THS-

 

 

1115513

\

6

>01”)

 

 

_UBRARY
MlCrup Qtate
University

 

 

 

This is to certify that the
thesis entitled

PHOTOVOLTAIC SYSTEM FOR STANDALONE/
GRID-CONNECTED APPLICATIONS

presented by

SANGMIN HAN

has been accepted towards fulfillment
of the requirements for the

MS. degree in Electrical Engineering

 

 

9‘

9/12”! #96"

V Majc“ Pi'ofessor’gTi-gnature
777% 1L, £010

Date

MSU is an Affirmative Action/Equal Opportunity Employer

o.—-—.—-q.—.—.—.—.-.-.-o--.--.-.—.-.-.—.—.-.-.-.P.—.-.-.-.-.-._.-.-—.—.------o-.-o-.--.-.-.—.—-—-.--.--—.-.--

 

 

 

 

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE

 

DATE DUE «y
mi It LU‘L
Jain

5’ 5Mp‘7i l 1% ’le‘ll

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 
 

 

 

5/08 K:IProj/Acc&Pres/CIR )ateDueindd

 

 

 

PHOTOVOLTAIC SYSTEM FOR
STANDALONE/GRID-CONNECTED
APPLICATIONS

By

Sangmin Han

A THESIS

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

MASTER OF SCIENCE
Electrical Engineering

2010

ABSTRACT

PHOTOVOLTAIC SYSTEM FOR
STANDALONE/GRID-CONNECTED APPLICATIONS

By
Sangmin Han

This Research develops design, analysis, and implementation of a series connected
PV system for stand-alone/grid-connected photovoltaic power generation. The
objective of the gateway dc-ac conversion system is to develop a low-cost conversion
system based on the intelligent controller to provide clean residential electricity. The
feature of the prototype includes: (l)the combination of dc-dc, dc-ac and/or ac-dc
converters for the flexible and uninterruptible energy utilization; (2)a smart algorithm
for optimizing the power consumption and defining priority on the demand system. A
buck converter is used for the MPPT implementation and presents the functions of
battery charger and step-down converter. The dc-dc converter realize phase shifting to
control power flow through a transformer with a MOSFET full bridge on the low
voltage side. In addition, a voltage doubler on the high voltage side is installed to
achieve enough high voltage to run the inverter. The operation principle, theoretical

analysis, simulation and experimental results are presented in this paper.

To my parents, Youngyul Han and Eunmo Kim

iii

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr. Peng for his guidance and encouragement
throughout my study. I am grateful of his support throughout the two years I spent
working on the Gateway project.

I would like to thank several individual who contribute to my research work. Dr.
Song and Dr. Min introduced me to power electronics and provided opportunities to gain
various technical and non-technical experiences through various experiment and
discussion. Dr. Wang and Dr. Yi developed the power stage used in this research and
provided fruitful discussions on the subject of solar power system level and the
simulation. I would like to extend my appreciation to my committee members for their
help. I would also like to express my appreciation to Dr. Yoo and Dr. Cha for helping me
studying at MSU and provided fruitful discussions on the subject.

It was a pleasure to work with PEMD colleagues, Wei, Craig, Dong, Uthane, Irvine,
Xi, Qin, George, Shuai and Matt for their help and cooperation.

I am indebt to my parents, who always provide endless love and encouragement
throughout my life.

Sangmin Han

May 2010

iv

TABLE OF CONTENTS

 

 

 

 

 

LIST OF TABLE vii
LIST OF FIGURES viii
CHAPTER 1 1
1.1 Motivations and Objectives of Research 1
1.2 Background 2
1.3 System Configuration 2

 

CHAPTER 2 5
2.1 DC-DC CONVERTER Topology
2.2 PV Source Model & MPPT Algorithm

 

 

 

 

 

 

 

 

2.2.1 Solar Cell Modeling ...................................................................................... 8
2.2.2 MPPT Algorithm ......................................................................................... 11
2.2.3 Perturb and Observe Approach ................................................................... 14

2.3 Design of DC-DC Converter 17
2.3.1 Inductance Value Selection ........................................................................ 17
2.3.2 Inductor Design .......................................................................................... 18
2.3.3 Capacitance of Cdcl Selection ................................................................... 19
2.3.4 Capacitance of va Selection ..................................................................... 20

2.4 DC-DC Converter Control Scheme 22
2.4.1 Simulation Result of Buck Converter ......................................................... 23
2.4.2 MPPT Simulation ....................................................................................... 27
2.4.3 Experimental Result ................................................................................... 28
CHAPTER 3 32
3.] Isolated DC-DC Converter Topology 32
3.1.1 DC—DC Circuit Selection ............................................................................ 33

3.1 .2 Controlled Rectifier and Uncontrolled Rectifier ......................................... 35
3.1.3Isolated DC-DC Converter Control ............................................................. 39

3.2 Design and Development of DC-DCFull-Bridge Converter 42
3.2.1 Turns Ratio Selection of Transformer ......................................................... 42
3.2.2 Leakage Inductance Selection ..................................................................... 43
3.2.3 Core and Winding Selection ....................................................................... 44
3.2.4Device Selection .......................................................................................... 49

3.3 Simulation & Experimental Result 49
3.3.1 Simulation Result ....................................................................................... 49
3.3.2 Experimental Result ................................................................................... 51

Chapter 4 54

 

 

 

 

 

 

 

4.1 Inverter Design 54
4.1.1 Overall Inverter Configuration ................................................................... 54

4.1.2 DC Input Stage ........................................................................................... 56

4.1.3 Bridge Stage ............................................................................................... 57

4.1.4 AC Filter Stage ........................................................................................... 60

4.2 Inverter Control Methodology 62
4.2.1 Overall Inverter Control ............................................................................. 62

4.2.2 Voltage Control ........................................................................................... 66

4.2.3 Current Control ........................................................................................... 67

4.3 Test Setup and Software Implementation 71
4.3.1 Test Setup ................................................................................................... 71

4.3.2 Software Development ............................................................................... 73

4.4 Simulation and Experimental Results 76
4.4.1 Simulation Results for Voltage and Current Control .................................. 76

4.4.2 Experimental Results for the Stand-Alone Mode Operation ...................... 81

4.4.3 Experimental Results for the Utility-Connected Operation ....................... 87
Chapter 5 97
Conclusion 97
Future work 99
BIBLIOGRAPHY 100

 

vi

LIST OF TABLES

Table 1 Test setup parameters ........................................................................................................ 88

Table 2 Inverter and grid power factor ......................................................................................... 96

vii

LIST OF FIGURES

Figure 1 Overall PV System Configuration ........................................................... 3
Figure 2 PV—Connected Converter (between PV array and 48V DC bus) ............. 5
Figure 3 Typical PV panel characteristic ............................................................... 6
Figure 4 Typical I-V and P-V characteristic of solar cell ...................................... 9
Figure 5 I-V curves for various irradiance .......................................................... 10
Figure 6 I-V curves for various temperature ........................................................ 11
Figure 7 I-V characteristics under wide operating conditions ............................. 12
Figure 8 Direct coupled method .......................................................................... 12
Figure 9 Basic components of a maximum power pointer tracker ...................... 13
Figure 10 Perturb & Observe(P&O) control action ............................................. 15
Figure 11 MPPT scheme for buck/boost converter ............................................. 22
Figure 12 Buck converter output current ............................................................. 24
Figure 13 Buck converter output current ripple .................................................. 24
Figure 14 Control block diagram of MPPT system ............................................. 25
Figure 15 dc-dc converter voltage control waveform ......................................... 25
Figure 16 DC-DC converter current control waveform ...................................... 26

Figure 17 Switch signal at duty cycle 0.5(blue: triangular wave, red: gate signal)

..................................................................................................................... 26
Figure 18 Simulation results of the MPPT(P&O) ............................................... 27
Figure 19 Buck converter output current, diode voltage and gate signal ............ 28
Figure 20 DC-DC converter control block diagram ............................................ 28
Figure 21 Buck/Boost converter lkW power experiment ................................... 29

viii

Figure 22 Inductor current, input output voltage & diode voltage ...................... 29
Figure 23 Current control transient waveform .................................................... 30
Figure 24 Voltage control transient waveform .................................................... 31

Figure 25 Grid-connected converter(between 48V DC bus and 240V AC bus) . 32

Figure 26 Isolated DC-DC Converter model ....................................................... 34
Figure 27 Two configurations of the secondary side in the isolated dc-dc

converter ...................................................................................................... 35
Figure 28 topology of the isolated dc-dc converter ............................................. 39
Figure 29 Control scheme for isolated dc-dc converter ...................................... 41
Figure 30 Power transfer curve of utilizing leakage inductance ......................... 42

Figure 31 Simplified equivalent circuit referred to the primary(low voltage side)

..................................................................................................................... 43
Figure 32 DC-DC converter transformer design ................................................. 47
Figure 33 Transformer short circuit test (leakage inductance test) ..................... 48
Figure 34 Transformer open circuit test (magnetizing current) ........................... 48
Figure 35 DC-DC High side voltage & voltage reference .................................. 50
Figure 36 Transformer Primary voltage & Secondary voltage ............................ 50
Figure 37 Transformer voltage & current after turning on inverter .................... 51
Figure 38 the experimental result of the dc—dc converter. .................................. 52
Figure 39 the experimental result of the dc—dc converter. .................................. 53
Figure 40 Relationship between A and B leg control signal ................................ 58
Figure 41 . PWM Switching Scheme for leg A of inverter .................................. 58
Figure 42 Combined PWM Switching for inverter ............................................. 60
Figure 43 Inverter control system ........................................................................ 63

ix

Figure 44 Picture of control board with input/output connection ....................... 65

Figure 45 Inverter voltage control ....................................................................... 67
Figure 46 Grid connected current control ............................................................ 68
Figure 47 Block diagram of product type PLL .................................................... 69
Figure 48 PLL simulation result .......................................................................... 69
Figure 49 PLL experimental result ...................................................................... 70
Figure 50 Power circuit of Inverter system connected with grid ........................ 71
Figure 51 Single line diagram of test setup ......................................................... 73
Figure 52 Simulation results for current control mode of operation ................... 77
Figure 53 Single diagram of simulation circuit grid-connected mode ................ 78
Figure 54 Simulation results for voltage control mode of operation ................... 79
Figure 55 Simulation Waveforms for Complete Inverter Operation ................... 80

Figure 56 Single diagram figure of current and voltage control: transition of

control mode ................................................................................................ 81
Figure 58 First category operation mode: stand-alone ........................................ 82
Figure 59 Experimental results of inverter.(Stand alone mode.) ......................... 83
Figure 60 Stand-alone mode power flow ............................................................ 84
Figure 61 L-R Load waveform at PF 0.73 ........................................................... 85
Figure 62 L-R Load waveform at PF 0.73 ........................................................... 85
Figure 63 Efficiencies of the Voltage-Controlled Inverter .................................. 86
Figure 64 Second category operation mode: grid-connected .............................. 87
Figure 65 Grid-connected mode power flow before turning on the inverter ....... 89
Figure 66 Inverter output voltage and current ..................................................... 90
Figure 67 Grid-connected mode power flow afier turning on the inverter ......... 90

Figure 68 Experimental results of inverter.(Grid connected mode.-1KW) ......... 91

Figure 69 Inverter output voltage, current and grid current ................................ 92
Figure 70 Experimental results of inverter.(Grid connected mode.-2KW) ......... 92
Figure 71 Inverter output voltage, current and grid current. ............................ 93
Figure 72 Inverter output voltage, current and grid current ................................ 94
Figure 73 Ac load voltage, current & Dc load voltage and current ..................... 95

Figure 74 Inverter output voltage, current and grid current(inductive ac load test)
..................................................................................................................... 96

xi

CHAPTER 1

1.1 Motivations and Objectives of Research

Renewable energy sources become a more and more important contribution to the
total energy consumed in the world. Because of their independence from limited fossil
and nuclear fuels and their low impact on the environment will become the only crisis-
proof and reliable energy supply within the next decades. Although PV energy has
received considerable attention over the last few years, the high installation cost of PV
systems and the low conversion efficiency of PV modules are the major obstacles to
using this alternative energy source on a large scale. Therefore, many studies are being
researched in order to minimize these disadvantages [l]—[5]. In 2009 the idea of the
Gateway project was introduced with the goal to develop a low cost sustainable source of
electrical power to residential applications. Development of a low-cost DC/AC inverter
and controller system to support the commercialization of a maximum '1 kW photovoltaic
cell module providing clean residential electricity (each input and output is rating on
1kW). The system also needs to provide an AC/DC inverter with the same capacity.
Ideally the optimization of the system will reduce the number of electronic components
in each inverter (it is requirement to have a single inverter for the system). A smart
algorithm for optimizing the power consumption and defining priority on the demand
system are also required. The design of the inverter should optimize a share DC bus
between the DC inputs and the rectified AC inputs considering the best trade—off between
efficiency and cost. It consists of a photovoltaic (PV) array for electricity generation,

battery for energy storage, digital control system to perform maximum power point

tracking (MPPT) and battery charge control, and various DC loads and AC loads.

1.2 Background

Highly reliable and efficient power processing systems allow exploitation of the
enormous potential of the renewable sources by transforming the maximum available
power into an electrical one, fed into the grid or converted into a high density energy
vector for being stored and used in another place or at another time, when the primary
source is not available. PV industry is growing at more than 30% per year and the cost of
PV energy will reach the break-even point very soon in many countries. The most
relevant goals of PV energy include 40% cost reduction of PV panels and of the power-
converter stage within five years, increased efficiency of both panels and converters, and
considerable improvement in converter reliability. These goals are driving the research
toward several directions: Higher efficiency of the power converter such as
transformerless topologies and converters based on silicon carbide devices, maximum
power extraction algorithms, integration of the power converter in the panel to reduce the
problem of minor energy production due to panel mismatch and advanced islanding
detection algorithms that help monitor grid connection and compliance with grid standard
and codes. These will have impact on safety, since undetected island in the power grid.
may cause hazards to user and utility workers. The main objective of our project is to
design and construct a PV based system that produces electric energy and operates in
dual mode, supplying stand alone AC loads or the grid, while minimizing its cost and size.
The system’s main properties are production of quality electricity from a renewable
source to reduce dependence on fossil fuels and the associated emissions of pollutants,

reduce cost of energy consumption by being able connect to the grid: sell energy and

charge the energy in storage batteries which are the most expensive system components
per watt. Our goal is to design and develop an 3 module series connected system that will
handle the task described and implement the Maximum Power Point Tracking (MPPT)
algorithm to achieve the maximum power possible from the PV panel and

synchronization with the utility grid.

1.3 System Configuration

The developed conversion system depicted in Figure 1 provides energy to both do and
ac load from PV panels and ac grid together. The PV panels are connected via the dc
input terminal; the utility ac grid is plugged in through the ac input. Besides, a 48 V
battery stack is paralleled with the dc output terminal on the purpose of energy storage. In
order to extract the maximum power of the PV array, the implementation of the
maximum power point tracking (MPPT) in stand-alone/grid-connected systems is
generally accomplished by the series connection of a dc—dc buck converter between the

PV array and the load or the energy storage element.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Buck-Boost
PV 48~96|V| CONVERTER 48Vdc
In" fi
va — LI Vodc
Cdcl
m := 51.1%} == ‘=-‘ ”may
i 3 3
DC-DC DC-AC LC AC GRID
CONVERTER Vdcfl 400v INVERTER FILTER 240v
. W .. - . 9 o
Cch __
. J 151} A 4%} ’43 ”<3 u
: NL _H [up
‘7ch ? 9| é ' ;;cr 79?:
' 1%} J51} “a“ J J J
4 V o o

 

 

 

 

 

 

Figure 1 Overall PV System Configuration

Since the desired dc voltage is 400 V which is much higher than the dc bus voltage of
48V, that is, a large step-up conversion ratio is required for the dc-dc converter, the use of
a transformer can allow better converter optimization. By proper choice of the
transformer turns ratio, the voltage or current stresses imposed on the switches and diodes
can be minimized, leading to improved efficiency and lower cost. On the other hand, the
transformer can provide galvanic isolation between the high voltage and the low voltage
to satisfy safety requirements. As shown in , a full-bridge in low voltage side (LVS) and a
half-bridge in high voltage side (HVS) coupled by a high-frequency transformer are used
for the dc-dc converter. [6] Proposed topology is shown in Figure 1 and a general
description is explained next chapters. The inverter module consists of one 1PM switches
along with 3000uF of input capacitor. The output of the inverter is fed to the inductor-
capacitor (LC) line filter. The series-connected, 2mH inductors and parallel 7.5uF
capacitors are used in order to achieve desired voltage and current ripple characteristics.
The output of the inverter is a single-phase, 60 Hz, 240V waveform. The test setup
includes: the inverter; a 240V, 60Hz AC source that emulates the utility grid; and one sets
of resistances representing the local load. The inverter and utility are connected to each
other through 1 single-phase breakers. When breaker is closed; the inverter operates in
the utility-connected mode. In that scenario, both the inverter and utility provide power to
the local load. When breaker is opened; the inverter operates in the stand-alone mode. In

this scenario, only the inverter provide power to the local load.

CHAPTER 2

2.1 DC-DC CONVERTER Topology

The dc-dc converter between the dc source and 48 V dc bus, termed as PV-connected
converter, is used to transmit electricity from photovoltaic array (variable) to the dc bus
(constant). Figure 2 shows the circuit of the PV-connected converter. It is a dc-dc buck
converter consisting of two switches with anti-parallel diode, an inductor and two
capacitors. As described in specifications, the maximum output power of PV array is 1
kW, so the capacity of the converter should match the demand. The PV-connected

converter will work in buck mode (variable input constant output) when power flows

 

 

 

 

 

 

 

 

 

 

 

 

 

from PV to do load.
48Vdc
, Di 3 AVA I'Kfl WAAr ;
PV 48~96[V| L”, 1.1
SH ._ Cdcl Vodc
va fl: SL131} :: -_:' Vbattery
va
1r 4 ‘1 3
pv

va* [LI

”PT 3, p, ' Pl PWM L» £2!

 

 

 

 

 

 

 

 

va fl “ I I

Figure 2 PV-Connected Converter (between PV array and 48V DC bus)

 

Given that the output voltage is Vodc = 48 V and the input voltage va varies in a wide
range from 48V to 96 V. The maximum input/output power is 1 kW. So, the maximum
output current is Iodc, max = lkW/48 V = 20.83 A, and the input current is [in = 1kW/(48
~ 96 V) = 20.83 ~ 10.41 A. The equivalent load resistance is R = 48 V / 20.83 A = 2.3 Q.

The basic relationship of the PV—connected converter in buck mode

is V0dc = V PV x Db“ ,and [Ode = I in / Dbu where Dbu is the duty cycle. Therefore,

Db” is in a range of 1.0 ~ 0.5.

2.2 PV Source Model & MPPT Algorithm

Figure 3 shows the current and voltage (I-V) characteristics of the typical PV array
used for residential power generation. These characteristics show the PV array behaving
as a current source, in the sense that the current remains constant as the voltage varies.
The standard Buck topology would have limited performance with a current source for its
electrical power input, since the output current of the converter could never be higher

than the maximum input current.

  

 

 

a -‘r —————— I
s V ‘
__-\. __ i

:12" :x‘ dP :
Q) I" '_—<0r
B .1 ' L dV i
o”
‘3‘ :\ l :
S 10 r-\:—l1-—--:
9-4 \ I
‘5 : ii :
o 5 : it—"g
. l. .

: . :l :

oO 5 20 25

10 15
output voltage(V)
Figure 3 Typical PV panel characteristic
The understanding of solar cell source behavior is necessary to specify the size of

solar array system or to study the stability of regulators. Placing the solar panels in an

optimal way is an effective measure to take to maximize the energy yield from a
photovoltaic-installation. Shading can rarely be completely avoided, certainly not in
urban or suburban environments. At higher latitudes the sun is often close to the horizon,
which makes the shading problem more severe than at low latitudes. Therefore,
especially for high-latitude locations and urban/suburban sites, a shading-tolerant system
should be chosen. Furthermore, the direction south/north (azimuth) and elevation angle
should also be considered. Any shading of the solar panels will lead to considerable
reduction in energy yield, even if just a small fraction of the panels is shaded.
Consequently, the case in which the solar cells in the generation system do not operate
under uniform generation conditions should be examined. The connection of photovoltaic
modules with different operating currents and/or voltages characteristic may result in the
performance of the array being less than the sum of the potential performances of
individual modules. When connected in series, the current flowing through the lowest
productive cell limits the entire array output. The problem arises when modules have
different sizes, which may be a result of individual modules not exposed to the same
lighting conditions, such as is the case with differently oriented modules or irregular
shading of the array. Mismatch is more likely in large systems than in independent arrays,
because individual modules may be oriented differently or they may be subject to varying
degrees of shading and heating.

Taking into account that real operating conditions of a solar array are difficult to
reproduce during tests, simulations to determine electrical characteristics will allow the
researcher to gain more insight for better understanding and design. Performance gain is

beneficial and can find applications in numerous areas such as the one described in

reference [6]. As the fine modeling of a cell's behavior is the basis for any solar generator,
the solar cell model needs to be properly determined. Methods for simulation of general
solar array configuration have been proposed; however, these implementations are
associated with special-purpose simulation tools, and detailed modeling has not been

performed.
2.2.1 Solar Cell Modeling

Solar cells are essentially a very large area p-n junction diode, where such a diode is
created by forming a junction between the n-type and p-type regions. As sunlight strikes
a solar cell, the incident energy is converted directly into electrical energy. Transmitted
light is absorbed within the semiconductor by using the energy to excite free electrons
from a low energy status to an unoccupied higher energy level. When a solar cell is
illuminated, excess electron-hold pairs are generated by light throughout the material;
hence the p-n junction is electrically shorted and current will flow. The equivalent circuit
of a solar cell is represented by four components: a light-induced current source, a diode
parallel to the source, a series resistor and a shunt resistor. The light-induced current is
due to the separation and drift of the photon-generated electron—hole pairs under the
influence of the built-in field. The corresponding current vs. voltage curves (I-V)
equation is:

5193152 £1132

V+IR
I=Iph—151(e "1” -1)-152(e "2” —1>———S—

Rs (2.1)

where, I and V are the solar cell output current and voltage, I ph is the generated

photo-current, and I S], 152 are the reverse saturation currents of each diode. The

parameters n1, n2 are diode ideality factors, T is the absolute temperature in Kelvin, k is

the Boltzmann’s constant 1.38x10—23J /K and q is the elementary charge

1.602 x10_19C . Consider a fixed environmental condition, a typical I-V characteristic
of a solar cell is shown in Figure 4. The operating point on power vs. voltage curve (P-V
curve) will depend on the solar array characteristic and the load. Assuming that initially
there is no load, the operating point will be at the far right at the open-circuit voltage,
V0c, of the solar array with zero current (V = Voc, I = 0). As the load increase, the
operating point will move up and to the left, i.e. voltage at the solar array terminal
decreased, while the power increases. As the load increases firrther, it will reach the
maximum power point (MPP), where the power drawn fi'om the solar cell is maximized.
The voltage at this point is denoted by the maximum-power voltage (Vmp), and the
current by maximum-power current (Imp). If the load increases beyond this point, the
voltage decreases and power drawn from the solar array decreases. Eventually, the
operating point will reach the far left at the short-circuit current, Isc, with zero voltage
output (V = 0, I = Isc).

Ill

ISC I VS. V

p————————-————_————

 

  

 

 

Vmp Voc V
Figure 4 Typical I-V and P—V characteristic of solar cell

Solar cell model parameters vary with environmental conditions, the two most

important effects being temperature and irradiance. Solar cell open-circuit voltage
decreases with increasing temperature, and the short circuit current is proportional to the
amount of irradiance. The current vs. voltage curves (I-V curves) for various irradiance
and temperature are shown in Figure 5 and Figure 6, respectively. The equations showing

modeling effects of irradiance and temperature on model parameters are:

Iph = 1ph_trefl:1+Ko(T—Tref)]

 

 

(2.2)
3
1 —1 T ex _qu 1 ”l—
s— s_tref Tref p nk Tref T (23)
Rs = Rs_tref [1" K3(T“ Tref )] (2.4)

where T ref is the cell reference temperature. E8 is the band gap energy of the

semiconductor. K0 is the short-circuit temperature coefficient, and K3 is the

resistance temperature coefficients. Each parameter value is varied relative to the value at

the cell reference temperature.

T = 25 deg C, irr(rated%) = 25,50,75,100

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

‘ """" 4 ““““ i """" f"" 4 J
r _,___,___'“‘
3.5 _______ .J: _______ J: _______ :L-:‘:-"'-\- lrr(rated%)=100
A 3 ....... _£ ....... .i ....... L____ __i ....... 1
c I I
{x 2.5 ------- -: ——————— 4. ....... .J ________
I
§ 2.._.______: _______ J______-L ..... 1c: ________
§ 1 s ------- A: ——————— a: _______ 1.--- ________
n r I \’L
1 ------ 1 ------- g--—--——' -------------
I
0.5 ------- -:---- lrr(rated%P25 l\ T ----- 4
I I
00 5» 1‘0 15 20 25

voltage(V)

Figure 5 I-V curves for various irradiance

10

 

T =2 25 deg C, irr(rated%) = 25,50,75,100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L... ----- + ------ ’
A 3.5 >- ------- i ------- '1 -------
V: 3 » ------- l ------- -«
v r
‘5 2.5 -------- l ------- -«
m _______ ' ______________ _'.- - 1 - _____
E 2 '- : -1 v . r-
3 1.5 L ------- ' ------ -. - — -1
D 1 ,_ _______ I _______ _, ._ __ _,
0.5 -------- ' ------- a
0o 5 25
voltage(V)
Figure 6 l-V curves for various temperature
2.2.2 MPPT Algorithm

The environmental condition under which a solar power system operates can be wide,
as shown in I-V curves in Figure 7. The current-voltage relation of a solar'array is
variable throughout the day, as it varies with environmental conditions such as irradiance
and temperature. In terrestrial applications, Low Irradiance, Low Temperature (LILT)
condition reflects morning condition where the sun just rises. A High Irradiance, High
Temperature (HIHT) condition might represent a condition near high noon in a humid
area. High Irradiance, Low Temperature (HILT) condition can represent a condition with
healthy sunlight in the winter. Finally, condition near sunset can be described by Low
Irradiance, High Temperature (LIHT) condition. For space application, LILT
characterizes a deep space mission or aphelion period, while HIHT condition is when

satellite orbits near the sun (perihelion).

11

HIHT

LILT

\ \.

V
Figure 7 I-V characteristics under wide operating conditions

 

For a uniformly illuminated array, there is only one single point of operation that will
extract maximum power from the array. In a battery charging system where the load seen
by the solar modules is a battery connected directly across the solar array terminals, the
operating point is determined by the battery’s potential. This operating point is typically
not the ideal operating voltage at which the modules are able to produce their maximum

available power.

 

 

Low current Low voltage
variation variation

 

 

 

 

 

IA

 

 

 

 

Figure 8 Direct coupled method
In the direct coupled method [7], in which the solar array output power is delivered
directly to the loads, as shown in Figure 8. To match the MPPs of the solar array as
closely as possible, it is important to choose the solar array I-V characteristic according

to the I-V characteristics of the load. A general approach for the power feedback control

12

is to measure and maximize the power at the load terminal, and it assumes that the solar
array maximum power is equal to the maximum load power. However, this maximizes
the power to the load not the power from the solar array. The direct-coupled method
cannot automatically track the MPPs of the solar array when the insulation or temperature
changes. The load parameters or solar array parameters must be carefully selected for the
direct coupled method. To be able to extract the maximum power from the solar array
and to track the changes due to environment, therefore, a maximum power point tracking
should be implemented. Devices that perform the desired function are known as
Maximum Power Point Trackers, also called MPPTs or trackers. A tracker consists of
two basic components, as shown in Figure 9: a switch-mode converter and a control with
tracking capability. The switch-mode converter is the core of the entire supply. The
converter allows energy at one potential to be drawn, stores as magnetic energy in an
inductor, and then releases at a different potential. By setting up the switch-mode section
in various topologies, either high-to-low (buck converter) or low-to-high (boost) voltage
converters can be constructed. The goal of a switch-mode power supply is to provide a
constant output voltage or current. In power trackers, the goal is to provide a fixed input
voltage and/or current, such that the array is held at the maximum power point, while

allowing the output to match the load voltage.

 

 

 

 

 

Figure 9 Basic components of a maximum power pointer tracker

l3

When properly applied, a maximum power point tracking control can prevent the
collapse of the array voltage under excessive load demand, particularly when supplying a
signals, such as the array current and voltage, to determine a proper direction to move the
operating point. Eventually, this continuously updated set point will fluctuate around the
voltage corresponding to the array peak power point. By adjusting the operating point of
the array to the point Vmp(maximum power point voltage), power output of the array is
maximized, and the most efficient use of the solar array may be realized.

For a system without MPPT, the voltage will quickly collapse to zero. This
phenomenon can be understood from the I-V characteristic of a solar array. The flatness
of the I-V curve on the left of the MPP implies that a small incremental increase in
current demand leads to large voltage change. A system with MPPT avoids the voltage
collapse by keeping the operating point near the MPP. On the I-V curve, the operating
point corresponding to the maximum-power point is around the “knee” region. Therefore,
unlike other power systems with stiff voltage sources, power conversion from solar array
sources with MPPT requires more robust design due to risks of an array voltage collapse
under peak load demand or severe changes in the array characteristics. The location of

the MPP of an I-V characteristic is not known a priori, and must be located.

2.2.3 Perturb and Observe Approach

As the name of the perturb-and-observe (P&O) states, this process works by
perturbing the system by increasing or decreasing the array operating voltage and
observing its impact on the array output power. The operating voltage is perturbed with
every MPPT cycle. As soon as the MPP is reached, V will oscillate around the ideal

operating voltage Vmp. Figure 10 summarized the control action of the P&O method.

14

The value of the reference voltage, Vref, will be changed according to the current
operating point. For example, for when the controller senses that the power from solar
array increases (dP > 0) and voltage decreases (d V < 0), it will decrease (-) Vref by a
step size C I, so Vref is closer to the MPP. The MPP represents the point where Vref and
scaled down Vsa become equal.

The oscillation around a maximum power point causes a power loss that depends on
the step width of a single perturbation. The value for the ideal step width is system
dependent and needs to be determined experimentally to pursue the tradeoff of increased
losses under stable or slowly changing conditions. In fact, since the AC component of the
output power signal is much smaller than the DC component and will contain a high
noise level due to the switching DC-DC converter, an increase in the amplitude of the
modulating signal had to be implemented to improve the signal to noise ratio (SNR),
however, this will lead to higher oscillations at the MPP and therefore increase power

losses even under stable environmental conditions.

 

 

 

 

 

 

 

 

 

 

 

Case dP (IV Action
‘l P l < 0 < 0 -
Case 4 Case 2 i :3 :3 :
4 > 0 > O -
Case 1
Case 3
v

 

 

7

Figure 10 Perturb & Observe(P&O) control action

Several improvements of the P&O algorithm have been proposed. One of the simplest

15

entails the addition of a ‘waiting’ function that causes a momentary cessation of
perturbations if the algebraic sign of the perturbation is reversed several times in a row,
indicating that the MPP has been reached. This reduces the oscillation about the MPP in
the steady state and improves the algorithm’s efficiency under constant irradiance
conditions. However, it also makes the MPPT slower to respond to changing atmospheric
conditions, worsening the erratic behavior on partly cloudy days. Another modification
involves measuring the array’s power P1 at array voltage V1, perturbing the voltage and
again measuring the array’s power, P2, at the new array voltage V2, and then changing
the voltage back to its previous value and remeasuring the array’s power, P1, at V1. From
the two measurements at V1, the algorithm can determine whether the irradiance is
changing. Again, as with the previous modifications, increasing the number of samples of
the array’s power slows the algorithm down. Also, it is possible to use the two
measurements at V1 to make an estimate of how much the irradiance has changed
between sampling periods, and to use this estimate in deciding how to perturb the
operating point. This, however, increases the complexity of the algorithm, and also slows
the operation of the MPPT. The MPPT techniques have much wider application than just
photovoltaic alone, since similar functionality of power output versus loading can be seen
in the I-V curves of other sustainable energy sources. Such sources are small water
turbines and wind-power turbines; however, the actual physics behind the IV curves for
the various sources are different. The voltage, current, and power produced by such
sources is variable in response to environmental conditions (insolation, pressure, or wind
speed) and dependent on the electrical impedance of the load. Under any combination of

environmental conditions, each of these sources is characterized by exactly one idea] load

16

impedance, which will result in operation at VMPPTand maximum power transfer.

2.3 Design of DC-DC Converter

2.3.1 Inductance Value Selection

There are three passive components. One is inductor and the other two are capacitors.
Generally, passive components are designed according to switching frequency and their
current ripple or voltage ripple requirements. This section focuses on the inductor design
and the capacitor design will be discussed in the later section. It is desired that the current
ripple through the inductor has a peak magnitude equaling to 20% of the full-load value

of the dc component current. For the buck converter, the current ripple is

 

 

 

 

 

D T (1— D )T

A1L1: bzuL S(va—V0dc)= 2:11 S 'Vodc

1 1 (2.5)
So, for the range duty cycle (1.0 ~ 0.5), we can select the inductance value:
A1L1 =M'Vod0510dcmax'20%:>141 Z (l—Dbu)TS ' odc

’ (2.6)
L12 (l—Dbu)TS 'VO 0: (l—Dbu)TS .48:2_'6_1_§ (27)
2 'Iodc,max - 20% 2 - 20.83 - 20% f5

For a switching frequency of 10 kHz, the inductance must be no less than:

2.618 2.618
L12: = 3
fs 10-10

 

 

= 0.2618 - 103(H) = 0.2618(mH)

For this inductance value, the current ripple is:

(l—Dbu)-TS (1—0.5).10‘4
: 2L 'VOdC :: _3
1 202618-10

 

-48 = 4.58(A)

 

a!“

17

2.3.2 Inductor Design

Use area product, A p , approach to design the inductor L1. First, calculate the A p :

4
_ L1 -1 2 pk -10 _ 0.000262x(20.83+4.584)2 x104
1’ BmJKu lx300x0.3

 

 

=18.72(cm4)
(2.8)

where Bm is operating flux density in [Tesla], J is current density in [A / cmz] , Ku is
copper fill factor. Second, select a powerlite C core from Metaglas because this kind of
core has high saturation flux density and low power loss. The part number is AMCC-25

which A p =18.72[cm4],Aw =8.4[cm2] .Then the minimum number of turns can be

calculated as

L -I
Nmin : 1 pk : 0.000262x25.4 z 26 (2.9)

Bm ‘Acore 1><2.7><10—4

 

 

Now use the magnetization curves of AMCC-25 to check the design. The selected
magnetizing force at peak power is 25 ><25.4 = 635[AT]. From the curve shown in Figure
11, we choose an air gap of 0.5 mm to prevent the core from getting in to saturation.

Then the inductance can be computed as follows:
L _ 2 _ 2 ~
I,cal — AL - N — 0.38 - 26 = 256.8(uH) (2.10)

In order to reduce the skin effect, Litz wire is the suitable selection for winding. Basically,

AWG 33 wire is the recommended wire for 10 kHz — 20 kHz application. Here we select

310-33 AWG Litz wire which insulated area, Acu = 0.146[cm2]. Then the window

utilization factor is

18

: N°Acu,J-ns = 26x0.146
Aw 8.4

= 0.45

 

K11
(2.11)

2.3.3 Capacitance of Cdcl Selection

For PV-connected converter (in buck mode), the output voltage ripple (peak-to-peak)

is expressed as

2
V -T V ‘T
.A odflzw'a-Dbufls zw'U—Dbu)

AV _ =(2
p p 8Ll'Cdcl 8Ll'Cdcl (2.12)
2 2 2
M_L_£_(l_0bu)=z’_(l_0bu)[_faj
Vodc 8 L1 'CdCI 2 fs
1
Where f6:

 

27rfLICdcl
Equation shows that the voltage ripple can be minimized by selecting a corner

frequency of the low-pass filter at the output such that fc<<fS(Generally, fs / fl. is

with a factor of at least 10. It is desired that the voltage ripple (peak magnitude) is

required to be 1%, hence, the necessary capacitance is calculated as

 

2
1%2A‘fodc :8LTSC -(1—Dbu)
odc l' dcl (2.13)

Cd 1> (1—1),,,,)TS2 _ (l—O.5)TS2 _ 2.4x103
6 _ _ __
3L1~1% 8-0.26-10_3-1% f,2

 

 

For a switching frequency of 10 kHz, the capacitance must be no less than

2.4x103 > 3.97x103
fs2 (10x103)2

 

Cdcl 2 = 240 x10—6(F) = 240(uF)

19

Considering the operating voltage is 48 V, a 3paralled lOO-V/lOO-uF film capacitor is
selected for Cdcl.

Hence, the voltage ripple is

 

Vodc-T52. (1_ Dbu)= 48x(1—0.5)x(10_4)2 =O.48(V)

A odc =
8L1 ’Cdcl 8x0.26x10_3 x 240x 10‘6 (2,14)

2.3.4 Capacitance of va Selection

C pv is designed to smooth the current from PV and keep the voltage ripple of output

of PV small to achieve good MPPT. For the capacitor C pv , the value can be calculated

in boost mode (Dbo = 0~0.5) as follows:

I'D-TS _ P-D-TS _P-D(1—D)oTS

va VPV'

 

 

 

A pv=(2-8va)= .
CP" VOdC CP" (2.15)

(>va _ P-D(1—D)-Ts _P-D(1—D)?--Ts
va Vodc' pv‘va Vodcz -va

 

 

 

where Dbo = 0~0.5, P = 1kW. It is desired that the voltage ripple (peak magnitude) is
required to be 1%, Hence

2
1%2‘3 odc =P'D bu(1_Dbu)'Ts

V 2 .
odc Vodc C pv (2.16)

 

 

va Z P'Dzbu(1—Dbu)'Ts
Vodc2'1%

The C pv has its maximum at D = 0.66. It means

20

1000 x 0.662 x (1 — 0.66) . TS _ 6.42
482 x 1% fs

 

pv2

For a switching frequency of 10 kHz, the capacitance must be no less than.

> 6.42 = 6.42

CV-
p fs 10x103

 

= 642(uF)

The highest operating voltage of C pv is about 96 V, so four 200-V/390-uF aluminum

electrolytic capacitor is chosen for C pv in case of bigger capacity solar cell.

Device selection

The maximum voltage output by PV is 96 V. However, the switch and diode shall
sustain the open-circuit voltage of the PV which is higher than the maximum output
voltage usually. The open—circuit voltage of 96 V PV is about 120 V. The switch should
be capable to sustain the voltage of 120 V. Generally, the voltage spike across switch is
considered below 50% of the maximum voltage applied to switch. So, we can selected
the rated voltage of the switch is

szitch = 96 x (150%) =144(V)

It is a buck converter from PV to do load, so the output current is not less than the input
current. The peak current through the switch occurs at maximum power of 1 kW. As
calculated above, the current ripple is 4.584 A (at D = 0.5). So, the peak current through
the switch is Ip = 20.83+ 4.584 = 25.4 [A]. The current rating of the switch should be at
least 1.5~2.0 times the peak current. So, two 150-V/50-A MOSFETs are selected for PV-
connected converter. 2 MOSF ET devices are connected in parallel per switch to reduce

the power loss.

21

2.4 DC-DC Converter Control Scheme

 

 

 

 

 

: A'A rm] Wit/‘1 - 48Vdc
PV 48~96[V] L”. u
5” Cdcl Vodc
va 3: SLJ :: -‘___ Vbattety
va
* : ~ I -

 

 

 

 

Figure 11 MPPT scheme for buck/boost converter
The converter will be controlled in the buck mode (variable input and constant
output). When the converter works in the continuous conduction mode (CCM), the input
and output voltages are related as follows:
Vodc = Dbu 'VPV (2.17)
Where Vodc is battery output voltage, VPV is PV voltage and Db“ is the duty cycle in
range 0.5 ~ 1. When the converter operates in the discontinuous conduction mode (DCM),

the input and output voltage relation is: For constant input voltage:

 

2
Vodc : D bu
V 1 1
PV Dzbu + _ (_0_dC_)
4 ILB,max (2.18)

Where I odc is buck converter output current and I L B max is max ripple in buck converter
inductor current.

Ts‘VPV

1L3,max =
8 ' L1 (2.19)

22

Where 7;. is switching period and L] is inductor value. For constant output voltage:

{ Iodc \
I 1 .
D = Vodc, ( LB,max)A Where, 1 = Ts Vodc
bu V LB,max
VPV 1-...__.0dc 2'14 (2.20)
1 VPV 1

 

 

The duty cycle is the unique control variable to adjust the operation point of the buck

converter. Based on above equation, we can get

dDbu :_ Vodc <0 01, dVPV :_ Vodc <0
dVPV VPVZ dDQl Dbuz

dDbu < 0
va

 

From above equation, we can also get the same conclusion:

Considering all equation we’ve discussed, Dbu and va have an inverse

proportion relation in both CCM and DCM. Thus a duty cycle reduction causes a PV
array output voltage increment if the battery voltage keeps invariable. Based on the
conclusion, a control scheme with maximum power point tracking (MPPT) technology
can be designed for buck converter, as shown in Figure 11. An inner current loop is

applied to limit current.
2.4.1 Simulation Result of Buck Converter

Figure 12 shows the output current transient waveform. As we calculated above, output

current is near 20.8[A] at le. Transient time is 0.03 see which satisfy our operation time.

23

 

 

 

 

Output current: SA/div ...,

 

 

 

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

 

 

 

0.038 0.068 0.18

Figure 12 Buck converter output current

 

 

 

 

u o o I
n I n I
I I I I
I a o I
D Q I .
I I I U
I
u I r n
u o I I
O I I C
u r o
I I l I
. u o I
. .
IIIIIIIIIIIIIIIIIIII IDIIIOIIII
I I l I
s r I n
n u o I
v o I I
l a a y
I t a I
r) .................. \ ................. I ................ . ................... t ..................
n a I o
u o
I c
0 I
I I
I '

Output current: 2A/div

 

 

 

 

 

 

Figure 13 Buck converter output current ripple

Figure 13 shows the buck converter output current ripple which corresponds to what
we calculated previous section. We set current ripple as 20% of output current which is
near 4.5[A]. This ripple value satisfies inductor saturation point. Figure 14 shows the
MPPT control algorithm in DC-DC converter system. By detecting PV current and
voltage, MPPT algorithm can deduce the va* voltage which value leads to maximum
power point of PV panel. va* compare with real PV voltage and(va) go through PI
controller with result of ILl* current. This value is updated every lmsec compare with

real output inductor current to achieve desired output current.

24

 

 

va* ILI *

MPPT PI PI —> PWM

IL]
7%

Figure 14 Control block diagram of MPPT system

 

 

SH

Ipv—>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

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

 

 

 

Voltage reference va*: 7V/div

oooooooooo

 

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

 

 

 

 

 

 

 

 

Figure 15 dc-dc converter voltage control waveform

FigurelS shows the voltage reference and real output voltage of buck
converter.Voltage control loop is outer loop to current control loop. Voltage loop is lkHz
period while current loop is 10times faster than voltage loop period. Figure 16 shows the
buck converter output current reference and real output current. This simulation is
achieved in order to get appropriate Kp and Ki value. Blue waveform shows the output
current reference which is near 20.8[A] and therefore red waveform follows its reference

current in 0.03sec.

25

Inductor current command
IL: 3A/div

Inductor current IL: 3A/div

 

Figure 16 DC-DC converter current control waveform

 

 

 

 

    
 

 

 

 

 

 

 

 

 

_ K
. 7 Gate signal

 

 

 

 

 

 

 

 

Figure 17 Switch signal at duty cycle 0.5(blue: triangular wave, red: gate signal)
Figure 17 shows the gate signal of both high side switch and lower side switch at the
duty cycle 0.5. This signal shows when input voltage is 96V and output voltage is 48V in

which to keep constant output voltage.

26

2.4.2 MPPT Simulation

In simulation, perturb and observation method is used to measure the PV array output
power and to change the duty cycle of the dc/dc converter control signal. By measuring
the array voltage and current using P&O method, the PV array output power is calculated
and compared to the previous PV array output power. Depending on the result of the
comparison, the duty cycle is changed accordingly and the process is repeated until the
maximum power point has been reached. From Figure 18, PV voltage is around 66V and
PV current is 15A. The implementation of the proposed method was only realized by
matlab simulink simulation. Experimental results was replaced by DC power source
which is same function as PV panel. However, this algorithm method controls a high
efficiency Buck-type dc/dc converter and performs all control functions required by the

MPPT process and battery charging, if required.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

70 MPPT Voltage Reference
601 -
50 PV Voltage
60-4
401
20 ‘__._i f
PV Current
165' , . 1 L . . , .
12*
PV Power
1000
900~
0T1 0315 012 0.25 0:3

Figure 18 Simulation results of the MPPT(P&O)

27

2.4.3 Experimental Result

 

 

Output Current:
10A/div

 

 

 

 

Diode voltage:

25V/div

 

 

 

A J_ A_L* ___.m AAA A
—_ ——

 

l l
l l
‘ t

 

 

 

 

 

“Mk——

, :
t—J lm' buy-fl

 

 

 

 

H“ 1"" I"
. l :
l
l . .
I e
. l l
M‘ VII-II) M

 

 

 

 

 

 

Signal voltage:5V/div

 

 

 

Figure 19 Buck converter output current, diode voltage and gate signal

We applied dc source with rectified grid voltage in order to replace PV source power.

Verifying DC-DC buck converter control system with this configuration was appropriate

since we only need to prove current controller of system. Figure 19 shows basic function

of buck converter with 0.7 duty cycle. Figure 20 shows experimental control block

diagram of DC-DC buck converter. However realizing MPPT control at the input side

have not applied. By changing reference voltage to output voltage instead of PV voltage,

control system give right duty cycle in order to follow reference output voltage.

V0dc* ILI *

 

Pl

 

 

 

Vodc

 

 

Pl

 

_q PWM

 

 

ILI

 

SH
"" SL

 

 

 

414%

Figure 20 DC-DC converter control block diagram

28

Figure 21 shows the buck converter inductor current and output voltage. In order to
verify lkw power, input voltage was 75[V], average inductor current was 20[A] and

output voltage was 48[V].

 

 

 

Input voltage: 25V/div ~

 

 

 

 

 

 

 

 

 

 

 

 

l

I

: Diode voltage: 25V/div Inductor current: 5A/div
:'

 

 

 

Figure 21 Buck/Boost converter 1kW power experiment

 

 

 

 

 

 

 

[ Input voltage: 25V/div Diode voltage: 25V/div
i i l *

 

A ‘Ak r .. A‘ A A _ LAAJIHA.
i w? “v ~ “7" 'U'v" ‘v T“ W'r’
.
r r u
.

ELM.

 

 

‘rF-v'r'i r"

 

 

Inductor current: 5A/div

 

 

 

 

 

 

.11....-..-..__-

 

Figure 22 Inductor current, input output voltage & diode voltage

29

From Figure 22, we can calculate inductor value by knowing current ripple and inductor

T 4
voltage. L = VL x 3T = 52[v] x M = 249.6[11H]. Inductor value is similarly same
A

10[A]

as what we have calculated at the previous section. At le, current ripple was near 5[A].

 

r
l

_ _ . A. *_‘_—A
A‘ A-..“

v—-—wwv_ V'“ 1" v vv

 

 

 

 

 

Output current: 5A/div

 

 

 

 

Figure 23 Current control transient waveform

Figure 23 shows output current which follows 18[A] current reference with 2.6[9]
load resistor. Output voltage is 50[V]. It takes 6msec to reach command reference value.
6mesc is current transient time that is appropriate for current control for PV panel since
PV don’t need in high speed feedback for MPPT.

Figure 24 shows output voltage which follows 50[V] voltage reference with load
resistor was 2.6[9]. It takes 60msec to reach command reference voltage. Transient
condition is stable since voltage loop is 10times slower than current loop. It reaches

steady state slower than current.

30

 

 

 

.. .nL- ‘ A A h“

 

 

 

 

 

/ Output voltage: 25V/div

 

Figure 24 Voltage control transient waveform

31

 

CHAPTER 3

3.1 Isolated DC-DC Converter Topology

The dc-ac converter between PV or 48 V dc bus and utility grid, termed as grid-
connected converter, is used to feed PV power into utility grid to supply ac load. On the
other hand, it also can absorb real power from utility grid and transmit to dc load when
PV power is not enough. Here, the power flow is bidirectional, so the grid-connected
converter must be designed bidirectional. Because 48 ~ 96 V dc voltage cannot direct
produce 240 V ac voltage by a single stage inverter, an additional step—up dc-dc converter
is needed to produce sufficient voltage (400 Vdc) for the inverter input. Since the utility
grid is a single-phase system, the following inverter should meet this demand. Figure 25.
shows the configuration of the grid-connected PCS. As shown in Figure 25, the grid-
connected PCS includes two stages: a dc-dc converter and a dc—ac inverter with a filter.

DC-AC
DC-DC CONVERTER INVERTER LC FILTER AC GRID

Vch 400v 340V

l

    

 

 

 

 

Figure 25 Grid-connected converter(between 48V DC bus and 240V AC bus)
Since the hoped dc voltage is 400 V which is much higher than the dc bus voltage of
48V, that is, a large step-up (or step-down for reversed power flow) conversion ratio is

required for the dc-dc converter, the use of a transformer can allow better converter

32

optimization. By proper choice of the transformer turns ratio, the voltage or current
stresses imposed on the switches and diodes can be minimized, leading to improved
efficiency and lower cost. On the other hand, the transformer can provide galvanic
isolation between the high voltage and the low voltage to satisfy safety requirements.
Since battery input voltage is almost fixed, transformer primary and secondary is equal
which in turn causes flat transformer current. Decreasing stress on the low and high
voltage capacitors. It also leads to lower losses in the semiconductor switches,

transformer, and capacitors.
3.1.1 DC-DC Circuit Selection

In general, there are two types of DC-DC converters: isolated and non-isolated.
Usually, non-isolated DC-DC converters like buck, boost, buck-boost, and Cuk converter
are used at low power applications. In the cases of distributed generation such as fuel cell
or photovoltaic system, utilizing these types of converters will result in high cost and low
efficiency. Moreover, in the circuit shown in Figure 26, isolation is needed. So this
chapter mainly focuses on comparisons of isolated dc-dc converters. The general diagram
of an isolated DC-DC converter is shown in Figure 26. For isolated DC-DC converters,
forward, push-pull, half bridge and full bridge can be considered as topology candidates.
In general, forward and push-pull converters are not suitable for high power applications.
Forward converters have restrained duty cycles and lossy resetting of excitation; in push-
puIl converters, the two halves of a center-tapped winding cannot be equal or
symmetrically wound, and the power switch on/off times as well as their forward voltage

drops are never equal.

33

 

Transformer
1"”‘3‘ DC/AC E 3 AC/DC
Capacitor - -

Figure 26 Isolated DC-DC Converter model

 

 

 

These irregularities, which exist in practice, can contribute to transformer core
saturation and result in the converter failure. In both forward and push-pull converters,
the voltage rating of the device is two times that of the half and full bridge. Though the
full bridge, compared with the half bridge, has more components, the current in the
devices and the transformer turns ratio can both be reduced to half. The current ripple in
the front end capacitor is also minimized. The voltage and current stresses on the devices
in the full bridge are the smallest in all of the four topologies. There are also several
topology options for the secondary side of the DC-DC converter, such as diode rectifier,
controlled rectifier, voltage doubler and controlled voltage doubler. All these four
topologies have the same device voltage rating. For the secondary side, which is directly
connected with the inverter module, the major concerns for the topology selection should
be efficiency, cost, the resulted transformer turns ratio and power control flexibility.
Voltage doubler circuits will give smaller transformer turns ratio. But it will also lead to a
large current ripple in the output capacitor. The current ripple will result in low efficiency,
and possible overheat of the capacitor and reduced lifetime of the capacitors. So in the
applications of distribute generation, a voltage doubler should only be used when the
power is less than 10 kW. By using the active switches, there will be more flexibility in

controlling the power flow and voltage. Active devices will also create more possibility

34

in the control the current waveforms over the transformer, thus the current stress on

switches will be depressed. Furthermore experimental results also show that active

control of the secondary side, in order to achieve a higher voltage transfer ratio than in

the non-controlled case, will little lower the system efficiency.

3.1.2 Controlled Rectifier and Uncontrolled Rectifier

The circuit structures of controlled and non-controlled case are shown in Figure 27(1)

and (2).

 

voltage voltage

current

2‘
fl
5
I-
n.
E“
a
'U
C
O
O
O
0
L
0
E
L.
O
.-
0
C
W
L
|.

 

 

 

 

[1
Ph hifl
ase S (3)

 

voltage

voltage

Transformer secondary Prlmary
current

   

<
1:

 

 

Vs

:1

4° l”—l’ H

 

Itr

il—lll—HL~
:4'L'4lL—

 

AVA\

Phase shift (4)

 

I
|
l a

Figure 27 Two configurations of the secondary side in the isolated dc-dc converter

The correspondent voltages and current waveform over the transformer are shown in

Figure 27(1) and (2). It is found that the voltage doubler and full bridge rectifiers yield

the same voltage and current waveforms. The only thing that will make a difference in

the current and voltage waveforms is whether the secondary is controlled or not. So, the
controlled voltage doubler and uncontrolled full bridge are shown in Figure 27(3) and(4),
respectively.

A. Uncontrolled Rectifier

In uncontrolled case, (9 is no longer controllable, the relationship between Hand VS,

 

 

is found as
V — V
= 7r- P S (3.1)
2 - V P
where VP > VS .
The power transferred over the transformer and RMS value of the current are
Vz—Vz-V
Pz—l—IVP‘Is’dt=( P S) S (3.2)
T 8- VP - f oLS

1'—-1Tflm—$my_%5 an
rms“ s "' '
T 0 12on-f-LS

 

It can be seen that 6 decides both the secondary side voltage, Vs , and the transferred
power, P . For the diode rectifier, VS cannot be equal to or larger than VP, thus, the

controllable ranges of output voltage and power are quite limited.
B. Controlled rectifier

In Figure 27, Ls is the leakage inductance of the transformer. The full bridge and
voltage doubler are controlled to switch at high frequency to produce two square voltage
waveforms with a phase shift, (9, as shown in Figure 27(3). The two square-wave

voltages are impressed across the both sides of the transformer, respectively. So the

current over the leakage inductance would look like the I ”as shown in the Figure 27(3).

36

As I ,r is a function of wt , where w is the switching frequency, the current

waveform can be analyzed as following: At 0 radian in Figure 27(3), the primary side
voltage becomes positive while the secondary side voltage remains negative, creating a
positive voltage across the leakage inductance making the current increase until wt = 6 ,
where the secondary voltage changes polarity from negative to positive. The current at

this point can be written as

VP + VS 0 + i,,(0), wt = 6 (3.4)
st

 

itr (6) =

where Vp and Vs are the amplitudes of the transformer primary side and secondary side
voltage. With proper control, VS can be smaller, bigger or equal to Vp . The peak current
will be minimized when Vs =Vp is met.

At wt = n+6, the secondary side voltage becomes positive while the primary side
voltage remains positive, making the voltage across the leakage inductance negative,
allowing the current to decay until wt=7r, where the primary voltage changes from

positive to negative,

 

it, (7:) = V1139 (7r — (9) + it, (0), wt 2 7r (3.5)

At this point, the primary side voltage becomes negative while the secondary side voltage
remains positive, creating a negative voltage across the leakage inductance, making the
current keep decreasing until wt = n+6, where the secondary voltage changes polarity

again,
W S

After the secondary voltage changes polarity, the secondary side voltage becomes

37

negative while the primary side voltage remains negative, making the voltage across the

leakage inductance positive allowing the current to increase to point 1, where

VP—Vs

 

it, (27!) = - (7r — 6) + i” (7: + 6?), wt = 272' (3.7)

Because is (6) = —is (7r+ 9) and is (7!) = —is (27f),the exact relationship between the current

and voltages can be found as equation (3.8) and (3.9) by solving equation (3.5)-(3.6)

21/125 ‘1' (VS — VP )7!

 

 

 

i 6? = 3.8

1,( ) 2st ( )

, 2V (9 + V — V 7:

11,47?) 2 S ( P s)

2st (3.9)
From the representation of is (wt) , the power transferred through the transformer can be
found as:
P: ‘l'g'VP'VS’ (;”+6)
2 (7t 'f'LS) (3.10)

Threesconclusions can be made based on the Figure 27(3) and the equations obtained
above. First, the power is transferred through the leakage inductance. Second, by
controlling the phase shift of the voltages between the two sides of the transformer, the
power transferred over the transformer as well as the secondary side dc voltage, Vs , can
be precisely controlled. Third, if the system is controlled to have Vs sz , the current
wave form would be rather flat, which means the peak current can be smaller. In
summary, for the controlled rectifier case, the voltage angle difference, 6, and the
secondary side voltage, Vs , are both controllable by properly coordinating the control of
the primary side and secondary side. The power transferred over the transformer, P , can

be variable at fixed secondary dc voltage, Vs . Vs can be equal to or larger than the

38

primary voltage Vp .
3.1.3 Isolated DC-DC Converter Control

The topology of the isolated dc-dc converter is shown in Figure 28. For isolated dc-dc
converter, the primary side refers to low voltage side and the secondary side refers to high
voltage side. And, forward power flow will be defined as from low voltage side to high
voltage side. The primary and secondary voltages are fully controlled with a phase shift.

So, the power transferred using this control strategy is given by the following equation.

DC-DC CONVERTER
Vch 400V

 

 

J

 

 

 

 

 

 

 

JE'} 5‘} CchT is J?

Figure 28 topology of the isolated dc-dc converter

 

Vch(V::1H)'¢'(7r—¢)

2-7r-fs-Ls-7r

 

(3.11)

Where (If is the phase shift angle of the secondary side, fS is the switching
frequency, L5 is leakage inductance of the transformer, n is turn ratio of transformer and

Vch, Vch are low voltage and high voltage side of dc-dc converter. Based on (3.11), we

can get

39

V
:15 Vch[———— M ) (rt—21>)
d¢= 2-7r-fs-Ls-7r

 

 

 

 

 

 

 

(3.12)
dVchZP-Z-n-fs-Ls-rr. -2 _ -1 Po2ozofsoLS-7r. ._1 _ -2
ds (Vdcfls ¢ (r: 11) + [WI] 16 (z ¢)
2n 2n
=P-2-7r-fs-Ls-7r. 2¢—7r
(chfi) ¢2(7r—¢)2 (3.13)
2n
dVch zP'z'”'fs'Ls'”.¢—2(fl,_¢)—l+P'z'fl‘fs'LS’”.¢—I(fl_¢)-2
d¢ (Later) (Van)
2n 2n
:P-2-7r-fs-Ls-7r. 2¢—7r (3.14)

 

(fig—CL) Per—(7)2
n

Since the phase shift varies in the range of ( 0 , %) under forward power flow, thus

11:“) (3.15)

d¢

Considering (3.12), a phase shift increase causes a transferred power increment. We
have to use isolated dc-dc converter to keep the 400V high voltage side (HVS) dc link
voltage constant if the whole system works on stand-alone state. Here, this situation is
considered firstly. Now, let us consider the dc voltage over capacitors Cdc2 which is
inverter dc link capacitance. Generally, the voltage of the Cch will increase when the
input power (Pin) is larger than output power (Pout). This means the voltage of the Cdc2
variation indicates power change or the difference between Pin and Pout. So, the voltage

control loop can be used to substitute the power control loop. Figure 29 shows the control

40

scheme of isolated dc-dc converter when working on stand-alone state (forward power
flow). Here, dc-link voltage Vch is feedback to obtain the phase shift to control the
power from primary side to secondary side. The dc-link voltage reference will be about
400 V. The isolated dc-dc converter will carry out the different control strategy in the
situation where the whole system is working on grid-connected mode. Because the HVS
dc link voltage Vch can be controlled by the inverter which will be discussed later, the
isolated dc—dc converter just need control the penetrating power by adjusting phase shift
angle between the full-bridge legs and the voltage doubler leg.

DC-DC CONVERTER
Vch 400V

I l Cdc2
J1: JH NL NH 1: E CH”?

'3

JP? J53: Cdczj: a CLJ

 

 

 

 

‘ l
l E Vch

 

 

 

 

 

 

 

 

 

Vch
' 7 Vch *

 

Figure 29 Control scheme for isolated dc-dc converter

Figure 30 shows power transfer curve of utilizing leakage inductance of transformer.
Since transformer leakage inductance is too small that we added additional leakage
inductance with lluH. By adding this leakage inductance, we can realize more stable
control. Deriving from power transfer formula, we can transfer lkw power at the 10usec
phase shift delay between low voltage side and high voltage side. If primary referred to

secondary voltage is negative, which is negative phase shift, power is transferred from

41

secondary side to primary side. By giving Phase shift control, we can control power

either forward or backward direction.

Transferred
o power(W)_

—90° —45°

. I"
U, u:

p—u

x 103

 

 

 

/ \

 

 

 

 

. kl!

 

 

 

 

 

 

 

y.____u__

 

 

 

“w

 

 

\

/

 

\xi/

/

 

 

‘7

 

 

 

 

 

 

 

 

= 0 45° 90°(25,usec)18'0°

Phase shift angle
(radian)

Input voltage = 48] V] Output voltage=400V
Leakage inductance=lllqu

V - V
= __H_L . 5 ( 7, ._ 5)

W ° Lleak ' 7r

*l[kW] z 6u sec

*0

ransfer

Figure 30 Power transfer curve of utilizing leakage inductance

3.2 Design and Development of DC-DCFull-Bridge Converter

3.2.1 Turns Ratio Selection of Transformer

The dc-dc converter stage is a typical dual-active-bridge dc-dc converter. The LVS

full bridge generates a square-wave voltage applied to the primary of the transformer, and

the HVS half bridge also generates a square-wave voltage applied to the secondary of the

transformer. Phase shift control and duty cycle control can be applied to obtain desired

performance. The current waveform is determined by the phase shift and voltage

42

relationship of Vch and Vch. The current through transformer and switches would
have minimum peak when the voltages of primary and secondary are almost same in
primary side (or secondary side) referred. The high voltage Vch is almost constant (400
V) but the low voltage Vch is 48[V]. In order to decrease the current spike, the turns
ratio is selected between 48/200. Here, the turns ratio is fixed at 1:4 (NI : Nh) with using

voltage doubler circuit.
3.2.2 Leakage Inductance Selection

Figure 31 shows the equivalent circuit referred to LVS of dc-dc converter.

 

 

  
   
 

DC-DC CONVERTER
Vch 400V
‘ , l:4=l:n=NL:NH - o
J“; J NL NH TCdcz fl

 

l
Vch Vch

.1335 J%} 7::Cch Ml}

v

 

 

 

 

 

Figure 31 Simplified equivalent circuit referred to the primary(low voltage side)

If only phase shift control is applied, the amount of power transferred through the
transformer can be calculated as

_ Vane—”Vii!” )-¢(n —¢>

p _ (3.16)
2asts7r

 

where ¢ is the phase shift between the two square voltage waveforms generated by the

43

LVS bridge and HVS bridge, respectively, f5 is the switching frequency, Lsis leakage
inductance of the transformer. Suppose the maximum power through transformer (dc-dc
converter) is PmaX , the expected phase shift angle at Pmax is¢,the Ls can be computed

as

 

_ Vch(K§‘j~1’i)-¢(n—¢)

Ls 2
27’ f SP max

(3.17)

So, the leakage inductance Ls shall not be greater than L when phase shift is n/2

s,max

and Vch is minimum.

 

 

 

 

V
L Vch ($636,111 ) (3 18)
s,max = -
8 ‘ f SP max
If the switching frequency is 10 kHz, Pmax is 2 kW, then
481?) 60uH
Ls,max = 4 n = = lSuH
8-1-10 -2000 n (3.19)
Here, the leakage inductance of the transformer is selected as 11 pH. And then
2
LS'Zfl ‘fS'Pmax 86.85
Va” 2 Vch = W — ¢>
(m) ° ¢(7r - ¢)
2" (3.20)

Thus, ¢= 0.78 when Vch = 48 V

3.2.3 Core and Winding Selection

The selected core for the transformer is Powerlite C core from Metaglas. The most

benefit from this core is ultra low power loss. The core has very high saturation flux

44

density. Like inductor, the high-frequency transformer can also be designed using area

product (A p) approach. Firstly, Ap is calculated as

_ 1000108 _ KPO _ 0.00528x2000x108 _ 52 8(cm4)
4778fKu Bf 2000><10><103 (3.21)

 

 

Where K = 0.00528 for E-U-I core, P0 is the output power in (W), C (= NJ) is

current capacity in (cm2 / A) which is the reciprocal of current density, n is transformer
efficiency, B is flux density in (Gauss), f is operation frequency in (Hz), and Ku is
winding factor. Here, we assume 1] = 90% for transformer. Secondly, the E core is

selected. from Magnetics Inc. 3 pairs of E cores are used. to reduce number of turns.

Selected core area product A p =55.7[cm4] , core area Acore=3.92[cm2] and window

area Aw: 8 [cmz]. In order to reduce the core loss, the magnetic flux density should be

small. Here, it is 2000 Gausses. Then the number of HVS turns, NH VS can be calculated

using Faraday’s Law.

 

8 8
NHVS = VSOCXIO = 200x") =21.25 use 16(turns)
K BAC xf 4
f 0re,efl 0 4x2000x3x3.92x10
l l
N =—-NHVS=—xl6= turns
LVS n 4 4( ) (3.22)

We could chosen N HVS = 20 but that would have made N HVS = 5,which means

that the secondary would probably either be a single section of Slayers or Ssection of
llayer. This would have greatly reduce the flexibility in designing a sectionalized
transformer winding.

The core loss per unity volume at B=0.2362T is

45

PAR = a - chd = 0.074x101'43 x 2.3622‘85 = 23(mW / cm3) (3.23)
The actual total core loss is Pcore = PL, R - Ve = 23x48x4 = 4.4(W) .The ac

resistance of LVS winding and dc resistance of HVS will be very large if there are just
one primary section(with 4 layers) and one secondary section(with 16 layers). So the
transformer winding should be reselected and rearranged. Generally, the transformer
winding can be interleaved to reduce power loss. The 4 LVS layers(tums) are split into 2
sets of 2 layers.And the 16HVS layers are divided into 3section,with the central section
having 8 layers and the outer two section having 4 layers. Window length is about 5.6cm,
we selected standard copper foil with standard width of 5.080m. The resistance per turn

of the windings in bout HVS and LVS shown in following equation.

p-MLT _ 2.2x10_6 x33
ACU,HVS 5.08 x 0.0254

 

 

rHVS@0.01" = = 5.63X10_4(Q/turn)

p-MLT _ 2.2x10‘6 x33
ACU,L VS 5.08 x 0.0533

Where 0.01” is thickness of copper foil at high side voltage and 0.021” is thickness of

= 2.68x10_4(§2/tum) (3.24)

 

 

rL VS@0.021" =

low side voltage. MLT is mean length per turn and ACU is copper foil width times

copper foil thickness. The equivalent resistance in both HVS and LVS is

RHVS@0.01" =rHVS x(l.lx8+lx8)=5.63><10-4 x(1.lx8+lx8)=9.45x10—3(Q)

RLVS@O_021.. = rHVS x1.2x4 = 2.68x10‘4 xl.2x4 =1.286x10_3(§2) (3.25)

. 2 2
The total copper power loss 18 PCU@0.021"+0.01"=1 HVSRHVS +1 LVSRLVS

=10.52 x 9.45 ><10_3 + 422 ><1.286x10_3 = 3.31[W] (3.26)

46

We consider of 1.5 for our application. So the modified total copper power loss is
PCU@0.021"+0.01" =1-5PCU =1-5-3-31= 5[W]
Total loss from calculation would be following value.
Ptotal = PcoreJoss +1) cu,loss 510ml] (3.27)
F igure32 shows the transformer with 230 uH magnetizing inductance and
0.1 uH leakage inductance. This small leakage inductance may cause some problem in
controlling the power transfer. Small leakage inductance can cause unstable power
transfer to high voltage side which brings controller to be more sensitive to feedback
system. Therefore, we made additional inductor beside transformer to achieve higher
leakage inductance with 11 uH . Our system needs to transfer maximum 2KW from dc to
ac side. Leakage inductance with additional wire satisfy enough margin to control the

power flow of dc-dc converter.

    

‘\ ‘ /’1u’

F igure 32 DC-DC converter transformer design

47

 

 

Transformer
current: 20A/div

 

 

 

 

 

 

Primary side
voltage: lV/div

 

 

 

 

 

 

 

 

 

 

 

 

 

1.‘
Figure 33 Transformer short circuit test (leakage inductance test)
I : s
1 I ii \ d

Primary side
voltage: 25V/div

 

 

 

 

[lam

 

 

. Secondary side
voltage: 250V/div

 

 

 

 

1

5‘5"} r “fig.
5' ' l "3'

 

 

  

i"; 4'7: 3 " . Vi Kg ',' if Transtormer
f ‘ V magnetizing
current: 5A/div

 

 

 

 

 

 

Figure 34 Transformer open circuit test (magnetizing current)

3.2.4 Device Selection

The maximum voltage across the switches on LVS is the open-circuit voltage (50[V])
of DC-DC buck converter. The voltage rating of the switch should not be less than it. The
voltage rating of the switches on LVS is selected at 150 V. The maximum voltage across
the switches on HVS is the dc-link voltage of 400 V. So the rated voltage of the switch

on HVS is selected as 600 V. As analyzed above, the maximum peak current through the
switches of LVS is 58.7[A] occurring at Vch = 48[V]. The peak current through HV S

switches is Ip(HVS) =58.7/4=14.67[A]. The current rating of the switches is at least 1.5 ~
2 times the peak current. So, 2 MOSF ET modules are selected for LVS full-bridge and
two MOSFET are paralleled per switch. And a 600-V/75-A IGBT—IPM is selected for

HVS half-bridge.

3.3 Simulation & Experimental Result

3.3.1 Simulation Result

From Figure 35,'before turning on the inverter, dc-dc converter charging the capacitor
on the high voltage side before it reaches 0.5 sec. Turning on the inverter at 0.53ec,
voltage on the high side drops suddenly, then turn into stable voltage after controller start
to get reference voltage of high side. From Figure 36, after charging the high side voltage
up to 360[V], Phase angle keeps almost zero in order to block current from dc side. After
turning on the inverter phase is keeping constant phase angle in order to transfer the
power that load needs as shown in Figure 37. When load(80 Q) is connected, 5A output

current is flowing.

49

 

 

 

 

 

 

 

 

 

 

 

 

 

__V V V “VA' 7
Inverter : 3 Inverter
Vref dclink 5 DC link voltage
100V/div .. ,, .. . . i .. l00V/div
0 0.2 0.4 0.6 0.3
(Time/sec)

Figure 35 DC-DC High side voltage & voltage reference

 

 

 

 

 

 

Transformer Primary voltage

primary voltage after leakage

20V/div inductance
(20V/divl

 

 

    
  
 

 

Fl

 

 

. Transformer
current(5A/div)

Ill
41:}

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

 

 

 

 

Ell—Jl:

Figure 36 Transformer Primary voltage & Secondary voltage
before turning on the inverter

 

 

 

50

 

 

Transformer Transformer

 

 

 

 

 

primary voltage secondary voltage
V (20V/div) ~— (250V/div) _.

 

        

 

 

Transformer ........ ,...... ......, ......................
secondary ‘ '
current(SA/div)

 

 

 

 

 

Figure 37 Transformer voltage & current afier turning on inverter

3.3.2 Experimental Result

Experimental results of 2 kW DC-DC converter is shown in Figure 38 and 39. In the
experiment, the input power to the dc-dc converter is supplied by a 48V dc source. The
transformer turns ratio is 1:4. The secondary side is a voltage doubler with two IGBTs
which is one phase part of 3phase IPM. The DC-DC converter is connected to a inverter
module. Before turning on the inverter, two phase leg of IPM(inverter side) is not
switching while the one phase leg of IPM(dc-dc converter side) is using free wheeling
diodes to establish 400V in dc-link. Once ac load is installed, one phase leg of IPM start
to switching for the phase shift purpose, (secondary side voltage refer to primary side
voltage) transferring power to ac load through inverter. It can be concluded that the full
bridge inverter cascaded with a high frequency transformer and controlled switch would

be the Optimized circuits for the isolated DC regulator in the PV inverter system. The

51

controlled rectifier has a larger controllable output voltage and power range. In Figure 39
and 40, the phase shifting of the voltages over the transformer is clearly shown; the
transformer current is not exactly flat. This reason is coming from voltage difference
between leakage inductance is exsiting. In Figure 39, the secondary side voltage is 405V,
which is little bit higher than desired reference voltage. The primary side voltage is 48 V.
This two voltages shows an equivalent boost ratio of 8.43, whereas the transformer turns
ratio is only 1:4 with voltage doubler. This verifies that the high side voltage can be

further boosted and precisely controlled by adjusting the phase shift angle.

 

 

Primary voltage: 50V/div 1 Primary current: 20A/div J

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r4
EMU“? ........
z—i
Phase shift
angle
Secondary voltage: Secondary current:
250V/div ' 10A/div

 

 

 

 

 

 

Figure 38 the experimental result of the dc—dc converter.

52

 

Primary voltage: 50V/div—l I Secondary voltage: 250V/div

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

Primary current: 20A/div l ....... 1 Secondary current: 10A/div

 

 

Figure 39 the experimental result of the dc—dc converter.

53

Chapter 4

4.1 Inverter Design

4.1.1 Overall Inverter Configuration

The inverter circuit converts a DC input (as would commonly be from a distributed
energy source afier proper conditioning) into a grid-suitable AC output. The DC input
voltage is nominally 400 V i 5%. Conversion is done by a power electronics bridge
circuit that operates to produce a pulse-width modulation (PWM) AC output, which for
this model will be nominally 240 V AC. To allow a seamless transition between utility
connected mode and stand-alone mode, the inverter must transition its operations from
current control to voltage control. The power circuit can be divided into four essential
parts: (1) DC input stage; (2) bridge stage; (3) AC filter stage; and (4) utility stage. A
number of small resistances are included throughout the model to make the model more
realistic and to facilitate convergence of the model in experiment. To meet the
requirement of the single-phase utility grid, the dc-ac inverter stage is realized using a
three-leg PWM converter. It is connected to the utility grid through the LC filter. Here,
they are termed as the grid-tied-VSC. It is noticed that the dc-ac inverter stage shares a
common two-switch leg with the dc-dc converter stage. There are two states for the dc-ac
inverter stage: when injecting real power into the grid, the PWM converter is operated as
a single-phase PWM inverter to transform the dc voltage into ac voltage. When absorbing
real power from the grid, the PWM converter is treated as a single-phase PWM rectifier
to transform the ac voltage into dc voltage. Although the ac load power is about 2 kW,

the capacity of the grid-tied-VSC is near 2 kW since the PV power is only about 1kW

54

and the dc load is about 1 kW. Actually, the capacity of the grid-tied-VSC just need be
designed to match the demand of the dc load or PV power production. According to the
specification, the battery is only charged by PV. Here, we suppose the battery has the
same power of 1 kW as the dc load. So, for the grid-tied-VSC, the extreme case is that
both of PV and battery supply to the ac load when the dc load are not in operation. That is,
the grid-tied-VSC should transmit 2 kW power from dc side (1 kW from PV and 1 kW
from the battery) to the ac grid at the extreme case. However, in order to ensure the unity
of the AC input power factor of the gateway grid unity, the grid-tied-VSC may need work
as a VAR compensator. Consider the worst case when the inverter needs to compensate
the 2 kW ac load from the power factor of 0.7 to nearly 1. Plus, it still supplies 1 kW
active power to the ac load if no dc load absorbs power from the PV. The required

apparent power is:

2
S = \[12 + [62—7— x \/1 — 0.72) = 012 + 2.032 = 2.27[kVA]

 

(4.1)
If the ac grid is temporarily unavailable and PV and battery is.designed to support the
load as they can. Assuming the condition that the 2 kW is all consumed by the ac load,

the maximum apparent power passing through the inverter can be:

S = M = 2.85[kVA]
0.7

Based on the above two cases, the capacity of the grid-tied-VSC is designed as 2.5[ kVA]

and LC filter is adopted in the grid-tied-VSC. Given that the nominal output voltage

(utility grid voltage) V0616: 240 V, the maximum voltage = 252 V and the minimum

voltage is 220 V. So the nominal output current I 0: 2 kW/240 V = 8.3 [A], maximum

55

output current is [0,max =2 kW/220 V=9.09[A]. And the dc-link voltage (input voltage)

should be selected as Vindc =400V. For the grid—connected PV system, the total harmonic
current distortion at the point of common coupling (PCC) shall be less than 5% of the

fundamental frequency current at rated output. So, how to design the LC filter is a very

important issue.

4.1.2 DC Input Stage

The inverter is designed to accommodate a DC input from a 400 V DC source. In the
model, the input consists of a controlled DC voltage source. The design also incorporates

an input filter capacitor of 1500[uF] to reduce input voltage disturbances.
DC-link capacitor (C dc2) selection

The dc-link capacitance is selected to guarantee that the 2nd harmonic is controlled

below 1% of the rated dc voltage. So

2
Cdc2 2 m ‘Ioac = Voac '1 oac
2 4w0-72-V0dc-1% 4w0-x/2-V0dc-l%

 

= 240)‘ 8'3 2 =12.67x10‘4 =1267[uF]
272 x 27rx 60x 400 XI% (4.2)

As stated in previous section, the dc-link capacitance must be designed taking into
account that: during the delay time of the voltage control loop, the output power demand
must be sustained by the dc-link capacitor energy. Hence the value of dc-link capacitance

turns out to be:

56

TrAPmaX
2ViNdCAVindC,maX (43)

 

C dc22

where response time Tr of the voltage control loop, is in the order of a few modulation
periods; APmax is maximum expected variation of the output power. Give that Tr = 10Ts

= 1 ms, APmax = 2.0 kW, voltage ripple below 1%, then

10"3 x 2000

2 = 12.25 x1040?) =1225[uF]
2x400x(400x1%)

 

C d62
(4.4)

4.1.3 Bridge Stage

The bridge stage is made up of IPM that are controlled using a unipolar switching
scheme. In a unipolar switching scheme, each leg of the bridge is controlled as a separate
pair. The two switches in each pair are controlled in a complimentary fashion. In other
words, when the top switch in a leg is closed, the bottom is open, and vice versa for the
next switching interval. The control for each leg of the bridge is accomplished by
comparing a control signal (on A) with a triangular waveform (sigl). This results in the

following logic signals being applied to the gates of the IPM (using leg A as an example):

onA< sigl :Al on andA2 off

onA > sigl. : A1 ofiand A2 on

The control signal (on B) for the B leg is the negative of on A, but is compared to the
same triangular waveform (sigl). This is shown below in Figure 40 with the frequency of

triangular waveform significantly reduced for illustrative purposes.

57

 

Figure 40 Relationship between A and B leg control signal
& triangular carrier waveform

This results in the following gate signals being applied to the leg B IPMs:

On B (or—onA) < sigl :81 on and BZ off
On B (or —on A) > sigl :Bl ofl°and BZ on

Figure 41 illustrates the results of this switching scheme for leg A. The bottom
graph shows the comparison of the control signal with the triangular waveform and the
other two graphs show the control logic being applied to the IPM gates. The IPM

switches are closed when their logic inputs are high.

WMUMJUUUMUUULH

 

Figure 41 . PWM Switching Scheme for leg A of inverter

58

When on A is less than sigl, A1 is commanded on (top graph) and A2 is off (middle
graph). The opposite also holds true, as shown. Combining the effects of switching the A
and B legs in this manner results in four output voltage states:

A1, B2 on : output = DC input

A2, B1 on : output = -DC input

A1, Bl on : output = 0

A2, 82 on : output = 0

Notice that when either both the top or both the bottom switches are on at the same time,
the output is zero and current flows through the forward biased diodes. Controlling the
IPMs in this manner generates an output with an average that follows the control signal
on A. Moreover, as long as the frequency of the control signal is much smaller than that
of the triangular waveform, the output can effectively be a scaled mirror of the control
signal. Thus, a controlled inverter output can be achieved by precisely controlling the

signal on A.

The on A signal is either generated using a sine function (with variable magnitude,
phase, or frequency components) or by a comparator output depending on whether the
inverter is in current or voltage control mode. Details on these modes and how they affect
on A are discussed in the next subsection. The result of this switching scheme is
illustrated in Figure 42 which shows the PWM positive and negative pulses in green and
red, respectively, along with their average values traced in blue showing it to be a
sinusoidal signal. Another thing to note is that the actual hardware used for the IPM
bridge stage can be configured to operate as either a single phase inverter or a three phase

inverter, as they actually consist of six IGBTs.

59

 

 

 

 

Figure 42 Combined PWM Switching for inverter
During single-phase inverter operations, two of the six switches are not the part of
inverter operation but these switches operate as phase shift fimction with full-bridge

converter in the isolated DC-DC converter.

4.1.4 AC Filter Stage

The bridge output goes into the filter stage, which contains the energy storage
elements necessary for circuit operation. The 2mH inductors and parallel, 7.5uF
capacitors are sized in order to achieve desired voltage and current ripple characteristics.
The output is a single phase, 60 Hz, 240 V AC waveform. PV system has certain source
referencing (grounding) requirements for voltage and current sources that need a ground
on the negative side of the input DC voltage. This arrangement does not accommodate
utility grounding, therefore improper current paths are created. To prevent this from
affecting the modeling performance, an idea] 1:] turns ratio transformer is placed on the

output. Actual system hardware implementation will not include this transformer.

60

Inductance and capacitance selection of LC filter

The base values can be calculated as

2 2
R =V =240 z30[Q]’ Cb: l =__1___
P 2000 W0 'Rb 27r-60-30

 

 

= 88.4[uF]
(4.5)

The selection of filter capacitance Cf is a trade-ofi‘ between reactive power in Cf and
equivalent inductance Lf. Filter capacitance Cf should be limited by the amount of

reactive power absorbed in the capacitor.

C
[.%J S O-I'Cb —) 2x0.1x88.4 =17.68[uF]
(4.6)

Firstly, the inductance Lf is chosen to limit the ripple at the VSC output current to 10% of
the rated amplitude value. For the filter inductance design, the VSC can be treated as a

buck converter at the switching frequency. So, we know

 

 

 

 

For unipolar PWM
(2Lf) Z Voacm ,Vindc ‘Voacm = Vindc _ Voacm .D._l_ (4,7)
2 ‘A’Lf,max Vindc ’fs 2 'AlLf,max f5
For bipolar PWM
(2Lf)2 Vindc + Voacm , Vindc “Voacm = Vindc "Voacm,D,_1_ (4.3)
2 'AiLfJnax 2' Vindc 'fs 2 'AiLf,max f S

For a switching frequency of 10 kHz, the filter inductance Lf must be no less than

240x72 400—240><\/2X 1

 

 

2L 2 . =l.367 mH

( f)for unipolar 2x13.3x\/2x10% 400 104 [ ]

(Hf) . 2 (400+240x\/2) .400—240xJ—2—x 1 21.82[mH]
for blpolar 2x11.5x \f2x10% 2X400 104

61

On the other hand, the inductance should not be selected too large to meet the demand of
the current tracking capability and the demand of the power transmission.

L f 2 910.5[uH]

Through out numerous experiment, Lf is selected as near 2 [ml-I] to make smooth ac

sinusoidal wave form. The resonant frequency if LC is

1 1
”#fo 27t\/2x10_3 x7.5x10‘6

 

 

f res : = 1-29lkHZ l

The resonant frequency is in the range nearly between twenty times the line frequency
and about one-tenth of switching frequency. The utility stage consists of a 240 V AC
source and a circuit breaker. The breaker is controlled with on/off switch facilitate
opening and closing during a simulation run. The utility source is a constant magnitude,
frequency, and phase AC source with some nominal series impedance. The state of the
breaker is represented by switching between logic level 0 or 1, depending on whether the

breaker is open (0) or closed (1).

4.2 Inverter Control Methodology

4.2.1 The Overall Inverter Control

The basic inverter controls can be broadly classified into two types: current control
and voltage control. When the inverter is connected to the utility, the grid controls the
amplitude and frequency of the inverter output voltage and the inverter itself operates in
the current control mode. Other types of inverter controls such as power flow control and

VAR/voltage control can be derived from the basic current control. In contrast, in case of

62

grid failure (grid faults, maintenance, etc.), the connected loads have to be supplied by
the inverter. In such a scenario, which is often referred to as islanded operation, the
inverter has to maintain the amplitude and the fiequency of the voltage so that the
connected loads are not affected by the utility interruption. The inverter operates in
voltage control mode for such scenario providing the reference voltage and frequency.
For inverter, forward power flow will be defined as from the dc-link. In the stand-alone
mode where the inverter is only connected to passive loads, the inverter shall control the
output ac voltage (Va and Vb) constant. The laboratory hardware setup consists of several
components as shown in Figure 43: inverter power and control circuits, DC voltage

source, utility grid, resistive—inductive-capacitive (RL) load bank, and electronic load

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'

bank.
DSP contorller ‘
: fl 1
g Auxiliary power Gate Driver Protection i
i 1
1 1 1
.. 4} i
' l
J J J i
i ii I AC
DC i 2: i .
side i Side
,3, a} a? I
E 1

 

 

 

IPM

 

Figure 43 Inverter control system

63

Powerex Intelligent Power Modules (IPMs) are advanced hybrid power devices that
combine high speed, low loss IGBTs with optimized gate drive and protection circuitry.
3phase,75A/600V IGBT type inverter and has monolithic gate drive & protection logic.
Detection, protection, status indication circuits for short circuit, over-temperature and
under-voltage is performed. The DC bus filter capacitor, voltage and current sensors,
auxiliary power supply, and gate drivers are on a common heat sink. An external DSP is

selected. for this particular design to have more flexibility.

The control board is designed with a DSP 2407 which contains the TMS320F2407
DSP by Texas Instruments. To simplify code development and shorten debugging time, a
C2000 Tools Code Composer Studio (Texas Instruments) driver is provided. The inputs
to the control board are DC bus voltage, phase currents and fault signals from the PV
system. Voltage and current measurement signals from the external sensors and mode
selection and startup signals from the external central processing unit (CPU). The main
outputs of the control board are the gate drive PWM signals for the PV system.
Additionally, the control board also provides enabling signals for the PV system, PWM
status signals to the external CPU. The local controller is designed to provide control and
communication functions for the associated gate signal. As the DSP module has
sufficient calculation power, we preferred to carry out all control calculations inside the
local controller in addition to the PWM generation for the switches. The control board
has 2 main input port for sensing voltage and current from those sensors. The power is
supplied to the control board by an external AC/DC converter that converts 240V, 60 Hz
AC into 15/48V DC. The signals from the voltage and current sensors are fed directly

into the control board, as it has several voltage and current conditioning systems that

64

convert the measured voltage/current into DSP input compatible quantities. The
connection between the PV system and DSP is obtained using a D-sub 25 pin connector
with ribbon cable. The higher level controller in the CPU is currently connected to the
DSP board through JTAG-compatible parallel port. An annotated picture of the control

board is shown in Figure 44.

External voltage and current sensors input

' v .511
U '. 1':
:il ll”; 7:11“ fag" :ir'fipn,‘

JTAG port ,
‘ I/O PORT

5/15V power ;
supply '

 

Communication
port

Figure 44 Picture of control board with input/output connection

Initially the DSP code was written in C programming language with TI’s Code
Composer Studio—all the results that are presented in this section are obtained using these
C—based control codes. The C codes were compiled and loaded into the DSP on-chip
random access memory (RAM) using Code Composer Studio. Then the controller was

activated using the serial interface. The program code is interrupt driven and runs at a

65

fixed sampling frequency of 10 kHz based on the interrupt set up. For the inverter
platform, the switching frequency was set at 10 kHz. Therefore, the service routine ran

once in each PWM cycle and was used to compute duty cycles for the next cycle.

4.2.2 Voltage Control

For the voltage control mode, the IPM switches are controlled using biipolar pulse-
width modulation (PWM) switching such that the inverter output voltage follows the
reference voltage. The voltage control loop block diagram is shown in Figure 45. The
voltage control system consists of a negative feedback loop containing a proportional
integral (PI) compensator We compared the actual voltage to the reference signal and the
error was then fed to a proportional controller. In this case the reference value is unity, as
it is assumed that utility voltage is constant. The output of the controller was then scaled

and added to a feed forward loop. Then V ref signal is fed to the gate driver block
where it is compared with a triangular waveform to determine the duty cycle. As V ref

increases, the average duty cycle of the gate driver circuit also increases raising the
average output voltage of the inverter. The final output of the controller is a duty cycle

value (D), as given by equation.

 

 

V K x V —V- )
D=[0.5+ ref ]+ v ( ref mv

2 x Vdc 2 x Vdc

(4.10)

where, V ref is the reference voltage, Vim, is the inverter output voltage, Vdc is the

input DC voltage and K v is the proportional gain. The initial gains of the compensator

are determined using the Ziegler—Nichols Oscillation method. Further tuning is achieved

66

by performing multiple simulations of several scenarios. Best performance is observed

when the Kp=0.3 and Ki=0.1 values. Duty cycle, D, is computed and then compared to a

10kHz triangular wave to generate a lOkHz PWM switching signal with positive duty

cycle of D. Based on the PWM switching pattern, IPM switches are turned on and off in a

bipolar fashion, such that two diagonally opposite switches in the H-bridge turn on and

 

 

 

 

 

 

 

 

 

 

     

 

   
   

  

off simultaneously.
DC-AC INVERTER
Vch 400V
J- '1‘ , l:
T i *1 03H szJ u
j i ’znr‘ i ,4
Vch
1“ 1 Cf LOAD
Cd 2 i: l T
C -= :1 )fi
11 *4 .4S4J S3J
1
Vac
S1 53 m \J G {1' Vac*
S2S4:-' fl 1 a
Figure 45 Inverter voltage control
4.2.3 Current Control

In the utility-connected mode, the inverter operates as a current—controlled source to

generate an output current based on a reference current signal. The inverter shall not

actively regulate the voltage at the PCC (point of common coupling). So, in the grid-

connected mode, the inverter does not control the output ac voltage but control the output

current (or output power). And at the same time, DC-DC converter should also keep the

dc-link voltage constant. When the inverter is utility connected mode, I ref (Iac*)

becomes the controlling variable.

67

DC-AC INVERTER

J. I
1 1
Cdc2 ’1—\ ii . f/ ;.‘1 SIJ SZJ .
i 1 M AC Grrd
I ‘ ‘ INN % /\

 

 

V

 

:
1

Vch :[C‘

 

 

 

 

 

 

 

 

Figure 46 Grid connected current control

The block diagram of the current control loop is shown in Figure 46. An external
voltage signal with 240V AC, 60Hz is first fed into a discrete single-phase phase-locked-
loop (PLL). The PLL is necessary to make sure that when utility voltage is present, the
inverter voltage is synchronized with the utility voltage. Discrete PLL is used to find the
phase angle of the utility voltage. Figure 47 shows the PLL structure we used for
synchronize grid current to grid voltage. First we generate 60Hz sine signal and make
product with real grid voltage. Once it go pass by digital low pass filter, average goes to
zero in few cycle. This means Vgrid and Vsyn voltage getting close enough to get
synchronized. Then next step is that different value is amplified and sum up with
fundamental radian velocity. Integration of radian velocity is phase angle. This angle
makes feedback voltage and continuing this same process until it get close with grid

voltage phase.

68

lwf V

1 SYN
- KAwl 71-»
n 01 -
Vf ' S.

Figure 47 Block diagram of product type PLL

Vgrid

 

Figure 48 PLL simulation result
Figure 48 shows the simulation result of PLL. Vsyn and Vgrid waveform become same

phase in a few cycle. Figure 49 shows the inverter output voltage and grid voltage
realizing product type PLL control. Inverter output voltage become same magnitude and

phase with grid voltage after operating PLL control.

69

 

 

I PLL Signal (5V/div) l

Inverter output _ t , -
Voltage (2 50V /div)G"d Voltage (250V/d1v)

 

 

 

Figure 49 PLL experimental result

The user defined rrns current amplitude (1 ac*) and phase angle (9) are then used to

generate the reference current waveform. This sine wave is compared to [ac and the

controlling output I_command is generated. I_command current then sum up with

*
V acand ij - I ac in order to make desired current. To ensure stability for high-duty
cycle conditions, a stabilizing saw tooth wave is added to the phase adjusted sine wave.
The final output of the current controller is a duty cycle value (D), and is given by

equation

 

D: 0.5+ Vac +Kix([ac*_1ac)Xfosw
2dec 2><Vdc

(4.11)

where, Vac is the inverter output voltage, Vet: is the DC link voltage, I ac* is the
reference current, I acis the inverter output current, Ki is the proportional gain, L is the

gain and is set to a filter inductance of 2[mH] and fswis the gain equal to a switching

70

frequency of 10 kHz. As with the voltage control loop, this duty value is then compared
to a lOkHz triangular wave to generate a lOkHz PWM switching signal with positive

duty cycle of D.

4.3 Test Setup and Software Implementation

4.3.1 Test Setup

The power circuit for the complete inverter system is shown in Figure 50.

DC-AC
DC-DC CONVERTER INVERTER LC FILTER

I V0211 400V

AC GRID

 

 

 

 

 

 

Figure 50 Powercircuit canverter system connected with grid

The DC input stage of the inverter consists of a 400V DC source along with a small
series resistance of 0.49. This series resistance is required in order to facilitate
convergence of the model in the simulation environment. The power electronics devices
are modeled based on the IPM module parameters used in the hardware setup. The IPM
module consists of six IGBT switches along with 1500uF of input capacitor. The output
of the inverter is fed to the inductor-capacitor (LC) line filter. The series-connected, 2mH
inductors and 7.5uF capacitors are used in order to achieve desired voltage and current

ripple characteristics. The output of the inverter is a single-phase, 60 Hz, 240V AC ms

71

waveform. . Though an optional DSP controller can also be included inside the module,
but in the laboratory setup an external DSP was selected for this particular design to have
more flexibility. Various low DC voltages are required to operate different parts of the
hardware. The PV system module requires 15V DC, the DSP board requires 5V DC, and
additional 48V DC is required for all the electronic integrated circuits (ICs). A small DC
power supply is used for generating 48V DC from 120V AC utility. The 5V DC is
developed from the additional power supply PCB board. A precision micro-power, low
dropout voltage reference IC is used for generating the voltage references for analog-to-
digital (A/D) converters in the DSP. The phase current sensors, DC bus voltage sensor,
and heat-sink temperature sensor are built on the PCB board. In addition to those sensors,
battery voltage and converter input current sensors are added to different parts of the
hardware to achieve control and protection requirements. The test setup includes the
inverter; a 240V, 60Hz AC source that emulates the utility grid; and one sets of
resistances representing the local load. The inverter and the local loads are connected to
each other through single-phase breakers. When breakers are closed; the inverter operates
in the utility-connected mode. In that scenario, both the inverter and utility provide power
to the local load. To test islanded operation of the inverter, the utility breaker is Opened so
that the inverter supplies power to the local load.

The maximum voltage across the switches in the grid-tied VSC is 400V. The peak
current through the switch occurs at maximum ac load power of 2 kW and at minimum ac
voltage of 220 V. The peak current through the switches is 13A.

As mentioned in previous section, a 600-V/75-A (IGBT-IPM) is selected.

In addition to the inverter platform, other components such as DC voltage source,

72

utility grid, RL load bank, and electronic load bank were connected in the test setup. The
DC supply for the inverter was from output of isolated DC-DC converter from previous
chapter. An ac power source is directly connected to 240[V] local grid. 1kW RL load
bank, mimicking the local load, was placed in parallel to the grid. The load terminal of
the inverter was connected to a 1kW/1Kva AC Load. A 1:1 isolation transformer was
placed between the inverter output and the grid. The single line diagram of the complete

test setup is shown in Figure 51.

 

 

 

 

CB 1:1 TR AC GRID

DC power Inverter and Line FCC @%
supply control Filter

 

 

 

 

 

 

 

 

 

 

 

 

 

Local Load
R: IkW,R-L:IkW/IKVA

 

 

 

 

 

 

 

 

Figure 51 Single line diagram of test setup

4.3.2 Software Development

Up to this point, the voltage and current control for the inverters are implemented in
the DSP controller. The program for the DSP is written using Texas Instruments Code
Composer Studio which offers robust, mature core functions with easy-to-use
configuration and graphical visualization tools for faster system design. The codes are
written based in C programming language and then compiled and load into the DSP on-
chip random access memory (RAM) using Code Composer Studio. Then the controller is
activated using the serial interface. In future, the on—chip flash memory will be used for
utilization of larger codes and retaining the compiled codes in the DSP. The program code

is interrupt driven and runs at fixed sampling frequency of 10 kHz based on the interrupt

73

set up. For the inverter platform, the switching frequency is set at 10 kHz. The actual
program codes for PWM calculations are written inside the interrupt service routine. This
service routine runs once in each PWM cycle and is used to compute duty cycles for the
next cycle. The controller was designed so that the control functions are interrupt driven.
After completing A/D conversion (ADC), an interrupt is generated and the program goes
to the interrupt service routine (ADC-PWM subsystem block).The ADC start occurs
based on the period interrupt of the PWM. Therefore, the interrupt service routine runs
once in each PWM cycle and is used to compute duty cycles for the next cycle. For the
experimental evaluation, the PWM operates at 10 kHz, therefore the interrupt routine is
called at every lOOus. A new fixed point PLL was developed which accepts a sinusoidal
signal and outputs the phase angle for the input signal. The phase angle output is a ramp
signal varying between 0 to 21; and is synchronized on the zero-crossing (rising) of the
input signal. The grid voltage signal is fed from the ADC-PWM subsystem block to the
PLL and the angle output from PLL is fed back to the same subsystem block. As the
calculations inside the ADC-PWM subsystem block are done asynchronously with a

sample time of 100us defined by the hardware interrupt setting, several rate transition
blocks are used in the model. Also the amplitude (I (10*) and the phase (0) of the current

reference are fed into the ADC-PWM subsystem block.

The ADC-PWM subsystem block is a triggered subsystem and activated when the
hardware interrupt occurs. Inside this subsystem, DSP’s ADC channels are configured to
perform analog-to-digital conversion of input signals. Both the ADC module A and B are
selected and the simultaneous conversion mode is used. The ADC start occurs based on

the period interrupt of the PWM using Event Manager A and an interrupt is posted at the

74

end of the A/D conversion. The output of the ADC goes into the scaling block where the
digital signals are converted to their corresponding analog values. These scaled voltage
and current measurements along with the PLL-calculated phase angle are then used for
the voltage and current controller. The output of the controller is the percentage duty
cycle value which is fed to the PWM block to generate lOkHz SPWM switching signals.
When the operational mode is selected as islanded voltage control, the voltage control
subsystem is activated. In the voltage control subsystem, the reference voltage signal is
generated first using the phase angle (Theta) information. An offset is added to the PLL-
calculated phase angle to take care of the phase shifis produced by sensors and control
board filters. The amplitude of the reference voltage is set to 240V AC rrns. A
proportional—integral controller (PI controller) was used in the voltage loop. The final
output from the voltage control subsystem is the percentage duty cycle and a saturation
block was used to limit its output between 5% and 95%. For the operational mode of
utility-connected current control, circuit breaker was closed and the current control was
activated. Again, frequency and phase of the utility voltage were obtained using the PLL
which was then used along with the user-defined phase angle and current amplitude to
calculate reference current waveform. The PI controller was finally used to generate

percentage duty cycle outputs.

75

4.4 Simulation and Experimental Results

4.4.1 Simulation Results for Voltage and Current Control

The inverter model is simulated in Simulink PLECS to allow researchers to observe
both voltage and current control. In the simulation, the inverter is turned on at time=0.053
in current control mode. The rms value of the reference current is set to 8A and the phase
angle is set to zero degrees. In the current control mode, breakers are closed and the
inverter operates in utility-connected mode. At 0.55, the utility breaker is opened, thus
transforming the inverter control to voltage control to maintain 120V, 60Hz AC voltage
to the local load. Finally the inverter is turned off at 0.955. During the simulation, the
local load is set as 100 resistive and the utility load is set at 6 Q resistive. The simulation
results for current control mode are shown in Figure 53-56. The waveforms for the
inverter voltage (V inverter), inverter current (I inverter), utility voltage (V s), and utility
current (Is) are given in Figure 52,(1)-(8). We can observe from these figures that the
waveforms are very clean with only some higher order current harmonics present in the
inverter current waveform. During utility connected mode, the inverter supplies current
equal to the reference current. The local load requires 12A current, of which 8A are
supplied by the inverter and the remaining 4A are delivered from the utility through

point-of—common-coupling (PCC). The PCC voltage and currents are denoted by Vg

and Ig respectively.

76

_.
C)

    

a:

4

Vinverter(V)

M

 

I inverter rms(A)

0 0.05 0.1 0.15 0.2 0.25

0 0.05 0.1 0.15 0.2 0.25

I inverter(A)
Is rms(A)

 

0 0.05 0.1 0.15 0.2 0.25

          

   

3
V)
E
in
NJ
N
_1
40
20 8
A m
1
20 3
1 2 1 1
.40
0 005 01 0,15 02 0,25 " 0 0.05 0.1 0.15 0.2 0.25
(4) (8)

Figure 52 Simulation results for current control mode of operation

Before the inverter is turned on, breakers are opened, and the utility supplies 20A rrns

77

to the utility load. When the inverter is operating, the utility supplies its own load along
with the residual current required by the local load. The rms calculations for the currents
are shown in Figure 52,(5)-(8). The simulation suggests that the inverter current control
loop does a very good job of maintaining an 8A rms current when operated in the utility

connected mode. Figure 53 Shows the grid-connected mode single diagram of simulation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8A 4A 24A
g g
*- N
' Local Utility _
‘ inverter Load Load % Grid
m
__ —l— _l_ ?

Figure 53 Single diagram of simulation circuit grid-connected mode

When the utility breaker is opened, the inverter switches to islanded operation to
supply power to the local load. The simulation shows that the utility keeps supplying the
utility load but there is no power flowing through the PCC. In the islanded mode of
operation, the inverter is voltage controlled to maintain 120V AC, 60Hz voltage for the
local load. The voltage and current waveforms in this mode of operation are shown in
Figure 54,(1)-(6). From Figure 54-(2), we see that the inverter is maintaining 120V AC
ms and supplying current to the local load. The utility voltage and current waveforms are
shown in Figure 54,(5)-(6) and it is evident that the utility is supplying power to the

utility load.

78

     

    

A
A 200 a
5:100 E
g L:
1; ° E
.E 400 g
> .5
"”06 >
A 20 2
‘3 10 1
g E
l- 0 h
1 i
.5 .10 o 5
_ >
.E g
'2'st 0.55 0.7 075 00 035 1-1 .6 0.55 0.7 0.75 0.0 0.05
(4)
40
20
A
$1 0
U)
H
.20
'“05 0.55 0.7 0.75 0.8 0.85 "'05 0.55 0.7 0.75 0.0 0.05
(5) (6

Figure 54 Simulation results for voltage control mode of operation

Finally, in Figure 55 the voltage and current waveforms from the complete simulation
sequence of operation. Figure 56 shows diagram of transition of two control mode. From
these waveforms, we see that the simulated transients during the change in mode of
operation from current to voltage control are small. The local load does not experience
any variation in voltage Figure 55-(5), when the inverter is disconnected from the utility.
Also from Figure 55-(4), it is evident that the power is transferring through the PCC only
when the inverter is utility connected. From the inverter current waveform, as in Figure
55-(2), we see that the transition from utility-connected to islanded mode is smooth. In
the utility-connected mode, the inverter is supplying 8A rrns current which is increased to

12A rrns when the inverter is in islanded mode supplying the local load.

79

I inverter(A) V inverter(V)

 

 

  
 
        

0 0.1 0.2 0.3 0.4 0,5 0.6 0.7 0.8 0.9 1
2m (2)
E 101
=0 D
> it!)
21D
0 0.5
(3) G' .
10 nd disconnected
Q 5
33 0
"‘ -5
-10
0 0.1 0.2 0.3 0.4 40.5 0.6 0.7 0.8 0.9 1
E m ( l
.5 111]
g 0
1.1 411)
> .m
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 t
A (5)
S.
'1:
a
O
1—1
0.4 0,5 0.6 0.7 0.8 [1.9 1
(6)
Time(s)

Figure 55 Simulation Waveforms for Complete Inverter Operation
It is important to note that, in Figures 52 to 55, only some representative results are
shown based on particular load settings and set points. The inverter model was tested
with a variety of different types of loads and different voltage and current reference
settings to verify its operation under various scenarios.

The inverter platform was initially evaluated using both islanded-mode voltage

80

control and utility-connected-mode current control. Based on successful operation of the

inverter in the voltage control mode, current control mode tested experimentally.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Iinverter 8A Igrid 4A Is 24A Iinv§“?.-1%
’Tl‘ 7'} ‘ 7 e H ‘ 7 7 ' <!
V' rt Local Ut '1 ' Grid Vinverter Local
inlée0v er ray 120v

Figure 56 Single diagram figure of current and voltage control: transition of control mode

Experimental Setup

For the voltage control tests, the DC bus voltage was set at 400V DC and the local
load resistance was set at 609 resistive which consrunes le at 240V. When the utility
circuit breaker is left open, the inverter supplies the local load and the utility supplies the
utility load. The local load resistance was varied from no load to fill] load. We observed
that the inverter output voltage operates close to 240V AC rrns for both the no load and
full load cases. In this test the DC bus voltage was set at 400V DC, the local load was set
with 600 and 0.2[H] L load was added. The amplitude of the reference current was set at

8.3A rrns and the phase was set unity power factor as we seeing from inverter side.

4.4.2 Experimental Results for the Stand-Alone Mode Operation
Series connected PV system can be grouped into two categories from the relationship
between the inverter output and utility grid. The first category is the situation where the
inverter is not connected with utility grid, which is named as “stand-alone”, as shown in
Figure 57. In this situation, because the output of PV and battery is limited, only part of
loads (critical loads) is admitted to connect. An MPPT algorithm is executed for PV to

obtain maximum power to meet the demand of the critical ac loads. If the produced

81

obtain maximum power to meet the demand of the critical ac loads. If the produced
electric power from PV is less or more than the load need, the battery is used to
compensate the difference. That is, the actual charging (or discharging) current depends
on the PV power production and the demand of onsite critical loads. In this case, the
isolated dc-dc converter shall use the constant dc link control strategy and the inverter
shall use the voltage/current control strategy. Because the buck converter should be
utilized to realize MPPT for PV if both the battery current and voltage do not exceed the
safe limit, it shall carry out the control shown in previous section. Otherwise, the buck
converter will carry out battery current or voltage limit control. It should be noticed that,
under stand-alone mode, when the battery is discharged so deeply that its voltage reaches

a minimum threshold, the load (or converter) shall be automatically disconnected (or shut

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

down).
va __ Pdc Pch
W: “ ‘ —* “r 51
PV _ — ‘5}
BUCK/BOOST I DC_load
BATTERY
P ac I I
8 I. a
AC_load
INVERTER DC-DC CONVERTER

Figure 57 First category operation mode: stand-alone
In islanded operation, the inverter has to maintain the amplitude and the frequency of
the voltage so that the connected local loads are not affected by utility interruption. The
voltage control loop for the inverter is designed for such islanded mode of operation.

When the utility voltage is present, software zero-crossing detection is used to obtain the

82

frequency and phase of the voltage reference, whereas the frequency is set to 60Hz when
there is no utility. We have found that the fixed amplitude of the reference voltage causes
bigger steady-state errors depending on the change in local load. To overcome this
problem, a slower control loop (running at 1/60 s) was designed that changes the
amplitude of the reference voltage depending on the output current. The final output of
the controller is a duty cycle value (D). The DC input voltage was set at 400V DC. The

screenshot of experimental results for the voltage control is given in Figure 58.

 

 

 

 

DC link voltage (l00V/div)

Inverter output
voltage (250V/div)

 

 

 

 

nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn

n u . - - u . . e

c - - r o n .

o a a o n a n n

‘ n . o u o n n

. n . . . 5 a . .

a o c o a n u n a

u o u o a u u - c
-------------------------------------------------------------------------------------------------

I u v I I - n I u

v . r u . u . n o

o n . n n n

n r - v .

u u n c n

n u u n o

AC Output current (10A/div)

 

 

 

 

 

Figure 58 Experimental results of inverter.(Stand alone mode.)

The local load is set at 600 at this time. The inverter is islanded from the utility and
no connection exists between the inverter and the utility as the circuit breaker is kept
open. We can see from the waveforms that the harmonics present in the inverter voltage

are low and the inverter voltage is in phase with the inverter current.

83

Stand—alone mode R-L load test.

Buck-Boost converter

L]
on"

48Vdc

 

m
Cch

 

 

 

1

 

 

 

 

 

DC-DC conve

 

 

 

 

 

 

rter

Vbattery
DC-AC AC
Load

converter L-C Filter

 

'_l—I .
W

WE

Cdc2:

 

       
 

11:11,}:

_L
Cf]:

I
17

Figure 59 Stand-alone mode power flow

To verify variety of load condition, we set the R-L load condition in stand-alone mode.

First we set the L with 0.2[H] using transformer to get high reactance at the primary

side of the L load. Picture 59 shows 609 Resistor is parallel with L load at the ac

output side. Since output voltage is 240V, we can calculate estimate real power and

reactive power of the load. Real power consumed by resistor is 960[W] and reactive

power at the L is near 763[VAR]. From knowing the P,Q value we can derive apparent

power which is l.22[K VA] with 38°lagging. This setup is appropriate for replacing

motor load in real situation. The theoretical power factor was 0.79.

Figure 60 and Figure 61 shows the waveform of inverter output voltage and output

current at the ac side. It shows almost similar angle between voltage and current as we

calculate before. Since apparent power is 1.22[K VA], the current is near 5.1[A]at the

84

load. It is obvious that with low power factor, inverter have to generate higher current

than at the power factor 1.

 

 

‘ DC link voltage (100V/div) L Inverter output voltage (100V/div) l

1
/

    
   
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L

Figure 60 L-R Load waveform at PF 0.73

Actual power factor value was 43° due to line impedence effect. From measuring the
time delay between voltage and current shown in Figure 61, we can derive power factor
of inverter. 1.99msec is phase angel delay time between voltage and current while

16.66msec is the time for 60Hz.

 

 

 

J Inverter output voltage (100V/div) I
. : I

   
   

 

 

 

 

 

Inverter output
current (SA/div)

 

 

 

 

 

 

 

 

Figure 61 L-R Load waveform at PF 0.73

The inverter in the voltage controlled mode was also run for difl‘ere

85

maximum load for the inverter is limited to 15A rms by the measurement and protection
circuits. Therefore the local loads were adjusted in such a way that the maximum inverter

output current did not exceed 12A rms.

 

‘ 100
i 95 .
1 ’3‘
,I 9:, 90 . M
i k.
I 2.
! u -
1 :5 +efficuency
I m 30 .
7s
70

0 500 500 2000

1000 , 1
Pout(\\ )

Figure 62 Efficiencies of the Voltage-Controlled Inverter
Finally the input-output efficiencies of the inverter are shown in Figure 62 over
varying load conditions. It is evident from the figure that the lowest efficiency of 86%
occurs at the light load condition and, as loading increases, efficiency also increases. The
range of efficiencies are good considering the fact that the inverter is operating at much
lower power levels than the full load of the IGBTs, which are rated for 75A continuous
current and 600V AC rms. Also the inverter platform is built to have some flexibility so

that different distributed energy sources can be connected to it in future.

86

4.4.3 Experimental Results for the Utility-Connected Operation

The second category in Figure 63 is the situation where the gateway grid is connected
to utility grid. Part of the PV produced power is transferred to the battery and dc loads,
and the excess power is transferred to utility grid and ac loads (termed as forward grid-
connected). On the other hand, the whole system will draw real power from utility grid if
the PV produced power is not enough to supply dc loads and the battery (termed as
reverse grid-connected). In this case, the isolated dc-dc converter shall use the constant
dc link control strategy and the inverter shall use the current control strategy. The buck

converter should realize MPPT control and the control strategy shall apply.

va . ' 3 Pdc Pch
—> t _ _ 5 <—> ——-> g R“:
7 . 1 , ,,,, ‘,7 _{ ‘11
1 :r t i 1 \_
PV . , 2 1 ' ,
BUCK/BOOST “E' ‘ DC_load
. , . "7 BATTERY ‘
| i w . | i
‘8 .; H 3g”; ~
I ' .II . ‘ ‘
Pac ‘ i . a L i
Qac INVERTER DC-DC CONVERTER

.g, e e .9.
AC_load
Figure 63 Second category operation mode: grid-connected
When the inverter is connected to the utility, the grid controls the amplitude and
frequency of the inverter output voltage, and the inverter itself operates in the current

control mode. The currents required by the local load is shared by the inverter and the

utility, where the inverter provides the amount of current based on the reference current.

87

In the current controller, software zero-crossing detection was used to obtain the
frequency and phase of the utility voltage, which was then used along with the user-
defined phase angle and current amplitude to calculate the reference current waveform.
After obtaining the reference current waveform the current controller was designed. The
final output of the current controller is a duty cycle value (D). The PWM generation is
the same as described in the voltage controller section. The DC input voltage was set at
400V DC. The current control loop was tested with different amplitude and phase

reference values.

Test setup for grid connected mode

The test parameter for grid connected system is shown in table 1. DC load of buck
converter is 100 which consumes 250W. DC source supply dc power to load and
surplus power is send to ac load or grid. Leakage inductance for the system is set up with

11 uH which could realize more stable power control with this value.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Component Value P/Q/PF
Rdc 100 250W
C dcz 3000uF each
Czdofl_fiequency 1.3KHz
Rae 609 960W (actual:l.3Kw)
L a c 0.2[H]
X Lac 75.399 763VAR (actual:1KVA)
P0vver_faetor PF ___ 360x 1.99 = 430 Cos43° = 0.731
of inverter srde 1666
AC_ grid _voltage 240V

 

Table 1 Test setup parameters

88

 

Situation #1 (Resistive load experiment)

Buck-Boost converter

 

PV 48~96lVl ‘ r“"1 u 48Vdc
va 3: SL JE Cch == __ Vbattery
va

 

 

 

 

 

 

 

 

DC-AC AC
DC-DC converter converter L-C Filter Load

JFVJFJ‘FU +-

Cdc2’
h—éllL Vch W7}: '- <9

V

 

 

 

I_—
TIT

 

 

 

 

 

 

 

 

J J: Cch N J I J J

 

 

 

 

Figure 64 Grid-connected mode power flow before turning on the inverter

To test the current controlled mode of operation, it is necessary to make sure that the
current control loop is working before actually making utility connection. In this test
setting, the inverter output terminals are connected with grid where grid resistance is very
small. First, the reference current is set to zero, then the system is turned on. After system
become stable, the reference current value is increased to a higher value.

Figure 64 shows that DC source only provides power to dc load and battery since
inverter is turned off. This situation is possible when PV power and battery energy is
insufficient or not available during the night time. Before turning on the inverter, only
grid can supply power to ac load as shown in Figure 65. Both PV power and battery
supply power to the dc load when PV don’t have enough power. It can be seen from the

figure that, the inverter current is nearly zero.

89

 

 

LGrid current (5A/div) 1

 

\\ ’/-t\ . "/—\\‘ _/..

 

 

 

 

I
[ Grid voltage (250V/div) I

lei-1'50»; 3:- L."l’ia; . 3 {pt-‘1 " ” 213‘3-2-‘L I?

- .4 .- . .
., .. I . .... . c; > .- u . 3»; , ....,, e. ....,.
‘ .. .ou’htd 1 "way ;5 “,3; ~ . .' s. ,. . "‘gwtfia‘JQwh-"F‘

 

T
l inverter current (SA/div) I

 

 

 

 

 

Figure 65 Inverter output voltage and current

Situation #2 (Resistive load experiment)

Buck-Boost converter
PV 48~96[V] [:3] LI 48Vdc

 

'fi

va Z: SLJE Cdcl == __ Vbattery
va

 

 

 

 

 

 

 

 

DC-AC ‘ AC
l DC-DC converter converter L-C Filter grid
J . J J
Lf I

Figure 66 Grid-connected mode power flow after turning on the inverter

 

S

 

 

 

 

 

In this situation, only inverter provide power to the load which means ac load power
is same as PV power. Figure 66 shows grid-connected mode power flow after turning on
the inverter. When system is connected to the grid with IKW active power demanded by

the ac load, system is operated with only the PV panel connected in the system and the

90

inverter is used to feed the maximum available power to the ac load. The power
transmitted into grid was 0 W since PV power is limited to lKW. DC link voltage is 400V,
line to line ac voltage is 240V and output ac current is 4A. Output power is around lKW
at 609 resistive load. It can be seen from the Figure 67 that, the grid current is nearly
zero. Only inverter provides the current to a resistive load which is about lKW and also

indicating that the inverter side power factor is equal to 1.

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

. . 5 3 5 DC link voltage (100V/div)
“W h ..‘t:‘#“ ..... ........
E 3 : Inverter
........ voltage (250V/dIV)
.. 2 : .
Ff fiwgfi—fi i #mfg : 3 fifiq
Q: ...... “t?— ............ fl ............. 1“ ...... My; ......... 2m’ui.. ......... f .....
at s -. a... s s ' ..a
...... InvertercurrentlloA/div)
: Myth“ fiiflhiiflw

 

 

 

Grid current (lOA/div)

 

 

 

 

H

 

Figure 67 Experimental results of inverter.(Grid connected mode-1 KW)

91

Situation #3 (Resistive load experiment)

Buck-Boost converter
PV 48~96[V] '32] L1 48Vdc

va 3F SLIEI} Cdcl == 17L- Vbattery
va

 

 

 

 

 

 

 

 

 

    

 

 

j
DC-AC AC
I DC-DC converter converter L-C Filter grid
Cdc2 i: KI} ")3
Cf
; a V .1 Vé

 

 

 

 

 

 

 

 

 

 

Figure 68 Inverter output voltage, current and grid current
When PV can transfer lKW as well as battery(lKW), whole system can supply 2KW

power to both ac load and grid as shown in Figure 68

 

 

I DC link voltage (100V/div) I
v

_. ..mw - --- ___. l
I

 

 

Inverter
voltage (250V/div)

\ r Inverter
current (10A/div)

 

 

 

 

 

 

 

 

. w”

I Grid current (10A/div) J

 

 

 

 

 

D
Figure 69 Experimental results of inverter.(Grid connected mode-2KW)

92

This situation can be possible when PV and battery power have big enough power to
supply all loads connected to system. During the peak sunlight in the day time, system
can send surplus power to grid for cost reducing purpose.From the experimental result,
lKW is consumed by the ac load and the remaining lKW is fed into the grid. Figure 69
shows the waveform that inverter output current is 8A and grid current is 4A. The phase
of both currents shows same direction which means that inverter is supplying power to
grid.

Situation #4 (only R load experiment)

Buck-Boost converter
PV 4M6[V] L1 48Vdc

  

 

 

 

 

 

 

 

 

va 3: S1,]? Cdcl I: 1‘:- Vbattery
va
* _
DC-AC AC

DC-DC converter converter L-C Filter grid

" I. Ill-cell'm'l—M -Cd|-c2—I—J“I—T’ *1: ”viii

EIIL W" “i .[M

 

 

 

 

 

 

T V

 

JE JE Cch N .l . J J

 

 

 

 

 

 

 

F igure 70 fl Inverter output voltage, current and grid current.

In this situation, inverter and the grid provide the power to the ac load as shown in
Figure 70. Both the PV system and grid are operated in parallel. When PV can not supply
enough power to ac load during the night time or when the battery has no energy to
discharge, grid should supply its power to ac load with inverter side power. Figure 71

shows the inverter output voltage and current waveforms and the grid current. After

93

reducing power of PV side to 500W, both inverter and grid is supplying active power to
lKW ac load. From the curve, we could observe that each current curve has 180 phase
difference representing that both current are flowing opposite direction in order to supply

power to ac load. Above experiment was based on Resistive load condition.

 

DC link voltage (100V/div)

.
. . . v .
L ALA. u A A a A A L ALLA M A A. A MAL-‘.. A “MMA AAAA _...__._
vfi— "—'— * "'vv'vv'vvvvvw‘v V‘V‘fvv‘wi w vvvv‘r ‘v v v m y 1 v
I I I t l .
. . n o . l a . o
. r . . . o . . q
. . v 9 -
. . . . .
. . - . e -

_ , Inverter
voltage (250V/div)

 

 

 

 

 

 

  
 
 
 

 

 

 

oooooo

 

 

.I:.. Inverter
; current (l0A/div)

 

 

 

 

........

 

 

 

Grid
current (10A/div)

 

 

 

 

 

Figure 71 Inverter output voltage, current and grid current

Situation #5 (Inductive load experiment)

Installing inductive load to see verity of load condition of system. Since inverter is
working with unity power factor control, grid power factor become lower. After turning
on the inverter, grid supply power to ac load and PV supply power to DC load(250W).
Inverter supply real power (520W) to ac load and at the same time grid supply real
power(780W) to ac load which consumes near 1.3 [KW]. Figure72 waveform shows DC

load current, voltage, AC load voltage and current waveform. Seeing from the inverter

94

side, we could observe that power factor is near 1 by using unity power factor control. L

load was about 0.15[H] in real experiment situation. Reactive power consumed by
inductor is near lKVA. Different phase angles is shown between grid voltage and glid
current due to inductive load. Figure 73 shows that grid current is 52 degree lagging to
voltage which means grid is supplying real power(780W) to the ac load as well as

reactive power( lKVA).

 

 

 

I Grid voltage (100V/div) I i , . . . Inverter output

\1 current (SA/div)

 

 

 

 

1F

 

5| .. I , . ‘.. . A. .
DC load voltage (25V/div)

 

 

 

 

 

 

I DC load current(SA/div) I
1

J

 

 

Figure 72 Ac load voltage, current & Dc load voltage and current

We could verify that inverter controlled with unity power factor with lack of power to
supply ac load, grid always supply both real power and reactive power to the R-L load. In
this experiment, since ac load consumes 1.3KW, both inverter and grid supplying real
power. In addition, grid supplying lKVA reactive power. As a result, grid shows 0.6]

power factor as shown in Figure 73.

95

 

 

Grid voltage

 

 

 

 

 

 

 

 

. ’; (100V/div)
i .

Grid current
(SA/div)

 

 

 

 

 

Inverter output
current (SA/div)

 

 

 

Figure 73 Inverter output voltage, current and grid current(inductive ac load test)

 

 

 

 

 

 

 

 

Inverter Grid
KW 520w - 780w
K VA 0 1 K VA
PF 1 -0. 61 pf
Degree 0° 52°
Table 2 Inverter and grid power factor

96

 

Chapter 5

Conclusion

A new low cost series connected PV system for Stand-alone/Grid-connected
applications has been presented in this paper. The operation, analysis, features, and
control scheme were explained. Simulation and experimental results for the 1kW
prototype are shown to verify the operation and control principle. A buck converter is
used for the MPPT implementation and presents the functions of battery charger and step-
down converter. The dc-dc converter realize phase shifting to control power flow through
a transformer with a MOSFET full bridge on the LVS.(low voltage side) In addition, a
voltage doubler on the HVS(high voltage side) is installed to achieve enough high
voltage to run the inverter. The feature of the prototype includes the combination of dc-dc,
dc-ac and/or ac-dc converters for the flexible and uninterruptible energy utilization and a
smart algorithm for optimizing the power consumption and defining priority on the
demand system. For energy sources such as PV or fuel cell that generate DC, an inverter
is required to convert DC into AC for utility connection and consumer use. Though low
power single-phase inverters are very common and available from different
manufacturers, these inverters mainly work as a controlled current source when they are
operating in the utility-connected mode. This thesis outlines the detailed modeling, design,
and testing of a single phase PV inverter system that operates in both islanded mode and
utility-connected mode. Both modeling and experimental results are given to demonstrate
the ability of the developed inverter to provide advanced control functions. We used the

Simulink PLECS simulation to develop the single-phase inverter model and the control

97

algorithms before they were implemented in hardware. The parameters and simulation
conditions for the model were carefully selected to accurately emulate the hardware. Also,
the control designs have been developed in such a way that they could be implemented
on a hardware DSP platform. In addition to the inverter power and control circuits, the
laboratory test setup consisted of several other components, such as a DC voltage source,
a utility grid and a RL load bank. Two techniques were demonstrated for DSP code
generation. Initially the DSP code was written in C- language using Code Composer
Studio and experimental results proved that the inverter successfully operates for all
functional modes, namely the islanded voltage control mode, and utility-connected
current control. Future work will concentrate on developing and testing additional
ancillary functions such as power quality control and seamless transfer for the inverter

associated with distribute energy sources.

98

Future work

First, the increase interest in novel architectures based on the adoption of switching
converters employing a PV module-dedicated decentralized MPPT fimction is giving rise
to new challenging problems. The distributed-MPPT method, which is also stimulating
some industries to the production of DMPPT dedicated devices, allows for reducing the
impact of the mismatching effect, but its implementation requires further studies in terms
of interactions among different systems, employing the MPPT function at the same time.

Second, Sic semiconductors have a high potential for enhancements of the electrical
conversion efficiency of PV systems. In addition, high switching frequencies inherent to
such devices will also enable decrease of weight and cost. The most significant
characteristics of Sic is the very high electric breakdown field, resulting in very low
specific on-state resistance even at higher blocking voltages, as can be observed below
for different switch technologies. This will allow the use of simplified circuits with fewer
semiconductors and power state as well as allow a considerable diminution of the heat-
sink size and cooling effort, again reducing the weight and cost of the converter.

Third, multilevel PWM converters are gaining popularity in the field of renewable
energies because of the need of connecting several distributed power sources, whose
power is continuously growing, guaranteeing at the same time good power-quality levels.
They can be used as rectifiers in the case of wind and hydroenergy, where the electricity
is generated by ac generator and inverters in the case of wind, solar, hydro and fuel cell
generation. Multilevel active rectifiers improve the absorbed ac currents, thus reducing
generator stress; at the same time, they allow precise output voltage regulation, thus

eliminating the need for a dc/dc convert

99

BIBLIOGRAPHY

[1] N. Femia, G. Petron, G. Spagnuolo, and M. Vitelli, “Optimization of perturb and
observe maximum power point tracking method,” IEEE Trans. Power Electron, vol. 20,
no. 4, pp. 963—973, Jul. 2005.

[2] M. G. Simoes and N. N. Franceschetti, “Fuzzy optimization based control of a solar
array system,” Proc. Inst. Electr. Eng.Electric Power Applications, vol. 146, no. 5, pp.
552—558, Sep. 1999.

[3] J. A. Abu-Qahouq, H. Mao, H. J. Al-Atrash, and I. Batarseh, “Maximum efficiency
point tracking (MEPT) method and digital dead time control implementation,” IEEE
Trans. Power Electron, vol. 21, no. 5, pp. 1273—1281, Sep. 2006.

[4] N.Mutoh, M. Ohno, and T. Inoue, “A method for MPPT control while searching for
parameters corresponding to weather conditions for PV generation systems,” IEEE Trans.
Ind. Electron, vol. 53, no. 4, pp. 1055—1065, Aug. 2006.

[5] J. H. R. Enslin and D. B. Snyman, “Combined low-cost, highefficient inverter, peak
power tracker and regulator for PV applications,” IEEE Trans. Power Electron, vol. 6, no.
1, pp.73—82, Jan. 1991.

[6] J in Wang, F.Z. Peng, Joel Anderson, Alan Joseph, and Ryan Buffenbarger “Low Cost
Fuel Cell Converter System For Residential Power Generation,” IEEE Trans. Power
Electronics, Volume: 19, Issue: 5, pp: 1315-1322, Sep. 2004

[7] K.Y. Khouzam, Optimum load matching in direct-coupled photovoltaic power
systems-application to resistive loads, IEEE Transactions on Energy Conversion, vol. 5,
no. 2, June 1990, pp.265 — 271.

100