generation may

also promote local business opportunities, and develop products and services based on

local raw materials and labor. Local employment can be improved by creating new jobs

related to distributed energy generation. This, in turn, causes a need for high-quality

education. Locality also means the absence of transmission lines, large power plants, and

fuel storage, which spoil the landscape. The environmental load is also reduced due to the

avoidance of additional energy required to compensate transmission losses.

4
1.3. Issues associated with DG

Despite the above mentioned benefits provided by DG, there are technical limits regarding

the degree to which DG can be connected. Indiscriminate application of individual DG systems

can cause as many problems as it may solve [11]. This is because the distribution system was

intended to cope with the conventional load supply by central generation, where power flows

radially from the transmission network. Changing the power flow causes problems since DG

does not behave the same way as conventional load. The issues associated with DG can be

summarized as [14, 17, 19-27]:

 Interface with AC System: The main DG requirement for interfacing with the utility

system is that it must not compromise the stability and reliability of the grid. The

interface must also guarantee compliance with already existing protection schemes.

 Protection: DG needs to be retrofitted in its protection algorithms to include the

contributions of micro-sources. DG can impact a variety of levels of short-circuit current.

The DG systems have to provide enough fault current to operate the protective devices,

including circuit breakers, fuses, and fault-protection relays. The addition of DG on a

circuit may need to be studied to establish whether changes are needed for coordination

or protection equipment. Different hardware setups may be required: the rating of the

breakers responsible to implement protection may need to be revisited, and surely they

will have to be controlled with a different algorithm.

 Power quality: DG can negatively affect power quality. Power quality refers to the degree

to which power characteristics align with the ideal sinusoidal voltage and current

waveform, with current and voltage in balance.

o Voltage Regulation: The primary objective of voltage regulation is to provide

each customer connected to the utility with voltage that conforms to limits voltage

5
 range for normal operation. The operating window for DG systems is 106-132

Vrms on a 120 Vrms base, that is, 88 - 110% of nominal voltage [28].

o Harmonics: Due to the power electronics and digital methods used to form the AC
waveform from DC, inverter-based DG technologies produce various harmonics
of the power system frequency. Standards states that the total harmonic distortion
(THD) must not exceed 5% of the fundamental 60 Hz frequency, nor 3% of the
fundamental for any individual harmonic [28].
o Power Factor: Power factor measures the apparent power that is generated when

the voltage and current waveforms are out of synchronism. Power factor is the

ratio of true electric power (watts), to the apparent power (kVA). Although not

strictly the case, power factor problems can be thought of as contributing to utility

system inefficiencies [29]. The DG inverter should operate at a power factor >

0.85 (lagging or leading) [28]. Most DG inverters designed for utility-

interconnected service operate close to unity power factor. Specially designed

systems that provide reactive power compensation may operate outside of this

limit with utility approval.

o Frequency: DG systems have to operate in synchronism with the utility. DG

systems installed in North America should have a fixed operating frequency range
of 59.3 - 60.5 Hz [28]. Systems installed in another country should follow the
frequency operating window standards of that country.
o DC injection: DC injection occurs when an inverter passes unwanted DC current
into the AC or output side of the inverter. The DG inverter should not inject dc
current > 0.5% of rated inverter output current into the ac interface under either
normal or abnormal operating conditions [28]. Inverter manufacturers generally
use one of two methods to prevent the injection of dc current into the utility
interface. One method is to incorporate an ac output isolation transformer in the
inverter. The other method, which uses a shunt or dc-current sensor, initiates

6
 inverter shutdown when the dc component of the current exceeds the specified
threshold.
 Instantaneous power tracking: The requested power from the load coming on-line is a

step function, while the prime mover in the micro-source always takes a finite amount of

time to ramp up to the newly requested value. Micro-sources have a slow response to

changes in commands and the inverter interface by itself does not provide any kind of

internal form of energy storage. This inertia-less system is not well suited to handle step

changes in the requested output power. If the connection to the grid is missing due to a

temporary malfunction then the need for some sort of storage is manifest. Storage is

required to satisfy the instantaneous power balance as a new load comes on-line without

penalizing the quality of other network quantities, such as bus voltage magnitude. Load

changes resulting in fast transients that exceed the ramping capability of generation

require storage availability from which to draw the required transient energy.

 Unidirectional Area Electric Power System (EPS) Relaying: Due to the traditional radial

nature of distribution systems, most protective devices on the EPS are unidirectional in

nature and respond to a given value of current without regard to the direction of flow of

that current. Because DG produces fault current of various magnitudes for faults on the

EPS, the traditional radial nature of the distribution system is disrupted and unidirectional

relays could operate improperly.

 Paralleling DG: One of the selling points for DG is that the local generation follows a

scalable system model, where due to the typical small size of the units, one can install as

many units as needed to satisfy the requests of the loads, without having too much of

extra capacity sitting idle. But this concept requires that the micro-sources can be

installed in parallel without any restrictions.

7
 Plug and Play Configuration: When looking into the future and envisioning micro-

sources being installed by the thousands, the need for plug and play configuration

becomes manifest. The plug and play mode of operation implies simplicity of installation

and therefore speeds up the process of diffusion of the DG in the system. The ideal case

would be purchasing a unit and plugging it in a three-phase socket, having power injected

immediately (after synchronization).

 Synchronization with the Grid: One of the most important issues of a DG system

connected to the utility network is the synchronization with the grid voltage vector. The

synchronization algorithm mainly outputs the phase of the grid voltage vector. The phase

angle of the utility voltage is a critical piece of information for grid connected systems.

This information can be used to synchronize the turning on/off of the power devices,

calculate and control the flow of active/reactive power or transform the feedback

variables to a reference frame suitable for control purposes. Because of these, the

accurate and fast detection of this phase angle is essential to assure the correct generation

of the reference signals.

 Loss of Main Detection (Grid condition detection) or Islanding Detection: One of the

major concerns in operating DG systems connected to the grid is the possibility of

islanding due to grid disturbances, intentional disconnect for servicing, accidental

opening, intentional disconnect from the utility, and an act of nature, among others.

Islanding is the condition in which a portion of the utility system, which contains both

load and DG, is isolated from the remainder of the utility system and continues to

operate. Islanding is either due to preplanned (intentional) events or due to non-planned

or accidental (un-intentional) events [10]. Some distinctions of islanding are:

8
 o non-intentional islanding occurs if it is not possible to disconnect the DG after the

fault, non-intentional islands must then be detected and eliminated as fast as

possible;

o intentional islanding refers to the formation of islands of predetermined or

variable extension; these islands have to be supplied from suitable sources able to

guarantee acceptable voltage support and frequency, controllability and quality of

the supply, and may play a significant role in assisting the service restoration

process;

o microgrids are seen as particular types of intentional islands basically operated in

autonomous mode, not connected to the supply system; the whole microgrid can

be seen from the distribution system as a single load and has to be designed to

satisfy the local reliability requirements, in addition to other technical

characteristics concerning frequency, voltage control and quality of supply.

In a deregulated market environment, current practices of disconnecting the DG

following a disturbance will no longer be a practical or reliable solution. As a result, the

IEEE Std. 1547-2003 states as one of its tasks for future consideration of the

implementation of stand-alone operation of DG (IEEE P1547.4: Guide for Design,

Operation, and Integration of Distributed Resource Island Systems with Electric Power

Systems) [28].

 Transitions from grid-connected to stand-alone: One of the problems to face during the

transfer to stand-alone is to ensure minimal to none transients at all the load bus. The DG

inverter should operate in grid-connected and stand-alone modes in order to provide

power to the emergency load during system outages. However, grid current controller

9
 (for grid-connected) and output voltage controller (for stand-alone) are switched between

the two modes, so the outputs of both controllers may not be equal during the transfer

instant, which will cause the current or voltage spikes during the switch process.

 Load shedding: Another problem to face during the transfer to stand-alone is to

automatically readjust the power output of the units to compensate for the missing quota

of injected power from the utility system. This issue brings a challenge because of the

assumption that there is no previous knowledge of the loading levels as the system

transfer to island and also because of the assumption that no communication may exist

between the micro-sources to enforce a coordinated behavior during and after the transfer

to island.

 Reclosing: Automatic circuit reclosers may also be deployed in the feeder circuit to

automatically clear faults and quickly restore service on the feeder. Reclosers reenergize

the circuit automatically at a predetermined time after a trip resulting from a feeder fault.

The response of the DG unit needs to be coordinated with the reclosing strategy and the

settings of the recloser isolation operations. Coordination and synchronization are

required to prevent possible damage to the equipment connected to the DG.

 Transitions from stand-alone to grid-connected: When the grid-disconnection cause

disappears, the transition from stand-alone mode to grid-connected mode can be started.

The voltage amplitude and phase of the DG have to be synchronized with the grid to

avoid hard transients in the reconnection. Once the synchronization process is completed,

the DG can be reconnected to the grid.

10
1.4. Scope of the Dissertation

This research presents the development and test of a control strategy for DG capable of

working in both stand-alone and grid-connected modes. The stand-alone control features an

output voltage controller capable of handling excess or deficit of generated power and

synchronization for grid reconnection with a seamless transition from stand-alone to grid-

connected operation modes. The grid-connected mode with current control is also enabled for the

case of power grid connection. This grid-connected control features an output current controller

capable of loss of main detection, synchronization with the grid, and seamless transition from

grid-connected to stand-alone operation modes with minimum interruption to the load.

A method to automatically switch between both modes of operation is described. This

method, based on a phase-locked loop (PLL), detects the power grid disconnection or recovery,

and switches the operation mode accordingly.

The proposed control strategy will have the following characteristics:

1. Interface with AC System

2. Power quality

a. Voltage Regulation

b. Harmonics

3. Synchronization with the Grid

4. Loss of Main Detection or Islanding

5. Transitions from grid-connected to stand-alone

6. Load shedding

7. Reclosing

8. Transitions from stand-alone to grid-connected

11
 Operation modes as well as the involved voltage and current controllers will be described in

detail. The design of an LCL filter to achieve attenuation of the switching frequency ripple in the

output voltage will be also described. A PLL for grid synchronization and loss of main detection

will be presented. Also, a re-connection algorithm and automatic mode switching will be

delineated.

1.5. Outline of the Dissertation

A brief description of the Distributed Generation Concept is presented in this chapter.

Chapter 2 presents a literature review of the techniques used for the control of an inverter unit in

grid-connected and stand-alone modes. The system description including circuit topology,

voltage source inverter, LCL filter, simulation set-up, and experimental set-up are given in

Chapter 3. Chapter 4 discusses the proposed control for grid-connected operation of DG. The

proposed control for intentional islanding operation is presented in Chapter 5. Conclusions and

the scope for future work are presented in Chapter 6.

12
CHAPTER 2. LITERATURE REVIEW

This section summarizes the techniques used for the control of an inverter unit in grid-

connected and stand-alone modes.

R. Tirumala, et al. [30] presented an algorithm for a utility interactive DG system with a

seamless transition between grid-tied (current controller) and off-grid (voltage controller) modes

of operation. A static transfer switch (STS) was used to disconnect and reconnect the load bus to

the grid. The control ensured a smooth voltage profile across the load when the grid was

disconnected or reconnected. The control algorithm ensured that the static transfer switch was

turned off before the inverter was shifted to the voltage-controlled mode. Also, the algorithm

ensured that the voltage applied across the load matched the load voltage just before

disconnection. The steps to perform this algorithm can be summarized as follows:

1. Detect a fault on the grid and give a turn off signal to the STS.

2. Monitor the magnitude and phase of the load voltage.

3. When the STS current goes to zero, switch the inverter to a voltage-controlled mode,

with the voltage reference being derived from the load voltage.

4. Ramp up the magnitude of the load voltage from its initial value to the rated value.

For the transfer between the voltage-controlled mode and the current controlled mode, the

inverter voltage matched the grid voltage both in magnitude and phase before the static transfer

switch was turned on. The STS would be turned on when there is essentially zero voltage across

it. Once the STS was turned on, the grid current would be slowly ramped up to prevent any

voltage spikes caused by the grid inductance. The steps to perform this algorithm can be

summarized as follows:

13
 1. Detect that the grid is within nominal operating parameters.

2. Adjust the load voltage to match the magnitude and phase of the grid voltage.

3. Once the load voltage is equal to the grid voltage, turn on the STS and switch from

voltage-controlled mode to current-controlled mode, with the reference current being

equal to the load current.

4. Change the reference current slowly to the desired current (both magnitude and phase).

L. Yunwei, et al. [31] proposed a unified controller for use with DG system. By regulating

the output voltage, the proposed controller controlled power flow in the grid-connected mode of

operation, enabling the operation of DG when the system islands, and resynchronized the DG

with the utility before reconnecting them. The presented controller responded fast, allowing the

controlled DG to transit smoothly between the grid-connected and islanding modes without

disrupting critical loads connected to it. The resynchronization process when the utility grid

returned back to normal operating conditions form was achieved by aligning the voltage at the

DG and utility ends of a STS separation device. This process was implemented by adding two

synchronization compensators to the real and reactive power control loop, as shown in the

dashed frames of Fig. 2.1 and Fig. 2.2. To summarize, the final block diagram representation of

their proposed unified controller is shown in Fig. 2.3.

14
 Fig. 2. 1 Real power compensator with synchronization function [31]

Fig. 2. 2 Reactive power compensator with synchronization function [31]

Fig. 2. 3 Overview of proposed unified controller [31]

15
 T. Thacker, et al. [32-33] proposed a switched-mode control with detection and re-closure

algorithms scheme to regulate between the grid-connected and stand-alone modes of operation.

Their proposed system configuration can be seen in Fig. 2.4.

VSI PCC control

Va_grid
ia va
ib vb
Vdc
ic vc

PCC
Da Dc grid

load

ABC 1  2
2  PLL 1  2 PLL
DQ0
Vd Vo 
Id Vq Dd
DQ0
Vo_g
Vd_g
Vq_g

Iq control and Dq 2
Io detection Do ABC

PCC control

Fig. 2. 4 System configuration for simulated and experimental results [32-33]

Their proposed detection scheme, Fig. 2.5, used passive sensing parameters (over/under voltage

& frequency, real/reactive power deviation) with a BPF loop being fed directly to the duty cycles

instead of being fed to the current references. The advantage of BPF going directly to the duty-

cycles is that the current limiters can still be implemented without special considerations. When

an islanding event was detected, the system was automatically disconnected from the grid and

switched to voltage mode control.

16
 Measurements

V > 1.1 Vpu no

or Islanding? no
V < 0.88 Vpu

yes
no

ΔP > 20%

ΔQ > 20%

freq > 60.5

or
freq < 59.3

yes

Islanding? yes

Fig. 2. 5 Islanding Detection Flowchart [32-33]

17
Their proposed re-closing scheme is seen in Fig. 2.6. Like the detection algorithm, this algorithm

was autonomous and incorporated into the control of the DG. The idea was that by measuring

four key parameters: the grid voltage, VSI voltage, VSI frequency, and the line angle difference

between the grid and VSI, the system could safely reclose to the utility without having to de-

energize the DG. The scheme first measured and detected if the grid voltage had been recovered

from its fault and was back in nominal operating conditions. With the grid back to nominal

conditions, the signal coordinate transforms and detection started to use the grid‟s line angle as a

reference. This caused the VSI‟s PLL to start tracking the grid. Next, the magnitudes of the VSI

voltage and grid voltage were compared to ensure that they were on the same order of magnitude

before re-closing. The VSI frequency was then checked to make sure that the VSI had not left the

nominal frequency range. Finally, the system checked to see that the phase angle difference

between the VSI and grid were near zero. Once all of these conditions were satisfied,

concurrently, the DG would re-close with the grid.

18
 Measurements

yes Vg < 1.1 Vpu no

or Islanding? yes
Vg > 0.88 Vpu

no

V ≈ Vg

yes

freq < 60.5

or
freq > 59.3

yes VSI θ
≈
Grid θ

yes

Islanding? no

Fig. 2. 6 Re-closure Flowchart [32-33]

19
 The work presented in [34] by H. Zeineldin, et al., proposed a control strategy to implement

intentional islanding of inverter based DG. The proposed method was based on designing two

control algorithms, one for normal operation and the other for islanded operation. The DG

provided constant power and constant voltage during normal operation. A hybrid passive

islanding detection method was implemented to detect DG islanding. The detection algorithm

sent a signal to declare an islanded operation and the DG switched to the voltage control mode.

To overcome a large delay in detecting islanding for matching load condition cases, the Rate of

Change of Frequency (ROCOF) was used, in parallel with the OFP/UFP, as another measure to

detect islanding. The algorithm stated by monitoring both the frequency and voltage at the PCC

and calculating the ROCOF. The time at which the frequency exceeded its thresholds and the

time when ROCOF exceeded its threshold were denoted by T1 and T2 respectively and were

stored. The difference between the actual time T and T1 and T2 respectively were calculated. If

the difference exceeded a certain predetermined delay time, islanding was declared and the DG

was operated in the islanded mode. The islanding detection algorithm was responsible for

sending a signal that switched the inverter to the suitable interface control. Once islanding was

detected, the DG switched to the voltage control interface. To assure safe islanded operation of

the DG, the output active power of the DG was monitored to assure that it was less than the DG

capacity. If the load on the island was greater than the DG capacity, the DG would become

overloaded and a signal was sent to disconnect it. Fig. 2.7 and Fig. 2.8 show the flowcharts of

their designed control strategy.

20
 frequency and voltage
measurement

df/dt calculation

59.3 < f > 60.5

or no
0.88 > Vpu > 1.1

store T1

df/dt > 0.1

yes Store T2 no

yes

T-T1 > delay1

no
no
T-T2 > delay2

yes

yes

Islanding
detection

Fig. 2. 7 Proposed islanding detection algorithm [34]

21
 frequency and voltage
measurement

Islanding detection
algorithm

no Operate DG
Is the DG In
Islanded P-V mode

yes

Operate DG As
voltage
controlled VSI

DGactive power
> Prated
no

yes

Disconnect DG

Fig. 2. 8 Flowchart of the control strategy implemented to produce intentional islanding

[34]

22
 In [14], a control strategy for grid-connected and islanded operation was designed and

implemented by Y. Zhou, et al. Their proposed controller solved the following issues: how to

detect system separation fast and accurately, how to control DG in an islanded power system to

stabilize frequency and voltage, and how the controller of the DG can be transferred through grid

connected operation to islanding operation and vice versa seamlessly in order to protect the

critical load. Voltage angle difference between the local power system and the main power grid

was measured. If the angle difference was increased to an abnormal value, which cannot happen

under grid connected operation, then it was determined that the local power system was

disconnected from the main power grid. Before synchronization, voltage magnitudes and phase

angles of the islanded system at the point of common coupling (PCC) and grid were measured.

When the magnitude and angle differences between the islanded system and the main grid

approached zero the synchronization began and the hybrid power system was connected with the

power grid again.

G. Fang and M. R. Iravani [35] reported a control strategy that provided control over

frequency and voltage permitting the operation of the DG unit in both grid-connected and

autonomous modes. The features of the control strategy presented were: it provided smooth

transition capability from grid-connected mode to autonomous (islanded) mode, and it included

an islanding detection capability. Reactive power control (upper path of Fig. 2.9), which drove

the system to an unstable condition during an autonomous mode of operation, was exploited for

islanding detection. To decrease the detection time, positive feedback based on the instantaneous

PCC voltage, was added to the controller as shown in Fig. 2.10. This positive feedback ensured

that the feedback process was only effective during transients, did not impact the dc steady state

error signal, and provided attenuations for high-frequency noise.

23
 Fig. 2. 9 Block diagram of voltage control [35]

Fig. 2. 10 Block diagram of voltage control with positive feedback [35]

In [36], presented by G. Iwanski and W. Koczara, the loss of mains was detected using two

methods. The first method was based on the voltage amplitude calculation and the second

method, used for detection of grid voltage failures, was based on the monitoring of the frequency

of the stator voltage provided by the grid. The mains outage detection implicated soft

disconnection of the generator from grid by STS opening, and transition to standalone operation.

24
This way the selected load was supplied uninterruptedly. On the other hand, recovery of the grid

voltage caused that the stator voltage was synchronized with the grid and the generator was

softly connected to the grid by closing the STS. Their strategy for synchronization of the stator

and grid voltages was:

1. The angle between the stator voltage vector and the grid voltage vector was calculated.

2. This angle was reduced by iterative turning of the synchronously rotating frame in the

direction of the grid voltage vector.

3. The original transformation angle, calculated by integration of reference angular speed,

was decreased iteratively in every computation cycle and in each cycle a new angle was

determined.

4. When the stator voltage was synchronized with the power grid, two PLL structures

were operated simultaneously. During synchronization, the reference dq frame connected

with the stator voltage rotates with an angular speed Ω while the grid voltage vector

rotates synchronously with the reference synchronous speed of the dq frame ( *s ).

5. The relative speed and position of the grid voltage vector and the synchronously

rotating frame were decreased to zero.

6. Thus, the synchronously rotating frame was synchronized with the grid voltage,

whereas the stator voltage was synchronized with the rotating frame controller and

consequently provided stator voltage and grid voltage synchronization.

7. If the angle was close to zero, the STS was closed and the generator was connected to

the grid.

They proposed two methods for the main outage detection and soft disconnection. The first

method was based on the voltage amplitude calculation. If the power of load connected between

25
the stator and breaking point was much higher or much lower than the power delivered to the

grid, the amplitude was increased or decreased rapidly after line breaking. This way the loss of

main was quickly detected. The second method used for detection of grid voltage failures was

based on the monitoring of the frequency of the stator voltage provided by the grid. The output

frequency could not be controlled using a power control method which was applied during the

grid connection operation. The frequency was continuously increased or decreased in case of the

main outage.

The aim of the research presented by R. Majumder, et al. [22, 37] was to set up a power

electronics interfaced DG system. A seamless transfer between islanded and grid connected

mode was proposed that uses an online load flow study. During resynchronization, the reference

voltage of the DG was changed to achieve the same voltage magnitude and angle at PCC as the

grid voltage while simultaneously supplying the total power demand of the DG in desired ratio.

Once the breaker was closed, the DG started to supply its rated power and the rest of the power

requirement was supplied by the grid. Their proposed sequence of control from a grid connected

operation to islanded mode and then again back to grid connected can be summarize as:

1. Detect the islanding by zero current in the breaker connecting the grid with the DG.

2. If islanded, switch to voltage control mode.

3. Calculate the reference voltage required for the DG to maintain the load power

requirement.

4. Check for the clearing of the fault.

5. If resynchronization is desired, detect the grid voltage and frequency.

6. Calculate the reference voltages for the DG to make the PCC voltage the same as the

grid voltage while continuing to supply the load requirement.

26
 7. Calculate the change in angle required to match the phase for proper synchronization.

8. Change the voltage reference to the new calculated value.

9. Connect the breaker.

10. Switch to the state feedback control mode.

Aiming to solve the transition problem from grid-tied mode to off-grid mode, H. Shengli, et

al. [38] presented a control strategy based on a controlled voltage method whether connected to

the grid or stand-alone. In grid-tied mode, the inverter worked as a controlled voltage source by

controlling the phase and magnitude of the filter capacitor voltage, the injection current to the

grid was regulated indirectly, so the output active and reactive powers were also successfully

controlled. Once the grid failed, the inverter regulated its output voltage amplitude quickly, while

the voltage fluctuated within permissible levels during this period, the grid currents forced

through the STS decreased to zero at a highly accelerated rate. The controller cancelled all the

gate drive pulses of the STS and reduced the inverter output voltage lower than the grid voltage

simultaneously, which introduced a reverse voltage across the coupling inductor and forced the

grid current to decrease to zero quickly. As soon as the grid current reached zero, the STS was

turned off and the controller promptly regulated the inverter output voltage to its rated value.

After disconnection from the grid, the inverter recovered its voltage amplitude to a rated level.

Based on the SST forced turn-off strategy, the inverter operated as a controlled voltage source

whether grid-tied or not. By controlling the phase and magnitude of the filter capacitor voltage,

the grid current was regulated indirectly; accordingly the output active and reactive powers were

also successfully controlled.

C.-S. Wu, et al. [4] introduced a power supply system designed with grid-connected

operation and autonomous operation modes. Through a novel anti-island (AI) detection method,

27
the current status of the electricity network was detected quickly and the operation mode of the

inverter was ascertained. Based on analysis of transition between grid-connected operation and

autonomous operation modes, the output conditions of the inverter were adjusted optimally and a

seamless transition process was acquired. All transitions were conducted in zero-crossing in

order to reduce the voltage ripple of the sensitive load. Their seamless transition from

autonomous operation mode to grid-connected operation mode in detail is as follows:

1) Detects the voltage, frequency of the electricity network and judges the operation

status of the grid;

2) Adjusts the output voltage of the inverter, and makes it equal to those of the electricity

networks;

3) After adjusting, translates the operation mode of inverter (from voltage source to

current source) and closes the switch. In order to reduce the voltage ripple on the

sensitive load, the initial current is given as the current of the load;

4) Increases the output current step by step up to maximization.

Similar to the transition process described above, the seamless transition from grid-connected

operation mode to autonomous operation mode in detail is as followings:

1) Detect a fault in the electricity networks (by means of AI method introduced in next

chapter);

2) Set the output current to that of load in order to reduce the voltage ripple impact;

3) Open the switch and change the operation mode of inverter.

For the AI detection two positive feedback mechanisms were established. One was magnitude of

voltage feedback (Fig. 2.11); the other was frequency of voltage feedback (Fig. 2.12).

28
 Vd sense Id ref P V
Fig. 2. 11 Diagram of magnitude of voltage feedback in dq frame

ω sense Iq ref Q ω
Fig. 2. 12 Diagram of frequency feedback in dq frame

A novel scheme for seamless transfer of MTG (Microturbine Generation) system operation

between grid-connected and islanding modes and vice versa was proposed by D. N. Gaonkar, et

al., [39]. The presented scheme used the estimated phase angle error obtained by the PLL for

islanding detection and re-synchronizing DG to grid. A converter control strategy for both grid-

connected and islanding modes was presented. The proposed seamless transfer scheme consisted

of a passive islanding detection and re-closure method. The presented islanding detection method

used the phase angle estimated by the PLL to detect the islanding condition. Their algorithm

devised for the detection scheme is shown in Fig. 2.13. Their re-closure scheme continuously

monitored the phase angle and terminal voltage magnitude to determine whether or not the

disturbance in the grid was over. This was necessary in order to synchronize the MTG system

and to connect back to the grid without any down-time. The re-closure algorithm continuously

monitored the terminal voltage of the grid and MTG system. Both voltage magnitudes were

compared. Both voltages had to be approximately equal to avoid large transients during re-

connection to grid. Once the voltages were approximately equal, the algorithm compared the Δθ

29
value obtained from the PLL with the set threshold limit. As long as these minimum

requirements were met (voltage and phase angle), there were no major issues in reconnection of

islanded systems to the utility. The re-closure algorithm for connecting the MTG system to the

grid, when the utility recovers from the disturbance, is shown in Fig. 2.14.

measurement

Is
no no
Δθ > Δθmin
Islanding

yes

Transfer control to
islanding mode

Fig. 2. 13 Islanding-detection scheme [39]

30
 measurement

Is
V < 1.1 pu and no Islanding
V > 0.88 pu exist

yes

Is
no
Vdg ≈ Vgrid

yes

Is
Δθ > Δθmin no

yes

Transfer control to
islanding mode

Fig. 2. 14 Re-closure scheme [39]

31
 Synchronization
restoration

Islanding mode Grid-connected

operation mode

Grid failure
IBS disconnection

Fig. 2. 15 Flowchart diagram of the modes of the VSI [3]

E* ES

iPCC Qc + +
Q -
X LPF Gq(s)
+ +
-
Q*=0 GS
o
P/Q
+90 decoupling
VPCC E

iPCC P Pc + φ Vref
X LPF Gp(s) E sin(ωt-φ)
+- +
P* φs
X LPF PI(s) φs ZG θG
VG(rms) + ES
VG GS’ PI(s)
-
Phase synchronization VC(rms) GS’

amplitude synchronization
Fig. 2. 16 Block diagram of the whole proposed controller using the synchronization
control loops [3]

32
 The intent of the research that has been presented in [40] by C. Chien-Liang, et al., was to

introduce a practical inverter based DG system that ensured smooth mode transfer between

islanding and grid-tie modes while maintained accurate current sharing and high quality output

waveforms. The key features of the proposed DG system include: a single current loop controller

designed for a grid-tie operation to reduce the steady-state error while maintaining system

stability; for islanding operation, a dual-loop control system for outer voltage loop and a

controller for inner current loop control is proposed to limit peak current magnitude under

transient, enhance voltage loop stability, and reduce the voltage steady-state error; a phase

synchronization is implemented with PLL and an automatic phase adjustment to synchronize the

output currents among the inverters; and, a proper mode transfer procedure with smooth current

transition are suggested to minimize the excessive electrical stresses. Their proposed procedure

to change from grid-tie to islanding mode is summarized as follows.

1) The upper level controller detects the fault on the voltage grid and extracts the current

information.

2) Through a CAN bus, the upper level controller provides the current information and

commands the current-controlled inverters to change their outputs so that the current on

the STS can be minimized to avoid mode transfer transient.

3) The upper level controller provides the turn OFF signal for the STS after a certain

waiting time.

4) Through the CAN bus, the upper level controller commands a selected inverter to

change from current-controlled mode to voltage-controlled mode at the next zero

crossing.

33
 5) That inverter regulates the bus voltage to a desired level and provides the output

current information.

Their procedure to change from islanding to grid-tie mode is summarized as follows.

1) The upper level controller detects if the grid voltage recovers, and keeps detecting the

grid voltage magnitude and phase information.

2) Through the CAN bus, the upper level controller provides the grid voltage information

and commands the inverter to adjust its voltage to track the grid voltage in both

magnitude and phase.

3) The upper level controller keeps monitoring both voltages. It turns on the STS once the

two voltages are synchronized in phase and magnitude.

4) Through the CAN bus, the upper level controller commands the inverter to change its

controller from voltage controlled mode to current-controlled mode at the next zero

crossing. It assigns the current references so that zero current goes through the STS

during transfer transient.

5) The current controlled inverters change the current reference to a desired level.

In [41], presented by L. Qin, et al., the transition from grid-connected to standalone operation

using a Solid State Relay (SSR) was discussed. A feed-forward that could boost the dynamic

response and benefited the transition from grid-connected to stand-alone was added to the

controller. They proposed to force the grid current to be zero by using a voltage amplitude

regulation method. Then, the SSR was turned off and the reference output voltage was recovered

to the rated value. The principle of zero current regulation was to retain the current control mode

in transition but change the current reference to zero. After the current dropped to zero, the SSR

was turned off and the system was switched to voltage control.

34
 A seamless transfer of grid-interactive inverters between grid-connected and stand-alone

modes was proposed by Y. Zhilei, et al. [42]. In grid-tied mode, an output voltage controller was

used for compensating the filter capacitor current, and a grid current controller was used to

control the grid current. In stand-alone mode, the output voltage controller was used to regulate

the output voltage and to set the output of the grid current controller to zero. Therefore, the

transfer between the two controllers did not exist with their proposed method. In order to realize

a seamless transfer between grid-tied and off-grid modes, they proposed a controller that

matched the magnitude, frequency, and phase of the grid voltage before connecting to the utility.

The detailed process of their seamless transfer between the two modes is as follows:

1) Off-grid mode to grid-tied mode:

a) Detect that the grid is operating in nominal condition.

b) Adjust the reference voltage to match the frequency and phase of the grid

voltage.

c) Change the voltage reference to the grid voltage at the positive zero-crossing of

the output voltage, which makes the output voltage equal to the grid voltage, even

in polluted grid voltage.

d) Once the load voltage is equal to the grid voltage, the STS is turned on.

e) Increase the reference grid current slowly from zero to the desired value (both

magnitude and phase).

2) Grid-tied mode to off-grid mode:

a) Detect a fault on the utility.

b) Decrease the reference current to zero at zero crossing of the grid current.

35
 c) Change the STS voltage from high level to low level at zero crossing of the

grid voltage. This will turned off the STS.

d) Synchronize the reference voltage with the grid voltage.

e) Change the reference voltage at zero-crossing of the grid voltage.

From this algorithm, the transfer between both modes was only the change of the reference

voltage between the two modes. The transfer between the output voltage controller and the grid

current controller did not exist in their proposed method.

36
CHAPTER 3. SYSTEM DESCRIPTION

3.1. Circuit Topology

Fig. 3.1 shows the proposed main circuit topology for grid-connected and stand-alone

operations voltage source inverter. The system consists of a DG unit that is modeled by a DC

source, a sinusoidal pulse-width-modulated (SPWM) voltage source inverter, an LCL filter to

achieve attenuation of the switching frequency ripple in the output voltage, a load, and the

controller needed to implement grid connected and stand-alone operations.

Voltage Source Inverter Utility

PCC Disconnect
Sa Sb Sc LI Lg Ia
VIa Switch VGa
a
Vdc b
c Grid
Sa’ Sb’ Sc’
Cf
LCL Filter Load
VIa VIb VIc Ia Ib Ic VI-ac VI-bc VI-ca VG-ab VG-bc VG-ca

Sa-a’

CONTROLLER Sb-b’
Sc-c’

Fig. 3. 1 Schematic diagram of the grid-connected inverter system

The voltage source inverter performs the interface function between the DC bus and the three

phase AC world. The inverter consists of power electronics devices actuated by gate pulses:

37
these devices are high frequency static switches that periodically apply repetitive pulses with

different widths and polarities at the AC terminals, with the goal to create an AC waveform of

desired magnitude and phase.

The controller includes control for grid-connected and stand-alone operations, PLL,

coordinate transforms, parallel interconnection to the grid (represented by ideal, sinusoidal

voltage sources behind line impedances) through the Point of Common Coupling (PCC)

breakers. The loss of main detection, load shedding algorithm, and re-connection to the grid are

also included.

3.2. Voltage Source Inverter

A typical DG system based on micro-source is depicted in Fig. 3.2. There are two basic

elements: the part where the electricity is generated (micro-source) and the part that performs

the interface with the rest of the system (voltage source inverter).

Three
Micro- Phase
DC
Output
source Link

Voltage Source
Inverter

Fig. 3. 2 Typical diagram of micro-source generation system

38
 All the power, DC or non-power frequency AC, generated by each DG must use an

electronic inverter (to convert DC voltage to a three-phase AC voltage with desired magnitude

and phase) to interface with the electrical power system. This power electronic interface provides

significant flexibility and permits the DG to function as a semi-autonomous power system [31].

There are two main requirements to convert DC voltage to a three-phase AC voltage: a

matrix of switches and a switching sequence. Fig. 3.3 shows the circuit that allows synthesizing

the AC voltage at the three-phase terminals starting from a DC voltage. From left to right, there

is the DC bus and then the matrix of six power electronics devices. Each device has bi-

directional current flow capability. The flow of current out of the DC bus is regulated by signals

sent to the gates to each of the devices present in the inverter.

Fig. 3. 3 DC to Three-Phase AC Inverter Diagram

The operation of this bridge requires two restrictions. First, two devices on the same leg may

never be conducting. Second, at least one device per leg must be conducting. The first constraint

39
comes from the fact that if for example, devices numbered 1 and 2 are conducting at the same

time then the DC source is shorted, resulting in very high currents that will destroy both devices.

The second constraint stems from the fact that the current flowing in each of the phases on the

AC side must have a path where to flow: if for instance neither device numbered 1 and 2 is

closed, then the current flowing on the AC side in the „a‟ phase would have no path where to

flow.

Consider the switching sequence shown in Fig. 3.4, where devices 1, 4, and 5 are closed: this

configuration satisfies both the above constraints. The voltages at the AC terminals are:

V  V (3.1)
ab dc

V  V (3.2)
bc dc

V 0 (3.3)
ca

From Eq. 3.1 to Eq. 3.3 it can be noticed that the output voltages can be either V , V
dc dc

or zero. By carefully selecting a switching sequence it is possible to make those three levels of

voltage appear in one particular phase. Furthermore, it is possible to create a three phase system

of sinusoidal voltages with desired amplitude and phase. This is called a six pulse operation of

the bridge. In reality the output waveform of the AC voltage is a squared wave with a particular

amplitude and phase at the fundamental frequency. In order to solve this problem, more complex

switching sequences with higher switching frequencies are needed.

One of these complex switching sequences is the pulse width modulation (PWM) technique.

This PWM technique works on an averaging effect using the switching frequency as a base

period where the averaging occurs. Fig. 3.4 shows how the basic principle of the PWM technique

works: it allows the instantaneous average output to be held closer to the desired fundamental

40
output. The voltage waveform is made of pulses of different widths: hence the name pulse width

modulation. The sinusoidal signal is compared with a sawtooth waveform (carrier). When the

carrier is less than the sinusoidal signal, the PWM signal will be in high state (1). Otherwise it

will be in the low state (0).

Fig. 3. 4 PWM Operation of the Bridge

(For interpretation of the references to color in this and all other figures, the reader is
referred to the electronic version of this dissertation)

3.3. LCL filter

For attenuating the current ripple produced at the output of the inverter, an LCL filter is

placed in between inverter and grid. An LCL filter can achieve reduced levels of harmonic

41
distortion with lower switching frequencies and with less overall stored energy [43]. Fig. 3.5

shows the LCL filter where L and L are inductors on inverter side and grid side respectively.
i g

LCL Filter
Inverter Li Lg
+ + + +
Vdc Vi Cf Vc Vg
_ _ _
_

Fig. 3. 5 LCL Filter configuration circuit

Z = Ls (3.4)
i i

Z =L s (3.5)
g g

1
Z = (3.6)
c sC
f

From the economical point of view, the design of the LCL filter is made in such a way that

the magnitude of the inductor should be minimum and the magnitude of the capacitor should be

maximum. By selecting large capacitor the filter volume is reduced [44-45].

3.3.1. Specifications for the design of LCL filter

The following are the parameters for the design of the LCL filter:

Output Power of inverter: P  10 kW

i  3ø

42
 Apparent Power of inverter: S  10 kVA
i

Output Voltage of Inverter: V  208 V

i ll

Switching frequency: f  10 kHz

s

V  400 V
dc
Input DC voltage:
(V  V / 0.612  208 / 0.612  339.869 V )
dc ll  rms

Grid Voltage: V  208 V

g ll

P
ˆ load  1 3.33kW
Inverter current: I    27.5 A  28 A
L V 120V
load  1

Inverter ripple current:

L  
Iˆ  10%  (0.10) Iˆ  (0.10)(28)  2.8 A
L

Reactive Power storage at C : Q  5%

f

3.3.2. Design of L
I

The ripple current depends on the DC link voltage, inductance, and the switching frequency.

The DC link voltage and switching frequency are constant, thus the inductance can be calculated

using the following procedure:

Iˆ
V L L (3.7)
L ˆT
s

where ˆ is the phase voltage duty cycle at maximum output and is given by:

ˆ 
2V


g  pk  1 2 208 2
 0.85
 (3.8)
V 3(400)
dc

43
 ˆ
ˆ ˆ
V 1  Vdc
I   L  (3.9)
L Lf 6 Lf
s s

 V 
L  (0.1416)  dc  (3.10)
 Iˆ f 
 L s 

 200 
L  (0.1416)    1mH (3.11)
 (2.8)(10 x103 ) 
 

3.3.3. Design of C f

Considering the capacitor stores 5% of the system power as reactive power, maximum

capacitance is:

Q  5%S (3. 12)

Q  5%S  (0.05)(10kVA)  500VA (3.13)

3 

Q
C  (3.14)
f
3 V 2 
i

Q 500
C    31 F (3.15)
f
3  V 2   (3)(1202 )(2 60)
i

3.3.4. Design of L
g

As most of the ripple is filtered by L & C , grid side inductor value is taken as half the
I f

value of L .
I

L  0.5L  L  0.5 1mH   0.5mH (3.16)

g I g

44
3.4. Simulation and Experimental Set-ups

3.4.1. Simulation Set-up

The performance of the proposed control strategies will be evaluated by computer simulation

using SABER. Fig. 3.6 shows the simulated system. This system will be tested under the

following conditions:

 Switching frequency, f , 10 kHz

s

 Filter inductor, L , 1 mH
I

 Filter inductor, L , 0.5 mH

g

 Filter capacitor, C , 31  F
f

 Output voltage 208 V , 3 @ 60 Hz grid connection, with V  400 V

rms  ll dc

 Output capacity 10 kW from DG

 Total load 10 kW

The load will be adjusted to consume 10 kW. The DG system will be simulated to supply 10

kW and zero reactive power. The system will be operated initially in grid-connected operation.

45
 PCC Utility
Voltage Source Inverter
Disconnect
Sa LI Lg Ia VIa Switch VGa
a
Vdc b
c
DG unit Sa’ Grid
Cf
LCL Filter Load
θ
Mode of Operation new
Controller for Grid-connected Operation grid-
Switch Da S Sa-a’
IDref + connected DQ
P Sb-b’
current + Db
- regulator + +
+ W S
ABC Dc
IQref - + M c-c’
ID IQ VD VQ Ia Ib Ic VIa VIb VIc
Controller for intentional islanding operation intentional
VDref + IDref + + islanding ABC θnew ABC
voltage current
- - - + DQ 
+ regulator + regulator +
- + V
VQref - IQref -
ID IQ θnewV
K PLL
VQ
VD K VD VQ f θold
VD > 1.1 ID IQ VD VQ
or VI-acVI-bcVI-caVG-abVG-bcVG-ca
VDpu VD < 0.88
Load shedding Re-synchronization
f > 60.5 Switch Mode θold
algorithm
f or of Operation θnew
f < 59.3
Loss of Main Detection

Fig. 3. 6 Simulated system

3.4.2. Experimental Set-up

The hardware prototype of Fig. 3.6 will be implemented for experimental verification. The

control, PLL, grid condition detection, and re-closure algorithms will be programmed using a

46
universal DSP control board developed at the Power Electronics and Motor Drives Laboratory at

Michigan State University. The system will be tested under the following conditions to

experimentally verify the simulation results:

 Switching frequency, f , 10 kHz

s

 Dead time 3μs

 Filter inductor, L , 1 mH
I

 Filter inductor, L , 0.5 mH

g

 Filter capacitor, C , 50  F
f

 Simulated Output voltage 104 V , 3 @ 60 Hz grid connection, with

rms  ll

V  200 V
dc

 Output capacity 2.5 kW from DG

 Total load 2.5 kW

The reason for simulating the output voltage is to ensure the algorithms and controllers are

functioning properly under low-power tests; such that there is a reduced risk of operator and

equipment damage if the system fails.

Shown in Fig. 3.7 are the inverter, the DSP board, the filter, and the rectifier.

47
 Fig. 3. 7 Experimental set-up

The DG will be started up in grid-connected operation mode, and then the separation device

will be opened. When the DG is disconnected from the grid it operates in stand-alone mode.

48
CHAPTER 4. PROPOSED CONTROL FOR GRID-CONNECTED OPERATION OF DG

4.1. Introduction

There are two modes of operation for the DG inverter – the grid-connected mode and the

stand-alone mode. Each mode has its own control requirements and control structure. In the grid-

connected mode of operation, the DG is connected to the utility. The utility, which is assumed to

be stiff, sets the voltage at the terminal of the DG inverter [10]. The inverter controls the power

being injected into the grid by controlling the injected current. Thus, in this mode, the inverter

operates in the current control mode.

In the stand-alone mode, the inverter supplies power to the load. It has to maintain the

voltage at the terminals of the load, irrespective of any changes in the load. Thus, in this mode,

the inverter operates in the voltage controlled mode.

This chapter discusses the grid-connected mode of operation and its controller design, while

the stand-alone mode will be discuss in the next chapter.

4.2. Circuit Topology

Fig. 4.1 shows the main circuit topology. This system consists of the micro-source

represented by the DC source, the conversion unit which performs the interface function between

the DC bus and the three phase AC world, and the LCL filter that attenuates the current ripple

produced at the output of the inverter. The controller presented provides constant DG output and

maintains the voltage at the PCC before and after the grid is disconnected.

49
 Voltage Source Inverter Utility
PCC Disconnect
Sa Sb Sc LI Lg Ia
VIa Switch VGa
a
Vdc b
c Grid
Sa’ Sb’ Sc’
Cf
LCL Filter Load
VIa VIb VIc Ia Ib Ic VI-ac VI-bc VI-ca VG-ab VG-bc VG-ca

Sa-a’

CONTROLLER Sb-b’
Sc-c’

Fig. 4. 1 Schematic diagram of the grid-connected inverter system

4.3. Controller

For grid-connected operation, the controller shown in Fig. 4.1 is designed to supply constant

current output in order to provide a pre-set power to the main grid [46]. An important aspect to

consider in grid-connected operation is the synchronization with the grid voltage [47]. For unity

power factor operation, it is essential that the grid current reference signal be in phase with the

grid voltage. This grid synchronization can be carried out by using a PLL [48-49]. Also, the PLL

is used to determine the frequency and angle reference of the PCC [3]. Fig. 4.2 shows the control

topology used.

50
 VIa
VIb
VIc
ABC

V V
PLL
new
old
Ia
Ib
Ic

ABC Synchronization
DQ new VD controller

VQ
new
ID

IQ

- + Sa-a’
Current + + DQ Da SPWM
IDref+ - Regulator Sb-b’
+ Db
IQref + ABC Sc-c’
Dc
Fig. 4. 2 Block diagram of the current controller for grid-connected

When using the current control, the output current from the filter, which has been

transformed into a synchronous frame by Park‟s transformation (Eq. 4.1) and regulated in DC-

quantity, is fed back and compared with reference currents I . This generates a current
DQref

error that is passed to the current regulator (PI controller) to generate voltage references for the

inverter. In order to get a good dynamic response V is fed forward. This is done because the
DQ

terminal voltage of the inverter is treated as a disturbance and the feed-forward is used to

compensate for it. The voltage references in DC-quantities, V , are transformed into a
DQref

51
stationary frame by the inverse of Park‟s transformation (Eq. 4.2) and utilized as command

voltages for generating high frequency pulse width modulated (PWM) voltages.

X   
 D   cos   cos(  2 / 3)  cos(  2 / 3)   X a 
 X   2  sin  sin(  2 / 3) sin(  2 / 3)   X  (4.1)
 Q 3  b
   1/ 2 1/ 2 1/ 2   X 
 X 0   c

where θ = ωt and ω is the frequency of the electric system.

X   
 a   cos  sin  1/ 2   X D 
 X     cos(  2 / 3) sin(  2 / 3) 1/ 2   X  (4.2)
 b   Q
X     cos(  2  / 3) sin(  2  / 3) 1/ 2   
 b  X 0 

4.3.1. Synchronization Controller for Grid Reconnection: Proposed Algorithm

When the DG is in islanded mode operation and the grid-disconnection cause disappears, the

transition from stand-alone to grid-connected mode can be started [42]. To avoid hard transients

in the reconnection, the DG has to be synchronized with the grid voltage [50-52]. The DG is

operated in synchronous island mode until both systems are synchronized. Once the voltage in

the DG is synchronized with the utility voltage, the DG is reconnected to the grid and the

controller will pass from voltage control mode to current control mode. This synchronization is

achieved by implementing the following algorithm:

o Assume that the phase difference between grid voltage and inverter voltage is given by:

  V  V (4.3)
G I

o In order to obtain information of  , two sets of voltage values are used:

3
k  V V  V V  V V   cos( ) (4.4)
Ia Ga Ib Gb Ic Gc 2

52
 3
g  V V  V V  V V   cos( )  3 sin( )  (4.5)
Ia Gb Ib Gc Ic Ga 4

where,

V V sin(t )
Ga Gm

V V sin(t  120o )
Gb Gm

V V sin(t  120o )
Gc Gm

V  V sin(t   )
Ia Im

V  V sin(t  120o   )
Ib Im

V  V sin(t  120o   )
Ic Im

Using the variables k and g, sin( ) can be found as:

4 2
g k
sin( )  3 3 (4.6)
3

Fig. 4.3 shows how sin( ) is used to obtain the compensated phase angle for which the grid

voltage and the DG voltage are synchronized. When the grid is recovered, a PLL observer,

whose output will generate the grid frequency and phase, processes the grid phase voltages. If

the amplitude and frequency of the voltages are within the limits defined by standards, a

synchronizing signal will be generated. When the sine of the phase error is between -0.04 and

0.04 radians, the voltages are considered to be synchronized and the control of the supply-side

inverter will be switched from stand-alone to grid-connected control mode.

53
 sync
sin    +
+
comp Voltage
+_ PI X 1/s PLL Controller

-0.04  sin  0.04 Synchronization Grid

condition
sin desired 
signal

Grid
recovered

Fig. 4. 3 Synchronization controller

4.4. Transfer Functions

Fig. 4.4 shows the block diagram of the DG interface control for grid-connected operation.

D Vd
F(s) Z(s)
+ Vin Id Vd
+ e K + K I1 + Vin 1 Filter Load
Id-ref _ P1
s V*
Id

Fig. 4. 4 Block diagram of the current controlled inverter

The PI controller produces a signal that is proportional to the time integral of the controller.

The transfer function of the PI controller is given by:

K
DK  I1 (4.7)
P1 s

where K is the proportional gain and K the integral gain.

P I

The inverter stage does not have any significant transient time associated with it and hence it

is modeled as an ideal gain. This ideal gain can be given by G  s   1 .

I

54
 The schematic circuit of the filter stage is shown in Fig. 4.5. It consists of an LCL filter and a

parallel RLC load.

L1 Vx L2 Id
+
Vin Vd
Cf
-
RL LL CL
Filter: F(s)
Load: Z(s)

Fig. 4. 5 LCL Filter and parallel RLC load

The transfer function of the filter is given by:

I (s) 1
F ( s)  d 
 
(4.8)
V ( s) s3C L L  s 2C L Z  s L  L  Z
in f 1 2 f 1 1 2

The transfer function of the load, Z(s), is given by:

V (s) 1
Z (s)  d  (4.9)
I (s) 1 1
d   sC
R sL L
L L

This can be re-expressed as:

sL R
Z ( s)  L L (4.10)
2
s R C L  sL  R
L L L L L

Using Fig. 4.4 and equations (4.7), (4.8) and (4.10), the transfer function of the current controlled

system is derived as:

55
 V ( s) FZD
H ( s)  d  (4.11)
I ( s) 1  FZ  FD 
d - ref

651.466 s
Z (s)  (4.12)
s 2  150.454 s  142117

F (s) 

6.4516 x1010 s 2  150.454s  142117  (4.13)
s5  150.454s 4  9.8219 x107 s3  1.456 x1010 s 2  5.5783x1013 s  0.4187

25
D  0.8  (4.14)
s

3.365 x1013 s 2  1.051s

H (s)  (4.15)
s6  150.108s5  9.83x107 s 4  6.62 x1010 s3  2.31x1013 s 2  7.58 x1015 s  2.28x1017

where R  4.33 , L  4.584mH , C  1.535mF , L  1mH , L  0.5mH , C  31 F ,

L L L 1 2 f

K  0.8 , K  25
P1 I1

4.5. Simulation Results

The performance of the proposed control strategies was evaluated by computer simulation

using SABER. Fig. 4.6 shows the simulated system.

56
 Voltage Source Inverter Utility
PCC Disconnect
Sa Sb Sc LI Lg Ia Switch
VIa VGa
a
Vdc b
DG unit c Grid
Sa’ Sb’ Sc’
Cf
LCL Filter Load

θnew
Mode of Operation Switch
Controller for Grid-connected Operation
IDref + + Grid-connected Da Sa-a’
current DQ
- + ++ Db Sb-b’
regulator SPWM
IQref - intentional ABC Dc Sc-c’
+
islanding

ID IQ VD VQ

Ia Ib Ic VIa VIb VIc

Voltage controller for

θnew stand alone operation ABC
ABC
DQ

VI-ac VI-bc VI-ca VG-ab VG-bc VG-ca
ID IQ V V

Re-synchronization PLL
θnew
θold
θnew
VD VQ f θold

Fig. 4. 6 Current controller for grid-connected operation

The R load was adjusted to consume 10 kW. The DG system was designed to supply 10 kW

and zero reactive power. The system was operated initially in grid-connected operation. Fig. 4.7

shows the voltages and currents at the PCC during grid-connected operation.

57
 V:t(s)
va
200.0
vb
0.0
V

vc
-200.0
A:t(s)
50.0 ia
ib
0.0
A

ic
-50.0 V:t(s)
grid_condition
4.0
V

2.0
0.0
0.0 25.0m 75.0m 0.125 0.175 0.225 0.275 0.325
t(s)

Fig. 4. 7 Phase voltages and currents during grid-connected operation

When the grid-disconnection cause disappears, the transition from stand-alone mode to grid-

connected mode can be started. The DG was operated in synchronous island mode until both

systems were re-synchronized. While in synchronous island mode, the synchronization controller

decreases or decreases the frequency to a limited value, as seen in Fig. 4.8. Also seen in that

figure are the voltages of the DG and grid; here the DG voltage can be seen to synchronize with

the grid and when the phase angle between the two voltages are approximately equal, the

algorithm reconnects the system to the utility and switches the mode of control from voltage

control to current control. As can be seen, the proposed algorithm successfully forces the voltage

at the DG to track the voltage at the grid.

58
 (V) : t(s)
Vg_a
200.0 Vg_a Va
Va
0.0
(V)

-200.0

6.0 (V) : t(s)

4.0 Grid-connected grid_condition

(V)

2.0
Intentional-islanding
0.0
0.6 0.61 0.62 0.63 0.64 0.65 0.66 0.67 0.68 0.69 0.7
t(s)

Fig. 4. 8 Synchronization for grid re-connection

Once the synchronization was completed, the DG was reconnected to the grid and the

controller was switched from voltage control mode to current control mode. Fig. 4.9 shows the

phase voltage V without and with the synchronization algorithm implemented. As can be
a

noticed, the transients are minimal and virtually negligible indicating that the algorithm avoids a

hard transient in the reconnection from stand-alone operation to grid-connected operation.

59
 (V) : t(s)
Va(w/o sync.)
200.0
0.0
(V)

-200.0
(V) : t(s)
200.0 Va(w sync.)

0.0
(V)

-200.0
0.50 0.52 0.54 0.56 0.58 0.60 0.62 0.64 0.66 0.68
t(s)

Fig. 4. 9 Phase voltage without (top) and with (bottom) synchronization

4.6. Experimental Results

A hardware prototype has been implemented for experimental verification. The control, PLL,

grid condition detection, and re-closure algorithms have been programmed using a universal

DSP control board developed at the Power Electronics and Motor Drives Laboratory at Michigan

State University. The DG is started up in grid-connected operation mode. Fig. 4.10 demonstrates

how the system line to line voltage and phase current behaves during grid-connected mode.

60
 Fig. 4. 10 Line to line voltages and phase current during grid-connected operation

Fig. 4.11 shows the process of synchronization where the line to line voltage at both ends of

the separation device is illustrated. At the beginning of the synchronization, both voltages are out

of phase. As can be seen, the proposed algorithm successfully forces the voltage at the DG to

track the voltage at the grid until the synchronization process is completed. Also shown is the

smooth transition of the currents.

61
Fig. 4. 11 Transition from stand-alone to grid connected operation

62
CHAPTER 5. PROPOSED CONTROL FOR INTENTIONAL ISLANDING OPERATION
OF DG WITH SEAMLESS TRANSITION FROM GRID CONNECTED OPERATION

5.1. Introduction

The control of the DG system is important in both grid-connected and stand-alone modes and

the system stability becomes very crucial during the transfer between these two modes. If the

system does not have a proper transfer procedure, severe transient voltages or currents will

occur, which may damage the entire system [40]. Fig. 5.1 shows a waveform showing this

situation after grid disconnection at time t . A seamless transfer can ensure smooth operation
1

and quick attainment of steady state.

Fig. 5. 1 Simulation waveforms of load voltages and grid currents during the grid
disconnection [38]

63
 In order to solve this transition problem, this chapter presents the development and test of a

control strategy for DG capable of working in intentional islanding connection mode with a

seamless transition from grid-connected to stand-alone operation modes. The control scheme

proposed is based on a voltage-controlled method for the stand-alone DG inverter. The stand-

alone mode with voltage control is featured with a grid condition detection algorithm to detect

the instant at which the DG is cut from the main grid, a load shedding algorithm to disconnect

part of the load in order to steer the power system from potential dangers, and a voltage

controller with seamless transition from grid-connected to stand-alone operation modes.

5.2. Grid Condition Detection

The instant at which the DG is cut off from the main grid must be detected in order for the

system to change between grid-connected to stand-alone modes [53]. This detection is achieved

by using a PLL which consists of Clarke‟s transformation (Eq. 5.1), Park‟s transformation (Eq.

5.2), a PI regulator, and an integrator [5, 54]. The schematic of the PLL is illustrated in Fig. 5.2.

V   2 3 1 3  V 
    ab  (5.1)
V   0 1 3  V 
   bc 

V    cos  sin   V 

 d  
cos   V 
(5.2)
Vq   sin 
   

64
 '
Vq ω ω' '
Va b c
abc  K p+ k I / s 1/s
+
+ Test switch mode
 dq0 ωs e t
Vα β 1/ (2) voltage of operation
f ' and
frequency
Vd = E '

Fig. 5. 2 PLL structure

The lock is realized by setting V to zero. A PI regulator can be used to control this variable
q

and the output of this regulator is the grid frequency [55]. As can be noticed in Fig. 5.1, in

addition to the frequency, the PLL is capable of tracking the magnitude of its input signals ( V
d

and V ), e.g. the grid voltages. These two parameters, frequency or voltage magnitude, are used
q

in the loss of main detection algorithm to detect the grid condition [28]. The algorithm sends a

signal that switches the inverter to the suitable interface control.

Fig. 5.3 shows the implementation of this algorithm. The output of a comparator block

activates a monostable circuit. If the monostable circuit stays in on-state for 0.16 seconds,

corresponding to 10 cycles, this is considered as an abnormal condition [28]. This abnormal

condition sends a signal that switches the inverter to the suitable interface control.

65
 Loss of Main Detection
Vdp.u. > 1.1
YES
or
Vd Vdp.u. < 0.88
monostable
f > 60.5 Switch
or YES Mode of
f f < 59.3 Operation

Fig. 5. 3 Loss of Main Detection

This switching between grid-connection and stand-alone operation modes is done by using an

over/under frequency method (OFP/UFP) or an over/under voltage method (OVP/UVP) that

cause the DG inverter to switch if the frequency or amplitude of the voltage stays outside of

prescribed limits. When the grid is disconnected, this causes that the voltage amplitude or

frequency changes, as shown in Fig. 5.4. When these changes in voltage or frequency exceed

certain thresholds, grid disconnection is considered and the operation mode is consequently

switched. The algorithm is shown in Fig. 5.5.

66
 Fig. 5. 4 Voltage or frequency change at grid disconnection

voltage or frequency
measurement

Vdpu < 0.88 f < 59.3

or or
Vdpu > 1.1 NO f > 60.5

Intentional islanding
operation
YES YES

Fig. 5. 5 Islanding Detection Algorithm

67
 While serving as good indications for islanding detection, the quick voltage or frequency

variations lead to a serious concern: the DG would operate out of the allowable voltage or

frequency range quickly after islanding occurs [56]. To avoid this, intelligent load shedding

algorithms need to be implemented in a DG system to make sure that the demand is within

available generation by disconnecting some least important loads [57].

5.3. Intelligent Load Shedding

Load shedding is defined as the process in which a part of the system load is disconnected

according to a certain priority in order to steer the power system from potential dangers [58-59].

During the grid connected operation, the DG is operated to provide the optimum power to the

grid according to many factors such as the availability of energy, energy cost, and so on [39].

The main grid is supplying or absorbing the power difference between the DG and local load

demand. When the main power grid is out (power outage), the DG that continues to inject pre-

determined optimum power can cause voltage and frequency transients depending on the degree

of power difference. The power difference makes the voltage and frequency drift away from the

nominal values [37]. When the voltage and frequency drifts have reached certain levels, it is

deemed that an islanding is occurring. This methodology is enough for islanding detection.

However, it is not enough for intentional islanding operation because often the local DG is either

less or greater than local load demand, and intelligent load shedding is needed. Therefore, it is

essential to have an analytical solution of the voltage and frequency transients locally for the DG

to have information and make decisions, and for intelligent load shedding to secure energy

delivery to sensitive loads.

68
 To develop the load shedding algorithm, a constant impedance load is used. Fig. 5.6 shows

the theoretical voltage transients for a constant impedance load under various active power

differences (from -50% to +50%) after main power outage, while Fig. 5.7 shows the theoretical

frequency transients under various reactive power differences. As seen from Figs. 5.6 and 5.7,

with no load shedding it would be insufficient for keeping the voltage or frequency within the

limits required.

Fig. 5. 6 Voltage Transients under various Active Power Differences

69
 Fig. 5. 7 Frequency Transients under various Reactive Power

The challenge is how to switch the DG inverter system to voltage control mode and bring the

voltage back to the normal range (0.88 – 1.1 Vpu) for intentional islanding operation. The

analytical solution of the simple case scenario shown in Fig. 5.6 provides a possible solution to

this challenge. Fig. 5.6 shows that the voltage change rate is related to the power differences

between the DG and load demand. The approach proposed is to detect the voltage change rate

and profile after the power outage and determine how much load shedding is needed before

going to the intentional islanding operation and switching to the voltage control mode. In order

to accomplish this, the system shown in Fig. 5.8 has been analyzed.

70
 S1
S2

Id R1 R2 Vg

Fig. 5. 8 System to implement load shedding

To determine the amount of load to be disconnected the following algorithm, based on Fig.

5.9, is proposed:

S2
R1 R2 Rpu=R1//R2
Id Idpu

Fig. 5. 9 System in per unit to implement load shedding

o Obtain the voltage amplitude expression for load shedding:

Using the circuit shown in Fig. 5.9, the expressions for the load voltages V ,V , and V
apu bpu cpu

can be found:

V I R sin(t   ) (5.3)
apu dpu pu

 2 
V I R sin  t     (5.4)
bpu dpu pu  3 

71
  2 
V I R sin  t     (5.5)
cpu dpu pu  3 

Using V ,V , and V an expression for the voltage amplitude can be found,

apu bpu cpu

2
V  V 2 V 2 V 2 (5.6)
pk 3 ab bc ca

2    
V  3I R cos2  t      cos2 t     sin 2  t     (5.7)
pk 3 dpu pu  3  6

   
V  2I R cos2  t      cos2 t     sin 2  t     (5.8)
pk dpu pu  3  6

o Derive the slope of the voltage amplitude

d  V (t )   2I R K
s 
pk  dpu pu
 (5.9)
dt    
cos2  t      cos2 t     sin 2  t    
 3  6

o Derive I
dpu

   
 s 2 cos 2  t      cos 2 t     sin 2  t    
 3  6
I  (5.10)
dpu
2 R K
pu

Using

R R / /R (5.11)
pu 1 pu 2 pu

72
and solving for R , where R represents the load to be shed,
1 1

 2 I K
dpu
R  (5.12)
1 pu    
s cos 2  t      cos 2 t     sin 2  t      2 I K
 3  6 dpu

where

       
K  sin  t     cos  t      sin  t     cos  t      sin t    cos t   
 3  3  6  6

(5.13)

5.4. Transition from Grid-connected to Stand-alone: Proposed Controller

The proposed voltage closed loop control for stand-alone operation with seamless transition

is shown in Fig. 5.10. The control works as voltage regulation through current compensation.

The controller uses voltage compensators to generate current references for the current

regulation.

73
 VIb
VIa

VIc
ABC

V V
PLL
new
old
Synchronization
VD

controller new
VQ

new
+ - IDref DQ
Voltage + Current _+ Da Sa-a’
VDref - - + SPWM
Regulator Regulator ++
VQref + IQref - ABC Db Sb-b’
K _ +
Dc Sc-c’
K
IQ
ID

VQ
DQ VD
new ABC
Load Shedding
Ib
Ia
Ic

Algorithm Rload shedding

Fig. 5. 10 Voltage controlled inverter

As shown, the load voltage, V and V , are forced to track their references ( V and
D Q Dref

V ) by using a voltage regulator (PI compensator). The outputs of this compensator, I

Qref Dref

and I , are compared with the load current ( I and I ), and the error is fed to a current
Qref D Q

regulator (PI compensator). The output of the current regulator acts as the voltage reference

signal that is fed to the sinusoidal pulse-width modulator (SPWM) to generate the high frequency

gating signals for driving the three-phase voltage source inverter. The current loop is included to

stabilize the system and to improve the system dynamic response by rapidly compensating for

near-future variations in the load voltages [11]. In order to get a good dynamic response V is
DQ

74
fed forward. This is done because the terminal voltage of the inverter is treated as a disturbance

and the feed-forward is used to compensate for it [60].

A compensation coefficient, K, is added to the current control loop as a feedforward control

unit. This feedforward control achieves accurate tracking of the sinusoidal reference. The

compensation coefficient processes the magnitude of the input voltage of the PWM modulator

and counterattacks the severe transitions influenced by the grid disconnection. Then, this input

voltage is fed to the PWM modulator to produce the drive signals for the inverter switches.

Therefore, the modulated wave does not include the transient components that resulted from the

grid disconnection.

5.5. Transfer Functions

Fig. 5.11 shows the block diagram of the DG interface control for intentional islanding

operation.

E D Vd
F(s) Z(s)
Vd-ref + + Vin Id Vd
e K I1 + b K + K I2 + Vin 1 Filter Load
K P1+ Id-ref _ s V*
_
P2
s
Vd Id

Fig. 5. 11 Block diagram of the voltage controlled inverter

75
 The PI controllers produce a signal that is proportional to the time integral of the controller.

The transfer functions of the PI controllers are given by:

K
EK  I1 (5.14)
P1 s

K
DK  I2 (5.15)
P2 s

where K are the proportional gains and K the integral gains.

P I

The inverter stage does not have any significant transient time associated with it and hence it is

modeled as an ideal gain. This ideal gain can be given by G  s   1 .

I

The transfer function of the LCL filter is given by:

I (s) 1
F ( s)  d 
 
(5.16)
V ( s) s3C L L  s 2C L Z  s L  L  Z
in f 1 2 f 1 1 2

The transfer function of a parallel RLC load, Z(s), is given by:

sL R
Z ( s)  L L (5.17)
2
s R C L  sL  R
L L L L L

Using Fig. 5.11 and equations (5.14) to (5.17), the transfer function of the current controlled

system is derived as:

V ( s) FZDE
H ( s)  d  (5.18)
V (s) 1  KF  FZ  FD  FZDE 
d - ref

By using K  0.9425 , R  4.33 , L  4.584mH , C  1.535mF , L  1mH , L  0.5mH ,

L L L 1 2

1 1
C  31 F , K  0.8 , K  25 , K  , K  , then,
f P1 I1 P2 0.8 I 2 25

76
 25
E  0.8  (5.19)
s

1
D  1.25  (5.20)
25s

H ( s) 

4.203x1013 s 2  31.282s  1 
s6  150.454s5  9.8219 x107 s 4  1.5601x1011 s3  7.7068 x1013 s 2  2.1418 x1016 s  4.0878x1014

(5.21)

5.6. Simulation Results

The performance of the proposed control strategy was evaluated by computer simulation

using SABER. Fig. 5.12 shows the simulated system.

The R load was adjusted to consume 10 kW. The DG system was designed to supply 10 kW

and zero reactive power. The system was operated initially in grid-connected operation. The grid

was disconnected at 0.3 seconds. After the island was detected the control mode was changed

from current controlled to voltage controlled operation. Fig. 5.13 shows the currents at the PCC

before and after grid disconnection without the seamless transition controller implemented, while

Fig. 5.14 shows the currents at the PCC before and after grid disconnection with the seamless

transition controller implemented.

77
 Voltage Source Inverter PC
Sa LI Lg Ia VIaC
a
Vdc b
DG unit c
Sa’ Cf
LCL Filter Load
Ia Ib Ic Mode of Operation Switch θnew
Sa-a’
Controller for Grid-connected Da
ABC DQ Db SPWM Sb-b’
Grid-connected
DQ θnew Operation intentional ABC Dc
ID IQ
islanding Sc-c’

Controller for intentional islanding operation Load shedding VIa VIb VIc
VDref IDref + algorithm
current +
+
voltage
- ++
- + regulator - + regulator Rload shedding ABC
VQref - IQref -
K
-
+

VD VQ K VQ
ID IQ VD V V
VI-acVI-bcVI-caVG-abVG-bcVG-ca
PLL
θnew
Re-synchronization
θold VD VQ f θold
θnew
Fig. 5. 12 Simulated systems

78
 (V) : t(s)
100.0 Va
grid-connected stand-alone
200.0 Vb
Vc
(V)
(V)

0.0 0.0
(V) : t(s)
grid_condition
-200.0
-100.0 (A) : t(s)
100.0 Ia
grid-connected stand-alone
Ib
Ic
(A)

0.0
grid_condition

-100.0
0.36 0.38 0.40 0.42 0.44 0.46
t(s)

Fig. 5. 13 From grid-connected to stand-alone operation with severe transients

(V) : t(s)
100.0 Va
grid-connected stand-alone
200.0 Vb
Vc
0.0
(V)

0.0
grid_condition
-200.0
-100.0 (A) : t(s)
100.0 Ia
grid-connected stand-alone
Ib
0.0 Ic
(A)

grid_condition

-100.0
0.34 0.36 0.38 0.40 0.42 0.44 0.46 0.48 0.50 0.52
t(s)
Fig. 5. 14 From grid-connected to stand-alone operation without severe transients

79
 Fig. 5.15 shows a line-to-line voltage and a phase current at the PCC before and after grid

disconnection with the seamless transition controller implemented. As can be noticed, the

proposed controller successfully suppresses the severe transients caused by the disconnection of

the grid.

(V) : t(s)
500.0 100.0 Vca
grid-connected stand-alone
(V) : t(s)
grid_condition
0.0 0.0
(V)

(V)

-500.0 -100.0 (A) : t(s)

100.0 Ic
grid-connected stand-alone
grid_condition

0.0
(A)

-100.0
0.33 0.34 0.35 0.36 0.37 0.38 0.39 0.4 0.41 0.42 0.43 0.44 0.45 0.46 0.47 0.48 0.49
t(s)

Fig. 5. 15 From grid-connected to stand-alone operation without severe transients

To keep the magnitude of the voltage in its normal operational range when there is a power

mismatch, the load shedding algorithm proposed was implemented. Fig. 5.16 shows the

theoretical voltage transients under a power difference without the load shedding algorithm

implemented. For this case, when the voltage is out of the normal operating point, the load

80
shedding algorithm cut off the power difference from the load and the voltage was brought back

to the normal range. Fig 5.17 shows that the suitable load disconnection results in voltage

recovery, compared to the case of “no load shedding”.

(V) : t(s)
200.0 Va
Va
150.0

100.0

50.0
(V)

0.0

-50.0

-100.0

-150.0

-200.0
0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.36 0.38
t(s)

Fig. 5. 16 Phase voltage Va without load shedding algorithm

(V) : t(s)
200.0
(0.304,169.21) (0.457,169.78) Va
150.0
100.0
50.0
(V)

0.0
-50.0
-100.0
-150.0
-200.0
0.2 0.24 0.28 0.32 0.36
t(s)

Fig. 5. 17 Phase voltage Va with load shedding algorithm

81
5.7. Experimental Results

The DG is started up in grid-connected operation mode, and then the separation device is

opened. When the DG is disconnected from the grid it operates in stand-alone mode. Fig. 5.18

shows a voltage and phase currents when the disconnection device is opened and the seamless

controller from grid-connected to stand alone is not implemented.

Fig. 5. 18 Transition from grid-connected to stand-alone operation with severe transients

Fig. 5.19 shows the grid voltage, the DG line to line voltage, and the phase current before

and after grid disconnection with the seamless transition controller implemented. As can be

noticed, the proposed controller successfully suppresses the severe transients caused by the

disconnection of the grid.

82
 grid-connected stand-alone
V_DG

V_grid Voltages (L-L)

100v/div

current

Fig. 5. 19 Transition from grid-connected to intentional islanding operation without severe

transients

Fig. 5.20 shows the line to line voltage for the grid and the DG when the system is operating

in the islanding mode. Also shown are the load currents for the same type of operation. Fig. 5.21

shows the line to line voltages of the DG and the phase currents when the system is operating in

the islanding mode. As can be seen, the proposed control scheme is capable of maintaining the

voltages within the designed levels.

83
 Fig. 5. 20 Line to line voltages and phase currents during intentional islanding operation

Voltages (L-L)
100v/div

current

Fig. 5. 21 DG line to line voltages and phase currents during intentional islanding operation

84
 To test the load shedding algorithm, a test case where the islanded network is supplying

330W and importing 330W of active power from the grid was analyzed. Starting from this point,

in steady-state, the DG is disconnected and the network will become islanded. As shown in Fig.

5.22 it can be noticed that the suitable load disconnection results in voltage recovery. A total load

of around 640 W is curtailed to 320 W through load shedding, which is within the DG

capabilities. It can also be noticed from Fig. 5.21 that the load shedding assists the voltage to

reach acceptable values above the threshold selected.

Fig. 5. 22 Implementation of the load shedding algorithm

85
CHAPTER 6. CONCLUSIONS AND FUTURE WORKS

6.1 Conclusions

This research presented the development and test of a control strategy for DG capable of

working in both intentional-islanding (stand-alone) and grid-connected modes. The stand-alone

control featured an output voltage controller capable of handling excess or deficit of generated

power and synchronization for grid reconnection with a seamless transition from stand-alone to

grid-connected operation modes. The grid-connected mode with current control was also enabled

for the case of power grid connection. This grid-connected control featured an output current

controller capable of loss of main detection, synchronization with the grid, and seamless

transition from grid-connected to stand-alone operation modes.

An automatic mode switch method based on a PLL controller was described in order to

detect the power grid disconnection or recovery and switch the operation mode accordingly.

The proposed control strategy has the following characteristics:

 Interface with AC System

 Power quality

 Synchronization with the Grid

 Loss of Main Detection or Islanding

 Transitions from grid-connected to stand-alone

 Load shedding

 Reclosing

86
  Transitions from stand-alone to grid-connected

The simulation results showed that the detection algorithm can distinguish between islanding

events and changes in the loads and apply the load shedding algorithms when needed. The

experimental results showed that the proposed control schemes are capable of maintaining the

voltages within the standard permissible levels during grid-connected and islanding operation

modes. In addition, it was shown that the re-closure algorithm causes the DG to resynchronize

itself with the grid. Also, the simulation and experimental results show that seamless transfer has

been achieved.

6.2 Future Works

There are some issues that could be explored in the future related to this work:

 Throughout this thesis, the DG unit was modeled by an ideal DC source and therefore,

the dynamic of the input source was neglected. As a future work, the effect of these

dynamic on the islanding detection and control of islanded systems can be studied.

 A situation particular to the three-phase signals is the concept of unbalance. The three-

phase signals are unbalanced if they either have unequal magnitudes or phase-

displacements unequal to 120 degrees. The performance of the three phase PLL

algorithm presented is not good when the utility voltage presents voltage unbalances. If

these unbalances are not taking into account, the information of the utility voltage,

obtained from PLL structures, can present undesired errors. Based on the theory of

symmetrical components, an unbalanced set of signals can be decomposed into positive-,

negative- and zero-sequence components. The positive-sequence component is a

87
 balanced set of signals. To eliminate the distorted utility voltage effects, a positive

sequence computation block can be used with good performance. This situation can be

considered as a future work.

 One of the selling points for DG is that the local generation follows a scalable system

model, where due to the typical small size of the units, one can install as many units as

needed to satisfy the requests of the loads, without having too much of extra capacity

sitting idle. This concept requires that DGs can be installed in parallel without any

restrictions.

 Storage need to be incorporated to satisfy the instantaneous power balance as a new load

comes on-line without penalizing the quality of other network quantities, such as bus

voltage magnitude. Load changes resulting in fast transients that exceed the ramping

capability of generation require storage availability from which to draw the required

transient energy. This inertia-less system presented is not well suited to handle step

changes in the requested output power.

88
APPENDICES

89
Appendix 1. Transfer Functions of the current controlled inverter

Fig. A1.1 shows the block diagram of the DG interface control for grid-connected operation.

D Vd
F(s) Z(s)
+ Vin Id Vd
+ e K + K I1 + Vin 1 Filter Load
Id-ref _ P1
s V*
Id

Fig. A1. 1 Block diagram of the current controlled inverter

The PI controller produces a signal that is proportional to the time integral of the controller.

The transfer function of the PI controller is given by:

K
DK  I1 (A1.1)
P1 s

where kp is the proportional gain and kI the integral gain.

The inverter stage does not have any significant transient time associated with it and hence it

is modeled as an ideal gain. This ideal gain can be given by GI(s) = 1.

The schematic circuit of the filter stage is shown in Fig. A1.2. It consists of an LCL filter and

a parallel RLC load.

90
 L1 Vx L2 Id
+
Vin Vd
Cf
-
RL LL CL
Filter: F(s)
Load: Z(s)

Fig. A1. 2 LCL Filter and parallel RLC load

The transfer function of the filter is given by:

I (s)
F (s)  d (A1.2)
V (s)
in

By Kirchhoff Current Law, the following expression is obtained:

V V
x in  sC V  I ( s )  0 (A1.3)
sL f x d
1

Equation (A1.3) can be re-expressed as:

V V
x  in  sC V  I ( s )  0 (A1.4)
sL sL f x d
1 1

91
Using Ohm‟s Law, the current Id is given by:

V 0
I (s)  x (A1.5)
d sL  Z ( s )
2

From equation (A1.5), the voltage Vx is given by:

V  I (sL  Z ) (A1.6)
x d 2

Substituting equation (A1.6) into equation (A1.4)

I ( sL  Z ) V
d 2
sL sL

f d 2 d
 in  sC I ( sL  Z )  I ( s )  0 (A1.7)
1 1

Equation (A1.7) can be re-expressed as:

sI L I Z V
d 2  d  in  I  s 2C I L  sC I Z  0 (A1.8)
sL sL sL d f d 2 f d
1 1 1

 Z L2  Vin
I  s 2C L  sC Z    1  (A1.9)
d
1 
f 2 f sL L sL
 1 1

92
V  L 
in  sL  s 2C L  sC Z  Z  2  1 (A1.10)
1
1 
I f 2 f sL L
d  1

 
V
in  s3C L L  s 2C L Z  s L  L  Z (A1.11)
I f 1 2 f 1 1 2
d

Then, the transfer function of the filter, F(s), is given by:

I (s) 1
F ( s)  d 
 
(A1.12)
V ( s) s3C L L  s 2C L Z  s L  L  Z
in f 1 2 f 1 1 2

The transfer function of the load, Z(s), is given by:

V (s)
Z (s)  d (A1.13)
I (s)
d

By Kirchhoff Current Law:

I I I I (A1.14)
d R L C
L L L

V V
I  d  d  sC V (A1.15)
d R sL L d
L L

93
Equation (A1.15) can be re-expressed as:

 1 1 
I V    sC  (A1.16)
d d R sL L
 L L 

I  
d   1  1  sC  (A1.17)
d  L
V R sL L
L 

Then, the transfer function of the load, Z(s), is given by:

V (s) 1
Z (s)  d  (A1.18)
I (s) 1 1
d   sC
R sL L
L L

This can be re-expressed as:

sL R
Z ( s)  L L (A1.19)
2
s R C L  sL  R
L L L L L

Using Fig. A1.2 and equations (A1.12) and (A1.19), the transfer function of the current

controlled system is derived as:

V ( s)
H ( s)  d (A1.20)
I ( s)
d  ref

94
I I e (A1.21)
d  ref d

eD  V * (A1.22)

V* V  V (A1.23)
d in

V FI (A1.24)
in d

I Z V (A1.25)
d d

From equation (A1.25),

V
I  d (A1.26)
d Z

Substituting equation (A1.25) into equation (A1.24)

V
V F d (A1.27)
in Z

Equation (A1.27) can be re-expressed as:

V
V  d (A1.28)
in FZ

95
Substituting equation (A1.27) into equation (A1.23)

V
V* V  d (A1.29)
d FZ

Substituting equation (A1.22) into equation (A1.29)

V
eD  V  d (A1.30)
d FZ

From equation (A1.30)

V 1
e   d V  (A1.31)
 FZ dD
 

From equation (A1.31)

V  1 V
I  d   1 d (A1.32)
d  ref Z  FZ  D

Equation (A1.32) can be re-expressed as:

V V V
I  d  d  d (A1.33)
d  ref FZD D Z

96
  1 1 1
I    V (A1.34)
d  ref  FZD D Z  d

I
d  ref 1 1 1
   (A1.35)
V FZD D Z
d

I
d  ref 1  ZF  FD
 (A1.36)
V FZD
d

Then, the transfer function of the voltage controller, H(s), is given by:

V ( s) FZD
H ( s)  d  (A1.37)
I ( s) 1  FZ  FD 
d  ref

Where

sL R
Z ( s)  L L
2
s R C L  sL  R
L L L L L

1
F (s) 
f 1 2 f 1 1 2 
s3C L L  s 2C L Z  s L  L  Z 

K
CK  I1
P1 s

97
R  4.33
L

L  4.584mH
L

C  1.535mF
L

L  1mH
1

L  0.5mH
2

C  31 F
f

K  0.8
P1

K  25
I1

651.466 s
Z (s) 
s 2  150.454 s  142117

F (s) 

6.4516 x1010 s 2  150.454s  142117 
s5  150.454s 4  9.8219 x107 s3  1.456 x1010 s 2  5.5783x1013 s  0.4187

98
 25
D  0.8 
s

3.365 x1013 s 2  1.051s

H (s) 
s6  150.108s5  9.83x107 s 4  6.62 x1010 s3  2.31x1013 s 2  7.58 x1015 s  2.28x1017

99
Appendix 2. Derivation of the Load Shedding Equations

The approach proposed is to detect the voltage change rate and profile after the power outage

and determine how much load shedding is needed before going to the intentional islanding

operation and switching to the voltage control mode. In order to accomplish this, the system

shown in Fig. A2.1 has been analyzed.

S2
R1 R2 Rpu=R1//R2
Id Idpu

Fig. A2. 1 System to implement load shedding

o Obtain the voltage amplitude expression before load shedding. The expressions for the

load voltages V ,V , and V can be found:

apu bpu cpu

V
a  I sin(t   ) (A2.1)
R d

b  I sin  t    2 
V
  (A2.2)
R d  3 

c  I sin  t    2 
V
  (A2.3)
R d  3 

In per unit:

V
apu
I sin(t   ) (A2.4)
R dpu
pu

100
V
bpu  2 
I sin  t     (A2.5)
R
pu
dpu  3 

V
cpu  2 
I sin  t     (A2.6)
R
pu
dpu  3 

Solving for the voltages:

V I R sin(t   ) (A2.7)
apu dpu pu

 2 
V I R sin  t     (A2.8)
bpu dpu pu  3 

 2 
V I R sin  t     (A2.9)
cpu dpu pu  3 

o Using V ,V , and V an expression for the voltage amplitude can be found,

apu bpu cpu

2
V  V 2 V 2 V 2 (A2.10)
pk 3 ab bc ca

 2 
V V V I R sin(t   )  I R sin  t     (A2.11)
abpu apu bpu dpu pu dpu pu  3 

  2 
V I R sin(t   )  sin  t     (A2.12)
abpu dpu pu   3 

   
V I R 3 sin  t     (A2.13)
abpu dpu pu   6 

 
V  3I R sin  t     (A2.14)
abpu dpu pu  6

 2   2 
V V V I R sin  t      I dpu R pu sin  t     (A2.15)
bcpu bpu cpu dpu pu  3   3 
  2   2  
I sin t      sin  t   
dpu pu   
V R (A2.16)
bcpu 3   3  

101
V I R   3 cos t     (A2.17)
bcpu dpu pu 

V   3I R cos t    (A2.18)

bcpu dpu pu

 2 
V V V I R sin  t    I R sin(t   ) (A2.19)
capu cpu apu dpu pu  3  dpu pu

  2  
V I R sin  t      sin(t   )  (A2.20)
capu dpu pu   3  

   
V I R  3 cos  t      (A2.21)
capu dpu pu   3 

 
V  3I R cos  t     (A2.22)
capu dpu pu  3

 
V2  3I 2 R 2 sin 2  t     (A2.23)
abpu dpu pu  6

V2  3I 2 R2 cos2 t    (A2.24)

bcpu dpu pu

 
V2  3I 2 R 2 cos 2  t     (A2.25)
capu dpu pu  3

V2 V 2 V 2
abpu bcpu capu
(A2.26)
     
 3I 2 R 2 cos 2  t      cos 2 t     sin 2  t     
dpu pu   3  6 

V2 V 2 V 2
abpu bcpu capu
(A2.27)
   
 3I R cos 2  t      cos2 t     sin 2  t    
dpu pu  3  6

2
V  V 2 V 2 V 2 (A2.28)
pk 3 ab bc ca

102
 2    
V  3I R cos2  t      cos2 t     sin 2  t     (A2.29)
pk 3 dpu pu  3  6

   
V  2I R cos2  t      cos2 t     sin 2  t     (A2.30)
pk dpu pu  3  6

o Derive the slope of the voltage amplitude

d  V (t ) 
s 
pk 
(A2.31)
dt

 2I R K
dpu pu
s (A2.32)
   
cos 2  t      cos 2 t     sin 2  t    
 3  6

where

       
K  sin  t     cos  t      sin  t     cos  t      sin t    cos t   
 3   3   6   6
(A2.33)

Solving for I
dpu

   
 s 2 cos 2  t      cos 2 t     sin 2  t    
 3   6
I  (A2.34)
dpu
2 R K
pu

Using R  R / / R and solving for R , where R represents the load to be shed,

pu 1 2 1 1

 2 I K
dpu
R  (A2.35)
1 pu    
s cos 2  t      cos 2 t     sin 2  t      2 I K
 3   6 dpu

103
Appendix 3. Transfer Functions of the voltage controlled inverter

Fig. A3.1 shows the block diagram of the DG interface control for intentional islanding

operation.

E D Vd
F(s) Z(s)
Vd-ref + + Vin Id Vd
e K I1 + b K + K I2 + Vin 1 Filter Load
K P1+ Id-ref _ s V*
_
P2
s
Vd Id

Fig. A3. 1 Block diagram of the voltage controlled inverter

The PI controllers produce a signal that is proportional to the time integral of the controller.

The transfer functions of the PI controllers are given by:

K
EK  I1 (A3.1)
P1 s

K
DK  I2 (A3.2)
P2 s

The inverter stage does not have any significant transient time associated with it and hence it

is modeled as an ideal gain. This ideal gain can be given by GI(s) = 1.

The schematic circuit of the filter stage is shown in Fig. A3.2. It consists of an LCL filter and

a parallel RLC load.

104
 L1 Vx L2 Id
+
Vin Vd
Cf
-
RL LL CL
Filter: F(s)
Load: Z(s)

Fig. A3. 2 LCL Filter and parallel RLC load

The transfer function of the filter is given by:

I (s) 1
F ( s)  d 
 
(A3.3)
V ( s) s3C L L  s 2C L Z  s L  L  Z
in f 1 2 f 1 1 2

The transfer function of the load, Z(s), is given by:

sL R
Z ( s)  L L (A3.4)
2
s R C L  sL  R
L L L L L

Using Fig. A3.9 and equations (A3.3) to (A3.4), the transfer function of the current controlled

system is derived as:

V ( s)
H ( s)  d (A3.5)
V ( s)
d  ref

V V  e (A3.6)
d  ref d

eE  I (A3.7)
d  ref

105
eE  I  b (A3.8)
d

bD  V  KI  V (A3.9)
d d in

V FI (A3.10)
in d

I Z V (A3.11)
d d

From (A3.11),

V
I  d (A3.12)
d Z

(A3.12) in (A3.10)

V
V F d (A3.13)
in Z

From (A3.13)

V
V  d (A3.14)
in FZ

106
(A3.14) in (A3.9)

V V
bD  V  K d  d (A3.15)
d Z FZ

From (A3.15)

 1 K1
b V  1  (A3.16)
d  FZ Z D

(A3.16) and (A3.12) in (A3.8)

V  1 K V
eE  d   1  d (A3.17)
Z  FZ Z D

From (A3.17)

 1 K 1 1 V
e     d (A3.18)
 FZD ZD D Z  E

(A3.18) in (A3.6)

V  e V (A3.19)
d  ref d

107
  1 K 1 1 
    1 V
d  ref  EFZD EZD ED ZE  d
V (A3.20)

From (A3.0)

V ( s) FZDE
H (s)  d  (A3.21)
V ( s) 1  KF  FZ  FD  FZDE 
d  ref

Using K  0.9425 , R  4.33 , L  4.584mH , C  1.535mF , L  1mH , L  0.5mH ,

L L L 1 2

1 1
C  31 F , K  0.8 , K  25 , K  , K  , then,
f P1 I1 P2 0.8 I 2 25

651.466 s
Z (s)  (A3.22)
s 2  150.454 s  142117

F (s) 

6.4516 x1010 s 2  150.454s  142117  (A3.23)
s5  150.454s 4  9.8219 x107 s3  1.456 x1010 s 2  5.5783x1013 s  0.4187

25
E  0.8  (A3.24)
s

1
D  1.25  (A3.25)
25s

V ( s) FZDE
H (s)  d  (A3.26)
V ( s) 1  KF  FZ  FD  FZDE 
d  ref

H ( s) 
 
4.203x1013 s 2  31.282s  1

s6  150.454s5  9.8219 x107 s 4  1.5601x1011 s3  7.7068 x1013 s 2  2.1418 x1016 s  4.0878x1014

(A3.27)

108
Appendix 4. Block Diagram of the 2407A DSP Controller

Fig. A4.1 shows the block diagram of the basic configuration for the LF2407A DSp

controller. The major interfaces of the board include the target RAM, SRAM memory, analog

interface, analog signals interface, PWM output signal interfaces, among others functions.

Fig. A4. 1 Block diagram of the 2407A DSP controller

109
REFERENCES

110
 REFERENCES

[1] A. Ajaja, "Reinventing electric distribution," Potentials, IEEE, vol. 29, pp. 29-31, 2010.

[2] E. M. Stewart, et al., "Analysis of a Distributed Grid-Connected Fuel Cell During Fault
Conditions," Power Systems, IEEE Transactions on, vol. 25, pp. 497-505, 2010.

[3] J. C. Vasquez, et al., "Adaptive Droop Control Applied to Voltage-Source Inverters

Operating in Grid-Connected and Islanded Modes," Industrial Electronics, IEEE
Transactions on, vol. 56, pp. 4088-4096, 2009.

[4] C.-S. Wu, et al., "Design of intelligent utility-interactive inverter with AI detection," in
Electric Utility Deregulation and Restructuring and Power Technologies, 2008. DRPT
2008. Third International Conference on, 2008, pp. 2012-2017.

[5] G. Franceschini, et al., "Synchronous Reference Frame Grid Current Control for Single-
Phase Photovoltaic Converters," in Industry Applications Society Annual Meeting, 2008.
IAS '08. IEEE, 2008, pp. 1-7.

[6] J. Selvaraj and N. A. Rahim, "Multilevel Inverter For Grid-Connected PV System

Employing Digital PI Controller," Industrial Electronics, IEEE Transactions on, vol. 56,
pp. 149-158, 2009.

[7] M. ROPP, et al., "Sandia Smart Anti-Islanding Project; Summer 2001: Task II
Investigation of the Impact of Single-Phase Induction Machines in Islanded Loads:
Summary of Results," SAND2002-1320; TRN: US200222%%275, 2002.

[8] T. C. Green and M. Prodanovic, "Control of inverter-based micro-grids," Electric Power

Systems Research, vol. 77, pp. 1204-1213, 2007.

[9] M. E. Haque, et al., "A Novel Control Strategy for a Variable-Speed Wind Turbine With
a Permanent-Magnet Synchronous Generator," Industry Applications, IEEE Transactions
on, vol. 46, pp. 331-339, 2010.

[10] J. C. Vasquez, et al., "Hierarchical Control of Intelligent Microgrids," Industrial

Electronics Magazine, IEEE, vol. 4, pp. 23-29, 2010.

[11] L. Yunwei, "Development of power conditioners and controllers for Microgrids,"

DOCTOR OF PHILOSOPHY, School of Electrical and Electronic Engineering, Nanyang
Technological University, 2005.

[12] R. A. Prata, "Impact of distributed generation connection with distribution grids - two
case-studies," in Power Engineering Society General Meeting, 2006. IEEE, 2006, p. 8 pp.

111
[13] N. Pogaku, "Analysis, Control and Testing of Inverter-Based Distributed Generation in
Standalone and Grid-Connected Applications," Ph.D., University of London, 2006.

[14] Y. Zhou, et al., "Grid-connected and islanded operation of a hybrid power system," in
Power Engineering Society Conference and Exposition in Africa, 2007. PowerAfrica '07.
IEEE, 2007, pp. 1-6.

[15] V. T. Consortium on Energy Restructuring. (2008, Distributed Generation Education

Modules. Available: http://www.history.vt.edu/Hirsh/CER/CER-Base.htm

[16] S. P. Chowdhury, et al., "UK scenario of islanded operation of active distribution

networks &#x2014; A survey," in Power and Energy Society General Meeting -
Conversion and Delivery of Electrical Energy in the 21st Century, 2008 IEEE, 2008, pp.
1-6.

[17] A. A. Bayod-Rújula, "Future development of the electricity systems with distributed

generation," Energy, vol. 34, pp. 377-383, 2009.

[18] A. A. Salam, et al., "A 50kW PEM fuel cell inverter-based distributed generation system
for grid connected and islanding operation," in TENCON 2009 - 2009 IEEE Region 10
Conference, 2009, pp. 1-5.

[19] M. Ciobotaru, et al., "Accurate and less-disturbing active anti-islanding method based on
PLL for grid-connected PV Inverters," in Power Electronics Specialists Conference,
2008. PESC 2008. IEEE, 2008, pp. 4569-4576.

[20] R. A. Walling and N. W. Miller, "Distributed generation islanding-implications on power

system dynamic performance," in Power Engineering Society Summer Meeting, 2002
IEEE, 2002, pp. 92-96 vol.1.

[21] P. Piagi, "Microgrid Control," PhD, Department of Electrical and Computer Engineering,
University of Wisconsin - Madison, Madison, 2005.

[22] R. Majumder, et al., "Power System Stability and Load Sharing in Distributed
Generation," in Power System Technology and IEEE Power India Conference, 2008.
POWERCON 2008. Joint International Conference on, 2008, pp. 1-6.

[23] S. Carley, "Distributed generation: An empirical analysis of primary motivators," Energy

Policy, vol. 37, pp. 1648-1659, 2009.

[24] T. S. Basso, "System Impacts from Interconnection of Distributed Resources: Current

Status and Identification of Needs for Further Development," NREL/TP-550-44727;
TRN: US200908%%14, 2009.

112
[25] A. Cardenas, et al., "An Active Anti-Islanding Algorithm for Inverter Based Multi-
Source DER Systems," in Power and Energy Engineering Conference, 2009. APPEEC
2009. Asia-Pacific, 2009, pp. 1-6.

[26] W. Fei and Z. Chengcheng, "Protection and Control for Grid Connected Photovoltaic
Power Generation System Based on Instantaneous Power Theory," in Circuits,
Communications and Systems, 2009. PACCS '09. Pacific-Asia Conference on, 2009, pp.
356-359.

[27] W. Fei and M. Zengqiang, "Passive Islanding Detection Method for Grid Connected PV
System," in Industrial and Information Systems, 2009. IIS '09. International Conference
on, 2009, pp. 409-412.

[28] "IEEE Standard for Interconnecting Distributed Resources With Electric Power
Systems," IEEE Std 1547-2003, pp. 0_1-16, 2003.

[29] L. Varnado and M. Sheehan, "Connecting to the Grid: A Guide to Distributed Generation
Interconnection Issues. ," Interstate Renewable Energy Council Connecting to the Grid
Project, 2009.

[30] R. Tirumala, et al., "Seamless transfer of grid-connected PWM inverters between utility-
interactive and stand-alone modes," in Applied Power Electronics Conference and
Exposition, 2002. APEC 2002. Seventeenth Annual IEEE, 2002, pp. 1081-1086 vol.2.

[31] L. Yunwei, et al., "Design, analysis, and real-time testing of a controller for multibus
microgrid system," Power Electronics, IEEE Transactions on, vol. 19, pp. 1195-1204,
2004.

[32] T. Thacker, et al., "Islanding Control of a Distributed Generation Unit's Power

Conversion System to the Electric Utility Grid," in Power Electronics Specialists
Conference, 2005. PESC '05. IEEE 36th, 2005, pp. 210-216.

[33] T. Thacker, et al., "Implementation of control and detection algorithms for utility
interfaced power conversion systems," in Applied Power Electronics Conference and
Exposition, 2006. APEC '06. Twenty-First Annual IEEE, 2006, p. 6 pp.

[34] H. Zeineldin, et al., "Intentional islanding of distributed generation," in Power

Engineering Society General Meeting, 2005. IEEE, 2005, pp. 1496-1502 Vol. 2.

[35] G. Fang and M. R. Iravani, "A Control Strategy for a Distributed Generation Unit in
Grid-Connected and Autonomous Modes of Operation," Power Delivery, IEEE
Transactions on, vol. 23, pp. 850-859, 2008.

[36] G. Iwanski and W. Koczara, "Autonomous power system for island or grid-connected
wind turbines in distributed generation," European Transactions on Electrical Power,
vol. 18, pp. 658-673, 2008.

113
[37] R. Majumder, et al., "Control of parallel converters for load sharing with seamless
transfer between grid connected and islanded modes," in Power and Energy Society
General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century,
2008 IEEE, 2008, pp. 1-7.

[38] H. Shengli, et al., "Control algorithm research on seamless transfer for distributed
resource with a LCL filter," in Electric Utility Deregulation and Restructuring and
Power Technologies, 2008. DRPT 2008. Third International Conference on, 2008, pp.
2810-2814.

[39] D. N. Gaonkar, et al., "Seamless Transfer of Microturbine Generation System Operation

Between Grid-connected and Islanding Modes," Electric Power Components and
Systems, vol. 37, pp. 174 - 188, 2009.

[40] C. Chien-Liang, et al., "Design of Parallel Inverters for Smooth Mode Transfer Microgrid
Applications," Power Electronics, IEEE Transactions on, vol. 25, pp. 6-15, 2010.

[41] L. Qin, et al., "Multi-loop control algorithms for seamless transition of grid-connected
inverter," in Applied Power Electronics Conference and Exposition (APEC), 2010
Twenty-Fifth Annual IEEE, 2010, pp. 844-848.

[42] Y. Zhilei, et al., "Seamless Transfer of Single-Phase Grid-Interactive Inverters Between

Grid-Connected and Stand-Alone Modes," Power Electronics, IEEE Transactions on,
vol. 25, pp. 1597-1603, 2010.

[43] K. H. Ahmed, et al., "Passive Filter Design for Three-Phase Inverter Interfacing in
Distributed Generation," in Compatibility in Power Electronics, 2007. CPE '07, 2007, pp.
1-9.

[44] J. Dannehl, et al., "Limitations of Voltage-Oriented PI Current Control of Grid-

Connected PWM Rectifiers With "LCL" Filters," Industrial Electronics, IEEE
Transactions on, vol. 56, pp. 380-388, 2009.

[45] O. Vodyakho and C. C. Mi, "Three-Level Inverter-Based Shunt Active Power Filter in
Three-Phase Three-Wire and Four-Wire Systems," Power Electronics, IEEE
Transactions on, vol. 24, pp. 1350-1363, 2009.

[46] B. Singam and L. Y. Hui, "Assessing SMS and PJD Schemes of Anti-Islanding with
Varying Quality Factor," in Power and Energy Conference, 2006. PECon '06. IEEE
International, 2006, pp. 196-201.

[47] G. Hernandez-Gonzalez and R. Iravani, "Current injection for active islanding detection
of electronically-interfaced distributed resources," Power Delivery, IEEE Transactions
on, vol. 21, pp. 1698-1705, 2006.

114
[48] H. Karimi, et al., "Negative-Sequence Current Injection for Fast Islanding Detection of a
Distributed Resource Unit," Power Electronics, IEEE Transactions on, vol. 23, pp. 298-
307, 2008.
[49] "IEEE Recommended Practice for Utility Interface of Photovoltaic (PV) Systems," IEEE
Std 929-2000, p. i, 2000.

[50] S. Alepuz, et al., "Control Strategies Based on Symmetrical Components for Grid-
Connected Converters Under Voltage Dips," Industrial Electronics, IEEE Transactions
on, vol. 56, pp. 2162-2173, 2009.

[51] M. Castilla, et al., "Linear Current Control Scheme With Series Resonant Harmonic
Compensator for Single-Phase Grid-Connected Photovoltaic Inverters," Industrial
Electronics, IEEE Transactions on, vol. 55, pp. 2724-2733, 2008.

[52] M. Brucoli, et al., "Modelling and Analysis of Fault Behaviour of Inverter Microgrids to
Aid Future Fault Detection," in System of Systems Engineering, 2007. SoSE '07. IEEE
International Conference on, 2007, pp. 1-6.

[53] A. Pigazo, et al., "Wavelet-Based Islanding Detection in Grid-Connected PV Systems,"

Industrial Electronics, IEEE Transactions on, vol. 56, pp. 4445-4455, 2009.

[54] T. Thacker, et al., "Islanding Detection Using a Coordinate Transformation Based Phase-
Locked Loop," in Power Electronics Specialists Conference, 2007. PESC 2007. IEEE,
2007, pp. 1151-1156.

[55] R. M. Santos Filho, et al., "Comparison of Three Single-Phase PLL Algorithms for UPS
Applications," Industrial Electronics, IEEE Transactions on, vol. 55, pp. 2923-2932,
2008.

[56] L. Qin, et al., "Islanding control of DG in microgrids," in Power Electronics and Motion
Control Conference, 2009. IPEMC '09. IEEE 6th International, 2009, pp. 450-455.

[57] S. Hirodontis and H. Li, "An adaptive load shedding method for intentional islanding," in
Clean Electrical Power, 2009 International Conference on, 2009, pp. 300-303.

[58] H. Gaztanaga, et al., "Real-Time Analysis of the Control Structure and Management
Functions of a Hybrid Microgrid System," in IEEE Industrial Electronics, IECON 2006 -
32nd Annual Conference on, 2006, pp. 5137-5142.

[59] G. Iwanski and W. Koczara, "DFIG-Based Power Generation System With UPS Function
for Variable-Speed Applications," Industrial Electronics, IEEE Transactions on, vol. 55,
pp. 3047-3054, 2008.

[60] J. M. Espi Huerta, et al., "A Synchronous Reference Frame Robust Predictive Current
Control for Three-Phase Grid-Connected Inverters," Industrial Electronics, IEEE
Transactions on, vol. 57, pp. 954-962, 2010.

115